JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY C. E. WOOD, JR. EDITOR L. I. NEVLING, JR. LAZELLA SCHWARTEN ASST. EDITOR CIRCULATION VOLUME XLII CAMBRIDGE, MASS. 1961 Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION ew York 1968 Ss 4k 700 DEC 121968 - } / / - j, { A, FA oL fy : « - 4 \Z X ¢ v& eo i“ r 4 (7?) DATES OF ISSUE No. 1 (pp. 1-114) issued January 17, 1961. No. 2 (pp. 115-262) issued April 17, 1961. No. 3 (pp. 263-372) issued July 14, 1961. No. 4 (pp. 373-477) issued October 13, 1961. Printed in U.S.A. s TABLE OF CONTENTS IvAN Murray JOHNSTON, 1898-1960. With portrait. By Richard A \ wet THE GENERA OF ERICACEAE IN THE SOUTHEASTERN UNITED STATES. aN Che ate, | OMS 1 Nao [7 Ae | rr eee See Pee STUDIES IN THE THEACEAE, XXX. Tue AFRICAN SPECIES OF TERNSTROEMIA. By Clarence E. Kobuski ...0..0.0000.0000000........ STUDIES IN THE Genus Coccotopa, X. New SPECIES AND A SuM- MARY OF DISTRIBUTION IN SouTH AMERICA. By Richard A. Je LOL 0 4 tr eR Ue ONte ted nt CN 25 OF SPN e THE GENERA OF THE CYRILLACEAE AND CLETHRACEAE OF THE SOUTHEASTERN Unirep States. By Joab L. Thomas .............. STUDIES IN THE GeNus CoccoLosa, XI. Nores oN THE SPECIES TN Reta Dy RiChard Ai omard meets ed SCHIZOCARDIA BELIZENSIS: A SPECIES OF PURDIAEA (CYRILLACEAE) FROM CENTRAL AMERICA. By Joab L. Thomas .................... STUDIES IN THE THEACEAE, XX XI. A New Species oF ADINANDRA FROM THE CELEBES. By Clarence E. Kobuski ..................... THe Botanica, Resutts or THE U.S. Commission oF INQUIRY TO SANTO Dominco IN 1871. By Richard A. Howard ............ COMPARATIVE ANATOMY OF THE LEAF-BEARING ee it, STRUCTURE AND DISTRIBUTION OF SCLERENCHYMA IN THE PHLOEM OF PERESKIA, PERESKIOPSIS AND QUIABENTIA. By NR A) 21111) eg Mc i ne Ne eco So eeten Nos Tue CyTouocy or ¢ SOME CULTIVATED Spécies oF VipurnumM. By POCO Ly, AOI IR: 2 odes ate ea eh eR Mr10cENE Foss Woops FroM THE COLUMBIA BASALTS OF CENTRAL WasuHincton. By U. Prakash and Elso S. Barghoorn ......... THE GENERA OF TURNERACEAE AND PASSIFLORACEAE IN THE SOUTH- EASTERN UNITED States. By George K. Brizicky .................... A Srupy or Hysripization In DowNINGIA (CAMPANULACEAE). By Carrol. Wood Jr toa oes STUDIES IN THE THEACEAR, XXXII. A Review or THE GENUS TERNSTROEMIA IN THE PHILIPPINE ISLANDS. By Clarence E. OTIS asus bs agst ps ee a Ne oe. Lear ann Nopan ANATOMY oF SoME ANDEAN COoMPposITAE— ASTEREAE. By Otto 7’. SOlDrig .o....ccccccccccccsccsccscscscececeecesccesescccesn. A REVISION OF THE Asi1aTIC Genus LINOSTOMA (THYMELAEACEAE). Dylorin 1 Nevin dye es 8, cen THE GENERA OF VIOLACEAE IN THE SOUTHEASTERN Unit ED > St ATES. Dy George I. Briar hies eee COMPARATIVE ANATOMY OF THE LEAF-BEARING Cactaceag, III. ForM AND DISTRIBUTION OF CRYSTALS IN PERESKIA, Popes OPSIS AND QUIABENTIA. By IJ. W. Batley .00000000000 ccc. 81 87 MtoceNsE Fosstu Woops FROM THE COLUMBIA BasaLTs OF CENTRAL Wasurnaton, Il. By U. Prakash and Elso S. Barghoorn ...... A Synopsis or THE GENUS COLUMELLIA (COLUMELLIACEAE). By George K. Brizichey .......c.c.ccccccssscesseseeesesseenessssenensenenesennenenenannee A REvIsION oF THE Asiatic GENUS ENKLEIA (THYMELAEACEAE). ABERRATION IN ACMOPYLE AND PopocAaRPUS DACRYDIOIDES. By Lucy M. Cranwell ..... STUDIES IN THE THEACEAE, XXXIII. VARIATION IN THE FRUIT OF TERNSTROEMIA KWANGTUNGENSIS. By Clarence E. Kobuski Tur Correct NAMES For “Dr1ospyROS EBENASTER.” By Richard FE fg: Re ENOL TE Pree rperenC nC T op rrr Ener one rrerr mere rect A New Species or Prnus rrom Mexico. By John W. Andresen and John HH, Beaman ....1....cscccisccissssnesscvesenrsesscens FIBERS IN THE SECONDARY XYLEM OF AESCHYNOMENE. By U. PYGRGGR sisnccissignssismns THE DIRECTOR’S REPORT. ..........c:::ccescccessssesseeseeeeeeeeeseeenseeeneaeeeseeenseeenes BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND STuDENTS, JULY 1, 1960-JUNE 30, 1961.00.00... STAFF OF THE ARNOLD ARBORETUM, 1960-1961 InpEx To Vou. XLII ie eS eee a i: IvAN Murray JOHNSTON JOURNAL OF THE ARNOLD ARBORETUM Vou. XLU JANUARY 1961 NUMBER 1 IVAN MURRAY JOHNSTON, 1898-1960 RIcHARD A. HOWARD With portrait * IvAN M. JoHNSTON was born in Los Angeles, California, on February 28, 1898. the son of William Murray and Etta Farnsworth Johnston. He died suddenly on May 31, 1960. His career was devoted to taxonomic botany, and, by combining field work with herbarium study, he was able to bring to his many publications an intimate knowledge of the plants about which he wrote. Few men of his generation did this so well, and fewer students today are developing this ability. Johnston’s interest in local flora and field observations was stirred while he was still a student in high school. He collected plants and sent speci- mens to the botanical leaders of this period, each with critical observations not usually associated with a student of his age. His college work was begun at Pomona College (1916-1918) before he transferred to the Uni- versity of California at Berkeley where he received his A.B. degree in 1919. During these formative years he was influenced by T. S. Brandegee, Alice Eastwood, Harvey M. Hall, W. L. Jepson, P. A. Munz, S. B. Parish and W. A. Setchell, each of whom can be associated with some aspect otf his future work. Hall and Parish guided Johnston to his work on the San Gabriel Mountains of California. Setchell and Brandegee to the work on Baja California and the desert vegetation of Mexico, and Jepson to the work on the family Boraginaceae which was to become Johnston’s most significant contribution. Munz was his collaborator on many papers and Miss Eastwood his ideal of botanical efficiency. Johnston also received his M.A. degree from the University of California (1922) before coming to Harvard in the Fall of that year to continue his work towards his doctorate which he received in 1925. During his undergraduate years he did field work in the San Gabriel mountains with S. B. Parish and in the vicinity of Claremont, California with P. A. cree Some of the troublesome specimens were sent to Mac- bride at the Gray Herbarium, and thus began a correspondence that led * This portrait was taken by Ernst Abbe in 1942 at the Gray Herbarium, Cam- bridge, Massachusetts. 2 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII eventually to Johnston’s application to Harvard for graduate work and a long association with the Gray Herbarium. In 1918, Johnston expressed his annoyance to Macbride that Amsinckia and Cryptantha were in a “mess.” Macbride subsequently named the specimen which had bothered him Cryptantha intermedia var. johnstonii. Another collection sent to Rydberg was named by him Rosa johnstonii. These two appear to be the first of nearly a dozen taxonomic epithets commemorating Ivan Johnston’s active botanical role. During the summers of 1919, 1920 and 1922 Johnston worked as an assistant at the Alpine Laboratory of the Carnegie Institution near Boul- der, Colorado. He served as assistant and companion to Dr. Harvey Hall in ecological studies involving the transplanting of paired species of re- stricted distribution from one environment to another to determine the effects of soils, altitude and genetic factors on the tolerance, adaptability and the morphological characteristics of the species. This early training created in him an awareness of soil types in relation to plant distribution which was to be present in all his botanical observations in future years and which, in many ways, was influential in his taxonomic decisions. The summer of 1921 Johnston spent on a California Academy of Sci- ences expedition to the Gulf of California as the official botanical collector. He wrote to Macbride, “The expedition touched on every island of the Gulf of California and at a number of localities on the Sonoran and penin- sular coasts. I had the time of my young life as you may imagine. I got about 1400 numbers which aggregate in the neighborhood of 8000 speci- mens.’’ Encouraged by Macbride’s subsequent interest, Johnston made his plans to go to Cambridge for graduate work. He was awarded an Austin Fellowship but chose instead employment offered by B. L. Robin- son as herbarium assistant in the Gray Herbarium. He was listed as such on the Gray Herbarium letterhead for the years of his graduate work. Johnston completed his graduate work in 1925 with a thesis on “The North American Species of Cryptantha.” He was awarded a Sheldon Travelling Fellowship by Harvard which he used for a long trip to Chile, departing from New York on October 1, 1925, and returning on May 31, 1926. His field work in Chile was both energetic and successful. Physically active and personable he made many friends on the boat trip south and received invitations to visit remote areas for his collecting. His interest in following the desert routes of Philippi, as well as entering into new areas, led him from the driest of desert areas to Andean locations at 17,000 feet. Many of his collecting areas have not been revisited as they represent remote locations difficult to reach. The work on these collections was to occupy him for many years and necessitated trips to European herbaria for study in 1927, 1929 and 1932-1933. The last trip to Europe, made on an award of a Guggenheim Fellowship, served as a wedding trip and established contact with many European botanists. Following the death of Charles Sargent, director of the Arnold Arbo- retum, in 1930, Prof. Oakes Ames was appointed supervisor of the botan- ical collections of Harvard University. Professor Ames secured the resig- 1961 | IVAN MURRAY JOHNSTON, 1898-1960 3 nation of Johnston as senior assistant at the Gray Herbarium and his appointment as research associate at the Arnold Arboretum on August 1, 1931. This appointment was to enable Johnston to continue his field work in South America and his work on the Boraginaceae. Ames had extensive plans for Johnston to return to Chile, to visit the central Andean region, and to extend the interests of the Arboretum to Argentina and Brazil. The program for exploration was delayed but Johnston’s intense interest in the flora of the desert areas of South America persisted to the end of his life. For the remainder of his professional career Johnston was associated with the Arnold Arboretum. To enable him to teach within Harvard Col- lege, Johnston was appointed a lecturer in 1934, and in 1938 he was given the title of Associate Professor. After the retirement of Dr. E. D. Merrill as director of the Arnold Arboretum in 1947, Prof. Karl Sax was made acting director and Dr. Johnston supervisor of the herbarium and library. He served as associate director of the Arnold Arboretum from 1948 until 1954. Johnston’s courses, offered as a teacher, were of varying quality. Most outstanding was a survey of the families of flowering plants where his field knowledge of so many families, coupled with his comprehensive knowl- edge of the botanical literature, made the presentation of great value to the many students who enrolled in or audited the course. He supervised the graduate work of three students, one Chinese, one Indian and one American. The encouragement of Ames, and later Merrill, for Johnston to return to the field work which he previously had done so productively and so well was not successful until 1938, when he completed the first of three trips (1940, 1941) to the intermontane plateau and the deserts of northern Mexico. These trips were made with the codperation of the Desert Labo- ratory of the Carnegie Institution. The most significant results of these trips were his papers on the relationships of floras of North and South American deserts and on the relation of vegetation types to soils, e.g., the gypsophile plants. Johnston started a series of papers on the species composition of the flora of the Coahuila area, and articles dealing with the pteridophytes, gymnosperms, monocotyledons, and the dicotyledons through the family Nyctaginaceae were published. The remainder of the collection was never studied. During World War II, Johnston was called upon as a civilian for his botanical knowledge. He worked first in Massachusetts on the problems of using plant materials as camouflage and later, in Panama, supplied botanical data for studies of means of defoliating tropical forests by the use of chemicals. His papers on the flora of San José Island and other areas of the Isthmus of Panama present aspects of this service. He was cited by the United States Army for these contributions. In the post-war years Johnston’s botanical interests became more varied. His work with the pollen of Lithospermum caused him to consider seri- ously devoting his time to the field of palynology. An opportunity for contract work with the United States Army in Panama changed his in- 4 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII terests again to ecology, and in 1955-1956 he made four trips to Central America collecting data. At the end of his career he had returned to a study of the flora of the southwestern United States and adjacent Mexico through his codperation with a project on the flora of Texas. A manu- script treatment of the Boraginaceae of Texas was almost complete at the time of his death. This, along with all his personal library and botani- cal records has been deposited at the Texas Research Foundation, Renner, Texas. It was during this post-war period that Johnston became involved in a controversy with Harvard University officials and many of his colleagues. He sided with those opposing the move of any portion of the herbarium, library or staff of the Arnold Arboretum to the new herbarium building in Cambridge. The move was completed in 1954, but the controversy was not resolved at the time of his death Ivan M. Johnston had been a member of the American Academy of Arts and Sciences, the American Society of Plant Taxonomists, the Botanical Society of America, the New England Botanical Club, Sigma Xi, Gamma Alpha, the International Association of Plant Taxonomists, and a corre- sponding member of the Chilean Society of Natural History and the Argen- tine Society of Natural Sciences. He is survived by his widow, Mildred Williamson Johnston, a son, Wil- liam, and a daughter, Elizabeth (Mrs. Frederick Milleker). BIBLIOGRAPHY * — 1918 — A few notes on the botany of southern California. Bull. So. Calif. Acad. 17: 64- 66 Some undescribed plants from southern California. Bull. So. Calif. Acad. 17: 63, 64. — 1919 — Contributions on southern Californian botany. Bull. So. Calif. Acad. 18: 18-21. The flora of the pine belt of the San Antonio Mountains of southern Cali- fornia. Plant World 22: 71-90; 105-122. — 1922 — The distribution of sie Sopeadins Pteridophytes (with P. A. Munz). Am. Fern Jour. 12: 69-77; 101-122; 13: 1-7. arse notes on a ts via southern California, I, II. (with P. A. nz). Bull. Torrey Club 49: 31-44; 349-359. Undescribed plants mostly from Baja California. Univ. Calif. Publ. Bot. 7: 437-446. — 1923 — Diagnoses and notes relating to the Spermatophytes chiefly of North America. Contr. Gray Herb. 68: 80-104 Studies in the Boraginaceae. 1. Restoration of the genus Hackelia. 2. The genus * Compiled by Lazella Schwarten. 1961] IVAN MURRAY JOHNSTON, 1898-1960 5 ae Ul 3. Novelties and new combinations in the genus Cryptantha. nopsis and redefinition of Plagiobothrys. Contr. Gray Herb. 68: ao fe — 1924 — Expedition of the California Academy of Sciences to the Gulf of California in 1921. The botany (the vascular plants). Proc. Calif. Acad. IV. 12: 951- 1218 Miscellaneous notes on plants from southern California, III. (with P. A. Munz). Bull. Torrey Club 51: 295-302. New plants of Portuguese West ja collected by Mrs. Richard C. Curtis. Contr. Gray Herb. 73: 31-40. pl. 1 On some South American Proteaceae. oe Gray Herb. 73: 41, 42. The Penstemons of southern California (with P. A. Munn), Bull. So. Calif. Acad. 23: 21-40 Studies in the Boraginaceae, II. 1. A synopsis of the American native and immi- grant Borages of the subfamily Boraginoideae. 2. A tentative classification of the South American Coldenias. Contr. Gray Herb. 70: 1-61. Studies i in the Boraginaceae. III. 1. The Old World genera of the Boraginoideae. s on miscellaneous American Boraginaceae. Contr. Gray Herb. 73: Taxonomic records concerning American Spermatophytes. 1. Parkinsonia and Cercidium. 2. New or otherwise noteworthy plants. 3. A neglected paper by Jean Louis Berlandier. 4. On the validity of Molina’s scientific names. Contr. Gray Herb. 70: 61-92 — 1925 — Further new plants a by Mrs. Richard C. Curtis in tropical Africa. Contr. Gray Herb. 75: 23- Miscellaneous ee on ee from southern California, IV (with P. A. Munz). Bull. Torrey Club 52: 221-228 The Oenotheras of pala South America (with P. A. Munz). Contr. Gray Herb. 75: 15- The Potentillas - tas California (with P. A. Munz). Bull So. Calif. ome undescribed netee Spermatophytes. Contr. Gray Herb. 75: 27-40. Studies in the Boraginaceae, IV. The North American species of Cryptantha. Contr. Gray Herb. 74: 1-114. Studies in the Boraginaceae, V. 1. Concerning the range and identity of certain American species. 2. New or otherwise interesting Asiatic Borages. Contr. Gray Herb. 75: 40-49 — 1926 — Some notes on the Chilean relatives of Plantago patagonica Jacq. Revista Chil. Hist. Nat. 30: 13-18 — 1927 — Studies in the Boraginaceae, VI. A revision of the South American Boraginoi- deae. Contr. Gray Herb. 78: 1-118. — 1928 — The botanical activities of Thomas Bridges. Contr. Gray Herb. 81: 98-106. Some undescribed American Spermatophytes. Contr. Gray Herb. 81: 85-98. 6 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Studies in sg Boraginaceae, VII. 1. The South American species of Heliotro- pium. 2. Notes on various Se Contr. Gray Herb. 81: 1-83. — 1929 — A collection of plants from the high Cordilleras of northwestern San Juan. Physis 9: 297-326 A new Chilean plant and some nomenclatorial changes. Revista Chil. Hist. Nat. 25-27. Papers on the flora of northern Chile. 1. The coastal flora of the departments of Chanaral and Taltal. 2. The flora of the Nitrate Coast. 3. oe species from the Cordilleras of Atacama. Contr. Gray Herb. 85: 1-17 pl. 1, 2 Some undescribed species from Peru. Contr. Gray Herb. 85: 172-180. — 1930 — Some notes on the flora of northern Chile. Revista Chil. Hist. Nat. 34: 228— 234 Studies in the Boraginaceae, VIII. 1. Observations on the species of Cordia and Tournefortia known from Brazil, Paraguay and Argentina. 2. Taxo- nomic notes concerning various Borages. Contr. Gray Herb. 92: 1-95. — 1931 — The flora of the Revillagigedo Islands. Proc. Calif. Acad. IV. 20: 9-104. A new species of Trigonotis from southwestern China. Candollea 4: 310, 311. New Spermatophytes from Mexico and Argentina. Contr. Gray Herb. 95: 53- The vascular flora of the Guano Islands of Peru. Contr. Gray Herb. 95: 26-35. pl. 6,7 — 1932 — New records for the flora of the Nitrate Coast. Revista Chil. Hist. Nat. 36: 4-8. Studies in the Boraginaceae, IX. 1. The Allocarya section of Plagiobothrys in the western United States. 2. Notes on various Borages of the western United States. Contr. Arnold Arb. 3: 1-102. — 1935 — Boraginaceae. Jn: Munz, P. A. A manual of southern California botany: 417- 436. The ie of San Felix Island. Jour. Arnold Arb. 16: 440-447. pl. 165. Studies in the Boraginaceae, X. The Boraginaceae of northeastern South Americ : Studies in the Boraginaceae, XI. The species of Tournefortia and Messer- schmidia in the Old World. 2. Notes on Brand's treatment of Cryptantha. 3. New or otherwise noteworthy species. Jour. Arnold Arb. 16: 145-205. — 1936 — Boraginaceae. /n: Pulle, A. Flora of Suriname 4 1: 306-333. study of the Nolanaceae. Proc. Am. Acad. 71: 1-87. Reprinted in Contr. Gray Herb. 112: 1-87. — 1937 — Studies in the Boraginaceae. XII. 1. Trigonotis in oe China. 2. Novelties and critical notes. Jour. Arnold Arb. 18: 1961] IVAN MURRAY JOHNSTON, 1898-1960 7 — 1938 — New or noteworthy plants from temperate South America. Jour. Arnold Arb. 248-263 Notes on some Astragalus species of Ecuador and Peru. Jour. Arnold Arb. 19: 88-96 Some undescribed species from Mexico and Guatemala. Jour. Arnold Arb. 19: 117-128 The species of Sisyrinchium in Uruguay, Paraguay and Brazil. Jour. Arnold Arb. 19: 376-401 — 1939 — New Fuchsias from southern Peru. Jour. Arnold Arb. 20: 241-244. New phanerogams from Mexico. Jour. Arnold Arb. 20: 234-240 The preparation of botanical specimens for the herbarium. 1-33. pl. 1-5. The rnold Arboretum, Jamaica Plain, Mass. Studies in the Boraginaceae, XIII. New or otherwise oe. species, chiefly from western United States. Jour. Arnold Arb. 20: 375-402 — 1940 — The floristic significance of shrubs common to North and South American deserts. Jour. Arnold Arb. 21: 356-363. w phanerogams from Mexico, IJ, II. Jour. Arnold Arb. 21: 67-75; 253-265. = in oe Scan XIV. Miscellaneous species from Asia, Malaysia, A a. Jour. Arnold Arb. 21: 48-66. Studies in Be eacie XV. Notes on some Mexican and Central American species of Cordia. Jour. Arnold Arb. 21: 336-355 — 1941 — Bibliographic data concerning Gay’s Flora de Chile. Darwiniana 5: 154-165. El significado floristico de los arbustos comunes a los desiertos norte y sudameri- canos. Revista Argent. Agron. 8: 114-121 Gypsophily among Mexican desert plants. Tour Arnold Arb. 22: 145-170. New Phanerogams from Mexico, IV. Jour. Arnold Arb. 22: 110-124 Preparacion de ejemplares botanicos para herbario. Traducido ‘del Ingles al Castellano por H. R. Descole y C. A. O’Donell. 1-49. pl. 1-5. Tucuman — 1942 — Boraginaceae. /n: Kearney, T. H. & Peebles, R. H. Flowering plants and ferns of Arizona. U.S. Dept. Agr. Misc. Publ. 423: 740-761. ee OO New phanerogams from Mexico, V. Jour. Arnold Arb. 24: 90-98. Noteworthy species from Mexico and adjacent United States. I. Jour. Arnold rb. 24: 227-236. Plants of eg eastern Chihuahua, and et Zacatecas and Durango, ur. Arnold Arb. 24: 306-339, 375-42 Publication ae for the botanical parts of the an Railroad reports. Jour. Arnold Arb, 24: 237-242. — 104s oa e Coahuila, eastern Chihuahua, and adjoining Zacatecas and Durango. . Jour. Arnold Arb. 25: 43-83; 133-182; 431-453. 8 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Publication dates of ia i Botany of the Voyage of the Bonite. Jour. Arnold Arb. 25: 481-4 — 1945 — A a a sketch of Latin America (with A. C. Smith). Jn: Verdoorn, F and Plant Sciences in Latin America: 337-349. The pemiaride of San José Island, Perlas group, Republic of PeORIOK. ae cal Warfare Service, San José Project Report 87: 1-104; 88: — 1947 — Astragalus in Argentina, Bolivia and Chile. Jour. Arnold Arb. 28: 336-409. os 1945 wae concibe In: Maguire, et al. Plant explorations in Guiana in 1944, chiefly e Tafelberg and the Kaieteur Plateau—V. Bull. Torrey Club 75: 5 Noteworthy species ia Mexico and adjacent United States, II. Jour. Arnold Ar 19 Studies in the Sie rr XVI. Species chiefly from Mexico and western United States. Jour. Arnold Arb. 29: 227-241. — 1949 — The Botany of San José Island (Gulf of Panama). Sargentia 8: 1-306. pl. 1-17. Studies in the Boraginaceae, XVII. A. Cordia section Varronia in Mexico and Central America. B. The eri of species proposed by Sessé and Mocifo. Jour. Arnold Arb. 30: 85-11 Studies in the Boraginaceae, XVII. Boraginaceae of the southern West Indies. Jour. Arnold Arb. 30: 111- — 1950 — Noteworthy aes - Mexico and adjacent United States, III. Jour. Arnold Arb. 31: Studies in he fee ene XIX. A. Noteworthy species from tropical America. ; dia § Gerascanthus in Mexico and Central America. Jour. Arnold Arb. 31: 172-187 — 1951 — Studies in the Boraginaceae, XX. Representatives of three subfamilies in eastern Asia. Jour. Arnold Arb. 32: 1-26; 99-122 Studies in the Boraginaceae, XXI. Sino-Indian species of Onosma. Jour. Ar- nold Arb, 32: 201-225; 344-368. — 1952 — sie in the Boraginaceae, XXII. Noteworthy species, chiefly Asian and outh American. Jour. Arnold Arb. 33: 62-78. oe in the Boraginaceae, XXIII. A survey of the genus Lithospermum. Jour. Arnold Arb. 33: 299-366. pl. 1-3. — 1953 — Boraginaceae. Jn: iy Trinidad & Tobago. Dept. Agriculture Trinidad & Tobago 24: 189-2 1961 | IVAN MURRAY JOHNSTON, 1898-1960 9 Boraginaceae. /n: Steyermark, J. A., et al. Contributions to the Flora of Venezuela. Fieldiana Bot. 28: 511, 512. Studies in the Boraginaceae, XXIV. A. Three genera segregated from Litho- spermum. B. Supplementary notes on Lithospermum. Jour. Arnold Arb. —16. Studies in the Boraginaceae, XXV. A revaluation of some genera of the Litho- spermeae. Jour. Arnold Arb. 34: 258-299. — 1954 — Studies in the Boraginaceae, XXVI. Further revaluations of the genera of the Lithospermeae. Jour. Arnold Arb. 35: 1-81. Studies in the Boraginaceae, XX VII. Some general observations concerning the Lithospermeae. Jour. Arnold Arb. 35: 158-166 — 1956 — co in the asia rae. XXVIII. New or pre interesting species merica and Asia. Jour. Arnold Arb. 37: 288- Ath study of the eee Canal Zone with fae reference to the Fort Sherman area and vicinity (with McCullough, C. R. & Parker, J. M.) North Carolina State College, Dept. Engineering Research. 1-267. — 1957 — Boraginaceae. Jn: Steyermark, J. A., et al. Contributions to the flora of Vene- zuela. Fieldiana Bot. 28: 1080-1082. Studies in the Boraginaceae, XXIX. Echiochilon and related genera. Jour. Arnold Arb. 38: 255-293. — 1959 — Some noteworthy American Borages. [Studies in the Boraginaceae XXX. ] Wrightia 2: 13-22. 10 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII THE GENERA OF ERICACEAE IN THE SOUTHEASTERN UNITED STATES ! CARROLL E. Woop, Jr. ERICACEAE Jussieu (HEATH FAMILy) Trees or shrubs to herbs or fleshy parasites, rarely trailing or scrambling vines [or epiphytes], almost all of acid soils and mostly with mycorrhizal associations. Leaves alternate or sometimes opposite or whorled, simple, ‘Prepared for a biologically oriented generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University which has been made possible through the support of George R. Cooley and the National Science Foun s treatment (which includes Monotropaceae and Pyrolaceae) follows the pattern peers in the first paper in she series (Jour. Arnold Arb. 39: 296-346. 1958) and continued through those in volumes 40 and 41. It should be _Tepeated that the area a by this work is bounded by gee includes signed for the species of this area. References which I have not seen are marked by an asterisk. The preparation of this study of Ericaceae has led into unexpected byways, and the omewh literature is formidable the other hand, the numerous published studies have be o provide real bases for an understanding of the complex biological problems, and the many papers have been most gratefully drawn upon outlines of the fam Many people have contributed generously to this study. In addition to the valued comments of my immediate colleagues and the careful help of Mrs. Gordon W. Dillon or information to R. B. Channell, G. R. Cooley, W. H. Duncan, R. J. Eaton, Mrs. - oe Henry, J. dere Miss a H. Lee, Mrs. E. O. Mellinger, J. D. E. Richardson, H. F. L. Rock, and E. T. Wherry. Henry T. Skinner has kindly read the manuscript on appara ndron, “4 A. Howard has generously given me access to his unpublished data on the nodal anatomy of members of the family, and L. Thomas has struggled valiantly with some frustrating cytological material. I hes ve also been privileged to che al a number of distributions against the mounted materials of the herbarium of the Department of Botany at the University of North Carolina at Chapel Hill. The eight generic illustrations are the continuing and careful work of Dorothy H. Marsh. 1961] WOOD, GENERA OF ERICACEAE 11 exstipulate, often coriaceous and persistent [frequently needle-like] ; nodes with one trace from one gap. Flowers bisexual (rarely functionally uni- sexual and the plants dioecious), characteristically regular, in some ir- regular (mostly by displacement), most commonly 5(4—7)-merous, typi- cally obdiplostemonous, the insertion hypogynous to epigynous. Inflores- cences basically racemose (racemes, panicles or corymbs), the individual flowers in the axils of leaves or bracts, the pedicels typically with 2 lateral bractlets. Perianth usually biseriate (the calyx reduced in Monotropa). Calyx 4-7-lobed or sepals distinct, usually persistent, valvate, imbricate or reduplicate in bud. Corolla of 4-7 sometimes distinct but usually united petals, often funnelform, campanulate or urceolate, convolute or imbri- cated in bud. Stamens usually twice or (less frequently) equal the num- ber of petals, usually inserted at the edge of a nectariferous, variously lobed disc [rarely epipetalous], distinct, the filaments often flattened, dilated, or S-shaped [united in some tropical Vaccinioideae]; anthers be- coming inverted in development, dehiscing from the base (the apparent apex) by longitudinal slits, clefts, or pores, often with terminal awns or abaxial spurs; pollen grains in tetrads or single. Gynoecium syncarpous, the carpels [2]3—7|-10], typically 4 or 5; stigma simple, often lobed; style single, with a stylar canal with as many rays as locules and continuous with the locules; ovary 1, superior, half-inferior or inferior, typically with as many locules as carpels, the placentation axile, but sometimes loculate below and 1-locular above, the placentation then axile below, parietal above, or with twice as many locules as carpels by the development of additional partitions; ovules anatropous to campylotropous, 1-integu- mented, numerous to 1. Fruit a loculicidal or septicidal capsule, berry, or drupe, the calyx usually persistent, rarely accrescent and fleshy. Seeds usually small, sometimes winged or tailed; embryo small (in parasitic forms very much reduced), straight, with fleshy endosperm. Base chromo- some numbers 6, 8, 12, 13, 19, 23. (Including Monotropaceae, Pyrolaceae, Vacciniaceae.) TYPE GENUS: Erica L. A family of about 75-80 genera and some 2000 species, widely dis- tributed, mostly in acid soils, from sea level to high elevations in the tropics, and throughout the temperate regions of both hemispheres (al- though largely lacking in Australia where replaced by Epacridaceae), some extending into the arctic. About 35 genera occur indigenously in the United States, 20 in our area. Erica L. and Calluna Salisb. are natural- ized to the north of our area. Ericaceae share the syncarpous gynoecium with single stigma, style, and ovary, the style with a fluted canal continuous with each locule; the flowers generally with sympetalous corolla; the androecium typically obdi- plostemonous, the usually distinct stamens from the edge of a nectarifer- ous disc; the anthers inverted either in development or at anthesis, open- ing by morphologically basal pores or slits or longitudinally; and the pol- len usually in tetrads (with the exception of Monotropoideae and a few others). Some of the apopetalous genera appear to be primitively so (e.g., 12 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Befaria, Elliottia) but others seem to be derived types (e.g., Ledum, Pyrola, Chimaphila, Monotropa). The group is predominantly woody with specialized forms herbaceous (e.g., species of Gaultheria, Pyroleae, Mono- tropoideae ). The family is further characterized by a well-marked combination of embryological features found only in the Ericales: ovules with a single integument and thin, evanescent nucellus, a more or less differentiated endothelium, a lack of parietal cells; embryo sac of the Polygonum type from a single archesporial cell which functions directly as the megaspore mother cell; endosperm cellular, the first two divisions transverse; the formation of endosperm haustoria at both ends of the embryo sac (with a few exceptions); single-layered seed coat formed from the outermost layer of the integument; endosperm fleshy, the embryo straight; anthers lacking a fibrous layer, the tapetum with multinucleate cells; pollen grains 2-celled and (in most instances) adhering in tetrads. Structurally and embryologically the Ericaceae are clearly related to Clethraceae and Cyrillaceae, on the less specialized side, and to Epacri- daceae and Empetraceae, on the more specialized, the five constituting the Ericales. Subfamilial categories are not well agreed upon, and the limits of the family itself have been variously interpreted, some authors separating Pyrolaceae, Monotropaceae, and Vacciniaceae. However, Pyroleae and Monotropeae appear to be related to Andromedeae and Arbuteae, and, with the exception of the herbaceous or parasitic habit and attendant modi- fications, the features are all those of Ericaceae (Henderson, Copeland). The Vaccinioideae differ only in the partially to completely inferior ovary (cf. Gaultheria) which becomes baccate or drupaceous in fruit (cf. Ar- buteae, Monotropoideae) ; separation as a family on the basis of this one character seems hardly justifiable. The Ericaceae are here treated as com- posed of subfamilies Ericoideae (including Rhododendroideae, Arbu- toideae, Pyroloideae), Monotropoideae, and Vaccinioideae; five tribes (by some regarded as eleven) occur in our area. The family includes both small, isolated, well-marked genera and large, polymorphic complexes with indistinct generic lines. Within genera species may be similarly distinct or complicated. Polyploidy is frequent in some genera. The, flowers of most are showy, those of the temperate areas being adapted for pollination by insects (or those of some tropical species by birds). The stamens of species with campanulate to urceolate pendent flowers are often appendaged in such a way that the disturbance of the anthers by an insect visitor and the sifting out of pollen are insured. In species with horizontal flowers and exserted stamens, on the other hand, the stamens are unappendaged but the pollen tetrads are often linked by viscin strands so that the entire mass is pulled from the anther as a cob- webby net. In Andromedeae the anthers frequently have disintegration tissue which produces powdery white areas (of apparently unknown func- tion) in the connective. 1961 } WOOD, GENERA OF ERICACEAE 13 Economically the group is notable chiefly as a source of showy orna- mental woody plants, but various species of Vaccinium (blueberries, cran- berries, mountain cranberries) are of increasing importance for their fruits. REFERENCES: Artopoeus, A. Uber den Bau und die Offnungsweise der Antheren und die Entwicklung der Samen der Ericaceen. Flora 92: 309-345. 1903. BaILton, H. Ericacées. Hist. Pl. 11: 122-210, 1892. [Includes Epacridaceae, Lennoaceae: Diapensiaceae, Empetraceae, Clethraceae, et al.] Bett, H. P., and J. Burcuity. Winter resting stages of certain Ericaceae. Canad. Jour. Bot. 33: 547-561. 1955. [Mostly concerned with state of de- velopment of ovules and stamens; includes most woody Ericaceae of eastern Canada, many of our genera. | BENTHAM, G., and J. D. Hooker. Vacciniaceae, Ericaceae, Monotropeae. Gen. Pl. 2: 564-608. 1876. BERGMAN, H. F. Internal stomata in ericaceous and other unrelated fruits. Bull. Torrey Bot. Club 47: 213-221. 1920. BREITFELD, A. Der anatomische Bau der Blatter der Rhododendroideae in Beziehung zu ihrer systematischen rene und zur geographischen Verbreitung. Bot. Jahrb. 9: 319-379. pls. 5, 1888. CANDOLLE, A. P. DE. Vaccinieae, Ericaceae. nan 7: 552-733; Pvyrolaceae. 772-776; Monotropeae. 779-781. 9. COPELAND, H F. The development of seeds in certain Ericales. Am. Jour. Bot. 0: 513-517. 1933. [Stages in development of seeds and ovules observed in Sarcodes, Chimaphila, Pyrola, Pterospora, Pleuricospora. | A study, anatomical and taxonomic, of the genera of Rhododendroideae. Am. Midl. Nat. 30: 533-625. 1943. [Many important data; extensive bibliography. | a aaa on the structure and classification of the Pyroleae. Ma- drono 9: 65-10 47 Cox er. ane in the comparative anatomy of the Ericales. I. Ericaceae — subfamily Rhododendroideae. Am. Midl. Nat. 39: 220-245. 1948. II. Ericaceae — subfamily Arbutoideae. bid. 40: 493-516. 1948. [Wood anatomy. | Don, D. An attempt at a new arrangement of the Ericaceae. Edinburgh New Philos. Jour. 17: 150-160. 1834. [The basic paper in the taxonomy of the bats | Drube, O. Pyrolaceae, Ericaceae. Nat. Pflanzenfam. IV. 1: 15-65. 1889. nee J. The biological significance of false witches’ fale: in ericaceous plants. Jour. Wash. Acad. Sci. 8: 527-532. 1918. [Fungi on Arbutus and Vaccinium; no evidence of a symbiotic association comparable to the mycorrhizal one. | ErRDTMAN, G. Pollen morphology and plant taxonomy. V. On the occurrence of tetrads and dyads. Sv. Bot. Tidskr. 39: 286-297. 1945. GopLey, E. J. Unisexual flowers in the Ericales. Nature 180: 284, 285. 1957. [In Epacridaceae (Cyathodes, Leucopogon); also mentions Gatltheria, Epigaea. | Gray, A. Ericaceae. Syn. Fl. N. Am. 2(1): 14-50. 1878. [An important re- vision. | 14 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Grevitiius, A. Y., and O. KrrcHNer. Ericaceae. Jn O. KirCHNER, E. Loew, & C. SCHROTER, Lebensgeschichte der Blitenpflanzen Mitteleuropas. 4(1): 1-164. 1923. Monotropaceae. /bid. 181-243. 1925. [Includes extensive bibliography. | s, A. Anatomie comparée de la moelle dans les Ericinées. Bull. Soc. Bot. Fr. 17: 11-16. 1870. Hacerup, O. Morphological and cytological studies of Bicornes. Dansk Bot. Ark. 6(1): 1-27. 1928. Zytodkologische Bicornes- anise Planta 32: 6-14. 1941. [Ericaceae (including Pyrolaceae), Empetra ; morphology and ede of the leaves in Ericales. Phytomor- phology 3: 459-464. 1953. Hara, N. Development of the leaf margin in the Ericaceae (a preliminary re- port). Bot. Mag. Tokyo 69: 442-446. pl. 12. 1956. HeNperSON, M. W. A comparative study of the structure and saprophytism of the Pyrolaceae and Monotropaceae with reference to their derivation from the Ericaceae. Contr. Bot. Lab. Univ. Penn. 5: 42-109. 1920 Ikuse, M. The presence of the viscid threads among pollen grains in - Phyllo- doceae, etc. of Ericaceae. Jour. Jap. Bot. 29: 146-148. 1954. [Cf. CopE- LAND, 1943.] Kearney, T. H. Report on a botanical survey of the Dismal Swamp Region. Contr. U. S. Natl. Herb. 5: 321-550. pls. 65-77. 1901. [499-503, leaf anatomy of Leucothoé axillaris, L. racemosa, Lyonia lucida, L. ligustrina, Kalmia angustifolia, Vaccinium arboreum, V. macrocarpon. | Kewuey, A. P. Mycotrophy in plants. xvi + 223 pp. Chronica Botanica, Wal- tham, Mass. 1950. [Includes a complete review of the literature on mycor- rhizae. Mattuews, J. R., and E. M. Knox. The comparative morphology of the stamen in the Ericaceae. Trans. Proc. Bot. Soc. Edinburgh 29: 243-281. 1926. Niepenzu, F. Uber den anatomischen Bau der Laubblatter der Arbutoideae und Vaccinioideae in Beziehung zu ihrer systematischen Gruppierung und geo- graphische Verbreitung. Bot. Jahrb. 11: 134-263. pls. 3-6. 1890. | Gen- eral and systematic account of leaf anatomy. | Patser, B. F. Studies of floral morphology in the Ericales. I. Organography and vascular anatomy in the Andromedeae. Bot. Gaz. 112: 447-485. 1951. II. Megasporogenesis and megagametophyte development in the Androme- deae. Ibid. 114: 33-52. 1952. III. Organography and vascular anatomy in several species of the Arbuteae. Phytomorphology 4: 335-354. 1954. Petrrisot, C. N. Développement et structure de la grain de quelques Erica- cées. Jour. Bot. Morot 18: 309-367, 386-402. 1904. (Note préliminaire. Ibid. 234-242.) Recorp, S. J. Woods of the Ericales, with particular reference to Schizocardia. Trop. Woods 32: 11-14. 1932. [See Cyrillaceae. | Rypserc, P. A. Pyrolaceae. N. Am. FI. 29: 21-32. 1914. SaMUELssoNn, G. Stiidien iiber die Entwicklungsgeschichte der Bliite einiger Bicornes-Typen. Sv. Bot. Tidskr. 7: 69-188. 1913. [A classical embryo- logical study. | Stapkov, A. N. On morphological characteristics of pollen grains of the heath family (Ericaceae). (In Russian.) Dokl. Akad. Nauk SSSR. 92: 1065- 1068. 1953.* 1961 | WOOD, GENERA OF ERICACEAE 15 SMALL, J. K. Monotropaceae, Ericaceae. N. Am. Fl. 29: 11— 18, 33-102. 1914. [ Does not include Vaccinioideae. | SmitH, A. C. The American species of Thibaudieae. Contr. U. S. Natl. Herb. 28: 311-547. pls. 1-19. STANLEY, O. B. Fat deposits in .cektain Ericaceae. Butler Univ. Bot. Stud. 2: 33-41. 1931. [Cassandra, meee Kalmia, Leiophyllum, Rhodo- dendron,. Vaccinium, Gaylussa TAKEMOTO, T., and N. YAHacr. sain on the constituents of Ericaceae. III. (In J etenese English summary.) Jour. Pharm. Soc. Japan 78: 304, 305. 1958. WarMING, E. The structure on biology of arctic flowering plants. I. Ericineae (Ericaceae, Pirolaceae). Morphology and biology. Meddel. Gronl. 36: 1-71. 1908. [See also H. os PETERSON. 2. The biological anatomy of the leaves and stems. /bid. 75-138. ] Wuerry, E. T. Observations on the soil acidity of Ericaceae and associated plants in the Middle Atlantic states. Proc. Acad, Nat. Sci. Philadelphia 72: 84-111. 1920 Correlation between vegetation and soil acidity in southern New Jer- . Lbtd. 113-119. Soil tests of Ericaceae and other Aaosee rene families in northern Vermont and New Hampshire. Rhodora 22: 33-49. 1920 Yano, B. Y. Pollen grain morphology in the cae Guar Jour. Taiwan Mus. 5: 1-24. 1952. [58 spp. in 20 genera described and figured. | ARTIFICIAL KEY TO THE GENERA OF ERICACEAE IN THE SOUTHEASTERN UNITED STATES (Based upon flowering material, with subsidiary characters from the fruit.) A. Plants woody or herbaceous, autotrophic, with green leaves; pollen in B. Petals distinct; ovary superi C. Plants woody, shrubs or “small trees; flowers 4—7-merous; ovary 3-7- — ae Oe axile; capsule septicidal. D. Ovary 2-, 3- 7-locular; ovules anatropous; seeds not winged; nants a E. Ovary 7- tose petals 7; stamens 12-20 (usually 14); flowers showy, in terminal racemes, producing quantities of sticky nec- LC) COMO DE Rw Nee et te, AUN 2 ae 1. Befaria. E. Ovary 2- or 3-locular; oe 5; stamens 10: ee small, in umbel-like inflorescences. ................ 4. Leiophyllum. D. Ovary 4- or 5-locular on the same plant; ovules flat, sessile on the massive placentae; rarely fruiting; seeds winged all around: deciduous shrub or small tree with racemes or panicles of white flowers with recurved petals. ............. 2. Elhiottia. . Plants herbaceous. evergreen, with slender subterranean rhizomes; flowers 5-merous; ovary 5- docular below, 1-locular above, the placen- tae axile below, becoming parietal above; capsule loculicidal. F. Inflorescence corymbose; flowers regular; filaments hairy, en- larged at or below the middle; style short, straight; capsule valves with smooth margins. .. eae 15. Chimaphila. O JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII F. Inflorescence racemose; flowers more or less irregular; filaments glabrous, not enlarged i in the middle; style declined; capsule valves with cobwebby margins. .................-...555. 16. Pyrola. B. Petals united (usually conspicuously 0) or, x appearing distinct, the ovary inferior (some species of Vaccinium). G. Stamens with disintegration tissue inane white, powdery) on the apparent abaxial side at the top of the filament or in the connective or upward o ae anther-halves, the anthers or filaments often ap- ae es (or if stamens aborted or not opening see 14. Epigaea) ; fruit a aa capsule (the calyx fleshy in Gaultheria, surround- ing the fruit and appearing berry-like). H. Plants low, subherbaceous, evergreen, with underground rhizomes; anther-halves with 2 terminal awns; flowers solitary in the axils of leaves, with a pair of bracteoles immediately beneath the calyx; calyx accrescent, fleshy, berry-like; seeds with hard coats. ee a eee eee 8. Gaultheria. . Plants erect, seperuansasly — aba or trees; calyx not accrescent; seeds with thin, loose I. Inflorescence rma elongate, racemose or paniculate; capsule margins not differentiated Anthers appendaged on the abaxial side with 2 stout, re- flexed spurs just above junction with filament, these lack- ing white disintegration tissue; each anther-half with a line of disintegration tissue extending upward above spurs; corolla glabrous; evergreen shrubs with cable eee axil- lary: racemes. s : 206. iawes he: Pieris. J. geal not + appendased on dbacial: side, “with or amerieaie nal a en) Corolla “puberulous without; deciduous trees with ter- minal panicles; stamens unappendaged, with a small area of disintexration tissue near apex of each anther- half; capsule ovoid-pyramidal. ...... 13. Oxydendrum. _ Corolla glabrous; deciduous to evergreen shrubs with terminal and axillary racemes; anthers with or without 2 or 4 terminal awns, with white disintegration tissue in angle between awns or this position when lac awns; capsule depressed-globose. ...... 7. Le ere I. Inflorescences of corymbose clusters (reduced racemes) in axils of leaves or bracts (sometimes thus a raceme of corymbs) of the same or preceding season. L. Stamens with 4 prominent terminal awns; disintegration tis- sue in a conspicuous patch below awns; shrub with white, broadly campanulate flowers on wood of — season. Zenobia. “a L. Stamens anew: terminal awns, wath: or ene reflexed spurlike appendages on upper part of filament (these, when present, with a margin of white disintegration tissue ex- tending upward onto the anther-halves); carpel midrib dii- ferentiated (colored or thickened), and to varying degrees splitting away separately from valves in fruit; evergreen 1961] WOOD, GENERA OF ERICACEAE 17 to deciduous shrubs with globose-urceolate to cylindric CORON et Rae on eb US ey en een. Tee ee 10. Lyonia. G. Stamens lacking white disintegration tissue on apparent abaxial side. (If not opening or aborted, see 14. Epigaea.) M. Ovary superior; stamens unappendaged; fruit a capsule. N. Corolla saucer-shaped, with 10 pouches in which rs anthers re MGM IDE yyw’. amo akaikdn 4 hae ole aloes . Kalmia. N. Corolla various, wiht pouches for the anthers. O. Flowers 4-merous; corolla campanulate or urceolate: sta- mens included; anthers opening by short slits; capsule sep- Ce MR nS oe ied whe ene 5. Menziesia. O. Flowers 5-merous. P, Stamens conspicuously exserted; anthers not a opening by terminal pores; flowers showy, in Os inflorescences from buds of the preceding season; cap- GUNG SSDI nes 5 i oa aea< odie ome 6. Rh ae P. Stamens included; capsule loculicidal. . Anthers tapering into tubular beaks, opening by ter- minal pores; flowers bisexual; corolla oblong-urceo- late; flowers solitary in the axils of reduced ee leaves; scurfy-leaved shrub of bogs. .......... © @) ; Anthers opening along the entire length; Sante fines tionally dioecious, the stamens aborted or not open- ing in carpellate flowers; corolla salverform; flowers in the axils of bracts; trailing broad-leaved ever- BlCeM es GUL Ue seca aes et as eel 14. Epigaea. M. Ovary inferior or half-inferior; fruit fleshy. R. Ovary half-inferior; flowers 4-merous; each anther-half with a bifurcate terminal awn; diminutive creeping subherbaceous plant of cold mountain bogs, the solitary flowers in the leaf axils: fruit white.f1 ou oe eee ees . 8. Gaultheria. . Ovary inferior; flowers 4- or 5-merous; ates without ter- minal awns but the anther-halves more or less prolonged and tubular; inflorescences various; fruit black, blue, or red (very rarely white S. Ovary 4- or 5-locular, or 10-locular above by the develop- ment of partial false partitions; ovules numerous; fruit a many-seeded, fleshy to leathery berry; anthers with or without spurs on the abaxial side. ....... . 19. Vaccinium. . Ovary 10-locular, 10-ovuled, becoming baccate-drupaceous with 10 hard nutlets; anthers unappendaged...... sss Saas @ 8:8, 2 ea Sta a he a 20. Gaylussacia. A. Plants herbaceous, parasitic, lacking chlorophyll, mie. reduced scalelike leaves, variously colored; pollen grains single. T. Petals united; calyx of 5 regular sepals; fruit a berry. bch as GOR Basie @ i & Sane Re RO CORAM 2 echo 17. Monotropsis. bie Pauls distinct, saccate at the base; calyx represented by 1-5 scale- or bractlike sepals; fruit a loculicidal oe .......... 18, Monotropa. x N 18 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Subfam. ERICOIDEAE Endlicher (‘Ericinae’) Tribe RHoporEAE D. Don ? 1. Befaria Mutis ex Linnaeus, Mantissa Pl. 152, 242. 1771. filaments free, elongate, dilated at the base, pubescent below; anthers dorsifixed, muticous, dehiscing by 2 oblique slits at the apex, without dis- * According to changes adopted by the IX International Botanical Congress (1959) modifying Article 19 of the International Code of Botanical Nomenclature (to read, in part, “The name of any taxon of a rank below family and abcve genus which in- cludes the type of the next higher taxon must be based on the same stem as the ame without an author. Other families may not fare so well. Under the principle of the repetition of generic or subgeneric names for included subgenera and sections, already in use in Article 22 of the Code and now to be embodied in Articles 17 and 19, pre- cisely the same taxon in large and complex orders, families, or genera may have alter- native names (and types) depending upon the taxonomic ideas of the author as to taxon between the rank of order and series would seem to lead only to confusion and obfuscation of formerly clear taxa and to work against the interests of nomenclatural fa) “= i = | 5 o t=] a, FX} > i= sn J ) = —< p | Q. = ~ “A ° = oO 3 a =, = exercise in total confusion, attempt to 1 the classifications adopted by Rehder, Sleumer, Seithe-von Hoff, and the present author!) 1961} WOOD, GENERA OF ERICACEAE 19 integration tissue; pollen tetrads with viscin strands. Stigma depressed- capitate, 7-grooved; style terete, elongate, bent to one side; ovary superior (or the base slightly immersed in the receptacle), 7-sulcate, 7-locular, with numerous elongate ovules on large axile placentae. Capsule erect, de- pressed globose, surrounded at the base by the persistent calyx and termi- nated by the elongate style, 7-valved, septicidal from the apex. Seeds numerous, elongate, narrow, curved, tapered, some angled by compression. (Bejaria Vent.) Type species: B. aestuans L. (Described from Mutis’ manuscript, the name, apparently intended in honor of Mutis’ friend José Bejar, professor of botany at Cadiz, misread and written consistently by Linnaeus as Befaria, rather than Bejaria.) Perhaps 10-30 species in two sections: Befaria racemosa Vent. in our area; B. cubensis Griseb. in western Cuba; four or more species in Mexico, Guatemala, and Honduras; the remainder in the Andes, from Venezuela to Bolivia, and on the disjunct table-topped mountains eastward to Mt. Roraima. Mansfeld and Sleumer recognized 20 species in South America. whereas Camp has suggested that as few as four basic species may account for the range of variation, the difficulties in delimiting taxa being ascribed to hybridization. Befaria racemosa, a shrub 1-2.5 m. tall, with sparsely hispid twigs and conspicuous, fragrant, white flowers in winter, spring, or early summer, occurs in flatwoods and sand-scrubs from Dade, Highlands, and Lee coun- ties, Florida, northward to Camden, Glynn, and Worth counties, Georgia. Nectar is produced copiously from the glandular inner surface of the calyx, hence the name tar-flower or fly-catcher. The species lacks a hypodermis in the leaves, in contrast to all other species, and differs further in the elongate raceme (or panicle) much exceeding the leaves. It is relatively invariable and isolated within the genus and is the sole member of sect. RAcEMOSAE Fedtsch. & Basilev. According to Mansfeld and Sleumer its closest relatives are B. glauca Humb. & Bonpl., of Colombia, and B. laevis Benth., of Mexico, the two later united under the former name by Camp. Befaria cubensis, geographically nearest to B. racemosa, is a very differ- ent and highly specialized plant, presumably most closely related to B. ledifolia Humb. & Bonpl., of Colombia. Copeland has shown that Befaria shares with Elliottia and Clado- thamnus viscin strands among the pollen tetrads and resorption tissue in the anthers, a combination found nowhere else, but differs from them in having haustoria at both ends of the endosperm and in the opening of the anthers by a brief slit. It is unique among genera of Rhodoreae in the trilacunar vascular supply to the sepals and in the opening of the anthers by the formation of a cleft between intact cells (rather than by collapse or resorption of cells). On the basis of these features, as well as the gross morphology of the flower, the genus was set apart by Copeland as a tribe and regarded as the most primitive of the Ericaceae. In wood structure, however, “Befaria |5 species of sect. BEFARIA] is more advanced than several of the lower genera of other tribes” (Cox). In a comparison of 20 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII Befaria with Clethra Copeland concluded, “The effect of this comparison is to show that while Clethra and the Rhododendroideae are descended from a common ancestor, this ancestor and all transitions between it and the groups named have disappeared, leaving the survivors quite isolated.” REFERENCES: Under family references see also CopELAND (1943, pp. 548-553), Cox (Eri- cales I, pp. 228, 229). Camp, W. H. Studies in the Ericales. A discussion of the genus Befaria in North America. Bull. Torrey Bot. Club 68: 100-111. 1941. [Recognizes 4 species and 1 variable taxon of hybrid origin. | FEDTSCHENKO, B. A., an _ A. BAsILevsKaJA. Revisio generis Bejaria Vent. (In English.) Not. Syst. Leningrad 6: 37-45. 1926. . Origin and geographical distribution of the genus Bejaria Mutis. (In Russian; English summary.) Zhur. Russ. Bot. Obshch. 11: 285-296. 1927.* Revision of the genus Bejaria Mutis. Bot. Gaz. 85: 299-322. 1928. [41 spp., 8 new, 36 in South America. | MAnsFELD, R., and H. SLeuMER. Revision der Gattung Befaria Mutis. Notizbl. Bot. Gart. Berlin 12: 235-276. 1935. SMALL, J. K. Befaria racemosa. Addisonia 15: 21-23. pl. 491. 1930. 2. Elliottia Muhlenberg ex Elliott, Sketch Bot. S. C. & Ga. 1: 448. 1817. Deciduous tree to 6 (9) m., or often a shrub to 3 m., with gray bark, the twigs chestnut-brown. Leaves alternate, petiolate, oblong or oblong- elliptic, the apex acute, the base cuneate, membranaceous, pale and pubes- cent beneath. Inflorescences terminating the growth of the season, elongate panicles or loose racemes 8-25 cm. long, the flowers slightly irregular, slender-pediceled, with deciduous, subulate bracts. Calyx small, 4(5)- lobed, persistent. Petals 4 (or almost as frequently 5 in the same inflores- cence), free, oblong-lanceolate, recurving, white. Stamens 8 (10), the filaments flattened, the anthers 2-lobed at the base, the apex apiculate, opening almost from base to apex by 2 V-shaped areas of resorption tis- sue; pollen tetrads with viscin strands. Stigma discoid-capitate, with 4 (5) grooves; style longer than the petals, thickened and incurved above, soon deciduous; ovary superior, 4(5)-locular, on a 4(5)-lobed disc (gyno- phore?); placentae large, the ovules 5—8 in each locule, strongly flattened, sessile. Fruit a depressed-globose 4(5)-lobed capsule (8-12 mm. broad), sometimes with a very short gynophore, septicidally dehiscent from above, the valves and enlarged placentae persistent; seeds 1-several in each locule (many ovules aborting), discoid, winged all around, strongly com- pressed between placenta and ovary wall, 3-4 mm. across, falling from the placenta. Type species: FE. racemosa Muhl. ex Ell. (Named for Stephen Elliott, 1771-1830, distinguished South Carolina botanist. ) One to three very distinct species, depending upon the generic con- cept; Elliottia racemosa in our area, two others in Japan (see below). Elliottia racemosa, with numerous showy panicles or racemes of 40-80 white flowers produced in June or early July, is one of the handsomest, 1961 | WOOD, GENERA OF ERICACEAE 21 as well as one of the rarest, of our native shrubs. It apparently is largely confined to isolated colonies in mediacid sandy soils (Norfolk sand) of the Coastal Plain in an area roughly bounded by Aiken County, South Carolina (where it is now presumed extinct), and Columbia, Screven, and Coffee counties, Georgia. Although usually seen as a shrub where sub- jected to burning, in its best development the plant is a small tree to 9 m. tall and grows in moist but always well-drained sands in mixed woods within a few hundred feet of streams. The plant shows a considerable range of tolerance, however, and some colonies are in rather dry locations, on oak ridges, in Pinus australis savannas, or in Pinus-Quercus woods. (In cultivation, given an acid and well-drained soil and an eastern or southern slope. it is hardy at least as far north as Philadelphia. ) . Elliottia. a-h, E. racemosa: 3; c, stamens, inner and outer views diagrammatic, * 10; f, five-locular ovary in cross section, showing divided placentae, sessile ovules, & 10; g, open 4-carpellate capsule after shedding of seeds. X 4; h, inner and outer views of thin-winged seed, X 4. At a number of the known stations Elliottia apparently seldom or never fruits, and a number of misconceptions concerning fruits and seeds have remained in the literature. (See Drude, Copeland; but see Wherry for an accurate description.) In at least two localities fruits are produced, how- 22 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII ever, and reproduction by seedlings has been found. (It is perhaps sig- nificant that these stations are relatively moist.) The lack of fruiting has led to the supposition that each colony is clonal and self-sterile, but no evidence of root-suckering or clonal connections between plants has been produced. (The plant suckers abundantly from the stem-base after burning or cutting, however.) Chromosome counts, further observations on the pollination mechanism, and artificial crossing experiments are sug- gestive lines of approach to the problem. Copeland observed in preserved material that the pollen is shed in the pendent bud, falling in four cylindri- cal masses from each anther and lodging in the apex of the closed corolla beyond the stigma. He supposed that as the style elongates the pollen is pressed against the spoonlike tips of the petals by the stigma, and then exposed as the flower rises to the horizontal and the petals recurve.’ Lack of unanimity has characterized the generic treatment of Elliottia racemosa and the two Japanese species, both of which have five sepals, three petals, six stamens, and a three-locular ovary. Tripetaleia paniculata Sieb. & Zucc. (E. paniculata (Sieb. & Zucc.) Benth. & Hook.), a shrub of Kyushu, Shikoku, Honshu, and Yezo, resembles E. racemosa in many respects, but for the smaller, more complex panicles of smaller flowers, the trimerous symmetry, the development of a definite gynophore, and the lack of a well-developed marginal wing on the seed; it may well be in- cluded in Elliottia, Tripetaleia bracteata Maxim. (E. bracteata (Maxim.) Benth. & Hook.), a smaller shrub of Honshu and Yezo, is considerably more modified, with comparatively large bracts, reduced, purely racemose inflorescences, subfoliose sepals, an obscure gynophore, and different seeds. Stapf regarded E. racemosa and T. paniculata as congeneric, but treated the third as Botryostege bracteata (Maxim.) Stapf, thinking it closer to Cladothamnus pyroliflorus Bong. (distributed from Oregon to Alaska) 24 number of the ecological data above were corroborated or supplied by Miss Clermont H. Lee, of Savannah, Georgia, who has also very kindly made additional field observations on the flowers and their opening. She has observed that the buds are upright at first, becoming pendent when about 4—-4.5 mm. long. At 5-7 mm. the pollen has started to fall from the stamens onto the underside of the stigma. Eventu- ally it works its way under and around the stigma, there being a space between it and the spoonlike petal tips. At the time of opening of the flower, most of the pol- 14 mm. long) has assumed a horizontal position. Some pollen may fall from the spoonlike tips of the petals and some from the stigma as the petals reflex. (Open attempt by Dr. Joab Thomas to secure meiotic chromosomal counts from material from this colony kindly fixed by Miss Lee was unsuccessful, for, although the buds were of the proper size, the pollen mother cells, for reasons unknown, showed a com- plete lack of divisions. 1961] WOOD, GENERA OF ERICACEAE a than to Elliottia. Copeland and Cox arrived at opposing conclusions re- garding Elliottia and Tripetaleia, but neither was able to study material of Botryostege. Whatever the taxonomic treatment followed, all four taxa are clear-cut, relict species of great morphological and phytogeographic interest. Copeland characterized his tribe Cladothamneae (which included only these plants) by the free petals, by the anthers opening throughout their length by resorption tissue, by viscin strands among the pollen tetrads, by endosperm without chalazal haustoria, and by the characteristic manner of discharging the pollen. Ledum L., associated with these genera by Drude, was removed to a position near Rhododendron. REFERENCES: See also under family references CopELAND (1943, pp. 533-559), Cox (Eri- cales I, pp. 230, 231), and IkuseE (1954). Harper, R. M. Notes on Elliottia racemosa. Pl. World 5: 87-90. pl. 12. 1902. . Two new stations for Elliottia. Ibid. 6: 60. 1903. —.. Ellhiottia racemosa again. Torreya 3: 106. 1903. : A phytogeographical ote of the Altamaha Grit Region of the coastal plain of Georgia. Ann. N. Y. Acad. Sci. 17: 1-415. 1906. [Elliottia, 187. pl. 19, fig. 2; pl. 20.) Henry, M. G. Elliottia racemosa. Natl. Hort. Mag. 20: 223-226. 1941. [In- cludes photographs of flowering and fruiting plants. KnicuT, W. A. A rare American shrub. Arnold Arb. Bull. Pop. Inform. IV. 6: 7-13. 1938. [Notes on ecology and details of Elliottia.] SARGENT, C. S. Elliottia racemosa. Gard. Forest 7: 206. fig. 37. 1894. . Elhiottia. Silva N. Am. 14: 29-32. pl. 712. 1902. SEALY, J. R. Elliottia racemosa. New FI. Silva 10: 154-164. figs. 49, 50. 1938. SMALL, - K. The rediscovery of Elliottia. Jour. N. Y. Bot. Gard. 2: 113, 114. 1 90 STAPF, 0. ae racemosa. Bot. Mag. 138: pl. 8413. 1912. Botryostege: a new genus of Ericaceae. Kew Bull. 1934: 191-195. 1 pl. 1934. [Includes discussion of relationships of Elliottia, Tripetaleia, and Botryostege. | oar H.W. “Rescuing Elliottia.” Bartonia 9: 11-15. 1926. new colony of Elliottia. Tbid. 10: 24-27. 1929. Wuerry, E. T. Discovery of Elliottia seed. Jbid. 17: 51. 1936. 3. Kalmia Linnaeus, Sp. Pl. 1: 391. 1753; Gen. Pl. ed. 5. 185. 1754. Evergreen or rarely deciduous shrubs or small trees to 12 m. Winter buds with about 2 outer scales. Leaves alternate, opposite or whorled, entire, usually coriaceous, petioled or rarely sessile, glabrous, pubescent, hirsute or glandular. Flowers in terminal or axillary corymbs or solitary in the axils of leaves and forming a loose terminal “‘raceme”’; each flower in the axil of a leaf or (in those species with corymbs) of a small, thick bract, and with a pair of bractlets at the base of the slender pedicel. Flowers 5-merous; base of receptacle impressed. Calyx synsepalous, deeply S-parted, persistent, or deciduous in fruit. Corolla sympetalous, saucer- shaped, with a short, narrow tube, shallowly 5-lobed, and with 10 pockets 24 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII in which the anthers are held. Stamens 10, the filaments slender, the anthers short, unappendaged, opening by apical slits, without white dis- integration tissue; anthers lodged in the pockets of the corolla in the bud so that the filaments are strongly recurved and held under tension by the expansion of the corolla and spring up suddenly when the corolla is touched; pollen tetrads with or without viscin strands. Style slender, straight; stigma flat, unexpanded, 5-grooved; ovary superior, 5-locular, with massive placentae and numerous ovules, on a conspicuous disc. Cap- sule subglobose to globose-ovoid, septicidal, 5-valved, opening from above, the style persistent on the central axis. Seeds numerous, minute, essen- tially cylindrical, with a loose coat. (Including Kalmiella Small.) Lrcto- TYPE spEctES: K. latifolia L.; see N. L. Britton, N. Am. Trees 756. 1908. (Dedicated to Pehr Kalm, 1716-1779, student and friend of Linnaeus, who travelled and collected in North America from 1748 to 1751.) About six species of North America, including K. polifolia Wangenh. (consisting of vars. polifolia and microphylla (Hook.) Rehd.), of cold bogs and alpine meadows from Labrador to Alaska, south to Pennsylvania, Minnesota, Colorado, Montana, and California; K. ericoides C. Wright ex Griseb. (including K. aggregata (Small) H. F. Copel. and Kalmiella simulata Britt. & Wilson, based on variations in pubescence and compact- ness of inflorescence), of Pinar del Rio and the Isle of Pines, Cuba; four species in our area. Kalmia latifolia, mountain laurel, widely distributed in acid _ soils, largely in deciduous woodland, from western Florida to Louisiana, north- ward to New England, New York, Ohio, and Indiana, is a shrub, or occa- sionally a tree to 10-12 m. tall, with elliptic or elliptic-lanceolate, mostly alternate leaves 5-10 cm. long and showy terminal corymbs of white to pink flowers 1.5—3 cm. across, the petioles glandular-pubescent. The plant is a popular horticultural subject and a number of variant forms have been named. (See Rehder, Holmes.) Plants with the pedicels less markedly glandular have been named var. laevipes Fern. Kalmia angustifolia L. var. carolina (Small) Fern. (K. carolina Small), a shrub to about 1 m. tall, with short-petioled oblong to lanceolate mostly opposite or ternate leaves, pubescent beneath even at maturity, axillary to terminal corymbs of deep pink or purple flowers about 1 cm. across, ranges from south- eastern Virginia southward through eastern North Carolina to eastern South Carolina, and occurs in scattered localities in western Virginia (Car- roll County), eastern Tennessee (Johnson County), western North Caro- lina. and northeastern Georgia (Rabun County). In southeastern Vir- ginia it intergrades with the northern var. angustifolia (southeastern Vir- cinia northward through eastern Pennsylvania and New Jersey to Labra- dor and Hudson Bay and west to northern Michigan) which differs in having the leaves glabrous beneath at maturity. Kalmia cuneata Michx. is a distinctive deciduous shrub with oblong-obovate, cuneate leaves, glandular beneath, and almost umbellate corymbs of white flowers about 1.5 across produced from the axils of leaves of the preceding year. One 1961 | WOOD, GENERA OF ERICACEAE 25 of the rarest of American shrubs, it is known from wet thickets or shrub bogs in only a few localities in southeastern North Carolina and from an early collection of Nuttall between Camden [Kershaw County], South Carolina, and Statesville | Iredell County], North Carolina. Kalmia hirsuta Walt. (Kalmiella hirsuta (Walt.) Small), a straggly shrub with thin, hirsute, alternate, persistent leaves, of the Coastal Plain from southeastern North Carolina (?) and South Carolina southward to Flagler, Putnam, and Alachua counties, Florida, and west to southernmost Alabama; and K. ericoides, a low, revolute-leaved shrub of Cuba. differ from other species in the deciduous calyx (cf. Lyonia) and in having the flowers borne singly in the axils of ordinary-sized or somewhat reduced foliage leaves (instead of bracts). The plants are otherwise entirely con- sistent with Kalmia both morphologically and anatomically and there seems to be no real justification for the removal of these species (as Kal- miella Small) from Kalmia. The six species of Kalmia are quite distinct morphological entities which are separated ecologically and/or geograph- ically and perhaps genetically. Artificial hybrids do not seem to have been attempted, and no natural hybrids are known. The chromosomal data reported by Hagerup (2 = 24 rather large chromosomes in K. latifolia and 2n = 48 chromosomes about half as large in K. polifolia) suggest an interesting pattern of polyploidy which should be investigated further. The most distinctive generic feature of Kalmia is the pollination device. In the flower bud the ten anthers are pushed upward by the growth of the filaments into corresponding pockets in the corolla. As the corolla opens and expands the anthers are held by the pockets and the filaments are bent backward under tension. The weight of a suitably heavy insect displaces the corolla, and frees one or more of the anthers; the filaments snap sud- denly upward, showering the visitor with pollen. In K. latifolia viscin strands occur in the pollen, which is hurled like a sticky net, while in K. angustifolia var. angustifolia (and probably K. polifolia) the pollen is powdery and dusts the visitor. The leaves of at least Kalmia latifolia and K. angustifolia are poisonous to many mammals when eaten. The toxic principle is thought to be an- dromedo-toxin which is presumably the same as that in other poisonous Ericaceae (including evergreen Rhododendron species, Ledum, Leucothoé, Lyonia, and Andromeda). A mycorrhizal association, without which growth does not occur, has been demonstrated in K. latifolia. Kalmia, Leiophyllum, and seven other small, definitely distinct and rather isolated genera are sometimes segregated as the tribe Phyllodoceae Drude. All are evergreen plants, mostly with ericoid or buxoid leaves: in those with pentamerous flowers the median petal is abaxial; and the seeds generally lack appendages (except in Kalmia). As defined by Copeland the tribe is characterized primarily by the presence of resorption tissue involving the epidermis of the anther and by the absence of filaments among the pollen tetrads. The latter characteristic does not hold. how- ever, for viscin filaments have been reported in Phyllodoce, Loiseleuria. and Kalmia. The closest relative of Kalmia both morphologically and anatomi- 26 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII cally appears to be Rodothamnus Reichenb. (including Kalmiopsis Rehd. iP comprised of two species, R. Chamaecistus Reichenb., of the eastern Alps, and R. Leachianus (Henderson) H. F. Copel., of the Siskyou Mountains, Oregon. Rhodothamnus, Kalmia, and Phyllodoce Salisb. are regarded as relatively primitive within the group. REFERENCES: Under family references see also BELL & BURCHILL, COPELAND (1943), Cox (Ericales 1), IkuseE, KEARNEY, MATTHEWS & mag PeLTRISOT, and STANLEY. Britton, E. G. Wild plants ‘needing protection. “American or mountain laurel” (Kalmia latifolia L.). Jour. N. Y. Bot. oa 14: 121-123. pl. 117. 1913. . A freak of the mountain he Ibid. 26: 187-190. 1925. [K. latifolia f. polypetala (Nicholson) Reh Brown, N. E. Kalmia cuneata. a Mag. 136: pl. 8319. 1910. Burtrick, P. L. Connecticut’s state flower, the mountain laurel, a forest plant. Marsh Bot. Gard. Publ. 1. 28 pp. Curtis, W. Kalmia hirsuta. Ibid. 4: pl. 138. 1795; K. latifolia, 5: pl. 175. K. glauca = K. polifolia), pl. 177. 1792; K. angustifolia, 10: pl. 331. 1796. Entows, E. M. A. A leafblight of Kalmia latifolia Jour. Agr. Res. 13: 199- 212. pls. 14-17. 1918. | Phomopsis kalmia FERNALD, M. L. A new southern Kalmia. Rhodora 42: 53, 54. 1940. [K. lati- folia var. laevipes Fern. | FLEMER, Wo., III. The propagation of Kalmia latifolia from seed. Bull. Tor- rey Bot. Club 76: 12-16. 1949. [Mycorrhizal grea Forses, E. B. Mountain laurel and rhododendron a ood sad so white tailed deer. Ecology 12: 323-333. 1931. [K. latifolia ego Haptey, W. J., and H. H. Haven, Jr. Alabama’s “sheep Kill” ei Kalmia latifolia L.—A preliminary study. Jour. Am. Pharm. Assoc. Sci. Ed. 35: 340-342. 1946.* Hoimes, M. L. Kalmia, the American laurels. Baileya 4: 89-94. 1956. [ In- cludes cultivars of K. latifolia and K. angustifolia; K. ericoides omitted. | LoveLL, J. H. The pollination of Kalmia angustifolia. Rhodora 36: 25-28. 1934. Marsu, C. D. The poisonous laurel. (Abs.) Jour. Wash. Acad. Sci. 20: 350, 351. 1930. [K. latifolia and K. angustifolia. | and A. B. CLawson. Mountain-laurel (Kalmia latifolia) ao sheep laurel (Kalmia oo as stock-poisoning plants. U. S. . Agr. Tech. Bull. 219: 1- Reuper, A. Note on . rte of Kalmia latifolia. Rhodora 12: 1-3. 1910. [Corolla shape and color, leaf shape variants. SARGENT, C. S. Kalmia. Silva N. Am. 5: 137-142. pls. 236, 237. 1893. ——. Kalmia maa var. myrtifolia. Gard. Forest 8: 315. fig. 44; K. cuneata. 434. fig. 1895. Trevett, M. F. pate of lambkill (Kalmia angustifolia) in agevere —— fields. Northeast. Weed Control Conf. Proc. 10: 193-198. 19 4. Leiophyllum (Persoon) Hedwig filius, Gen. Pl. 313. 1806. Evergreen shrubs, prostrate to erect and to about 1 m. tall with crowded, small (to 1.5 cm. long), leathery, alternate to opposite, oblong to orbicu- lar leaves, the petioles short, articulated at the base. Inflorescences ter- 1961] WOOD, GENERA OF ERICACEAE 27 minal umbel-like simple or compound corymbs, the flowers borne singly in the axils of reduced leaves, each with 2 small lateral bracts at the base of the slender pedicel. Flowers 5-merous; base of receptacle impressed. Calyx lobes 5, free nearly to the base, persistent. Corolla of 5 spreading, equal, white or pinkish petals. Stamens 10, conspicuously exserted, equal- ing or exceeding the petals; filaments filiform, glabrous; anthers small, almost globular, opening by 2 lengthwise slits, unappendaged, blue to purple or pink, without disintegration tissue; pollen tetrads without viscin strands. Stigma not expanded; style filiform, about as long as the stamens, straight, tapering into the more or less glandular 3- or 2(rarelv 4 or 5)- locular, superior ovary which is seated on a conspicuous disc; ovules numerous, axile. Capsule ovoid. 3- or 2(rarely 4 or 5)-valved from the apex, the central axis persistent, pointed, the valves nearly smooth to sharp-tuberculate on the back, the style adhering to one valve or broken away. Seeds numerous, brown, minute, plump, unappendaged, the seed coat reticulate. (Ledum?+ Leiophyllum Pers. Syn. Pl. 1: 477. 1805; Dendrium Desv. Jour. Bot. Desvaux II. 1: 36. 1813; Ammyrsine Pursh; Fischeria Sw.) Type species: Ledum thymifolium Lam. = Leiophyllum buxifolium (Berg.) Ell. (Name from Greek, leios, smooth, and phyllon, leaf, in allusion to the shining, leathery leaves.) — SAND-MYRTLE. A single variable species (sometimes treated as two or three) of open habitats on acid rocks and sands from the higher summits of the Smoky and Blue Ridge mountains of North Carolina, Tennessee, and northern- most Georgia, via several disjunct stations to the coastal plain of south- eastern North Carolina and northeastern South Carolina, and of the pine barrens of southern New Jersey. Most of the characters used by Small (habit, petal shape, disc, style/ Ovary proportions, capsule valves) in distinguishing three species in the genus do not hold, and even those of glandular pubescence, leaf shape and leaf arrangement adopted by Camp in delimiting two species (one com- posed of two varieties) are not altogether constant. The course of Rehder, who recognized a single species, L. buxifolium, composed of three varieties, seems to be most in line with the information presently available. Varietas Hugeri (Small) Schneid. (L. Hugeri (Small) K. Schum.), considered by Camp to be the basic entity, is a decumbent to erect shrub to about | m. tall, primarily of the Coastal Plain of southern New Jersey and the Caro- linas (but with a few disjunct stations inland in the latter states and eastern Kentucky) with glandular hairs on the pedicels and with mostly alternate leaves which tend to be oblong and slightly apiculate. Varietas prostratum (Loud.) Gray (L. Lyonit (Sweet) Sweet [nomen subnudum] sensu Small, Camp, e¢ al.), prostrate to almost 1 m. high depending upon the habitat, and with glandular pedicels but mostly opposite, elongate- oval to orbicular leaves which are rarely apiculate, is confined principally to exposed ledges and summits in the higher mountains of North Caro- lina, Tennessee, and northernmost Georgia. Varietas buxifolium, pri- marily of southern New Jersey, has the leaves and leaf arrangement of 28 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII var. Hugeri but has glabrous pedicels. It should be noted that vars. buxi- folium and Hugeri may occur together in New Jersey; that at least some plants with glabrous pedicels occur in the Carolinas with var. Hugeri; that one New Jersey collection with glandular pedicels has predominantly opposite leaves; that all of the mountain plants have glandular pedicels but that in herbarium materials leaf arrangement is somewhat variable and there are possible evidences of ecological influences; and that leaf shape is far from a dependable characteristic in this genus. As Camp has noted, the type of var. Hugeri, from Table Mountain, Pickens County, S. C, is, unfortunately, biologically closer to var. pros- tratum than to the glandular-pediceled plant of the Coastal Plain; the name is retained tentatively, however, in the sense of Camp, until further information on variation and geography is obtained. Ecological and population studies are very much in order. Of particular interest in an understanding of the genus is the relationship of alternate vs. opposite leaves to genetic and ecological factors. Transplant experiments and ob- servations on sprouting after fire and the effect of rapidity of growth on leaf arrangement should provide pertinent information. Little seems to have been recorded of the biological features of the genus, except that the numerous small flowers (April to June) are very attractive to insects. Leiophyllum has been associated with Ledum in the past, largely on ac- count of the distinct petals, but the relationships of the latter appear to be instead with Rhododendron. In wood anatomy Leiophyllum lies be- tween Ledothamnus Meissn. and Loiseleuria Desv. REFERENCES: Under family seipeci see also COPELAND (1943, pp. 567-569), Cox (Eri- cales I, p. 233), and STANL Camp, W. H. Studies in the Ericales III. The genus Leiophyllum. Bull. Tor- rey Bot. Club 65: 99-104. 1938 FERNALD, M. L. Lesaphyilum versus Dendrium. Rhodora 29: 225-227. 1927. Fox, W. B., R. K. Goprrey, and H. L. BLromauist. Notes on vaaiachairndn of North Carolina plants — IIL Ibid. 54: 165-182. 1952. [176, var. Hugeri reported from Gaston Co., | Hooker, J. D. Leiophyllum ane Bot. Mag. 110: pl. 6752. 1884. [See also Epwarps, Bot. Reg. 7: pl. 531. 1821. SMALL, J. K. Shrubs and trees of the southern states. —IV. Bull. Torrey Bot. Club 28: 356-361. 1901. [360, 361, Dendrium Hugeri described. | 5. Menziesia J. E. Smith, Pl. Icon. Ined. 3: pl. 56. 1791. Deciduous shrubs (to about 2 m.) with the general aspect of Rhodo- dendron § Anthodendron. Branches erect or spreading, the leaves alter- nate, crowded at the ends, the blades thin, broadly elliptic to obovate, apiculate, often strigose, ciliate, entire, gland-tipped, short-petioled. Win- ter buds swollen, with several outer scales. Flowers 4[5]-merous, appear- ing with the leaves, in sub-umbellate (to short-racemose) clusters termi- nating the branches of the previous year, nodding, the pedicels slender, be- 1961 | WOOD, GENERA OF ERICACEAE 29 coming erect in fruit. Calyx small, flat or saucer-shaped, circular or shal- lowly 4[5]-lobed, glandular-ciliate, persistent. Corolla yellowish white at the base, rose at the tip [or greenish to purplish], campanulate or urceolate [or cylindric and sometimes irregular], the tube stout, [often] swollen, the lobes 4[5], rounded and much shorter than the tube, often erect. Stamens 8 [or 5 or 10], included [or exserted]; filaments subulate, flattened; anthers narrow, linear to linear-sagittate, unappendaged, with- out disintegration tissue, opening (through collapse tissue) by short slits broadened above; pollen tetrads with viscin strands. Stigma truncate; style columnar, included; ovary 4[5]-locular, superior, obscurely lobed, glandular-hairy [glabrous or pubescent], commonly ovoid; ovules many, axile, elongate; disc 8[10]|-lobed. Capsule ovoid, oval or obovoid, firm, rather thick-walled, septicidally 4-valved; the style broken or persistent on the central axis; seeds numerous, linear, abruptly pointed or tailed. TYPE spPEcIES: M. ferruginea J. E. Sm. (Named in honor of Archibald Menzies, 1754-1842, who served as surgeon-botanist on Vancouver’s great expedition of survey, 1790-1795, in the course of which he collected the type species. A small genus of disjunct distribution including about five or six species: three or four in Japan and Sakhalin (Menziesia ciliicalyx (Miq.) Maxim. vars. ciliicalyx and multiflora (Maxim.) Makino, M. purpurea Maxim., and M. pentandra Maxim.); M. ferruginea vars. ferruginea and glabella (Gray) Peck in western North America (Alaska to northern California, Alberta, Idaho, and Wyoming) ; and M. pilosa (Michx.) Juss. in the eastern United States, in rocky upland deciduous woods from Pennsylvania south- ward to West Virginia, Virginia, North Carolina, northern Georgia, and eastern Tennessee. Menziesia pilosa is easily recognized by the azalea-like aspect, the ter- minal clusters of nodding, tetramerous, campanulate, yellowish-white to orange-pink flowers about 6-7 mm. long, and the glandular or glandular- setose, septicidal, ovoid capsule 3-4 mm. long. It is closely related to the western American species, from which it differs primarily in details of pubescence of bud scales, leaves, pedicels, calyx, filaments, and ovary and in shape and size of fruit. Both eastern and western species are far from stable in most of these features and the full range of variation has not been adequately described in either group (e.g., the undersurfaces of the leaves of M. pilosa vary from densely pilose with short, nonglandular hairs in much of its range to almost glabrous with scattered glandular hairs along the veins in some high-mountain localities in North Carolina). The American species are characterized by reduced calyces. regular corollas, and tetramerous symmetry, while the Japanese have more promi- nent calyces, more or less irregular corollas (some almost cylindric and about 1.5 cm. long), and prevailingly pentamerous symmetry (although with a tendency toward tetramery and reduction in the number of sta- mens). The genus presumably stands nearest Rhododendron from which it differs most obviously in the included stamens and the opening of the 30 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII anthers by short slits, rather than by pores. In Tsusiophyllum Maxim. (monotypic, Japan), in which the stamens are ten, the corolla 5-lobed, and the ovary 3-locular, the anthers open by longer slits, so that in this respect Menziesia stands between Tsusiophyllum and Rhododendron. In wood anatomy, however, Tsusiophyllum is more highly specialized than either and is on a level with Ledum L. which was placed near Rhododendron by Copeland. REFERENCES: See also under family references CopELAND (1943, pp. 589-592), Cox (Eri- cales I, p. 243), and PELtRIsor. Cormack, R. G. H., and A. L. GorHam. Effects of exposure to direct sun- light upon the development of leaf structure of two deciduous shrub species. Canad. Jour. Bot. 31: 537-541. 1953. [M. glabella and Lonicera glauces- cens, Sims, J. Menziesia ferruginea. Bot. Mag. 37: pl. 1571. 1813. [Plant figured is M. pilosa. | 6. Rhododendron Linnaeus, Sp. Pl. 1: 392. 1753; Gen. Pl. ed. 5. 1754, “Rhododendrum.” Evergreen or deciduous shrubs (rarely trees) [sometimes epiphytes], with alternate, entire (rarely ciliate-serrulate), glabrous to tomentose, lepidote, or strigose leaves. Buds with several to many imbricate scales. Flowers pedicellate, usually in terminal corymbs [sometimes solitary or few, rarely from lateral buds]. Calyx 5-parted, various, often very small [rarely 6-10-parted], persistent. Corolla white to yellow, pink, purple or deep red, rotate to campanulate or funnelform [sometimes tubular], usu- ally irregular, with a 5{rarely 6-10]-lobed limb [rarely some of the lobes divided to the base], the median lobe adaxial; deciduous. Stamens 5—7— 10{—20], usually exserted from the corolla [except in some tropical mon- tane groups], the filaments usually unequal, the anthers unappendaged, muticous, without white disintegration tissue, opening (through collapse tissue) by apical pores; pollen tetrads with viscin strands. Stigma capi- tate; style slender, curved; ovary 5|—10—20]-locular, superior, free from the calyx; ovules numerous in each locule. Fruit a septicidal capsule, usually ovoid to oblong; seeds numerous, minute, flat and more or less winged to fusiform |or caudate]. Basic chromosome number 13. (Includ- ing Azalea L. sensu Desv., Azaleastrum Rydb., Biltia Small, Hymenanthes Blume, Rhodora L.) LEcTOTYPE spEcIES: R. ferrugineum L.; see N. L. Britton, N. Am. Trees 752. 1908. (Name from Greek, rhodon, rose, and dendron, tree; the ancient Greek name of Nerium Oleander L.) — Ruopo- DENDRONS, AZALEAS. A genus of 800 or more species of the colder and temperate parts of the Northern Hemisphere, centering in eastern and southeastern Asia, and especially the Himalayas, extending southward on the high mountains of Malaysia, the Philippines, and New Guinea to Queensland, and with a 1961] WOOD, GENERA OF ERICACEAE 31 much smaller center of development in eastern North America, a few species in the mountains of Europe, and R. lapponicum (L.) Wahlenb. (2n = 26, 52) circumboreal. About 21 species occur in continental North America, including 17 species in four sections in the southeastern United States. Most authors with more than a regional viewpoint have maintained Rhododendron as a polymorphic genus difficult to divide into large, well- marked natural units and susceptible of only artificial division into genera of convenience. Both Copeland and Cox have proposed dismemberment of the inclusive genus on anatomical grounds, but very few (including almost none of the taxonomically critical) species have been examined in this respect. The subgeneric classification is still in a state of flux, and from four to eight subgenera and as many as 20 sections or some 43 in- formal ‘‘series” have been recognized. The conservative course of empha- sizing morphological and genetic similarities, rather than differences, is followed here and the subgeneric categories are taxonomically those of Rehder. The species of the southeastern United States fall into four well- marked groups which have been placed in three genera (Rhododendron, Biltia, Azalea). The relationships of these are not directly with each other but with species elsewhere, and the much greater diversity of the Asiatic species bridges the discontinuities seen on a purely regional basis. Subgenus RHODODENDRON (subg. Eurhododendron Drude) comprises evergreen [or rarely deciduous] species with flowers produced in terminal umbel-like racemes on shoots of the preceding year, the stamens 5-20. Two sections are represented in our area. Section Ponticum G. Don (§ Leiorhodium Rehd.; Hymenanthes Blume) has glabrous to tomentose persistent leaves, |[glabrous,] glandular or hairy ovaries, and 10-20 sta- mens. Rhododendron catawbiense Michx. and R. maximum L. (both 2n = 26) occur in the eastern United States, the former largely in the southern Appalachians. Most closely related to R. ponticum L. (2n = 26). of Europe and Asia Minor, and to R. macrophyllum D. Don ex G. Don (2n = 26), of California to British Columbia, the two are distinctive morphologically (leaf shape, flower size and color, indument of ovary, etc.), ecologically, and seasonally, hybridizing occasionally (to produce R. wellesleyanum Waterer ex Rehd.) when ecological and seasonal barriers break down. Rhododendron catawbiense is interfertile with a number of other species of the section and has been used extensively in the pro- duction of hardy ornamental hybrids. Section RHODODENDRON (§ Leipipherum G. Don), with leaves and ovary lepidote with entire-margined, patelliform scales, and stamens 10, is rep- resented by three closely related magenta- to white-flowered taxa (sub- sect. Caroliniana Sleumer) ranging entirely within our area. Rhododendron carolinianum Rehd. (2m = 26) occurs mostly at higher altitudes in the mountains of North Carolina, Tennessee, South Carolina, and northeast- ern Georgia. Although usually with pink or white corollas, a clear yellow form (f. /utewm Frisbie) has been described from a cultivated plant pre- sumably collected in western North Carolina. Rhododendron minus a2 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Fic. 2. Rhododendron. a-e, R. carolinianum: a, flowering branchlet. « 1%; x 4 2 flower, X %; c, flower. lateral view, corolla and four stamens removed. ee 1s : : 2 A view, corolla and three stamens removed, * 1; i, branchlet with fruits of two seasons, flower buds and vegetative bud, & %; j, capsule showing dehiscence by l-q sepiicidal slits, X 11%; k. seed. hilum at lower ba ca. X 20. JR. lee oe l, fiowering branchlet, * 14; m, flower, * anther, inner view. stigma, X 10; p, capsule. * 11%; q, seed, hilum : Rare end, ca. X20. 1961 | WOOD, GENERA OF ERICACEAE 33 Michx. (R. punctatum Andr., R. Cuthbertii Small) (2n = 26) is a plant of lower altitudes, primarily in the Piedmont, in scattered localities from central Alabama to southwestern Georgia and northward to North Caro- lina, while R. Chapmanii Gray is of much more restricted distribution, occurring in scattered colonies in pinelands and on bluffs in the Apalachicola region of western Florida and in northeastern Florida. The latter two are similar in conformation of the corolla, while R. minus and R. carolinianum are similar in leaf shape and habit (although the former is a taller, more straggling and later flowering shrub than the latter). The three appar- ently are largely separated geographically and ecologically, although the ranges of R. carolinianum an . minus seem to approach one another closely in northeastern Georgia and some intergradation may occur. Subgenus ANTHODENDRON (Reichenb.) Rehd.* includes deciduous to evergreen species with the flowers in leafless corymbs at the ends of shoots of the previous season, the stamens 5-10. Four sections are recognized, two in our area, two entirely Asiatic. Section Tsutsust Sweet ($ Tsut- sutsi G. Don) includes many of the commonly cultivated azaleas: R. in- dicum (L.) Sweet, R. Simsii Planch., R. obtusum (Lindl.) Planch., among others, as well as many artificial hybrids. Section RuHopora (L.) G. Don, with flowers from a terminal bud, the leafy shoots from separate lateral buds and stamens 10 or : includes five species: three of Asia; R. canadense (L.) Torr. (2n = and unique in the genus in the deeply divided corolla) entirely aris the ene area to the north of our range; and R. Vaseyi Gray (Biltia Vaseyi (Gray) Small) (2” = 26), restricted to a limited area of western North Carolina t 900-1600 m. altitude. Rhododendron Vaseyi is quite distinct in the pale-pink, 2-lipped, rotate-campanulate corolla with very short tube, the uppermost corolla lobe exterior in the bud, and the stamens 7 (5 or 6). but, as Rehder (1926) has shown, Biltia Small cannot be maintained on these bases. The species is isolated among the American taxa and is probably most nearly related to the Japanese R. Albrechtii Maxim. Both species reportedly have been crossed with R. Schlippenbachii Maxim. (very similar in floral structure, but of the Asiatic sect. BRACHYCALYX Sweet *Seithe-von Hoff (1960, p. 312) adopts subg. Azalea (L.) Planchon (FI. des Serres Oe 5c 1893 - as the correct name for this taxon, maintaining ie Rend (Jour. Arnold Arb. 2: 156-159. 1921) was incorrect in typifying Azalea L. by procum- bens L. (Loiseleuria procumbens (L.) Desv.) and that A. pontica L. (ehadaienara’ luteum Sweet) should be rai: instead as lectotype species. However, Rehder’s arguments seem relevant, and, as he pointed out, Azalea was effectively typified by Salisbury who removed the ie species to Rhododendron, Stee only A. procum- bens in Azalea. Since this species is also the type of the conserved name Loiseleuria, Azalea L. is automatically unavailable. Planchon, moreover, published this combina- tion as a section, rather than a subgenus. It is mentioned only briefly as “un sous- genre auquel nous réserv ons le nom sous-genérique Azalea. IO NEV EE, sect. een Tsutsia,’ receives ‘auilar ireatutent: Planchon noting that it is anne af he “Azalées de l’Inde, encore plus clairement un simple sous-genre du type Rhododendron pour lequel nous prosposerons le nom de 7sutsia.” The parallel is clear, and “sous- genre” is to be translated as “sectio” in this instance. 34 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII [§ Sciadorhodion Rehd. & Wils.] in which the flowers are from terminal buds with leafy shoots from the axils of the lower scales of the same bud). Section ANTHODENDRON (§ Pentanthera G. Don), with the inflores- cence from the terminal bud and leaves from lateral buds below, stamens 5, leaves mostly deciduous, and pubescence, if present, setose or the hairs flattened, includes the European R. luteum Sweet (R. flavum G. Don), the Japanese R. japonicum (Gray) Suringar, the Chinese R. molle (Blume) G. Don, the western American R. occidentale (Torr. & Gray) Gray, and about 13 species centering in the southeastern United States but extend- ing northward to New England, southern Quebec, central New York and westward to Arkansas and eastern Texas. The group includes the red- to yellow-flowered R. Bakeri (W. P. Lemmon & McKay) Hume (including R. cumberlandense E. L. Braun), R. flammeum (Michx.) Sarg. (R. spe- ciosum Sweet), R. calendulaceum (Michx.) Torr., R. austrinum (Small) Rehd., and R. prunifolium (Small) Millais; the pink-tubed to pink- flowered R. roseum (Loisel.) Rehd., R. nudiflorum (L.) Torr., and R. canescens (Michx.) Sweet; and the essentially white-flowered R. ala- bamense Rehd., R. atlanticum (Ashe) Rehd., R. viscosum (L.) Torr., R. serrulatum (Small) Millais, R. oblongifolium (Small) Millais, and R. arborescens (Pursh) Torr. These basic species are distinctive, differing morphologically in characters of bud scales, corollas, leaves, and pubes- cence, but hybridization in many areas has so blurred the lines between species that precise identification of individual specimens is difficult or impossible. With the exception of R. calendulaceum (2n = 52), all are diploids (2n = 26) which produce largely fertile hybrids. However. con- siderable hybridization seems to have occurred between R. calendulaceum and diploid species; in most instances the hybrids are tetraploid. Dif- ferences in geographical distribution, ecology, and flowering time appear to be the chief barriers to hybridization, and wherever these overlap per- plexing hybrid swarms, sometimes involving three or more species, may occur. Only the late-flowering and isolated R. prunifolium, of southwest- ern Georgia and adjacent Alabama, is not known to hybridize with other species in the wild, but R. « gladwynense M. G. Henry, its artificial hybrid with R. serrulatum, is fertile and vigorous. Other combinations of allopatric species also produce fertile hybrids, and R. * gandavense (K. Koch) Rehd. is noteworthy as a showy artificial hybrid swarm involving at least R. luteum, R. calendulaceum, and R. nudiflorum and probably R. viscosum and other species. Chromosome numbers have been reported for about 360 species of Rhododendron. The species of subg. RHODODENDRON § PontTiIcuM, subg. AZALEASTRUM, and subg. ANTHODENDRON appear to be diploids (2” = 26), with the two exceptions noted above, but subg. RHODODENDRON § RHODODENDRON includes a number of series with varying degrees of poly- ploidy (about 44 known tetraploids, 22 hexaploids, 1 octoploid, 1 duo- decaploid), apparently correlated with altitude in the high mountains of Asia. A large number of species are in cultivation, especially in the milder 1961 | WOOD, GENERA OF ERICACEAE 35 climates of Great Britain and the northwestern United States, and many hybrids have been made both between closely related species and even those of different subgenera (e.g.. RHODODENDRON * ANTHODENDRON = subg. * AZALEODENDRON). All of the intersubgeneric hybrids appear to be sterile, however, and some are weak or inviable. Self-compatibility to -incompatibility have been recorded for various species (e.g., R. cataw- biense) and hybrids. The genus appears to be proterogynous throughout. The horticultural and technical literature on the genus is enormous. REFERENCES: The vast number of references has been reduced here primarily to those of either general interest or dealing specifically with the southeastern United States. See also Rhododendron Yearb. Roy. Hort. Soc. 1-7, 1946-1953, continued as Rhododendron and Camellia Yearb. and Quart. Bull. Am. Rhododendron Soc.; Rehder. Man. Cult. Trees Shrubs, Bibl. Cult. Trees Shrubs, and in Bailey, Stand. Cyclop. Hort.; also under family references CopELAND (1943, pp. 592- 619), Cox (Anatomy of Ericales. I, pp. 239-243), MATTHEWS & KNox (1926, pp. 245-247), PELtTRIsoT (1904). STANLEY (1931). ALEXANDER, E. J. Azalea viscosa glauca. Addisonia 11: 35, 36. pl. 370. 1926. Azalea calendulacea. Ibid. 13: 17, 18. pl. 425. Azalea arborescens. Ibid. 19: 17, 18. pl. 617. 1935. BALDSIEFEN, W. Deciduous azaleas from cuttings. Proc. Pl. Propagators Soc. 8: 172-175. 1958. Bowers, C. G. ria and azaleas. Their origins, cultivation and de- velopment. i-xv + 549 pp. Macmillan Co., New York. The Hee eat of pollen and oe strands in Rhododendron cataw- biense. Bull. Torrey Bot. Club 57: 285-314. pls. 11- Braun, E. L. The red azalea of the eee. see 43: 31 —35. 1941. [R. cumberlandense. | BRITTON, = - Wild plants needing protection. 6. “Wild Azalea” (Azalea nudi- flora Jour. N. Y. Bot. Gard. 14: 79-81. pl. 114. 1913. 14. Great laurel or rose ge (Rhododendron maximum L.). Ibid. 23: 137, 138. pl. 277 1922. Carn. S. A.. and J. D. O. Mriter. Leaf structure of Rhododendron catawbiense Michx. grown in Picea-Abies forest and in heath communities. Am. Midl. Nat. 14: 69-82. 1933 CrarKE, J. H. A selected list of Rhododendron literature. Quart. Bull. Am. Rhododendron Soc. 10: 163-168. 1956. CoccesHALL, R. Rooting Ghent azaleas under plastic. Am. Nurseryman 107(11): 7, 8, 29. 1958. (See also Arnoldia 20: 1-7. 1960.) [Method applicable to American deciduous species. | Coker, W. C. The distribution of Rhododendron catawbiense with remarks orm. Jour. Elisha Mitchell Sci. Soc. 35: 76-82. pls. 19-22. 1919. [Stations in the Piedmont of N. C.; also comments on R. maximum, R. minus. | Azalea atlantica Ashe and its variety luteo-alba n. var. Ibid. 36: 97-99. pls. 1.7. 1920 A remarkable new Rhododendron. Ibid. 51: 189. 190. pls. 53. 54. 1935. [R. Ashleyi = form of R. maximum. | CoviLte, F. V. The effect of aluminum sulphate on rhododendrons and other acid-soil plants. Rep. Smithson. Inst. 1926: 369-382. pls. 1-13. 1927. 36 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Cowan, J. M. A survey of the genus Rhododendron, Rhododendron Yearb. Roy. Hort. Soc. 4: 29-58. figs. 3, 6-19. ; The Rhododendron leaf: a study of the epidermal appendages. xi + 120 pp. Oliver & Boyd, London. 1950. CreecH, J. L. An embryological pena in the Rhododendron subgenus Antho- dendron Endl. Bot. Gaz. 116: 234-243. 1955. [Embryology of R. (§ Antho- dendron) japonicum, R. (§ oo mucronatum, and hybrid. Davipson, H., an . P. Watson. Terat signa? pag of photoperiod on Rhododendron catawbiense Michx. Proc. Am. . Hort. Sci. 73: 490-494. _ ‘Oo > 1959. Duncan, W. H. Ecological comparison of leaf structures of Rhododendron punctatum Andr. and the ontogony |sic| of the epidermal scales. Midl. Nat. 14: 83- 96. 1933. | R. carolinianum in the Smoky Mts., Tenn.] FERNALD, M. L. Some forms of Rhododendron atlanticum. In Another cen- tury of additions to the flora of Virginia. Rhodora 43: 619-624. pl. 692. 1941 Foster, A. S. Structure and behavior of the marginal meristem in the bud scales of Rhododendron. Am. Jour. Bot. 24: 304-316. 1937. [R. ponticum, R. x Loderi, R. grande Foster, H. Rhododendron carolinianum naturalized in New England. Rho- dora 49: 116. 1947. Gorpon, H. D. Mycorrhiza in Rhododendron, Ann. Bot. II. 1: 593-613. pl. 23. 1937 one A Rhododendron (Azalea) Vaseyi. Bot. Gaz, 8: 282. 1883. |Addi- nal notes on the species. See also Bowers, Natl. Hort. Mag. 15: 202, iat 205. 1936: SARGENT, Gard. Forest 1: 377. fig. 60. 1888; SKAN, Bot. Mag. 132: pl. 8081. 1906; SmaLt, Addisonia 16: 97, 98. pl. 525. 1931; SMITH, Bull. Torrey Bot. Club 15: 164, 165. 1888. } Harpikar, S. W. On Rhododendron poisoning. Jour. Pharmacol. Exp. Therap. 20: 17-44. 1922.* [See also ForBes under Kalmia. HARSHBERGER, J. W. Thermotropic movement of the leaves of Rhododendron maximum L. Proc. Acad. Nat. Sci. eae 1899: 219-224. 1899 Hayes, S., J. Keenan, and J. M. Cowan. A survey of the anatomy of the Rhododendron leaf in relation to the taxonomy ot the genus. Notes Bot. Gard, Edinburgh 21: 1-34. 1. Henry, M. G. A new hybrid deciduous Rhododendron. Rhodora 55: 205-208. pl. 1193. 1953. [R. & gladwynense. Hume, H. H. Azaleas: kinds and culture. ix + 199 pp. Macmillan, New York. 1948. Hvutcuinson, J. Evolution and classification of Rhododendrons. Rhododen- dron Yearb. Roy. Hort. Soc. 1: 42-47. 1946. The distribution of Rhododendrons. /bid. 2: 87-98. 1947. ILTIs, H H. Studies in Virginia plants. I]. Rhododendron maximum in the Vir- ginia coastal plain and its distribution in North America. Castanea 21: 114-124. 1956. {See aiso Wild Flower 32: 57-66. 1956. JANAKI AMMAL, E. k. Polyploidy in the genus Rhododendron. Rhododendron Yearb. Roy. Hort. Soc. 5: 92-98. 1950. ———, I. C. Enocu, and M. BrincGwarter. Chromosome numbers in species of Rhododendron. Ibid. 78-91. |360 species, total of 550 counts. repre- sentatives of 42 “series”; counts mostly undocumented, although source in cultivation given. | 1961 | WOOD, GENERA OF ERICACEAE 37 LAawrRENCE, G. H. M. The status of Rhododendron nudiflorum and R. roseum. Baileya 2: 1, 2. 1954 LeacH, D. G. The re-creation of a species. Quart. Bull. Am. ee Soc. 12: 188-191. 1958. [R. & Furbishii W. P. Lemmon pro. sp. = arborescens X Bakeri.| See also Gard. Jour. N. Y. Bot. Gard. 9: 3, 4. Tes Lee, F. P. Additional ge oe the azaleas and rhododendrons of the Blue Ridge Mountains. Quar . Am. Rhododendron Soc. 12: 74, 75. 1958. The azalea ae xu ++ 324 pp. Van Nostrand, Princeton, N. J. 1958. | es Luteum and Canadense subseries, pp. 161-171.! , B. Y. Morrison, M. Perkins, and F. Wetss. The azalea mee Natl Hort. Mag. 31: 1-148 (Index, 293-303). 1952. (Also reprinted separately.) Lemmon, W. P. Notes on a study of the southeastern azaleas with description of two new species. Bartonia 19: 14-17. 1938. [A. fastigifolia, A. Bakert. | A Azalea from the mountains or Georgia. Jbid. 21: 5, 6. 1942. [ 4. Furbishii; see LEACH. | Lr, H. L. Chromosome studies in the azaleas of eastern North America. Am. Jour. Bot. 44: 8-14. 1957. [271 plants representing all of the native species of § Anthodendron and many hybrids. | Lovett, J. H., and H. B. Lovett. The pollination of Rhodora. Rhodora 34: 213, 214. pl. 222, 1932. [R. canadense. | Mittats, J. G. Rhododendrons, in which is set forth an account of all species of the genus Rhododendron (including azaleas) and the various hybrids xi + 268 pp. Longmans, Green & Co., London, 1917. Rhododendrons and the various hybrids. 2nd series. xii + 265 pp. 1924. Morrison, B. Y. Rhododendron Chapmanii A. Gray. Natl. Hort. Mag. 18: 48-50. 1939. Ramseur, G. S. A natural stand of Rhododendron * wellesleyanum Waterer ex Rehder in the southern Appalachians. Jour. Elisha Mitchell Sci. Soc. 75: 131. 1959. [In the Black Mts., western N. C., alt. ca. 1700 m. | Reuper, A. Rhododendron carolinianum, a new rhododendron from North Carolina. Rhodora 14: 97-102. 1912. [See also G. V. Nash, Addisonia 1: ae eke an . Azalea or Loiseleuria. Jour. Arnold Arb. 2: 156-159. 1921. SARGENT, C. S. Rhododendron. Silva N. Am. 5: 143-150. pls. 238, 239. 1893. [R. maximum. | Sax. K. Chromosome stability in the genus Rhododendron. Am. Jour. Bot. 17: 247-251. 1930. [Undocumented counts of 16 spp., 9 hybrids. | ScHRENK, H. von. Branch cankers of Rhododendron. Missouri Bot. Gard. Rep. 18: 77-80. pls. 5, 6. 1907. [On R. maximum. |] SEITHE-vON Horr, A. Die ene der Gattung Rhododendron L. und die MOglichkeit ihrer taxonomischen Verwertung. Bot. Jahrb. 79: 297-393. pls. 6-11. 1960 SHAMMARELLO, A. M. The propagation of rhododendrons by stem cuttings. Proc. Pl. Propagators Soc. 7: 85-87. 1957.* Stus. J. Rhododendron calendulaceum. Bot. Mag. 41: pl. 1721. 1815. [See also 47: pl. 2143. 1820; 62: pl. 3439. 1835. R. catawbiense, 39: pl. 1671. 1814; R. maximum, 24: pl. 951, 1806; R. minus, 49: pl. 2285. 1822; R. speciosum, 5: pl. 180. 1792; R. Vaseyt, 132: pl. 8081. 1906. | 38 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII SINCLAIR, J. The Rhododendron bud and its relation to the taxonomy of the genus. Notes Bot. Gard. Edinburgh 19: 267-271. 1937. [223 spp.; con- volute buds in lepidote species, revolute in elepidote. | SKINNER, H. T. In search of native azaleas. Morris Arb. Bull. 6: 1-10, 15-22. 1955 (Abridged, Rhododendron Camellia Yearb. Roy. Hort. Soc. 11: 9-28. 1956). [A general account of extensive collecting, including many com- ments on species and hybridization. | . An evaluation of our eastern wild azaleas. Arb. Bull. 22: 80. 81, 106. 1959, ’, H. Camp. Luteum subseries. 7m Ler et al., Azalea Handb. pp. 29-36. [§ Anthodendron. | SLEUMER, H. Ein System der Gattung Rhododendron L. Bot. Jahrb. 74: 511- 533. 1949. [Eight subg., 20 sects.; see also Gartenflora 86: 103-105. 1937.] Florae Malesianae Precursores XXIII. The genus Rhododendron in Malaysia. Reinwardtia 5: 45-231. 1960. ———. The genus Rhododendron L. in Indochina and Siam. Blumea Surpl. 4: 39-59, : SMALL, J. K. A Georgia Rhododendron. Torreya 2: 9, 10. 1902. [R. Cuthbertii Small = = R, - us. STEVENSON, J. editor. The species of Rhododendron. i-viii + 861 pp. ane rate Society, Edinburgh. 1930. [Species of Malaysia and New Guinea omitted; species arranged in 43 series; no key to series; § Ponticum by H. F. Tagg; § Rhododendron by J. Hutchinson; subg. Anthodendron, Azaleastrum, etc., by A. Rehder. | Totten, H. R. A station for Rhododendron ati in eastern Florida. Proc. Fla. Acad. Sci. 7: 105. 1944 [1945.] [Clay Coun VEILLET-BarToszEwskA, M. Ericacées. Développement ri Yembryon chez le Rhododendron ferrugineum L. Compt. Rend. Acad. Sci. Paris 244: 1952- 1954. 1957. . Développement de l’albumen chez le Rhododendron ferrugineum L. Bull. Soc. Bot. Fr. 106: 17-20. 1959. Warp, F. K. Rhododendron seeds, with special reference to their classification. Jour. Bot. 73: 241-247. 1935 Observations on the classification of the genus Rhododendron. Rhodo- dendron Yearb. Roy. Hort. Soc. 2: 99-114. 1947, Wuerry, E. T. The American azaleas and their varieties. Natl. Hort. Mag. 22: 158-166. 1943. Witson, E. H. The “Indian Azaleas” at Magnolia Gardens. Jour. Arnold Arb. 2: 159, 160. 1921. [Charleston, S. C.] ; . A monograph of azaleas. Rhododendron subgenus Anthodendron. Publ. Arnold Arb. 9: 1-219. 1921. [The basic monograph for subg. Anthodendron. } Tribe ANDROMEDEAE Endlicher 7. Leucothoé D. Don, Edinburgh New Philos. Jour. 17: 159. 1834. Evergreen or deciduous shrubs (or small trees), with alternate, short- petioled, serrulate or crenulate to entire leaves; winter buds small. with several outer scales. Inflorescences racemose, from axillary buds on wood 1961] WOOD, GENERA OF ERICACEAE 39 of the preceding season; flowers in the axils of small, persistent bracts [in L. Grayana one or more flowers in the axils of leaves, the inflores- cence terminating a short leafy branch of the same season from an axillary bud]; pedicels with two opposite bracteoles at the base or beneath the calyx. Calyx of 5 nearly distinct sepals imbricate or quincuncial in bud. Corolla ovoid or cylindric, glabrous, white [to red], with 5 short lobes. Stamens 10, on the base of the corolla; filaments straight or S-shaped, expanded at the base, glabrous to villous, unappendaged; anthers opening by pores, obtuse or 2- or 4-awned at the apex (the awns sometimes minute), with white disintegration tissue in the angle between the awns (or in this position when the awns are lacking) ; pollen tetrads without viscin strands. Stigma expanded, somewhat peltate and 5-lobed to truncate; style straight, included or exserted from the corolla; ovary superior, 5-locular; placentae undivided, at the top of the axis, with the anatropous to campylotropous ovules mostly pendulous on the outer and/or under side. Fruit a 5-lobed, depressed-globose capsule, 5-valved, loculicidal, without thickened sutures, the calyx and placentae persistent; seeds pendulous, scobiform, angled, with a loose, thin testa, sometimes winged. (Including Eubdotrys Nutt., Agarista D. Don ex G. Don.) Typr species: L. axillaris (Lam.) D. Don. (A poetic name, for Leucothoé, daughter of Orchamus, King of Babylon, mentioned by Ovid.) A genus of about 50 species, variously placed in about six genera, two subgenera, or seven sections; about four species in eastern Asia, 35 in South America, one in Central America, one in California and Oregon, and five (in dines sections) in the eastern United States. Section LEUcoTHOE, composed of evergreen species with bracteoles at the base of the pedicels, erect, glabrous to short-pubescent filaments, and anthers with four very short (in ours) to long awns, is represented in the southeastern United States by two closely related species. Leucothoé axillaris (L. Catesbaei (Walt.) Gray, but not sensu Gray or most authors; L. platyphylla Small), with abruptly pointed leaves with petioles to about 1 cm. long and with rather broadly ovate or rounded bracts and sepals broadly ovate (hence still imbricate at anthesis), is confined to low woods on the Coastal Plain from Florida westward to Louisiana and northward to southeastern Virginia. Leucothoé Fontanesiana (Steud.) Sleum. (An- dromeda Fontanesiana Steud., L. editorum Fern. & Schub., L. Catesbaet sensu most authors), with decidedly long-acuminate leaves, longer petioles, and the bracts lanceolate-acuminate, the sepals ovate-oblong and acutish (hence scarcely imbricate at anthesis), is a plant of cool, moist woods (especially along streams) in the mountains of southwestern Virginia, west- ern North Carolina, eastern Tennessee, and northern Georgia. The two species, which appear to be completely isolated, are in need of further study, for, although the leaf shapes usually are distinctive, those of bracts and sepals appear to be less so. Leucothoé Fontanesiana is far hardier than L. axillaris and is important horticulturally northward to New Eng- land 40 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XLII The closest relative of these two species is L. Griffithiana C. B. Clarke (including L. tonkinensis Dop), of Indochina, Burma, Yunnan, and south- eastern Tibet. The placing of this species, which is amazingly similar to ours, in sect. OLIGARIsTA Sleum. on the basis of the greater development of the four awns on the anthers unfortunately obscures another of the close relationships between species of Asia and North America. Other re- lated species which fit well into sect. LeucotHo# are L. Davisiae Torr. ($ Acranthes Sleum.), of California and Oregon, and L. Keiskei Miq. ($ Paraleucothoé Nakai), of Japan. Section Acast1a DC. (Agarista D. Don ex G. Don), a distinct section of clearly related species in which the leaves are persistent and entire, the anthers without awns, the filaments S-curved and villose, the stigmas truncate, and the bracts and bractlets narrow, is primarily South American with its center of differentiation in Brazil. Leucothoé mexicana (Hemsl.) Small, ranges from Honduras to Mexico, and the similar L. populifolia (Lam.) Dippel (L. acuminata (Lam.) G. Don), 2n = 24, occurs in swamps and low hammocks on the Coastal Plain from central Florida northward to South Carolina. Among our species, L. populifolia is easily recognized by the entire, lanceolate to ovate-lanceolate, persistent leaves, the loose racemes of long-pediceled white flowers in April or May, as well as by the other characters of the section. Section Eusorrys (Nutt.) Gray (Eubotrys Nutt.) includes only two deciduous species of the eastern United States with axillary inflorescences developing the summer before flowering (as in Cassandra and Pieris), bracteoles on the pedicel just beneath the calyx, and anthers conspicu- ously 2- or 4-awned at the apex. One, L. racemosa (L.) Gray, with 4-awned anthers and wingless seeds, is a plant of the Coastal Plain from Massachusetts to Florida and eastern Texas. Within this species three varieties (racemosa, elongata (Small) Fern., and projecta Fern.) have been proposed on the basis of pubescence of branchlets, development of the inflorescence, and size of corolla. The other, L. recurva (Buckley) Gray, with recurved racemes, 2-awned anthers, and winged seeds, is a mon- tane species, mostly of rocky, open woods with other Ericaceae, Pinus, and Quercus, from western Virginia to eastern Tennessee and northern Georgia. On the basis of floral anatomy and embryology these two species were the most primitive of the 22 of the Andromedeae studied by Palser. The exact relationship of these and L. Grayana Maxim. (§$ Eubotryoides Nakai), of Japan, to the remainder of the genus is in need of thorough study. Leucothoé is distinguished by the axillary racemes (those of L. Grayana terminating short, leafy branches from axillary buds of the preceding season). the imbricate sepals, the stamens with terminal awns (or these lacking), the white areas of disintegration tissue in the connective lobes extending to the base of the awns (or the position they should occupy if present), the lack of spurs on the filaments, the dry capsular fruit lacking thickened margins, and the seeds with loose seed coats, sometimes winged. The genus appears to be most closely related to Gaultheria (g.v.); when 1961 | WOOD, GENERA OF ERICACEAE 41 awned, the stamens are strikingly like those of both Gaultheria and Zenobia. The observations of Palser that “the species of Leucothoé have one character in common which separates them from other species studied — the position of the placentae against the outer carpel walls with the ovules borne on the inside surfaces,” need to be extended, for other species do not conform. In L. axillaris, L. Fontanesiana, and L. Davisiae the stalked placentae extend along the top of the locule and bend downward, bearing ovules only on the lower, inner surface; and in L. racemosa and L. re- curva, the placentae are like those of Cassandra: columnar with the end pressed against the ovary wall, the ovules borne around the margins. How- ever, in L. populifolia, L. Griffithiana, and L. Grayana, the ovules are borne on the outer as well as lateral surfaces of the stalked placentae. REFERENCES: Under family references see also Cox (Ericales II), KEARNEY (1901), and PALSER (1951, 1952). BAKER, J. G. Leucothoé Davisiae. Bot. Mag. 32: pl. 6247. 1876. DE ie ILDEMAN, E. Leucothoé recurva A. Gray. Ic. Sel. Horti Thenensis 1: 74. pl. 40. 1900. Detailed aeeiton. | ae aE L. Last survivors in the flora of tidewater Virginia. Rhodora 41: 465-504, 529-559, 564-574. pls. 570-583. 1939. [Vars. of L. racemosa, 553, 554, pl. 578.] . Botanical specialties of the Seward Forest and adjacent areas of south- eastern Virginia. /bid. 47: 93-142, 149-182, 191-203. pls. 876-911. 1945. [L. axillaris var. ambigens Fern. and discussion of LZ. axillaris and L. Fon- tanesiana. | ——— and B. G. ScHusert. Studies of American types in British herbaria. Ibid, 50: 149-176, 181-208, 217-233. pls. 1097-1117. 1948. [Andromeda Catesbaei Walt. is L. axillaris (Lam.) D. Don; L. editorum Fern. & Schub. | Hara, H. Observationes ad plantas aia orientalis. sed Jap. Bot. 11: 622-633. 1935. |Hubotryoides (Nakai) Hara, gen. nov., for L. Grayana Maxim. | MiyajiMa, S., and S. Taker. Ueber eee die wirksamen Inhaltstoffe von Leucothoé Grayana. II. Jour. Agr. Chem. Soc. Japan 12 (Bull.): 73 1936.* SARGENT, C. S. Leucothoé recurva. Gard. Forest 9: 224. fig. 33. 1896. ae H. Ein Bestandteil der Blatter von Leucothoé Keiskei Miq. (In Japa- ; German summary.) Jour. Pharm. Soc. Japan 59: 619, 620. 1939.* ae J. Andromeda Catesbaei. Bot. Mag. 45: pl. 1955. 1818. |[L. Fontanesi- ana. | SLEUMER, H. Ericaceae americanae novae vel minus cognitae III. Notizbl. Bot. Gart. Berlin 13: 206-214. 1936. [Key to sects. of Leucothoé, 211-213.] . Studien uber die Gattung Leucothoé D. Don. Bot. Jahrb. 78: 435-480. 1959. | Key to sects. and spp.; detailed treatment of Centr. and S. Am. spp. of § Agastia. | SMALL, J. K. Leucothoé Catesbaei. Addisonia 4: 61, 62. pl. 151. 1919. YAMASHITO, M. On grayanotoxin, the poisonous constituent of the leaves of Leucothoé Grayana Max. I. Sci. Rep. Tohoku Univ. Ser. Math. Chem. -hys. 21: 537-544. Soe 42 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII 8. Gaultheria Linnaeus, Sp. Pl. 1: 395. 1753; Gen. Pl. ed. 5. 187. 1754. Dwarf, prostrate, evergreen undershrubs [to upright shrubs or rarely small trees] with broad, alternate [rarely opposite], short-petioled, usually serrate leaves; winter buds ovoid, with several outer scales. Flowers bisex- ual [or rarely both bisexual (?) and carpellate, the latter with much re- duced stamens], 5(rarely 4)-merous, [in terminal panicles or axillary racemes or clusters or]| in ours solitary in the axils of leaves; pedicels with 2 bractlets [at the base or middle or] immediately beneath the calyx [or the bracteoles several, scattered]. Calyx 5(4)-parted, [usually] accres- cent and becoming fleshy. Corolla campanulate to urceolate, white to pink, glabrous or hairy within. Stamens 10 (8), on the base of the corolla; filaments expanded at the base, hairy or glabrous; anthers with white dis- integration tissue on the abaxial side at the base of the bifid terminal awn on each anther-half [or in this position when awns are lacking], or dis- integration tissue lacking in the most reduced species, the anthers opening by terminal pores; pollen tetrads without viscin strands. Stigma obtuse or truncate; style straight; ovary superior (and free from the calyx) or rarely partly inferior, 5(4)-locular, the axile placentae with 5—10[—many] ovules; disc 10-lobed. Fruit a 5(4)-valved loculicidal capsule inclosed by and rarely united with the usually colored (red, white [to bright blue or black]) and fleshy calyx, appearing berry-like; seeds usually numer- ous, small, hard, angled, wingless. (Including Chiogenes Salisb.) TypE SPECIES: G. procumbens L. (Named in honor of Jean-Francois Gaulthier, ?1708-1756, botanist and court-physician at Quebec.) A genus of perhaps 150 species, many of high mountains: in Asia ranging from Japan, China, and the Himalayas, southward to New Guinea, Australia, Tasmania (three species), and New Zealand (about six species) ; in North America from Labrador to British Columbia, southward through Mexico and the West Indies to Patagonia. Gaultheria has not been revised as a unit, but Airy-Shaw has placed the species of continental Asia in five sections which include the species of North America. Section Gaut- THERIA (§ Eugaultheria Airy-Shaw) occurs in America as two widespread species primarily of the East; sects. AMBLYANDRA Airy-Shaw and BrossAr- opsis Airy-Shaw are represented in the West by G. ovatifolia Gray and G. humifusa (Graham) Rydb. and by G. Shallon Pursh, respectively. The two species of the West Indies and those of Central America also belong to sect. BROSSAEOPSIS. Section GAULTHERIA (dwarf undershrubs with mostly small leaves and solitary flowers with two opposite bracteoles immediately beneath the calyx) includes the most reduced species in the genus. Series Procum- bentes Airy-Shaw comprises only the pentamerous, red-fruited G. pro- cumbens (2n = 24), nomenclaturally typical but morphologically an iso- lated type in the genus. Widespread in eastern North America (New- foundland to Manitoba, south to Minnesota, Georgia and Alabama [in the mountains], and southeastern North Carolina), the species is prob- 1961] WOOD, GENERA OF ERICACEAE 43 Fe 7 Pee ag, ie = &. a POA A PAE! GLEE: SOM Fic. 3. Gaultheria. , G. procumbens: a, x 4; b, flower, X 3; c, stamen, X 10; d, ee. oalee view, eae ee of acidic: as awns, X<-LORne: calyx, disc, gynoecium, X 6; ovary, disc, calyx, vertical section, x 6; g, fruit with accrescent calyx, Ng : -h, fruit and calyx in vertical section, « 3; i "seed x 20. j-l, G. hispidula: j, fruiting shoot, from bel a; k, en, outer view, showing lack of disintegration tissue, < 20; 1, immature fruit in vertical section — note two calyx lobes at top, & 4 ably most closely related to G. (§ Leucothoides) pyroloides Miq., of Japan. Series Hispidulae Airy-Shaw (which seems to have little relationship to Procumbentes) includes only two tetramerous-flowered, small-leaved, deli- cate, creeping species which represent ‘“‘the last word in reduction in the genus.” Gaultheria hispidula (L.) Muhl. var. hispidula (Chiogenes his- pidula (L.) Torrey & Gray) is a white-fruited plant, mostly of cold woods and bogs to the north of our range (Labrador to New England and Penn- sylvania, westward to Minnesota, Idaho, and British Columbia) but ex- tends sporadically southward in cold mountain bogs to western North Carolina. Varietas japonica (Gray) Makino, of central and northern 44 JOURNAL OF THE ARNOLD ARBORETUM [| VOL, XLII Japan, differs only in the more obovate leaves with more cuneate base and in the shorter and less deeply bifid awns of the anthers. The second species, the awnless, red-fruited G. suborbicularis W. W. Sm., occurs in the high mountains of Yunnan. In both, the calyx tube enlarges into a berry-like structure which almost completely surrounds and is united with the capsule, while the calyx lobes persist almost unchanged at the top. The appearance thus given of a berry from an inferior ovary (in flower the ovary is about half-inferior) has led to the erroneous associa- tion of G. hispidula (as Chiogenes) with the Vaccinioideae. A third series, Trichophyllae Airy-Shaw, consists of eight or more pentamerous species, of the Sino-Himalayan ranges. Closely related to (and perhaps congeneric with) Gaultheria is Pernettya Gaud. (Mexico to Patagonia, New Zealand, and Tasmania) which differs consistently only in the baccate (vs. capsular) fruit, although usually, but with exceptions in both genera, in the dry vs. fleshy calyx. Natural hybrids ( Gauithettya Camp) have been found in New Zealand, Mexico, and the Falkland Islands. Gaulthettya wisleyensis (Marchant) Rehd. (G. Shallon Pursh [2n = 88] x P. mucronata (L. f.) Spreng. |2” = 66]) is a heptaploid (2m = 77) which is partially fertile (presumably through pairing within the polyploid parental genomes?) producing aneuploid off- spring (e.g., 2” = 70, 71, 79). Such hybridization was suggested by Cal- lan as a possible method of origin of new species; Camp suggested inde- pendently that the Mexican P. ciliata (Cham. & Schlecht.) Small and perhaps P. hirsuta (Mart. & Gal.) Camp may have resulted from a back- cross between a Gaulthettya and the Pernettya parent or from genetic segregation within the hybrid itself. A number of interspecific hybrids, some fertile, are known in Gaultheria. Base chromosome numbers of 11, 12, and 13 have been reported in Gaultheria (2n = 22, 24, 26, 44, 88, 96), and of 11 in Pernettya (2n = 22, 44, 6 Gaultheria shows with other genera reticulate relationships which hardly support the retention of the tribe Gaultherieae as a natural unit apart from the Andromedeae. Airy-Shaw notes G. codonantha Airy-Shaw, from As- sam, as nearest to the prototype of the genus, and indicates a close agree- ment “in the morphology of the calyx, corolla, filaments, anthers and seeds,” as well as in the inflorescence, between this species and Zenobia and Lyonia § Lyonia and § Maria. He further calls attention to the re- semblance of G. (§ Leucothoides) Griffithiana Wight to Leucothoé Grif- fithiana C. B. Clarke, remarking that the former provides an almost per- fect link between Gaultheria and Leucothoé § Leucothoé. He would re- gard Gaultheria as an advanced type ‘adapted for seed-dispersal by the agency of birds.” REFERENCES: Under family references see also ARTOPOEUS, nea & BURCHILL, BERGMAN, Cox (Ericales II), MatrHews & Knox, and PELTR ApramMs, L. The dwarf gaultherias in California. "eke 2: 121, 122. 1934. [G. naga G. ovatifolia. | Arry-SHaw, H. K. Studies in the Ericales. IV. Classification of the Asiatic 1961 | WOOD, GENERA OF ERICACEAE 45 species of Gaultheria. Kew Bull. 1940: 306-330. 1940. [Includes sects., series applicable to N. Am. spp.|; VI. Further notes on Gaultheria dumicola W. W. Sm. /bid. 1948: 109, 110. 1948; VII. Illustrations of four scarce Asiatic gaultherias. /bid. 158-161; X. An undescribed Asiatic Gaultheria, with a key to the Sino-Himalayan species of section Leucothoides, Ibid. 1952: 171-174. 1952. ALEXANDER, E. J. Gaultheria procumbens. Addisonia 12: 53, 54. pl. 411. 1927. BESANT, J. W. Gaultherias. New Fl. Silva 11: 211-218. 1939.* BrIDEL, M., and S. Gritton. La glucoside a salicylate de méthyle du ale procumbens L. est le monotropitoside. Jour. Pharm. Chim. VIII. 9: 1-16. 1929.* [See also Compt. Rend. Acad. Sci. Paris 187: 609-611. 1928.) Burtt, B. L., and A. W. Hity. The genera reg ily and Pernettya in New Zealand, Tasmania, and Australia. Jour. Linn. Soc. Bot as 611-644. map. 1935. [Important comments on generic eee hybrids, Brook, P. ee i ae of Pernettya macrostigma. New PLitel 51: 388-397. pl. 9 CALLAN, a The cytology of Gaulthettya ae. Sarasin in a new mode of species formation. Ann. Bot. II. 5: 579-585. Camp, W. H. Studies in the Ericales. IV. Néies on Chimaphla. ae and Pernettya in Mexico and adjacent regions. Bull. Torrey Bot. Club 66: 7-28. 1939. [Includes Gaulthettya. | Cuovu, Y. L. Floral morphology of three species of Gaultheria. Bot. Gaz. 114: 198-221. 1952. [G. procumbens, G. ovatifolia, G. Shallon; includes em- bryology Hom, T. Medicinal plants of North America. 11. Gaultheria procumbens L. Merck Rep. 17: 1-14. 1908. LeBeDEv, D. V. Gibridy autopoliploidov i vidoobrazovanie (Hybrids of ee polyploid and the formation of species). Priroda Leningrad 36(3): 62, 7.* [X Gaulthettya wisleyensis. | een T. M. Host-endophyte relationships in eee of Pernettya macrostigma. New Phytol. 56: 247-257. pl. 6. 1957 Muttican, B. O. A probable bigeneric hybrid between Gauitheria and Pernettya. Jour. Roy. Hort Soc. 64: 125-127. figs. 14-16( 717). 1939. [P. mucronata < G. Shallon = X Gaulthettya wisleyensis. | Stus, J. Gaultheria procumbens. Bot. Mag. 45: pl. 1966. 1818. SLEUMER, H. Ericaceae americanae novae vel minus cognitae II. Notizbl. Bot. Gart. Berlin 12: 277-294. 1935. [Includes key to Gaultheria in Mexico and Guatemala, 285-287. | als ion der Gattung Pernettya Gaud. Jbid. 626-655 e Malesianae Beas a XV. The genus Goulthena in Malay- sla. ete an at 163-188. He VEILLET-BARTOSZEWSKA, M. mbryogénie des éricacées; développement de Vembryon chez . ee Shallon Pursh. Compt. Rend. Acad. Sci. Paris 248: 720-722. 5 9. Zenobia D. Don, Edinburgh New Philos. Jour. 17: 158. 1834. Deciduous to half-evergreen, glabrous, often glaucous shrubs with alter- nate, short-petioled serrulate-crenulate to entire, veiny, coriaceous leaves; buds ovoid, obtuse, with several outer scales. Flowers showy, nodding. fragrant, in corymbs from axillary buds on the upper part of twigs of the 46 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII preceding season; individual flowers in the axil of a short bract, the pedicel with 2 lateral bracteoles at the base. Calyx lobes 5, valvate, thickish, per- sistent and somewhat accrescent in fruit. Corolla white, broadly cam- panulate, obtusely 5-lobed. Stamens 10; filaments abruptly enlarged be- low, flattened; anthers opening by oblong pores, each anther-half sur- mounted by a pair of slender ascending awns and with a conspicuous area of white disintegration tissue in the connective; pollen tetrads without viscin strands. Stigma truncate, simple; style columnar, about as long as the corolla; ovary superior, 5-locular, 5-lobed, placentae from about the middle of the axis in each locule, with numerous ovules; disc 10-lobed. Fruit an erect depressed-globose loculicidal 5-valved capsule, the valves, axis, and placentae persistent; seeds numerous, angled, not winged. TypE AND SOLE SPECIES: Z. speciosa (Michx.) D. Don = Z. pulverulenta (Bartr. ex Willd.) Pollard. (Name fanciful, for Zenobia, queen of Pal- myra, 267-272 A. D., who conquered Egypt and most of Asia Minor.) Fic. 4. Zenobia. a—h, Z. pulverulenta: a, flowering twig and new growth, * 4; b, flower with most of corolla removed — not e white areas of disintegration tis- tation and disc, semidiagrammatic, X 5; f, ovary, cross section, semidiagram- matic, X 5; g, open capsule, & 4; h, seed, & 15 1961] WOOD, GENERA OF ERICACEAE 47 A single isolated species of damp, sandy or peaty, pine savannas, shrub-bogs, or swamp margins on the Coastal Plain from southeastern Virginia to northeastern South Carolina. Zenobia pulverulenta is a showy shrub with racemes of (1—)5—10-flowered corymbs of fragrant, campanu- late, white flowers about 1.5 cm. across borne from May to mid-June. In f. pulverulenta the undersides of the leaves (in addition to stems, pedicels, calyces, and ovaries) are heavily glaucous, while in f. »itida (Michx.) Fern. (Z. cassinefolia (Vent.) Pollard; Z. speciosa (Michx.) D. Don. var. nitida (Michx.) Rehd.) the leaves are green beneath. Al- though the two have been treated as separate species, they occur together, along with intermediates, and there is no question that only a series of forms of a single distinctive species is represented. Both are showy but not often cultivated plants which are quite hardy at least as far north as Massachusetts. Zenobia is characterized by the combination of inflorescence, 4-awned stamens, relatively large campanulate corollas, capsular fruit, and angled, unwinged seeds. There seem to be no close relatives, but the placentation and inflorescence resemble Lyonia § Maria while the corolla and stamens are reminiscent of Gaultheria codonantha, suggesting interrelationships with these genera. (See Gaultheria.) Neither the chromosome number nor observations on the biology of the plant seem to have been recorded. REFERENCES: Under family references see also ARTOPOEUS and PELTRISOT. FERNALD, M. L. A century of additions to the flora of Virginia. Rhodora 42: 355-416, 419-498, 503-521. pls. 626-649. 1940. [Discussion of forms, nomenclature, and distribution of Zenobia, 471-473, map 16; see also 385, 3 : Linpey, J. Andromeda dealbata. Bot. Reg. 12: pl. 1010. 1826. [An aberrant form with deeply lobed corollas. Pierce, S. A. Zenobia pulverulenta. Gard. Chron. III. 119: 148. 1946. [Notes on the plant, its culture and propagation. PotiarD, C. L. The genus Zenobia D. Don. Bull. Torrey Bot. Club 22: 231, 232. 1895 Stms, J. Andromeda pulverulenta. Bot Mag. 18: pl. 667. 1803; Andromeda cassinefolia. Ibid. 25: pl. 970. 1807. 10. Lyonia Nuttall, Gen. N. Am. Pl. 1: 266. 1818, nom. cons. Evergreen or deciduous shrubs [rarely small trees], with terete or angled branches. Leaves alternate, short-petioied, entire or shallowly toothed or serrulate, glabrous, or with hairs or peltate scales; winter buds ovoid, with 2 outer scales. Inflorescences axillary (sometimes appearing to be termi- nal, but falsely so), corymbose, [racemose] or paniculate, the racemes sometimes contracted into axillary clusters; each flower in the axil of a small leaf or bract and with 2 lateral bracteoles at the base of the pedicel, these sometimes quickly deciduous. Calyx 5(rarely 4—8)-lobed, the lobes valvate or reduplicate in bud, persistent or rarely deciduous in fruit. 48 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII Corolla cylindric-campanulate to urceolate or globose-urceolate, with 5(4-8) short lobes, white to pink, glabrous to hairy or lepidote. Stamens 10 (rarely 8-16); filaments flattened, often S-shaped, glabrous to hairy or roughened, with or without a pair of short, spurlike appendages on the back near the apex; anthers obtuse, the lobes parallel, lacking apical awns, dehiscent by large terminal pores, always with a white line of dis- integration tissue on the back of each lobe extending at least along the apex of the filament and along the upper edge of the appendages when present; pollen tetrads without viscin strands. Stigma truncate to capi- tate; style columnar to fusiform, straight, not exserted; ovary superior, 5-locular, the placentae large, undivided; disc an enlargement of the ovary wall, variously developed. Capsule subglobose to ovoid, 5-angled, locu- in dehiscence; placentae persistent at the top of the columella; seeds scobiform, with a loose, thin testa. (Xolisma Raf.; not Lyonia Raf., 1808, or Lyonia Ell., 1817, nomina rejicienda; including Arsenococcus Small, Desmothamnus Small, Neopieris Britton.) Lectotype species: L. ferrugi- nea (Walt.) Nutt.; see I.C.B.N. 261. 1956, and Rickett & Stafleu, Taxon 9: 75. 1960. (In commemoration of John Lyon, 17?—1818, early Ameri- can botanist and explorer of the southern Appalachians, “who fell victim to a dangerous epidemic amidst those savage and romantic mountains which had so often been the theatre of his labours.” A genus of perhaps 40-50 species, in three or four sections, about ten in Asia (Kashmir to Japan, south to Malaya), perhaps 30 or more in the Greater Antilles, one or two in eastern and southern Mexico, and five in the United States. The Asiatic species belong to sect. Prertpopsis (Rehd.) Airy-Shaw, with 1-sided, usually elongated racemes, filaments with or with- out appendages; leaves entire, persistent or deciduous), which might well be merged with sect. Marta. Although the genus has been divided into a number of genera, the distribution of the characters used is so reticulate, when all of the species are considered, that segregates are essentially based on single-character differences. The species of the southeastern United States are distinctive, but the genus as a whole (especially § Lyonta) is in need of a thorough monographic study. Section Maria (DC.) C. E Wood ® (Xolisma § Maria (DC.) Rehd.), as defined here, esa only three isolated species which are placed together on the basis of the fila- ments appendaged near the summit, the nonlepidote pubescence, and the e proper combination under Lyonia does not appear to have been made previ- ously for this section. This combination, with additional synonymy is yonia Nutt. sect. Maria (DC.), comb. nov. Leucothoé D. Don, sect. Maria DC. Prodr. 7(2): 602. 1839; Andromeda L. sect. Maria (DC.) Gray, Man. Bot, North. . S. 266. 1848; Pieris “ang Maria (DC.) Benth. & Hook. Gen. Pl. 2: 588. 1876; Lyonia subgen. Maria (DC.) Drude, Nat. Pflanzenfam. IV. ht 44. 1889; Nolisma Raf. sect. Maria (DC.) Rehd. Jour. Arnold Arb. 5: 55. 1924; Arsenococcus Small in Small & Carter, Fl. Lancaster Co. 218. 1913; NXolisma Raf. sect. Arsenococcus (Small) Rehd. Jour. Arnold Arb. 5: 54. 1924. Typr SPECIES: Lyonia mariana (L.) Don. 1961] WOOD. GENERA OF ERICACEAE 49 axillary corymbose (fasciculate) inflorescences. Lyonia mariana (L.) D. Don (Neopieris mariana (L.) Britton), 2n = 24, ranging mostly on the Coastal Plain from Florida to Missouri, Arkansas, and eastern Texas, and Fic. 5. Lyonia. a-i, L. mariana: a, branchlet, X 12; b. flower with bract Se bractlets, < 3; c-e, lateral, inner and outer views of stamen and anthers si show filament appendages, aisinten dion tissue and anther dehiscence, * 10; gynoecium, lateral view — note glandular development of base of ovary, < Ss g, ovary, vertical section, semidiagrammatic, x 5; h. opened ee partial separation of thickened valve-margins (carpel midribs), * 4; i, seed, 20. j-l, L. lucida: j, flower, X 3; k. lateral and outer views of an nthers — n anbeny jue with margin of disintegration tissue, &X 10; 1. leaf. to show marginal x Y. m-o, L. Ge m. immature and opened capsules — note sepa- ene. of differentiated ag . X 4; n. 0, outer view of anthers of two collec- tions to show appendages. p-s. a Sraiticosa’ p. flowering branchlet. * %:; a. flower, X 3; r, stamen. itera view. < 10; s, opened capsule. one differentiated midrib (valve-margin) partially remoy a x4 t. L. ferruginea: outer view of anther. showing lines of distention tissue, 10 : 50 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLII northward to eastern Pennsylvania, New Jersey, and southern Rhode sland, is notable for the calyx lobes deciduous with the leaves in fall, while L. lucida (Lam.) K. Koch (Desmothamnus lucidus (Lam.) Small, Neopieris nitida (Bartr.) Britton), 2” = 24, distributed from western Cuba northward on the Coastal Plain to swaliensarn Virginia and Louisi- ana, is unique in the shining, persistent, coriaceous leaves with an intra- marginal vein. Although L. ligustrina (L.) DC. is unique in the genus in the axillary or pseudoterminal inflorescences which are small panicles (racemes of corymbose fascicles) aggregated near the end of the growth of the preceding year, its separation as a monotypic section is doubtfully justifiable. Within the broad range of this species (New England to New York, West Virginia, Kentucky, Arkansas, Oklahoma, and eastern Texas southeastward to Florida) a number of not very sharply limited geograph- ical varieties in need of further study have been defined: vars. ligustrina; salicifolia (Wats.) DC.; capreaefolia (Wats.) DC.; foliosiflora (Michx.) Fern.; and pubescens (Gray) Bean (Arsenococcus frondosus (Pursh) Small) Section Lyonita (Xolisma Raf.), with the flowers in dense axillary fascicles (as in § Marta), the filaments unappendaged, the leaves per- sistent and lepidote, and the thickened part of the sutures separating as a whole from the rest of the capsule (sometimes seen in other sections as well), is primarily of the Greater Antilles, with two species, L. fer- ruginea (Walt.) Nutt. and L. fruticosa (Michx.) G. S. Torrey on the Coastal Plain from Florida to South Carolina. Lyonia ferruginea also occurs in Mexico along the eastern Sierra Madre from San Luis Potosi south to Oaxaca. The former is a shrub or small tree with leaves much rolled on the margin, while the latter is always a shrub of stricter habit with nonrevolute leaves and flowers as much as two months later. Matthews and Knox (1926) and Anthony (1927) have doubted that any real distinction exists between appendages on the back of the anthers (spurs) and those at the summit of the filament (used by Rehder [1924] in separating Pieris and Lyo onia). Their contention as to the lack of a shown that those of Pieris lack the white disintegration tissue which is always present on the upper aie of the appendages of Lyonia.’ Some ®°The additional difference of a supposed absence of appendages on the filaments of this species does not hold, for ae on which are comparable with those of L. of tissue which disintegrates, arise from the filament below its point of attachment with the anther . . . The disintegrating areas continue upward. These spurs are ob- servable only on young filaments, as they have disintegrated completely on mature ones and only the ragged lateral margins of the filaments remain.” These appendages are pale present in boiled-up flowers of the material which I have examined. ‘ when the filaments of Lyonia lack the awnlike appendages near the sum mit, a tine of white disintegration tissue is present on each anther-half, in some species of § Lyon1a forming an inverted ‘“‘V” at the junction of filament and anther. (See also footnote 6.) Confusion of the two types of appendages may have come in 1961] WOOD, GENERA OF ERICACEAE 51 of the other anatomical distinctions made by Palser (e.g., vertically vs. horizontally S-shaped filaments) do not hold, but the conspicuously paler, usually thickened sutures of the fruit, exclusively axillary inflores- cences from the wood of the preceding season, two bracteoles at the base of the pedicels, and two large outer scales of the winter buds set Lyonia off as a natural group of species quite distinct from Pieris. REFERENCES: Under family oS see also Cox (Ericales II), KEARNEY, PALSER (1951, 1952), and PELTR Arry-SHAW, H. K. Lyonia macrocalyx. Bot. Mag. 160: pl. 9490. 1937. [In- cludes comments on § Pieridopsis and Pieris bracteata. ANTHONY, J. A description of some Asiatic phanerogams. Notes Bot. Gard. Edinburgh 15: 239-246. 1927. [Pieris macrocalyx; merges Pieris and olisma. | FERNALD, M. L. The varieties of Lyonia ligustrina. Rhodora 43: 624-629. 1941. [In “Another century of additions to the flora of Virginia.’ GREENE, E. L. A name explained. Torreya 4: 173, 174. 1904. [Derivation of X jaa Raf. ] Kucera, C. L. The genus Lyonia in Missouri. Rhodora 55: 155. 1953. [L. mariana. | Lovett, J. H., and H. B. Lovett. Pollination of the Ericaceae: Chamaedaphne and Yolisma. Rhodora 37: 157-161. 1935. [L. ligustrina.] REHDER, A. New species, varieties and combinations from the herbarium and the collections of the Arnold Arboretum. Jour. Arnold Arb. 5: 49-59. 1924. [49-55. takes up and defines Xolisma Raf., sects., new Ee (See also ibid. 20: 425, 426. 1939, for correct name for L. luc SARGENT, C. S. Andromeda. Silva N. Am. 5: 129-132. ni ae 1893. [L. ferruginea. | Stms, J. Andromeda mariana (a.) ovalis. Bot. Mag. 37: pl. 1579. 1813. SLEUMER, H. Neue Xolisma-arten von Hispaniola. Repert. Sp. Nov. 36: 270- 273. 1934. [Seven new spp. | SMALL, J. K. Desmothamnus lucidus. Addisonia 11: 51, 52. pl. 378. 1927. Yasue, M., and Y. Kato. Studies on the constituents of Lyonia ovalifolia pe et Zucc. var. elliptica Hand.-Mazz. oi Japanese; English summary.) Pharm. Soc. Japan 79: 403-405. 1959 11. Pieris D. Don, Edinburgh New Philos. Jour. 17: 159. 1834. Evergreen shrubs [small trees] rarely vines, with alternate [rarely opposite], short-petioled, coriaceous, broad, entire, crenulate, or toothed eaves; winter buds ovoid, with several narrow outer scales. Flowers white, in terminal panicles or axillary racemes produced the preceding summer, or racemes from axillary buds of the preceding growing season; pedicels each in the axil of a bract and with 2 alternate to opposite brac- part from the stamens of Pieris bracteata W. W. Sm. which were used as an example of both awned and appendaged anthers by Matthews & Knox and by Anthony. The plant has been found by Airy-Shaw to be not a Pieris but a form of Vaccinium For- restit Diels (or a closely related species) with foliaceous bracts (Bot. Mag. 160: pl. 9490. 1937). 52 JOURNAL OF THE ARNOLD ARBORETUM [| VOL, XLII teoles from above the base to beneath the calyx. Calyx deeply 5-lobed, the firm, thick-edged, ovate lobes valvate in bud, persistent in fruit. Corolla ovoid-urceolate to ovoid, with 5 short lobes. Stamens 10, included; fila- ments more or less vertically S-shaped and flattened; anthers with a pair of stout, deflexed spurs (without disintegration tissue) on the back just above junction with filament, otherwise unappendaged but with white dis- integration tissue on the abaxial side of each of the often divergent anther-halves, each anther-half opening by a large oval to V-shaped pore; pollen tetrads without viscin strands. Stigma flat; style straight, as long as the corolla; ovary superior, 5-lobed, 5-locular, each locule with a short-stalked, pendent placenta on the axis near the top of the locule, the outer, lateral and lower surfaces with pendulous ovules; disc 10-lobed, with nectariferous lobes extending upward between the filaments. Capsule subglobose, loculicidally 5-valved, the sutures not thickened; seeds scobi- form, nearly all pendulous, with a loose, cellular coat. (Including Am- pelothamnus Small, Arcterica Coville.) Type SPECIES: P. formosa (Wall.) D. Don. (Name poetic, for Pieris, a Muse.) A small genus of about eight species, five in eastern Asia, one in western Cuba. and two in the southeastern United States. Pieris floribunda (Pursh) Benth. (Andromeda florib-rida Pursh) is a handsome shrub (to about 1.5 m.) of moist to dry, more or less open, rocky woods and mountain slopes from Virginia and eastern West Virginia to northern Georgia, eastern Tennessee, and Alabama. The inflorescences terminate the growth of the season in June or July; the small white flowers expand the following April or May. Although without any very close relatives, P. floribunda appears to be related to P. japonica (Thunb.) G. Don (2n — 24), P. formosa (Wall.) D. Don (2n = 24), P. taiwanensis Hayata, and P. Swinhoei Hemsl., all of Asia. Pieris floribunda and P. japonica are two of the best of the ornamental Ericaceae for the eastern United States. Pieris phillyreifolia (Hook.) DC. (Ampelothamnus phillyreifolia (Hook.) Small), is a curious plant of cypress ponds, sphagnum bogs, or wet pine- lands from southernmost Georgia to central Florida, western Florida, and southern Alabama, in which the inflorescences are axillary and appar- ently produced just prior to flowering (January to March). The plant, which may be a shrub to about | m. tall, may climb to 10 m. as a vine by growing beneath the outer bark of Taxodium ascendens Brongn. and sending out leafy stems at intervals. No rootlets or other organs of sup- port are formed, and the species seems to climb only this Taxodium, in which the outer bark is composed of long, parallel, easily separable fibers. Although segregated as Ampelothamnus Small, largely on the basis of the axillary racemes and the strongly S-shaped flattish filaments, this species belongs in Pieris. The details of floral anatomy and the characteristic anthers are precisely those of Pieris, the flattening and curving of the fila- ments are only a matter of degree, and P. cubensis (Griseb.) Small, of Pinar del Rio, an undoubted member of the genus, seems also to have 1961 | WOOD, GENERA OF ERICACEAE 53 exclusively axillary racemes. The production of a number of deciduous scale-leaves on the lower part of the new growth, although seldom seen in P. floribunda, is well developed in P. phillyr cifolia and is a character- istic pattern in P. japonica and its close allies. Fic. 6. Pieris. a—g, P. floribunda: a, flowering ae mate x ™%; b, flower, x 3; c. flower with half of corolla and calyx removed — note nectariferous ihe of disc between stamens, 6; d, outer views of Weecus from outer (left) and inner staminal whorls — note ARES Hed tissue along each anther-half above spurs. X 10; e, inner ia of stamen from inner staminal whorl to show pores, edge of H ehteseiuon issue on each anther-half, * 10; f, eas ont e 4: g, seed, Mm Osher phillyreifolie: stamen from outer staminal whorl, < 1 The limits of Pieris have been a source of great confusion (see Lyonia), but the genus seems to be a natural unit consisting of evergreen species with persistent, coriaceous, serrulate, undulate, or bristle-margined leaves: winter buds with several narrow outer scales; racemose inflorescences either aggregated into terminal panicles or from the upper leaf axils; pedicels with bracteoles above the base; valvate sepals with fiber strands: anthers with stout deflexed spurs lacking disintegration tissue but with disintegration tissue ,in the connective above; and capsules without thick- ened or otherwise differentiated margins. The appendaged anthers are very distinctive and are in no way to be confused with those of Lyonia. REFERENCES: Under family references see also Cox (Ericales II). MatrHews & Knox, PALseR (1951, 1952), and PELTRISOT. 54 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Arry-Suaw, H. K. Pieris formosa var. Forrestii. Bot. Mag. 157: pl. 9371. 1934. [Includes comments on the generic limits of Pieris. BriweL, M., and A. Kramer. L’asebotoside et son identité avec le phlorizoside. Bull. Soc. Chim. Biol. 15; 531-551. 1933.* [P. japonica. | Epwarps, S. Andromeda floribunda. Bot. Reg. 10: pl. 807. 1824. Harper, R. M. A unique climbing plant. Torreya 3: 21, 22. 1903. [P. phil- lyreifolia. | Hooker, W. J. Andromeda phyllyreifolia. Hooker Ic. 2: pl. 122. 1837 Murakami, S., and M. Fuxupa. Studies on the constituents of Pieris japonica D. Don. III. Inconsistency of the occurrence of asebotin. (In Japanese; English summary.) Jour. Pharm. Soc. Japan 75: 603, 604. 1955.* Sms, J. Andromeda floribunda. Bot. Mag. 37: pl. 1566. 1813. SKAN, S. A. Pieris formosa. Bot. Mag. 135: pl. 8283. 1909. SMALL, J. K. Pieris floribunda. Addisonia 5: 5, 6. pl. 163. 1920. Stapr, O. Pieris taiwanensis. Bot. Mag. 149: pl. 9016. 1923. [Includes com- ments on confusion among Pieris, Andromeda, Lyonia, Cassandra, etc. Tamura, K. Studies on the structure of asebotin, a component of Andromeda japonica Thunb. Bull. Chem. Soc. Japan 11: 781-785. 1936.* Wixie, D. The genus Pieris, Arb. Bull. 11(2): 6, 7, 38. 1948.* { Horticul- ture. 12. Cassandra D. Don, Edinburgh New Philos. Jour. 17: 158. 1834. Low, evergreen shrubs of sphagnum bogs, peaty swales, and pond mar- gins. Leaves alternate, elliptic to lanceolate, short-petioled, entire or crenulate, more or less appressed to the stem; buds small, with several outer scales: vegetative parts, pedicels, and calyx (especially underside of leaves) scurfy with peltate scales. Flowers small, white, nodding, short- pediceled, solitary in the axils of the small upper leaves. forming almost horizontal, apparently 1-sided, 10-30-flowered, leafy racemes. Calyx small, 5-lobed, subtended by 2 persistent bractlets, aestivation quincuncial. Co- rolla oblong-urceolate, with 5 short lobes. Stamens 10, included; filaments flat, tapering upward; anthers terminating in 2 elongated tubes opening by a terminal pore, each sometimes prolonged beyond the pore as a minute triangular awn, otherwise unappendaged, lacking white disintegration tis- sue; pollen tetrads without viscin strands. Stigma slightly expanded; style straight, exserted from the corolla; ovary superior, 5-lobed, 5-locular, the placentae columnar, undivided, 1 in each locule from the center of the axis, the end flattened against the ovary wall, bearing about 15 campylotro- pous ovules around the edge; disc dark green, 10-lobed, nectariferous. Capsule depressed-globose, with slightly thickened sutures, loculicidally 5-valved, but the wall splitting into 2 layers, the inner 10-valved; seeds about 10 in each locule, small, somewhat compressed laterally, brown, shining, wingless. (Not Cassandra Spach, 1840, = Leucothoé; Chamae- daphne Moench, Meth. Pl. 457. 1794; not Chamaedaphne Mitchell, Diss. Brevis Principiis Bot. Zool. 1769 |= Mitchella L.|.*) TYPE AND SOLE ® Chamaedaphne Moench (1794) is clearly a later homonym of Chamaedaphne Mitchell, first published in 1748 and taken by Linnaeus as the basis for Mitchella L. (1753) but republished validly in 1769. The earliest legitimate name appears to be 1961] WOOD, GENERA OF ERICACEAE 55 SPECIES: Andromeda calyculata L. = Cassandra calyculata (L.) D. Don. (A poetic name, for Cassandra, a daguhiet of Priam, King of Troy, and Hecuba. ) Cassandra calyculata (Chamaedaphne calyculata (L.) Moench) is of almost circumpolar distribution, ranging from Scandinavia, the Baltic States and northern Poland to aovihérg Russia and northern Asia, Kam- chatka, Manchuria, and northern Japan; and from Alaska to Labrador and Newfoundland, southward to British Columbia, Alberta, Minnesota, the northern parts of Illinois, Indiana, and Ohio, New York, Pennsylvania, and New Jersey, and southward i in more or less isolated iecsnues mostly on the Coastal Plain, in Delaware, Maryland, North Carolina (there also inland in Henderson County), and South Carolina. Three geographical varieties have been distinguished: var. calyculata in Eurasia: var. an- gustifolia (Ait.) Rehd., the too similar widespread American form; and var. latifolia (Ait.) Fern., a dwarfish broad-leaved form of Labrador, New- foundland, and Nova Scotia, westward in the north to Mackenzie. FIG. Cassandra. a-i, C. calyculata: a, baa twig. & 1%; b. flower, & 2 c, inner and lateral views of stamens, : calyx and gynoecium with disc, x 6; vertical section — note columnar placentae ola Nika ovules, 5; g, ovary. diagrammatic cross section, * ee ned capsule — note five outer. ten inner valves, X 5; i, seed, raphe to left, x 2 A low (exceptionally to 1.5 m.) and much-branched evergreen shrub with coriaceous, scurfy leaves, Cassandra is one of the most abundant and characteristic plants of bogs in the glaciated portions of eastern North Cassandra D. Don (1834) by which this plant was known to botanists of the nine- teenth century. Although conservation may be advocated by some, such a course trusion of Harrimanella upon this classic bevy. (pages 28, 29, in “Some Historical a pects of Plant Taxonomy,” Rhodora 44: 21-43. 1942). 56 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII America. It shows a wide tolerance, being an early invader of bogs, but often persisting long after the establishment of tree species. In some bog areas the plant is completely dominant, forming a closed association with Sphagnum and several other mosses. South of the glacial boundary, with the exception of southern New Jersey, the colonies appear to be rare and scattered. The horizontal inflorescences terminate the growth of the year, and flower buds develop in late summer; the plants flower from March in the South to early July in Labrador; a partial second flowering sometimes occurs in the fall. The reduced leaves of the inflorescence are deciduous after maturity of the fruit (that portion of the shoot then dying) but the larger leaves below persist another year. Growth of the shoot is continued after flowering by one or more axillary buds below the inflorescence. The small, brown seeds have a large raphe of spongy tissue and appear to be capable of floating. The roots have been reported to lack mycorrhizae under bog conditions but to show such an association under drier condi- tions in cultivation. The leaves are poisonous, containing andromedotoxin, Cassandra is a well-marked and apparently isolated genus distinctive in the bracteoles beneath the calyx, the stamens, the inflorescence, peltate scales, fruit, and seeds. The closest relationship seems to be with An- dromeda, Pieris, and Lyonia. REFERENCES: Under family references see also BELL & BURCHILL, Cox (Ericales I1), GRE- VILLIUS & KIRCHNER (pp. 57-62), PALSER (1951, 1952), PELTRIsoT and STAN- LEY. FERNALD, M. L. Chamaedaphne oe. (L.) Moench, var. latifolia (Ait.), comb. nov. Rhodora 47: 390, 391 Gates, F. C. The relation of snow cover to winter killing in Chamaedaphne calyculata. Torreya 12: 257-262. 1912. . Winter as a factor in the xerophily of certain selec ericads. Bot. Gaz. 57: 445-487. 1914. [Includes many data on Cassan . s of northern lower Michigan. Ecol. “pil 12: 213-254. 1942. |The ee association, 238-240. | Lems, K. Ecological study of the peat bogs of eastern North America. III. Notes on the behavior of ec ie calyculata, Canad. Jour. Bot. 34: 197-207. 1956. Lovett, J. H., and H. B. Lovett. Pollination of the Ericaceae: Chamaedaphne and Xolisma. Rhodora 37: 157-161. 1935. [Cassandra and Lyonia ligus- trina. | SEGADAS- Tae. F. Ecological study of the peat bogs of eastern North America. II. The pa retivis oe community in Quebec and panos. Canad. Jour. Bot. 33: —68 Sims, J. Andromeda calyculata var. pi aaisy heh. Mag. 32: pl. 1286. 1810. 13. Oxydendrum A. P. de Candolle, Prodr. 7: 601. 1839. A deciduous tree to 25 m., with a tall, straight trunk to 25-50 cm. in diameter, the bark deeply fissured. Winter buds small, axillary. Leaves membranaceous, oblong-lanceolate, petiolate, serrulate, pale beneath (remi- 1961 | WOOD, GENERA OF ERICACEAE ay niscent of Prunus Persica). Flowers bisexual, 6-8 mm. long, in drooping panicles (terminating the growth of the year) of 6 or more slender, de- clined racemes, these 1-sided (by the bending of the pedicels); each flower in the axil of a minute, deciduous bract, each pedicel with 2 minute, alternate, persistent bractlets. Calyx deeply divided into 5 separated lobes, valvate in bud. Corolla white, cylindric-ovoid to ovoid, puberulous, with S short, spreading or recurved lobes imbricate in bud. Stamens 10, hypog- ynous, free from the corolla; filaments broad, stout, swollen at the base; anthers appressed to the style, fixed near the base, linear-oblong, opening from the apex to above the middle by slitlike pores, the connective near anther apex with a region of white disintegration tissue: pollen tetrads without viscin strands. Stigma simple, unexpanded: style columnar, straight, slightly exserted; ovary superior, 5-locular, extending upward around the base of the style; ovules numerous on outer, lateral and upper surfaces of the conspicuous placentae, all near bottom of locule and ex- tending upward. Pedicels recurving in fruit, the ovoid-pyramidal capsule erect, ca. 5 mm. long, crowned by the persistent style, loculicidally 5- valved; seeds numerous, elongate, ascending, the seed-coat pointed at the ends; embryo minute. Type AND SOLE sPECIES: O. arboreum (L.) DC. (Andromeda arborea L.). (Name from Greek, oxys, sour, and den- ~e a re VN edn ic ae A NS oO Ry dy = Vabigs I GAS Mee ae AN 5 7 nd ol: % ? LRT YEN eae yO eM Fic. 8. Oxydendrum. a-g, O. arboreum: a, flowering branch. « 14; b, flower, 2; c, flower, in section, 6; d, outer, inner, and lateral views of stamens — note area of disintegration tissue on upper part of outer surface of each anther-half. ; €, portion of raceme with immature fruits, X 2; f. opened capsule, one 10. >< valve removed — note deeply immersed style, X 4; g, seed, & 58 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLII dron, tree, referring to the acid taste of the leaves.) — Sourwoop, TITI, SORREL-TREE. Oxydendrum arboreum is of wide distribution in the eastern United States in acid, well-drained soils in dry woods, on bluffs and ridges, in hammocks and creek bottoms, from southern Pennsylvania to northern and western Florida and to southern Ohio, southern Indiana and Illinois, and Mississippi and southeastern Louisiana. One of the largest of the Ericaceae, it is a handsome, slow-growing tree of great ornamental value both for its panicles of small, white flowers in June or July and for the showy red autumn coloration. It reaches its greatest size on the western slopes of the Smoky Mountains of eastern Tennessee. The plant is easily raised from seeds and transplants easily. It is hardy northward at least to eastern Massachusetts. The species is an isolated one among the Ericaceae, apparently with no close relatives. Largely on the basis of the rather advanced evolutionary level of the wood anatomy in comparison with other members of the An- dromedeae, Cox concluded that Oxydendrum is “the survivor of a com- paratively long evolutionary sequence paralleling that of the other tribes,” and placed it in a separate tribe, Oxydendreae. It is anatomically notable for the tapered receptacle with traces to all floral organs completely dis- tinct. REFERENCES: See also under family references Cox (Ericales II), PAtser (1951, 1952), and PELTRISOT. BaLpwin, J. T. Cytogeography of Oxydendrum arboreum. Bull. Torrey Bot. Club 69: 134-136. 1942. [Includes distribution map; 31 collections showed Crovis, J. F. A note about Oxydendrum. Castanea 24: 51-53. 1959. [Calls attention to pendent flowers vs. erect fruit. Esson, J. G. The sourwood —a neglected tree. Jour. N. Y. Bot. Gard. 51: 12-15. 1950. [Horticultural merits. ] Lewis, C. E. Tips for better landscapes; the sorrel tree. Am. Nurseryman 105(4): 14. 15. 1957.* Nasu, G. V. Oxydendrum arboreum. Addisonia 4: 37, 38. pl. 139. 1919. SARGENT, C. S. Oxydendrum. Silva N. Am. 5: 133-136. pl. 235, 1893. Sms, J. Andromeda arborea. Bot. Mag. 13: pl. 905. 1806. 14. Epigaea Linnaeus, Sp. Pl. 1: 395. 1753; Gen. Pl. ed. 5. 186. 1754. Prostrate evergreen shrubs with creeping stems. Leaves alternate, oval to elliptic, petiolate, coriaceous. Plants functionally dioecious (at least in E. repens), the flowers in short terminal and axillary bracteate spikes, the pedicels with 2 bractlets at the base. Flowers either carpellate (with rudimentary stamens, usually only filaments) or apparently bisexual, but with unexpanded stigmas and functionally staminate. Sepals 5, imbricate, conspicuous, green. Corolla salverform, with 5 spreading [or recurved | lobes, in staminate flowers about 1/3 larger than in carpellate. Stamens 1961] WOOD, GENERA OF ERICACEAE 59 of staminate flowers 10; filaments adnate to the base of corolla tube, half as long to as long as corolla tube; anthers elongate, the halves joined almost to the tip, lacking white disintegration tissue, opening by longi- tudinal slits (those of carpellate flowers, when present, not opening); pollen tetrads with viscin strands. Stigma 5-lobed, radiate and glandular in carpellate flowers, the lobes erect and dry in staminate flowers; styles half as long to as long as corolla tube; ovary superior, 5-locular, glandular- hairy; placentae 2 in each locule, axile below but parietal above (cf. Pyroleae, Monotropeae) with numerous anatropous ovules. Capsule de- pressed-globose, fleshy, glandular, hirsute, loculicidal, the valves thin; seeds numerous, minute, hard, brown, shining, on white, succulent placen- tae filling the locules. Type species: FE. repens L. (Name from Greek, e/7, on, and gaca, earth, in allusion to the creeping habit.) — TRAILING ARBU- TUS. Two species, Epigaea repens, of wide distribution in eastern North America (Labrador to northern Florida, Alabama, Tennessee, Indiana, Michigan, Minnesota, and Saskatchewan), and £. asiatica Maxim., of Japan to Formosa. Although closely related, E. asiatica differs from FE. repens in the more acute, coriaceous leaves, the longer pedicels (to 1.5 cm. vs. 5 mm.), the greater number to flowers ver inflorescence, the broader corolla tube and shorter lobes, and other details. Epigaea appears to be most closely related to the monotypic Orphanidesia Boiss., of the moun- tains inland from the southeast corner of the Black Sea, a similar low ever- green with 1—3 flowered racemes and a larger, cup-shaped corolla, but with anthers opening as in Epigaea. Within the Andromedeae both genera seem to be isolated, although, on the basis of wood structure, Cox placed Epr- gaea, Agauria, Cassiope, and Enkianthus together as Cassiopeae Cox. Epigaea repens is noteworthy for its functional dioecism, the plants bearing either carpellate flowers with only rudimentary stamens (usually only the filaments developed) or seemingly bisexual but functionally staminate flowers with well-developed stamens and apparently functional embryo sacs, but stigmas which are not receptive to pollen. Although fila- ment and style length vary, pollen grains are of a single size, and experi- mental evidence is all against any form of heterostyly. Epigaea astatica does not appear to have been studied in this respect, but two artificial hybrids with E. repens (one a backcross to E. repens), both were of the pistillate type. The roots do not develop root hairs, and a mycorrhizal association is obligate; as a result, the species transplants with difficulty. Although the plant seems to tolerate a fairly wide range of moisture and soil acidity (pH 7.6—4.5), constant moisture, good drainage, some shade, and acid soil apparently are required for optimum growth. The plant was formerly gathered in large quantities for the fragrant pink to white flowers produced in earliest spring, and has been very nearly exterminated in some areas. The fruit requires 40-55 days to mature, and the seeds presumably are dispersed by ants, which have been observed attacking the fleshy placentas. 60 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLII REFERENCES: Under family references see also BELL & BURCHILL, BERGMAN, Cox (Ericales Il), Ikuse and aioe (1951, 1952). ANDREWS, M. © Antonie von Epigaea repens L. Beih. Bot. Centr. 19(1): 314-320. i . . 1906. The embryo-sac and pollen of Epigaea repens. Ibid. 44(1): 264-266. 27. ARLAND, F. L. Growing the trailing arbutus. Jour. N. Y. Bot. Gard. 46: 102— 109. 1945. [E. repens. | Barrows, F. L. Propagation of Epigaea repens L. I. Cuttings and seeds. Contr. Boyce yee Inst. 8: 81-97. 1936. II. The endophytic fungus. /bid. 11: 431-440. 194 . How to Nene trailing arbutus. Pl. Gard. II. 3: 49-53. 1947. BasTIN, E. 5. Structure of Epigaea repens. Am. Jour. Pharm. 67: 213-236. 1895.* CouNcILMAN, W. T. The root system of Epigaea sil and its oe to the fungi of the humus. Proc. Natl. Acad. Sci. 9: 92: CovIL_e, F e use of acid soil for raising ie of re mayflower Epigaea repens. Science 33: 711, 712. 1911 The cultivation of the mayflower. Natl. Geogr. Mag. 27: 518, 519, Ok 1915 Epwarps, S. Epigaea repens. Bot. Reg. 3: pl. 201. 1817. Gray, A. Vieiemmancities in Epigaea. Am. Jour. Sci. III. 12: 74-76. 1876. (Reprinted, Am. Nat. 10: 490-492. 1876.) Hatstep, B. D. Notes upon Epigaea repens. Bull. Torrey Bot. Club 18: 249, 250. 1891. [Comments on supposed heterostyly. | Hoim, T. Medicinal plants of North America. 73. Epigaea repens L. Merck Rep. 22: 144-146. 1913.* Lancpon, F. E. A study of Epigaea repens. Asa Gray Bull. 2: 1-3. 1894. | Observations on dioecism. | Mvtuican, B. O. Epigaea hybrids. Jour. Roy. Hort. Soc. 64: 507-510. figs. 114-116. 1939. [E. X& intertexta Mulligan E. astatica & repens. | STAPF, O. Epigaea asiatica. Bot. Mag. 154: pl. 9222. 1930. STEVENS, N. E. Dioecism in the trailing arbutus, with notes on the ep haa of the seed. Bull. Torrey Bot. Club 38: 531-543. 1912. [Plants fun ally dioecious; no evidence of heterostyly; mal. | STOKER, F. st saa Sud seceoealy tiie Jour. Roy. Hort. Soc. 65: 210. 211. figs. 55,5 | ription from living plants in cultivation. | Warp, L. F. oe, ecco in Epigaea repens. Am. Nat. 14: 198-200. on- embryo-sac Lan pennty nor- Weep, C. M. The guests of the mayflower. Pop. Sci. Monthly 45: 17-23. 1894.* Witson, K. E. Double flowers of the Epigaea repens. Bot. Gaz. 15: 19, 20. 1890. Witsoxn, W. P. Observations on Epigaea repens, L. Contr. Bot. Lab. Univ. Penn. 1: 56-63. pl. 8. 1893. {Dioecism; — 1000 plants studied; in- cludes illustration of various tvpes of flowers. | 1961] WOOD, GENERA OF ERICACEAE 61 Tribe PyroLeaE D. Don 15. Chimaphila Pursh, Fl. Am. Sept. 1: 279, 300. 1814. Low, suffruticose to nearly herbaceous plants with long, scaly, creeping subterranean rhizomes and short, evergreen aerial shoots with alternate to subopposite or almost whorled, coriaceous, often shining, serrate, short- petioled leaves. Rhizomes with distant scale-leaves, each with an axillary bud and an associated root; roots reported to form adventitious buds. In- florescence a stalked, terminal 1—8-flowered corymb, each flower in the axil of a deciduous [or persistent] bract, the pedicels lacking bracteoles. Flowers 5-merous, regular, fragrant, nodding. Sepals 5, united at the base, persistent. Petals 5, distinct, white to pink, concave, orbicular, forming a saucer-shaped corolla. Stamens 10; filaments enlarged and hairy at or below the middle; anthers extrorse in bud but at anthesis inverted by the inflexion of the apex of the filament, opening by a pore at the apparent apex (the base) of each of the divergent anther-halves (cf. Arbutus, Arcto- staphylos); pollen in tetrads, 3-colpate. Stigma peltate, 5-lobed; style very short, top-shaped, immersed in the summit of the ovary; ovary su- perior, depressed-globose, 5-locular below, the placentae axile, but 1-locu- lar above, the placentae becoming parietal (cf. Epigaea) ; ovules numer- ous, anatropous; disc circular, nectariferous, at the base of the ovary. Capsule depressed-globose, erect, 5-lobed, loculicidally 5-valved, the valves smooth on the edges; seeds small, numerous, with a thin cellular coat, an endosperm of a few large cells and a minute, undifferentiated embryo. LECTOTYPE SPECIES: C. maculata (L.) Pursh; see Britton & Brown, Ill. Fl. North. U. S. ed. 2. 2: 672. 1913. (Name taken from the common name ‘“‘wintergreen,” from Greek, cheima, winter, and philos, friend.) — WINTERGREEN. A well-marked boreal genus of about four or five species, including Chi- maphila umbellata (L.) Bart., of circumboreal distribution with disjunct stations southward in Mexico and Guatemala and a close relative, C. domi- nigensis Blake, in Hispaniola; vs japonica Miq., in eastern Asia; C. Men- ziesit (R. Br.) Spreng. (2n = 26), from British Columbia to California, Sonora, and Jalisco; and C. maculata in eastern North America and Cen- tral America. A number of geographical variants have been recognized in the wide- spread C. umbellata. The plant of eastern North America is var. cisat- lantica Blake, which occurs from eastern Quebec to western Ontario, southward to Virginia, North Carolina, and Georgia, Ohio, Michigan, northeastern Illinois, and Minnesota; var. umbellata (2n = 26) is of Eurasian distribution. Chimaphila maculata ranges widely in dry woods, from southern New Hampshire westward to Michigan and northeastern Illinois, southward to Georgia, Alabama, and Tennessee, and, as a more robust form (apparently with larger flowers and fruits), from the mountains of Chihuahua and Sonora southward to Guatemala, Honduras, and Costa Rica. 62 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII The genus shares many features with Pyrola, from which it is readily distinguished by the corymbose inflorescence, enlarged staminal filaments, presence of a disc, and fruit with non-cobwebby valve margins. REFERENCES: —. S. F. A new Chimaphila from San Domingo. Jour. Bot. 52: 169. 1914. C. dominigensis, closely related to C. umbellata. | The varieties of Chimaphila umbellata. Rhodora 19: 237-244. 1917. Camp, W, H. Studies in the Ericales. IV. Notes on Chimaphil1, Gaultheria and Pernettya in Mexico and adjacent regions. Bull. Torrey Bot. Club 66 939. 7-28. CopeLAND, H. F. Observations on the structure and classification of the Pyro- leae. Madrono 9: 65-102. 1947. HENDERSON, M. W. A comparative study of the structure and saprophytism of the Pyrolaceae and Monotropaceae with reference to their derivation from the Ericaceae. Contr. Bot. Lab. Univ. Penn. 5: 42-109. 1920. Horm, T. Medicinal plants of North America. 28. Chimaphila umbellata L. Merck Rep. 18: 143-145. 1909.* . The flower of Chimaphila. Rhodora 29: 1-6. 1927. Lemmon, R. S. Propagation of the native species of Chimaphila. Jour. N. Y. Bot. Gard. 39: 129, 130. 1938. [Propagation by seeds and runners unsuc- cessful; success with cuttings with a stub taken just before young leaves reached full size; a summer and winter required before rooting began. | MEEHAN, T. Chimaphila umbellata. Meehans Monthly 7: pl. 9. 1897. Chima- phila umbellata. Ibid. 9: 1, 2. MOUvILLEFARINE, E. Le Chimaphila maculata Pursh aux environs de Paris. Bull. Soc. Bot. Fr. 49: 281-284. 1903. ScHWaARTING, A. E., and L. D. HuNer. A cai study “ agi a um- bellata. Jour. Am. Pharm. Soc. Sci. Ed. 32: -187. 19 Stms, J. Pyrola umbellata. Bot. Mag. 20: pl. a 1804. a maculata. Ibid. 23: pl. 897. 1806 16. Pyrola Linnaeus, Sp. Pl. 1: 396. 1753; Gen. Pl. ed. 5. 188. 1754. Low, smooth, evergreen perennial herbs, with slender, creeping sub- terranean rhizomes (as in Chimaphila) and erect shoots bearing a cluster of petioled, coriaceous, persistent basal leaves and/or a raceme of | 2]6—20 ascending, spreading or nodding flowers on an upright, more or less scaly- bracted scape; flowers in the axils of bracts, lacking bractlets on the pedi- cels. Flowers 5-merous, nodding, irregular |or regular]. Calyx 5-parted, vrei Petals 5, concave, distinct, deciduous, [greenish-|white | to ], imbricate in aestivation. Stamens 10; filaments glabrous, not en- nent usually bent, the anthers then clustered on the upper side of the flower; anthers extrorse in the bud but in flower inverted by the inflexion of the apex of the filament, opening by a pore at the apparent apex (base) of each of the anther-halves, constricted [or not] below the pores, some- times mucronate at the apparent base; pollen in tetrads, 3-colpate. Stigma of 5 narrow, erect lobes extending from the top of the style which forms a rim beneath the lobes; style [short to] long, [straight to] declined with the apex curved upward; ovary 5-locular below and the placentae axile, 5" a 1961] WOOD, GENERA OF ERICACEAE 63 but 1-locular above, the placentae becoming parietal; ovules numerous, anatropous; disc lacking. Capsule 5-lobed, loculicidally 5-valved, the valves cobwebby on the edges; seeds minute, as in Chimaphila. (Excluding Ramischia Opiz.) Lectotypr species: P. rotundifolia L.; see Britton & Brown, Ill. Fl. North. U. S. ed. 2. 2: 668. 1913. (Name Latin, diminu- tive of Pyrus, pear, from the somewhat similar leaves.) A circumboreal genus, sometimes overestimated at 40 species, distributed southward along the higher mountains to Central America (Guatemala) and southeastern Asia (Sumatra). About six species occur in northeastern North America. The circumboreal Pyrola rotundifolia L. (2n = 46) reaches our area as var. americana (Sweet) Fern., distributed from Quebec to Ontario, southward to North Carolina, Tennessee, Kentucky, northern Indiana, Michigan, Wisconsin, and Minnesota. A report of P. elliptica Nutt. from Tennessee is unconfirmed, the species apparently not being known with certainty south of West Virginia. The few chromosome numbers thus far determined support the separa- tion of Ramischia Opiz (Pyrola secunda L., 2n = 38) and of Moneses S. F. Gray (M. uniflora (L.) A. Gray, 2n = 26), but not that of Erzle- benia Opiz (P. minor L., 2n = 46), from Pyrola in which the diploid numbers are 46 or 92. Both Pyrola and Chimaphila show tendencies toward vegetative reduc- tion which culminate in the herbaceous Moneses uniflora, in which the rhizome has been eliminated, the aerial shoots arising from adventitious buds on the roots. The four genera of the tribe form a natural group which seems to have its closest affinities with Andromedeae and Arbuteae, al- though there can hardly be a direct relationship. Aside from the tendency toward reduction of the plant body and of the size and complexity of the seed and embryo, the group has no combination of features which distin- guishes it from the Ericaceae proper. REFERENCES: See also under family references CopPELAND (1933, 1947), PELTRISOT, WARM- NG. Anpres, H. Piroleen-Studien. Beitrage zur Kenntnis der Morphologie, Phyto- geographie und allgemeinen Systematik der Pirolaceae. Verh. Bot. Ver. Brandenburg 56: 1-76. 1914 Studien zur speziellen Systematik der Pirolaceae. IV. oe der Gat- tung Ramischia Opiz. Repert. Sp. Nov. 19: 209-224. 192 Camp, W. H. Aphyllous forms in Pyrola. Bull. Torrey Bot. a 67: 453-465. 1940. FERNALD, M. L. The American representatives of Pyrola rotundifolia. Rhodora 6: 197-202. 1904. ———. The American varieties of Lake a pee 22: 49-53. 1920. . Pyrola rotundifolia and P. americana. Ibid. 1 FurtH, P. Zur Biologie und Mikrochemie einiger ae Arten. Akad. Wiss. Wien. Sitz-ber. 129: 559-588. pl. 1. 1920.* Hoi, T. Pyrola aphylia: a morphological study. Bot. Gaz. 25: 246-254. pl. 17. 64 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII 1898. [Notes adventitious buds from roots of several species; also data on Chimaphila. | Hu ten, E. The amphi-Atlantic plants and their phytogeographical momrvar Sv. Vet-akad. Handl. IV. 7(1): 1-340. 1958. [p. 142, maps 123, rotundifolia, P. grandiflora, P. asarifolia; p. 268, map 250, P. as P. rentfolia. Inouye, H., and Y. Kanaya. Studies on the constituents of pirolaceous plants. VII-VHUI. (In Jae English summary.) Jour. Pharm. Soc. Japan 78: 298-303. 1958. Lovett, H. B., and J. H. Lovett. Pollination of the Ericaceae: IV. Ledum and Pyrola. Rhodora 38: 90-94. 1936. [P. elliptica, P. americana. | SKAN, S. A. Pyrola uliginosa, P. bracteata. Bot. Mag. 143: pl. 8710A, B. 1917. Sources, R. Embryogénie des Ericacées. Développement de Vembryon chez le Pyrola ‘ie aie L. Compt. Rend. Acad. Sci. Paris 209: 635-637. 1939. Witcke, J. Dissemination of seeds of Indianpipe and Pyrola. (In Dutch.) ine Nat. 57: o 11. 1954.* Subfam. MONOTROPOIDEAE Gray (‘Monotropeae’) Tribe MoNoTRoPEAE D. Don 17. Monotropsis Schweinitz in Elliott, Sketch Bot. S. C. & Ga. 1: 478. 1817 Low, smooth, reddish to purplish-brown (violet, pink, or white) plants with the aspect of Monotropa Hypopithys, lacking chlorophyll, the stem bent, emerging from the ground center part first, the plant hidden be- neath fallen leaves, flowering in early spring. Rhizome lacking, erect flowering stems from adventitious buds from the roots, as in Monotropa. Stems with spirally arranged obovate, obtuse, scalelike bracts, the flowers 1-16 in a dense terminal raceme, each flower in the axil of a bract and the pedicel with 2 bracteoles. Flowers 5(4—-8)-merous, nodding, very fra- grant. Calyx of 5 oblong-lanceolate erect, distinct, persistent, scalelike sepals. Corolla campanulate, sympetalous, 5-lobed, slightly 5-gibbous at the base, the lobes ovate, 1/3-1/4 the length of the tube, persistent. Stamens 10; filaments filiform; anthers short, with a conspicuous con- nective, unappendaged, 4-locular at first but becoming 1-locular, horizontal at first but becoming inverted in flower (as in Pyrola and Chimaphila) and opening by 2 terminal pores in the apparent apex; pollen grains single, 2-colpate. Stigma large, capitate, not evidently lobed; style short, thick: ovary superior, 5(4)-locular at the base, each locule with a divided axile lacenta, becoming t-locular above, the placental halves then becoming parietal forming 5(4) conspicuous placentae; ovules anatropous, the in- tegument of 2 layers of cells, embryo sac of the Polygonum type; disc 10-lobed, the lobes in pairs, each clasping the base of a petalad stamen. Fruit a subglobose berry remaining inclosed in the persistent corolla. filfed by the enlarged placentae and the numerous minute, ovoid, un- appendaged, red or orange seeds; seeds with an endosperm of a few cells and a globular (4-celled?) embryo. (Schweinitzia Nutt., Cryptophila 1961 | WOOD, GENERA OF ERICACEAE 65 Wolf.) Type species: M. odorata Schwein. in Ell. (Name composed of Monotropa plus Greek, opsis, appearance, resembling Monotropa.) — SWEET PINE-SAP. A distinctive genus of one or two species of the southeastern United States. Monotropsis odorata ranges at least from eastern Maryland and Virginia to central North Carolina, eastern Tennessee, northern Georgia, and northern Alabama. Since the plant flowers and fruits under a cover of leaves it is usually overlooked, and it is possible that its range is much wider, especially southeastward. The cycle of the plant above ground extends from the fall of one year, when the flowering shoots first appear beneath the fallen leaves, through anthesis early the following spring (March—April), to fruiting in July or early August. Baldwin has demon- strated the changing proportions of corolla tube/lobes from bud to an- thesis and has effectively disposed of M. Lehmaniae Burnham as the im- mature fall phase of M. odorata. The status of M. Reynoldsiae Gray, de- scribed from scrub-oak thickets near St. Augustine, Florida, and known elsewhere only southward in the Indian River region, is dubious but un- settled. In M. odorata the colored corolla is as long as, shorter, or longer than, the oblong sepals, while the white corolla of M. Reynoldsiae is about twice as long as the ovate or ovate-lanceolate sepals. Wolfe, in his careful study of M. odorata in Alabama, found, however, a wide range in size (4-20 cm.) and color (reddish pink to brownish red, light lavender to darker shades, pink, white with pink vascular bundles, or pure white) of the plant and a wide variation of the length of the sepals (7-13 mm.). As in Monotropa, neither size nor color of plant or corolla can be regarded as diagnostic, and corolla/sepal ratio should be treated with the utmost suspicion. Pollen is shed in bud on the inner surface of the corolla lobes, and pol- lination is by insects attracted by the strong, spicy fragrance at anthesis. Seed disperal apparently is by ants which attack the berry with its fleshy, sweetish placentae; the seeds are relatively plump and hard, well adapted to this habit, in contrast to the slender, tailed, wind-dispersed seeds of Monotropa. (Compare other baccate-fruited vs. dry-fruited Monotro- poideae; Epigaca,; Gaultheria and Leucothoé.) Monotropsis is of particular interest for the 2-bracteolate pedicels, the well-developed calyx of five sepals, the sympetalous corolla, and the rela- tively unspecialized anthers. Its relationships are with Monotropa, Pitvo- pus Small, Monotropastrum Andres, the two last of which also have bac- cate fruit. All four are more specialized than Pterospora Nutt., Sarcodes ‘orr., and Allotropa Gray, which constitute the tribe Pterosporeae. Ptero- spora is especially notable for the appendaged anthers which strongly suggest those of Arbuteae. Aside from the reductions and specializations to be expected in connection with the parasitic habit, there seems to be no basis for setting the Monotropoideae apart as a separate family, for the details of morphology, anatomy, and embryology fit excellently with those of other Ericaceae. 66 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII REFERENCES: A number of references pertaining to ee other than ours are lal here. See also CopELAND (1941) under Monotropa. BaksHI, T. S. Ecology and morphology of Pterospora andromedea. Bot. Gaz. 120: 203-217. 1959. BaLpwin, J. T., Jr. Monotropisis Lehmaniae not a real species. Rhodora 59: 259-262. 1957. BurNuHaM, S. H. A new species of Monotropsis. Torreya 6: 234, 235. 1906. [| M. Lehmaniae. | CopELAND, H. F. The acne of the flower of Newberrya. Madrono 2: 137- 142. 1934. |[Hemitom . On the genus Pityopus Ibid. 3: 154-168. 1935 . The reproductive structures of Pleuricospora. tid 4: 1-16. 1937. . The structure of Allotropa. Ibid. 137-153. ———. The structure of Monotropsis and the 22 ie of the Monotro- poideae. /bid. 5: 105-119. 1939. Domin, K. Vergleichende Studien iiber den Fichtenspargel mit Bemerkungen uber Morphologie, Phytogeographie, Phylogenie und systematische Gliede- rung der Monotropoiden. Sitz-ber. Bohm. Ges. Wiss. Math.-Nat. Cl. 1915: 1-111. 1915 Doyet, B. E., and L. sr Goss. Some details of the reproductive structures of Sarcodes. Madrono 6: 1-7. 1941. Gray, A. Chloris spe -americana. Mem. Am. Acad. Arts Sci. II. 3: 1-56. pls. 1-10. 1848. [M. odorata, 15-20, pl. 2.] HeENpeERSON, M. W. A comparative aude of the structure and saprophytism of the Pyrolaceae and Monotropaceae with reference to their derivation from the Ericaceae. Contr. Bot. Lab. Univ. Penn. 5: 42-109 ' HoweELtL, J. T. A year with Pityopus. Leafl. West. Bot. 6: 57-61. 1950. McDovealt, D. T., and F. E. Lioyp. Roots and mycorrhizas of some Mono- tropaceae. Bull. N. Y. Bot. Gard. 1: 419-429. pls. 10-12. 1900. | Ptero- spora, Sarcodes, Monotropa. | Putt, C. D. Notes on Monotropsis odorata. Rhodora 11: 153. 154. 1909. SMALL, J. K. Monotropaceae. N. Am. FI. 29: 11- “18. 1914. SPAWN, W. Notes on Monotropsis. Bartonia 20: 7-10. 1940. WoLr, W. Notes on Alabama plants. A new pena plant. Am. Midl. Nat. 8: 104-129. 1922. [Extensive observations on M. odorata, as Crypto- phila pudica. | 18. Monotropa Linnaeus, Sp. Pl. 1: 387. 1753; Gen. Pl. ed. 5. 183. 1754 Low, fleshy, tawny to reddish to translucent white herbs, lacking chloro- phyll, blackening in drying. Main root thick, horizontal, with a surround- ing mass of short, fleshy, intertwined secondary roots, each incased in a mycelial sheath; rhizomes lacking, the stems from adventitious (endog- enous) buds from the roots. Flowering stems with scales or bracts, the tip nodding at first (erect in fruit), with 1-15 flowers in the axils of bracts, bractlets absent. Flowers 5-merous or the terminal 5-merous, the others 3- or 4-merous. Calvx(?) of 1-5 lanceolate, deciduous, bractlike scales. Corolla appearing campanulate, of 5(3 or 4) erect, scalelike. distinct petals 1961] WOOD, GENERA OF ERICACEAE 67 saccate at the base and tardily deciduous. Stamens 10 or 8; filaments subulate, the outer whorl (opposite petals) shorter than the inner; an- thers short, kidney-shaped, becoming 1-locular at anthesis, opening trans- versely through 2 curving slits; pollen grains single, 2- or 3-colpate. Stigma funnel-shaped to discoid, obscurely lobed opposite the petals; style colum- nar, thick and fleshy; ovary superior, 10—8-lobed, 5- or 4-locular at the base and 1-locular above with 5 or 4 bilobed parietal placentae; ovules anatropous, very numerous; disc 10- or 8-lobed, the lobes in pairs clasp- ing the bases of the petalad stamens. Capsule ovoid, 10—8-grooved, erect, loculicidally 5- or 4-valved, the very thick placentae covered with numer- ous minute, elongate, tailed seeds with a very loose coat, very few large endosperm cells and a 9—2-celled embryo. (Hypopitys Hill.) LEctoTyPE species: M. uniflora L.; see Britton & Brown, Ill. Fl. North. U.S. ed. 2. 2: 674. 1913. (Name Greek, monos, one, and tropos, turn, the summit of the stem nodding, bent to one side.) —— INDIAN PIPE, PINE-SAP. A small, but widely distributed, genus of parasitic herbs of the North- ern Hemisphere, including two or more species, both occurring in our area. Section Hypopitys Gray (Hypopitys of authors), with several-flowered racemes, the terminal one usually 5-merous, the others 3- or 4-merous, the sepals 3-5 (mostly as many as the petals), the anthers opening by a continuous line into two very unequal valves, and the style longer than the ovary, includes M. Hypopithys L. (Hypopitys lanuginosa (Michx.) Nutt., H. americana (DC.) Small), pine-sap, a woodland plant distributed from Newfoundland to British Columbia, southward to Iowa, Louisiana, and Florida in the East and to Arizona in the West, with a wide Eurasian range, as well. The species is quite variable in size of plant, color, pubes- cence, indentation of margins of bracts, and proportions of ovary and style, on the basis of which a number of species have been erected. Only a single species seems to be represented in North America, but a careful study of variation is needed. Diploid chromosome numbers of European material of M. Hypopithys and M. Hypophegea Wallr. (M. Hypopithys var. glabra Roth), which differ in pubescent vs. glabrous filaments, styles, and inner surface of the petals, have been reported as 48 and 16 respec- tively. Section Monotropa (§ Eumonotropa Gray), differing in the 1-flowered shoots, the more reduced calyx of 1—4 scales, the transverse anthers opening by two short slits, and the short, thick style, includes only M. uniflora L. (M. Brittonii Small), Indian pipe. A frequent woodland plant of North America from Newfoundland and Quebec to British Columbia, southward to Oregon, Montana, Minnesota, Missouri, Oklahoma, Texas, and Florida, the species is disjunct southward in Central America in the mountains of central Mexico, Guatemala, and Columbia, and occurs in Japan, central China, and the Himalayas, as well. A number of variants in color, size, and other minor characters have been proposed, but only a single wide- ranging species seems to be involved. Monotropa Hvpopithys is segregated by some as Hypopitys Hill, on 68 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII the basis of the several flowers, the stem anatomy (a flattened cylinder vs. a ring of ten bundles), and the number of grooves on the pollen grains (2 vs. 3), but the plants are so suar in all other features that they are here maintained in a single genus.” The wide range of habitats from which Monotropa has been recorded (moist woods to sand dunes to sphagnum bogs, but always in association with trees) and the tight mycorrhizal underground masses of the plant are more in keeping with a parasitic habit than with the saprophytic mode often assigned to it. In a recent study Bjorkman found that the same mycorrhizal fungus occurs in M. Hypopithys and adjacent trees; that the mycorrhizal masses of Monotropa are associated with the roots of trees (although there is no direct connection between Monotropa and tree); and, on the basis of trenching and radioactive tracer experiments, that it is probable that ‘“‘Monotropa obtains its energy requirements from the trees via the mycorrhizal fungus common to them bot There is still some question as to whether the “seals” of Monotropa are bracts or sepals REFERENCES: Under Monotropsis see especially CopELAND (1939), Domrn, HENDERSON, and McDoveat & Lioyp; under family references see BERGMAN and PELTRISOT. ByORKMAN, E. On the metabolism of Monotropa Hypopitys L.: an epiparasite on forest tree roots. (Abs.) Proc. IX Int. Bot. Congr. 2: 35, 36 CaMPBELL, D. H. Monotropa uniflora as a subject for demonstrating the em- bryo-sac. Bot. Gaz. 14: 83. 1889. {Whole mounts of placenta and ovules in 3% sugar solution. | CopeLanp, H. P. Further studies on Monotropoideae. Madrono 6: 97-119. 1941. [Includes details of M. Hypopithys, M. uniflora, Pterospora, Hemi- tomes. | Davis, J. J. Occurrence of Indian pipe. Torreya 14: 162. 1914. [In sphagnum in a swamp of Picea mariana and Larix. D’Hupert, E. Recherches sur le sac embryonnaire des plantes grasses. Ann. Sci. Nat. Bot. VIII. 2: 37-128. pls. 1-3. 1896. GRIER, N. nee Preserving Indian pipes without discoloration. Rhodora 24: 225, 226 ie K, i . The nutrition of Indian pipe. Turtox News 35: 188-190. 1957. | Parasitic on the mycorrhizal fungus. Hutt, E. D. Occurrence of the Indian pipe (Monotropa uniflora) ina xerophytic habitat. Torreya 14: 101-105. 1914 . [Among an open stand of Quercus velutina on dunes, Lake Co KAMIENSKI, F. Die = Snr tre a der Monotropa Hypopitys, L. Bot. Zeit. 39: 457-461. 1881. . Les organes végétatifs te eh Hypopitys L. Mém. Soc. Natl. Sci. Nat. Math. Cherbourg 24: 0. pls. 1-3. 1882. Kocu, L. Das Entwicklung des Stil von Monotropa Hypopitys L. Jahrb. Wiss. Bot. 13: 202-252. pls. 9-11. 1882. it) se who wish to separate them must find another generic name, for Hy popitys Hill is illegitimate, having been proposed as a substitute for Monotropa. Hill wrote, “This author [Linnaeus] takes away its received name peers and calls its mono- iropa.” Hill’s Hypopitys includes both Linnaean specie 1961] WOOD, GENERA OF ERICACEAE 69 MEEHAN, T. Monotropa uniflora. Meehans Monthly 12: 133, 134. pl. 9. 1902. NIEUWLAND, J. A., and A. D. Stavin. Preservation of Monotropa and similar plants without discoloration. Proc. Indiana Acad. Sci. 38: 103, 104. 1928. {Saturated solution of sulfur dioxide in alcohol. | Oveva, C. Le Monotropa Hypopitys L. Anatomie et biologie. Bull. Soc. Bot. Fr. 57: 639. 1910 Suipata, K. Die Doppelbefruchtung bei Monotropa uniflora. Flora 90: 61-66. 1 pl. 1902 [1901]. SMALL, J. K. Monotropa Brittonii, Addisonia 13: 35, 36. pl. 434. 1928. Witcke, J. Dissemination of seeds of Indianpipe and Pyrola, (In Dutch.) Levende Nat. 57: 9-11. 1954.* Subfam. VACCINIOIDEAE Endlicher (‘Vaccineae’) Tribe VACCINIEAE D. Don 19. Vaccinium Linnaeus, Sp. Pl. 1: 349. 1753; Gen. Pl. ed. 5. 1754. Evergreen or deciduous, often stoloniferous shrubs [sometimes epi- phytic], rarely trees. Leaves alternate, short-petioled, entire or serrate; winter buds small, ovoid, with 2 or several outer scales. Flowers solitary in the axils of bracts or leaves, the more or less well-defined inflorescences thus racemose; inflorescences from buds of the preceding season, rarely reduced to 1 or 2 flowers; pedicels usually with 2 bracteoles. Calyx adnate to the ovary, the lobes 4 or 5, rarely obsolete. Corolla white or greenish to red, cylindric, urceolate or campanulate, 4- or 5-lobed or sometimes 4-parted to the base. Stamens 8 or 10; anthers with or without spurs on the back, lacking white disintegration tissue in the connective, the halves narrowed upward into tubes, opening by a terminal pore, without awns; pollen tetrads without viscin strands. Stigma small, simple or somewhat capitate; style straight; ovary [partly to] completely inferior, 4- or 5- or falsely 10-locular (then usually 10-locular above. 5-locular below) by the development of ingrowths of the ovary wall which unite with the axile, more or less enlarged and pendulous placentae: ovules usually numerous on the surface of the placenta; disc conspicuous, on the surface of the ovary. Fruit a [5—|many-seeded berry crowned by the persistent calyx lobes: seeds usually small. (Including Batodendron Nutt., Cyanococcus (Gray) Rdyb., Herpothamnus Small, Hugeria Small, Oxycoccus Hill, Polvcodium Raf., Rhodococcus (Rupr.) Avorin.) LECTOTYPE SPECIES: V. Myrtillus L.; see Britton & Brown, Ill. Fl. North. U. S. ed. 2. 2: 698. 1913. (Vaccinium, the classical name of V. Mvyrtillus.) —- DEERBERRIES, BLUEBERRIES, and CRANBERRIES. A polymorphic genus of perhaps 150 species, mostly of the Northern Hemisphere, from the mountains of the tropics northward to the Arctic Circle. The genus is composed of a number of more or less well-marked subgenera and sections, these often of restricted geographical distribution; but the classification is not yet well worked out and opinions as to the segregation of sections, subgenera, or genera vary. Vaccinium is treated ~ conservatively here (cf. Rhododendron), for many reticulate relationships 70 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII are evident between the various subgenera. (Note also natural and artificial hybrids between sections and subgenera.) Seven subgenera occur in the United States, subg. VaccinruM (subg. Euvaccinium Klotzsch), including sects. Vaccinium (§ Mvyrtillus Koch) and Vittz-Ipara (Moench) Koch, entirely to the north and west of our range. Vaccinium is taxonomically difficult and much herbarium material is inadequate for study. Collectors should note accurately height of plant, glaucescence of various parts, corolla shape and color, fruit color, habitat, and presence of other species or individuals of varying size in the same or adjacent habitats. Subgenus BaToDENDRON (Nutt.) Klotzsch (Batodendron Nutt.), in which the flowers are in bracteate racemes, the corolla campanulate and 5-lobed, the stamens included, the anthers spurred on the back, the style exserted from the corolla, the pedicels jointed beneath the flower. the ovary incompletely 10-locular, and the berry black, leathery, and _per- sistent, includes Vaccinium cubense Griseb. (and perhaps other species), of Cuba and Hispaniola, V. arboreum Marsh., of the southern United States, and V. leucanthum Cham. & Schlecht. and V. stenophyllum Steud., of Central America. Vaccinium arboreum (including var. glaucescens (Greene) Sarg.), farkleberry or sparkleberry, is an easily recognized shrub or small tree (to ca. 9 m.) with a wide range from Florida westward to Texas (Edwards Plateau) and northward to southeastern Virginia, Illi- nois, southern Missouri, and Oklahoma. The common southeastern form has lustrous, persistent leaves; in the western part of the range a elauces- cent phase appears and leaf proportions, size of bracts, shape of flower, pubescence, and glandularity change to varying degrees. A careful study of variation is needed. Small observed that this species is tolerant of calcareous soil. Subgenus Potycopium (Raf.) Sleumer (Polycodium Raf., Picrococcus Nutt.), the deerberries, in which the corolla is campanulate and open in bud, the stamens exserted and spurred on the back, the pedicels not jointed, the flowers in leafy or bracted racemes, the ovary incompletely 10-locular, the fruit greenish to nearly black and insipid, is a well-marked group ranging from Florida to Maine, westward to central Canada and Texas, and in Mexico along the Sierra Madre Oriental from Puebla to Nuevo Leon. About 27 “species” have been described, but the number is un- doubtedly far lower. The taxonomy of the group is chaotic and no practical treatment exists. All of the characters used are variable, including leaf size (note dimorphism between flowering and vegetative shoots), size of plant, pubescence, glaucescence, fruit color, and presence or absence of stipitate glands. The corolla is open in the young bud and continues to enlarge (as in Halesia), making size-comparisons difficult. Transplant and crossing experiments, population and cytological studies are an absolute necessity before the biology of this group can possibly be understood. The few chromosome counts (2n = 24) suggest that the polyploidy so characteristic of subg. CyaNococcus may be lacking here. Subgenus HERPOTHAMNUS (Small) Uphof (Herpothamnus Small), in- 1961] WOOD, GENERA OF ERICACEAE 71 cludes a single isolated species in our area, V. crassifolium Andr. (H. crassifolium (Andr.) Small), 21 = 24, a creeping or tailing, narrow- and coriaceous-leaved evergreen with condensed 1—5-flowered axillary racemes, the flowers in the axils of small bracts, the pedicels with 2 bracteoles of almost the same size as the bracts, the ovary 5-locular, the stamens 10 with hairy filaments and unspurred anthers with short tubules, the berries black, sweet. The species is localized in moist, sandy habitats on the Coastal Plain from South Carolina to southeastern Virginia. Camp has called attention to the curious nodular enlargement of the roots and has suggested a relationship to a few montane South American pendent epi- phytic species in which a similar development occurs, observing further that a species epiphytic at lower altitudes may become terrestrial at higher, with an even greater development of the root enlargements. Subgenus Cyanococcus (Gray) Klotzsch (Cyanococcus (Gray) Rydb.), the blueberries, with flowers in racemes or clusters from axillary buds, corollas urceolate to cylindric or ovoid-campanulate with 5 short lobes, anthers spurless, filaments pubescent, ovary incompletely 10-locular, and berries blue or black and often glaucous, includes about 24 species of eastern North America (about 16 in the Southeast), perhaps 10 of Japan and Korea, and probably V. Selerianum (Loesen.) Sleum., of Chiapas, Mexico. Extensive field studies and careful morphological comparisons, along with a series of cytological and hybridization studies, have revealed many of the underlying causes of the taxonomic difficulties encountered in this subgenus in eastern North America. (See especially Camp [1945 | and Darrow & Camp [1945] for biology, discussion, keys to North Ameri- can species and hybrids, chromosome numbers. ) In the United States the subgenus comprises about nine diploids (2 = 24), twelve tetraploids (including V. corymbosum L., a hybrid complex of varying combinations of about six tetraploids to the north of the glacial boundary), and three hexaploids (including V. Ashei Reade, a complex alloploid involving about five tetraploids). Interspecific hybridization (often extensive) occurs at the diploid, tetraploid, or hexaploid levels, the principal barriers being geographical and ecological isolation with no interspecific sterility between homoploids. Segregation within alloploid complexes adds further complications, for some of the derivatives may simulate the parental species. Autoploidy is postulated for six species which appear to be more robust tetraploid counterparts of diploids; three other tetraploids (V. myrsinites Lam., V. marianum Wats., V. fuscatum Ait.) and the hexaploid V. Constablaei Gray appear to be of alloploid origin. Pentaploid hybrids, but no triploids, have been found in the wild. Individual plants may be self-sterile, but others of the same species may be partly or largely self-fertile. Vaccinium hirsutum Buckl. (2n = 48), of the mountains of eastern Tennessee and western North Carolina, is aberrant in the subgenus in the glandular pubescence of flowers and fruit. Camp has raised the in- triguing possibility (based on some early hybrids made by Coville) that this species may have arisen as an alloploid between subg. CyANococcus 72 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII and subg. Potycopium and has suggested that a pubescent- leaved, black- fruited V. vacillans Torr. and a pubescent-leaved, nonglaucous, gland- bearing form of subg. Potycoprum be used in attempting to resynthesize V. hirsutum. Blueberries are a commercial fruit crop of ever-increasing importance, and a number of cultivars including both selections from the wild and a series of artificial hybrids (begun by Coville) have been named. The low, diploid V. angustifolium Ait. (2n = 24) is the most important commercial species in the Northeast, while commercial plantations of highbush types on the Coastal Plain, from eastern Massachusetts southward to Florida and ag are largely V. australe Small, southeastern highbush blue- berry (2m = 48), or (especially to the soul V. Ashei, rabbit-eye blue- berry. Vaccinium corymbosum is cultivated to a slight degree. Subgenus Oxycoccoiwres (Benth. & Hook.) Sleumer (Hugeria Small), shrubs in which the leaves are deciduous, the flowers solitary in the axils of leaves, nodding on slender pedicels, the corolla deeply 4-cleft and the lobes revolute, the anthers with very long terminal tubules and usually spurless, the filaments short, hairy, the ovary 4-locular, and the fruit red to black, insipid to acid and pleasant tasting, comprises only Vaccinium japonicum Miq. (including vars. sinicum (Nakai) Rehd., of China, lasio- stemon Hayata, of Formosa, and japonicum, of Japan and Quelpaert Island) and V. erythrocarpum Michx., of the southern Appalachians. Vac- cinium erythrocarpum, mountain cranberry, a shrub to 2 m., is a plant of cool montane habitats from eastern Tennessee, and western North Carolina. to western Virginia and eastern West Virginia. The fruit varies widely in taste from insipid to pleasantly acid, and in color from red to black (f. nigrum Allard). Although this subgenus and subg. Oxycoccus are simi- lar in many floral details, the two are undoubtedly of quite independent derivation and are not closely related. The anthers are usually spurless, but Hooker figured a plant with short spurs, and in an occasional indi- vidual some (but not necessarily all) anthers of a flower may bear (one or) a pair of very short divergent spurs Subgenus Oxycoccus Drude (Oxycoccus Hill), the cranberries, is a well-marked group of about three trailing, small-leaved, evergreen bog species of the Northern Hemisphere with 4-merous flowers, the corolla parted nearly to the base, the lobes linear-oblong and revolute, the ovary 4-locular, the anthers with very short filaments and very long tubules: spurless on the back, the flowers articulated to slender 2-bracteolate pedi- cels, axillary or in short racemes, the berry bright red and acid. Vaccinium macrocarpon Ait. (O. macrocarpus (Ait.) Pers.), cranberry, a diploid (2n = 24), ranges from Minnesota to Newfoundland and southward on the Coastal Plain to Delaware, southeastern Virginia, and eastern North Carolina, and in the mountains to West Virginia, western Virginia, north- eastern Tennessee, and western North Carolina.'’? A second diploid, V. "There is also in the Gray Herbarium a small aarti — of Vaccinium macrocarpon, labeled in Asa Gray’s hand, ‘Little Rock, Dr. Hasse, 1886,’ which is presumably the basis for reports of this species a emails 1961 | WOOD, GENERA OF ERICACEAE ve microcarpum (Turcz.) Schmalhaus. ex Busch (O. microcar pus Turcz.) eland. The tetraploid and relatively variable V. Oxycoccos L. (O. quadripetalus Gilib.) combines the characters of the diploids and has a virtually cir- cumpolar sar cance overlapping in part that of both diploids. In Am eastern North it extends ieee in cold bogs in the moun- tains to paetione: nes West Virginia.'' Isolated, sterile, hexaploid plants (2n = 72), distinguishable primarily i ae size, have been re- ported from Europe. Hagerup thought these to have resulted from the hybridization of V. Oxyvcoccos and V. microcarpum, while Camp has sug- gested instead a derivation directly from the tetraploid and has further postulated both the sporadic occurrence of similar hexaploids throughout the range of V. Oxycoccos and the possible occurrence of fertile hexaploids. Vaccinium macrocarpon is an important commercial plant, the crop of cranberries from Massachusetts, New Jersey, and Wisconsin (and to a lesser extent Oregon, Washington, Maine, and Rhode Island) amounting to 10-20 million dollars annually. REFERENCES: Under family references see also BELL & BURCHILL, DUFRENOY, KEARNE MattHews & Knox, PELTRISOT, SMITH, and STANLEY; under ce em see Cain & PotzcER; for sie information on Vacciniuan. macrocarpon see Cra berries. the national cranberry magazine (Wareham, Massachusetts, 1936+ ) '' Fernald (Gray. Man. Bot. ed. 8. 1950) records Vaccinium Oxycoccos var. ovali- folium Michx. from North Carolina. Apparently this record is based on a sheet in the Gray Herbarium which bears two si eri both labeled in Gray’s handwriting. To the left is a fruiting specimen of V. Oxycoccos with ‘“‘N. Carolina, Croom” be- neath it; to the right are two owed shoots of V. macrocarpon and a character- istic blue label “Herb. Asa Gray” (on which the smaller label “Syn. Fl. N. Amer.” has been pasted) with “Vaccinium macrocarpon. W. New York,” also in Gray’s hand. Another sheet bearing five collections of V. Oxycoccos has two small fruiting pees mens which are associated with a similar label with “Vaccinium Oxycoccus. W. New York.” Croom’s botanizing seems to have been confined mostly to the Coastal Plain, especially around New Bern, North Carolina, and Tallahassee and Aspalaga, Florida, with observations obtained in the course of his travels between these areas. (See Jour. N. Y. Bot. Gard. 21: 164. 1920.) His published botanical notes do not list montane plants (with the exception of Pinus pungens which he recorded from Pilot Mountain, Stokes County, North Carolina), and I have been unable to find any evidence that he collected in the high mountains of western North Carolina where . Oxveoccos might possibly im expected. Furthermore, he listed Oxycoccus macrocar- pus in his Catalogue of Plants Native or Naturalized in the Vicinity of New Bern, North Carolina (1837). In view of these circumstances and of the two labels bearing “W. New York,” it seems likely that an error in mounting or labeling is involved and that the specimen of V. macrocarpon should be attributed to Croom, w hile that of V. Oxycoccos should be assigned to New York. It is also significant that Gray himself (Syn. Fl. N. Am. 2(1): 26. 1886) did not admit V. Oxycoccos south of the mountains of Pennsylvania, whereas V. macrocarpon was given a range carne Newfoundland and North Carolina. Until more convincing evidence is presented, seems os to regard Pocahontas County, West Virginia, as the approximate nae limit of V. Oxycoccos and to exclude this species from the flora of the southeastern United a 74 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Appoms, R. M., and F. C. Mounce. Notes on the nutrient requirements and the histology of re cranberry (Vaccinium macrocarpon Ait.) with special ref- erence to mycorrhiza. Pl. Physiol. 6: 653-668. pls. 11, 12. 1931. | See also ibid. 7: 643-656. 1932.] ALLARD, H. A. Vaccinium erythrocarpum — what is the fruit color? Castanea 12: 117, 118. 1947. [See also ibid. 13: 127. 194 8.] Ammons, N., and E. L. Core. Huckleberries, blueberries and cranberries of West Virginia. Castanea 10: 103-109. 2 AsHe, W. W. Polycodium. Jour. Elisha Mitchell Sci. Soc. 46: 196— 213. 1931. [One of Ashe’s numerous contributions to confusion: useful only in show- ing the range of variability in the subgenus. | Avorin, N. A. On the genus Rhodococcum (Rupr.) gen. nov. (Vacciniaceae). (In Russian.) Bot. Zhur. 43: 1719-1724. 1958. |V. Vitis-Idaea segregated; Rhodocinium intermedium (Ruthe) Alava for V, Myrtillus > Vitis- Idaea; tribe sintieed oun published by Small, Man. Southeast. Fl. 1006. 1933) maintain Barn, H. F. Production of ene mycorrhiza in the cultivated cranberry. Tour, Agr. Res. 55: 811-835. pls. 1-10. 1937. [Includes critical review of the literature on mycorrhiza in the Ericaceae: bibliography. | ——— and H. Dermen. Sectorial poly ploidy and phyllotaxy in the cranberry (Vaccinium macrocarpon Ait.). Am. Jour. Bot. 31: 581-587. : Bett, H. P. Determinate growth in the blueberry. Canad. Jour. Res. C. 28: 637-644, 1950. [V. angustifolium; the same species in other studies by this author . The growth cycle of the blue-berry and some factors of the environ- ment. Canad. Jour. Bot. 31: 106. 1953. ———. The development of the blueberry seed. /bid. 35: 139-153. 1957. and J. BurcHILL. Flower development in the lowbush blueberry. /bid. 33: 251-258. 1955. and E. C. Girrin. The oe blueberry: the vascular anatomy of the ovary. /bid. 35: 667-673. Camp, W.H. Hugeria a. Addisonia 21: 3, 4. pl. 674. 1939 . On the structure of populations in the genus Vaccinium. Brittonia 4: 189-204. 1942. . A survey of the American species of Vaccinium, subgenus Euvaccinium. Ibid. 205-247. A preliminary consideration of the biosystematy of Oxycoccus. Bull. Torrey Bot. Club 71: 426-437. 1944. he North American blucbersies with notes on other groups of Vacci- niaceae. Brittonia 5: 203-275. 1945. and C. L. Gitty. Polypetalous forms of Vaccinium. Torreyva 42: 168- 173. 1942 [1943]. Ciark, J. H., and S. G. Gripert. Selection of criterion leaves for the — tion of blueberry varieties. Proc. Am. Soc Hort. Sci. 40: 347-351. 1942 Covitte, F. V. Experiments in blueberry culture. U. S. Dep. Agr. Pl. ee Bull. 193: 1-100. pls. 1-18. 1910. Directions for blueberry culture, 1921. U. S. tig ig Bull. 974: 1-24. pls. 1-29. 1929. [See also 334: 1-16. 1915; U. S. Dep. Agr. PI. Industry Circ. 122: 3-11. 1913.] Blueberry chromosomes. Science 66: 565, 566. 1927. [Reports on hybridization experiments. | 1961] WOOD, GENERA OF ERICACEAE 75 Darrow, G. M. Rest period requirement for blueberries. Proc. Am. Soc. Hort. Sci. 41: 189-194. 1942.* —-. au varieties of blueberry. Yearb. U. S. Dep. Agr. 1943/47: 300-303. 1947 : “The big six ass varieties for northern states. Natl. Hort. Mag. 35: 162-165. 1956 Blueberry growing. U. S. Dep. Agr. Farmers Bull. 1951: 1-33. 1957.* ’, H. Came. Vaccinium hybrids and the development of new horti- cultural material. Bull. Torrey Bot. Club 72: 1-21. 1945 —. H. E. FiscHer, and H. DERMEN. Chromoccae numbers in Vac- cinium and related groups. /bid. 71: 498-506. 1944 , E. B. Morrow, and D. H. Scott. An evaluation of aDEPE SECU blue- berry crosses. Proc. Am. Soc. Hort. Sci. 59: 277-282. 1952. DERMEN, H. The mechanism of colchicine- induced cytohistological changes in cranberry. Am. Jour. Bot. 32: 387-394. 1945 . Periclinal cytochimeras and hese ee in cranberry. /bid. 34: 32-43. 1947. and H. F. Barn. Periclinal and total eh ae in cranberries by colchi- cine. Proc. Am. Soc. Hort. Sci. 38: 400. A general cytohistological study of pela polyploidy in cranberry. Am. Jour. Bot. 31: 451-463. 4, Eaton, E. L. The spread of blueberry seed through ear and by migrating robins. Proc. Am. Soc. Hort. Sci. 69: 293- 295. 1957. Eccert, F. P. Shoot emergence and flowering habit in the lowbush blueberry (Vaccinium angustifolium). Ibid. 288-292. FERNALD, M. L. The variations and distribution of American cranberries. Rhodora 4: 231-237. 1902. Fuint, E. M. Structure of wood in blueberry and huckleberry. Bot. Gaz. 65: 556-559. pls. 10, 11. 1918. [V. corymbosum, V. angustifolium, Gaylussacia sp., Rhododendron sp. | GLEISBERG, W. Systematisch-kritische Vorarbeit fiir eine Monographie der Spezies Vaccinium Oxycoccus L. Bot. Arch. 2: 1-34. 1922 . Vaccinium Oxycoccus L., ein weiterer Beitrag zur Typenfrage der Art. Ibid. 130-139. Vergleichend-anatomische Untersuchung des Blattes der Vaccinium Oxycoccus-Typen. Ber. Deutsch. Bot. Ges. 40: 139-147. 1922. GOHEEN, A. C. The cultivated highbush blueberry. Yearb. U. S. Dep. Agr. 1953: 784-789. 1953. [Diseases; see also BERGMAN, H. F. Disorders of cranberries. /bid. 789-796. | Hacerup, O. Studies on polyploid ecotypes in Vaccinium uliginosum L. Here- ditas 18: 122-128. 1933 . Studies on the ceninee of polyploidy. IV. Oxycoccus. Ibid. 26: 399-410. 1940. Hatt, I. V. The apes of the lowbush blueberry. Nova Scotia Fruit Growers Assoc. ee bs io tap r a in lowbach blueberry. Canad. ce Bot. 35: 934. pl. 1. L057: a myrtilloides, V. angustifolium, V. Britto . Plant populations in blueberry stands eee ed ae abandoned hay- fields and woodlots. Ecology 40: 742, 743. 1959. [New Brunswick. | Hooker, J. D. Vaccinium erythrocarpum. Bot. Mag. 51: pl. 7413. 1895. 76 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Hooker, W. J. Vaccinium corymbosum, V. angustifolium. Ibid. 62: pls. 3428, 3434, 1835. Karo, E. A. Pollination of the lowbush blueberry. Nova Scotia Fruit Growers Assoc. Rep. 94: 93-97. 1957.* Kramer, A., and A. L. ScHRADER. Significance of the pH of blueberry leaves. Pl. Physiol. 20: 30-36. 1945. Lonctey, A. E. Chromosomes in Vaccinium. Science 66: 566-568. 1927. LoveLL, H. B. Pollination of the Ericaceae: VI. Vaccinium al on Mt. Katahdin. Rhodora 44: 187-189. 1942. [By 3 spp. of Bon ; LoveELL. Pollination of Vaccinium Spite sada. Ibid. 39: 60-63. pl. 456. 1937. [V. angustifolium Ait. | Mautstepe, J. P., and D. P. Watson. An anatomical study of adventitious root development in stems of Vaccinium corymbosum. Bot. Gaz. 113: 279-285. Sz. Newcomer, E. H. Chromosome numbers of some species and varieties of Vaccinium and related genera. Proc. Am. Soc. Hort. Sci. 38: 408-470. 1941. Porsitp, A. E. The cranberry in Canada. Canad. Field-Nat. 52: 116, 117. 1938 RITCHIE, J. C. A natural hybrid in Vaccinium. I. The structure, performance and chorology of the cross Vaccinium intermedium Ruthe. New Phytol. 54: 49-67. 1955. [V. Myrtillus * V. Vitis-Idaea.| II. Genetic studies in Vaccinium intermedium Ruthe. Jbid. 320-335. pls. 6, 7. [Fls. homog- amous in both parental spp. | . Vaccinium Mortillus L. Jour. Ecol. 44: 291-299. 1956. Roperts, R. H., and B. E. SrrucKMEYER. Growth and fruiting of the cranberry. Proc. Am. Soc Hort. Sci. 40: 373-379. 1942.* —. Blossom induction of the cranberry. Pl. Physiol. 18: 534-536. 1943. Rosrinson, C. B. Polycodium. Bull. Torrey Bot. Club 39: 549-559. 1912. [Dis- cussion of generic limits; an attempt at delimitation of species. | Rozanova, M. A. A survey of the literature on the genera Vaccinium and Oxy- coccus. (In Russian; English summary.) Bull. Appl. Bot. Genet. Pl. Breed. VIII. 2: 121-186. 34. SARGENT, C. S. Vaccinium. Silva N. Am. 5: 115-120. pl. 230. 1893. [V. arbo- reum. | . Vaccinium hirsutum, Gard. Forest 2: 364, 365. 1889 SAWYER, W. H., Jr. Stomatal be caeraaes of the cultivated cranberry, Vaccinium macrocarpon. Am. Jour . 19: 508-513. pl. 41. 1932. SEYMOUR, F. C., ne as a genus. Am. Midl. Nat. 49: 935, 936. 1953. Stms, J. Vaccinium crassifolium. Bot. Mag. 29: pl. 1152. 1809. [Also V. myr- sinites, 37: pl. 1550. 1813; V. arboreum, 38: pl. 1607. 1814; V. macro- carpon, 52: pl. 2586. 1825.]| SLEUMER, H. Die Arten der Gattung Vaccinium L. in Zentral- und Siidamerika. Notizbl. Bot. Gart. Berlin 13: 111-140. 1936. Vaccinioideen-Studien. Bot. Jahrb. 71: 375-510. 1941. [Vaccinium, 408-493, including key to 33 sects., revision of spp. of India and East Asia. | SMALL, J. K. Polycodium floridanum. Addisonia 17: 35, 36. pl. 562. 1932. SmitH, B. E. Notes on Polycodium of North and South Carolina. (Abs.) Am. Jour. Bot. 36: 803, 804. 1949. |See also Jour. Tenn. Acad. Sci. 24: 168. 1949. ] STEVENS, N. E. The development of the oo in Vaccinium corymbosum. Bull. Torrey Bot. Club 46: 465-468. 1961] WOOD, GENERA OF ERICACEAE a | VEILLET-BARTOSZEWSKA, M. Ericacées; développement de l’embryon chez le Vaccinium Myrtillus L. Compt. Rend. Acad. Sci. Paris 246: 824-826. 1958. WasscHER, J. Vergelijkend-anatomische onderzoekingen bij eenige Vaccinium- soorten. (In Dutch; English summary.) Ned. Dendr. Ver. Jaarb. 15: 65- 112. 1947. [Comparative-anatomical studies of 11 spp. of Vaccinium. | Wetcu, W. H. A contribution to the phytoecology of southern Indiana with special reference to certain Ericaceae in a limestone area of the Blooming- ton Quadrangle. Proc. Indiana Acad. Sci. 38: 65-83. 1929. [Includes data on Gaylussacia baccata, Vaccinium vacillans, Epigaea repens. | 20. Gaylussacia Humboldt, Bonpland & Kunth, Nova Gen. Sp. Pl. 3: ed. fol. 215, ed. quart. 275. pl. 257. July 1819, nom. cons.'* Evergreen or deciduous, branching, often stoloniferous shrubs with the aspect of Vaccinium. Leaves alternate, short-petioled, entire or toothed, membranaceous to coriaceous, often with glandular hairs; winter buds ovoid with about 3 outer scales. Flowers in axillary, usually few-flowered and sometimes much condensed racemes from buds of the preceding grow- ing season, bracteate, the bracts persistent to deciduous; pedicels with (1 or) 2 bractlets. Calyx adnate to the ovary, 5-lobed, persistent. Corolla white, greenish to red, ovoid or campanulate, 5-lobed. Stamens 10, in- serted on the base of the corolla; filaments short; anthers unappendaged on the back, each half narrowed upward into more or less of a tube and opening by a terminal pore or short slit, awnless, without disintegration tissue; pollen tetrads without viscin strands. Stigma more or less capitate and 5-angled; style straight; ovary inferior, 10-locular, each locule with a single axile, campylotropous ovule; disc conspicuous, on the surface of the ovary. Fruit a berry-like drupe with 10 smooth, 1-seeded lenticular nutlets (pyrenes), crowned by the persistent calyx lobes, black or blue, sometimes glaucous. (Including Buxella Small, Decachaena Torr. & Gray, Lasiococcus Small.) Type species: G. buxifolia HBK. (Named in honor of Joseph Louis Gay-Lussac, 1778-1850, eminent French chemist and physicist.) — HUCKLEBERRIES A genus of perhaps 50 species of eastern North America and of South America, with the center of diversity in southeastern Brazil; lacking in Central America and the West Indies. Three well-marked sections in- cluding six to nine species occur in eastern North America; the applica- tion of subgeneric categories to the South American species is not yet well worked out. Section Gaytussacta (§ Eulussacia Benth. & Hook.), includes two species of our area with clavate- -stipitate gland hairs in which the glandular portion is usually without a conspicuous secretion pellicle. Gaylussacia Mosieri Small (Lasiococcus Mosieri (Small) Small), with conspicuously glandular-hirsute hypanthia, is a plant mostly of sandy bogs from western 2 Conservation superfluous. Adnaria [odorata] Raf., 1817, against which Gay- lussacia was conserved, has been ae he Camp to be 'Styrax [americanum Lam.]. See Castanea 6: 82. 1941 and Taxon 9: 78 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Florida and southwestern Georgia to eastern Louisiana. The widespread G. dumosa (Andr.) Gray (Lasiococcus Small), with inconspicuous short- stalked glands on the hypanthium, ranges from southern Florida north- ward mostly on the Coastal Plain to Newfoundland, and westward to eastern Louisiana, with scattered stations inland across all but north- western Alabama, in Coffee County, Tennessee, in the mountains of north- ern Georgia, North Carolina and Virginia and West Virginia, and in Penn- sylvania. Southward a plant of drier soils, northward the species usually occurs in bogs. Considerable variation occurs in the length of the hairs on the hypanthium, those of some collections appearing to be almost in- termediate in length with G. Mosieri; G. orocola (Small) Camp, described from a single collection from Flat Rock, Henderson County, North Caro- lina, may be one of these. The variation and interrelationships of these three taxa need further study. Section DEcAMERIUM (Nutt.) Benth. & Hook. (§ Decachaena (Torr. & Gray) Sleumer), including in our area plants with the leaves bearing capitate subsessile glands with a conspicuous golden secretion layer, and usually with the inflorescence bracts deciduous, is represented by G. bac- cata (Wangenh.) K. Koch (2n = 24), wide-ranging from Newfoundland to Manitoba, southward inland to Iowa, Kentucky, and northern Georgia, in which the leaves are golden-glandular on both leaf surfaces, and by three or four species in which the glands are confined to the lower leaf surface. Gaylussacia ursina (M. A. Curtis) Torr. & Gray ex Gray, with pubescent stamen-filaments, is a clear-cut forest species of the mountains of eastern Tennessee, western North Carolina and South Carolina, and northern Georgia. The relationship of G. frondosa (L.) Torr. & Gray ex Torr., a shrub to 2 m. high, confined mostly to the Coastal Plain from Florida to Louisiana and northward to southeastern New York, Massa- chusetts, and southern New Hampshire, to two similar stoloniferous shrubs, the very glaucous G. nana (Gray) Small (1-4 dm. high) and the very hairy and nonglaucous G. tomentosa (Gray) Small (to 1 m. high), both ranging from southern Georgia and Florida to southern Alabama, is un- certain. Camp tentatively maintained the two latter at the rank of species, suggesting that chromosomal counts might clarify the interrelationships (as in Vaccinium § Cyanococcus). Section Vitis-IpArA Benth. & Hook. includes only Gaylussacia brachy- cera (Michx.) Gray (Buxella Small), the box-huckleberry or juniper- berry, a low, evergreen species completely lacking glandular hairs, which occurs in scattered localities on the coastal plain of Delaware and Mary- land and on mostly north-facing slopes in Pennsylvania, southeastern West Virginia (the area of greatest abundance), eastern Kentucky, and eastern Tennessee. Most colonies seem to consist of single individuals (one in Pennsylvania some eight acres in extent estimated to be 1200 years old) which are largely self-sterile. (Gaylussacia baccata has also been observed to be self-sterile. The flowers in this species are homog- amous and are cross-pollinated by Apis and various andrenid bees. ) Although Small considered the North American species to represent 1961 | WOOD, GENERA OF ERICACEAE 79 three genera distinct from South American Gaylussacia, Camp has shown numerous interlocking relationships (e.g., the remarkable similarity of G. Mosieri and the Brazilian G. pseudogaultheria Cham. & Schlecht.). Camp concludes, in part, “Considered alone the North American material consists of three well-marked groups, but when critically examined each of these gives clear indication that it has been rather directly derived from South American material, having counterparts in the common plexus of the genus. . . . It is certainly clear that any separation into subgeneric eat ROrS on a “getovaphic basis violates the fundamental phylogeny of the group.” On the basis of the 10-locular, 10-ovulate ovary and the fruit with 10 smooth, 1-seeded nutlets, Gaylussacia is sometimes segregated as a separate tribe, Gaylussacieae Small. REFERENCES: Under family references see also BELL & BuURCHILL, BERGMAN, and STANLEY; under Vaccinium see AMMONS & Core and WELCH Carn, S. A., and J. E. Poneces. A comparison of leaf tissues of Gaylussacia bac- cata (Wang.) C. Koch. and Mah vist vacillans Kalm. grown under ferent pacaus Am. Midl. Nat. 14: 97-112 mparison of leaf ace of Gaplccnen baccata grown under dif- Ferent ae Ibid. 24: 444-462. 1940. Camp, W.H. Studies in the Ericales. I. The genus be pare in North America north of Mexico. Bull. Torrey Bot. Club 62: 129-1 1935. . Studies in the Ericales: A review of the North American Gaylussacieae ; nae remarks on the origin and migration of the group. /bid. 68: 531-551. 194 ae in the Ericales: the search for Adnaria odorata Raf., and Ar- butus obtusifolius Raf. Castanea 6: 80 1941. [Adnaria = Styrax americanum Lam., not Gaylussacia; pees = Vaccinium arboreum Marsh. | CoviL_Le, F. V. The threatened ae of the box huckleberry, Gaylussacia brachycera. Science 50: 30— 1919. DE WILDEMAN, E. Pies resinosa Torrey et Gray. Ic. Sel. Horti Thenen- sis 4: 139-141. 1903. [G. baccata. | FERNALD. M. L. The northern variety of Gaylussacia dumosa. Rhodora 13: 95-99. 1911. [ Var. Bigeloviana Fern. | Gray, A. Chloris boreali-americana. coe Am. Acad. Arts Sci. II. 3: 1-56. ee 1-10. 1848. [G. ursina, 49-55, «| Gray. F. W. Scores of stations for ieee brachycera in West Virginia. oe 22: 17; 18) 1922. LoveLLt, H. B. Pollination of the Ericaceae: V. Gaylussacia baccata. Rhodora 42: 352-354. 1940 MACKENZIE, K. K. The scientific name of our common huckleberry. Torreya 7: 60. 1907. [G. baccata (Wangenh.) K. Koch. Rosinson, B. L. A blue-fruited huckleberry. Rhodora 2: 81-83. 1900. [G. resi- nosa f. glaucocarpa = G. baccata f. glaucocarpa (Robins.) Mackenz. | Sts, J. Se ale bustin Bot. Mag. 23: pl. 928. 1806; Vaccinium dumo- sum, 27: pl. 1808; Vaccinium resinosum (B8.). 32: pl. 1288. 1810. [G. burtiolta. . mine G. baccata. | 80 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Steumer, H. Vaccinioideen-Studien. Bot. Jahrb. 71: 375-510. 1941. [Gaylus- sacia, 382-385. ] SmaLL, J. K. A new gopher-berry from the Gulf States. Torreya 27: 35, 36. 1927. [G. Mosieri.] Wuerry, E. T. The box huckleberry as an illustration of the need for field work. Bull. Torrey Bot. Club 61: 81-84. 1934. [Summarizes data on distribu- tion, etc. } 1961] KOBUSKI, STUDIES IN THEACEAE, XXX 81 STUDIES IN THE THEACEAE, XXX THE AFRICAN SPECIES OF TERNSTROEMIA CLARENCE E. KoBusKI THE GENUS Ternstroemia is the largest genus of the family Theaceae with a nearly equal distribution of taxa in the tropical and subtropical regions of both the Eastern and Western Hemispheres. In 1942 and 1943, three papers * dealing with the genus in the Western, Hemisphere were published. The author was very much handicapped in that work since the study was made during the 2arly war years and the European types of the genus were not available to him. The present study is a part of one encompassing the remaining members of the group, those of the Eastern Hemisphere. As in the treatment of the western species the results will be published in a series of papers based on regional distribu- tion. There are several reasons for this manner of presentation. First, the taxa are generally quite endemic within a small area of distribution. The characters available for clear-cut distinction between taxa are very few and variable. Hence, a regional treatment is very helpful to any future worker, since he need consider only the taxa from a proscribed area. By the same token it is helpful to the author especially in making a key. To make a workable key to the complete genus may be possible — but hardly probable. An intense study of the group shows that there are many in- tangibles which the author learns to recognize but cannot use in presenting a key. This is especially true in the Indonesian species, which have been studied and will be treated later. The genus Ternstroemia was first proposed by Linnaeus f. in 1781. Various other names have been applied to this genus, such as Taonabo Aublet (1775), Dupinia Scoparius (1777), Hoferia Scoparius (1777), Tonabea Jussieu (1789), Amphania Banks (1821), Reinwardtia Korthals (1840), Llanosia Blanco (1845), Erythrochiton Schlechter (1846), Voelckeria (1847) and Mokofua Kuntze (1891). Taonabo is the nomen prius of this group of synonyms and until a few years ago was used by several of the American botanists who followed the rule of strict priority laid down by the ‘“‘American Code” of nomenclature. However, the name Ternstroemia has been included in the International Code of Botanical Nomenclature as one of the nomina conservanda and the name Taonabo listed as a nomen rejiciendum, thus ending any controversy which might have existed over the correct name. The description below applies to the entire genus, the portions in- applicable to the African species being inclosed in brackets. * Jour. Arnold Arb. 23: 298-343, 464-478. 1942; 24: 60-70. 1943. 82 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Ternstroemia Mutis ex Linnaeus f. Suppl. Pl. 39. 1781. Flowers hermaphroditic, dioecious [or androdioecious]. Sepals 5, im- bricate, persistent. Petals 5, connate at the base |free or joined to the middle or above], stamens 15—co, 2—several-seriate, rarely 1-seriate; fila- ments connate, the outer filaments adnate to the base of the corolla; anthers oblong or linear, as long as or longer than the filaments, dehiscing longitudinally, the connective usually projected into an apiculate or caudate appendage, rarely muticous. Ovary 2[3, rarely 1]-loculate, the ovules in each locule 2/—20], rarely solitary, pendulous from the apex of the placenta on more or less evolute funiculi. Style 1, entire [or rarely 2- or 3-parted], stigma or stigmata punctiform or evolute, entire or lobate. Fruit irregularly dehiscent, often explosively so. Seeds usually 4 | rarely 2-16]; testa opaque, smooth, rarely plicate-rugulose, yellow, usually cov- ered with reddish, many-celled papilli; endosperm pulpy, + developed, seldom lacking; embryo bent nearly double, oily, with half-terete or flat cotyledons. Glabrous trees or shrubs with branches usually verticillate or sub- opposite. Leaves spirally disposed, congested or verticillate at the apex of the current year’s growth, usually coriaceous, rarely chartaceous, entire or subentire. Flowers axillary, solitary, bracteolate, the bracteoles 2 [rarely 4], opposite or subopposite, placed immediately below the calyx, rarely with one somewhat removed and alternate on the pedicel. TYPE SPECIES: Jernstroemia meridionalis Mutis ex Linn. f. KEY TO SECTIONS AND SPECIES Flowers hermaphroditic. Leaves thick-coriaceous. Stamens 15-20, arranged in a single series. Ovary semiglobose, 2-loculate, each locule 2-ovulate. ire 2- punctate. Angola TERNSTROEMIA. 7. africana. Flowers dioecious. Leaves chartaceous-coriaceous. Staminate flowers: Stamens many, ca. 60, arranged in several series. Pistillodium negligible or lacking. Pis- tillate flowers: Stamens much reduced in character, arranged in two series. Ovary 2-loculate, each locule 2-ovulate. Stigma bilobed, each lobe rounded and convex, ca. 2 mm. across. Tanganyika. § ErytHrocuiton. T. polypetala. Ternstroemia africana Melchior, Notizbl. Bot. Gart. Berlin 8: 659. 1924; Nat. Pflanzenfam. ed. 2. 21: 142. 1925.— Exell, Jour. Bot. 74: Suppl. 21. 1936. — Exell & Mendonga, Consp. Fl. Angol. 1: 131. 1937 Adinandrella congolensis Exell, Jour. Bot. 65: Suppl. 30, fig. 1927. ‘“‘congolense.” Tree, glabrous throughout; branchlets generally terete, longitudinally striate in drying, reddish or reddish gray, the younger branchlets angled- striate in drying. Leaves coriaceous, obovate or obovate-elliptic, (4—)8—10 cm. long, (2—)4—5 cm. wide, obtuse or rounded at the apex, rarely some- what emarginate, cuneate at the base, the margin lightly recurved, en- tire, occasionally slightly subserrate near the apex, the midrib canaliculate 1961] KOBUSKI, STUDIES IN THEACEAE, XXX 83 above, elevated below, the veins ca. 7 pairs, inconspicuous on both sur- faces; petiole 5-8 mm. long, canaliculate above, lightly and distinctly glandular-denticulate along the margin. Flowers hermaphroditic, solitary, axillary; pedicel 3—4.5 cm. long, somewhat thickened at the apex; brac- teoles 2, subopposite, immediately below the calyx, oblong, ca. 4 mm. long, 1.6-1.8 mm. wide, obtuse; calyx-lobes 5, coriaceous, 7—7.5 mm. long, ca. 6 mm. wide, suborbicular, rounded at the apex (sometimes emar- ginate and mucronulate), unguiculate at the base, chartaceous at the mar- gin; corolla-lobes 5, membranaceous, lanceolate-ovate, 8—8.5 mm. long, 3—3.5 mm. wide, lightly connate at the base; stamens 15—20, arranged in a single series, the filaments 1-2 mm. long, adnate to the base of the corolla, the anthers 3.54.5 mm. long, narrowly lanceolate, projected at the apex into a subulate connective 0.5-1 mm. long. Ovary semiglobose to ovoid, 4-5 mm. long, 2—4 mm. diameter, 2-loculate, each locule 2-ovu- late; style ca. 4 mm. long; stigma punctate. Fruit not known. Angola: Zaire. at sea level, near the river Congo, Sumba, Lower Congo, J. Gossweiler 8592 (BM, type of Adinandrella congolensis and lectotype of T. afri- cana), Dec. 2, 1921 Ternstroemia africana was described by Melchior in 1924. Three years later, Exell, working with a duplicate of the type and not knowing of Melchior’s earlier disposition of the Berlin specimen, proposed Adinan- drella as a new genus. Later, in 1936, Exell agreed with Melchior’s treat- ment of the specimens and designated his Adinandrella as a synonym of T. africana. I had an opportunity to study the type of Adinandrella con- golensis while at the British Museum in the summer of 1960 and am very thankful for Exell’s “‘first mistake,” as he termed it, because up until then I had not known of either the existence of Exell’s species or its relation- ship to Melchior’s T. africana. The original specimen with which Melchior worked at Berlin has been destroyed, hence the specimen in the British Museum should become the type for 7. africana, as well as Adinandrella congolensis. An attempt to obtain additional material this past year from the her- barium at Luanda, Angola, has been unsuccessful, leading one to believe that the species has been collected only once. As a result, no dissections of the flowers of this species have been available to me. However, since the two descriptions of Melchior and Exell are so alike, and, as I believe, accurate, I do not feel the lack too greatly. Exell refers to the stamens as arranged in triads but does not appear certain of this arrangement. Melchior makes no comment on this arrange- ment. They both may be right. Since the stamens are arranged in a single series and are adnate to the base of the corolla and since there are approximately fifteen stamens it just may be that there are three sta- mens to each corolla-lobe. However, Melchior, who has studied the group more extensively than Exell and is more accustomed to the multiple series of stamens usually found in the genus, would give little attention to such an arrangement. accepting it as a single series, as would I. 84 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII The main differences between the two African species are actually those between the two sections as found in the Asiatic species and the American species. This species from Angola, on the west coast of Africa, resembles the American species and belongs to the section TERNSTROEMIA, to which most of the American species belong. On the other hand, 7. polypetala, from eastern Africa, belongs to the section ERYTHROCHITON and resembles the majority of the Asiatic species. However, this suggestive geographical arrangement is of no true significance since there are members of both sections in both hemispheres. The environments of the two species are strikingly different. Although both are found at approximately 5° S. Lat., T africana grows at sea level along the river Congo, while 7. polypetala was collected at an altitude of about 5400 feet in the mist forest of Tanganyika. Ternstroemia polypetala Melchior, Notizbl. Bot. Gart. Berlin 11: 1095. 34. Tree 10-12 m. high; branchlets verticillate or subverticillate, terete, eray. Leaves verticillate, disposed at the apex of the branchlets, charta- paler beneath, rounded or obtuse at the apex, cuneate at the base, the margin involute, sparingly glandular-denticulate, the midrib depressed on the upper surface, elevated below, the veins ca. 7 pairs, conspicuous below because of red color of veins against lighter background of leaf-blade; petiole up to 7 mm. long, appearing winged because of fine tapering de- currence of the leaf-blade. Flowers dioecious, axillary, solitary; pedicel ancipitous, 2—3 cm. long, thickening somewhat toward the apex; bracteoles 2. alternate below the calyx, lanceolate, ca. 4 mm. long, 2 mm. wide, ridged on the median of the dorsal surfaces, a single gland at the apex; calyx- lobes 5, rounded, unequal, the outer two lobes thicker, ca. 3 mm. long and 3.5 mm. wide. the inner lobes 3.5—4.0 mm. long, 4.5 mm. wide, the margin subscarious. Staminate flowers: corolla-lobes basically 5 (occasionally 2-5 additional smaller modified petals), obovate, ca. 10 mm. long, 6.5-7 mm. wide, rounded at the apex, connate at the base; stamens ca. 60, in several series, very unequal in length (3, 4, 5 and 6 mm. long), the longer ones, in the outer series, with filaments 1.5 mm. long, connate, and adnate to the base of the corolla, the anthers ca. 3 mm. long, the connective project- ing into a mucro 1.5 mm. long, the inner, smaller stamens more abundant, proportionally smaller in filaments and anthers, the mucro about the same as the longer outer stamens; pistillodium lacking. Pistillate flowers (not seen) fide Melchior: Stamens much reduced, fewer in number, in two series; ovary conical ovoid, 2-loculate, each locule 2-ovulate, the apex contracted into a short, entire style; stigma 2-parted, each part convex and rounded, ca. 2 mm. across. Fruit (not seen) conical, ca. 1.5 cm. long, 0.8-1.0 cm. diameter, often crowned by the persistent style and stigma; seeds normally 4, two to each locule, occasionally less in number by abor- tion, ca. 9 mm. long and 5.5 mm. across, covered with a yellow, mealy aril. 1961] KOBUSKI, STUDIES IN THEACEAE, XXX 85 Tanganyika: Morogoro District, northwest side of the Uluguru Mts., near Lupanga, mist forest at 1800 m. alt., H. J. Schlieben 3152 (vy, lectotype), Dec. 22, 1932 (tree 10-12 m., flowers yellowish white). The collector of the type states in his notes that both flowers and fruit were collected and Melchior in his original description fortunately de- scribed both, for the original specimen has been destroyed. However, Yale University has a duplicate specimen of Schlieben 3152 which has been available to me for study. This duplicate, which now must become the lectotype, is woefully lacking in both flowering and fruiting material. Only a single, unattached staminate flower with a broken ae was found in a packet on the sheet. Since this flower was unattached, I ven- tured to use it in a dissection for my study and carefully re- eee the material for future reference. It is odd that Melchior in his original description designated the species as belonging to the section ADINANDRELLA, the same name used generically seven years previously for Exell’s Angolan species. To my mind the species belongs to the Asiatic section ERYTHROCHITON as I do not find enough variation to warrant the creation of a new section. I am certain that Melchior, like myself, had never associated Exell’s Adinandrella with Ternstroemia or he would have mentioned it in his treatment of the spe- cies. Melchior does not designate the section ADINANDRELLA as new, and, knowing his fastidious methods of presentation, I feel that it probably had been initiated in some previous publication. If so, I have never en- countered it in my studies. The name 7. polypetala was selected by Melchior because of the many ‘“Detals” found in the flower. The single dissection made by the present author showed five basic petals, obovate in shape, rounded at the apex and measuring about 10 mm. in length and 6.5—7 mm. in width. These are typical petals for the genus in arrangement and size, especially when compared with the arrangement and size of the five calyx-lobes. There were also six other “pseudo-petals,” much smaller, measuring only 6 mm. long and 3.5 mm. wide, arranged between the basic petals and the stamens and appearing at first as a second petal series — which they are not. In Ternstroemia, the five petals, like the calyx-lobes, always increase in size and change in texture from the smaller, thicker, outer ones to the larger, more membranaceous, inner one. In 7. polypetala, this is true of the five basic petals, but the inner ‘“‘pseudo-petals” are all very membranaceous, definitely involute, and in an arrangement inconsistent with that of the true petals. The ‘“pseudo-petals” are bunched in three groups: (1) a single ‘“yseudo-petal” is between two outer petals; (2) two are together, one in front of the other, but again between two outer petals; and (3) three are together, two in front of an outer petal and the third between that and the next outer petal. Throughout the genus the stamens are always connate at the filament base, and in the outer series the filaments are adnate to the base of the corolla. In flowers in which there are more than a single series, the sta- mens in the outer series are usually larger. The inner “pseudo-petals” 86 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII T. polypetala are a part of the outer series of stamens and, therefore, are likewise joined with the regular petals at the base. At the points where the inner “pseudo-petals” are found, the stamens are joined to them, rather than to the outer petals. Thus the arrangement, size, shape and stamen attachment all indicate that the inner “pseudo-petals” are modified stamens. Regrettably, only a single staminate flower was available, and the only known pistillate flowers have been destroyed. I cannot but feel, however, that the interpretation of the ‘‘pseudo-petals” as modified stamens is correct and that pistillate flowers will also prove to have only five basic petals since the stamens have already been recorded as reduced to staminodia in these flowers. However, this is an assumption which cannot be verified until more ma- terial is available. 1961] HOWARD, STUDIES IN THE GENUS COCCOLOBA, X 87 STUDIES IN THE GENUS COCCOLOBA, X. NEW SPECIES AND A SUMMARY OF DISTRIBUTION IN SOUTH AMERICA RicHARD A. Howarp THE GREATEST NUMBER OF SPECIES of the genus Coccoloba occurs in South America. In the preceding paper of this series critical notes were presented on the 159 species reported to occur in South America. These 159 species, after critical study, were reduced to 77 either worthy of recog- nition or so incompletely known that they could not be referred else- where with confidence. The unnecessarily large number of species previ- ously described can be attributed to several factors, including a tendency on the part of many botanists to identify or concern themselves only with local or regional floras, thereby omitting consideration of species from adjacent regions; the failure to understand the morphological variation found on single plants; and the failure to recognize the complete or partial unisexuality of the flowers in the genus. The species of Coccoloba are mainly recognized on the appearance of the leaves. The particular identify- ing characteristics often seem to defy description and none has been found which lends itself to tabulation or analysis, largely due to the paucity of collections and the lack of critical field studies. A full and ample col- lection of a single species for complete understanding of its morphology should include detailed notes on the habit of the plant and all its branches. Specimens should be collected to show the leaves and shoot systems of adventitious growth as well as normal growth. Functionally staminate and pistillate flowering branches are needed along with fruiting branches. Specimens should be taken from several trees within a population, where possible, to show the effects of the environment. Such ideal material is gathered only by the monographer and then usually only after bitter ex- perience with a difficult group. Recently an experienced botanical traveler returned from a motor trip across the state of Goyaz in Brazil. He asked, ‘What is the attractive Coccoloba with wand-like branches and brilliantly colored leaves, so common everywhere?” The answer could only be, “It might be a new species,” for no such habit data are available for any species of the genus in South America. The following new species are not based on the ideal material suggested above, but, as is usually the case, on the specimens on hand. After a study of many hundreds of herbarium sheets of the genus from South America it is clear that no appropriate names exist for these plants. As I have indicated in previous papers the number of reliable characteris- tics useful in identification and classification of the species of this genus are few. The flowers and their parts show little variation, beyond that ex- 88 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII pected, in anthesis. The perianth segments vary in size and shape with their position in the bud. The functional stamens are always shorter than the perianth segments but appear to mature at different times and there- fore increase in length at different rates, until all are approximately equal in length when the flower is fully open. The sterile stamens may equal the mature fertile stamens in length or may all be rudimentary, or one or two may be long and the remainder short, or the reverse situation may occur. The sterile pistil from the staminate flower may be as large as the fertile pistil in a flower with abortive stamens or may be very small and rudimentary. Determining the fertility of a pistil in an herbarium speci- men, however, is extremely difficult; such observations really should be made by the collector. The styles and stigmas in both functional and non- functional pistils Rated in size and shape, particularly so in flowers clearly staminate in func A description sane the words, “perianth segments ovate-oblong, 1.5-2 mm. long, 0.5—1 mm. broad, functional stamens with filaments 1.5— 2 mm. long, functional pistil 1.5 mm. long, styles 2 or 3, stigmas punctate, occasionally lobed, non-functional stamens and pistils rudimentary, 0.5—-1 mm. long,” would apply to nearly every species described. The following descriptions therefore are limited to the most reliable characteristics which can be used to define the specimens cited. Coccoloba colombiana, sp. nov. Arbor, 4-6 m., ramulis teretibus, adpresse tomentosis; ocreis subcoria- ceis, 2 cm. longis, tomentosis, deciduis; petiolis 1.5—-2 cm. longis, supra basim ocrearum gerentibus, tomentosis; laminis obovato-oblongis vel obo- vato-ellipticis, 10 < 6 vel 14 x 10 cm. longis latisque, supra glabris, subtus tomentosis vel strigosis in nerviis vel axillis, apice acutis, , basi cuneatis asymmetricis, nerviis primariis 8-12, arcuatis; inflorescentibus 13-17 cm. longis, rhachibus striatis, puberulentis et tomentosis, bracteis ovatis, puberulentis, ca. 0.5 mm. longis, ocreolis translucentibus, ca. 1 mm. longis, floribus masculinis 3 vel 4 per nodulum, floribus femineis 1, raro 2 per nodulum; fructu ignoto. Colombia. Without definite location, Vageler s.n. (B—-holotype). ATLANTICO: Barranquilla, Cabica, Bro. Elias 1082 (w), 1469 (w). The specimen selected as the holotype was collected by H. Vageler on November 29, 1923, in Colombia. The specific location in German script is “Robles,” a location which cannot be found in current atlases. The specimen had been studied by Burret, who applied two unpublished names to the specimen, one of them honoring Vageler. The collection Elias 1082 in the Vienna herbarium was identified with the same unpublished name. Elias 1469 was identified by Standley as ‘““Coccoloba leptostachya,” which is neither similar nor apparently related. Coccoloba colombiana is similar to C. lehmannii and C, nitida, differing from both in the density of the flowers on the inflorescence, the origin of 1961] HOWARD, STUDIES IN THE GENUS COCCOLOBA, X 89 the petiole above the base of the ocrea, and the asymmetrical nature of the leaf base. Coccoloba duckéi, sp. nov. Arbuscula subscandens ramulis teretibus, crispo-ferrugineo-tomentosis; ocreis membranaceis, 1.5—2 cm. longis, pubescentibus: petiolis supra basim ocrearum gerentibus, 3-5 mm. longis, pilosis; laminis ovato-oblongis, pu- berulentis vel glabratis, tenuiter coriaceis, 5 & 3.5 vel 10 & 5.5 cm. longis latisque, apice acutis vel subrotundatis, basi pare vel subcordatis, nerviis primariis 8 vel 10; inflorescentibus racemosis, 10-14 c m. longis, rhachibus striatis puberulentis, bracteis ovato- omnes 2.5 mm. longis, nigris, puberulentis vel ad marginem ciliatis, ocreolis membranaceis, 2.5 mm. longis; floribus masculinis 3 vel 4 per nodulum; floribus femineis et fructu ignoto. Brazil. AMAzoNaAS: Boa Vista, Ducke 1358 (a—-holotype, F, NY. US). This species is named in honor of its collector, an explorer of the Ama- zon area, Adolpho Ducke. The species resembles Coccoloba obovata, from which it differs in having a densely flowered inflorescence, smaller leaves, and shorter petioles. Coccoloba llewelynii, sp. nov. Coccoloba bolivarana Williams, Trop. Woods 68: 39. 1941, nomen nudum; Howard, Jour. Arnold Arb. 41: 220. 1960. Arbor 8 m., ramulis sericeo-tomentosis glabrescentibus; ocreis trans- lucidis, 12 cm. longis, glabratis, deciduis; petiolis ad basim ocrearum gerentibus, 2-4 mm. longis, sericeo-tomentosis vel glabratis; laminis obo- vato-ellipticis vel oblanceolatis, 8 & 3 vel 12 & 6.5 cm. longis latisque, puberulentis, costa tomentosa glabrescente, apice acutis vel abrupte breve acuminatis, basi contracto-rotundatis vel truncato-subcordatis, nerviis primariis 6-8; inflorescentibus tenuibus 8-12 cm. longis, nodulis dis- tinctibus, rhachibus sparse tomentosis vel glabratis, floribus masculinis 1 ad 3 per nodulum, floribus femineis singularibus, bracteis late ovatis, membranaceis, 1 mm. longis, ocreolis bifidis, erectis, membranaceis; fructu ovoideo, 6 mm. longo, 4 mm. diametro, lobis perianthii coronatis, 1 mm longis; acheniis (oe ence Venezuela. ANZOATEQUI: Taguatagua, Pittier 14893 (us). BoLtvar: El Tigre, near Rio Cuchivero, alt. 90 m., Llewelyn Williams 13374, June 18, 1940, (F- holo- type, K). Monacas: Caicara, Foster Smith 239 (us). The collection by Williams has been referred to previously under the invalid name Coccoloba bolivarana (Howard loc. cit.). Williams described this plant as a tree of eight meters with many trunks. The Smith collection is described as a bush. The species is distinct among South American Coccolobas through the short petioled obovate-elliptic leaves and the inflorescence with distinctly separated flowering nodules. In appearance 90 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII the specimens are similar to C. obovata and C. venosa. Coccoloba llewely- nit differs from C. obovata in the type of inflorescence and in the point of attachment of the leaves. It also differs from C. venosa in the inflores- cence and the fruit. The species is named for its distinguished collector of plants in the American tropics, Llewelyn Williams. Coccoloba orinocana, sp. nov. Arbuscula vel arbor parva, 2-5 m.; ramis majoribus lacuniatis, horno- tinis teretibus, dense ferrugineo- -pubescentibus; ocreis 3 cm. longis, mem- branaceis, tomentosis vel pilosis, deciduis; petiolis leviter supra basim ocrearum gerentibus, 10-15 mm. longis, tomentosis; laminis oblongis, late ovatis vel oblongo-obovatis, 7 * 3.5 vel 12 *& 8.5 cm. longis latisque, frequente bullatis, subtus tomentosis vel pilosis, supra pubescentis in costa nervisque, apice rotundatis vel obtusis, basi rotundis vel leviter cordatis, nerviis primariis 9-14, arcuatis; inflorescentibus 18-25 cm. longis, rhachi- bus puberulentis, nodulis confluentibus, floribus masculinis 3 per nodulum, floribus femineis singularibus, bracteis late ovatis, 0.5 mm. longis, pu- sok rings ocreolis membranaceis, 1-2 mm. longis, puberulentis; fructu voideo, 4 mm. longo, 4 mm. diametro, lobis perianthii imbricatis, 1/4 Al ia hypanthium, acheniis nitidis, cinnamomeis Venezuela. BArINAS: Pedroza, Bernardi 1955 (Ny). Botfvar: Rio Pargueni, Wurdack & Monachino 39759 (a-holotype, NY); Rio Parguaza below Raudal Maraca, Wurdack & Monachino 40994 (a, NY); Rio Parguaza, Cerro Negro Parado below El Carmen, Wurdack & Monachino 40973 (a, NY). GuArtIco: El Sombrero, Pittier 12367 (A, G, US) This species is superficially similar to Coccoloba caracasana but differs in the softer leaves which are commonly bullate, the arcuate ascending nervation, and the fruiting perianth which is not lobed to the base. Coccoloba portuguesana, sp. nov. Arbor parva, ramulis teretibus, dense brunneis crisposo-tomentosis; ocreis membranaceis, 1—1.5 cm. longis, dense tomentosis, deciduis, petiolis supra basim ocrearum gerentibus, 5-12 mm. longis, tomentosis; laminis obovatis, 5.5 & 3 vel 16 & 11 cm. longis latisque, coriaceis, rugosis, inter nerviis bullatis, subtus puberulentis sed crisposo-tomentosis vel pilosis in costa axillisque, apice obtusis, subapiculatis, basim versus cuneatis, minute cordatis, nerviis primariis 8—11; inflorescentibus 6-30 cm. longis, rhachi- bus multo-striatis, puberulentis, bracteis ovatis, 1 mm. longis, puberu- lentis, ad marginem ciliatis; ocreolis translucidis, 1-2 mm. longis, pu- berulentis; floribus masculinis 3—5 per sod wlaci. floribus femineis 1—2 per nodulum; pedicellis fructiferis 0.75-1 mm. longis, fructu ovoideo, 3 mm. longo, 3 mm. diametro, hypanthio pauce manifesto, lobis perianthii basim versus distinctibus; acheniis cinnamomeis. Colombia. Without specific location, Goudot “31!"(P). Venezuela. Portvu- ESA: Ospino, Pittier 12024 (a-holotype, G, M, NY). ZAMORA: Sabaneta de Barinas, Mell s.n. (NY 1961] HOWARD, STUDIES IN THE GENUS COCCOLOBA, X 91 Lindau (Bot. Jahrb. 13: 222. 1890) cited the Goudot collection in a list of specimens seen as “C. spec.” The collecting locality is uncertain but the specimen, although smaller in size than the Pittier and Mell material, is clearly referable here. Coccoloba portuguesana resembles C. charitostachya, C. warmingii and C. latifolia in one or several characteristics. It differs from C. charito- stachya in the cuneate leaf base, longer inflorescence and the fruiting perianth which is divided to the base. The species is distinct from C warmingii in the ridged inflorescence axis and the pubescence and is clearly distinct in fertile material from C. latifolia through the racemose inflorescence and the smaller leaves. Coccoloba steinbachii, sp. nov. Arbor, 6 m., ramulis striatis, glabris; ocreis 1.5-2 cm. longis mem- branaceis, glabris; petiolis 1.5-3.5 cm. longis, supra basim ocrearum gerentibus, ad basim glabris, ad laminas pilosis; laminis ovatis vel ovato- oblongis, (5 & 3.5) 8 & 5.5 vel 14 & 9.5 cm. longis latisque, tenuiter coriaceis, inter nerviis bullatis, costa supra puberulente, subtus nerviis primariis tomentosis vel pilosis, apice abrupte acutis, basi obtusis, trun- catis vel leviter cordato-truncatis, nerviis primariis 10-14, arcuato-ad- scendentibus: inflorescentibus 32-44 cm. longis, rhachibus, bracteis et ocreolis puberulentis, bracteis ovatis, ca. 1 mm. longis, ocreolis mem- branaceis, 1-1.25 mm. longis, floribus masculinis (2—)7—13 per nodulum, floribus femineis 1—2 per nodulum,; fructu ignoto. Bolivia. Junction of the Rio Beni and Rio Madre de Dios, Rusby 1380 (GH, K, NY); San Carlos, Mapiri, Buchtien 1768 (Ny); Santa Cruz, Vista, Steinbach 5617 (a-holotype, G, GH, NY), Rio Palometillas, Stemmbach 6604 (A) The type was collected in flower on April 25, 1921. Coccoloba stein- bachii, named for its collector, is superficially similar to C. tiliacea of Argentina. Plants of C. tiliacea produce both peltate and nonpeltate leaves. Specimens of C. steinbachtt may eae the nonpeltate leaf types of C. tiliacea but differ in having more numerous veins, the leaf blades bullate between the veins, a longer aan and shorter pedi- cels. Britton and Rusby (Bull. Torrey Club 27: 129. 1900) referred Rusby 1380 to “Uvifera illhaensis” from which C. steinbachii differs in having much longer inflorescenses, longer petioles and more numerous veins to the leaves. Coccoloba wurdackii, sp. nov. Frutex adscendens, ramulis teretibus striatisque, dense ferrugineo-pu- berulentis; ocreis 5 mm, longis, rigidis, persistentibus, subampliatis; petio- lis ad basim ocreis bevenuinne. robustis 5-8 mm. longis, puberulentis; laminis late ovatis vel ovato-oblongis, 9 & 4 vel 12 XX 9 cm. longis 92 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII latisque, crasse coriaceis, glabris vel rare puberulentis, in sicco dense reticulatis, costa curvata, apice late acutis, basi cordatis vel rotundis, nerviis primariis 5 vel 8; inflorescentibus 10-15 cm. longis, rhachibus, bracteis, ocreolis dense puberulentis; floribus masculinis 2-5 per nodulum, floribus femineis singularibus, bracteis late ovatis, ca. 0.5 mm. longis, 1 mm latis, ocreolis subampliatis, bifidis, ca. 1.0 mm. longis; fructu immaturo, ovoideo basi rotundato, lobis perianthii brevibus, coronatis. mbia. VAupES: Rio Guainia opposite Maroa, Maguire, Wurdack & Keith Pe (A-holotype. Ny). Venezuela. Amazonas: Maroa, Maguire, Wurdack & Bunting 36443 (xv), Maguire & Wurdack 35693 (A, NY This species is named in honor of Dr. John Wurdack, who along with Dr. Bassett Maguire, has contributed greatly to our botanical knowledge of northern South America. Coccoloba wurdackii is similar to both C. schomburgkii and C. marginata. It differs from the former in its liana abit, longer inflorescences, and larger, reticulately veined leaves. From C. marginata the species may be distinguished by its shorter, broader, and thicker ovate leaves and heavier inflorescence axis. Coccoloba zuliana, sp. nov. Ramuli eae ocreis deciduis; petiolis ad basim ocrearum gerenti- s, 2-3 mm. longis, glabris, laminis obovatis, 2.5 & 1.5 vel 5.5 & 2.5 mm. longis ae membranaceis, apice acutis, basi cuneatis, subtus in axillis sparse tomentosis, nerviis primariis 5—8; inflorescentiis 3—4 cm. longis, rhachibus striatis, bracteis ovatis, ad marginem ciliatis, ocreolis membranaceis, ca. 1 mm. longis; floribus ignotis; pedicellis fructiferis singularibus, erectis, adscendentibus, 2 mm. longis, fructu ovoideo, 4 mm. longo, 3 mm. diametro, hypanthio non-manifesto, lobis perianthii ad basim distinctibus, imbricatis, adpressis; acheniis triangularibus, ferrugineis. Venezuela. ZULIA: Perija, Tejera 140 (us—holotype, GH). This species is known from a single collection but is clearly distinct from the other species of Coccoloba in Venezuela and in South America. The small leaves and short inflorescences with well developed peduncles mark this species. The peduncles, while only 2 mm. long, appear even longer in contrast to the thin rachis. No collector’s data are available regarding the size or habit of the plant. The specimens have also lost all the ocreae. Clusters of hairs occur in the axils of the veins on the lower leaf surface. The remainder of the leaf surface, the remnants of the ocreae, and the inflorescence rachis show resinous dots which may represent hair bases or excretions of some form. Coccoloba zuliana appears to be re- lated to C. obtusifolia, differing in the shape of the leaves, the straight peduncles and the ovate fruit. In leaf shape C. zuliana resembles some collections of C. peruviana but differs in the size of the leaves, as well as the length of the peduncles. 1961] HOWARD, STUDIES IN THE GENUS COCCOLOBA, X 93 The distribution of the species of Coccoloba in South America. Numerals on the map refer to the number of species recognized in each state of Brazil and in the other countries of South America. No species have been reported from Chile. The dotted line indicates the recorded limit of distribution in northern Argentina. 94 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII DISTRIBUTION WITHIN SOUTH AMERICA Nearly half of the species of Coccoloba reported from South America are inadequately known. The flowers of one sex or the other, the fruits, or the leaves from adventitious shoots may be needed for a complete un- derstanding of the individual taxa. All attempts to produce a key to the species of the genus occurring in South America have been unsatisfactory. For the present, therefore, it appears most useful to list the species occur- ring in defined political and geographical areas. The largest number of species (44) occurs in Brazil, with 20 species reported from the state of Rio de Janeiro. Venezuela has 27 species. The largest number of in- adequately known species also occurs in these two areas, especially the vicinities of Rio de Janeiro and Caracas. Many species from the neigh- borhoods of these two major cities are known only from the type collections made one hundred to two hundred years ago. Modern collectors have not located additional material. Whether the species or the habitats have been eliminated by the growth of these urban centers, or whether repre- sentatives of these species would be located by careful scrutiny of the remaining vegetation, only time will tell. Collections representing the following species have been cited in this or the preceding paper of this series, or have been received recently from South American botanists. For ease in locating specific names the generic abbreviation has been omitted throughout. Argentina argentinensis, cordata, paraguariensis, tiliacea Bolivia ijabensis, longipes, meissneriana, mollis, persicaria, peruviana, spinescens, steinbackié Sliacéa Brazil ACRE: acuminata, confusa, densifrons, duckei, lepidota, marginata, mollis, ovata, — amie schomburgkii, spruceana s: alagoen: AMAZONAS: one confusa, msi duckei, marginata, mollis, ovata, Pe eae peruviana, schomburgkii, spruce Banta: alnifolia, alagoensis, arborescens, ities cruegeri, ilheensis, laevis, marginata, mollis, ochreolata, ovata, pipericarpa, plantagi ined, ramosissima, rosea CEARA: latifolia, mollis EspPIrITO SANTO: ochreolata Goyaz: densifrons, latifolia, marginata, mollis MarANHAO: latifolia, mollis, sparsifolia Marto Grosso: brasiliensis, cujabensis, longipes, mollis, peruviana Minas GERAES: acrostichoides, alagoensis, brasiliensis, cerifera, declinata, glaziovti, grandiflora, lanceolata, marginata, mollis, pipericarpa, ‘sticticoulis, WarmingiL ARA: acuminata, confusa, densifrons, ovata, parimensis PARANA: glaziovii PERNAMBUCO: alnifolia, laevis, mollis 1961] HOWARD, STUDIES IN THE GENUS COCCOLOBA, X 95 Prauny: mollis Rio DE JANEIRO: acrostichoides, — alnifolia, arborescens, confusa, hee gardneri, glaziovii, grandiflora, laevis, mollis, ochreolata, peltata, arpa, ramosissima, rigida, salicifolia, schwackeana, ee warmingit Ro GRANDE: cordat SANTA CATARINA: arborescens, warmingit SAo PauLo: confusa, cordata, glaziovii, mosenii, warmingu British Guiana acuminata, charitostachya, cruegeri, excelsa, gymnorrhachis, latifolia, lucidula marginata, ovata, parimensis, savannarum, schomburgkii, sparsifolia, striata Colombia cuminata, caracasana, colombiana, coronata, densifrons, dugandiana, filipes, re ees nitida, vata, obtusifolia, ovata, padiformis, portuguesana, trianaet, uvifera, venosa, see a Dutch Guiana conduplicata, excelsa, gymnorrhachis, latifolia, marginata, mollis, uvifera cuador acuminata, coronata, densifrons, mollis, obovata, ruiziana French Guiana excelsa, latifolia, lucidula, marginata, mollis, parimensis Paraguay cordata, cujabensis, guaranitica, paraguariensis, spinescens Peru acuminata, coronata, densifrons, gracilis, mollis, nutans, padiformis, parimen- sis, peruviana, ruiziana Uruguay argentinensis, cujabensis Venezuela acuminata, caracasana, coronata, cruegeri, dugandiana, bei fallax, lati- folia, peeiiee lehmanniz, jucidult: llewelynii, marginata, obtusifolia, orinocana, diformis, parimensis, ier Sea schomburgkii, pecs striata, swartzii, uvifera, venosa, wurdackit, zuliana 96 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII THE GENERA OF THE CYRILLACEAE AND CLETHRACEAE OF THE SOUTHEASTERN UNITED STATES ! JoasB L. THomas CYRILLACEAE Lindley (CyrtLia FAMILy ) Deciduous or evergreen shrubs or small trees, leaves simple, alternate, en- tire, exstipulate, short-petioled [sessile], glabrous. Inflorescences terminal or axillary racemes, each flower in the axil of an obovate, spatulate, or lan- ceolate, caducous or persistent bract; pedicels with [or without] 2 lanceo- late or oblanceolate bracteoles. Sepals 5, rarely 6 or 7, equal [unequal], {quincuncially| imbricate in bud, ovate, lanceolate, or deltoid, persistent. Petals as many as the sepals, spatulate to oblong-elliptic, white or pinkish white [deep pink or violet]. Stamens 5-10, the insertion hypogynous, fila- ments terete or flattened and petaloid below the middle; anthers versatile, dehiscing through longitudinal slits [or morphologically basal pores]; pol- len grains single, tricolporate (or tetracolporate), with wide longitudinal furrows which narrow in polar view to produce a Y-shape. Gynoecium syncarpous, ovary superior, 3—5-locular, each locule containing 1-3 ovules; style short, stigma 2—5-lobed [or not lobed]. Fruit indehiscent, globose or distinctly 3—5-winged, often devoid of seeds. Seeds without seed coats: endosperm cellular; embryo straight, cylindrical. A small family of three genera and fourteen species, distributed along the Coastal Plain of the southeastern United States, in the West Indies, Central America, and northern South America. Cyrilla and Cliftonia, each con- taining but a single species, occur in our area. The third, Purdiaea Planch., is composed of 12 species, oo in Cuba, British Honduras, Guate- mala, and northern South America Most authors who have ceesed a phylogenetic classification of the Cyrillaceae have placed this family either near the Celastrales or the Ericales, with a few suggesting a relationship with both of these groups. The Cyrillaceae have been shifted from one to the other many times, and “Prepared for a biologically oriented generic flora of the southeastern United ates, a joint project 9! the Arnold Arboretum and the Gray Herbarium made pos- sible through the support of George R. Cooley and — char Science Foundation, and under the direction of R. C. _ and C. E d, Jr. The scheme follows that outlined at the beginning of the series (Jour. i Arb. 39: 296-346. 1958). ts The figures of Cyrilla and Clethra were drawn nd Dorothy H. Marsh, in part from matcrial collected by R. B. Channell and H. F. L. Ro 1961 | THOMAS, CYRILLACEAE AND CLETHRACEAE 97 have been treated as a subfamily of both the Celastraceae and the Eri- caceae. They were also occasionally changed from one of these groups to the other in later editions by the same author (Endlicher, 1840, 1841; Lindley, 1835, 1847). As the family has been studied in greater detail, however, the evidence has become increasingly strong for a close relation- ship between the Cyrillaceae and members of the Ericales, particularly the Clethraceae. The Cyrillaceae have a well-documented fossil record, with cyrillaceous pollen grains occurring as far back as the Upper Cretaceous, thus indicating that the family is a relatively old one. Although these pollen grains cannot be definitely assigned to any of the living members of the Cyrillaceae, they are quite similar to the pollen of Cyrilla and Cliftonia. The best documented fossil record is in the Brandon Lignite of Vermont. The most abundant wood and the second most abundant pollen in this fossil flora is that of Cyrilla (Spackman, Traverse). The wood, and particularly the pollen, from this deposit is very similar to that of living material — re- markably so in view of the age of the deposit, estimated as late Upper Oligocene by Barghoorn. The Cyrillaceae are distinguished from related families by the outer whorl of stamens opposite the sepals, the dry, indehiscent fruit, the ovary with 1-3 ovules per locule, and seeds without seed coats. REFERENCES: BaiLton, H. Ilicacées. Hist. Pl. 11: 211-220. 1892. [Includes Cyrilla and Cliftonia. | Beatvisace, L. Contributions a l'étude anatomique de la famille des Ternstroe- miacées. 470 pp. Tours, 1920. [Includes discussion of the anatomy, partic- ularly of the leaves, of Cyrilla and Cliftonia. | BeNTHAM, G., and J. D. Hooker. Cyrilleae. Gen. Pl. 2: 1225, 1226. 1876. CANDOLLE, A. DE. Cyrilla. Prodr. 17: 292. 1873. CopELAND, H. F. Observations on the Cyrillaceae particularly on the reproduc- tive structures of the North American species. Phytomorphology 3: 405 411. 1953, ENpLIcHER, S. Gen. Pl. 1413. 1840. [Ericaceis affines. Cyrilleae _ Ilicineae. Enchiridion Bot. 577-579. 1841. [Includes Cyrilleae. ] Gitc. E. Cyrillaceae. Nat. Pflanzenfam. III. 5: 179-182. GriseBAcH, A. H. Erlaiuterungen ausgewahlter Pflanzen des tropischen Amerikas. Abh. Ges. Wiss. Gottingen 9: 3-58. 1860. [Cyrileen, 45.] Hermscu, C., Jr. Comparative anatomy of oe secondary xylem in the “Gruni- ales” and “Terebinthales” of Wettstein with reference to taxonomic group- ing. Lilloa 8: 83-198. 1942. [Includes a discussion of the anatomy and relationships of the Cyrillaceae. | Linptey, J. Celastraceae. Nat. Syst. Bot. 2: 119. 1836. [Includes Cyrilla and Cliftonia. | . Cyrillaceae. Veg. Kingdom ed. 2. 445. 1847. Monr, C. T. Cyrillaceae. Plant Life of Alabama. Contr. U. S. Natl. Herb. 6: 601. 1901. SARGENT, C. S. Cyrilla, Cliftonia. Silva N. Am. 2: 1-8. pls. 31, 52. 1891. SpacKMAN. W. The flora of the Brandon Lignite: geological aspects and a com- 98 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLIx parison of the flora with its modern equivalents. Sear ets Ph.D. Thesis, Biology Dept., Harvard University, Cambridge, Mass., STEEVES, M. W. Pollen and spores of the Raritan and Magothy (Cretaceous) formations of Long Island. Unpublished Ph.D. Thesis, Biology Dept., Har vard University, Cambridge, Mass., 1959. Tuomas, J. L. A monographic study of the Cyrillaceae. Contr. Gray Herb. 186: 1-114. 19 Torrey, J., and A. Gray. Celastraceae. Fl. N. Am. 1: 256. 1838. [Includes Cyriila and Cliftonia. | TRAVERSE, A. Pollen analysis of the Brandon Lignite of Vermont. U. S. Bur. Mines Rep. Invest. 5151. 107 pp. 1955. Upuor, J.C. Cyrillaceae. Nat. Pflanzenfam. ed. 2. 20: 1-12. 1942. KEY TO THE GENERA OF CYRILLACEAE Stamens 5, filaments terete; fruit globose or subglobose, longitudinally bisulcate OP TEWMIORE, ay 6csc charg eg otc beans pe pad dnck (4k kk 1. Cyrilla. Stamens - filaments broad and petaloid below the middle: fruit distinctly 3—-S-winged. 260... c cece eee. 2. Cliftonia. 1. Cyrilla Garden ex Linnaeus, Mantissa Pl. 5. 1767. Shrubs or small trees [up to 25 m. tall], usually branching near the base; bark light gray or brown. Roots shallow, usually running §hori- zontally a few inches beneath the surface of the soil. Leaves simple, alter- nate, entire, exstipulate, short-petioled, subcoriaceous to coriaceous, dark green, shining above, pale beneath, often clustered toward the ends of branches, densely reticulate, the midvein prominent beneath, prominent or depressed above. Flowers in slender, glabrous, erect or pendulous ra- cemes, borne near the ends of branches, in the axils of leaves or leaf scars of the previous growing season; pedicels articulate at the point of attach- ment to the rachis, persistent; bracts lanceolate, persistent; 2 bracteoles borne above the middle of the pedicel, alternate or opposite, lanceolate, persistent. Sepals 5, rarely 6, free or occasionally coalescent laterally at the extreme base, persistent, ovate-lanceolate to deltoid, imbricate in bud. Petals as many as the sepals, white or pinkish, inserted on a small glandular disc, membranaceous laterally and above the middle, thickened and glandular on the inner surface medially and below the middle. Sta- mens 5, opposite the sepals; filaments terete, subulate; anthers versatile, attached near the middle on the dorsal side, 2-lobed, the lobes free below the point of attachment to the filament, united above. Ovary sessile, 2- or 3(rarely 4)-locular, ovules pendulous, anatropous, 1-3 in each locule: stigma 2—4-lobed, the lobes equal to the number of locules. Fruit globose to ovoid or obpyriform, dry, indehiscent, often devoid of seeds, not more than 1 seed developing in each locule. Seeds pendulous, elongate; embryo elongate, cylindrical, the cotyledons short, inferior, the radicle superior. TYPE SPECIES: C. racemiflora L. (Named in honor of Dominico Cirillo, an Italian physician and professor at Naples.) — LEATHERWOOD, TITI. 1961] THOMAS, CYRILLACEAE AND CLETHRACEAE 99 A genus of a single species, Cyrilla racemiflora L., easily distinguished from other Cyrillaceae by the pair of persistent bracteoles on the pedicel, petals medially thickened and glandular, five stamens opposite the sepals and three ovules per locule. It is a plant of acid, and usually sandy, soil along the margins of swamps and small streams and in wet pinelands distributed in our area along the Coastal Plain from Florida to southeastern Virginia and to southeastern Texas. It also occurs in Mexico, British Hon- duras, the West Indies, and northern South America. oe) DO Oy EX ee RE? teks BEE Ss Muss Nore ea tived; wiles Na Ur Age ABH ne ath i. ae C. racemiflora: a, fruiting branchlet, x ie b, flower frui 10; d, nearly DUAL de 1 5: mature fruit in ver rtical section — note a ee (not 132 fruit, semi- sine filling left rene three ovules in right locule, in left locule sae by stippled diagrammatic cross section, po osition of seed area, embryo omit tted — note parenchymatous development of pericarp, 10 100 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII This widespread species shows a high degree of local variation in the mountains of northern South America, in Oriente Province, Cuba, and in the southeastern United States, especially in northern Florida. On the basis of this variation, some authors have divided the group into several species, including C. parvifolia Raf. and C. arida Small in the Southeast. The different variation patterns are graded, however, with intermediate forms connecting the various extremes. Cyrilla reproduces prolifically by means of adventitious shoots which arise from the roots. These root sprouts spread out radially from an in- dividual, and as they increase in size and become established, they, in turn, give rise to additional sprouts. In some areas, clones of this sort become quite extensive and may cover an acre or more. The buds which give rise to the shoots are exogenous in origin, developing in or near the cork cambium, and apparently do not develop before the initiation of secondary activity. The buds remain in the root for three to four years before emerging as a shoot. During this time the bud establishes vascular connection with the root and increases in size at a rate which is just suffi- cient to keep pace with the secondary activity of the root. This sprouting mechanism is the primary means of reproduction in the Cyrillaceae, for the abundant fruits produced each year rarely contain seeds. Parthenocarpic fruits of this type are found, almost without ex- ception, toward the middle of a large clone, thus indicating that Cyrilla must have a rather high degree of self-sterility. Apparently self-fertiliza- tion provides sufficient stimulus to initiate fruit development, but seeds are developed only when a plant receives pollen from a different individual. Cyrilla is occasionally cultivated as an ornamental in the southeastern United States and in a few botanical gardens outside that area on the East Coast and in Europe. It is valued for its graceful racemes of white flowers which bloom over a period of a month or more and are followed by small yellow fruits which are also attractive and persist until late autumn. In some areas Cyril/a is an important source of nectar for honey known locally as “‘titi” honey, in reference to one of the colloquial names for Cyrilla, and said to have a delicate and very delicious flavor. REFERENCES: BARGHOORN, E. S. Geological and botanical study of the Brandon Lignite and its significance in coal petrology. Econ. Geol. 45: 344-357. and W. SPACKMAN. A seeps study of the flora of the Brandon Lignite. Am. Jour. Sci. 247: 33-39. 1947. FERNALD, M. L. Overlooked ies. atk and novelties in the flora of east- ern North America. Rhodora 46: 46. 1944. | Notes on C. racemiflora var. subglobosa. | LUNDELL, C. L. Cvyrillaceae. Fl. Texas 3: 109-111. 1943. Masters, M. T. Cyrilla racemiflora. Gard. Chron. I]. 30: 198. 1901 SHUTTLEWoRTH, R. Cyrilla parvifolia. In Nasu, Notes on some Florida plants Ti. Bull. tg Bot. Club 23: 101, 102. 1896. StoctTFMYER, V. T., and F. L. O’Rourke. Vegetative propagation of Cyrilla. Am. Nurseryman 76(11): 5, 6. 1942 1961 | THOMAS, CYRILLACEAE AND CLETHRACEAE 101 2. Cliftonia Banks ex Gaertner f. Fruct. et Sem. Pl. 3: 246. 1805. Evergreen shrubs or small trees, usually branching near the base, rarely becoming up to 12 m. tall; bark reddish brown or gray, thick and spongy on the trunks of older trees; young stems dark red, smooth, becoming gray and scaly after 1-2 years. Leaves simple, alternate, entire, exstipu- late, short-petioled, subcoriaceous or coriaceous; light green, shining above, pale, glaucous beneath; oblanceolate-elliptic or rarely obovate-oblanceo- late, apex acute or slightly emarginate. Flowers in terminal or axillary racemes, glabrous; rachis attenuate, ridged, each ridge terminating in a small, concave mound at the base of a bract; bracts white or occasionally pinkish, usually turning brown above the middle, membranaceous, obo- vate-spatulate or spatulate, slightly concave and thickened basally, articu- late at the base, caducous well before anthesis; 2 bracteoles borne alter- nately on the pedicel, usually shedding shortly after anthesis. Sepals 5, rarely 6 or 7, white or pinkish, deltoid, persistent, membranaceous at an- thesis, becoming foliaceous at maturity. Petals white or pinkish, im- bricate in bud, occasionally weakly clawed near the base, longitudinally 3-veined. Stamens 10 in 2 whorls of 5, the outer opposite the sepals and inserted lower on the receptacle than the inner which is opposite the petals; filaments laterally expanded and petaloid below the middle, narrowing abruptly, becoming terete and subulate above; anthers versatile, attached slightly below the middle. Ovary borne on a small, slightly concave, gland- ular nectariferous disc, 3—5-locular, each locule with a single, pendulous ovule: stigma massive, subsessile, 2—5-lobed. Fruit 2—5-winged, dry, in- dehiscent, usually devoid of seed, but sometimes with as many as 5. Seeds consisting of an elongate embryo with short cotyledons, surrounded by a mass of cellular endosperm. Type species: Cliftonia monophylla (Lam.) Britton ex Sarg. (Named in honor of D. G. Clifton, Chief Justice of West Florida) — BUCKWHEAT TREES. A genus of a single species, Cliftonia monophylla, distributed along the Coastal Plain of southern Georgia, Alabama, Mississippi, and northern Florida. Cliftonia apparently has a very narrow ecological tolerance, as well as geographical range, occurring only in very wet, sandy, acid soils. It is nearly always found growing in association with Sarracenia L. Cliftonia is very distinctive and is easily distinguished from related genera by the winged fruit which resembles that of Fagopyrum, the mas- sive, 3—5-lobed, subsessile stigma, and the filaments which are laterally expanded and petaloid below the middle. In spite of its distinctiveness, Cliftonia was originally described as a species of Ptelea by Lamarck. appar- ently on the basis of the winged fruits. The fruits of Cliftonia, as of Cyrilla, are usually parthenocarpic, and the plants reproduce primarily by root sprouts. In contrast to Cyvilla, however, Cliftonia shows very little variation throughout its range and has been treated as a monotypic genus by every author who has dealt with this group. The narrow geographic range and the lack of variation indicate that Cliftonia is probably a relict form in the present-day flora. 102 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII REFERENCES: BRITTEN, J. Note on the history of Cliftonia. Jour. Bot. 43: 282-284. 1905. NuttTaLL, T. Buckwheat Tree. N. Am. Sylva 2: 39-44. 1865. RENDLE, A. B. Notes on Myricaceae. Jour. Bot. 41: 87. 1903. [Includes notes on the origin - the name Cliftonia. | Watson, S. A point in nomenclature. Bull. Torrey Bot. Club 14: 167. 1887. |On the aes name of Cliftonia. | WILLDENOow, K. L. Mylocaryum (Cliftonia). Enum. Pl. Berol. 454. 1809. CLETHRACEAE Klotzsch (Waite ALDER FAMILy) Deciduous shrubs or small trees with alternate, simple, exstipulate, serrate-margined leaves. Flowers in terminal racemes or panicles, bisexual, regular, the insertion hypogynous. Rachis, bracts and pedicels densely clothed with stellate [simple] trichomes. Sepals 5, persistent, quincun- cially imbricate, united at the extreme base, free above. Petals 5, free, alternate with the sepals, deciduous. Stamens 10 [rarely 12], in 2 whorls of 5, the outer whorl opposite the petals, the inner opposite the sepals; filaments glabrous or pubescent, adnate to the corolla at the extreme base; anthers extrorse in bud, becoming inverted and introrse at anthesis by a straightening of the filament, dehiscing by means of apical (morpho- logically basal) slitlike pores; pollen grains single, tricolporate. Gynoe- cium syncarpous, the ovary superior, 3-locular, with axile placentation; ovules numerous, small, anatropous; style single, dividing near its apex into a 3-lobed stigma. Fruit a 3-valved, loculicidal capsule. Seeds numer- ous, small [occasionally winged], with fleshy endosperm and very thin seed coats, the embryo cylindrical. Type Genus: Clethra L. A small family composed of a single genus, Clethra, with 30-40 spe- cies primarily of tropical or subtropical Asiatic and American distribu- tion. Schizocardia Smith & Standley, from Central America, was described in the Clethraceae, but is now known to be a species of Purdiaea (Cyril- laceae). The Clethraceae were first raised to family status by Klotzsch on the basis of the regular flowers with separate petals, anthers with poricidal dehiscence, three-lobed stigma, and loculicidal capsule. Most subsequent authors have agreed with this treatment, and generally the Clethraceae have been considered to be primitive members of the Ericales. However, recent authors have suggested that certain of the characteristics of Clethra which are often considered as primitive, may be derived. Among these Copeland (1943) included the tricarpellate ovary with loculicidal dehis- cence, and Kavaljian pointed out that the corolla of separate petals may be a derived condition in the Clethraceae, as was previously suggested for this condition in the Pyroleae and Monotropoideae by Copeland (1947). Also, the vascular supply to the sepals in the species of Clethra wluch have been studied show a single trace with a single associated gap, a condition generally considered to be derived, not primitive. The seeds 1961] THOMAS, CYRILLACEAE AND CLETHRACEAE 103 of Clethra, with seed coats only a single cell layer thick, also indicate a reduction REFERENCES: BaILLton, H. Ericacées. Hist. Pl. 11: 122-210. 1892. [Includes Clethra as series Clethreae. Britton, N. L. Clethraceae. N. Am. FI. 29: 3-9. 1914. CANDOLLE, A. P. pe. Clethra. Prodr. 7: 588-590. 1839. [Under tribe An- dromedeae of Ericaceae. | CopELAND, H. F. A study, anatomical and taxonomic of the genera of Rhodo- dendroideae. Am. Midl. Nat. 30: 533-625. 1943. [Comments on Clethra under “Bejaria.”’ | servations on the structure and classification of the Pyroleae. Ma- drono 9: 65-102. 1947 Drvupe, O. Clethraceae. Nat. Pflanzenfam. IV. 1: 1, 2. 1897. Gray, A. Clethraceae. Syn. Fl. N. Am. 2: 44, 45. 1878. KAVALJIAN, L. G. The floral morphology of Clethra alnifolia with some notes on C. a and C. arborea. Bot. Gaz. 113: 392-413. 1952. KuotzscH, J. Stiidien uber die natiirliche Kies Bicornes Linné. Linnaea 24: 1-88. ie Lamarck, J. Clethra. Encycl. Meéth. Bot. 2: 45, 46. 1786. [C. tomentosa, 46.] Sm1TH, A. C., and P. C. STANDLEY. Schizocardia. a new genus of trees of the family Clthracee Trop. Woods 32: 8-14. Tuomas, J. L. Schizocardia belizensis: a species of Purdiaea (Cyrillaceae) from ee America. Jour. Arnold Arb. 42: 110. 111. 1961 1. Clethra Linnaeus, Sp. Pl. 1: 566. 1753; Gen. Pl. ed. 5. 188. 1754. Deciduous shrubs or small trees, leaves alternate, simple, obovate, ovate or oblong, with serrate margins, Flowers in terminal racemes or panicles, fragrant. Bracts subtending the flowers, lanceolate, caducous or persistent, usually dropping shortly after anthesis; densely clothed with stellate or simple trichomes or both. Pedicels ridged and somewhat flattened laterally, increasing considerably in length during the maturation of the flower. Sepals 5, equal, quincuncially imbricate (with 2 sepals that are totally external, 2 totally internal, and 1 with 1 edge external and 1 edge internal), united at the extreme base, free above, lanceolate, elliptic, or ovate, the apex acute or obtuse, persistent in fruit. Petals 5, free, arranged in a quincuncial pattern in bud, spreading apart at anthesis, white or occasion- ally pinkish, alternate with the sepals, glabrous [ciliate margins], obovate or obovate- elliptic. Stamens 10, in 2 whorls of 5, the outer whorl oppo- site the petals, arising at a slightly lower level and often somewhat shorter than the inner whorl which is opposite the sepals; filaments glabrous or pubescent, somewhat flattened, laterally expanded and adnate to the corolla at the extreme base; anthers extrorse in bud, becoming inverted and introrse at anthesis by a straightening of the filament, 2-lobed, sagit- tate. the lobes separate above the middle for about one-half their length, but remaining together below the middle, becoming attenuate toward the base, the two thecae of each lobe becoming confluent shortly before the 104 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII pollen is shed, dehiscing by means of apical Spams aon send basal) pore- like slits; pollen grains single, tricolporate. Gynoecium syncarpous, pubes- cent to hirsute, ovoid to subglobose, with a small practice at the apex at the point of connection with the style; 3-locular, each locule containing a single, pendent, placental arm which arises from a central column of tissue extending longitudinally from the receptacle to near the apex of the ovary; ovules numerous, small, anatropous, borne on the outer surface of the placental arm; style single, cylindrical, dividing near the apex into 3 lobes which are stigmatic at their tips, shorter than the stamens at anthesis, becoming equal to or longer than the stamens shortly after an- thesis. Fruit a globose or subglobose, 3-lobed, 3-valved, loculicidal cap- sule, more or less inclosed at maturity by the persistent calyx. Seeds numerous, small, irregularly ovoid [occasionally winged], with very thin seed coats. Type species: C. alnifolia L. (Name Greek, Clethra, alder, alluding to the resemblance of the leaves of the type species to those of an alder.) — SWEET PEPPERBUSHES. A genus of 30-40 species, represented in our area by only Clethra alni- folia L., C. acuminata Michx., and C. tomentosa Lam. Clethra tomentosa, treated by some authors as a variety of C. alnifolia, differs from that species in having leaves that are densely tomentose on the lower surface and sepals that are longer and obtuse, rather than acute. De Candolle divided Clethra into two sections: EUCLETHRA (S$ Clethra) with exserted stamens and styles, and Cuexrarra, with included stamens and styles. Drude described a third section intermediate between these to include a single species, C. arborea, with stamens included but with the style exserted. Hu, in a revision of the Chinese species, added still an- other section, again to include only a single species, and divided sect. CLETHRA (§ Euclethra DC.) into four series. According to this scheme, the species in our area belong in sect. CLETHRA, ser. Alnifoliae. In common with a majority of the Ericales, species of Clethra are usu- ally found growing in acid soils. Clethra alnifolia is distributed in acid swamps and low, moist woods and pinelands along the Coastal Plain from Maine to Texas, with occasional extensions into the Piedmont, par- ticularly in the Carolinas, Clethra tomentosa (or C. alnifolia var. tomen- tosa Michx.) is more restricted, occurring only on the Coastal Plain from southern North Carolina to Louisiana. The two thus overlap geographically in this area. Clethra acuminata, distinguished from the other species in our area in having oval or oblong, acuminate leaves, solitary racemes, and pubescent filaments, is geographically isolated from the other species, occurring on the Appalachian Plateau and inner Piedmont, from southern West Virginia to northern Georgia. Chromosome numbers have been reported by Hagerup for C. arborea (# = 8), and C. alnifolia (n = 16). Other genera of the Ericales for which a base number of 8 has been reported are Calluna and Monotropa in the Ericaceae (Hagerup, Love & Live) and Melichrus in the Epacridaceae (Smith-White). Several other genera of the Epacridaceae, however, have 1961] THOMAS, CYRILLACEAE AND CLETHRACEAE 105 been reported to have a base number of » = 4, which could easily be the original base number for the Clethraceae. IG, 2; sake on alnifolia: a, iin nea some ss es and racemes removed, d, anthers, in and outer views, ovary and base of in one section, em oe nai canal in hoes 5 diagrammatic cross section, X 15; g, mature capsule, seeds shed — a ot aaa x 8 The unusual structure of the stamens in the Clethraceae has been the subject of considerable discussion, and there is still no universal agree- ment as to whether the anther is in the normal position in bud and in- verted at anthesis or vice versa. A majority of recent authors is of the opinion that the anthers are inverted at anthesis, so that the apparent apex is the morphological base. This interpretation, which is that followed in the present treatment, is based on the pattern of the vascular trace in the anther. At anthesis, the stamen trace bends downward just beyond the point at which it enters the anther. There is, however, a small ex- tension of the trace which continues upward in the connective tissue for a short distance. In the normal pattern of stamen vascularization, in taxa in which the stamens do not undergo a 180° rotation, the stamen trace passes longitudinally through the anther toward the apex. A pattern similar to this is found in the stamens of Clethra before anthesis. Thus 106 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII the anther is interpreted as being in the normal position in bud and in- verted after anthesis. Hu disagrees, stating that the 180° rotation of the anther is merely the result of a straightening of the filament, and only after the filament has straightened is the anther normally oriented. In this connection, she suggests that the anther rotation is comparable to the unfolding or unrolling of a leaf. Several species of Clethra, including the three species in our area, are cultivated as ornamentals in many parts of the world. They are easily propagated either by seeds or by soft-wood cuttings. The cultivated species are prized for the spicy fragrance of their blossoms, and for their mid- or late-summer flowering period —a time when few plants are in flower. Clethra alnifolia {. rosea Rehd., a pink-flowered form which occurs sporadically in the wild, is commonly found in cultivation. REFERENCES: BENTHAM, G., and J. D. Hooker. Clethra. Gen. Pl. 2: 603, 604. 1876. BRITTON, E. G. Clethra alnifolia. Addisonia 1: 23. 1916. Fox, H. M. Three hardy native American shrubs. Gard. Chron. III. 122: 89, FuLiinc, E. H. Two midsummer flowering shrubs — Abelia and Clethra. Jour. N. Y. Bot. Gard. 28: 237-242. Hacerup, O. Morphological and cytological studies of Bicornes. Dansk Bot. Ark. 6(1): 1-27. 8. Hanpy, L. H. A pink-petaled form of Clethra alnifolia. Rhodora 9: 195, 196. 1907. Hu, S. Y. - ees of the genus Clethra in China. Jour. Arnold Arb. 41: 164— 190. LECHNER, - asia Untersuchungen tber die Gattungen Actinidia, Sau- rauia, Clethra und Clematoclethra mit besonderer Beriicksichtigung ihrer Stellung i im ii Beih. Bot. Centr. 32(1): 431-467. 5, Martin, A. C. The comparative internal morphology of seeds. Am. Midl. Nat. 36: rw 1946. [C. alnifolia, 585.] Meissner, C. F. Clethra. In Martius, Fl. Brasil. 7: 165-170. 1863. Renper, A. Einige neuere oder seltenere Gehdlze. Mitt. Deutsch. Dendr. Ges. 1907(16): 69-76. [1908.] [C. alnifolia f. rosea, 75.] SARGENT, C. S. Clethra alnifolia var. tomentosa. Gard. Forest 4: 64. 1891. Witson, C. L. The phylogeny of the stamen. Am. Jour. Bot. 24: 686-699. 1937. 1961] HOWARD, STUDIES IN THE GENUS COCCOLOBA, XI 107 STUDIES IN THE GENUS COCCOLOBA, XI. NOTES ON THE SPECIES IN ASIA RIcHARD A. HOWARD THE GENUuS Coccoloba, as recognized today, i is native only in the West- ern Hemisphere, although several epithets in the genus, based on speci- mens from Asia, were published in the eighteenth and nineteenth centuries. All of these, however, either have been referred to other genera of the Polygonaceae or represent nomina nuda which cannot be identified. A few species of Coccoloba have been introduced into cultivation in Asia, and C. uvifera, at least, has become established in several areas. The following species were described and based on specimens from Asia. I have not examined the types of any of them Coccoloba adpressa Meisner ex Steudel, Nomencl. ed. 2. 1: 390. 1841. Steudel attributed to Meisner the combination Coccoloba adpressa (Labill.) Meisner, based on Polygonum adpressum Labillardiére (Nov. Holl. Plant. Spec. 1: 99. fig. 127. 1804). Meisner, however, referred to the species as Muehlenbeckia adpressa (Plant. Vasc. Gen. 2: 227. 1840) and it is so treated in modern Australian floras. Coccoloba asiatica Loureiro, Fl. Cochinch. 239. 1790. Jackson (Index Kewensis 1: 573. 1895) referred this name to Polygo- num asiaticum. In a study of Loureiro’s work, Merrill (Trans. Am. Philos. Soc. II. 24: 143. 1935) stated that Loureiro “described the same species twice under Coccoloba, C. asiatica ... and C. cymosa ...” Merrill referred both species from Cochinchina to Polygonum chinense L. Coccoloba australis Forster, Prodr. 29. 1786. The Forster specimen from New Zealand was transferred to Muehlen- beckia by Meisner (Plant. Vasc. Gen. 2: 227. 1840). This genus has been conserved, with M. australis (Forst.) Meisn. designated as the type species. Coccoloba chita ? (Hort.), Cat. Pl. Hort. Bot. Bogor. 91. 1866, nomen nudum This epithet, published without description, was not included in later editions of the catalogue. No material bearing this name has been pre- served. 108 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLII Coccoloba crispata Buchanan-Hamilton in Roxburgh, Fl. Ind. 2: 292. 1832 Meisner (DC. Prodr. 14: 130. 1856) listed this species with question in the synonymy of Polygonum chinense L. var. thunbergiana Meisner. He gave credit for this placement to Wallich in notes on record in the Grif- fith herbarium. Hooker (FI. Brit. Ind. 5: 45. 1886) accepted the syn- onymy under P. chinense without question. The original description stated that the plant is a native of Nepal and was growing in the Botanic Garden at Calcutta. Coccoloba cymosa Loureiro, Fl. Cochinch. 240. 1790. Jackson (Index Kewensis 1: 573. 1895) referred this epithet to Poly- gonum cymosum without citing an author for the latter name, but Index Kewensis lists Polygonum cymosum Roxburgh, Hort, Beng. 88. 1814 [nomen nudum | and FI. Ind. 2: 289. 1832 (which has been referred to P. chinense L.), as well as Polygonum cymosum Treviranus, Nov. Act. Nat. Cur. 13: 177. 1866, referred to Fagopyrum cymosum Meisner. Merrill (loc. cit. A concluded that the Loureiro species was to be associated with P. chinense L Coccoloba dubia Hort. ex Trimen, Hort. Zeyl. 67. 1888, nomen nudum. No specimens are available to document this name which has not been uesd in more recent literature. Coccoloba indica Wight ex Meisner in DC. Prodr. 14: 131. 1856, pro Lindau (Bot. Jahrb. 13: 220. 1890) listed “C. indica mss.” in a collec- tion of excluded species at the end of his monograph. He failed to acknowl- edge that earlier Meisner had cited this manuscript name, attributed to Wight, in the synonymy of Polygonum chinense L. var. ovalifolium Meis- ner. Coccoloba platyclada (F. Mueller) F. Mueller, Bot. Mag. 89: pl. 5382. 1863. Mueller originally described Polygonum patycladum [sic] (Trans. Phil. Inst. Vict. 2: 73. 1858) and later sent to Hooker, for publication in the Botanical Magazine, an excellent drawing and description. Since Hooker added only two lines of commentary, he should not be considered the author of this combination or of the article, as he is frequently cited. Mueller corrected the obvious typographical error in the specific name in synonymy and transferred the species to Coccoloba. Meisner (Bot. Zeit. 23: 313. 1865) considered the species as ‘““Muhlenbeckia platyclados” and 1961] HOWARD, STUDIES IN THE GENUS COCCOLOBA, XI 109 many modern floras use the epithet Muehlenbeckia platyclada. Bailey (Gentes Herb. 2: 58. 1929) discussed the generic limits involved and es- tablished a former section of the genus Polygonum as the new genus Homa- locladium for the species. The original publication data attributed the plant collection to Shepherd from New Caledonia. In the Botanical Magazine (doc. cit.) Meisner clari- fied the origin by commenting, ‘This remarkable plant was discovered at Wanderer Bay, Solomon’s Islands, by Mr. Milne, during Captain Den- ham’s Voyage of H.M.S. Herald, and a living specimen of it was com- municated to Messrs. Shepherd, of the Darling Nurseries, of Sydney, from whence the plant was received at the Melbourne Botanic Garden.” Coccoloba totnea Buchanan-Hamilton in D. Don, Prodr. Fl. Nepal. 75. 1825. Hooker saw the Hamilton specimen and referred the species to the synonymy of Polygonum molle D. Don (FI. Brit. Ind. 5: 50. 1886). Danser (Bull. Jard. Bot. Buitenz. III. 8: 233. 1927) referred both species to the synonymy of P. paniculatum Blume. INTRODUCED SPECIES Coccoloba caracasana Meisner in DC. Prodr. 14: 157. 1856. A specimen was received from Dr. Otto Degener (s.2. [A]) who reported that the species was being used as a hedge on the campus of the University of Hawaii, Honolulu, Oahu. Coccoloba uvifera (L.) Linnaeus, Syst. Nat. ed. 10. 1007. 1759. The common sea grape has been introduced as an ornamental shrub or tree for seaside plantings in many islands of Polynesia, and on Kauai, in Hawaii, I have observed plants which, having persisted after cultivation, are becoming established spontaneously. The following references are to Coccoloba uvifera in the Pacific area: Rarotonga (Wilder, Bishop Mus. Bull. 86: 47. 1931) Nuku Hiva, Marquesas (Brown, ibid. 130: 67. 1935); Vaitepaua, Makatea (Wilder, ibid. 120: 21. 1934); Batavia, Java (Dan- ser, Bull. Buitenz. Bot. Gard. III. 8: 241. 1927). I have seen the follow- ing specimens: Hawaii. OAHU: Degener 8986 (Nv), Thompson s.n. (NY). India. Calcutta, cult. Calcutta Bot. Gard. s.2. (Ny). Indochina. Hanoi, Tonkin, Petelot 3979 (NY); Saigon, Poilane 2435 (A, NY). Indonesia. Bogor, Wurjantoro 1 (A). New Caledonia. Noumea, Franc 1320 (A, NY). Tuamotu. Makatea, Wilder s.n. (Ny). 110 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII SCHIZOCARDIA BELIZENSIS: A SPECIES OF PURDIAEA (CYRILLACEAE) FROM CENTRAL AMERICA Joas L. THOMAS ABOUT THIRTY YEARS AGO specimens of an unusual and very distinctive tree were collected in British Honduras by William A. Schipp and sent to the Field Museum of Natural History. It soon became evident that this tree was unknown in the Central American flora, and the specimens were eventually distributed under a provisional name as a new genus and spe- cies, Schizocardia belizensis, without any indication of relationship. Later, Smith and Standley (1932) described the plant under this name as repre- senting a second genus in the Clethraceae, a relationship suggested by the collector. Recently, Dr. C. E. Wood suggested that I look at specimens of Schizo- cardia, indicating that the plant resembled members of the Cyrillaceae. Examination of these specimens revealed that this plant fits well within the logical generic limits of Purdiaea, in the Cyrillaceae, and that it should be transferred to that genus. Purdiaea belizensis (Smith & Standley) Thomas, comb. nov. Schizocardia belizensis Smith & Standley, Trop. Woods 32: 9. 1932. DISTRIBUTION. British Honduras. Dist. Stann Creek: Nineteen Mile, Stann Creek Valley, W. A. Schipp 965 (type, F; isotypes, A, GH); near Middlesex W. A. Schipp 443 (a, F); Silk Grass Creek Reserve, P. H. Gentle 2987 (a); Cockscomb Branch, D. Stevenson 4 (¥); Temax-Sarstoon, NV. S. Stevenson 164 (¥). Guatemala. Dept. El Petén: northeast of Poptun, F. B. Lamb 96 (Fr). Characteristics of this plant which place it in the genus Purdiaea and exclude it from the Clethraceae are as follows: a greatly exaggerated quin- cuncial type of sepal insertion, with the exterior sepals considerably larger than and completely inclosing the interior sepals; an ovary composed of 5 carpels and 5 locules, with each locule containing a single, pendulous ovule; a single style with an unbranched stigma; and a dry, indehiscent fruit which is usually devoid of seeds. The pollen grains, also, are almost identical in shape, size and structure to those of other species of Purdiaea. The Clethraceae, represented by the single genus Clethra, is characterized by equal sepals, a 3-carpellate, 3-locular ovary with numerous ovules in each locule, usually a 3-lobed stigma, and a loculicidal capsule with nu- merous small seeds. The Cyrillaceae and the Clethraceae have many characteristics in com- mon, however, and of the three genera in the Cyrillaceae, Purdiaea appears to be the one most closely related to the Clethraceae. The inclusion of a 1961] THOMAS, SCHIZOCARDIA BELIZENSIS 111 species of Purdiaea in the Clethraceae is a further indication of the close relationship between these two families. Purdiaea belizensis differs from other members of the genus in having distinctly laciniate sepals and axillary racemes. In all other known species of Purdiaea the sepals are entire, and in the specimens which I have ex- amined the inflorescence is a terminal raceme. However, in Cliftonia mono- phylla (Lam.) Britton ex Sarg., of the Cyrillaceae, both terminal and axil- lary racemes have been observed, and it is not improbable that this situa- tion also obtains in some species of Purdiaea. The present species is very distinct, and its relationships with other species of Purdiaea are obscure. It seems to be most closely related to the South American P. nutans Planch., particularly on the basis of its scarious sepals and leaves with uniformly pinnate venation. The latter character was previously thought to occur in Purdiaea nutans only (Thomas, 1960). In size and shape of the leaves and the sepals and in general aspect Pur- diaea belizensis also resembles P. nipensis Marie-Vict. & Leon. The latter species seems to be the nearest connecting link between P. cubensis (A. Rich.) Urb., of Pinar del Rio, and the other Cuban species of Purdiaea, all of which are restricted to Oriente Province. Thus, in some respects the present species is intermediate between two groups of species, with P. nutans and P. cubensis forming one group, and the species in Oriente Prov- ince forming the other (oc. cit., p. 45, fig. 23). The recognition of a species of Purdiaea in British Honduras and Guate- mala is particularly interesting in terms of the geographical distribution and relationships within the genus. Previously, the genus was known only from Cuba and northern South America, and the South American species seemed to be most closely related to the species from Pinar del Rio. This seemed backward, at first, in terms of the logical distribution pattern, but the recognition of Purdiaea belizensis in Honduras and Guatemala adds further evidence to substantiate this relationship. Thus, Purdiaea becomes another of an increasingly large number of plant groups distributed in northern South America, British Honduras and vicinity, and the Wes ndies REFERENCES SmitH, A. C., and P. C. STANDLEY. Schizocardia, a new genus of trees of the family Clethraceae. Trop. Woods 32: 8-11. 1932. Tuomas, J. L. A monographic study of the Cyrillaceae. Contr. Gray Herb. 186: 1-114. 1960. 112 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII STUDIES IN THE THEACEAE, XXXI A NEW SPECIES OF ADINANDRA FROM THE CELEBES CLARENCE E. KoBuskKI Adinandra eymae, spec. nov. Arbor vel frutex (habitus ignotus). Ramuli teretes, glabri, ut videtur rubro-brunnei; gemmis terminalibus conicis, glabris. Folia coriacea, gla- bra, obovata vel oblongo-obovata, 5.5—8.5 cm. longa, 2.5—3 cm. lata, supra brunnea (in siccis), subtus castanea, punctata, apice obtusa, rotundata vel emarginata, basi cuneata, margine integra, venis 8-10 paribus subtus obscuris, costa supra canaliculata, subtus elevata, petiolis ca. 3-5 mm longis. Flores solitarii; pedicellis teretibus, glabris, 1.5—2 cm. longis; bracteolis ‘i persistentibus, glabris oppositis vel suboppositis, inaequalibus, sepalis 5, imbricatis, inaequalibus, rotundatis, crassis, duobus exterioribus ca kabus interioribus ca. 9-10 mm. gis, ca. 9 mm. latis; petalis 5, le obovatis, inaequalibus, ae rotundatis, duobus exterioribus 8-9 mm. longis, 5-6 mm, latis, tribus interioribus 9-11 mm. longis, 7-9 mm. latis; staminibus ca. 35, 1-seriatis, 5—6 mm. longis, fila- mentis inaequalibus, 2—3 mm. longis , pubescentibus, ad basim corollae ad- natis, antheris glabris, aequalibus, ca. 3 mm. longis, connectivo pubescente; ovario globoso, glabro, 3-5 mm. longo, ca. 5 mm. diametro, 5-loculari, multiovulato, apice abrupte in stylo attenuato, stylo integro, glabro, ca. 3 mm. longo, stigmate capitato-punctiformi, indistincte 5-lobo. Fructus non visus. Celebes: Subdivision Enrekang, between Pintealon, Pokapindjang and Tina- bang. on ridge, alt. 2600-3000 m., P. J. Eyma 588 (a, type; L, isotype), June 16, 1937 (buds cherry-red). —_ The striking feature of this species is the complete glabrosity throughout the whole plant, except for the pubescence on the stamens, a feature which immediately associates it with the genus Adinandra. The closest relative is undoubtedly the widespread A. dumosa Jack, so far unreported from the Celebes, although recorded as distributed from the Malay Peninsula south into Indonesia, Sarawak and British North Borneo. In A. dumosa, the corolla lobes are much longer (up to 16 mm. long), connate 3—5 mm, at the base and bluntly acuminate at the apex. The stamens are nearly twice as long, measuring up to 12 mm. in length, and the style may measure up to 15 mm. long, easily five times the length of that found in A. eymae. 1961 | KOBUSKI, STUDIES IN THEACEAE, XXXI 113 In general appearance, the specimen cited above agrees very well with Eyma 613, a fruiting specimen of Ternstroemia collected the same day at the same locality. Eyma, the collector, undoubtedly thought that he had collected flowering and fruiting material of the same species (I did also), since he labeled them both Ternstroemia. However, on dissecting the flowers of Eyma 588, the type of this species, one finds that the ovary is five-loculate with many minute ovules in each cell, the placentation is parietal, and the stamens are hirsute. The terminal bud is conical. All these are generic characters used to separate Adinandra from Ternstroemia. It is a pleasure to dedicate this species to the memory of the collector, Pierre Joseph Eyma, a promising botanist whose life ended so tragically in a Japanese prisoner of war camp near Palembang, Sumatra, on July 25, 1945 : rs JOURNAL OF THE ARNOLD ARBORETUM VoL. XLII APRIL 1961 NUMBER 2 THE BOTANICAL RESULTS OF THE U.S. COMMISSION OF INQUIRY TO SANTO DOMINGO IN 1871 RicHARD A. HOWARD IN THE GRAY HERBARIUM there is a small collection of specimens which were prepared by Charles Wright, C. C. Parry, and H. Brummel during a difficult period of United States-Latin American relations nearly a century ago. The specimens bear little information beyond the name, and yet approximately forty of these are ‘‘new species” which were never published. Supporting this collection is a 130-page handwritten manuscript entitled “Flora Domingensis” and attributed to Asa Gray and Charles Wright. This, too, was never published, although with a revision of only a few pages, the manuscript, with Latin descriptions of the new species, geo- graphical localities, and dates would have been ready for the printer. One wonders why the manuscript prepared with such care was not published; why the specimens with incomplete labels and often conflicting numbers were so haphazardly distributed to herbaria; why two such competent collectors as Wright and Parry left so little record of their expedition to- gether; and who was the botanist H. Brummel, who is today unknown among the collectors of West Indian vegetation. Asa Gray, in a necrology of Charles Wright (Am. Jour. Sci. 31: 17. 1886) states, “The small collection made in this, his last distant botanizing, was not of much account.” Parry’s biographers refer to the Santo Domingo trip as a not very successful expedition, and Urban states (Symb. Ant. 3: 143. 1902) that the botanical results of this expedition made during an unfavorable part of the year and on a hasty trip were not outstanding. These comments seem inappropriate when applied to an expedition in an area which later yielded hundreds of new species to Eric Ekman; to a collection of nearly 700 numbers; and to a manuscript which might have been one of Charles Wright’s outstanding publications. From many sources, particularly the letters of Wright, Gray, Torrey, and Oliver in the historical files of the Gray Herbarium, it has been pos- sible to piece together notes and comments to supply this documentation of the botanical results of the U.S. Commission of Inquiry. I am grateful to Dr. Reed Rollins, director of the Gray Herbarium, for permission to publish this report on an historical document in the Gray Herbarium files; 116 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII to Mrs. Lazella Schwarten, librarian of the Gray Herbarium and the Arnold Arboretum, who has been most helpful in tracing missing items; to Dr. John Reeder, of Yale University, who located Wright, Parry and Brummel collections in the D. C. Eaton herbarium, and, finally, to Mrs. Katherine Hall and Mr. Theodore Dudley for their assistance in methodi- cally leafing through herbaria to locate the general collections cited. Many others in the United States and abroad have assisted and their help has been appreciated. This study developed in the course of work on a flora of the Lesser Antilles through a consideration of certain species which occur in that area. This floristic project is supported by grant G-4441 from the National Science Foundation and for this assistance grateful acknowledgement is made. HISTORY OF THE COMMISSION OF INQUIRY The island of Hispaniola, comprising the countries of Haiti and Santo Domingo (now the Dominican Republic), lies in an important position in the Caribbean. In the 19th century, at the time of the Commission of Inquiry, it was subjected to the interests of European countries and was torn by internal strife. There was a desire on the part of some groups in Hispaniola and of many parties in the United States for the annexation of the country to the United States. Negotiations for annexation had been carried on during the administration of President Andrew Johnson (1865— 1869) to bring about at least the acquisition of Samana Bay to serve as a strategic naval station —a guard post for the Mona Channel, the gate to the Caribbean sea and the Isthmus of Panama. Soon after the inaugura- tion of President Grant in 1869, the question of annexation was revived. The government of Santo Domingo sent an envoy to President Grant to solicit his consideration in the matter of Dominican affiliation with the United States. To ascertain the true state of affairs on the island Grant dispatched General Orville Babcock as a confidential agent to Santo Domingo. From Babcock’s report Grant became convinced, first, that what had seemed a fabulous account of agricultural and mineral resources was true: second, that it would be advantageous to obtain Santo Domingo because of its value as a mercantile and naval station; third, that the people of Santo Domingo truly were desirous of their country’s annexation to the United States: and finally, that by annexing Santo Domingo it would be possible to strengthen the Monroe Doctrine, since Santo Domingo thus would avoid domination by European powers. Working from these prem- ises, Grant set about to have passed a treaty for the annexation. Grant’s efforts met with bitter opposition from European traders who wished to monopolize the Dominican trade, from the aggressive Negro party in Hispaniola which hoped to gain control over the established Dominican government, as well as the entire island, and from some American groups whose spokesman, Charles Sumner, led the fight against annexation. When Babcock’s reports were challenged, Grant established 1961] HOWARD, COMMISSION TO SANTO DOMINGO 117 through an act of the U.S. Senate a new party of investigation. The Com- mission was headed by Dr. Samuel Howe, a noted philanthropist, Benjamin Wade, a former senator, and Andrew White, the president of Cornell University. These three, with their supporting assistants, secretaries and scientists, called themselves the U.S. Commission of Inquiry to Santo Domingo. They visited Santo Domingo during the latter part of January through early March, 1871, and submitted their report, which was pub- lished as the Executive Document 9, 42nd Congress, Ist Session, 1871. The goal of the Commission was to survey the natural resources; the nature of the health, education, and government of the people; the nature of foreign claims to the area; and the attitude of the people to annexation. These goals were met fully. Although the report for the most part sup- ported the idea of annexation, the proposal itself was defeated in the U.S. Senate. In due time even the lease on Samana Bay was abandoned, and some years later a naval base was established in Guantanamo Bay, Cuba, as an adequate substitute. ACTIVITIES OF THE COMMISSION The Commission of Inquiry consisted of twenty-two official representa- tives traveling at government expense and ten representatives of the press who were given transportation. The scientists, in the order and with the title given in the official report, were: Prof. W. P. Blake, geologist and mineralogist; Prof. C. C. Parry, botanist; Dr. W. Newcomb, naturalist; A. R. Marvine, assistant geologist and mineralogist; E. Waller, assistant mineralogist and chemist; J. S. Adam, assistant mineralogist and chemist; Prof. H. A. Ward, zoologist and paleontologist; C. Wright, botanist; and H. Brummel, botanist. There is no information on how these men were selected, but it is apparent from the titles throughout the report that Parry was considered to be the principal botanist and that Wright and Brummel were considered as aides. At the time of the expedition, Parry was not a professor but was officially botanist for the U.S. Department of Agricul- ture, having been appointed to that post in 1868. Brummel apparently was an employee of the Department of Agriculture, and nothing can be found on his professional career either prior to or after the trip. Charles Wright, well known as a botanical collector, had completed his last trip to Cuba in the summer of 1867 and was operating a farm in Wethersfield, Connecticut, during the summers and working as an assistant to Asa Gray, at the Gray Herbarium in Cambridge, Massachusetts, in the winter. Both Parry and Wright were well known to Gray, and it is possible that he sug- gested Wright for the trip. Wright was 60 years old and in poor health at the time of the expedition, and one wonders what inducement led him to join the Commission. The large number of reporters accompanying the Commission indicates the delicate nature of the investigations and the interest of the newspapers and the public in their findings. There is even a suggestion of intrigue in the official report of the trip. The assistant to the confidential secretary 118 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII of the Commission was discharged and sent back to New York, with the note being made in the report that this man had misrepresented himself and was, in fact, an assistant editor of an important New York newspaper. Although the reporters were supposed to be observers, the Commission delegated several of them to make special trips to gather information, with power to conduct interviews on behalf of the Commission. Letters from these reporters were published in many papers in the United States while the expedition was in the field. Feeling about the trip and the principle of annexation was high in the United States. John Torrey and Asa Gray, both to be involved in an aftermath of the Commission’s trip, expressed opinions against the principle of annexation of Santo Domingo. he Commission left New York on January 17, 1871, aboard the U.S. Steam Frigate “Tennessee” heading for Santo Domingo City. Supplies of coal ran low en route, and, on January 24, the “Tennessee” stopped in Samana Bay, where it was joined by the cutter ““Nantasket.” Coal had been reported along the shores of Samana Bay, but on investigation this proved to be a low-grade lignite. While at anchor in Samana Bay, the Commission and its scientists were divided into parties with different assignments. One party investigated the shoreline and the anchorage of the bay, while another went to the interior of the Samana peninsula. A third party was sent overland to Santo Domingo City with a message for the secretary of foreign affairs to announce the impending arrival of the “Tennessee” and the Commission in the capital. Still another group was sent overland to the capital to investigate the resources en route, and a last party was sent to Puerto Plata by the north coast. The “Tennessee” remained in Samana Bay until January 29th, when it departed for Santo Domingo City, arriving there on the 31st. On January 26th, the Commission resolved “that the botanists attached to the expedition be requested to examine and report to the Commission regarding the trees, plants, roots, and grains and their vegetable products of that part of the island adjacent to the Bay of Samana, especially with regard to such trees and plants as may be of commercial value, or in any way decidedly useful to man.” According to the published report, Parry was the botanist to make this trip in a party headed by the geologist Blake. Contrary to the report, Wright’s letters to Asa Gray (January 31, 1871; February 5, 1871) indicate that he also made this trip. The party left the anchorage off Samana on January 26th and proceeded to Punta Corozos, Punta Mangle, Punta Grigri, Los Robalos, Cabeza de Toro, and Santa Capuza, where they spent the night. On January 27th the party visited Punta Gorda, where they investigated the reported coal seam, and then proceeded to Canitas, the mouth of the Yuna river, and back to Cabeza de Toro for a night anchorage. The following day, January 28th, the Blake party visited Punta Corozos and returned to the “Tennessee” off Samana. Parry wrote the official report of this trip entitled “Report on the botanical features, agricultural products and timber growth of the peninsula of Samana.” In it he described the cultivated crops, grasses, fibers, fruits, palms and timber trees. He also stated, “The short interval allowed for 1961] HOWARD, COMMISSION TO SANTO DOMINGO 119 botanical examinations on the peninsula of Samana has been improved by a very fair local collection of plants numbering about two hundred species.” On January 27th, while Parry and Wright were in the field, the Com- mission resolved that Prof. Ward, the zoologist, make a trip along the coast in the direction of Cape Cabron, ‘‘not exceeding ten days,” and then go by land to Santo Domingo City “making a tour not exceeding a week,” and that ‘Professor Wright and Mr. Brummel be requested to accompany Professor Ward, in order to complete the examination of the vegetable products of the peninsula.’ According to the official report, this party departed before Parry and the Blake party returned. Again, Wright’s letters to Gray are in conflict with the published report, for Wright wrote that he sailed with the “Tennessee” and collected in the vicinity of Santo Domingo City while Brummel was traveling overland with Ward. The Ward expedition left the Samana anchorage and spent January 29th and 30th at Punta Cacao. They were prevented by high seas from rounding the tip of the Samana peninsula and landed instead at “Port Francais” (Jan. 31) and went inland to Las Galeras and the Bahia del Rincon, re- turning to ‘Port Francais’ the following day, February Ist. If this part of the report is true, it must have been a difficult and rapid trip and not a collecting expedition. On February 2nd, Ward and party, with Brummel along, crossed Samana Bay, stopping at Cocal San Lorenzo. There are two specimens in the Wright, Parry and Brummel series attributed to San Lorenzo and these must have been gathered by Brummel. The party was in Savana de la Mar on February 3rd and then proceeded up the Yuna river on February 4th and 5th to Almacen (February 6th). They traveled overland to San Francisco de Macoris and on to La Vega and Cotui, all on February 7th, and on to Cevicos (February 8th), through the mountains of eastern Hispaniola to Savana la Grande, and arrived at Santo Domingo City on February 9th. Meanwhile aboard the “Tennessee,” anchored in very rough waters off Santo Domingo City, on February 4th the Commission asked Blake to make geological investigations about the capital and to organize an ex- pedition to cross the central range of mountains to Puerto Plata, in a period not to exceed two weeks. Wright was to accompany him and to report on the vegetation of the interior. In the same resolution Parry was directed to examine the area around Santo Domingo City and to report. Wright’s letters reveal that Parry was still suffering from a cold and that Wright took extended walks around Santo Domingo City; his letters describe the vegetation in considerable detail. On February 8th the report reveals that Blake accepted the directive to cross the island but that the time allotted was extended and the rendezvous time in Puerto Plata was set for March Ist. On February 9th the botanists attached to the Commission were literally going in all directions. Parry and members of the Commission were on a leisurely trip by boat fifteen miles up the Rio Ozama and then another 15 miles up a branch to the westward before returning in the late afternoon. Wright’s letters and notes indicate that his party headed westward to the 120 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII Rio Haina to begin a trek across the island. Brummel and party returned to Santo Domingo City on that day and apparently remained aboard the “Tennessee” to write up their report entitled, “Notes on the agricultural resources of Samana Peninsula and the Vega Real.” There is a suggestion that Brummel did not collect botanical specimens on his trip with the Ward party, and his report is in contrast to others published. It is written in the most general terms and contains only one botanical scientific name. Brummel’s name does not appear again in the official report of the activities of the Commission. While the Blake party, including Wright, left Santo Domingo City to make their way to Puerto Plata, the Commission, with Parry and Brum- mel, remained in Santo Domingo City and concluded its work. The cutter “Nantasket” left the capital anchorage on February 14th for Puerto Plata, where some members of the Commission were to proceed inland to La Vega, meet the Blake-Wright party in Puerto Plata on March Ist and then rendezvous with the “Tennessee” at Port au Prince on March 7th. Other members of the Commission, with Parry and apparently Brummel on board, sailed with the “Tennessee” on February 14th, arriving at Ocoa Bay on the 23rd. Members of the Commission went inland to the town of Azua, and the “Wright, Parry and Brummel” specimens from Azua must have been collected by Parry or possibly Parry and Brummel during that visit. The “Tennessee” left Ocoa Bay on February 28th and arrived at Port au Prince on March 9th. WRIGHT'S TRIP FROM SANTO DOMINGO CITY TO PUERTO PLATA Wright’s manuscript of the “Flora Domingensis” carries many refer- ences to plants collected at “interior savannas,” ‘‘wet vallies of the interior,” “pine woods of the interior,’ most with dates of collections and a few plants with specific seperaphical locations. The official report, however, gives barometric readings with dates and places. It is possible to coordi- nate these two sets of data to derive the following itinerary of the crossing of Hispaniola. February 9, Santo ea City, Rio Haina, Santa Rosa, Arroyo Lebrun. February 10-11, Madri February 12, Arroyo te Ces La Puerta. February 13, Loma Laguneta, El Aguacate. February 14, Rio Maimon, Hato del Banao. February 15, Arroyo Yuma, Rio Yuna, Rio Jima. February 18, Santo Cerro. February 18-19, Moca. February 21-25, Santiago (February 23. Rio Yaqui del Norte). February 26-28, Arrenquillo River, El Limon, Loma de Bajabonico. March 1-3, Puerto Plata 1961] HOWARD, COMMISSION TO SANTO DOMINGO 121 Even with modern roads, this trip of about 105 airline miles would be an arduous one by foot or by horseback. In 1871, with few established paths, these deep with the mud remaining from the rainy season, it was an heroic undertaking in the time allotted for the 60-year-old Wright. A reporter who had come by boat to Puerto Plata to visit La Vega com- mented, “The only road from Puerto Plata to the interior is a bridle path to Santiago. In a direct line Santiago is only 18 miles, but by path it is 60 miles.” He comments that the rainy season was scarcely over and that the paths were muddy and slippery. On his arrival in Puerto Plata, Wright received instructions to report on his trip immediately. His pique clearly shows in the first paragraph of his report. “The journey was made at a season of the year when fewest plants are in flower and not many in fruit. Called to join this Commission almost at a moment’s warning, no time was afforded to obtain books suitable or sufficient to determine the plants found there; and the only books brought were left on the ship from a desire to reduce the luggage to the smallest possible dimensions. Moreover, no time could be saved to examine the plants in their fresh state, and to study them carefully, without neglecting the specimens gathered, by the careful preservation of which they might be more critically examined at a future time. Even the inquiries for the vernacular names of trees met with re- ceived unsatisfactory answers. The muleteers and guide seemed to know few of the trees of the mountain regions. Even when inquiries were addressed to the inhabitants, answers, whether thoughtless or intentional, so absurdly wrong were given that I despaired of gaining much reliable information which would serve to connect with certainty the vernacular names of the plants with their scientific ones. A more particular report must consequently be postponed till a critical examination of the specimens can be made.” This proved to be a forecast of trouble to come The rendezvous at Port au Prince between the ‘‘Nantasket’”’ and the “Tennessee” occurred on March 9th. This date is also recorded as the departure date for Kingston, Jamaica, where the ‘‘Tennessee” arrived March 11th. The ‘‘Nantasket’’ apparently returned to Puerto Rico or to Saint Thomas. While the ‘‘Tennessee” re-coaled in Kingston harbor, Wright had the opportunity of visiting the government Forestry Station at Cinchona in the Blue Mountains. He refers in later letters to his knowledge of the growth of quinine trees in Jamaica. It was in this area that Wright must have collected an unnumbered specimen of Vaccinium meridionale (us) which has bothered monographers. The species is not known from Hispaniola but does occur in the vicinity of Cinchona. From Jamaica the ‘‘Tennessee’’ proceeded to Charleston, South Caro- lina, arriving there March 26th. The Commission held meetings aboard the ship en route and on March 19th, while at sea between Cape San Antonio de Cuba and Key West, they resolved ‘“‘that the collections made by the scientific gentlemen who have accompanied this expedition be inventoried by them and deposited 1 in the Smithsonian Institution, subject to the disposition of Congress.” That all was not harmonious in the acceptance of this resolution is attested by the comment in the official 122 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLII report, “During the discussion of said resolution the Commission took a recess until 7 o’clock p.m.” Whether Parry wanted the specimens for the herbarium in his charge at the Department of Agriculture or whether Wright felt his collections, or all of them, should come to the Gray Herbarium for his study can not be resolved. In any case, the issue was not settled amicably, as is obvious from a note in the official report regard- ing a meeting of the Commission at the Arlington Hotel, Washington, D.C., on April 3, 1871. It is noted there that the Commission “received the following letter dated April 3 from Joseph Henry, Secretary of the Smithsonian Institution: ‘Sir: On reply to the inquiry you make as to the disposition of specimens collected at the expense or under the auspices of the United States Government, I have the honor to inform you that, by the law of Congress organizing the Smithsonian Institution, it is the official curator of all collections of natural history, geology, etc. belonging to the United States, and that in accordance with this enactment all the speci- mens collected by the Wilkes, Gilliss, Rodgers, Perry and other naval expeditions, and the Pacific Railroad Boundary and geological surveys, are now in its custody. “ laid tenuior, . (us) 334. Lobelia domingensis, 6 (K, US) 300. “Sperm é sp.,’ 335. Ammania latifolia, 21 (us) 301. *Zobelia panne 19 is us) 336. Cordia sulcata, 19 (GH, US) 302. Lobelia cliffortiana, 20 (us) 337. Lantana camara var. mista,” 19 303. “‘Conradia reticulata,” 4 338. “Heliotropium parviflorum,” 20 304. “Ardisia crenulata,” 19 339. blan 305. Dipholis salicifolia, 19 (us) 340. “Scutellaria havanensis,” 19 306. “Jacquinia ruscifolia,” 24 341. Ocimum gratissimum (Us 307. Citharexylum fruticosum (us) 342. Leonotis nepetaefolia, 19 (us) 308. Jacquinia Eggersii (us) 343. Hyptis pectinata, 19 (us) 308. Citharexylum fruticosum (Us) 344. Scutellaria havanensis, 15 (us) s manuscript Wright recognized this collection as a “new species” which he called Jsodorea pungens citing in synonymy Ernodea pungens Lam. Robinson had access to the Wright manuscript and the ast te in the Gray Herbarium, althou igh ve W a TIsodorea pungens (Lam.) Robinson (Proc. Amer. Acad. Arts Sci. 45: 401. 1910). Wright’s specimen was collected on cliffs by the river, near Santo Domingo City, on February 3. “This material is inadequate for accurate determination but appears to be Psycho- tria pinularis. “The Wright manuscript ascribes a specific name attributed to Jacquin to this collection. I can find no reference to the publication of this epithet. * Wright’s manuscript cites “Solanum verbascifolium” for this number; however Francey in his monograph of Cestrum (Candollea 286. 1936) afta a soectinien fain the Chicago Natural History Museum as given abov * This collection, as well as 357, was cited by ‘Moldenke (An alphabetized list of citations 3: 1146, 1949) as entiaal: The herbarium source of the material is not stated 1961] 345. 346. 347. 348. 349. 350. oor: 02. $50. 354. So. 356. 351. 335; 309: 360. 361. 362. 363. 364. 365. 366. . Justicia sessilis, 20 (us) . Ruellia domingensis, 21 (us) . Justicia pectoralis, 19 (GH, US) . Cordia serrata, 21 (us . Jacquemontia Siz. ods: 374. ato. 376. Oy fs 378. HOWARD, COMMISSION TO SANTO DOMINGO blan ae suaveolens, 19 (us) Hyptis capitata, 19 (us) Hyptis lantanifolia, 9 (us) Ocimum gratissimum, 19 (US) “Salvia occidentalis” Hyptis americana, 19 (us) Hyptis pectinata, 19 (us) Stachytarpheta jamaicensis, 19 H, US Priva lappulacea, 20 (us) Petitia domingensis, 7 (US) Clerodendron aculeatum (Us) “Cornutia pyramidata,” 15 Acalypha iene (us) Acalypha se , US) Blechum rey Dicliptera oe "30 (us) Ruellia tuberosa (us) Teliostachya alopecuroides, 9 (u S Barleriola solanifolia, 2 (us) *Ruellia coccinea, 19 (K, US) Justicia pectoralis (us) pentantha, 20 US Jacquemontia nodiflora (us) Merremia umbellata, 20 (us) “Tpomoea umbellata,” 20 Rivea corymbosa, 20 (us) Merremia quinquefolia (us) Jacquemontia nodiflora (us) Ipomoea eriosperma (us) k 379. blan 380. 381. 382. 383. 384. 38D: 386. “Tpomoea campanulata,” 19 “Tpomoea cathartica” “Tpomoea fastigiata” “Tpomoea cathartica” Ipomoea acuminata (us) Ipomoea acuminata (Us) “Tpomoea fastigata” 387. 388. 389. 390. 391. 592, 139 *Ipomoea setifera,” 19 (GH, K) Ipomoea tiliacea (us Ipomoea tiliacea (US) Rivea campanulata, 19 (us) Cuscuta americana, 20 (us) “Limnanthemum humboldtia- um,” 10 . Schultesia heterophylla, 6 (us) . Micranthemum nuttallii, 18 (us) . *“Micranthemum sp. nov.,” (GH, K . “Utricularia pusilla” . a acl citrifolia, 6 . ek aiemontian neritfolia” . Psychotria undata (GH, US) . “Ipomoea fastigiata”’ . Rhabdadenia berterii, 21 (us) . “Tabernaemontana neritfolia,” . Distictis lactiflora («) S) . Rhabdadenia paludosa, 19 (us) . Asclepias nivea (us . “Echites umbellata” . Echites umbellata (us) . Urechites lutea, 15 (us) . Rauvolfia nitida, 20 (us) . Anagadenia berterii, 21 (us) . “Echites repens,” 19 . Exostema longiflorum (us) . “Metastelma leptocladon,” 21 . “Piper dilatatum,” 21 . Metastelma Picardae (us) ete amygdalzifolia, 7 (us cone pallidum,” 20 . Nectandra antillana, 9 (us) . “Phoebe montana,” 21 . blank . Linociera domingensis, 21 (us) . “Nectandra willdenoviana”’ 420. 421. Ocotea leucoxylon, 19 (us) “Nectandra willdenoviana” °° This collection bears an unpublished specific name ee to the peninsula of Samana where the specimen was collected January 27, 1871, Neither this specimen nor any of the other Convolvulaceae here have been cited by House or other monog- raphers of the family. pein had prepared a complete description for publication but was unable _ 21 supply the date or location of the collection. The genu us requir es a mon graphic eerie before the cited specimens can be determined. 140 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII 422. Licaria triandra, 10 (us) 453. icp romia glabella, 19 (u 423. Trichostigma octandrum, 21 454. Dendrophthora peat EN US, US) 424. “Pisonia aculeata,” 9 455. Dendrophthora marmeladensis, 425. Pisonia aculeata, 25 (us) 426. Rapanea guianensis (us) 456. “Dendrophthora wrightii,” 21 427. Rapanea ferruginea, 9 (GH, US) 457. Dendrophthora flagelliformis, 15 428. “Chamissoa altissima,” 20 (us 429. Iresine angustifolia, 20 (us) 458. Phoradendron anceps, 15 (us) 430. Alternanthera geniculata (us) 459. *Phoradendron cerinocarpum 431. “Rousselia lappulacea,” 19 459A. Phoradendron chrysocarpum ™ 432. Pilea repens, 19 (us) US) 433. Fleurya aestuans (us) 460. Phoradendron hexastichum, 21 434. Urera baccifera, 19 (us) (us) 435. Ficus trigonata, 8 (us) 461. *Phoradendron mucronatum 436. Ficus citrifolia, 8 (us) GH, K, US 437. Ficus pe naie 8 (US) 462. si lal mucronatum, 9 438. Ficus trigona (us) 439. “Marcgravia fee aie 463. art ae flagelliformis, 15 440. Piper aduncum, 19 (us) (us) 441. Piper aduncum, 19 (us) 464. Phoradendron dichotomum, 10 442. Piper dilatatum (us) US) 443. Piper jacquemontanum,” 19 465. Phoradendron piperoides," 25 (GH, US) (us) 444. Piper marginatum, 21 (Mo, Us) 466. Phoradendron dichotomum, 24 445. Piper hispaniolae ™ (us) (US 446. Piper parryanum," 19 (us) 467. Phoradendron antillarum, 21 447. Piper dilatatum (us) (us) 448. Piper parryanum, 19 (GH, MO, 468. Dendropemon purpureus (us) US) 468A. Dendropemon uniflorus (us) 449, a distachya, 19 (us) 469. i a alatus (GH, K, 450. “Peperomia caulibarbis 451. Peperomia serpens, 19 (us) 470. ti. donee alatus, 2 (GH, 452. Peperomia serpens,” 19 K, US. 7 a collection was cited by Trelease (Repert. Sp. Nov. 23: 307. 1927) as given abov 2 Trelease selected the specimen in the U.S. National Herbarium as the holotype of this species which he described as new (loc. cit. 309). Wright’s manuscript gives one collection number but two locations for the material: Puerto Plata, February 26, and err January 27. I have been unable to determine the Griginal locality of the holoty = Trelease cited Wright, Parry & Brummel 446 and 448 in oe this species as new (loc, cit. 311). Number 446 was selected as the holotype in the U.S. National Herbarium. Both collections were made in Samana on January 28, 1871. ” Trelease sae cit. 323) cited this collection number w hout indicating the loca- tion of the spec © Wright, Po - Brummel 459 is cited by Trelease (The Genus Phoradendron. Univ. Ill. Bull. 13: 139. 1916) as the holotype of Phoradendron cerinocarpum. The specific epithet is attributed to Wright and occurs in the manuscript. Unfortunately, mixed i ported to have been collected between Moca and Santiago, on February 21, and on the road from Santo Domingo City to Puerto Plata, on ina 13, a date when — was in the vicinity of Loma Laguneta and El Aguaca ited under this name by Trelease (loc. cit. 147). 1961] HOWARD, COMMISSION TO SANTO DOMINGO 141 471. Suriana maritima, 13 (us) 502. */lex aff. macfadyenii,” 21 (us) 472. Coccoloba swartzti, 21 (GH, kK, 503. Croton origanifolius (GH, US) US) 504. “Croton flavens,” 2 473. Coccoloba swartzii, 21 (GH, US) 505. Croton pallidus (GH) 474. Se fuertesii, 21 (GH, 506. Croton discolor, 24 See US) 507. Croton linearis, 13 (US) 475. KE sccolbe nodosa, 21 (GH, kK, 508. Euphorbia prostrata (us) NY, US) 509. Euphorbia buxifolia, 13 (us) 476. *Coccoloba incrassata, 2 (GH, 510. “Euphorbia heterophylla, et fo) K, US 511. Euphorbia hirta, 19 (us 477. *Coccoloba diversifolia (GH, 512. Euphorbia heterophylla, 15 (us) 513. Solanum antillarum, 21 (us) 478. Exostema caribaeum, 2 (US) 514. Peperomia obtusifolia,” 9 (us) 479. Pisonia aculeata (us) 515. Erithallis fruticosa 480. ee achryanthoides, 19 516. Stigmaphyllon lingudatu (us) 517. Exothea paniculat (us) 481. te aestuans, 21 (us) 518. *Cuphea ae = ve 482. Pilea repens, 17 (us) 519. Cuphea parsonsia, 16 (us) 483. Rousselia humilis, 19 (us) 520. “Miconia fothergilla,” 9 484. Tabebuia berteri, 6 (us) 521. Miconia macrophylla, 1 (us) 485. Tabebuia berteri (us) 522. Tetrazygia bicolor, 6 (us) 486. Forsteronia corymbosa (us) 523. Heterotrichum umbellatum, 21 487. Rhabdadenia ae (us) 1S 488. “Echites repens,’ 524. Ossadea acuminata, 6 489. *“Guettarda sp. nov.” * 525. Miconia elata, 21 (us) 490. Jatropha gossypifolia (us) 526. Miconia rubiginosa, 9 (GH, US) 491. Adelia ricinella, 2 (us) 527. Dieffenbachia seguine, 19 (us) 492. Acalypha setosa, 19 (GH) 528. Costus cylindricus, 19 (GH, US) 493. Acalypha glechomifolia (us) 529. Pistia zpabiinie 7 (us 494. Clerodendron aculeatum (us) 530. Hypoxis ere 495. Argithamnia candicans, 15 (us) Sol: ee ee 6 (us) 496. “Phyllanthus nobilis,” 25 531A. Heteranthera spicata (us) 497. “Drypetes alba,” 19 532. Cipura palludosa, 6 (us) 497. Psychotria revoluta, 19 (kK, US) 533. Anthurium scandens, 4 (us) 498. Drypetes alba (us) 534. ee dice (us) 499. “Drypetes sp. nov.,” ™ 15 535. “Anthur * 19 500. Drypetes alba, 15 (us) 536. Anthurin iun igen * 19 501. Drypetes alba, 21 (us) 537. Cissampelos pareira (us) *“No material has been seen of this number. Wright described a new species in his manuscript citing 237 and 489. The former has been referred to Gu esha preneloupii. n * Wright compared his material with “Drypetes incurva Mull.” a Giacediee it was perhaps a distinct species. The specimen was collected at Puerto Plata on February . No material has been located. * Wright’s notes indicate that he collected this specimen on February 15 when he trav velled i in the vicinity of the Arroyo Yuma, the Rio Yuna and the Rio Jima 'Trelease had annotated this sheet with a new specific name honoring Brummel This specimen has not been cited by monographers of the Lythraceae. The published specific ndme referred to Santo Domingo. The specimen was Pilkeeed * stg eed annas of the interior at Madrigal” on February 11, 1871 “ Wright uses in his manuscript a name attributed to Schott which appears to be ame No description is given and no specimens have been located. 142 JOURNAL OF THE ARNOLD ARBORETUM 538. Dioscorea polygonoides, 19 (us) ) 539. Smilax populnea, 15 (us 540. Smilax domingensis (us) 541. Callisia monandra, 20 (us 542. Tradescantia geniculata (us) gans, 19 (GH, US) 543. Commelina ele 544. Commelina elegans (us) 545. Commelina diffusa, 20 (us) 546. Callisia monandra, 21 (us) 547. Thrinax parviflora, 20 (us) 548. blank 549. Coccothrinax sp.,“ 14 (us) 550. Copernicia sp.,“ 2 (us) 551. Pleurothallis gelida, 5 (us) 552. Epidendrum rigidum, 19 (us) 553. “Aeranthus sp.,” 9 557. “Spiranthes elata,” 2 557A. Pelexia sp. (us) 558. “Pelexia setacea” 559. Malaxis spicata, 19 (us 560. Syringodium fliforme, . (us) 561. “Epidendrum bifidum 562. Ponthieva glandulosa, 6 on 563. Epidendrum wrightii, 21 (us) 564. “Epidendrum broughtonioides,” 565. “‘Polystachya luteola” 566. Oncidium variegatum, 19 (us) 566A. “Epidendrum bifidum 567. “Cranichis muscosa” 568. “Spiranthes apiculata,” 6 569. Ponthieva ekmanii, 5 (us) 570. “Epidendru US 572. “Aeranthus §sp.,’ 19 573. Epidendrum peti 19 (us) US) 574. Bletia patula, 5 (Gu 574A. Bletia patula (vu s) 575. Xyris caroliniana, 6 (us) 576. Cyperus peruviana, 9 (US 577. Cyperus (us 578. “Rhynchospora florida,” 21 579. Cyperus haspan, 6 (us) 580. “Scirpus exiguus,” 6 554. Tetramicra parviflora (GH, US) blank an 556. Ponthieva glandulosa, 5 (us) m nocturnum,” 22 571. Epidendrum strobuliferum, densicaespitosus, 581. 582. 583. 584. 585. [VOL. XLII Rhynchospora — (us) *Fleocharis sp. no Rhynchospora pear (us) “Fimbristylis brizoides,” 6 “Rhynchospora barbata”’ 586. blank 587. 588. 589. 590. 591, 592. 593. 594. 595. 596. 597. 598. 599. 600. 601. 602. 603. 604. 605. 606. 607. 608. 609. 610. 611. 612. 613. 614. 615. 616. 617. 618. . Phragmites communis (us) . Paspalum virgatum, 19 (us) . Cenchrus brownii (us) . Digitaria villosa (us) . Reynaudia filiformis, 6 (us) . “Cenchrus viridis,” 19 . Panicum stenodes, 6 (us) 5. Andropogon virgatus, 6 (us) 628. Eleocharis capitata, 9 (us) “Rhynchospora vahliana”’ “Rhynchospora glauca,” 19 Fimbristylis dichotoma, 6 (vs) Rhynchospora barbata (us) Eleocharis retroflexa, 19 (us) Cyperus nanus, 15 (us) Scleria secans, 19 (us) Cyperus elegans, 15 (us) Eleocharis geniculata (us) “Eleocharis sere "19 Fuirena umbellata, 19 (us) Rhynchospora pusilla, 19 (us) Eleocharis chaetaria, 6 (us “Dichromena pusilla,” 19 “Kyllingia brevifolia,” 19 Mayaca fluviatilis, 6 (us) “Carex scabrella,” 15 Pharus latifolia (us) “Pharus latifolius,” 19 “Paspalum compressum,” 19 Chloris sagraeana, 19 (us) Eragrostis ciliaris, 19 (us) Chloris inflata (us) “Orthopogon labiaceus,” 19 Lasiacis patentiflora, a (us) Lasiacis sloanei, 19 (us) *Hyparrhenia hirta, 12 (us) ) . Arundinella confinus, y (us) . Panicum diffusum, 21 (us) “Tricholaena insularis” * The available material of these palms is inadequate for accurate determination. =A specimen collected in “springy places in savannas of the interior, Feb. 10,” according to Wright’s manuscript. No specimens have been located. 1961 | 629. 630. 630. 631. a! nn _ co bo oR oe) Re Ree ee PWNnNPr DOAWANAIANAP WN HE HOWARD, COMMISSION TO SANTO DOMINGO . Aspidium scolopendrioides (vu) . Tectaria heracleifolia (us . Tectaria martinicensis (us, YU) . Pityrogramma sulphurea (vs, . Trichomanes krausii (us, Yu) . Trichomanes lineolatum (us) 143 *Andropogon saccharoides, 6 GH, “Panicum dichotomum,” 21 Digitaria horizontalis, 19 (us) 21. Trichomanes krausii (us) Andropogon bicornis (us) 632. Tricholaena insularis (us) Paspalum paniculatum (us) 633. *Andropogon saccharoides, 6 634. (us) Series II “Lycopodium mexicanum” Ananthacorus angustifolia (us) 22 Adiantum deltoideum (US, YU) 23 Adiantum cristatum (US, YU) 24. Adiantum obliquum (us, YU) Lindsaea portoricensis (us) 25 Adiantum fragile (us, YU) 26 Dryopteris serra (US, YU) 27 . Pteris longifolia (us, YU) . Pityrogramma tartarea (US, YU) 28 . “Gleichenia pubescens” 29 . Dryopteris tetragona (yu) . Polypodium salicifolium (us) 30 . Polypodium astrolepis (us) 31 Series III Aristolochia aff. pentandra™ (kK) Bidens leucantha (us) Guazuma ulmifolia (us) Hura crepitans, 19 (us) Hymenaea courbaril (us) Ipomoea batatas, 19 (us) Ipomoea pes-caprae (US) Series IV Ananas sativus, 19 Annona cherimola Annona muricata *° Determination by Oliver, 1871 Probably collected near Cinchona, Jamaica. . Anemia hirsuta (vv) . Anemia adiantifolia (vu) Cheilanthes microphylla (vs, YU) . Asplenium serratum (us, YU) . Alsophila aquilina (us) . Cyclopeltis semicordata (US, YU) . Asplenium dentatum (us) . Pityrogramma calomelaena (vs, YU . Asplenium cristatum (us) . Dryopteris dentata (us) . Cyathea arborea (us, YU) 33. Odontosoria aculeata (US, YU) . Alsophila aquilina (us . Selaginella plumosa (us, YU) Lycopodium cernuum (us, YU . “Lycopodium complanatum” Tresine celosia (us) Juniperus gracilis (us) Mimosa pudica (us) Nopalea cochenillifera, 10 (GH) Pinus occidentalis (us Vaccinium meridionale ™ (us) Zamia pumila (us) Annona palustris Sebastiana corniculata, 9 144 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII COMPARATIVE ANATOMY OF THE LEAF-BEARING CACTACEAE, II STRUCTURE AND DISTRIBUTION OF SCLERENCHYMA IN THE PHLOEM OF PERESKIA, PERESKIOPSIS AND QUIABENTIA I. W. Batrey ! VARIOUS FORMS OF SCLERENCHYMA, as commonly defined (e.g., Eames and MacDaniels,” Esau*), occur in the leaf-bearing genera Pereskia, Pereskiopsis and Quiabentia: (1) libriform fibers are present in the sec- ondary xylem of stems and roots, (2) strands of aggregated slender fibers in the primary phloem of stems, (3) concentric layers of sclereids in periderm of stems and roots, and (4) highly diversified forms of sclereids in the secondary phloem of Pereskia. It is significant in this connection that, under normal circumstances, sclerenchyma does not occur in the leaves and pith of the three genera. The libriform fibers of the xylem and the sclereids of the periderm will be dealt with in subsequent papers of this series. In the following pages, I shall confine myself to a discussion of sclerenchymatous elements that are formed in the phloem of the three genera. The leaf-bearing Cactaceae have mixtures of very dense and exceed- ingly soft tissues and are difficult to section. After experimenting with various methods of embedding following preliminary softening with such reagents as hydrofluoric acid, I have in general obtained the most useful preparations by sectioning stems and roots of living plants or specimens preserved in formalin-acetic-alcohol fixative without the use of preliminary treatments. Although the transverse and longitudinal sections thus ob- tained with an adequately sharpened knife and sliding microtome are relatively thick, they are suitable for critical visual examination. Further- more, they have the advantage, not only of being obtained simply and rapidly, but also of retaining starch, crystals, and other cell inclusions which may be much modified or eliminated during softening and em- bedding. It must be admitted, however, that they are not of the best quality for photographic illustration. FIBERS OF THE PRIMARY PHLOEM Strands of aggregated primary phloem fibers occur in the stems of Pereskiopsis and Quiabentia, as well as in those of Pereskia (Fics. 1a, ‘This investigation was hig ig by a grant from the National Science Foundation. * Eames, A. and I eae An introduction to plant anatomy. Ed. 2 McGraw- Hill Book Co., N. oe “Esau, K. Plant Anatomy. eee Wiley and Sons, N. Y., 1953. 1961] BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, II 145 7a). Particularly in the more succulent parts of Quiabentia and some species of Pereskiopsis, the number of fibers in each strand tends to be reduced, Furthermore, the maturation of fibers in the former genera tends to be precocious, whereas in Pereskia the final stages of maturation may at times be delayed until more or less extensive cambial activity has occurred. The individual fibers are slender, elongated elements of the general external form illustrated in Fic. 12. Their diameter commonly varies from 20 to 50 microns and their length from a few hundred microns to more than 2000. The marked variability in length occurs, not only in strands from different parts of a plant, but also in adjacent fibers of a single strand. Fibers having normal, lignified secondary walls may be internally septate or nonseptate, but they do not store starch as the cor- responding septate and nonseptate libriform fibers of the secondary xylem so commonly do. Those of QOuiabentia and Pereskiopsis tend to form relatively thick secondary walls and contain few if any internal septa at maturity (Fic. 13). On the contrary, in the case of Pereskia, many of the fibers have thinner secondary walls and are internally septate (Fic. 14). It is significant that part or all of such septate fibers in a strand may ultimately form internally a chain of sclereids having lignified, multi- layered walls (Fic. 15). Occasionally, the protoplast of one of the sclereids may divide leading to the formation of two smaller sclereids included within a larger one. This unusual phenomenon of cells-within- cells, which is of infrequent and sporadic occurrence in septate libriform fibers of secondary xylem (Fic. 6) is of considerable interest from physio- logical and developmental points of view and merits detailed investigation. For example, the two or three successive waves of lignified wall formation indicate that lignification is not necessarily an immediate precursor of loss of potentialities or degeneration of the protoplast as has sometimes been assumed. The orientation of cellulosic microfibrils in the broad central or Ss layer of the secondary wall of septate and nonseptate fibers, of both the primary phloem and the secondary xylem, varies from approximate parallelism to the long axis of the cell to helices of varying pitch. Clues to such varia- tions may be obtained by studying the orientation of the slit-like pits, by the distribution of slip-planes, and by examining thin transverse, longi- tudinal and diagonal sections in polarized light. In Fic. 6, a very thin, perfectly transverse section in polarized light between crossed nicols, the tenuous outer or S; layer of the four adjacent fibers is strongly birefringent. The broad central or Ss is dark, whereas in longitudinal section it is birefringent. This indicates that the orientation of cellulosic microfibrils in the S. layer of the secondary wall is approximately longitudinal, as does the orientation of slit-like pits and the distribution of slip-planes visible in longitudinal sections. Internal to the normal S; and So» layers of the septate fibers in Fic. 6 are transverse sections of the multilayered walls of included sclereids, the cellulosic layers of which are alternatingly birefringent and dark in polarized light. In such transverse sections as 146 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Fic. 6, it is not possible to determine whether the original secondary wall of the fibers has a tenuous S3 layer or not. It should be emphasized in this connection that these cells-within-cells provide interesting material for analysis by electron microscopy. Occasionally a cell of prosenchymatous form on the margins of a strand of primary phloem fibers divides precociously before secondary wall formation (Fic. 16). Subsequently such cells may form a chain of sclereids (Fic. 17) comparable developmentally to wood parenchyma strands or the strands of short tracheids that occur in certain of the Pinaceae. (See glossary of terms used in describing woods.*) FORMS OF SCLERENCHYMA IN THE SECONDARY PHLOEM OF PERESKIA Under normal conditions of growth, schlerenchyma does not occur in the cortex and secondary phloem of stems and roots of species of Pereskiopsis and Quiabentia that I have investigated. In striking con- trast to absence in these genera is the occurrence (particularly in larger stems and roots) of more or less numerous and diversified forms of sclereids in the secondary phloem of all species of Pereskia (including Rhodocactus) of which I have examined adequate material. The form and the distribution of the sclereids is consistently different in three distinct categories of putative species of Pereskia. 1. Species with Diffusely Distributed Sclereids in the Secondary Phloem In Pereskia aculeata Mill., P. pititache Karw., P. conzatti Britt. & Rose, P. autumnalis (Eichlam) Rose, P. nicoyana Web., P. weberiana K. Schum., and P. diaz-romeroana Card., the sclereids in the secondary phloem of large stems and roots are abundant and diffusely distributed (Fics. 1-3). In transverse sections, these cells, which vary in diameter from less than 18 to more than 100 microns, have conspicuously multi- layered secondary walls which frequently occlude the lumen of the cell. As indicated in Fic. 5, the numerous layers of the secondary wall (which may exceed 50 in the case of the largest cells) are alternatingly bire- fringent and dark or feebly birefringent in polarized light. (Compare with the internal sclereids shown in Fic. 6. As seen in longitudinal sections (Fic. 4), these cells have a much elongated prosenchymatous form, the most slender ones having a length of a few hundred microns, whereas the broadest ones may attain at times a length of more than two thousand microns. In many of these cells the lumen tends to be occluded toward the ends of the cells and to be much reduced in the central part (Fic. 18). However, some of the cells having thinner multilayered walls become septate and may subsequently form a “International Association of Wood Anatomists. Glossary of terms used in describ- ing woods. Tropical Woods 36: 1933 1961] BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, II 147 chain of shorter sclereids internally. The pits of the secondary wall tend to be restricted to the broadest central part of the cell (Fic. 18) and to be more or less irregularly distributed. It is significant in this connection that many of the pits are of the so-called ramiform type characteristic of many sclereids (Fic. 19). From a developmental point of view, such pits are actually of a coalescing, rather than of a ramifying type, for, the orifices of two or move simple pits in the first-formed layers of the sec- ondary wall unite to form a single opening in the inner or last-formed layers of the wall. A limited number of sclereids may occur at times in the expanded cortex of much enlarged stems (Fic. 1b). These vary from cubical to various bizarre forms. However, where they have a conspicuously longer axis, it is oriented at various angles and not consistently parallel to the long axis of the stem. 2. Species with Sclereids Aggregated in Longitudinally Oriented Clusters In large older stems of Pereskia sacharosa Griseb., P. grandifolia Haw., P. bleo DC., P. corrugata Cutak, and P. tampicana Web. aggregations of sclereids commonly occur in the secondary phloem (Fics. 7, 8). These aggregations, as seen in longitudinal sections of the stem, are of elongated form, frequently of more or less fusiform outline (Fic. 9), and are oriented parallel to the long axis of the stem. They are composed of sclereids which vary markedly in size, form and complexity of internal structure. Although in general the sclereids have a more or less exten- sively elongated form (with their major axis oriented parallel to the long axis of the stem), they frequently assume aberrant shapes (Fic. 23), apparently due to pressures exerted by adjacent cells during the earlier stages of their excessive enlargement. In some cases, the sclereids have a thick, lignified, multilayered wall and a much restricted or occluded lumen (Fic. 20). In other cases, the sclereids first form a multilayered wall followed by transverse septation and subsequent differentiation of two, three, or more smaller sclereids internally (Fic. 21). Particularly in the case of the largest sclereids, which may attain a diameter of 200 microns and a length of 1000 microns, the protoplasts of the included sclereids may in turn divide, leading to the formation of a second set of still smaller sclereids (Fic. 22). Such compound sclereids of the cells- within-cell type commonly assume many diverse forms internally depend- ing upon the timing and frequency of internal divisions and upon the frequency of diagonally rather than transversely oriented septa. There- fore, in macerations of the phloem, it may be difficult at times to deter- mine with certainty whether certain of the smaller isolated sclereids developed directly from less expanded parenchymatous elements of the secondary phloem or have been released from large compound sclereids by the disrupting forces of maceration. Strands of aggregated sclereids, at least in many cases, are infrequent 148 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII or absent, even in much enlarged roots of this category of pereskias, and commonly do not occur in young stems. Thus, I suspect that such pereskias as Pereskia bahiensis Giirke and P. moorei Britt. & Rose, which have very large, relatively thin, characteristically pinnately veined leaves, will ultimately prove to belong in this category of species when adequate material of the largest and oldest stems can be obtained and studied. 3. Species with Ordinary “Stone Cells” Aggregated in Massive Irregular Clusters In both the secondary phloem and cortex of older stems and roots of Pereskia colombiana Britt. & Rose, P. guamacho Web., P. cubensis Britt. & Rose, and P. portulacifolia Haw., simple, comparatively small sclereids occur in more or less massive aggregations of irregular form (PLATE IV). As indicated in longitudinal sections (Fic. 11), these aggregations do not have a consistently elongated form oriented parallel to the long axis of the stem or root. Furthermore, the sclerotic masses do not contain large compound sclereids but are composed of ordinary “stone cells” as regards size and form. These typical “stone cells’? have profusely pitted, thick, lignified, multilayered walls (Fics. 24-26). When they have a conspicu- ous major axis (Fic. 26), it tends to be oriented at right angles to, rather than parallel with, the long axis of the stem or root. It is evident from a developmental point of view that these sclereids differentiate from parenchymatous elements of the cortex and secondary phloem without excessive expansion and without internal septation. DISCUSSION AND CONCLUSIONS The occurrence of various forms of sclereids in the secondary phloem of pereskias and the absence of such cells in comparable tissue of Pereskiopsis and Quiabentia are obviously of some significance from a generic point of view, and parallel the presence or absence of glochids in these leaf-bearing genera. However, such differences must ultimately be considered in relation to close similarities that occur in other cells, tissues, and parts of the plants. Obviously a synthesizing discussion should be deferred to the last paper of this review. The consistent differences in the form and distribution of sclereids in three distinct categories of pereskias are of considerable taxonomic sig- nificance, not only in the discussion of putative genera, e.g., Rhodocactus, but also in dealing with putative species and varieties. Evidence from the first category of pereskias is indicative of relationship of Pereskia aculeata to P. diaz-romeroama and P. weberiana, of Bolivia, and likewise to P. conzattii, of Southern Mexico, and P. autumnalis and P. nicoyana, of Central America. Evidence from the second category of pereskias is indicative of affinities between P. sacharosa, P. grandifolia, P. bleo, P. corrugata, and P. tampicana; that from the third category of pereskias of affinities between P. colombiana and P. guamacho, of northern South 1961] BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, II 149 America, and the West Indian endemics P. cubensis and P. portulactfolia. The elongated or prosenchymatous form of the sclerenchymatous ele- ments in the first category of pereskias raises the question whether the cells should be classified as phloem fibers with lignified, multilayered secondary walls or as elongated forms of sclereids. The occurrence of ramifying” or coalescing simple pits in the thick secondary walls of these cells is a characteristic feature of many sclereids rather than of typical fibers. Furthermore, there are numerous transitions between short and variously elongated sclereids in the second category of pereskias. Admittedly, it is difficult to classify plant cells in rigid compartments, owing to the frequent occurrence of intergrading or transitional forms of morphological characteristics. However, in view of the well-known diversities in the size and form of sclereids (see Foster ®) I am inclined to refer to more or less elongated cells in Pereskia having thick, lignified, multilayered walls with ‘‘ramiform” or coalescing pits as elongated forms of sclereids rather than as fibers. The frequent occurrence of sclereids within fibers of the primary phloem and the formation of successive generations of smaller sclereids within large ones of the secondary phloem provide significant material for physiological and developmental investigations, as well as electron micros- copy. ACKNOWLEDGMENTS I am indebted to the American Philosophical Society for the loan of a Wild microscope. Since acknowledging the assistance of various individ- uals in the first paper of this series, I have received material of Pereskia sacharosa from Dr, Abraham Willink and of P. portulacifolia from Dr. José Jiménez. I am greatly indebted to these individuals for the time and effort involved in securing adequately preserved specimens for me. The drawings in this and the preceding paper of this series were made by Mr. Elmer W. Smith. EXPLANATION OF PLATES PLATE I Fics. 1-3. DIFFUSE DISTRIBUTION OF SCLEREIDS IN TRANSVERSE SECTIONS OF sTEMS. 1, Pereskia aculeata [Aw 9912] cluster of primary phloem fibers at A, a sclereid in the cortex at B, X 80. 2, P. nicoyana | Rodriguez 662|, * 80. 3. P. diaz-romeroana | Cardenas |, X 80. ° Foster, A. S. Practical plant anatomy. Ed. 2. D. Van Nostrand Co., N. Y., 1949. 150 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII PLATE II . 4-6. FREE AND INCLUDED SCLEREIDS. 4, Pereskia nicoyana | Rodriguez Pk ‘ongituaina section of a stem showing elongated diffusely distributed sclereids in the secondary phloem, 80. 5, P. aculeata |Aw a transverse section of a aed in polarized light between crossed nicols, X 260. 6, Trans- verse section of septate fibers with included sclereids deisel in polarized light between crossed nicols, & 1200. PLATE III Fics. 7-9. AGGREGATED SCLEREIDS OF THE SECOND CATEGORY OF PERESKIAS. 7, Pereskia grandifolia [Castellanos |, transverse section showing two aggrega- tions of sclereids in the secondary phloem, primary phloem fibers at A 80. 8, P. sacharosa [Tucuman], transverse section of a stem showing a agerezat on of sclereids i in the secondary phloem, 80. 9, P. grandifolia | Castellanos |, fanet tudinal section of the secondary phloem showing elongated aggregation of sclereids, & 80 PLATE IV Fics. 10-11. AGGREGATIONS OF “STONE CELLS” IN THE THIRD CATEGORY OF PERESKIAS. 10, Pereskia cubensis {Atkins Garden], transverse section of a root showing massive clusters of sclereids in the secondary phloem, x 80. 11, P. guamacho [Steyermark], longitudinal section of a stem showing irregular masses of sclereids in the outer secondary phloem and cortex, < 80. PLATE V Fics. 12-17. DIAGRAMMATIC ILLUSTRATIONS OF PRIMARY PHLOEM FIBERS AND SCLEREIDS. Included protoplasts are stippled, primary walls and septa are black, first-formed secondary walls of fibers are white, and the secondary walls of sclereids are multilayered. 12, Characteristic form of immature primary phloem fibers prior to secondary wall formation. 13, Part of a nonseptate fiber having thick secondary wall and reduced lumen. 14, Part of a septate fiber having a relatively thin secondary wall. 15, Part of a septate fiber with included sclereids. 16, Precocious septation of an elongated cell prior to the formation of strand sclereids shown in Fig. 17. PLATE VI Fics. 18-26. DIAGRAMMATIC ILLUSTRATION OF THE FORMS OF SCLEREIDS IN THREE CATEGORIES OF PERESKIAS. Included protoplasts are stippled, primary walls and septa are black, and the secondary walls of sclereids are multilayered. 18, Form of elongated, nonseptate sclereids in the first category of pereskias. 19, Part of the secondary wall of such a sclereid showing “ramiform”’ or coalesc- ing pits. 20, Single, nonseptate sclereid from an aggregation in the second cate- ory of pereskias. 21, Compound sclereid from the same source showing three septa and four included smaller sclereids. 22, Doubly compound sclereid from the same source showing septation of such included sclereids as illustrated in Fig. 21 and the maturation of a second set of still smaller included sclereids. 23, Small laterally deformed sclereid. 24-26, Typical ‘‘stone cells” from aggregations in the third category of pereskias. Jour. ARNoLp Ars. VoL. XLII PLATE I aang i: dp fe has ie - ei ‘ty = oy : BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, II PLATE II ARNOLD ArB, VoL. XLII JOUR. BEARING CACTACEAE, II ANATOMY OF LEAF- AILEY, B Jour. ARNOLD Ars. VoL, XLII Puate III EPR fires - * CacTAcEaE, II Jour. ARNOLD ARB. VoL. XLII PLATE BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, « ee ys => 6S SS = LSS Ee i Be Serie re ee HESS neh bret ——> cee aaa ee meee OO sates oe ~ ee © hey wy CGE BEES Septet anes + ae kil oye OF i ae - bine ert 3 : ah te tet <* . LB gr Sire own & , N s,s eS ie eS, a ~ Sie: i pvt te iad ae? Ct MrocENE Woops & BARGHOORN, RAKASH Pp PLATE 6 Jour. Arnotp Ars. VoL. XLII ae % b Pes Woops N, MIOCENE & BARGHOOR PRAKASH ~ PLATE XLIT 1 VOL. ARNOLD ARB. Jour. ce ek CS ee ie ae a Skt? Ee See. < ’ te till “hele sgh Os cy e ° >. : ‘ ‘ Pica r eS Pn b,' hee zi he Ns oe @ 722 4) ° t Bets bey ee, eee + t* > ou q = ie B= ‘= S k : ay He “hg > F is = & Se Pa od ER eae err ats Cm “a ; : wd A ~ | or 7 E f By WE Fenty, * ‘Fearne BZ Si moe ge < Pan’ * ” To , x oboe fee . hn the, Fics. 13-19. Cross sections of leaves to show general outlines. 13, Para- strephia lucida (Venturi 4750, rk x ca. 22; 14, Parastrephia teretiuscula (Cardenas 86, GH), X ca. 50; Parastrephia phyliciformis (Cabrera 7816, P), X ca. 45; 16, Chillotrichio pes Keidelit (Fabris & Marchioni 1748, Lp), * ca. 15; 17, Chiliotrichium diffusum (Eyerdam, Beetle & Grondona, cH), x ca. 20; 18, Parastrephia quadrangularis (West 6088, GH), young leaf, ca. 50; 19, Chiliotrichium diffusum (Eyerdam, Beetle & Grondona 24098, GH), x ca. 20— compare with Fic. 17 from the same collection layer of palisade parenchyma succeeded by one or two layers of spongy parenchyma. The lower epidermis is thin and has a thin cuticle. In all species, the abaxial surface has a thick cap of uniseriate multicellular hairs. Nardophyllum lanatum (Fic. 8) has hairs also on the upper leaf surface, but not so profusely as on the lower. The glandular trichomes differ from species to species. In Chiliotrichium diffusum (Fics. 12, 17, 19) they are numerous and restricted to the upper surface, where they are 1961 | SOLBRIG, ANATOMY OF ANDEAN COMPOSITAE 285 embedded in small depressions. In Nardophyllum lanatum they are scarce and confined to the adaxial surface. The glandular hairs of V. genistoides are peculiar: they are usually uniseriate, consisting of a long stalk (formed by two elongated cells) and a round multicellular head; they extend in rows up to 2 mm. long on the margins and the upper surface. In Cihilio- trichiopsis Keidelii (Fics. 6, 16) and C. ledifolia the glandular hairs are of the common uni- or biseriate, multicellular type already described; they are on the lower surface only in the former species, and on both surfaces in the latter. Aside from the glandular-hair differences, there are minor variations from species to species in the extent of inrolling of the margins, shape and size of the midvein, and lateral expanse of the blade. Nardophyllum genistoides is of special interest. Here the midvein region scarcely pro- trudes and the margins are only slightly bent. These conditions are very similar to those found in the following type, but there are some important structural differences. Type D. Three species in two genera are treated here: Aywacophora deserticola, Chiliophyllum densifolium, and C. fuegianum. The leaves are small, 3-10 mm. long and 1—3 mm. wide, and linear to oblanceolate. They are thickish in texture and are covered both adaxially and abaxially by an equally thick epidermal layer and a cuticle. A layer of palisade paren- chyma lies below the epidermis, which envelopes the leaves on both sides; in the middle of the lamina are one to three layers of spongy parenchyma interspersed with the vascular bundles. Aylacophora deserticola (Fic. 10) and Chiliophyllum fuegianum are hairy on both surfaces; in C. denst- folium hairs are completely absent. All three species have glandular hairs on both leaf surfaces. The leaf margins are straight or slightly curved in C. fuegianum, the midvein region protrudes only slightly if at all and no “inner cavity” is found. Vascular bundles. The vascular bundles of all species investigated are normal collateral bundles. All species of Parastrephia have an adaxially extended bundle-sheath associated with the midvein, enveloping a duct that is probably a resin canal (Fics. 18, 20, 27). In P. lucida this duct is small, but in all other species it is very large and conspicuous. Except in P. guadrangularis, no resin canals were observed in primary stems. Species of Parastrephia and Lepidophyllum lack fibers in association with the vascular bundles. All other genera studied showed some xylary fibers, usually situated at each end of the bundle, or sometimes also between phloem and xylem, or even dispersed between the conducting elements themselves. The midvein of Chiliotrichiopsis ledifolia is surrounded by a bundle-sheath formed by one row of extremely large cells of about the same diameter as the radius of the bundle. THE SHOOT The primary stems are terete or slightly angled in transection except in species of Nardophyllum or Aylacophora. In all other genera the thick 286 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII ry oe 0 Se Fpe Orr SH v, {7 m i : ies “ ‘a yy anti Ka eee wus mae On) ) SS s = uf me (yA = ony, 73 Om, () sue a 1K RA —— ue Wy ive ARS “aie: 18 i PRY PAO 6 SS tH te AW SS oo” = seats THY yay rH SS 077 ee L_) eae t — aniain mw ‘@ reat NM J i] U x q ‘ Ne NeIN too Wh Z®, S 24 oe Ov Se =N “J a C2 4K y LON ZENS, NY ed — VA! yQ p s N Y OVP 4 \ SS ioe OY SIS ae Lr SPE OU Pe e242 ROT OY PAR OTD C} PB ES SEA cD CAL SS Oo eo 8 ine ASR “ace 2, TSE V2 ro s: \ RSH OW, GY eT Wasese: STAKES PSK SL YY & Soh ke \ \ Qj pr) \\ 2° \ a th eK at YJ Gi\ \ IL \ | | J1 ) ¥ J) \ eae | \ N N 20. Camera lucida drawing of cross section of leaf of Parastrephia lucida Fic. (Venturi 4750, GH), showing anatomical detail; hairs drawn schematically. 1961 | SOLBRIG, ANATOMY OF ANDEAN COMPOSITAE 287 epidermis usually is covered by a thick layer of glandular and nonglandu- lar hairs like those present in the leaves (Fics. 21 & 23). The cortex is composed of thin, isodiametric cells, and occupies one-third to one-fourth of the diameter of the stem, surrounding a cylinder of collateral bundles. The pith consists of large, thin-walled cells at the center and smaller, thin- or sometimes thick-walled, fiber-like cells on the periphery. In Parastrephia phyliciformis the pith is unusually large, occupying over three-quarters of the diameter of the stem. In all species of Nardophyllum the stem is angled or ribbed in tran- section. The epidermis, cortex, conducting tissues, and pith are similar to those just described. In addition, well-developed strands of sclerenchyma fibers along the ribs add rigidity to the stems. The pith in Nardophyllum is also sclerified, especially in N. armatum. In N. bracteolatum (Fic. 24) and Aylacophora deserticola (Fic. 22), deep grooves, reminiscent of the “inner cavity” of the leaf, and also lined with a thick cap of hairs, occur between the ribs. The epidermis is thick and heavily cutinized on the blunt, exposed surface of the ribs, but is thin and only slightly cuti- nized at the inside of the grooves, where stomata are present. The ribs are abundantly sclerenchymatous peripherally, while the grooves are lined with photosynthetic cortical parenchyma. It is clear that the primary shoots in these species are specialized for photosynthetic activity. In Aylacophora deserticola, in which the leaves are few and short-lived, photosynthesis must take place chiefly in the stem. Nodal anatomy. The number of traces and gaps at the node has been considered to be a character of possible evolutionary significance, espe- cially in woody plants, and may sometimes provide useful taxonomic in- formation. In Compositae, the basic type of node is trilacunar with three traces; this type is basic also for most of the species studied here. Since type of node is usually very consistent in related groups — whole genera and sometimes even whole families — it is interesting to note that in eight of the species studied the node is unilacunar with a single trace, presum- ably by reduction. Unilacunar nodes have been reported previously in Compositae only for upper leaves in certain genera of Madineae (Carl- quist, 1959). Even more remarkable, however, is the occurrence in Para- strephia, Nardophyllum, and Chiliotrichium of species with a trilacunar node and others with a unilacunar node. Parastrephia phyliciformis, P. lucida, and P. quadrangularis have a trilacunar node; P. lepidophylla and P. teretiuscula are unilacunar. Lept- dophyllum cupressiforme has a unilacunar node with one trace subtending each of the opposite leaves. Nardophyllum armatum, N. genistoides, N. bracteolatum, and N. chiliotrichioides all have a trilacunar node with three traces; V. bryoides and N. obtusifolium are unilacunar with a single trace. Aylacophora deserticola, Chiliotrichiopsis Keidelti, and C. ledifolia have Fic. Camera lucida drawing of stem and leaf cross section of Bhd die ely (Cardenas 86, GH), showing cellular detail. Each division of sc bo a 2) CO JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII bes "4 ( Sin = Chef hee Tighe: rt tid i Gtk ee ay “+ * : '* iy ag k Fics. 22-24. Stem cross sections. 22, Aylacophora eeenee oC 11053, LP), X ca. 15; 23, Parastrephia teretiuscula (Cardenas 86, ca. 20; 24, Nardophyllum bracteolatum (Ruiz Leal 7212, Lp), & ca. “1s Fics. 5-27. Details of leaves of / Vardophyllum bryoides igen pkalag ini dona 24083, GH). 25, Primordia in differ ee Brat of develop 50; 26, detail of glandular eure x ca. 100; , detail of leaf st ae hairs and resin canal, * c the trilacunar node common to the family. Chiliotrichium rosmarinifolium also has a trilacunar node, but that of C. diffusum is unilacunar. Final y, both Chiliophyllum densifolium and C. fuegianum have a unilacunar node with a single trace 1961 | SOLBRIG, ANATOMY OF ANDEAN COMPOSITAE 289 DISCUSSION In the preceding paragraphs anatomical characters of leaf and node have been described. Their significance from the point of view of adapta- tion, evolution, and taxonomy remains to be seen. Cushion and mat plants have been the object of intensive investigations in the past, especially studies of an ecological or anatomical nature. An attempt to review the extensive literature is not pertinent to the present study and therefore only some salient information will be pointed out. Hauri and Schroter (1914, p. 619) defined cushion plants as “peren- nials, herbaceous or woody, usually green chamaephytes, of spheric, hemi- spheric, or shingle-like, compact growth. The branches are numerous, with short internodes, covered more or less to the base and without in- terruption by small, more or less rigid, sessile leaves, withered in various ways. The branches end on the same level and are either closely appressed or the intervening cavities are filled by foreign material. In this manner is produced a certain strength, compactness, and closeness of the indi- viduals, which consist of a living, appressed cover over a spongy mass of decaying material created by the plant itself’ (free translation). If we adhere to this definition, none of the species studied can be considered cushion plants. However, Hauri and Schroter also recognized the exist- ence of forms which do not fit their definition but which are nevertheless usually referred to as cushion or cushion-like plants. For these “‘cushion- like’ types they proposed nine different categories. The plants studied here would come under the grouping of ‘“spherical-shrubs” (Kugel- strducher), characterized by the loose growth (for cushion plants) of the branches, allowing the passage of air and light into the interior of the plants. In a later study, Hauri (1916) investigated the anatomical struc- ture of leaves and stems of 73 species in 29 genera and 16 families. It is interesting to note that he found a strong convergence in anatomical characters: sessile leaves; a thick epidermis and usually a thick cuticle, or a thick layer of hairs, or both; a well-developed palisade layer; and a relatively slight sclerification of the stem. In this last respect, Nardo- phyllum and Aylacophora, of the seven genera investigated, do not agree. Rauh (1939) presented a very thorough and complete analysis of cushion plants from a morphological point of view. According to him, cushion plants can be defined as “perennial, evergreen plants, with no well-defined central axis and numerous radial, stratified branches, the ends of which form a more or less compact surface due to a similar amount of growth in length” (free translation, p. 269). He divided the “growth- form system of cushion-plants” into seven main types: I. True cushion plants; II. Creeping cushions; III. Turf cushions; IV. Rosette cushions; V. Succulent cushions; VI. Cushion mosses; VII. Monocarpic (‘‘hap- axanthe”) cushion plants, that is, annual or biennial forms. He further subdivided these into subtypes. Type I, true cushion plants, was divided into high (radial-high) cushion plants and flat (radial-flat) cushion plants; the first subtype is further subdivided into cushion trees, spheric shrubs, 290 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII and hollow cushions. The species studied in the present investigation all fall under the category of spheric shrubs. Although Rauh did not undertake any anatomical study, he reviewed the different leaf forms of cushion plants and grouped them into different types. Even if a detailed analysis of this classification is not in order here, a few comments may be made about his curved-leaf (Rollblitter) type. He characterized it as that in which the lamina, instead of flatten- ing out at maturity, maintains the position it had in the bud. He indi- cated further that leaves with revolute margins are found only in Fran- kenia and Anthobryum, among cushion plants. As can be seen from the present study, revolute margins are more frequent among cushion-like plants (cf. also Cabrera, 1951; Carlquist, 1958). Although, on super- ficial inspection, the small leaves and inrolled margins might appear to be persistent juvenile forms, this view cannot be supported if the special- ized anatomical structure of the leaf is taken into account. Likewise, the contention that sessile leaves in Anthobryum are the result of “arrested growth” in the petiole of a related and ancestral Frankenia must be dis- carded. A final word might be added with regard to Rauh’s classification of the leaf types in cushion plants. Although a classification of leaf types can be helpful, a formal and complicated subdivision based exclusively on morphological types, regardless of phylogenetic origin, is not conducive to better understanding. When related species, sometimes even of the same genus, have to be separated to fit the classification, it does nothing to clarify taxonomic relationships. Rauh’s proposed body types of cushion plants also suffer from the same defect. Espinosa (1932) made a very careful and complete study of high- montane plants from the Cordillera de los Andes, especially the pdramos of Ecuador. He recognized four main vegetative types: cushion plants, rosettes, bushes, and spiny bushes (Dornstréucher). Rosettes, cushion plants, and shrubs are commonest in the puna and pdramos, spiny shrubs in the deserts of northern Chile and southern Pert, although they are fairly common also in the puna and in the “dry puna.” Of more interest here are his anatomical investigations. He found that high-cordilleran plants generally are protected against desiccation by a thick epidermis and cuticle and a thick layer of hairs. Although none of the species studied by Espinosa shows leaves with the same cross section as those studied here, species with involute- and revolute-margined leaves were found, and also plants with leaves appressed to the stem, and some with large interior air chambers. Similar air chambers were reported for different species of the subantarctic islands of New Zealand by Cockayne (1909). No data on nodal anatomy were presented by Espinosa. By way of summary, although none of the species studied by me can be considered a “typical” cushion plant, there seems to be general agree- ment that my material belongs ecologically to a cushion-like category. ‘ome o1 the foliar features, such as thick epidermis, thick cuticle and compact palisade, appear to be relatively common among both cushion plants and cushion-like plants. On the other hand, the peculiar revolute- 1961] SOLBRIG, ANATOMY OF ANDEAN COMPOSITAE 291 margined leaf of some species of Parastrephia and Nardophyllum is less common and is only indirectly associated with a cushion habit. This type of leaf is widespread in Ericaceae (Hagerup, 1953), where it is frequently associated with dry habitats, but it is not very common outside that family. Hagerup (1953), in reviewing the leaves of Ericaceae, considered the revolute projections to be not the true margins of the lamina, but a later outgrowth. This view was challenged by Hara (1956) on ontogenetic grounds. My own observations on shoot apices and primordia appear to confirm Hara’s views, although they were not always decisive. Even at very early stages the primordia show configurations correlated with the final shape of the leaf (Fic. 25), and, although there is an increase in size of the revolute projections, they can hardly be interpreted as a “later outgrowth.” The point is, in any case, a rather academic one, which de- pends largely on the particular definition of leaf margin. Leaf venation in cushion plants has been studied very little. From my study of Raoulia (Solbrig, 1960), a New Zeland genus of true cushion plants, it would appear that a strong reduction in leaf venation may accompany the cushion habit in this genus. A similar conclusion is reached in the present study. Nodal structure in cushion plants has also been neglected. The presence of two nodal types in the same genus is remarkable, and suggests greater flexibility in this character than has usually been assumed. The correla- tion with other structures suggests that in these Compositae the unilacunar condition is derived, adding to the body of evidence presented by Bailey (1956) that this is a common situation in advanced families of Angio- sperms. That the trilacunar situation is always the primitive condition in Compositae cannot be said with certainty. Although cushion plants are found in most high-montane areas of the world, they are most abundant in certain portions of the Andes of South America, especially the pdramos of Colombia and Ecuador and the puna from Peru southwards, and in Patagonia (Andine Dominion of Cabrera, 1953). In these areas are found about 50 per cent of the pulvinate species known (Hauri and Schroéter, 1914). Hauri and Schroter recorded 34 families and 78 genera containing a total of 338 species of cushion plants. Rauh (1939) put the number at about 400 species. Families with the largest number, according to these authors, are Compositae (15 genera, 52 species), Caryophyllaceae (12 genera, 49 species), Umbelliferae (4 genera, 34 species), and Cruciferae (2 genera, 34 species). The inde- pendent acquisition of this habit by more than 300 species of so many different families in different parts of the world suggests that this form of growth has some adaptive advantage, and this has been generally assumed. On the other hand, it is not very clear what the exact adapta- tion might be. The early interpretation was that cushion plants were xerophytic. When it was discovered that they often grow in moist habitats, the concept of “physiological xeromorphy” was developed (Hauri and Schroter, 1914; Hauri, 1917). Experiments by Walter (1931) indicated 292 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII for cushion plants on Pikes Peak, Colorado, a low osmotic pressure which discredited the idea of “physiological xeromorphy.” Rauh (1939) pre- sented a lengthy discussion of the subject. It suffices to say that differ- ent factors probably are operative, and that, although there is undoubtedly a convergence in morphological characteristics, there is no necessity to assume an equal convergence in the physiological responses of the plants. Only careful experiments on a larger number of plants could elucidate this problem. It is also quite clear that the plants show definite local adaptations within the broad environmental limits where they occur. Al species considered in the present investigation grow in dry to very dry environments (Cabrera, 1958; Dusén, 1905: Fries, 1907; Skottsberg, 1916; Soriano, 1948), although such species as Chiliotrickium difiusum are found also in moister environments (Skottsberg, 1916). A few words should be said about the possible taxonomic implications of the present work. Cabrera (1954) has considered the relationships of Nardophyllum and concluded that it is closely related on the one hand to the South African genus Pteronia L.. and to Chiliotrichium and Para- strephia on the other. Nardophyllum differs from Parastrephia and Chilio- trichium, as well as from Lepidophyllum, Chiliophyllum, and Chilio- trichiopsis, by its tubular heads with all flowers perfect, the other genera aving ovulate ray flowers, which can be either ligulate or tubular. An- atomically, Nardophyllum is separated from the genera mentioned by the possession of cortical sclerenchyma. It shares the tubular head and corti- cal sclerenchyma with the genus Aylacophora; the two genera differ in pappus and shape of achenes, characters which are generally variable within the tribe. There is also remarkable similarity in stem structure between Aylacophora deserticola and Nardophyllum bracteolatum: the leaf structure, on the other hand, is quite different in Aylacophora and Nardophyllum. In any case, it seems safe to assume that Avlaco phora deserticola is closely related to Nardophyllum, Cabrera (1954) separated Lepidophyllum and Parastrephia because of the opposite leaves, well-developed ray ligules, shortly papillose or glabrous achenes and paleaceous pappus of Lepidophyllum, contrasted with the alternate leaves, tubular or shortly ligulate ray flowers, hairy achenes, and bristly pappus of Parastre phia. Anatomically, Lepidophyllum is similar to Parastrephia terctiuscula, the main differences relating to the opposite leaves of the former. It seems likely that the similarities are due to con- vergence in these two reduced species. Chiliophyllum and Chiliotrichiopsis differ in pappus and achenes: hairy pappus and weakly pubescent achenes in Chiliophyllum versus scaly pappus and strongly pubescent achenes in Chiliotrichiopsis (Cabrera, 1954). Anatomically, Chiliophyllum has a unilacunar node. Chiliotrichio p- sis, a trilacunar node. This character is difficult to evaluate taxonomically in this group of genera since in Nardophyllum, Parastrephia, and Chilio- teichium, both unilacunar and trilacunar nodes occur in the same genus. The leaves of Chiliophyllum are less well developed than those of Chilio- trichio psis. 1961 | SOLBRIG, ANATOMY OF ANDEAN COMPOSITAE 293 The flowers of Chiliophyllum are also very similar to those of Chtlio- trichium, the only difference being the color of the ligulate flowers, which are yellow i in Chiliophyllum, but white in Chiliotrichium (Cabrera, 1954). In view of the situation in other genera of the tribe (Keck, 1958) this is a poor distinction. Anatomically there is not much difference, aside from the fact that one species of Chiliotrichium, C. rosmarinifolium has a tri- lacunar node. LITERATURE CITED Battey. I. W. 1956. Nodal anatomy in retrospect. Jour. Arnold Arb. 38: 269- 287. CasrerA, A. L. 1937. ae argentinas nuevas o interesantes. Notas Mus. La Plata Bot. 2: 171- 1944. Compuestas ek anas nuevas o criticas. /bid. 9: 243-259 1945, ole del género Lepidophyllum (Compositae). Bol. Soc. = aa TQ fa) ee OD img iss) jo) a —_ 951. nie nuevo género de Compuestas. /bid. 4: 129-132. 1953a. Un nuevo género de Astereas de la Republica Argentina Ibid. 1953b. Esquema fitogeografico de la Republica Argentina. Revista Mus. La Plata IT. 8: 87-168. —, 1954a. Las especies del género Nardophyllum. Notas Mus. La Plata Bot. 17: 55-66. 1954b. Compuestas sudamericanas nuevas 0 criticas II. Ibid. 71-80. ———. 1958. La vegetacion de la Republica Argentina. i La vegetacion de la ‘Puna Argentina. Revista Invest. Agr. 11: 317-412. CarLoutst, S. 1958. Anatomy of Guayana Mutiseae. Part Il. Mem. N. Y. Sot. Gard. 10: 157-184. — 1959. Studies on Madinae: anatomy, cytology, and evolutionary rela- tionships. Aliso 4: 171-236. Cockayne, L. 1909. The ecological botany of the Subantarctic Islands of New Zealand. Jn: Ch. Chilton, ed., The Subantarctic Islands of New Zea- land. 1: 182-235 Dustx. P. 1905. Die Pflanzenvereine der Magellanslinder nebst einem Beitrage zur _ Okologie der Magellanischen Vegetation. Sv. Exped. Magellansland. ESP1Nos R. "1932 _ Okologische Studien tiber Kordillerenpflanzen. Bot. Jahrb. 65: 120-211. FRIES, 4 1905. Zur Kenntnis der alpinen Flora im nordlichen Argentinien. Nov. Acta Reg. Soc. Sci. Upsal. IV. 1: 1-205. Hacerup. 0. 1953. The morphology and systematics of the leaves in Ericales Phytomorphology 3: 459-464. Hara, N. 1956. Development of a Ban margin in the Ericaceae (a preliminary report). Bot. Mag. Tokyo 69: Havurl, H. 1916. saben an m, Beeeteee mete morphologischen und okol logischen Notizen. Beih. Bot. Centralbl. 33: 293. —— anp C, ScHRoTeR, 1914. Versuch einer Gear ae siphonogamen Polsterpflanzen. Bot. Jahrb. 50(Suppl.): 618-656 JoHANSEN, D. A. 1940. Plant microtechnique. McGraw- Hill & Co., New York. 294 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Keck, D. D. 1958. Taxonomic notes on the California Flora. Aliso 4: 101-114. Mayr, E, 1942. Systematics and the origin of species. Columbia Univ, Press, New Yor METCcALFE, C. R. AND L. CHALK. 1950. Anatomy of the Dicotyledons. Vol. 2. Clarendon Press, Oxford. Rauw, W. 1939 Uber polsterformigen Wuchs. Nov. Acta Leopl. II. 7: 266-508. REICHE, C. 1902. Flora de Chile. Vol. 3. Imprenta Barcelona, Santiago de “hile SIMPSON, G. G. 1953. The major features of evolution. Columbia Univ. Press, New York. SKOTTSBERG, C. 1916. Die oe langs pe lati de los Andes S. vom 41° S. Br. Sv. Vet-Akad. Handl. 56: SotsriG, O. T. 1960. Leaf venation and pubescence in oe a Raoulia (Compositae). Jour. Arnold Arb. 41: 259-269, SoRIANO, A. 1950. La vegetacién del Chubut. Revista Argent. Agr. 17: 30-66. STEBBINS, G. L. 1950, Variation and evolution in plants. Columbia Univ. Press, New York. Watter, H. 1931. Die Hydratur der Pflanze und ihre physiologische- okologische Bedeutung. Jena. GRAY HERBARIUM, HARVARD Cisse 1961] NEVLING, REVISION OF LINOSTOMA 295 A REVISION OF THE ASIATIC GENUS LINOSTOMA (THYMELAEACEAE) Lorin I. NEVLING, JR. THE SUBTRIBE LINOSTOMATINAE of the Thymelaeaceae is composed of three closely related genera which have been combined taxonomically in various ways in the past. Previous treatments of these genera have not been inclusive, and, therefore, a preliminary survey to establish generic limits has resulted in considerable realignment of the species. Two of the genera, Linostoma Wall. ex Endl. and Enkleia Griff., are restricted to the area from India to Indochina, Malay Peninsula, and adjacent islands, while the third, Lophostoma Meissn., is restricted to the Amazon basin. Numer- ous examples of a disjunction between southeastern Asiatic and south- eastern North American groups have been studied and reported but an Asian—Amazonian disjunction is unusual. For this reason an attempt will be made to define as precisely as possible the interrelationships of Linostoma, Enkleia and Lophostoma. In order to pursue these interrela- tionships it is believed essential to revise each of the genera concerned prior to generalizing about the subtribe as a whole. Accordingly, this re- vision of Linostoma is the first of a contemplated series. SPECIAL MORPHOLOGY Some of the more interesting morphological features found within the genus Linostoma during the course of this taxonomic investigation are presented in the following discussion, A few of these features have not been reported previously for the genus, and others are emphasized to indicate problems requiring additional investigation. Also included are several anatomical elements which have been employed previously as taxonomic characters and which I feel demand at least a summary review. Finally, additional information can be found in the body of the descrip- tions and in the discussion following each species. It must be stressed that the number of specimens examined (indeed available) was far too few to obtain any but the very roughest type of information. A more thorough investigation which would include a study of the variation in individual plants in the field and under experimental conditions is necessary as the first step to an understanding of the varia- tion exemplified in the herbarium specimens. However, even this brief study gives minor support to the taxonomic delimitations presented in this revision. The leaves and flowers were prepared for study by gentle heating in water plus a commercial detergent until softened and then were cleared with a 5% solution of sodium hydroxide (Foster, 1950). The hydroxide 296 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII treatment often left the specimens in a too softened condition. This was remedied by treatment with chloral hydrate, which tends to harden the tissue and at the same time accounts for some additional clearing. In practically all instances it was necessary to bleach the specimens in 50% “Clorox” to obtain final clarity. The time necessary for the clearing processes varied widely from specimen to specimen. The cleared leaves were stained with a 1% safranin solution in 95% ethyl alcohol for venation studies or with 5% ferric chloride and tannic acid in 70% ethyl alcohol for the study of cell types. The latter staining was particularly useful in studying the stomata. All specimens were de- hydrated appropriately and mounted in Gum Damar or “Diaphane.” As in the clearing processes, the staining times varied quite widely. Cleared flowers were studied both unstained and stained: when stained, a 1% solution of safranin in 95% ethyl alcohol was used. I personally feel that the stained specimens are more useful, particularly when the finer vascularization is being studied. The stained flowers were dehydrated and mounted in Gum Damar. Leaf cross sections and petioles were prepared for study by boiling in water or by clearing as in the procedure outlined above. They were em- bedded in paraffin through a tertiary butyl alcohol series and sectioned. A safranin-fast green staining schedule was followed. In addition, free-hand sections stained with potassium iodide and sul- phuric acid, as well as with phloroglucin and hydrochloric acid (Foster 1950) were later found to be useful in determining the degree of lignifica- tion of certain tissues. This technique has been employed with stems, petioles, leaves and bracts. Vegetative Morphology. The vegetative axis is monopodial, but the precocious development of axillary buds often obscures this nature. Sev- eral types of branching usually can be observed within any single rela- tively large collection. The simplest pattern is one in which dichotomous branching, with equal development of both axes, appears to have taken place. This type is the result of the destruction of the apical bud fol- lowed by development of the opposite axillary buds. An aborted ter- minal bud often may be observed in the fork of the dichotomy but it is gradually obscured with time. A second type of branching, also appearing to be dichotomous but unequally so, is the result of the continued growth of the main axis accompanied by the development of one of the axillary buds of the subtending node. The most striking pattern, however, is in the form of a trichotomy. This pattern is formed by the continued growth of the main axis accompanied by the simultaneous development of both axillary buds of the subtending node. The members of the trichotomy may be developed equally or unequally, and, in addition, the lateral axes may be slightly subopposite, depending on the exact position of the axil- lary buds. In a few instances, the branching patterns described above are not clear-cut because the leaves subtending the developing axillary buds some- 1961] NEVLING, REVISION OF LINOSTOMA 297 times do not appear to be present. These leaves are then borne on the developing axillary branch beyond the point of external furcation. They are usually near the division and are unpaired, in contrast to the leaves of the normal axis. The trace which supplies these leaves is initiated be- low the external furcation as is the trace of the axillary bud. Immediately below the external furcation the division of leaf and axillary bud traces is not yet complete. For some unknown reason the leaf is then “carried out” onto the developing axillary branch which it actually subtends. This phenomenon is more prevalent in the inflorescence and is further discussed there. The axillary branches themselves are of interest in that they may be ascending or become weakly to strongly uncinate, facilitating climbing or scrambling. The “hooks” are similar in appearance to those found in species of Cardiospermum (Sapindaceae) or in Schefflera caudatifolia Merrill (Araliaceae). There are reports (Metcalfe & Chalk, 1950) based on stelar anomalies indicating the possible origin of genera with shrubby or treelike habits from lianalike ancestors. It is because of the develop- ment of the uncinate branches that the habit of individual species appears to be so diverse. I suspect that the plants of Linostoma are erect shrubs or trees which are not above scrambling if the opportunity presents itself. The presence of an extraxylary fiber sheath surrounding the stele often is used as a diagnostic characteristic of the family. These fibers are often employed as twine in various parts of the world. They have relatively thick walls of cellulose and, in this genus, may become slightly lignified after a number of years. The lack of strong lignification of the fibers may contribute in part to the flexibility of the stems. The structure of the leaf was investigated because of the widespread use of the venation pattern as a means of distinguishing between Linos- toma and Enkleia. The leaves are dorsiventral. The upper epidermis, covered by a rather thick cuticle, is composed of relatively large tabular cells in L. persimile and L. pauciflorum and is devoid of stomata. A hypo- dermis of large, thin-walled cells is also found in these two species. In L. decandrum a single epidermal layer, comparable to the double layer of the other species, is found. The palisade layer immediately below is compact and composed of more or less isodiametric cells which are at most only a fifth the size of the epidermal cells. This palisade appears to be three or four cells thick. Within this layer or the one following, occur some large, thin-walled cells which contain druse-like crystals. These crystals are most prevalent in L. decandrum, The spongy paren- chyma layer below the palisade is extremely loose, with very large air spaces. The cells composing this layer are comparatively large and gen- erally are branched. The lower epidermis is composed of tabular cells which, in the case of L. pauciflorum, are conspicuously bullate on the outer periclinal wall. This peculiar papillation of the lower epidermis gives a characteristic farinose appearance to the lower leaf surface when viewed under low magnification. The cells of the lower epidermis, which is in- 298 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLII terrupted by numerous stomata, are consistently smaller than those of the upper epidermis. In Linostoma persimile and L. pauciflorum the stomata are similar, and it is impossible to distinguish these two species on the basis of stomatal type or size (Text Fic. I, 2, 3). These stomata are of the ranunculaceous type (accessory cells morphologically similar to the surrounding epidermal cells). This stomatal type has not been reported previously for the genus, although it is well known in other genera of the family. The guard cells are borne in the same plane as the epidermal cells. =” 1 2 3 Text Fic. I. Camera lucida drawings of the lower leaf epidermis . Linostoma. 1, L. decandrum (Parry 410); 2, L. persimile (Garrett 1397); L. pauci- florum (Kerr 11820A). ” Stippling pent stoma on different is from sur- rounding cells. The stomatal condition is more complex in Linostoma decandrum. In- stead of a pair of accessory cells as in the previous species, there is a cluster of small cells completely surrounding the stoma. Each cell is awl- shaped and bends slightly over and beyond the guard cells so that an urceolate structure is formed. The guard cells are situated at the base of this pit and are, in addition, borne in a plane above that of the lower epidermis (TExT Fic. I, 1; also Solereder Fig. 174, B, C and repeated in Metcalfe and Chalk, Fig. 284, B, E) The pinnate venation is essentially similar in all species of the genus. The primary lateral veins are very numerous and close to each other (several millimeters apart at most) and are more or less straight, parallel, and simple, but sometimes dichotomize. Each vein is subtended by an extraxylary fiber-sheath which is often more extensive than the vein often better developed than the contributing lateral veins (except in L. persimile). Its proximity to the margin varies slightly among the species. The secondary veins, after departing from the primary veins, run more or less parallel to them, The very fine appearance of the venation is due to this parallel orientation of veins and veinlets. The veinlets are also subtended by the extraxylary fibers. These fibers react strongly with potassium iodide and sulphuric acid to give a deep blue color but do not give a color reaction to phloroglucin and hydrochloric acid. On this basis 1961] NEVLING, REVISION OF LINOSTOMA 299 it appears that they are composed of cellulose and are not lignified. The fiber sheath extends almost to the end of the veinlet where it is often re- placed by variously shaped sclereids. In Linostoma persimile these sclereids do not extend to the vein-ending, but in both L. decandrum and L. pauci- florum they extend either to the vein-endings or slightly beyond them (see Text bie. IT). Cross sections of the petiole were studied to check information in Met- calfe and Chalk (1950) indicating the presence of intraxylary phloem in petioles of Lophostoma and Enkleia but not Linostoma. In all species of Linostoma the xylem, at mid-petiole, is arc shaped and often is inter- rupted by rays a single cell broad. In L. decandrum and L. pauciflorum it appears that phloem can be found both above and below the xylem-arc, thus indicating the presence of intraxylary phloem within the genus, con- trary to previous reports. In L. persimile the phloem is found only below the xylem-arc. At best this seems to be only an academic point and appears to lack taxonomic significance. The fibers which usually accompany the vascular bundles either are poorly differentiated in the petioles or are lacking altogether. When present, they have a negative color reaction with phloroglucin and hydro- chloric acid. However, they stain very dark blue with potassium iodide and sulphuric acid indicating the cellulosic nature of the cell walls. Reproductive Morphology. At the time of the initiation of the flowering response the vegetative pattern in the terminal portion of the shoot becomes modified. The changes which take place in the external morphology of the shoot at this time possibly are controlled by the same physiological forces which cause the induction of the flowering response. The same basic terminology which I applied to the inflorescence of Daphnopsis in a previous paper (1959, pp. 262, 263) is used here. The inflorescence is considered as being composed of the basic units of primary peduncle, rachis, secondary peduncles and_ flowers with their pedicels. There are a few additional features in the inflorescence of Linostoma not found in Daphnopsis, and these are discussed below. The flowering shoot is composed of many inflorescences, the sum of which appears to form a single large and highly branched, terminal in- florescence. The flowering response destroys the apical dominance of the stem apex, and the axillary buds of several subtending nodes begin de- velopment. This complex change must take place during the initiation of the affected parts and prior to their subsequent elongation. The re- sulting precocious development of the axillary branches is responsible for the decidedly peculiar appearance of the inflorescence. The main stem axis appears to trifurcate (see also discussion of branching patterns) with- out good cause, i.e., without subtending leaves. This pattern is due to the initiation of the leaf traces and the axillary bud traces shortly below the external furcation, followed by rapid elongation to the extent that the leaf is “carried out” onto the axillary shoot which it actually subtends. The external relationships of vegetative shoot, axillary branches and 300 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII leaves are obscured. The developing axillary shoots may undergo several such divisions until a large paniculiform structure is developed. As a result, the primary peduncle often bears near its base the reduced, dis- placed leaf of the subtending node. This displaced leaf is homologous to the bract found on the compound inflorescences of some species of Daphnopsis (especially D. americana). At the apex of the primary peduncle is borne a single small bracteole, which is usually caducous but occasionally persistent. This structure is further discussed following L. pauciflorum. In addition, the primary peduncle bears a pair of bracts. These bracts are opposite to alternate and are borne at or below the middle of the primary peduncle (Linostoma decandrum and L. persimile), or from the middle to the summit (L. pauciflorum). These bracts appear to be ex- panded fully, or nearly so, by the time the flowers attain anthesis. This contrasts with the situation in Enkleia in which the expansion of bracts is delayed until shortly after anthesis. The function of these bracts is uncertain, although Domke has used the very appropriate term “Flug- brakteen” to describe them and Ridley has included this genus in his book (1930) on plant dispersal under the heading “Bract Wings.” Speak- ing of the bracts in Linostoma he says (p. 92), “They become pale and papery when the fruit is ripe, and are detached, adhering to the fruit, so as to be blown away separately.” He attributes the small size of the plants of L. persimile to the drupe being too heavy to be borne away by the bracts, whereas, in L. decandrum and L. pauciflorum the drupe is light enough to be easily drifted. Finally, he says (p. 93) “Here the re- duction of the number of fruit produced on the spray, and the persistence of the terminal pair of leaves, has converted a bird-dispersed plant into a wind-dispersed one, and has allowed the plant to be successful in climbing to the top of the forest.” Theoretically, I disagree with Ridley because it appears, on the basis of herbarium material, that all separations of fruit plus peduncle with bracts are mechanical accidents. In these cases the peduncle at the point of separation is irregularly broken and does not seem to be as easily separated as at the articulation zone between pedicel and secondary peduncle. Ridley has seen the plants in nature and what appears to be theoretically improbable may occur in actuality. The shape and structure of the bracts are in several respects different from those of the foliage leaves. The bracts vary from ovate to oblanceo- late but are of a shape and size usually quite dissimilar from that of the leaves. In addition, they are much thinner than the leaves and are dis- tinguished further by their white to cream color. In spite of their extreme thinness, the bracts retain the same morphological components (i.e., epi- dermis, palisade and spongy parenchyma) and in approximately the same ratio as the leaves Another, and perhaps more striking, characteristic of the bracts is the relative spaciousness of the venation. The venation pattern is, in fact, more like that which one would expect to find in leaves, i.e., more reticu- late. The number of primary veins is greatly reduced from the number 1961] NEVLING, REVISION OF LINOSTOMA 301 found in leaves. In addition, the number of secondary, and particularly smaller, veins is much lower. The lack of development of the smaller veins accounts for the appearance of spaciousness of the venation. As in the leaf, a submarginal vein is present in all species. The extraxylary fiber sheath associated with the veins differs from that of the leaves in one striking aspect. In contrast to the leaf, the sheath supporting the minor veins is not constituted entirely of fibers but often has an outer layer composed of relatively short sclereids. These sclereids are irregularly shaped, with at least one surface noticeably enated. his characteristic allows microscopic identification of leaves and bracts (see TEXT Fic. e fiber sheath reacts to the lignification and cellulosic tests in the same way as in the leaves. e color reaction indicating cellulosic walls lacking lignification is less intense than in the leaves. In like manner, the color reaction indicating lignified xylary elements also is much less Text Fic. II. Photomicrographs of fine venation of leaves and bracts of Linostoma. 1, L. i leaf Ww allich 4203); 2, L. decandrum, bract (Wallich 4203); 4 . persimile, leaf (Garrett 1436); 4, . persimile, bract (Garrett 1436); io pauciflorum, leaf (Hallier B.2261); i, pee raeite bract (Hallier B.2261). Note numerous crystals in leaf of L. ante 302 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII intense, and it seems that the tissues of the bracts are developed much less than those of the mature leaves. The flowers are arranged on the rachis in an indeterminate fashion (the lowermost flower blooming first). The flower clusters are umbelliform to subracemiform, depending on the relative length of the rachis. I sus- pect (without substantiating evidence) that the condensation of the rachis and the resulting clustering of flowers is due to selective pressures in favor of the formation of a pseudanthemum. The flowers are bisexual, regular, pentamerous, perigynous, and pedicel- late. They are probably conspicuous at anthesis by their coloration and by the presence of the light-colored subtending bracts. The terminology applied to the parts of the thymelaeaceous flower has varied with the individual author’s personal inclinations. Accordingly, I am retaining the same terminology that I used in Daphnopsis (Nevling, 1959) and that for the same reasons. Further gross morphological in- formation, beyond that presented in the following discussion, concerning the flower may be found in the specific descriptions. The calyx tube is appendicular in origin, that is, composed of the fused bases of calyx, corolla, and androecial members. It is always tubular, although the exact shape is variable, and surrounds the superior ovary. The exterior is glabrous, while the interior varies from entirely villous to glabrous. The trichomes are exclusively unicellular and unbranched. The tube is vascularized by ten distinct veins which are marked ex- ternally by conspicuous ridges. Near the orifice the traces to the androe- cium depart, one to each stamen. Above the departure of androecial traces the vascular pattern becomes exceedingly obscure due to the proliferation of traces and to the overlying pubescence. I am not satis- fied with my results in this critical area of the calyx tube. The size and shape of the five calyx lobes vary among the species. The position of the lobes at anthesis is a more important taxonomic character. In Linostoma decandrum they are strongly reflexed, in L. pauciflorum they are spreading, and in L. persimile they are erect to slightly spread- ing. In all species the lobes are vascularized by three main veins. The lateral veins often proliferate at the orifice, and all send off numerous branches which often anastomose, particularly towards the lobe apex. Extraxylary fibers are present even on these small veins. The ten petals are inserted in pairs at the orifice of the calyx tube in an alternisepalous position just above the outer staminal whorl. Two decandrum the clavate petals are united at their bases by an inconspicu- ous annulus; (2) in L. pauciflorum and L. persimile the clavate or liguli- form petals are free at their bases. In all species the petals are glabrous and erect. The shape of the petal apex has been found to be a useful taxonomic character. It is dependent on differential growth patterns (particularly differential elongation of separate whorls). The elongation patterns are reflected in the relative position of petals, anthers and stigma in the 1961] NEVLING, REVISION OF LINOSTOMA 303 mature flower. In this respect two exsertion orders are found: (1) stigma, anthers and petals in Linostoma decandrum and L. pauciflorum (petal apex obtuse or irregular); (2) petals, stigma and anthers in L. persimile (petal apex acute). The distortion of the petal apex in Linostoma decandrum occurs as the pairs of petals develop in the alternisepalous position. In the bud their apices become appressed to the base of the developing alternisepalous anther. Elongation of the petal apparently occurs at a slightly faster rate than that of the developing staminal filament, with the result that each petal apex clasps one of the lower lateral margins of the alternisepalous anther, thus deforming the petal apex. Although the filaments subse- quently elongate and the petals undergo further elongation, the apices never recover from the initial contact with the anther. As a result, the petal apex is often cup-shaped or truncated with one lateral margin (that toward the calyx lobe) noticeably longer than the other. This same situa- tion is found in L. pauciflorum to a much lesser degree, with the result that the petal apices are never as strikingly deformed as in L. decandrum. The acute petal apex of L. persimile is due to the lack of interference of petal elongation by the androecium. It is impossible to say whether this condition is advanced or primitive in relationship to that found in L. decandrum and L. pauciflorum. In Linostoma persimile and L. decandrum vascular tissue is lacking in the petals. This is really quite surprising, for petal length in these species is 3.5 and 3.5—6 mm. respectively. It is difficult to understand why fleshy structures of this size require no vascularization. In L. decandrum the petals are inserted between the midrib supplying the calyx lobe and the lateral veins, which at this level have already divided several times. It does not appear to be associated more intimately with the lateral veins than with the midrib as might be expected. The petals of Linostoma pauciflorum are vascularized by a single trace. I have observed an exception in one petal in which the trace divided near the base and two well-developed traces were formed. I have been unable to determine the connection of the petal trace with a calyx tube trace. The trace is usually 5—10 tracheids in diameter, and these tracheids have annular to helical thickenings. Toward the petal apex the tracheids be- come noticeably shorter and the trace finally ends blindly. The androecium consists of ten stamens inserted on the calyx tube just below the petals. The stamens are inserted either in two whorls (L. per- simile and L. pauciflorum) or in a single whorl (L. decandrum). The five stamens of the upper whorl are in the antisepalous position, while those of the lower whorl are alternisepalous. The single whorl of L. decandrum is in reality two closely intermeshed whorls, for the traces supplying the alternisepalous stamens depart from the lateral commissural calyx traces (in respect to the position of these in the calyx lobe) some distance below those of the antisepalous stamens. The traces supplying the antisepalous stamens are continuous with those of the median calyx traces. 304 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XLII Text Fic. III. Camera lucida drawings of tg see of stipe and ovary of Linostoma. 1, L. decandrum (Parkinson 316); 2, L. persimile (Garrett 1436) ; 3-5, L. pauciflorum (Ridley 1868, Hallier B.2261, itatiland heheh Stippled por- tion is the approximate area occupied by the ovule; all v fat bundles, with the exception of the hook-shaped ovular trace in figs. —5, are in the ovary wall; only the bundles in the ovary wall facing the viewer are represented. The filaments are filiform, and those of the antisepalous stamens are usually longer than those of the alternisepalous ones. They are glabrous, except in a single anomalous flower in which they are hispid. Vasculariza- tion is by a single vascular bundle which either ends blindly somewhat below the connective tissue or enters only the lower half. The anthers are exserted, except for those in the alternisepalous whorl of Linostoma persimile. They are 4-lobed and 4-loculed in cross section. Dehiscence is longitudinal, The connective is basally inflated in L. pauci- florum, and to a lesser degree in L. decandrum, so that the pollen sacs are displaced to its lateral margins. In L. persimile there is no basal in- flation but the connective is produced beyond the pollen sacs. The pollen of the three species is, for all practical purposes, uniform. It is polyporate and highly sculptured. The grains vary between 40 and 60 micra in diameter, and, although an insufficient number have been 1961 | NEVLING, REVISION OF LINOSTOMA 305 measured, it appears that those of Linostoma persimile are the smallest and those of L. pauciflorum are the largest The disc is borne at the base of the calyx tube and closely surrounds the stipe supporting the ovary; in many previous descriptions the disc has been considered to be absent in the genus. It consists of irregular lobes which are free or connate and sometimes adnate to the calyx tube. The number of lobes is extremely variable within flowers of the same in- florescence and therefore is of little taxonomic consequence. Externally the disc appears fleshy and is glabrous, except in Linostoma decandrum in which it may support a few trichomes. Since I have had little success in clearing the disc, which scarcely attains 0. if mm. in height, I have been unable to (termine the vascularization, if a The gynoecium is composed of a siniele ok oresumably of the pseudo- monomeric type, i.e., one fertile and one sterile carpel. It is borne on a stipe (gynophore) ahh varies in length among the species. In the specific descriptions in the taxonomic treatment the length of stipe and ovary are combined because of the gradual transition between them. The ovary is superior, ellipsoid to obovoid, and unilocular. It is usu- ally densely sericeous, although occasionally sparsely so in Linostoma pauciflorum. The trichomes are unicellular, erect, and quite shiny (at least in herbarium specimens). The pattern of vascular supply to the ovary seems quite distinct among the species (see Text Fic. III). Linos- toma pauciflorum is immediately distinguishable by the well-developed ovular trace (further discussed under the species), the presence of only two traces at the base of the stipe, and few lateral traces. Linostoma persimile can be distinguished from L. decandrum by the more reticulate secondary venation arising from the main veins. The apparent lack of an ovular trace in the last two mentioned species is curious and requires fur- ther investigation. In all species the single, anatropous, two-integumented ovule is semipendulous from the upper locule wall. The style is borne terminally, although it sometimes appears to be eccen- tric, in which event the trichomes must be removed before its true position can be seen. It is filiform and is vascularized by two relatively large traces which stand opposite each other. The style is generally glabrous, except in a few specimens of Linostoma pauciflorum where a few trichomes can be found. The stigma is capitate or mammiform. (See also discussion following L. persimile.) The flowers of all three species are pedicellate. The junction of pedicel and secondary peduncle is clearly marked by an articulation. In the area of articulation parenchyma cells may be formed at the expense of other tissues. A longitudinal section through the point of articulation in Linos- toma decandrum has been figured by Leandri (Fig. 23, 3,4. 1930). In general, the fruits are preserved very poorly and I do not feel quali- fied to comment upon them at this time. The longitudinally ruptured calyx tube is persistent in Linostoma decandrum and can be seen at the base of the fruit. In L. persimile and L. pauciflorum the calyx tube is accres- cent and surrounds the fruit. 306 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII GEOGRAPHY The three species have a composite range extending from India to Indo- china, the Malay Peninsula, and the island of Borneo. The distribution is not complicated because of the small number of species. Distributions of the individual species have been plotted on Map I, which is a part of one of Goode’s Series of Base Maps published by the University of Chicago Press. The symbols indicate places of collection only and are in no way indicative of the relative frequency of the plants. At the present time too little is known of the ecological situations in which the plants grow to give any rational discussion but in only one species, Linostoma decandrum, apes the distribution appear to be unusual. The disjunction shown by this pecies is, in all probability, a real one with a lack of suitable habitats es the two areas. There seems to be no morphological basis for the erection of any infraspecific categories in this case. MATERIALS This revision is based on specimens from the following herbaria, the abbreviations for which are taken from Lanjouw & Stafleu, /ndex Her- bariorum, Part I, Ed. 4 (Regnum Vegetabile 15. 1959). A Arnold Arboretum of Harvard University, Cambridge BM British Museum, Lon 8 Botanical Museum and Herbarium, Copenhagen G Conservatoire et Jardin botaniques, Genéve GH Gray Herbarium of Harvard University, hiimtinee K Herbarium, Royal Botanic Gardens, Kew i Rijksherbarium, Leiden M Botanische Staatssammlung, Munchen P Muséum National d’Histoire Naturelle, Paris U Botanical Museum and Herbarium, Utrecht US U. S. National Museum, Smithsonian Institution, Washington Ww Naturhistorisches Museum, Wien I wish to take this opportunity to thank the directors and curators of the above institutions for the many courtesies extended to me. This revi- sion would not be possible without their co-operation. I also wish to thank my colleagues of the Arnold Arboretum and Gray Herbarium for helpful suggestions, technical assistance, and constant encouragement. TAXONOMY Linostoma Wall. ex Endl. Gen. 331. 1837 (Type: L. decandrum (Roxb.) Wall. ex Endl.). Linostoma Wall. Cat. n. 4203. 1831, sine descript. Linostoma sect. Eulinostoma Meissn. in Mart. Fl. Bras. 5(1): 72. 1855 (Type: L. decandrum (Roxb.) Wall. ex. Endl.). 1961] NEVLING, REVISION OF LINOSTOMA 307 Linostoma subg. Nectandra Kurz, Jour. As. Soc. Bengal 39(2): 83. 1870 (Type: L. decandrum (Roxb.) Wall. ex. Endl.). Trees, shrubs or lianas, often scrambling, the stem monopodial, al- though frequently dichotomizing or trichotomizing, the axillary branches sometimes modified for climbing, the bark containing many fibers, becom- ing rugose with drying. Leaves opposite (or sometimes approximate) to alternate, simple, pinnately veined, entire, petiolate, estipulate. Inflores- cences borne from the terminal portions of young stems, subracemiform, umbelliform, or rarely the flowers solitary, the primary peduncle bearing a pair of bracts and a single bracteole. Flowers bisexual, pentamerous, peri- gynous; calyx tube urceolate, ellipsoid or tubular, nonarticulated, ribbed, red, yellow, greenish white or white, glabrous without, glabrous or variously pubescent within; calyx lobes 5, subequal, quincuncial, erect to reflexed at anthesis; petals 10, ligulate to clavate, sometimes with a minute an- nulus at the base, glabrous; stamens 10, inserted in 1 or 2 whorls, the upper whorl antisepalous, the lower whorl alternisepalous, the anthers filamented, basifixed, longitudinally dehiscent, the connective sometimes produced beyond the pollen sacs or inflated; disc of irregular lobes, some- times basally connate, sometimes adnate to the calyx tube; gynoecium single, pseudomonomeric, superior, variously stipitate, unilocular with a single anatropous ovule, the style terminal, the stigma capitate or mammi- form. Fruit drupaceous, the calyx tube sometimes persistent or accrescent. KEY TO THE SPECIES a. Leaves to 10 cm. long, thick-coriaceous, the apex acute, acuminate or sub- caudate; inflorescences 3-12-flowered; calyx tube villous within; calyx lobes generally as long as or longer than the tube, strongly reflexed: calyx per- sistent. but not accreseent in-fruit. .....54.%.i4.e0<00e es 1. L. decandrum. a. Leaves to 5 cm. long, thin-coriaceous, the apex mucronulate and acute, ob- tuse, nee or retuse; inflorescences 1-—6-flowered; calyx tube villous or glabrous within; calyx lobes at most half as long as the tube, erect or spread- ing; calyx re chat in fruit. b. Bracts sunepposite, borne at the middle of the primary peduncle or below; calyx lobes erect; petals exserted beyond the stigma and anthers; connective produced beyond the pollen sacs but never inflated; style Oo Oe ee len a fe ns Ad OS 2. L. pe ersimile. b. Bracts opposite to alternate, at least one of the pair borne near the sum- mit of the primary peduncle; calyx lobes oe stigma and anthers exserted beyond the petals; connective inflated so as to separate the pollen sacs at least basally but never produced ed the pollen sacs; style“6-20 mim: JONg. 25. c.dsi cs pedaw ewer ewes dancer Se a paucifiorum. 1. Linostoma decandrum (Roxb.) Wall. ex Endl. Gen. 331. 1837. Nectandra oe poe Hort. Beng. 90. 1814 (Type: Roxburgh s.n.!); Fl. Ind. 2: 425. Linostoma as (Roxb.) Wall. Cat. n. 4203. 1831, sine descript. 308 JOURNAL OF THE ARNOLD ARBORETUM [VoL XLII Subscandent to scandent shrubs or lianas (?) to 15 m. tall; young stems terete, dark reddish brown, glabrous except for a few short trichomes in the leaf axils, the lenticels round, becoming horizontally elongate; axillary branches normal or rarely modified and unciform, the unciform branches usually conspicuously swollen basally. Leaf blade lanceolate or elliptic to oblong-elliptic, 3.5-10 cm. long, 1.5—3.5 cm. broad, acute or acuminate to subcaudate at the apex, cuneate to obtuse at the base, thick- coriaceous, glabrous, darker above than beneath, the costa immersed above, elevated beneath, the primary lateral veins many, inconspicuous, parallel and more or less straight, the submarginal vein well developed, almost coinciding with the margin; petiole canaliculate, rugose, glabrous, 3-5 mm. long. Inflorescences borne terminally on the young shoots, generally involving several nodes, often compound; each inflorescence 3—7(-12)-flowered, subracemiform to rarely umbelliform, glabrous, the primary peduncle 2.7—5.5 cm. long, often with the reduced leaf of the subtending node near the base, the rachis 1-5 mm. long, the secondary peduncles 2—7 mm. long, often distally dilated, the bracts opposite or subopposite, borne at the middle of the primary peduncle or below, ovate, elliptic or oblong-elliptic, 2.5-4.5 cm. long, 1-2 cm. broad, obtuse to acute at the apex, truncate to rounded at the base, almost membranaceous, glabrous, creamy white, venation much more lax than that of the leaves, I . Distribution of the species of Linostoma: dots, L. decandrum; tri- angles, L. persimile; squares, L. pauciflorum. 1961] NEVLING, REVISION OF LINOSTOMA 309 subsessile, the bracteole borne at the summit of the primary peduncle, narrowly linear, 0.3-1 cm. long, caducous. Calyx tube fusiform, 5—7(—13) mm. long, 2—2.5 mm. in diameter at the orifice, glabrous and reddish without, villous and white within but often glabrous in the lower half, the trichomes protruding from the orifice; calyx lobes linguiform, 6.5—8.5 mm. long, 2—2.5 mm. broad, as long as or longer than the calyx tube, glabrous, strongly reflexed; petals inserted at the orifice, clavate, erect, fleshy, exserted, 3.5-6 mm. long, apex irregular, their bases connate by means of an obscure annulus; stamens exserted beyond the petals, inserted in a single whorl immediately below the petals, the filaments filiform, glabrous, the antisepalous filaments 6.5—9 mm. long and slightly longer than the alternisepalous ones, the alternisepalous filaments 4—8 mm. long, the anthers oblong, about 1 mm. long, 0.75 mm. broad, the connective basally inflated but not produced beyond the pollen sacs, the pollen sacs appearing confluent and hippocrepiform following anthesis; disc of small, irregularly shaped lobes, more or less adnate to the calyx tube, less than 0.5 mm. tall, each lobe glabrous or more often bearing a num- ber of short trichomes; stipe and ovary ellipsoid, about 3 mm. tall, densely sericeous, the style filiform, 8-11 mm. long, sericeous at the base, other- wise glabrous, the stigma mammiform, exserted beyond the anthers, about 0.75 mm. in diameter; pedicel 3-4.5 mm. long, glabrous. Fruit (imma- ture?) ovoid, 8-12 mm. long, 4-8 mm. in diameter, sericeous, dull red, the longitudinally and unilaterally ruptured calyx tube more or less per- sistent at the base; pedicel 5-7 mm. long, somewhat swollen. aera ee bere Bot. Ges. Regensburg 3: ¢. 7. 1841; lon, Hist. Pl. 4. figs. 72, 73. 1877; Pflanzenfam. III. 6a: 232. 82. A-E, oe ene Fl. Gen . Indo- Chine 5: 173, fig. 16. 1915; Bibl Bot. 27(111): t.2, fig. 11. 1934. DistTRIBUTION. Eastern India, East Pakistan, Burma, Thailand, Viet- m, Laos, and Cambodia, the plants associated with evergreen scrub or forests, Pea at altitudes up to 200 meters. Flowering specimens col- lected in January, April, June, November and December, specimens in fruit collected in December and February. Roxburgh (FI. Ind. 2: 425. 1832) says that the plants flower in October and the seeds ripen in Janu- ary, February and March. India. West BENGAL: Calcutta, (cultivated?), Gaudichaud 324 (GH). MANI- PuR: Piphima, Meebold 5164 (c). Assam: Sialsuk, Lushai Hills, Parry 410 (kx). East Pakistan. EAst BENGAL: Sylhet, Wallich 4203 (BM, G, L, M), Rox- burgh “1813” (BM, G), “1830” (kK), without collector, (G, L, w); Chittagong Division, Cowan s.n. (us); Chittagong ye Dr. sig collector 106 (us), 220 (G, L), Hooker & Thomson sn. (BM, GH, K, L, M, P, U, W). Burma. KacHIN: Myitkyina District, Tagwin, Bee oe (kK); Myitkyina District, hills w. of Hopin, Parincon 347 (Kk); Bhamo, Cubitt 1047 (A). Thailand. CuHonsurI: Chonburi, on trail between Mapkhla and Thapsai, Smitinand 3637 (L); Khow Diu near Sriracha, Collins A (BM), 1797 (BM, Us). NAKON Rat- CHASIMA: Hui Reng, Kerat (Korat), Kerr 17588 (BM); Kao Lem, Korat, Put 3573 (pM). Rayonc: Ban Pe, Rayawng (Rayong), Put 2763 (BM). Vietnam. 310 JOURNAL OF THE ARNOLD ARBORETUM [voL, XLII ANNAM: Nha Trang, Alleizette sn. (L); Tourane, Clemens & Clemens 3465 (p); Col de Braian, Pnom-sapoum, by Blao, Poilane 23698 (kK). Laos. CHAM- PASSAK: Bassac, Thorel 2612 (Pp). Cambodia. Insula Phu Quoc, Pierre 1470 (Pp), 19794 (pm). Without Locality. Wallich 2247 (c), Bruce s.n. (w) There has been considerable confusion and error as to the proper cita- tion of authorship of the genus and of this, the type species. The generic name was first used by Wallich in 1831 (Cat. n. 4203); this publication consists of the name Linostoma decandrum and the citation “Hort. Beng.” From this scrap of information it is evident that Wallich intended a “new combination” based on Nectandra decandra Roxburgh (Hort. Beng. 90. 1814). Roxburgh included here, in addition to the epithet, the common name, the place of collection, a two-symbol description and the time of flowering. It was not until Endlicher in 1837 (Gen. 331) published a Latin diagnosis for Linostoma that either the genus or Wallich’s “com- bination” became valid. The citation of the type species thus should be Linostoma decandrum (Roxburgh) Wallich ex Endlicher. As a result of Roxburgh’s publication (1814) he is given credit or rather discredit for an illegitimate genus (Nectandra Roxb. non alior) which is usually cited in the generic synonymy. However, in a subsequent publication (Fl. Ind. 2: 425. 1832) he maintains (despite Wallich’s com- bination of which he probably was not aware) the name Nectandra de- candra, furnishes generic and specific descriptions and attributes the genus to Jussieu! Indeed, Jussieu, in an earlier work (Gen. Pl. 86. 1791), had published an account of the genus Nectandra which was based, in turn, on Nectandra Berg. and which included a generic description and a single species based on Gnidia sericea L. The illegitimate genus Nectandra Berg. (Descript. Pl. Cap. 131. 1767) is composed of three previously de- scribed 4-merous species which are not congeneric with our 5-merous Linostoma decandrum. Thus Roxburgh did not describe a new genus but merely expanded the concept of a previously existing one. That he was in error is of little consequence. Since the citation in generic synonymy of all genera in which species of a particular genus have been misplaced leads to absurdity, these misplacements are adequately and properly cited within the specific synonymy. There is additional confusion concerning the type specimen. To the best of my knowledge a type has never been selected, although most authors in discussing this species include the citation of Wallich 4203. As a result of examining a number of specimens, all labeled Wallich 4203, I have come to the conclusion that these collections are a heterogeneous group. Four I believe to be authentic Wallich collections: Munich with a penciled label “4203,” probably written by Wallich; Leiden with a pen- ciled “4203”; British Museurn with the penciled label ‘4203, Wallich Nov. 1829,” also probably written by Wallich; and finally a Geneva speci- men with a printed label ‘“Wallich 1829.” There are a number of other specimens which may also be Wallich collections, for they bear the ‘“Cata- logue” number, but I suspect that these specimens were either collected after the publication of the “Catalogue” or that the number was added 1961] NEVLING, REVISION OF LINOSTOMA 311 to earlier collections as a means of specific identification, regardless of the collector. Some examples are specimens at Leiden and Geneva which bear the “Catalogue” number, but lack collection data, and a specimen from Kew which bears the ‘‘Catalogue” number, but the date 1830. There is also a collection at Vienna which bears a printed label upon which in red ink the “Catalogue” number, epithets, place of collection and collector (H. Bruce) have been added. Therefore, it seems that considerable caution must be used in evaluating specimens bearing the “Catalogue” number. However, in spite of past inference, the Wallich collections do not con- tain the type among them! The type must be selected from the collections of Roxburgh made prior to 1814. I have seen two specimens which prob- ably qualify: one from the Geneva herbarium with a handwritten label, “Nectandra decandra Roxb., Ex Sylhet montibus,” and, on the same label and written at a later date in another hand, ‘“‘Linostoma dicandrum. Wal- lich”; a second specimen from the British Museum bears a label written in ink, almost certainly by Roxburgh, “Nectandra decandra probably a Dais, or in all events a Thymeleae.”” Added below in pencil is the nota- tion ‘““Roxburgh 1813.’ Inasmuch as the British Museum specimen ap- pears to be best qualified I select it as the lectotype. It is composed of two flowering branchlets and a packet containing inflorescences, flowers and very immature fruit. 2. Linostoma persimile Craib, Kew Bull. 1911: 452.1911 (Type: Kerr 814!). Shrubs to 1.5 m. tall; young stems slender, terete, light brown to red- dish brown, glabrous, the lenticels inconspicuous; axillary branches nor- mal or sometimes slightly recurved. Leaf blades rotund, broadly elliptic to obovate, 2-3 cm. long, (0.8—)1—2.5 cm. broad, atesoaulite and ob- tuse, truncate, or retuse at the apex, cuneate at the base, thin-coriaceous, glabrous, darker above than beneath, the costa immersed above, elevated beneath, the primary lateral veins many, parallel and more or less straight, the submarginal vein inconspicuous, almost coinciding with the margin; petiole shallowly canaliculate, slightly rugose, glabrous, 1-2 mm. long. Inflorescence borne terminally on the young shoots, involving a few to sev- eral nodes, sometimes compound; each inflorescence 3—6-flowered, um- belliform, glabrous, the primary peduncle 2—4 cm. long, often bearing the reduced leaf of the subtending node near the base, the rachis about 1 mm. long, the secondary peduncles at most 0.5 mm. long, distally dilated, the bracts subopposite, borne at the middle of the primary peduncle or below, ovate to elliptic, 2-3 cm. long, 1-2 cm. broad, acute and mucronulate at the apex, obtuse to subcordate at the base, mem- branaceous, glabrous, with a venation more lax than that of the foliage leaves, the bracteole borne at the summit of the primary peduncle, linear, about 1 mm. long, caducous. Calyx tube tubular, 8-10 mm. long, 2—3.5 mm. in diameter at the orifice, glabrous without, glabrous within, yellow; calyx lobes linguiform, 4—4.5 mm. long, 1.5-2.5 mm broad, much shorter than the calyx tube, glabrous, erect or slightly spreading; petals inserted 312 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII at the orifice, liguliform, erect, exserted beyond the anthers and the stigma, 3-3.5 mm. long, gradually tapering to the acute apex, fleshy; stamens inserted in two distinct whorls, the filaments filiform, glabrous, the anti- sepalous filaments about 1 mm. long, the alternisepalous filaments to 0.5 mm. long, the anthers oblong, 0.75—1 mm. long, 0.5 mm. broad, the con- nective obtusely produced beyond the pollen sacs, the upper antisepalous filaments inserted just below the petals, exserted, the lower alterni- sepalous filaments inserted about an anther’s length below the anti- sepalous ones, included; disc of minute irregular lobes, adnate to the calyx tube, glabrous; stipe and ovary ellipsoid, about 2 mm. long, 0.75 mm. in diameter, densely sericeous, the style filiform, 6—-7.5 mm. long, glabrous, the stigma capitate, exserted beyond the stamens; pedicel 1.5—3 mm. long. Fruit accrescent, surrounded by the persistent calyx, the calyx more or less narrowly lageniform, 11-13 mm. long, 5-7 mm. in diameter, the lobes erect, the drupe broadly fusiform, 7-9 mm. long, 6-7 mm. in diameter, sericeous at least towards the apex; pedicel slightly swollen. DistRIBUTION. Burma and Thailand. Collected at altitudes of from less than 50 meters to 1650 meters. Found in deciduous forest or open “pine jungle.” The collections indicate flowering in September, Novem- ber, and December, with fruiting known in November and December. Burma. KaARrENNI: Karen Country and Hills, Kurz 990 (Mm). Thailand. CHIENGMAI: Camp Hoi Chan Kiang, Doi Sootep, Rock 123 (us); summit o Doi Chom Cheng, Doi Sootep, Rock 1173 (us); Doi Sootep, Garrett 1397 (A, L, US), 1436 (kK, L), Kerr 814 (x-holotype, BM), sm. (BM). Without precise locality: Tung Quang, Petchabuni (?), Kerr 20584 (BM), Marian 2759 (xk). The collections Kerr 20584 and Marian 2759 have not been mapped for I am not certain of the spelling of the locality on the handwritten labels, which read either “Petchabuni” or ‘Petchaburi.” Since a distance of about 250 miles separates these two cities, it seems best not to make a judgment at this time. If an itinerary of Kerr’s travels exists, the location could probably be determined. The holotype has with it a packet containing a rather unusual flower. This flower, which is nearly mature, has two petals (both lateral with respect to the single calyx lobe which they flank) which have become connate below by their lateral margins. In addition, these fused petals have produced a smaller subsidiary petal between them. The stamen standing in front of this petalaceous anomaly is unusual in that the fila- ment is almost hispid, in striking contrast to the glabrous filaments of the remaining nine stamens. The position of the stamen, i.e., standing within this petalaceous anomaly, provides evidence by its gross morphologi- cal position, of the true petalaceous nature of this structure. A number of flowers in several collections display an included stigma. of mechanism for the prevention of self-pollination, specifically, proter- andry. However, Burrows (1960) has described a somewhat comparable 1961 | NEVLING, REVISION OF LINOSTOMA a55 mechanism in the genus Pimelea and his observations and conclusions may be applicable to this species. In essence, he found that in bisexual flowers early pollination inhibited elongation of the style, with the result that fertilization may take place and the style remained included, late pollination inhibited elongation to a lesser degree, even though fertilization could no longer occur, and, finally, the lack of pollination allowed the style to elongate greatly and thus the stigma to become well exserted. 3. Linostoma pauciflorum Griffith, Calcutta Jour. Nat. Hist. 4: 234. 1844 (Type: Griffith 4376!). Linostoma leucodipterum Hallier f. Med. Rijksherb. 44: 28. 1922 (Type: Hallier f. B.2261!). Linostoma longiflorum Hallier {. ibid. 29 (Type: Haviland 1759!). Shrubs, trees, or climbers, to 25 m. tall; young stems filiform to slen- der, terete, light to dark brown, glabrous except for a tuft of short trichomes in the leaf axils, the lenticels inconspicuous; axillary branches normal or gently to strongly uncinate, becoming swollen at the base. Leaf blade ovate, lanceolate, narrowly to broadly elliptic or oblong- elliptic, 1-5 cm. long, 0.5-2.5 cm. broad, mucronate and acute, obtuse, truncate or retuse at the apex, cuneate to obtuse at the base, thin- coriaceous, glabrous, but sometimes appearing farinose beneath, darker above than beneath, the costa plane to immersed above, plane to ele- vated beneath, the primary lateral veins numerous, more or less straight and parallel, inconspicuous, the submarginal vein well developed, coin- ciding with the margin; petiole canaliculate (although sometimes scarcely so), 1-3 mm. long, rugose, glabrous. Inflorescences borne from the axils of the terminal portions of the young stems, simple or compound, gla- brous except for occasional trichomes near the summit of the primary peduncle; each inflorescence 1—5-flowered, umbelliform, the primary peduncle 2—18(-30) mm. long, often with the reduced leaf of the sub- tending node near the base, the rachis to 0.5 mm. long, the secondary peduncles obsolete, to 0.5 mm. long, somewhat dilated distally, the bracts borne from the middle to the summit of the primary peduncle, opposite, subopposite, or alternate, foliaceous or reduced and bracteolate, linear, lanceolate, ovate, elliptic or oblanceolate, 2-35 mm. long, 2-1 m. broad, mucronulate and acute, obtuse, or truncate at the apex, acute, obtuse, or subcordate at the base, membranaceous, glabrous, white, the venation more lax than that of the foliage leaves, the bracteole borne at the summit of the primary peduncle, linear, to 10 mm. long, some- times caducous, or extremely reduced and represented by a small lump of tissue in the axil of which stand a few trichomes. Calyx tube tubular to ellipsoid, 7-25 mm. long, about 0.5 mm. in diameter at the orifice, glabrous and green to white without, glabrous or partially to completely villous within, a few trichomes exserted beyond the orifice; calyx lobes lanceolate or elliptic to oblanceolate, 3-12 mm. long, 1.5—3 mm. broad, usually glabrous, spreading; petals inserted at the orifice, narrowly clavate, 314 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII fleshy, exserted, erect, 5.5-7.5 mm. long, obtuse to irregular at the apex, glabrous; stamens inserted in two close whorls just below the petals, ex- serted beyond the petals, the filaments filiform but enlarging distally, 4.5-10 mm. long, glabrous, the antisepalous filaments somewhat longer than the alternisepalous ones, the anthers more or less oblong, 0.5—1 mm. long, 0.5—0.75 mm. broad, the connective enlarged so as to displace the pollen sacs to its lateral margins but not produced beyond the pollen sacs; disc annular, with irregular lobes, 0.25—-0.5 mm. high, glabrous; ovary and stipe oblong-obovoid or obovoid, 3—7.5 mm. long, sparsely to densely sericeous, the style filiform, 6-20 mm. long, glabrous or with occasional trichomes, the stigma capitate, exserted beyond the stamens; pedicel 5—8 mm. long. Calyx tube accrescent in fruit, ellipsoid but tapered more sharply above than below, 12-15 mm. long, 6—7 mm. in diameter, the drupe ovoid. ILLUSTRATION. Flora Malesiana I. 6: 26. fig. 11. 1960 (as L. longiflorum and L. pauciflorum). DistRIBUTION. Thailand, south to Singapore and east to Borneo; col- lected from sea level to 200 meters. Flowering in March, April, June, August, September, and October; collected in fruit in December. Found in evergreen scrub, wastelands, and primary peat forests; the Hillard col- lection is from ‘‘undergrowth on rubber estate.” According to Kerr the root is sometimes used for poisoning fish. Thailand. Pattranr: Yala, Kerr 7258 (pm); Kok Po, Kerr 15073 (BM). SONGKHLA: Hat Zai, Kerr 13667 (BM). SURATTHANI: Surat, Kerr 11302 (Bm), Kerr 11820A (Bm). LocALiIty UNCERTAIN: Tako, Qungsuam (?), Kerr 11870 (pm). Malaya. Kepan: Langkawi Islands, Curtis s.n. (ep), Hillard KL1411 (a); Langkawi, Palo Lingsong (?), Curtis, 1890 (sm). PERAK: Pangkore Island, Scortechini 983 (K), Burkill & Shah HMB 240 (x, L). Pertis: Chapnema (?), Ridley 14916 (sm). YaAta: Pulau Penang, Telok Bahang, Toi, 1905 (BM), Sinclair SFN 39304 (1). WitHout LocaLity: Maingay 1308/3 (L). Singapore. Griffith 4376 (x-holotype of L. pauciflorum); Changi, Ridley 1858 (BM); Botanic Gardens, Kiah SFN 37931 (BM, L). Sarawak. Kuching, Haviland 1759, Oct. 1892 (1-holotype of L. longiflorum; BM, GH, K), Brooke 9696 (G, L); Binatang, J.A.R. Anderson 9047 (L); Tanjong Po, Brooke 10603 (G); near Long Kapa, Mount Dulit (Ulu Tinjar), Richards 1312 (A, Kk). British North Borneo. Sandakan and vicinity, Ramos 9 (a, us); Gaya Island, Haviland 1440 (kK). Borneo. W. Borneo: ect Sungai Déjméla und Gantng K(e)lamm, Hallier B.2261 (isotypes of L. leucodipterum, L, P). My treatment of this group of specimens as a single species requires graphically, the specimens can be divided into two groups, one restricted to the island of Borneo and the other to peninsular Malaya. I shall dis- cuss the relationship of the Bornean specimens to each other and to peninsular specimens (when of some significance). Hou (1960) has recognized two species from Borneo, Linostoma pauci- florum and L. longiflorum. His key to these species is based upon differ- 1961] NEVLING, REVISION OF LINOSTOMA 315 ences in the bracts, particularly the size and the degree to which they cover the lower half of the calyx tube, as well as the length of the calyx tube. On the basis of his key I have seen two collections which clearly can be placed with L. longiflorum; they are Haviland 1759 (type at LL. longiflorum) and Anderson 9047. In addition, Jahri 166 may belong with this unit, but the Leiden specimen examined lacks inflorescences. It has been annotated by Hou as L. longiflorum, and he may have seen a more complete specimen from the Bogor herbarium. The leaves of the two “species” are extremely variable in shape and size, although the leaves of Haviland 1759 tend to be the broadest of all Bornean collections (but no broader than those found on some of the peninsular specimens). There appears to be no certain way to separate two species on the basis of either leaf shape or size. As already indicated in the introductory comments, the structure of the leaf is rather constant throughout the genus, but several character- istics in this connection are of interest. The stomata of this species are similar to those of Linostoma persimile but quite different from those of L. decandrum. The lower leaf surface often appears farinose under mag- nification. This phenomenon is due to the papillate shape of the outer periclinal wall of individual cells of the lower epidermis. This condition can be demonstrated to a greater or lesser degree in most of the collections, either peninsular or insular and including the type of L. longiflorum, but is not found elsewhere in the genus. Finally, it must be mentioned that Hallier in the original description of L. longiflorum made reference to a fungal infection in the type. It was thus a surprise to find the leaves of Anderson 9047 similarly affected. None of the Bornean collections refer- able to L. pauciflorum (by Hou’s system) appears, on the basis of cleared leaves, to be infected, and to draw a conclusion is almost irresistible. The basic structure and size of the inflorescence is essentially the same in both “species,” although there are certain differences which require discussion. In Linostoma longiflorum the number of flowers per inflores- cence is one, whereas in L. pauciflorum it is rarely one, generally two (Borneo), and sometimes as many as five (peninsular). There does not seem to be any real correlation between the number of flowers per in- florescence and their size as one might expect to find. The position of the bracts near the summit of the primary peduncle is the same in all Bornean specimens. In the peninsular specimens, in which the bracts are alternate, one of the pair is borne at the summit of the primary peduncle. This distal position of the bracts gains importance because it is unique within the genus (other species having the bracts borne at the middle of the primary peduncle or below). The importance of the bracts themselves is magnified by their use as a “key” character. In Linostoma longiflorum the bracts vary in length from 2-12 mm. and their shape is usually linear or linear-lanceolate. The bracts of specimens referable to this species which I have cleared seem to have a poorly differentiated vascular system, perhaps an indication that growth of the bract had ceased before attainment of “mature” form. The 316 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII bracteole, in contrast, is well developed. In L. pauciflorum the bracts vary from 8-35 mm. in length, and the shape, although variable, is con- siderably broader than that found in L. longiflorum. The bracteole, in contrast, is poorly developed and is sometimes reduced to only a small lump of tissue in the axil of which stands a tuft of trichomes. It is tempt- ing to consider a type of developmental anomaly which indiscriminately “stunts” the growth of the bracts (L. longiflorum) on the one hand and growth of the bracteole (L. pauciflorum) on the other. 990008 2 Text Fic. IV. Variation in bract size and shape in two collections of Lino- — pty pe 1, Hallier B.2261, from Borneo; 2, Haviland 1759, from iwak (Natural size) Outlines of a series of bracts showing extremes in variation from two collections representing Linostoma longiflorum (Haviland 1759) and L. pauciflorum (Hallier B.2261) are presented in Text Fic. IV. It is of more than casual interest that a size overlap occurs and that the largest bract in the Haviland collection is quite similar in shape to the smallest bract found in the Hallier collection. Additional collections from Borneo will be necessary before this problem of size variation and overlap can be evaluated. It should be obvious at this point that the degree to which the bracts cover the base of the calyx tube is determined by the length of the bracts. This does not mean that the key does not work in the area of overlap. There are generally enough bracts on the available speci- mens to indicate strikingly to which “species” the specimen belongs, but I think that it does detract from the validity of this key character. Generally speaking, the vegetative portions of a plant may be expected to be more variable than reproductive structures. To check this general- ity a study of i nae and, to a lesser degree, anatomical character- istics of the flower was made. As in the case of the leaves, the sample is extremely Sai but the floral characteristics showed a striking homo- 1961] NEVLING, REVISION OF LINOSTOMA 317 geneity which cannot be easily dismissed. The calyx tube is the same shape and has the same type of pubescence in all Bornean collections. In flowers of insular specimens the calyx tube is variously villous within and is in sharp contrast to those of the peninsula in which the calyx tube is glabrous within. This difference is not considered taxonomically sig- nificant. The length of the calyx tube (a key character) is variable but it does not have a comparable overlap as found in bract size. In Linostoma pauciflorum it varies from 13-15 mm. in length (insular) and from 7-15 mm. in length (peninsular), whereas in L. longiflorum it varies from 17-25 mm. in length. Tube measurements must be taken from flowers at anthesis as there is considerable elongation of the tube immediately prior to an- thesis. The occasional oblanceolate shape of the calyx lobes is a charac- ter shared by both “species” and is unique within the genus. The margin of the calyx lobe is sometimes furnished with a few very small trichomes in L. pauciflorum (both insular and peninsular) but I have not seen them in the two collections referable to L. longiflorum. The petals of the two are alike in shape, size, insertion, and degree of exsertion. Indeed, in the original description of Linostoma longiflorum Hallier says of petals and stamens, ‘“‘fere ut in sp. praecedente” (L. leuco- dipterum,* synonymous with L. pauciflorum). The most striking point is that the petals are vascularized by a single vascular strand, a condition that is once again unique within the genus (the other species lacking vas- cularization of the petals altogether). Similarities within the androecium are also to be found as indicated by Hallier’s original description of Linostoma longiflorum. The insertion of the stamens and their relative exsertion is the same in all specimens. The peculiar swelling of the filament towards its summit, is common, as is the very expanded connective. (The connective is enlarged to the extent that the pollen sacs are displaced to its lateral margins and one wonders whether or not the dehiscence can be spoken of any longer as introrse.) Examina- tion of the pollen shows similar size and sculpturing of the exine (which is, however, rather constant throughout the genus). The disc is the same type in all specimens. The gynoecium is notable for the very long stipe (longest in the genus) which gradually tapers into the ovary. The pubescence of the stipe and ovary is similar and differs from other species in that the trichomes tend to be shorter and blunter. However, it is in the vascularization of the ovary that the soundest factual omen net is found. That is not to say there is no variation in the specimens of this species, for there cer- tainly is, but it is the kind of variation that is indicative of a basic rela- tionship. A series of camera lucida drawings made from cleared and stained ovaries from Ridley 1858 (representative of peninsular L. pauct- florum), Hallier B.2261 (isotype of L. leucodipterum) and Haviland 1759 (isotype of L. longiflorum) is presented in Text Fic. III, 3-5. In all three examples a well-developed ovular trace is present, in contrast to the very poorly developed or obsolete (?) ovular trace of other species * Holotype deposited at Bogor and not seen by me, isotypes at Leiden and Paris. 318 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII (see Text Fic. III, 7, 2). In addition, the number of accessory traces is notably decreased and a trend of reduction within the series itself seems a remote possibility. In summation, the combination of shared morphological and anatomical characteristics which are often unique within the genus is overwhelming evidence of the conspecific nature of this group. In addition, I think that the dissimilarities noted, which are of a quantitative and not a qualitative type, fall within the range of variation that could be expected within a single species of this family. STATUS UNDETERMINED Psilaea Miquel, Fl. Ind. Bat. Suppl. 355. 1860 (Type: P. dalbergioides). Psilaea dalbergioides Miquel, ibid. (Type: Teysmann ?). This genus and species were based by Miquel on a Teysmann collection from Siboga [sic, Sibolga], Sumatra. Presumably the specimen would be deposited in the herbarium at Utrecht but no record can be found of it there at present. Kurz (1870, p. 82, 83) in a discussion following the description of Linostoma siamense says that he has authentic specimens of the type species (Psilaea dalbergioides) and “cannot see how the species should differ from Linostoma pauciflorum, Griff.’ From his short discussion two important conclusions can be drawn: (1) that Kurz did not distinguish between Linostoma and Enkleia; (2) that he had an authentic specimen during his curatorship at the Royal Botanic Gardens, Calcutta. I have not yet determined whether this specimen is extant at Calcutta. This species has been referred repeatedly to the synonymy of Linostoma pauciflorum. I feel that this action is the result of Kurz’s comments, and that it involved no subsequent investigation. Inasmuch as Kurz did not distinguish between Linostoma and Enkleia I prefer to defer a decision regarding this name until the type is located or authentic specimens from Sumatra may be examined. I have not seen any specimens from Sumatra referable to the genus Linostoma to date, although Hou (1960) gives Simalur as a locality for L. pauciflorum. Linostoma sect. Psilaea (Miq.) Hallier f. Med. Rijksherb. no. 44, 28. Hans Hallier (1922) when treating the genus Linostoma split it into two sections: Eulinostoma Meissner, including only L. decandrum, and Psilaea, presumably based on Miquel’s Psilaea dalbergioides as indicated by parenthetical credit, including Hallier’s two new species L. leucodipterum and L. longiflorum. Although his brief descriptions would fail to hold today and infrageneric categories, with the exception of species, are con- sidered unnecessary in view of the small number of species, his concept was in essence correct. Linostoma decandrum, on the basis of several 1961] NEVLING, REVISION OF LINOSTOMA 319 a iam characteristics, stands isolated from L. persimile and L. pauciflorum Linostoma incertae sedis A collection from Brunei, Ashton BRUN 573, has caused me consider- able concern. It is an immature flowering specimen which is somewhat abnormal due to a fungal infection. The original identification was ‘‘Lino- stoma leucodipterum Hall. f.,” subsequently modified to ‘“Linostoma pauciflorum Griff.” by Hou (annotation, 1958, and Identification Lists of Malaysian Specimens, July, 1960). Superficially it does resemble this species, but closer examination shows some differences, such as general pubescence, position of the inflorescence, number of flowers per inflores- cence, lack of a disc, and pubescence and shape of the ovary. I should not be surprised to find that it represents an undescribed species but believe that any taxonomic designation at this time would be premature. I have, however, taken the liberty of including here a description which the reader will appreciate as being lacking in many respects Climber to 25 m. tall; young stems generally slender, terete, ferrugi- neously tomentulose and glabrescent, the lenticels sa eos axillary branches normal. Leaf blades ovate to lanceolate, 2—4.5 cm. long, 1—2 cm. broad, acute to obtuse at the apex, cuneate to obtuse = the base, chartaceous, glabrous, except for occasional trichomes beneath, darker above than below, the costa immersed above, elevated beneath, the primary lateral veins many, parallel and more or less straight, the submarginal vein well developed, coinciding with the margin; petiole canaliculate, sparsely tomentose, 1-3 mm. long. Inflorescences simple, borne termi- nally on the young stems or axillary branches, (5—) 10—16-flowered, sparsely tomentose, the primary peduncle 2—3 cm. long, the rachis 1-2 mm. long, the secondary peduncles about 0.5 mm. long, the bracts (with fungal in- fection) borne below the middle of the primary peduncle, opposite or subopposite, narrowly elliptic, 2-3 cm. long, 0.5-1 cm. broad, obtuse at the apex, cuneate at the base, membranaceous, sparsely puberulent both surfaces, the bracteole borne at the summit of the primary peduncle, linear, 0.5-2 mm. long. Immature flowers: calyx tube short-tubular, 1-3.5 mm. long, 0.5—1.5 mm. in diameter at the orifice, sparsely strigose without, glabrous within, cream-colored; calyx lobes ovate, 1 mm. long and broad, erect, glabrous within; petals inserted at the orifice, linguiform, 1-2 mm. long, ca. 0.75 mm. broad, acute at the apex, glabrous, exserted beyond the stamens; stamens borne in 2 whorls, the antisepalous whorl inserted just below the petals, exserted, the alternisepalous whorl about an anther’s length below the antisepalous whorl, included, the filaments 0.5—0.75 mm, long, glabrous, the anthers oblong, to 0.5 mm. long and broad; disc absent; ovary ovoid, ca. 1 mm. tall, densely hirsute, style and stigma poorly differ- entiated; pedicel to 4 mm. long. Brunei. Collected in September in primary forest, altitude 50 meters, Andulau F.R., Ashton BRUN 573 (1). 320 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII EXCLUDED TAXA LINOSTOMA subg. INOS TOMA poate Jour. As. Soc. Bengal 39(2): 83. 1870, as to species cited. nor Gri LINOSTOMA sect. LOPHOSTOMA Mein in Mart. Fl. Bras. 5(1): 72. 1855 (Type: L. calophylloides Meissn.) Lophostoma Meissn. LINOSTOMA ALBIFOLIUM Rodr. Nalini 1885-88. ed. 2. 67. t. 20. 1891 = Lopho- — aul (Rodr.) Gilg in Engl. & Prantl Nat. hada III. 6a: 232. dere ANDAMANICA Hutchinson ex C. E. Parkinson, Forest Flora Andaman Isl. 2291. 1923 = Enkleia sp. LINOSTOMA CALOPHYLLOIDES Meissn. in Mart. Fl. Bras. 5(1): 72. 1855 = hostoma calophylloides (Meissn.) Meissn. in DC. Prodr, 14: 600, 1857. LINOSTOMA DINIzII (Huber ex Ducke) Lemee, Fl. Guyan. Fr. 3: 108, 1953 = Lophostoma dinizii Huber ex Ducke, Arch. Jard. Bot. Rio de Janeiro 1: 51. oo SCANDENS (Endl.) Kurz, Jour. As. Soc. Beng. 39(2): 83. 1870 = Enkleia ‘oven 3 Lane DENS var. CAMBODIANA Lecomte, Nat. Syst. 3: 127. 1915. (Type: Pierre 511!) = Enkleia sp. LINOSTOMA SIAMENSE Kurz, Jour. As. Soc. Beng. 39(2): 82. 1870 (Type: Teysmann 5986!) = Enkleia sp. LINOSTOMA THORELIT Lecomte, Not. Syst. Paris 3: 127. 1915 (Type: Thorel 2823!) = Enkleia sp. LITERATURE CITED Burrows, C. J. Studies in oe The breeding system. Trans. Roy. Soc. New Zealand 88: 29-45. DomkgE, W. oehlaareaila iiber die systematische und geographische Glieder- ung der Thymelaeaceen. Bibliot. Bot. 27(111): 1-151. Foster, A. S. ee Plant Anatomy. D. Van Nostrand Company, Inc. New York. Hou, D. Tones In: Flora Malesiana I. 6: 1-48. 1960. Kurz, S. On some new or imperfectly known Indian plants. Jour. As. Soc. Bengal 39: 1-91. 1870. LeanprI, J. Recherches Anatomiques sur les Thyméléacées. Ann. Sci. Nat. Bot. X. 12: 125-237. 1930. Metcatre, C. R. and L. CHatk. Anatomy of the Dicotyledons. 2 Vol. Oxford University Press. 1950. NevLinG, L. I., Jr. A revision of the genus Daphnopsis. Ann. Missouri Bot. ard. 46: 257-35 58. 1959, Riptey, H. N. The Dispersal of Plants Throughout the World. L. Reeve & Co., Ltd. Ashford, Kent. 1930 SOLEREDER, H. Systematische Anatomie der Dicotyledonen. p. 806-813. Stutt- gart. 1899 1961] BRIZICKY, GENERA OF VIOLACEAE 321 THE GENERA OF VIOLACEAE IN THE SOUTHEASTERN UNITED STATES * GEORGE K. BrizICcKy VIOLACEAE Batsch, Tab. Affin. Reg. Veg. 57. 1802, ‘Violariae.’ (VIOLET FAMILY) Herbs [to shrubs (sometimes scandent) or trees] with alternate [rarely opposite], simple, usually stipulate leaves. Flowers [regular or] irregular, bisexual [rarely unisexual or polygamous], 5-merous (except gynoecium), solitary to few in the leaf axils [or in racemose or cymose inflorescences ]. Sepals distinct [rarely connate at base], imbricate [or open] in bud, usually persistent. Petals distinct, hypogynous, usually imbricated, un- equal |or equal], the lowermost larger than the others and often spurred, saccate, or gibbous at base. Stamens 5, alternipetalous, hypogynous, the lowermost two [or all the stamens] often with a gland-, [scale-], or spur- like nectary on the back (abaxial side) [the glands sometimes connate into a “‘staminal sheath’]; filaments distinct [or connate]; anthers 2- locular at anthesis, introrse, longitudinally |very rarely transversely] de- hiscent, connivent or connate into a sheath around the ovary, the broad connective usually apereer beyond the anthers into a membranaceous rot) simple [rarely 3—5-lobed]|; style simple, terminal, sometimes flexuous in the lower part, [filiform to] columnar or clavate, with variously shaped apex (‘‘stylar head’’); ovary superior, 1-locular, with 1—-many anatropous ovules on each of 3 parietal placentae. Fruit a loculicidal, 3-valved cap- ‘Prepared for a generic flora of the oe United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University which has been pee dasa through the support of Georg re R. Cooley and the National Science Foundation. 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 40-42 (1959-1961). It should be repeated that the area covered by this work is bounded by and includes North Carolina, Tennessee, Arkansas, and Louisiana. 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. results of the study of Violaceae are somewhat mor e voluminous cin had exceptional necessity. The key for the sections is based chiefly on the shape of the stylar head, a character which at ne in some cases is liable to subjective perception lite i author is indebted to ie. Carroll E. Wood, Jr., for his criticism and valuable Redon: and to Mrs. Gordon W. Dillon, for her careful help in the preparation of the manuscript 322 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. XLII sule [or a berry, very rarely nutlike]. Seeds with endosperm, often carun- culate. Typr Genus: Viola L. A family of 18 [primarily tropical] genera with about 800 species dis- tributed all over the world. Only Viola, Melicytus, and Hymenanthera occur chiefly or exclusively in the temperate zone, the first being of wide distribution, the two latter centered in New Zealand. Of the remaining genera nine are tropical American, two tropical African, one South Pacific (New Caledonia, Fiji, New Guinea), one Hawaiian, and two exclusively (Rinorea, with ca. 260 species) or primarily (Hybanthus, with ca. 80 species) pantropical. The family appears to be a very natural one, espe- cially the subfamily Violoideae Melch. (including 17 genera), which is subdivided into the tribes Rinoreae, with regular or slightly irregular flowers (e.g., Rinorea, Gleospermum, Rinoreocarpus), and Violeae, with markedly irregular flowers (e.g., Hybanthus and Viola). All the genera appear to be entomogamous, but pollination data have been reported only for Viola, Cleistogamy is known only in Viola and Hybanthus, which have also been investigated in regard to embryology. The cytogenetics of Viola have been studied extensively; however, there are available only a few chromosome counts for five additional genera, i.e., Rinorea, 2n = 24 (1 sp.), Decorsella, 2n = 20 (monotypic), Hybanthus 2n = 24 (1 sp.), Melicytus, 2n = 32, 64, 96 (4 spp., 3 vars.), and Hymen- anthera, 2n = 32, 64 (6 spp.). Taking into consideration these and the chromosome numbers in Viola (q.v.) one may assume that the polyploidy may be of wide distribution in the family and probably played an important role in the origin of species. Anchietea (South America) and Decorsella (Africa) are notable for capsules dehiscing when very young (almost im- mediately after the pollination of the ovary) so that the seeds mature uncovered. The family is closely related to Flacourtiaceae and through the latter with Turneraceae, Malesherbiaceae, and Passifloraceae. A relationship with Cistaceae seems probable. The economic significance of the family is mainly based on the orna- mental value of numerous species of Viola. REFERENCES: For chromosome numbers of genera beyond our range see BEUZENBERG & Hair, New Zealand Jour. Sci. 2: 533-535. 1959; MANGENoT & MANGENOT, Bull. Jard. Bot. Bruxelles 27: 653. 1957, HEILBorN, O. Bidrag till Violaceernas cytology. Sv. Bot. Tidskr. 20: 414-419, 1927. |Hybanthus parviflorus (L. f.) Baill., 2n = 24; chromosome counts on Viola. Knutu, P. Handbuch der Bliitenbiologie. 2(1). 1898 [Violaceae, 137-149]; 3(1). 1904 [Violaceae, 501]. Metcuior, H. Violaceae. Nat. Pflanzenfam. ed. 2. 21: 329-377. 1925. [ Viola by W. Becker. } . Die phylogenetische Entwicklung der Violaceen und die natiirlichen 1961] BRIZICKY, GENERA OF VIOLACEAE g23 Verwandtschaftsverhaltnisse ihrer Gattungen. Repert. Sp. Nov. Beih. 36: 83-125. 1925. Reicue, K., and P. Taupert. Violaceae. Nat. Pflanzenfam. II. 6: 322-336. Taytor, F. H. Comparative anatomy of the secondary xylem in the Violaceae and Flacourtiaceae. (Abstract.) Am. Jour. Bot. 25(Suppl.): 20s. 1938. Key To THE GENERA OF VIOLACEAE Sepals not auricled at base; lowermost petal gibbous at base; ear connate, the abaxial glands of the two lowermost stamens connate. .. Hybanthus. Sepals auricled at base; lowermost petal spurred at base; anthers Te i the abaxial spurs of the two lowermost stamens distinct, ee ees 2. Viola. 1. Hybanthus Jacquin, Enum. PI. Carib. 2. 1760, nom. cons.” Herbs [to shrubs] with alternate [opposite] leaves. Flowers small [or medium-sized], greenish white [white or blue to violet], pediceled, soli- tary to few in leaf axils [or in racemose, rarely cymose, inflorescences]. Sepals distinct, imbricate in bud, nearly equal. Petals distinct, imbricate, slightly [or considerably | unequal, the lowermost always largest and gibbous or saccate at base. Stamens with filaments distinct [or connate |, each of the 2 [rarely 4| lowermost bearing a nectar gland [or spur| on back (abaxial side), the glands connate into one 2-lobed [or entire] nectary [or distinct]; anthers connate [or connivent] into a sheath inclosing the ovary, each connective prolonged beyond the anther into a membranaceous appendage. Gynoecium 3-carpellate; stigma small, simple; style usually club-shaped, the apex hooked [or slightly curved], with stigmatic opening directed forward [abaxially]; ovary 1-locular; ovules 2- integumented with a thick nucellus, 3 [to numerous] on each placenta. Capsule 9[—many |- -seeded; seeds relatively large, globose [or ovoid], carunculate, with copious endosperm and large, straight, axile embryo with flat coty- ledons. (Including Cubelium Raf., lonidium Vent., Solea Spreng.) TYPE species: H. havanensis Jacq. (Name from Greek, #ybos, hump-backed, and anthos, flower, referring to the gibbous flowers.) A largely pantropical genus of about 80 species centered in tropical South America but extending into the Temperate Zones in Uruguay, Ar- gentina, and Chile, in southern Australia, and in North America. The northernmost species is Hybanthus concolor (T. F. Forst.) Spreng. (Cubelium Raf., Jonidium Benth. & Hook. f. ex S. Wats., Solea Ging. ex DC.), green ‘violet, a perennial herb with alternate leaves and small greenish-white flowers solitary in two’s or three’s in the leaf axils, which ranges from Georgia to Mississippi, Arkansas, and Kansas, northward to Wisconsin (?), Michigan, southern Ontario, New York, and Connecticut. Two other species, H. verticillatus (Ort.) Baill. (H. linearis (Torrey) Shinners) and H. attenuatus (Humb. & Bonpl.) G. K. Schulze reach their 2 Conservation unnecessary; see Rickett & Stafleu, Taxon 8: 314. 1959, 324 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII northern limits westward beyond our range in Kansas and Colorado and in southern Arizona, respectively. The occurrence of small, cleistogamous flowers with petals reduced or wanting and with two or five glandless stamens has been recorded in H. concolor and in the South American H. communis (St. Hil.) Taub. and H. Ipecacuanha (L.) Baill. Only H. parviflorus (L. f.) Baill., 2n = 24, has been investigated cytologically. The rhizome and roots of H. I peca- cuanha have sometimes been used as a substitute for true ipecac (Cephaé- lis Ipecacuanha (Brot.) A. Rich., and C. acuminata Karst., Rubiaceae), an emetic. Roots of several other species are used locally in South America and Asia as emetics and against diarrhea. Although the genus includes spiny shrubs to rosette plants, the delimitation of subgeneric categories awaits a much-needed taxonomic revision. Hybanthus is presumably most closely related to Viola (Melchior). REFERENCES: See also under family references HEILBORN (1927) and Metcuior (Nat. Pflanzenfam., 357-360). Anprews, F. M. Development of the embryo-sac of Hybanthus concolor. Bull. Torrey Bot. Club 37: 477, 478. 1910. BEAUVISAGE, G. L’inuline dans les Jonidium, Etude anatomique du faux Ipéca- cuanha blanc du Brésil (Jonidium Ipecacuanha). Bull. Soc, Bot. Lyon II. 6: 12-24. 1 pl. 1888. BERNOULLI, G. Zur Kenntniss dimorpher Bliithen. Bot. Zeit. 27: 17-19. 1869. |H. communis (as Ionidium commune St. Hil.). cleistogamy, 18, 19. Core, E. J. Cleistogamous flowers on Solea concolor. Asa Gray Bull. 6: 50. 1898. | H. concolor. | NIEUWLAND, J. A. Cleistogamy in Cubelium. Am. Midl. Nat. 4: 501, 502. 1916 Raju, M. V. S. Seed development and fruit dehiscence in Jonidium suffruti- cosum Ging. Phytomorphology 8: 218-224. 1958. |= H. enneaspermus (L.) F. Muell.] SAYEEDUD-Din, M. Some common Indian herbs with notes on their anatomical characters. X. Jonidium suffruticosum Ging. (Violaceae). Jour. Bombay Nat. Hist. Soc. 43: 475-477. pls. 1-3. 1942. |H. enneaspermus. | 2. Viola Linnaeus, Sp. Pl. 2: 933. 1753; Gen. Pl. ed. 5. 402. 1754. Herbs [rarely shrubs or subshrubs], often acaulescent, with alternate, petioled [or sessile] leaves. Flowers usually on 1-flowered, solitary pe- duncles borne in the axils of cauline leaves or (in acaulescent species) directly from the rhizome and stolons; cleistogenes often following the normal, chasmogamuus flowers. Sepals subequal, auricled at base. Petals unequal, the lowermost the largest and spurred at base. Stamens with short, distinct filaments, the 2 lowermost bearing spurlike [or wartlike | nectaries projected into the spur of the lowest petal; anthers usually con- uivent into a sheath inclosing the ovary, each connective produced at apex into a conspicuous, membranaceous appendage: pollen 3(4—6)-col- pate, spheroidal to ellipsoidal, smooth, medium-sized to large. Stigma 1961] BRIZICKY, GENERA OF VIOLACEAE 322 simple, terminating a stylar beak or situated on the abaxial side, rarely on the summit, of the beakless stylar head; style + columnar, curved + forward (rarely upward) at apex, terminating in a beak (rostrum), or enlarged upward into a + capitate, sometimes margined and helmet-like [or variously appendaged], abaxially beaked or beakless stylar head; ovules 3—many on each placenta, 2-integumented, with a thick nucellus. Seeds globular to obovoid, usually smooth, often carunculate; endosperm copious; embryo straight, axile, with rather thick, plano-convex cotyle- dons. LectotyPe spEcIES: V. odorata L.; see Britton & Brown, Illus. . North. U. S. ed. 2. 2: 546. 1913. (Name classical Latin, Viola, the divanmnated form of Greek, iolon, a diminutive of zon, a violet, or some other flower. A genus of about 400 species, of world-wide distribution, centered in the North Temperate ie restricted to mountains in the tropics. A few species occur in the Arctic Zone in the Old World and in the Subarctic Zone of North America; ae southernmost limits are South Africa, New Zealand (3 spp., 2 endemic), and Tierra del Fuego. Although the Andes of South America abound with endemic species exhibiting a great diversity of floral patterns, only a small number are known from the tropical moun- tains of the Eastern Hemisphere. The Hawaiian Islands are notable for eight endemic, mostly frutescent (single-stemmed) species (often with racemose inflorescences) which represent a special section, NOSPHINIUM W. Becker (e.g., V. tracheliifolia Ging., V. robusta Hilleb., 2n = ca. 76, V. mauiensis Mann, 2n = 82, ca. 85, ca. 86). Nearly 80 species in four of fourteen sections recognized by Becker (1925) are represented in North America. The tentative classification followed below is essentially that of Clausen (1929 and especially 1951) and is based mainly on floral characters (especially the type of stylar head), chromosome numbers, and, to a lesser degree, crossing relationships. The sections given in the key are divided into subsections within which the species are connected by fertile to partially sterile hybrids and between which connections are usually through sterile hybrids. Intersectional hybrids are difficult to produce, and only two instances of vigorous but completely sterile intersectional hybrids are known. No intersectional hybrids involving sect. MELANIUM have been produced. Chromosome numbers have been reported for about 160 species. KEY TO THE SECTIONS OF VIOLA IN THE SOUTHEASTERN UNITED STATES A. Upper portion of style variously shaped, beaked or beakless, without a lip- like appendage (‘‘lip”) at base of the stigmatic opening; rhizomatous peren- nials, usually with bractlike stipules. Style enlarged upward, capitate, beakless, the stigma within a minute opening on the lower (abaxial) side of the stylar head; flowers usually — or white with bases of petals and/or spur yellow; usually caules- cent. pense beled age ke pa ia ae ee ey aas Sect. CHAMAEMELANIUM. ad 326 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII B. Style various, beaked at apex, very rarely beakless; flowers blue to violet, creamy, or white (without any yellow). C. Style clavate, depressed-capitate and often margined, beaked on the lower side of the stylar head (or in V. pedata obliquely concave at summit, beakless); acaulescent. ............ ect. PLAGIOSTIGMA. r, apex beaked and + bent or curved forward, rarely mae straight pointing forward and upward; acaulescent or CRUMBRRR,. § xsd ead bobs ea hy ene bale ees ect. VIOLA. A. Upper portion - style capitate-globular, beakless, bearing on the lower side a wide stigmatic opening with a “lip” at its base; flowers bluish white, creamy, or lavender (or yellow and/or blue to dark violet); usually caules- cent annuals or biennials with large, leaflike, pinnately dissected, + lyrate, REO i ins 05 5 to cr gee an 4 eho ba 8s Cae OnS Sect. MELANIUM. ‘o) NM o < Oo ~ a = c = =] pee] Section CHAMAEMELANIUM Ging. (sensu Clausen), including North American and Eurasian species forming a polyploid series on a base of six (2n = 12, 24, 36, 48 (60), 72), is considered to be the most primitive of the genus (Clausen, 1929). Six subsections occur in North America, three in our area. Subsection Nudicaules W. Becker comprises yellow- flowered caulescent violets with one to few basal leaves and stems leafy only toward the summit. The four eastern North American species, V. pubescens Ait. (2n = 12), V. eriocarpa Schwein. (V. pensylvanica Michx., fide Fernald) (2n = 12), V. hastata Michx., and V. tripartita Ell., occur in our area. Putative natural hybrids intermediate in morphology and ecology between the first two species have been recorded (Russell, 1960), and the lines between the two are often blurred. Subsection Canadenses W. Becker (Erectae, Canadenses W. Becker) includes North American white-flowered caulescent violets (‘Canada violets”) represented in our area by V. canadensis L. (2n = 24), lacking stolons and ranging from Alabama to Tennessee and North Carolina and west- and northward far beyond our area, and the stoloniferous V. rugulosa Greene, which reaches its southern limit in northwestern North Carolina and eastern Tennessee. Experimental hybrids between the two are vigorous and fertile, and both intergrade in areas of overlap, suggesting their treatment as parts of a single species (Russell, 1960). Cleistogamous flowers are not produced in this subsection. Subsection Orbiculares Pollard sensu stricto (subsect. Rotundifoliae Clausen, ined. 1929, not Borbas, 1892) includes only V. rotundifolia Michx., 2n = 12, a yellow-flowered stemless violet, which extends southward along the Appalachians to Georgia and Tennessee. The species is relatively invariable, and there are no indications of hybridiza- tion or introgression with other species. This subsection seems to form a transition to sect. PLAGIosticMA (Clausen, 1929). No natural hybrids between subsections have been found, and experimental crosses have either failed or produced dormant seeds or weak seedlings. Section PiaciosticMa Godr. (sensu Clausen, 1951) (§ Nomimium Ging., in part; § Nomimium subsect. Plagiostigma auct.) comprises Eura- sian and North American white- to violet-flowered acaulescent species forming a polyploid series based on twelve (2n = 24, 48, 72, 96), with 1961] BRIZICKY, GENERA OF VIOLACEAE oot some exceptions (2m = 44, 54, 56). Subsection Primulifoliae Gershoy (§ Nomimium, Stolonosae Kupff. sensu W. Becker, in part; § Plagiostigma subsect. Stolonosae Clausen, 1951; Stolonosae, Primulifoliae Gershoy) includes usually stoloniferous white violets with slender, cordlike rhizomes, cleistogamous flowers with erect peduncles, and ellipsoid green capsules. Viola lanceolata subspp. lanceolata and vittata (Greene) Russell (V. vit- tata Greene), both 2m = 24; V. Macloskeyi F. E. Lloyd subsp. pallens (Banks) M. S. Baker (V. pallens (Banks) Brainerd), 2m = 24; and V. primulifolia L. subspp. primulifolia and villosa (Eaton) Russell (including V. rugosa Small), both 2n = 24, represent the group in our area. These species seem to intercross freely, and putative natural hybrids of all of them have been recorded. Experimental hybrids between V. Macloskeyi subsp. pallens and V. lanceolata, as well as V. primulifolia, have been reported to be fertile. Subsection Blandae Pollard, in part (Stolonosae ser. Blandae Gershoy, Incognitae Clausen), is composed of the closely related Viola incognita Brainerd and V. blanda Willd., which differ from the preceding white- flowered species in the prostrate peduncles of cleistogamous flowers, the ovoid, purplish or reddish capsules, and the chromosome number (2” = 44). Primarily northeastern, both reach their southern limits in our area. Where the two are in contact ‘there is morphological intergradation, due to relatively unrestricted intercrossing’’ (Russell, 1955). Subsection Boreali-Americanae W. Becker, the acaulescent ‘“‘blue”’ violets, includes exclusively North American species with fleshy, thickened rhi- zomes without stolons, blue-to-violet flowers (white in albinic forms), and a common chromosome number (2n = 54). About ten of the twenty- five species recorded from our area are primarily or exclusively south- eastern. Others (as Viola papilionacea Pursh [including V. Priceana Pol- lard], V. palmata L., V. cucullata Ait., V. sagittata Ait., and V. fimbri- atula Sm.) are of wide distribution in eastern North America. The species intercross freely, and numerous natural and experimental hybrids, the latter vigorous and usually fertile, are known. Introgression seems to be frequent. At least some of the species probably should be reduced to subspecific rank. This subsection is one of the most difficult groups taxo- nomically, both because of general morphological similarities and because “frequent hybridization tends to obliterate not only the morphological differences, but also the more sharply marked ecological preferences” (Russell, 1960). Experimental hybrids between subsects. Primulifoliae and Boreali-Americanae, 2n = 39, 51, are sterile. Subsection Pedatae Pollard differs from the preceding primarily in the clavate beakless style, obliquely concave at the summit and bearing the stigma within a small protuberance near the center of the cavity, as well as in chromosome number (2” = 56). The single species, Viola pedata L., widely distributed in eastern North America, is considered to be most closely related to V. Brittoniana Pollard (Boreali-Americanae). The ab- sence of cleistogamous flowers has been regarded as characteristic, but these have been recorded by Hills (1946) in plants cultivated in England. 328 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII No natural hybrids are known, and experimental attempts have failed. However, Russell and Bowen concluded (on the basis of vegetative features alone!) that a form with unusual leaf variation in North Carolina was probably the result of hybridization and subsequent introgression between V. pedata and V. primulifolia subsp. villosa (subsect. Primulifoliae). Section VIoLA (§ Nomimium Ging., in part; § Rostellatae Boiss. sensu Clausen, 1951) includes North American, Eurasian, and North African species of a polyploid series based on ten (2m = 20, 40, (60), 80). Sub- section Uncinatae Kupff. comprises acaulescent, often stoloniferous violets with the style curved in a hook at the tip and the seeds not forcibly scat- tered from the capsule. Viola odorata, sweet or English violet, 2x = 20, has been introduced and has become locally established in various parts of the United States. The North American—Eurasian subsect. Rosulantes Borb. includes caulescent blue violets with tufted stems and rosulate basal leaves, rather slender styles beaked and bent at the apex, and seeds forcibly ejected from the capsules. The Eurasian species are diploid or tetraploid, but the North American, with the exception of the western Viola Howellii (2n = ca. 80), are diploid. Viola conspersa Reichenb., V. rostrata Pursh, and V. striata Ait., all 2n = 20 and linked by partially fertile hybrids, reach their southern limits in our area, while V. Walteri House, 2n = 20, is primarily southeastern. Viola adunca Sm., of more northern distribu- tion, apparently also belongs to this group. Section MELANIUM Ging. includes about 50 almost exclusively Eurasian and North African species exhibiting a variety of chromosome numbers in a modified six- or ten-series. Only subsect. Tricolores W. Becker sensu Clausen (1929) (Calcaratae, Tricolores W. Becker) including species with chromosome numbers based on x = 6 (2n = 14, 16, 24, 26, 34, 36, 48) is represented in North America. Viola Rafinesquii Greene (V. Kitaibeliana R. & S. var. Rafinesquii (Greene) Fern.), 22 = 34, a small annual (or biennial) with bluish-white, creamy, or rarely lavender flowers, occurs from Georgia to Texas and northward beyond our limits. Being distinct from the related European species morphologically and from V. Kitaibeli- ana, 2n = 14, 16, 24, 36, 48, especially in the chromosome number, it is also supposedly the only species of sect. MELANIUM which produces cleis- togamous flowers (Gershoy, 1934). The nativity of this species in North America has been neither proved nor disproved. Viola arvensis Murr., 2n = 34, introduced from Europe, has been recorded as naturalized in North Carolina and northward to Newfoundland, Michigan, and Alberta. Viola tricolor L., 2n = 26, from Eurasia and North Africa, and especially its garden forms of hybrid origin (pansies, “violas,” etc.), are commonly cultivated. The Tricolores as a group are notable for the occurrence of polyploidy and polysomaty The violets, in general, seem to show a preference for acid soil, the cere’ of acidity varying with each species, although growth of plants “is not correlated with the soil acidity alone, but is influenced by a combination of this factor with moisture, texture, and fertility” (Mc- Cullough, 1941). Vegetative reproduction by rooting stolons (e.g., in 1961 | BRIZICKY, GENERA OF VIOLACEAE 329 Primulifoliae and Blandae) or by the detachment of rooting lateral, superficial branches of the rhizome (e.g., in oreali-Americanae), more rarely by adventitious buds on roots (in Pedatae), permits the survival of sterile hybrids. Mycorrhiza have been found in Viola palustris L. and V. biflora L. Although apparently often self-fertile, the chasmogamous flowers are adapted to cross-fertilization, self-pollination being prevented by the position of the stigma in regard to the anthers, and/or by configuration of the stigma, or by proterandry (V. lanceolata, V. Macloskevi). Thysa- noptera (especially Thripidae) and various Diptera, Hymenoptera, and Lepidoptera have been recorded as pollinators, the short-tongued species pollinating the short-spurred flowers, the long-tongued the long-spurred violets. These insects usually are in a reverse position (the head down) when sucking the nectar. Details of pollination differ somewhat in differ- ent sections. Viola arvensis is self-fertilized, and self-pollination occurs early, in the flower buds. In the great majority of the species of sects. CHAMAEMELANIUM, VIOLA, and Praciosticma, the chasmogamous flowers are followed by reduced, mostly apetalous, closed, highly fertile, cleistogamous flowers with 2-5, + reduced, spurless stamens and flexuous and much curved styles, so that the stigma is brought into contact with the anthers. The pollen grains germinate within the anthers, and the pollen tubes penetrate the anther walls into the stigma. The presence of dimorphic pollen grains in both chasmogamous and cleistogamous flowers in V. odorata var. praecox Greg- ory is noteworthy; only one kind of these pollen grains is germinable in the chasmogamous, and only the other in the cleistogamous, flowers (Madge, 1929). Semicleistogamous flowers forming a transition between chasmo- and cleistogamous flowers are often observed. In many hybrids the sterility involves only the chasmogamous flowers, seeds being produced by the cleistogamous flowers. Apomixis, although presumed by some earlier taxonomists, has not been proved in Viola. Myrmecochory has been reported in several species. Many species are valued ornamentals. Roots of Viola tricolor and V. odorata contain a saponin and an alkaloid, violin, and are official (ex- pectorants) in some countries. A medicinal preparation, traviolin, is ob- tained from the leafy stems of V. tricolor in the U.S.S.R. Viola odorata is extensively cultivated in Europe for use in perfume manufacturing. REFERENCES: The vast number of references has been reduced here primarily to those either of general interest or dealing specifically with the southeastern United States. Under family references see also KNUTH (1898), MELCHIOR (1925), REICHE & TAUBERT (1895). Antony. R. D. Asexual inheritance in the violet. N.Y. Agr. Exp. Sta. Bull. 76: 3-55. 1920.* Barrp. V. B. Wild violets of North America. 225 pp. including 76 col. pls. Berkeley and Los Angeles. Univ. Calif. Press. 1942. 330 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Baker, M. S. Studies in western violets —JI. Sections Chamaemelanium and Nomimium, Madrono 3: 51-57. 1935. [Includes descriptions of floral structures and pollination in these sections. | BAMForD, R., and A. GersHoy. Studies in North American violets. II. The cytology of some sterile F, violet hybrids. Vt. Agr. Exp. Sta. Bull. 325: 1-53. pls. 1-22. 1930. Beat, W. J. Seed-throwing - neal oo 4: 230. 1902. [See also G. Schendell, Urania Jena 15: Z| Becker, W. Viola. Nat. nce its a 2. 21: 363-376. 1925. Buss, M. C. A contribution to the life-history of Viola. Ann. Bot. 26: 155- 163. pls. 17-19. 1912. [Embryology; includes V. cucullata, V. fimbriatula, V. odorata, V. pedata, V. pubescens. Bop, H. C., and A. GersHoy. Studies in North American violets. IV. Chro- mosome relations me fertility in diploid and tetraploid species hybrids. Vt. Agr. Exp. Sta. Bull. 378: 1-36. pls. 1-15. 19 BRAINERD, E. Hybridism in the genus Viola. Rhodora 6: 213-223. 1904. II. Ibid. 8: 6-10. 1906. III. Ibid. 49-61. pls. 66-7 —— oo law of dominance in the hybrids of a Ibid. 9: 211-216. 19 ; iE evolution of new forms in Viola through hybridism. Am. Nat. 44: 229-236. 1910. Further notes on the stemless violets of the South. Bull. Torrey Bot. Club 38: 109. pl. 1. 1911. . The caulescent violets of the southeastern United States. /bid. 191- 198. | Includes key. | . Violets of North America. Vt. Agr. Exp. Sta. Bull. 224: 1-172. 1921. . Some natural violet hybrids of North America. /bid. 239: 1-205. 1924. | Includes references to his numerous papers on hybridization. | Cain, S. A., and P. DANserEAvu. A study of the stemmed yellow violets of eastern North America (four species) on the basis of mass collections. Am. Philos. Soc. Yearb. 1951: 166. 1952. [Progress report; see also ] Lévésque & Dansereau, Ann. ACFAS 23: 98. 1957.* CLAUSEN, A Studies on the collective species of Viola tricolor L. Bot. Tidsskr. 37: 201-221. pls. 1, 2. 1921. II. Ibid. 37: 363-416. 1922. pene: and cytological investigations on ae tricolor L. and V. arvensis Murr. Hereditas 8: 1-156. pi 19 on-Mendelian inheritance in Viola. are 9: ae 256. 1927. . Chromosome number and the relationship of species in the genus Viola. Ann. Bot. 41: 677-714. 1927 ———. Chromosome number and ielatiouship of some North American species of Viola. Ibid. 43: 741-764. ; ———. Male sterility in Viola orphanidis. Hereditas 14: 53-72. 1930. . Cyto-genetic and taxonomic investigations on Melanium violets. Ibid. 15: 219-308. 1931. [See also Proc. 6th Int. Congr. Genet. 2: 346-349. 1932.] ———. Stages in the evolution of plant species. 206 pp. Cornell Univ. Press. Ithaca, New York. 1951. [Viola, 16-19, 54-61, 143-158. ] Cooper, C. C., and D. P. Watson. Influence of daylength and temperature on the growth of greenhouse violets. Proc. Am. Soc. Hort. Sci. 59: 549-553, 1952.* 1961] BRIZICKY, GENERA OF VIOLACEAE g54 DeTJEN, L. R. Peloria in Viola primulaefolia Linn. Torreya 20: 107-116. 1921. [See also Jour. Hered. 16: 387-390. 1925.] Dopp, J. D., and A. GERsHoy. Experiments in the grafting of species in the genus Viola. Bull. Torrey Bot. Club 70: 91-103. 1943. [There are prob- ably no general grafting limits within the genus. | Enpo, T. Biochemical and genetical investigations of flower color in Viola tricolor L. I. Inter-relationships of pigment constituents occuring in ten varieties. Jap. Jour. Bot. 14: 187-193. 1954. Dominance relationships of flower color in varietal F, hybrids of Swiss giant pansy, Viola tricolor maxima Hort. Mishima Natl. Inst. Genet. Rep: 7; 60,51, 1957. Evans. L. T. Chasmogamous flowering in Viola palustris L. Nature 178: 1301. 1956," FERNALD, M. L. Another century of additions to the flora of Virginia. Rhodora 485-553, 559-630. pls. 670-695. 1941. [V. pensylvanica Michx. is conspecific with and preferred to V. eriocarpa Schwein., 616, 617.| Studies of eastern American plants. 2. Rhizome-characters in and minor forms of Viola. Ibid. 51: 51-57. pls. 1133-1136. 1949. (Contr. Gray Herb. 169.) Fortunk, H. Des Violariées. Etude spéciale du genre Viola. Thesis. 91 pp., 2 pls. Montpellier. 1887. [Primarily comparative anatomy. FotTHERGILL, P. G. Studies in Viola. I. The cytology of a naturally occurring population of hybrids between Viola tricolor L. and Viola lutea Huds. Genetica 20: 159-186. 1938.* [See also New Phytol. 43: 23-35. 1944.*] Fox, L. E. Viola conspersa in Louisiana. Torreya 41: 109. 1941. Gates. B. N. Carunculate seed dissemination by ants. Rhodora 45: 438-445. 1943. [V. cucullata Ait., 440, 441.] Gersuoy, A. Studies in North American violets —I. General considerations. Vt. Agr. Exp. Sta. Bull. 279: 1-18. 1928. [II. See Bamford, R., and A. Gershoy.] III. Chromosome numbers and species characters. /bid. 367: 1-91. pls. 1-27. 1934. [IV. See Bold, H. C., and A. Gershoy. | Gray. A. Memoranda of a revision of the North American violets. I. Bot. Gaz. 11: 253-256. 1886. II. bid. 289-293. GREENE. E. L. Mutations in Viola. Leafl. Bot. Observ. Crit. 1: 182-187. 1906. [For notes and numerous new taxa see also Pittonia 3-5. 1896-1902; Leafl. Bot. Observ. Crit. 1, 2. 1903-1910; Am. Midl. Nat. 3: 79-85. 1913.] Griscom, L. Notes from northwestern Florida. Rhodora 38: 48-50. 1936. [Critical notes on some species; new combinations; V. rugosa Small in- cluded in V. primulifolia. | Huts. L. D. Cleistogamy in some Viola species. Gard. Chron. III. 120: 31. 1946. [Formation of cleistogamous flowers in V. sagittata and V. pedata cultivated in England. | Hotm. T. Comparative studies of North American violets. Beih. Bot. Cen- tralbl. 50(2): 135-182. pls. 1, 2. 1932. __ Contributions to the knowledge of the germination of some North American plants. Mem. Torrey Bot. Club 2: 57-108. pls. 5-19. 1891. [Includes “V. palmata var. cucullata.” See also Brainerd, Science 25: 941- 944. 1907.] Horn. W. Untersuchungen iber die cytologischen und genetischen Verhaltnisse beim Gartenstiefmiitterchen Viola tricolor maxima hort. (= V. wittrocki- 4 ana Gams), einer polyploiden Bastardart. Ziichter 26: 193-207. 1956.* 332 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII House, H. D. Notes on southern violets —I. Torreya 6: 171-173. 1906. II. Ibid. 7: 133-136. 1907. Jones, G. N. Viola eriocarpa vs. V. deieaasean Rhodora 61: 219, 220. 1959, [V. pensylvanica Michx. is a synonym of V. pubescens Ait. | KaczMAREK, R. M. ee. aoe (Greene). An ecological and anatomical study. Am. Midl. Nat. 4: 74-88. pls. 4-10. 1915. [V. ach- lydophylla Greene, ecology and anatomy of the seedling. | KraeMER, H. The morphology of the genus Viola. Bull. Torrey Bot. Club 26: 172-183. 1899. Leavitt, R. G. Seed dispersal of Viola alae ag Rhodora 4: 183, 184. 1902. McConneLi, T. A.,. and N. H. Russett. Flowering dates of Viola sororia Willd. and V. hee ta Michx. "7 “different latitudes. Proc. Iowa Acad. Sci. 66: 178-184. McCu.toucu, H. M. Sais in soil relations of species of violets. Am. Jour. Bot. 28: 934-941. 1941. Mapce, M. Spermatogenesis and fertilization in the cleistogamous flowers of Viola odorata var. praecox Gregory. Ann. Bot. 43: 545-577. pl. 12. 1929. Mancu, D. Meiosis in an F, Viola hybrid and its reciprocal. Am. Jour. Bot. 24: 678-683. 1937. [V. ambigua < odorata and its reciprocal. ] Marvin, J. W. Cell size and ae size in two violet species and their hybrid. Bull. Torrey Bot. Club 63: 17-32. pl. 1. 1936. [V. conspersa, V. papiliona- ced. Meyer, F. J. Bau und Ontogenie des W asserleitungssystems der vegetativen Organe von Viola tricolor var. arvensis. Thesis, 38 pp. Marburg. 1916. [Review in Bot. Centralbl. 141: 211, 212. 1919.] Miter, P. M. An experimental study of variability in a wild population of violets. Diss. Abs. 16: 1568, 1569. 1956.* Miyajr, Y. Betrachtungen uber die Chromosomenzahlen von Viola, Violaceen und verwandten Familien. Planta 11: 631-649. 1930. [Review in Bot. Centralbl. 160: 122, 123. 1931.] Moore, D. M. Population studies on Viola lactea Sm. and its wild hybrids. Evolution 13: 318-332. 1959 Murriti, W. A. The violets of Alachua County, Florida. Jour, Elisha Mit- chell Sci. Soc. 56: 367-370. 1940. [See also Ecology 21: 512, 513. 1940.] A new Florida violet. Quart. Jour. Fla. Acad. Sci. 12: 61, 62. 1949. NIEUWLAND, J. A. Notes on cleistogamous flowers of violets —II. Am. Midl. Nat. 3: 198-200. 1914. . Cleistogamous flowers in the pansies. /bid. 4: 464-466. 1916. | Cleisto- gamous flowers in V. Rafinesquii. | d R. M. Kaczmarek. Studies in Viola —I. Am. Midl. Nat. 3: 207— 217. 1914. | Four genera, two new; segregation is mainly based on cleistog- amous flowers. | Potrarb, C. L. The acaulescent violets. Bot. Gaz. 23: 53, 54. 1897. . Further observations on the eastern acaulescent violets. /bid. 26: 325— 342, 1898. [See also Proc. ae Soc. Wash. 10, 13, 15, 16. 1896-1903: Bull. Torrey Bot. Club 24. 1897, | RussELL, N. H. An example of ie reedies hybridization between Viola papi- lionacea Pursh and V. nephrophylla Greene. Proc. Iowa Acad. Sci. 59: 134-140. 1952. . Three field studies of hybridization in the stemless white violets. Am. Jour. Bot. 41: 679-686. 1954 1961 | BRIZICKY, GENERA OF VIOLACEAE 333 _ Variation in leaf pubescence in Viola ae Brainerd and V. renifolia Guy, Proc. Iowa Acad. Sci. 61: 151-160. _ Local introgression between Viola uae Ait. and V. septentrionalis Greene. Evolution 9: 436-440. 1955 a ee variation in Viola rotundifolia Michx. Castanea 20: 144- 153. 195 The hee of the North American acaulescent white violets. Am. Midl. Nat. 54: 481-494. 1955. _ Hybridization between Viola pedatifida and its relatives. Castanea 21: 133- 140. 195 . Regional variation patterns in the stemless white violets. Am. Midl. Nat. 56: 491-503. 195 —. Technique for species comparison in violets. Proc. Iowa Acad. Sci. 63: 157-160. 1956. The nature of Viola missouriensis Greene. Ibid. 65: 96-101. 1958. _____ The violets of Tennessee. 1. Keys to the species and distribution maps. Castanea 23: 63-76. 1958. The violets of Minnesota. Proc. Minn. Acad. Sci. 25/26 (1957/1958): 126- 191. 1960. [Includes notes on hybridization and introgression. | and W. R. Bowen. A hypothesis to account for unusual leaf variation in V hee pedata L. Proc. Iowa Acad. Sci. 67: 189-194. 1960. M. CooperRIweR. Prediction of an introgressant in Viola. Am. Midl co 54: 42-51. 1955. [Introgression between V. sagittata Ait. and V. sororia Willd. is presumed. | and J. GraHaM. Midwestern ieee studies of Viola papilionacea Pursh. Proc. Iowa Acad. Sci. 65: —109. 1958. and A. C. Risser, Jr. The hy oe nature of Viola emarginata (Nuttall) Leconte. Brittonia 12: 298-305. 19 Scutnz, H. Die Veilchenriechstoffe. Panag Chem. Org. Naturstoffe 8: 146- 206. 1951. [Review with 171 references covering the odoriferous sub- stances of violets. | Scnibrer. G. Untersuchungen iiber die Polymorphie einheimischer Veilchen. Planta 43:°537-565. 195 SHAVER, J. M. Heterophyllous leaves in the ~~ lobed Viola Egglestonii Brainerd. a Tenn. Acad. Sci. 34: 40-46. 195 SovuEcEs. E. C. Embrvyogénie des Violacées. Sea de l’embryon chez le Viola ae L. Compt. Rend. Acad. Sci. Paris 205: 169-171. 1937." VALENTINE. D. H. Variation and polymorphism in Viola. Proc. Roy. Soc. Lon- don B. 145: 315-319. 1956 Cytotaxonomy of the rostrate violets. Proc. Linn. Soc. 169: 132-134. 1958. Watters. S. M. Observations on varieties of Viola odorata L. Bot. Soc. Exch. Club Brit. Isles Rep. 12(1943/1944): 834-839. West, G. Cleistogamy in Viola petiag with especial reference to the cytolog- ical aspects. Ann. Bot. 44: 09. pls. 11, 12. Wittrock, V. B. ee 1 oe Viola tricolore (L.) aliisque speciebus sectionis Melanit observationes SLO NPROIO RICA, biologicae, systematicae. (In Swedish.) Acta Horti Berg. 2(1): 3-142. pls. 1-14, in part color. 1897. II. A contribution to the ne any of the pansies having special ref- erence to their origin. (In Swedish; English summary.) /bid. 2(7): 3-77. 1 pl. 1896. 334 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII COMPARATIVE ANATOMY OF THE LEAF-BEARING CACTACEAE, III FORM AND DISTRIBUTION OF CRYSTALS IN PERESKIA, PERESKIOPSIS AND QUIABENTIA * I. W. Battery SINCE THE PUBLICATION of Schleiden’s classical paper (1845) on the anatomy of Cactaceae, many references, frequently more or less casual or incidental, have been made to the occurrence of crystals in various representatives of the family. The most comprehensive of these papers is that of Lauterbach (1889) who utilized crystals of calcium oxalate in correlation with mucilaginous and other structures in an attempt to dif- ferentiate three categories of the Cactaceae. Inadequate attention in investigations of cacti has been given to ranges of variability of selected diagnostic characters, not only in different taxa, different collections of the same taxon, and different plants of the same clone, but also in different parts of the same plant during successive stages of its growth to maturity. It seems advisable, accordingly, to determine to what extent crystals provide reliable and significant taxonomic evidence in the case of the leaf-bearing genera of the Cactaceae. CRYSTALS OF LEAVES Druses of calcium oxalate occur in the leaves of all putative species of Pereskia, Pereskiopsis, and Quiabentia of which I have succeeded in ob- taining material (Fics. 1-12). This is true. regardless of whether the plants were grown in greenhouses, botanical gardens, or in their native habitats in the wild. The druses vary widely in size, form and number, as they likewise do in the size, form and number of their constituent crystals (Fics. 13, 15, 16). The conspicuous differences in size (Fic. 1 vs. 4), in number (Fic. 3 vs. 1 and 4), and in form (Fic. 15 vs. 16) suggest at first sight that crystals might be of considerable utility in the differentia- tion of taxa. Unfortunately, this proves to be uncertain and difficult when the ranges of variability in leaves are taken into consideration. Con- spicuous differences such as are illustrated in Fics. 1—12 may occur at times in different leaves of the same taxon, clone or individual plant. For example, the conspicuous differences between Fic. 7 and Fic. 8 occur in two leaves from the same herbarium specimen. It is evident that differ- ences may occur in leaves during successive stages of their maturation, in * This investigation was supported by a grant from the National Science Founda- tion. I am indebted to the American Philosophical Society for the loan of a Wild microscope. 1961] BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, III 335 leaves of different sizes, in leaves of plants of different ages, and in those of plants grown under differing environmental conditions. Therefore, the task of securing strictly comparable material of different taxa becomes a difficult and extremely laborious one. Although a majority of the crystals of calcium oxalate in the leaves of these genera occur in the form of druses having radially oriented individual crystals of varying forms and sizes, aberrant: types of crystallization are of not infrequent occurrence. Particularly where the crystals are very abundant (Fic. 3), some of the cells of the mesophyll may contain a single crystal, several independent crystals or aggregations of crystals that are not perfectly oriented radially. CRYSTALS OF CORTEX AND PITH Druses of conspicuously stellate (Fics. 13, 15) or more rotund (Fics. 14, 16) form occur in the pith and cortex of all three genera of the leaf- bearing Cactaceae. The size, form, and number of the druses is highly variable, as in the case of the leaf. One or more independent crystals of varying size and form may occur at times in some of the cells of the pith and cortex. It is possibly significant taxonomically that there appears to be marked exaggeration of this tendency in the much expanded pith of large basal stems of Pereskia conzattii Britt. & Rose (Fic. 17) and in the cortex of ae older stems of Pereskia colombiana Britt. & Rose and P. guamacho The most ices! and reliable difference, and one of considerable phylogenetic and taxonomic significance, is the concentration of rotund forms of druses in the outermost layer of the cortex of Pereskiopsis and QOuiabentia. In the case of Pereskia sacharosa Griseb., P. grandifolia Haw.., P. bleo DC., P. corrugata Cutak, P. tampicana Web., and allied taxa, young stems of varying diameter, before periderm formation, have a thick cuticle (Fic. 25). The cells of the cortex decrease in size toward the epidermis. Although their cell walls increase somewhat in thickness toward the exterior of the cortex, no clearly differentiated collenchymatous layers are formed subtending the epidermis. The druses are diffusely distributed in the cortex (Fic. 25). In comparable stems of Pereskiopsis and Quta- bentia conspicuous layers of subepidermal collenchyma are formed (Fic. 27). The outermost layer of the collenchymatous zone is packed with druses of rotund, rather than conspicuously stellate, form (Fics. 19, 23, 26). The larger druse-bearing cells tend to bulge the relatively tenuous cuticle and slender epidermal cells outward (Fic. 27). Apparently with increasing succulence in Pereskiopsis and Quiabentia the inner cells of the cortex tend to become thinner walled, whereas the outer ones become thicker walled and collenchymatous. The first cork cambium or phellogen usually arises by periclinal divisions of epidermal cells and the periderm intervenes between the remainder of the epidermis and the outermost layer of crystal-bearing collenchyma. Thus, the latter layer persists after 336 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII periderm formation has occurred and may be detected for some time in older stems. In the case of Pereskia aculeata Mill., P. colombiana, P. guamacho, P. cubensis Britt. & Rose, P. portulacifolia Haw., P. diaz-romeroana Card., P. weberiana K. Schum., P. conzattit, P. autumnalis (Eichlam) Rose, and P. nicoyana Web., periderm formation commonly tends to be precocious and the epidermis and the outer layers of the cortex are rapidly modified. Although the stems of these pereskias do not form such clearly and strongly differentiated collenchymatous layers as occur characteristically in Peres- kiopsis and Quiabentia, they do exhibit at times a tendency for the outer cortical cells to have thicker walls. Particularly in P. diaz-romeroana and P. weberiana some druses of rotund form may occur in the outermost cells of the cortex. This suggests that in these species there may be a tendency toward structural changes such as become exaggerated and dominant in Quiabentia and Pereskiopsis. In view of the occurrence of so-called crystalliferous hypodermal layers in other cacti, the phenomenon merits detailed consideration in subse- quent discussion of salient trends of phylogenetic specialization in the Cactaceae as a whole. CRYSTALS OF THE SECONDARY PHLOEM AND XYLEM Druses of varying degrees of abundance occur in the secondary phloem of the three leaf-bearing genera. Although usually of spherical form, they may at times exhibit curious deviations in elongated parenchymatous ele- ments (Fic. 21). In such cells the druses may have a markedly elongated form and have conspicuously larger crystals at their upper and lower ex- tremities (Fic. 18). As in the case of the leaf, pith and cortex, cells having one or more independent crystals are of sporadic occurrence. However, it is again noteworthy in this connection that in large stems of P. colombiana and P. guamacho cells with single crystals may predominate in the older secondary phloem, druses being relatively infrequent in comparison. Pronounced differences in forms of crystallization of calcium oxalate occur in the rays of the secondary xylem. Where the rays are composed of cells having thick, strongly lignified walls, as in many pereskias, crys- tals either are absent or, when present, occur in a cell singly (Fic. 22) or as several independent crystals, druses being absent. On the contrary, where the rays or parts of them are composed of thin, unlignified cells, a varying number of the thin-walled cells form druses (Fic. 20). It is significant in this connection from a phylogenetic point of view that changes in the rays and forms of crystallization within them appear to be asso- ciated in some way with tendencies toward increasing succulence of stems and roots This is another phenomenon which deserves detailed considera- tion in subsequent discussion of salient trends of phylogenetic specializa- tion in the Cactaceae as a whole. 1961] BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, III 337 SPHEROCRYSTALS AND STRIATED ISOTROPIC BODIES Mobius (1885) early recorded the occurrence of spherocrystals of cal- cium oxalate in Cactaceae. His illustrations of the structural features of these crystalline bodies are reproduced by Kohl (1889), and two of them by Solereder (1899) and by Metcalfe and Chalk (1950). A few finely striated bodies that are birefringent in polarized light occur sporadically in some of my collections of leaf-bearing cacti that are pre- served in formalin-acetic-alcohol. Where they occur in small cells, and in one case in the lumen of vessels close to the cambium, they tend to be fan-shaped or hemispherical in surface view, rather than spherical as in large-celled tissue (Fic. 29). Where a cluster of cells contains them, the striations of each body are radially oriented toward the center of the aggregation of cells. In the case of the larger spherical forms, not only are they radially striated, but also at times they exhibit concentric struc- ture as well, thus resembling certain of the spherocrystals illustrated by Mobius. They also show at least superficial resemblances to the finely striated central parts of some of the more rotund forms of druses (Fic. 16). Furthermore, particularly in the case of the crystalliferous subepidermal layer of Pereskiopsis, the druses not only exhibit a radially striated cen- tral part (Fic. 23) but also not infrequently are concentrically layered internally (Fic. 28). Such rotund forms of druses differ from the sphero- crystals (Fic. 29) in being jacketed by a layer of large, more or less protuberant crystals. This raises a question in need of detailed investi- gation. Are there transitional forms of crystallization of calcium oxalate between typical stellate druses and structures commonly referred to as spherocrystals? Or are there fundamental differences in chemical com- position involved ? In contrast to these relatively infrequently occurring anisotropic bodies are the dark brown, striated, isotropic ones commonly visible during early stages of the clearing of dried leaves in 3 per cent sodium hydroxide at 56° C., ie., prior to their bleaching and dissolution. Although these bodies vary greatly in size and form, they tend to be more or less conspicuously striated. Some of them exhibit concentricities as well as radial striations (Fic. 30), thus in surface view resembling spherocrystals (Fic. 29). Others are composed of small cells the brown contents of which have striations oriented radially toward the center of the composite mass (Fic. 32). Where the constituent cells of the mass are larger, the radial orientation of the striations is more clearly visible (Fic. 33). In surface view such a mass of isotropic composition bears at least a superficial resemblance to Mobius’ figures 10-13 of spherocrystals. Under more pro- longed treatment in sodium hydroxide brown structureless globules fre- quently tend to exude from the striated bodies (Fic. 31), or the whole body may lose its striated structure. Obviously, such bodies are not com- posed of crystalline calcium oxalate, although they occasionally may con- tain some in the form of “crystal sand.” It should be emphasized in these connections that the brown bodies in dried leaves differ from the sphero- 338 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XLII crystals dealt with by Mobius (1885), Kohl (1889), Lauterbach (1889), and Michaélis (1896) in being isotropic rather than birefringent in polar- ized light. It is important to determine not only the chemical and physical con- stitution of crystalline and isotropic bodies that occur in material of cacti, but also to obtain reliable evidence regarding the occurrence of such bodies in normal living tissue prior to desiccation, preservation in alcohol or formalin-acetic-alcohol and subsequent treatments with varying reagents. At present, there is adequate evidence regarding the common occurrence of druses of calcium oxalate in normal unmodified tissues of various Cac- taceae, as well as in the leaf-bearing ones. However, there is evidence which indicates that the druses are not composed solely of calcium oxalate, but contain organic substances as well, particularly in their central parts. For example, when druses are treated with haematoxylin prior to the use of reagents which dissolve constituents such as residues of nuclei, proto- plasts, etc., their centers stain very dark (Fics. 14, 19, 20, 23). Further- more, if the calcium oxalate is dissolved prior to such treatment, a deeply stained central part of the druse persists. (Compare Fics. 23 and 24.) The spherocrystals studied by Mobius are stated to occur, at least in certain cases in living tissue, whereas those observed by Lauterbach and Michaélis occurred in material preserved in alcohol The formation of dark brown isotropic bodies in the leaves of Pereskia, Pereskiopsis, and Quiabentia appear to be induced by desiccation or other modifying treat- ments. However, their formation in all three genera is indicative of the presence of similar organic substances and may ultimately prove to be of some taxonomic significance in comparisons between the leaves of Cac- taceae and those of other families of the dictotyledons, e.g., those included in the Centrospermae. Thus, their chemical composition and factors in- volved in their formation merit detailed investigation. CONCLUSIONS The ranges of variability in the size, form, and number of druses in the leaves of Pereskia, Pereskiopsis, and Quiabentia are so extensive as to render their use difficult and excessively laborious in differentiating the three genera and the species which occur within them. The most constant and reliable differences from phylogenetic and taxo- nomic points of view are those that occur in the outer cortex of young stems prior to periderm formation. Pereskiopsis and Quiabentia, in con- trast to Pereskia, have conspicuous collenchymatous layers, the outermost of which is packed with druses of rotund form subtending the epidermis. Another difference of considerable phylogenetic and physiological in- terest occurs in the rays of the secondary xylem of both stems and roots. Where the rays are composed throughout of cells with thick, heavily lignified walls, as in less modified forms of woody dicotyledons, deposi- tion of calcium oxalate occurs in the form of single crystals or several independent ones, aggregation into druses being absent or of rare occur- 1961] BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, III 339 rence. On the contrary, where the rays or parts of them are composed of cells with thin unlignified walls, such rays or parts contain more or less numerous druses, The changes in wall thickness, lignification and form of crystallization appear to be correlated in some manner with tendencies toward increasing succulence, as in Pereskiopsis and Quia- entia The changes in the outer cortex of Pereskiopsis and Quiabentia and in the rays of the leaf-bearing cacti appear to be significant in any subse- quent discussion of salient trends of phylogenetic specialization in the Cactaceae as a whole LITERATURE CITED Kou, F. G. 1889. Anatomisch-physiologische Untersuchung der Kalksalze und Kieselsaure in der Pflanze. N. G. Elwert, Marburg. LAUTERBACH, C. 1889. Untersuchungen uber Bau und Entwicklung der Sekret- behalter bei den Cacteen. Bot. Centralbl. 37: 257-264, 289-297, 329-336, 369-375, 409-413. MetTcaLre, C. R., & L. CHALK. 1950. Anatomy of the dicotyledons. Clarendon Press. Oxford. MicHaELIs, P. 1896. Beitrage zur vergleichenden Anatomie der Gattungen Echinocactus, Mamillaria und Anhalonium. Doctoral Thesis, Halle. saa M. 1885. sported stalle von Kalkoxalat bei Cacteen. Ber. Deutsch. . Ges. 3: 178-18 ee M. J. ae sepia zur Anatomie der Cacteen. Mém. Acad. Imp. Sci. St. Petersbourg 4: -380. SOLEREDER, H. 1899. ee Anatomie der Dicotyledonen. F. Enke, Stuttgart. EXPLANATION OF PLATES PLATE I s. 1-6. Parts of leaves cleared in 3 per cent sodium hydroxide at 56°C., dehydrated, and mounted in diaphane. Druses photographed in polarized light between crossed nicols at a magnification of 150. 1, Quiabentia zehntneri (Britt. & Rose) Britt. & Rose [| Rose, N.Y. Bot. Gard. 42830]; 2, Pereskiopsis gatesii pe Lager Cactus Inc.|; 3, Pereskiopsis ot Britt. & Rose [Safford, US. Herb. 2263439]; 4, Pereskia bleo DC. [Pennell 4759 (cH)]; 5, eee Sieg ara [Jiménez 2578 (us)]; 6, sna aculeata [Duss 3073 (GH) |. PLATE II . 7-12. Parts of leaves cleared in 3 per cent sodium hydroxide at 56°C., dehydrated, and mounted in diaphane. Druses photographed in polarized light between crossed nicols at a magnification of 150. 7, Pereskia aff. guamacho [H. H. Smith 1886 (GH) ]; 8, Another leaf from Smith 1886; 9, Pereskia grandi- olia [Mexia 4129 (GH) |; 10, Pereskia grandifolia Ele 7270): 11, Pereskia conzatti | Boke]; 12, Pereskia nicoyana [Mo. Bot. Gard.]. 340 JOURNAL OF THE ARNOLD ARBORETUM [voOL. XLII PLATE III Fics. 13-18. Druses photographed in green light at a magnification of 425. 13, Pereskia diaz-romeroana [Cardenas |, stellate druse from leaf; 14, Pereskia tampicana | Boke B-17|, druse from cortex of stem; 15, Pereskia aff. guamacho [Smith 1886 (GH) |, stellate druse from leaf; 16, Pereskia grandifolia | Mexia 4129 (GH) |, druse from leaf; 17, Pereskia conzattii | Dressler|, crystals from pith of large basal stem; 18, Pereskiopsis chapistle Britt. & Rose {Boke B-3}, druse from phloem parenchyma. PLATE IV Fics. 19-24. Varied forms of crystals in stems. 19, Pereskiopsis chapistle [Mo. Bot. Gard.], tangential longitudinal section stained in haematoxylin and Sudan III and mounted in glycerin, showing druses in outer layer of collenchyma subtending the epidermis, as seen in polarized light, X 150; 20, Pereskiopsis chapistle | Boke B-3}, se ieee longitudinal — of xylem ‘tained j in haema- toxylin and safranin and mounted in diaphane, showing stellate druses in thin- walled, unlignified ray ete as seen in ‘elas light, & 150; 21, Pereski- opsis chapistle | Boke B-3|, tangential longtitudinal section of phloem, showing ruses in elongated parenchymatous elements, as seen in polarized light, x 150; 22, Pereskia colombiana Britt. & Rose | Romero], tangential coleee ma section of xylem showing single crystals in thick-walled lignified ray cells, in polarized light, & 150; 23, Pereskiopsis chapistle |Mo. Bot. Gard. ], as druse from Fig. 19 photographed in green light, & 425; 24, Pereskiopsis scandens Britt. & Rose | N.Y. Bot. Gard. 50085], organic residues of such a druse as illus- trated in Fig. 23 after removing the crystals of calcium oxalate in a mixture of equal parts of 10 per cent nitric and 10 per cent chromic acids, staining in haematoxylin and safranin and mounting in glycerin, photographed in green light, X 425. PLATE V Fics. 25-27. Transverse sections of stems. , Pereskia sacharosa | Mo. Bot. Gard.], stained with haematoxylin and nme ‘mounte d in diaphane, showing thick cuticle, conspicuous epidermis and diffuse distribution - druses in cortex, photographed in green light, X 150; 26, Pereskiopsis scandens Britt. & Rose [N.Y. Bot. Gard. 50085], unstained section mounted in diaphane and photo- graphed in polarized light, ees concentration of druses in subepidermal layer of outer collenchyma, « 150 , The same, section stained in haematoxylin and safranin, mounted in ee ae photographed in green light, showing collenchymatous layers in contrast to Fig. 25, and tenuous epidermis bulging outward owing to the development of druses shown in Fig. 26, 150. PLATE VI Fics, 28-33. Crystalline and isotropic bodies in green light; Fig. 28, 1000; es 29, X 500; Figs. 30-33, & 260. 28, Pereskiopsis aquosa (Web.) Britt. & euare Bot. Gard.], concentrically layered druse; 29, Pereskia guamacho avr Gard.], anisotropic “sphaerocrystal” from pit ith: 30, Pereskia aff. gua- macho |Smith 1886 (GH)]|, brown, radially striated, concentrically layered, ee mass, foliar; 31, Pereskia aculeata {| Ferreira 2316 (GH) |, exudation of structureless brown droplets from radially striated, isotropic mass, foliar: 32, Pereskia grandifolia [Moran 7270], mass of small cells with striated, brown, isotropic contents, foliar; 33, Pereskia colombiana [Record 16495 (GH) |, mass of large cells with striated, brown, isotropic contents, foliar. Jour. ARNoLD ArB. VoL. XLII PLATE I i. BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, ITI Jour. ARNoLD Ars. VoL. XLII BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, III] ’ x . ~ ; " “yy Jour. ARNOLD Ars. VoL. XLII PiaTeE III 2 BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, III Jour. ArRNotp Ars. VoL. XLII PraTe IV te, Aes t ot. oy a a8 ae: a" {24 onfaonc BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, III Jour. Arnotp Ars. VoL. XLII PLATE V 95 e eS 7 ' @ @ oO 2 i 7 tot, ) a we rt rr a Lo : ‘\ «2 +. i rs ‘ a a 4 .“ ‘ou —— 7 home BAILEY, ANATOMY OF LEAF-BEARING CACTACEAE, III PuaTeE VI Jour. ARNOLD Ars. VoL. XLII LEAF-BEARING CACTACEAE, III ANATOMY OF BAILEY, 1951 | PRAKASH & BARGHOORN, MIOCENE WOODS, II 347 MIOCENE FOSSIL WOODS FROM THE COLUMBIA BASALTS OF CENTRAL WASHINGTON, II U. PRAKASH AND Etso S. BARGHOORN IN A PREVIOUS PAPER (1961) the authors described in detail twelve species of woods, representing nine genera (Acer, Diospyros, Quercus, Liquidambar, Carya, Albizzia, Gleditsia, Platanus, and Ulmus) from the Columbia Basalts of the area of Vantage, Washington. The silicified woods described represent a small fraction of an extensive collection made in 1954 by Mr. Jay O’Leary, while he was a student in Harvard College. The present investigation is a continuation of the systematic study of this collection. In this paper we are concerned with the identification and description of six additional species representing four genera not recorded in the first paper. It is probable that the total number of genera preserved in the basalts of the Vantage area may be in excess of 30 of which more than two-thirds are genera of dicotyledonous trees. In a previous com- munication, Beck (1945) listed 27 genera determinable in the Vantage and nearby stratigraphically equivalent areas. It is of interest to note that only slightly more than one-third of these genera currently exist in the flora of the general region today (Wolfe and Barghoorn, 1960). Although the age of the flora with which our study is concerned is generally regarded to be Upper Miocene, the high percentage of genera represented in this fossil flora but now exotic to the region would tend to indicate a lower stratigraphic position in the Miocene sequence (Barg- hoorn, 1951; Wolfe and Barghoorn, 1960). This reconsideration of the age of the Vantage (Russell Forests) flora is also supported by compari- son of generic composition with that of the Latah flora, a large flora the composition of which is known from leaf impressions rather than silicified woods. The Latah flora is comprised of genera of which only slightly more than one-third are extant in the geographic region today. Statistical examination of thirteen floras of the western United States generally re- garded as Upper or Late Miocene, on the other hand, indicates that be- tween 50 to 67 per cent of their genera are yet in the geographic region of the fossil floras. FAGACEAE 1. Quercus sahnii Prakash & Barghoorn, sp. nov. (Fics. 16-19) Growth rings: Distinct; wood ring-porous. Transition between early and late wood abrupt. (Fic. 19). Vessels: Large vessels in the early wood (Fic. 19), 192-416 p in tan- gential diameter (mean diameter 269 »), solitary, forming a conspicuous band 1 or 2 pores in width, thin walled, vessels round to oval, occasionally 348 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XLII elliptical in cross-section. Vessel elements 240-540 ,» long. Perforation plates exclusively simple, usually horizontal. Intervascular pit-pairs alter- nate, bordered and round to oval in shape. Tyloses absent. Late-wood vessels small (Fic. 19), 30-111 » in tangential diameter (mean diameter 0 »), usually scattered in radially aligned, flame-shaped tracts of light- colored tissue, chiefly solitary, open, thick walled (Fic. 17), rounded in cross section, Perforation plates and intervascular pit-pairs similar to early wood vessels. Parenchyma: Abundant, paratracheal, metatracheal-diffuse, and meta- tracheal. Paratracheal parenchyma intermingled with the tracheids, (1) forming part of the conjunctive tissue between the early wood vessels and the rays, (2) composing most of the tissue in the flame-shaped tracts in which late wood vessels are embedded. Metatracheal-diffuse parenchyma (Fics. 17, 19) restricted to fibrous tracts and exhibiting a tendency towards aggregation into short, irregular lines of metatracheal parenchyma towards the outer part of the ring. Crystalliferous parenchyma frequent, com- posed of somewhat enlarged cells containing single crystals (Fic. 18). Xylem rays: Homogeneous (Kribs Homogeneous Type 1), composed of procumbent cells; both narrow and broad (Fic. 16); narrow rays very numerous, uniseriate, 2-27 cells high; broad rays 15-28 cells wide and very high, separated by numerous narrow rays (Fic. 16) often dissected into smaller units. Ray cells oval to polygonal in tangential section and 10-30 » in diameter, sometimes with crystals. Imperforate tracheary elements: Libriform fibers and vasicentric tra- cheids. Thick-walled fibers (Fics. 17, 19) well developed through the spaces among the flame-shaped tracts of late wood vessels, polygonal in cross section, 12—21 » in diameter, nonseptate; pits simple. Vasicentric tracheids intermingled with parenchyma, (1) composing most of the con- junctive tissue between the early-wood vessels and the rays and (2) forming part of the flame-shaped tracts in which late-wood vessels are embedded; pits bordered. Marertat. A single specimen of silicified, mature, secondary xylem measuring 6.5 « 2.5 & 7cm Hototype. Palaeobotanical Collections, Harvard University, No. 54968. This species is named in commemoration of the late Professor Birbal Sahni, founder-director of the Institute of Palaeobotany, Lucknow, India. There is close agreement in all details of structure of this wood with that of Quercus. Among the different types of oaks, the fossil wood be- longs to the red-oak group as indicated by the presence of thick-walled, rounded vessels in the late wood. A survey of the available woods of Quercus indicates that the fossil does not show close resemblance to any of the species represented in the wood collections of the Arnold Arboretum (86 species) nor to published descriptions of 15 other species (Sudworth & Mell, 1911, pp. 22-56, figs. 71-48; Kanehira, 1921a, pp. 67-71, Pl. 8. figs. 44, 45; Kanehira, 1924, p. 52; Kanehira, 1940; Pearson & cis 1932, pp. 981-996, figs. 299-304; Reyes, 1938, pp. 60-62, Pl. 4, figs. 1, 1961] PRAKASH & BARGHOORN, MIOCENE WOODS, II 349 Metcalfe & Chalk, 1950, pp. 1311-1314, figs. 313G, L; Greguss, 1947, pp. 41-46, figs. 18-27). Fossil woods assigned to Quercus have been listed by Edwards (1931) and Boeshore and Jump (1938) and were also referred to in our previous paper (Prakash & Barghoorn, 1961). More than 500 species and varieties of Quercus, ranging from low shrubs to stately forest trees, are known. There are about 50 species of Quercus native to the United States. The species belonging to the red-black oak group are widely distributed almost throughout this country. Those growing in the western United States are QO. Kelloggii Newb., of the mountain ranges of California and western Oregon, Q. tomentella Engelm., on the islands along the coast of southern California and Guadeloupe Island and extending farther south in Mexico, Q. douglasii Hook. and Arn., a tree of wide range in California, and vari- eties of O. dumosa Nutt., on the coastal islands of southern California and Mexico (Little, 1953). Species of the red-black group of the genus are not known from the state of Washington. HIPPOCASTANACEAE 2. Aesculus hankinsii Prakash & Barghoorn, sp. nov. (Fics. 20-23) Growth rings: Barely distinct; wood diffuse-porous (Fic. 22). Vessels: Small, 35-80 p in tangential diameter (mean diameter 54 y), more crowded in the early spring-wood, decreasing slightly in size through the late wood, where they are fewer and quite apart (Fic. 22). Vessels solitary and in radial multiples of 2—5 (mostly 2 or 3) cells, thin walled, the solitary vessels oval-angular to elliptical-angular in cross section. Ves- sel elements up to 500 » long with oblique, often tailed ends; spiral thick- enings occasionally present. Perforation plates exclusively simple. Inter- vascular pit-pairs (F1c. 21) bordered, alternate, orbicular to oval or angular where crowded, 5-7 p» in diameter, apertures oval to slitlike, sometimes oblique. Vessels open. Parenchyma: Appearing to be terminal and paratracheal. Terminal parenchyma forming a continuous or interrupted line. Paratracheal paren- chyma quite scanty. Xylem rays: Homogeneous, composed of procumbent cells (Fic. 23), uniseriate, 2-32 (mostly 8-15) cells high, showing a tendency for storied arrangement; up to 16-20 rays per mm. Ray cells thin-walled, oval to vertically elongated in tangential section. Imperforate tracheary elements: Fiber tracheids rectangular to poly- gonal in cross section (Fic. 20), 12-23 » in diameter, walls thin to slightly thick; nonseptate; pits bordered. MateriAt, A single specimen of silicified, mature, secondary xylem measuring 8 X 2 & 6 cm. Preservation poor. Ho.otyPe. Palaeobotanical Collections, Harvard University, No. 56240. The species is named after Mr. Frank W. Hankins, of Collegeville, Penn- sylvania, who kindly gave us the fossil wood for study. 350 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII The combination of structural features exhibited by this wood indicates its affinity with the genus Aesculus and its nearest affinity with the extant species Aesculus pavia L. Thin sections of the woods of eighteen species of the genus Aesculus were examined in the course of identification, in- cluding Asiatic and European, as well as North American, representatives of the genus. The fossil under consideration differs from A. pavia in only minor details, the most conspicuous of which are the less frequent occur- rence of vessel multiples throughout the growth ring and greater crowding or compaction of vessels in the early wood. Aesculus hankinsii, as far as the authors are aware, represents the first record of this genus occurring in the form of fossil wood. Its identification supports the conclusions of Beck (1945) on the presence of Aesculus in the Vantage and adjacent areas. The genus, as presently constituted, in- cludes thirteen species of trees and shrubs of temperate regions, occur- ring in southeastern Europe, India to Japan, and North America (Hardin, 1957). Aesculus pavia is a shrub, sometimes attaining the dimensions of a small tree, occurring in woods and along streams from eastern North Carolina to Florida, westward to the Edwards Plateau of Texas, and north in the Mississippi Valley to southern Mlinois. In present geographic distribution the nearest species is A. californica (Spach) Nutt., a small tree of the California Coast Ranges and Sierra Nevada foothills (Hardin, 1957) JUGLANDACEAE 3. Juglans fryxellii Prakash & Barghoorn, sp. nov. (Fics. 12-15) Growth rings: Distinct, delineated by narrow band of thick-walled, flattened fibers; wood diffuse-porous to semi-ring-porous (Fics. 13, 14). Vessels: Large in early wood, gradually decreasing in size through the late wood with a marked tendency toward diagonal arrangement (Fics. 13, 14), 51-208 » in tangential diameter (mean diameter 142 uw), Chiefly soli- tary, sometimes in radial groups of 2 or 3 cells, evenly distributed and closely placed, thin walled, the solitary vessels oval, elliptical, or some- times slightly irregular in cross section (probably due to pressure during mineralization). Vessel elements of varying lengths, with oblique end walls. Perforation plates not ascertainable, the vessels mostly being plugged with gummy or resinous deposits (Fic. 13). Intervascular pit- pairs (Fic. 12) bordered, alternate, or slightly irregular, 10-14 » in diam- eter, usually angular where crowded, pit apertures oval to linear. Parenchyma: Metatracheal and terminal. Metatracheal parenchyma (Fics. 13, 14) in uniseriate (rarely locally biseriate), somewhat wavering and more or less parallel tangential lines which are almost always continu- ous, sometimes small and broken. Terminal parenchyma uniseriate and continuous. Parenchyma cells sometimes enlarged and crystalliferous, usually with black, resiniferous substance, Xylem rays: Homogeneous to decidedly heterogeneous, 1—3 cells wide (Fic. 15); uniseriate rays few, 2-11 cells high; multiseriate rays mostly 1961] FRAKASH & BARGHOORN, MIOCENE WOODS, II 351 2 or 3 cells broad and 0.56 mm. high; multiseriate rays mostly hetero- eneous, sometimes homogeneous and composed of procumbent cells; multiseriate heterogenous rays with procumbent cells in the broad cen- tral part and one to several rows of upright cells at the uniseriate ends; end-to-end ray fusion sometimes present; up to 7-12 rays per mm. Ray cells thin walled, oval to vertically elongate in tangential section, 10-15 p in diameter. Imperforate tracheary elements: Fiber tracheids, polygonal to slightly irregular in cross section (Fic. 13), 12-22 » in diameter, the walls slightly thickened; nonseptate; pits bordered. MarerIAL. Few, small specimens of silicified, mature, secondary xylem. Hotorypr. Palaeobotanical Collections, Harvard University, No. 55045. This species is named after Mr. Roald Fryxell, Curator, Ginkgo Petrified Forest Museum, Vantage, Washington. The combination of structural features of the fossil, after extensive comparison with modern woods, indicates its affinity with the genus Juglans. Among the 15 species of this genus available for direct comparison, the closest approximation in structure is to be found in Juglans nigra L. and J. < notha Rehder, to both of which the fossil wood shows a close resemblance, with only some differences in details of ray structure and parenchyma distribution. One other conspicuous feature of the fossil is the presence in the vessels of a brownish-black substance plugging the vessel lumina. Fossil woods assigned to the family Juglandaceae are known from a wide range of localities in the Northern Hemisphere, in both Old and New Worlds. The genera have been designated as Juglandinium Unger (1845), Juglandoxylon Kraus (1882), Jugloxylon Stopes & Fujii (1910), Juglans, Pterocarya, and Carya. Species described are Juglandinium longiradiatum Vater (Kaiser, 1890), from the Lower Senonian of Germany; J. mediter- raneum Unger (Kaiser, 1890), and J. schenkit Felix (Kaiser, 1890), both from the Tertiary of Hungary; J. triebelii (Caspary ) Nagel (1915), from the (?) Tertiary of Prussia; J. wichmannii (Hofmann) Kaiser (1890), and J. zuriense (Falqui) Edwards (1931), from the Miocene of Sardinia; Juglandinium spp. (Kaiser, 1890; Schonfeld, 1930; Edwards, 1931), from the Lower Senonian of Brunswick, Upper Cretaceous of western Canada, and the Oligocene of Saxony; Jugloxylon homanum Stopes & Fujii (1910), from the Upper Cretaceous of Japan; Juglandoxylon prince ps (Ludw.) Miiller-Stoll (1938) and J. schadleri Hofmann (1952), from the Upper Oligocene of Prambachkirchen, Austria; Juglans sp. (Beck, 1942a) from the Oligo-Miocene of Oregon; Carya sp. (Beck, 1942b) and C. tertiara (Prakash & Barghoorn, 1961), from Vantage, Washington; Carya protojaponica Watari (1952) and Pterocarya rhoifolia Sieb. & Zucc. (Watari, 1952), both from the Miocene of Japan; and Pterocarya sp. (Beck, 1942c) from the Miocene of Washington. Juglans, with about 15 generally recognized species, is indigenous to China, Japan, India, Persia, the United States, Mexico, Central America, 352 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII the West Indies, and the Andean region of South America from Venezuela and Colombia to Argentina (Stark, 1953). Of the six species native to the United States, Juglans cinerea L. is widely distributed throughout most of the eastern half of the United States. J. nigra has much the same range as J. cinerea (Munns, 1938, maps 61, 62; Little, 1953). The four western species are Juglans microcarpa Berlandier, J. californica S. Wats., J. major (Torr.) Heller and J. hindsii Jeps. None is native to the state of Washington (Little, 1953). LEGUMINOSAE 4. Leguminoxylon occidentale Prakash & Barghoorn, sp. nov. (Fics. 24-27) Growth rings: Distinct due to presence of terminal parenchyma; wood diffuse-porous (Fic. 27). Vessels: Large to medium sized, 70-225 » in tangential diameter, chiefly solitary (Fic. 27), often paired and rarely in radial or tangential rows of 3 cells, almost always encircled by sheaths of parenchyma cells, thick walled, the solitary vessels mostly elliptical, sometimes circular in cross section. Vessel elements about 200-450 » long with usually truncate, some- times tailed ends. Perforation plates exclusively simple. Intervascular pit-pairs (Fic. 25) bordered, alternate, oval to orbicular or angular where crowded, about 8-10 » in diameter with lenticular apertures, often coales- cent. Vessels almost always plugged with dark brown gummy substance. Parenchyma: Paratracheal and terminal. Paratracheal parenchyma forming vasicentric sheaths 2 or 3 (4 or 5) cells thick round the vessels (Fics. 24, 27), mostly becoming aliform with short wings, often con- fluent and connecting 2 or 3 pores tangentially or rarely forming radial extensions uniting vessels between the same pair of rays. Terminal paren- chyma 1-3 (sometimes more) cells thick, forming a sharply defined line (Fic. 24) at the growth ring. Crystalliferous parenchyma rarely present. Xylem rays: Homogeneous (Kribs Homogeneous Type 2), composed of procumbent cells, 1-5 cells wide (Fic. 26), unstoried; uniseriate and biseriate rays very rare, the uniseriate rays only few cells high; multi- seriate rays mostly 3-5 cells broad and 0.75 mm. high; up to 6-10 rays per mm. Ray cells circular to oval in tangential section and 10-15 yp in diameter. Imperforate tracheary elements: Libriform fibers polygonal in cross section, 8-15 yw in diameter, walls thick (Fic. 24) with a small lumen; nonseptate; pits apparently simple MATERIAL. A single small specimen of silicified, mature, secondary xylem. HoLotype. Palaeobotanical Collections, Harvard University, No. 55076. The structural features of this fossil clearly indicate its affinity with the woods of Leguminosae and most closely with Acacia, Cassia, and other genera of similar wood structure. However, owing to the diversity, great 1961] PRAKASH & BARGHOORN, MIOCENE WOODS, II 255 number of species, and range of structural variability within these genera, it does not seem possible to establish with any assurance its precise generic position. The nearest approach to the structure of the fossil which we have been able to establish is the genus Acacia and, within this genus, the species A. ferruginea DC. Our survey included the study of thin sections of the wood of nineteen species of Acacia and published descrip- tions of nine additional species (Kanehira, 1921b, pp. 89-90, Pl. 18, fig. 106: Pearson and Brown, 1932, pp. 439-445, 450-452, figs. 150, 151, 154; Kribs, 1959, pp. 60-62, figs. 156, 158, 161, 394). One aspect of the fossil which renders its identification more difficult is the tangential compression failure which preceded mineralization, thus exaggerating the ellipticity of the vessels as seen in transverse section. In view of these facts, it seems more desirable to designate the fossil to family, rather than to genus. Its affinity with the Leguminosae seems unquestionable. If the wood actually belongs to the genus Acacia, it is of some interest to note the present geographic distribution of the genus in relation to the fossil occurrence. The genus comprises some 450 species of trees and shrubs, of which 300 species are found in Australia alone, while the remaining species are widely dispersed through southeastern Asia, tropical and southern Africa, and tropical America, with extensions into temperate South America and the southern United States. Its nearest occurrence to the Vantage area is in southern California. NYSSACEAE 5. Nyssa eydei Prakash & Barghoorn, sp. nov. (Fics. 1—5) Growth rings: Fairly distinct, delineated = a narrow band of flattened, thick-walled fibers; wood diffuse- -porous (Fic. | Vessels: Small, 45-90 yp in tangential diameter (mean diameter 62 ,), solitary (Fic. 1) and in small radial multiples and groups, thin-walled, polygonal to oval (Fics. 1, 2) in cross section, quite uniform in size, numerous and fairly evenly distributed. Vessel elements up to 1700 yp long with tapering ends. Perforation plates exclusively scalariform with numerous thin bars (Fic. 5). Intervascular pit-pairs bordered, opposite, oval-rectangular, usually 5-10 » in diameter with short linear apertures (Fic. 4). Vessels open. Parenchyma: Not abundant, mostly diffuse, sometimes in short meta- tracheal rows (Fic. 2), occasionally paratracheal. Parenchyma cells mostly rectangular, sometimes crystalliferous. Xylem rays: Heterogeneous, 1-3, rarely 4 cells wide (Fic. 3); uni- seriate rays quite common, 2-29 (often 4-12) cells high; multiseriate rays mostly 3 cells broad and 1.75 mm. high; uniseriate rays and margins of multiseriate rays composed of few to many rows of mostly upright and square cells; vertical ray fusions common; up to 12-15 rays per mm. Ray cells thin-walled, round to vertically elongated in tangential section. 10-20 p in diameter. Imperforate tracheary elements: Fiber tracheids mostly rectangular to 354 JOURNAL OF THE ARNOLD ARBORETUM [voL. XLII polygonal in cross section (Fic. 2), large, 20-45 » in diameter, walls thin with a fairly large lumen; nonseptate; pits bordered with slitlike apertures. Matertat. A single specimen of silicified, mature, secondary xylem measuring 8 & 7 & 1.5 cm. Hototype. Palaeobotanical Collections, Harvard University, No. 54972. The species is named for Mr. Richard H. Eyde, who is currently working on the fossil endocarps of Nyssa. A survey of the structural features of the woods of Nyssa indicates that the nearest affinity of the fossil is with Nyssa ogeche Bartr. The structure of Vyssa eydei and that of N. ogeche are virtually identical in all major features. A few minor differences exist, however. The inter- vascular pit-apertures often tend to fuse in Nyssa ogeche, a feature not found in N. eydei; the metatracheal parenchyma is more frequent in Nyssa eydei than in N. ogeche, and the rays tend to have longer uni- seriate wings which frequently fuse. Despite the widespread occurrence of Nyssa and related genera repre- sented as pollen and endocarp remains, the wood of Nyssa is very poorly represented in the fossil record. Its existence in the Russell Forests flora was noted by Beck (1945). The form genus Nyssoxylon Midel includes the genera Nyssa and Davidia. Nyssoxylon japonicum Midel (1959), from Japan, although somewhat resembling Nyssa eydei, differs in pos- sessing tyloses in the vessels and in the narrower 1- or 2-seriate rays in contrast to rays 1-3 (rarely 4) seriate in Nyssa eydei. In view of the rather limited distribution of Nyssa in both North America and Asia and since Nyssa ogeche is readily separable on the basis of wood structure from other species of the genus, the occurrence of a species closely resembling N. ogeche in the Tertiary of the north- western United States is quite unexpected. Nyssa ogeche is a large swamp tree of currently limited distribution in swamps of the Coastal Plain of the southeastern United States, extending from South Carolina to north- western Florida. ULMACEAE 6. Ulmus baileyana Prakash & Barghoorn, sp. nov. (Fics. 6-11) Growth rings: Distinct; wood ring-porous. Transition between early and late wood abrupt (Fics. 7, 8). Vessels: Large vessels in the early wood, 160-256 p in tangential diam- eter (mean 190 ,), solitary, arranged in three to four rows, rarely asso- ciated with small vessels (Fic. 7, 8), thin walled, the solitary vessels oval to elliptic in cross section. Vessel elements 130-460 p» long. Perforation plates exclusively simple, horizontal to sometimes slightly oblique. In- tervascular pit-pairs (Fic. 6) bordered, alternate, orbicular or angular where crowded, 9-11 y in diameter, with circular apertures; vessel-paren- chyma pits numerous, more or less arranged in vertical rows, with fairly large oval apertures. Late-wood vessels small (Fic. 7), 50-160 p» in 1961] PRAKASH & BARGHOORN, MIOCENE WOODS, II g55 tangential diameter (mean 81 ,), arranged in more or less continuous, wavy, tangential or oblique bands, usually polygonal due to clustering, occasionally round to oval in shape. Vessel elements with spiral thicken- ings. Perforation plates and intervascular pit-pairs similar to early-wood vessels, except smaller. Thin-walled tyloses common. Parenchyma: Paratracheal and metatracheal-diffuse (Fics. 8,9). Para- tracheal parenchyma occurring as few cells contiguous to the large vessels of early wood (but never forming a continuous sheath) and marginal to and occasionally included in the clusters of wavy bands of late-wood vessels and vascular tracheids. Diffuse parenchyma sparse, scattered among the fibers and occasionally in contact with the xylem rays. Parenchyma cells sometimes crystalliferous. Xylem rays: Homogeneous (Kribs Homogeneous Type 2), composed of procumbent cells (Fics. 10, 11), 1-6 cells wide; uniseriate rays rare, 2-10 cells high; multiseriate rays often fusiform, mostly 4 or 5 cells broad and 1.13 mm. high; up to 5-7 rays per mm. Ray cells slightly thick walled, circular to oval or vertically elongated in tangential section and 10-17 » in diameter. Imperforate tracheary elements: Libriform fibers and vascular tracheids. Libriform fibers polygonal in cross section, 10-22 p in diameter, wall mod- erately to very thick (Fic. 9), with a small to fairly large lumen, non- septate; interfiber pits simple. Vascular tracheids occasionally present both in the early-wood porous zone and in the clusters of late-wood ves- sels. MaterIAL. A single specimen of silicified, mature, secondary xylem measuring approximately 12 X 6 * 6 cm. Hototvpr. Palaeobotanical Collections, Harvard University, No. 54918. This species is named in honor of Irving W. Bailey, Professor of Plant Anatomy, Emeritus, Arnold Arboretum, Harvard University. Structural features of this wood indicate its affinity with the genus Ulmus, two species of which have been described in our previous paper. Sufficient differences in structure exist, however, to justify its separation as a distinct species. The fossil wood closely resembles the wood of the modern species U. rubra Mihl. (U. fudvu Michx.). It differs, however, from U. rubra in the presence of tyloses in the vessels, the larger late- wood vessels, the smaller amount of paratracheal parenchyma associated with the vessel clusters in the late wood, and the more crystalliferous wood parenchyma. In addition, the multiseriate rays are slightly broader and higher in U. baileyana than in U. rubra, and uniseriate rays are rare. DISCUSSION In our first paper (Prakash & Barghoorn, 1961) on the Vantage fossil woods, we indicated that the plants determined generically comprise an eastern American, a western American, and an Asiatic element. The present observations tend to confirm this geographic pattern. Of the genera 356 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLII described, Ulmus and Nyssa occur in the eastern United States, whereas Aesculus, Juglans, and Quercus are found in both the eastern and western parts of the United States. The leguminous wood which is quite com- parable to the genus Acacia indicates a more southern or Asiatic affinity. between Nyssa evdei and N. ogeche. The latter species is sharply de- limited today by both temperature and edaphic conditions such as obtain in the humid swamp-forests of the southeastern United States. It is quite probable, in view of the ecological association of the Vantage genera, that the climate was featured by both higher annual average temperatures and by fairly high summer rainfall. The decimation of the mesophytic forests of the Miocene of central Washington was brought about by a combination of factors of which gradual refrigeration, culminating in Pleistocene glacia- tion was a major factor, In addition, re-entrance into this region in post- glacial times of certain of these genera which possess cold tolerance has been prevented by the barrier of aridity induced by the Cascade uplift. In addition to the kind help of those acknowledged in the preceding paper, the authors are also much indebted for that of Mr. Clarence Scam- mon, who aided the first author greatly in his field excursions in the Van- tage area. The first author also wishes to acknowledge his sincere appre- ciation to the National Academy of Sciences, Washington, D. C., for the financial assistance given during this investigation. DEPARTMENT OF BIOLOGY A 74 ND BOTANICAL Museum. HARVARD UNIVERSITY LITERATURE CITED BarGHOoorN, E. S. 1951. Age and environment: A survey of North American Tertiary floras in relation to paleoecology. Jour. Paleontology 25: 736-744. Beck, G. F. 1942a. Additions to the Gray Ranch wood flora. Geol. News Lett. Portland, Ore. 8: 38-91. -. 1942b. Tertiary juglandaceous woods: 1. Hickory (Carya). Jbid. 103-106. . 1942c. Tertiary juglandaceous woods: 2. Pterocarya. Ibid. 8: 143-145. ———. 1945. Ancient forest trees of the sagebrush area in central Washington. Jour, Forestry 43: 334-338. BorsHore, I., and J. A. Jump. 1938. A new fossil oak wood from Idaho. Am. Jour. Bot. 25: 307-311. 1961] PRAKASH & BARGHOORN, MIOCENE WOODS, II 55% sae oe W.N. 1931. Fossilium Catalogus II: Plantae Dicotyledones (Ligna). Pars 17: 3-96. ee P. 1947. The identification of ri sac dicotyledonous trees and shrubs based on xylotomy. Buda Harpin, J. W. 1957. A revision of the er Hippocastanaceae — II. Brit- tonia 9: 173-195. Hormann, E, 1952. Pflanzenreste aus dem Phosphoritvorkommen von Pram- bachkirchen in Oberésterreich. Palaeontographica 92: 122-183. KaIserR, P. 1890. Die fossilen Laubhélzer. Wiss. Beilage. Jahresber. Realpro- gymn. Schonebeck a.E KANEHIRA R. 1921a. Identification of the important Japanese woods by ana- tomical characters. Taihoku, Formosa. 1921b. Anatomical characters and identification of Formosan woods with critical remarks from the climatic point of view. Taihoku, Formosa. 1924. oe of Philippine woods by anatomical characters. Taihoku, Form 1940. ie ee Formosan woods. Bull. For. Exp. Sta. Taiwan 2 Kraus, G. 1882. Beitrige zur Kenntnis der fossilen Holzer. 1. Holzer aus den sea gone Siciliens. Abh. Nat. Ges. Halle 16: 77-91. Kriss. D. A. 1959. Commercial foreign woods on the American market. Uni- versity ae Penn. Littte. E. L. 1953. Check list of native and naturalized trees of the United States (including Alaska). U.S. Dep. Agr. Handb. 41. MaApet, Ertka. 1959. Ein fossiles Nyssa-Holz aus Japan, Nyssoxylon japont- cum n. g., n. sp. Senckenbergiana lethaea 49(3/4): 211-222. Metcatre, C. R., and L. CHaLk. 1950. Anatomy of the dicotyledons. 2 vols. Oxford. Miu ier-STott, W. R. 1938. Die jiingsttertiare Flora os Eisensteins von Dernbach im Westerwald. Beih. Bot. Centralbl. 58B: 376. Munns. E. 1938. The distribution of important forest i. of the United States. U.S. Dep. Agr. Misc. Publ. 287. NAGEL, K. 1915. Peart Foss. Cat. Pl. 6: 1-87 Pearson. R. S.. and H. P. Brown. 1932. Commercial timbers of India. 2 vols. Calcutta. PRAKASH. U.. and E. S. BarcHoorn. 1961. Miocene fossil a ae ee the Columbia basalts of central Washington. Jour. Arnold Ar b. 42: —203. Reyes, L. J. 1938. Philippine woods. Philip. Dep. Agr. Tech. Bull. 27-449. SCHONFELD, E. 1930. Sichsische Braunkohlenholzer. Untersucht bie bestimmt von G. Schinfeld. Sitzber. Nat. Ges. Isis. Dresden 1929: 63— STARK, E. W. 1953. Wood anatomy of ay indigenous to ae United States. Agr. Exp. Sta. Purdue Univ. Bull. 595: 4-42 Sropes. M. C., and K. Fuyu. 1910. Studies on ihe structure and affinities of Cretaceous plants. Phil. Trans. Roy. Soc. London 201: 1-90. SupwortH, G. B., and C. D. MELL. 1911. eco. i important Ameri- can oak woods. U. S. Forest Serv. Bull. Uncer, F. 1845. Synopsis plantarum fossilium. iene Wararr, S. 1952. Dicotyledonous woods from the ne along the Noes Sea — of Honsyu [Honshu] Jour. Fac. Sci. Univ. Tokyo Bot. 6( 1-3) Q7- WOLFE, : \ and E. S. BARGHOORN. 1960. Generic change in Tertiary floras in relation to age. a Jour. Sci. Bradley Vol. 258-A: 388-399. 358 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII EXPLANATION OF PLATES PLATE I Wood of Nyssa eydei: 1, transverse section, X 37 —note vessel Fics. 1-5. and parenchyma distribution; 2, transverse section magnified to show the cellu- lar details, & 90 — note parenchyma distribution; 3, tangential section showing 5, scalariform perforation plate, ‘s 110. Fics. 6, 7. Wood of Ulmus baileyana: 6, intervascular pitting x 120; transverse section, * 20—vnote vessel dis- tribution in late wood forming cas bands. Pate II Fics. 8-11. Wood of Ulmus set dans 8, sina oe slightly a to show early wood, 30; 9, transverse section magnified to show fiber s ture, X 90; 10, tangential section in low power suaee shape, size, and distribu. tion of xylem ra rays, X 3 angential section magnified to show ray structure, x 80. Fics. 12- vi Wood of pete fryxellii: 12, safeiouan Alas pitting, X 160; 13, transverse section magnified to show structural det ails, X 60; 14, transverse section in low power, X 27 — note parenchyma distribution. PLaTeE III Fic. 15. Wood of Juglans fryxellii: 15, SS section in low power, X 60 hape, size, and nature Of vem ys. Fics. 16-19. _— of zeke a ection note large vessels in early wood and smaller vessels in late wood panied is in flame-shaped tracts. Fics. 20-22. Wood of Aesculus hankinsii: 20, transverse section highly magnified to show the fiber cells, X 90; 21, interva scular pitting, 40; 22, transverse section in low power to show vessel distribution, 35. PLATE IV Fic. 23. Wood of Aesculus hankinsit in pic ciniets section showing uniseriate xylem rays, X 90. Fics. 24-27. Legumin siete occidentale: 24, transverse sec- tion magnified, X 60—note parenchym rl cells; 25, intervascular pitting, X 160; 26, tangential section, x 27, transverse section in low power, X 26—note vessel and parenchy ma distributi ion > PLATE I Jour. ARNOLD Ars. VoL. XLII = ee ee ia Sed aod a jeoaueres cevun re ‘$ > » ee = ‘éé +a ef Rt ree reeve aon a. x dhe arres Zit gases aoe Saas ane et gee re, he Rs SERRE SIO a -° in eee We. <7 eestor niet ‘naneden Prd rd ae ee ee eae ean LET SRAEaRNRS be Bb gi " 8 a ae x fin avgag see a2 2! teak rid i 8 de we FG eeY Bs eH > & % = iain ot +0 gene z 4 , ‘ 5 oe Se indiana ea pest Ny ners ¥ 9 vee 5 ar od “ ¥ hati 60} dll sed be. si see 2a A Ge ; OF un. * = Fa Spee cates ro ww alan i * r* = ses ee “ie -¢* "F @e Fs é :. wo ge et ag a ¢ a ‘ * , waccs® * 1 Oe Oe ea ae ae: orseterer : E nce oi oat * : ee te one gn tes ae ee eee ad ; oe * bd a ire é "4 « LILO, HT pore | Te sapniagil Si saceenietee See OE EM We ® ewe te PRT a ‘ @ ',%,” TEXANS ges eNOS OE CREA Pog ccewny. ceamet or Tie © eee ods. &@* SO? a. 2- & Sere €. oo ae a a ee nb Nia d Phe . é reed SRA 5 eee Mais A Sl re cose ieee , MiocENE Woops, II PRAKASH & BARGHOORN I] PLATE XLIT ARNOLD ARB. VOL. UR. ya - os ae ee is if id Sone. MI0cENE Woops, II BARGHOORN, RAKASH & P PLaTE III VoL. XLII ARNOLD ARB. JOUR. oops, II W MIOCENE ? N ASH & BARGHOOR PRAK Nee eft en. ae 4% oh PLATE IV 5 ae ‘“ # DET gee 2 wy Se i, we toe > XLII ARNOLD ARB. VOL. Jour. S RAKASH & BARGHOORN, MIOCENE Woops, II P Ge 1961] BRIZICKY, SYNOPSIS OF COLUMELLIA 36 A SYNOPSIS OF THE GENUS COLUMELLIA (COLUMELLIACEAE) GEORGE K. BrizICKY THE GENUS COLUMELLIA was founded by Ruiz and Pavon in 1794. Since 1801, when Ruiz referred the genus to Oleaceae (Jasmineae), the relationships of this Andean South American group of three or four woody species have been a perplexing problem for several generations of taxono- mists. With its peculiar combination of opposite, exstipulate leaves; bi- sexual, epigynous flowers; somewhat irregular, sympetalous corollas; two stamens with plicate and contorted anthers resembling those of some Cucur- bitaceae; two-carpellate, imperfectly two-locular ovaries; and imperfectly four-locular capsular fruits, Columellia is indeed a unique genus. It has been considered to be related to or has been placed in a whole series of families, including the Saxifragaceae, Lythraceae, Onagraceae, Ericaceae- Vaccinioideae, Ebenaceae, Styracaceae, Oleaceae, Gentianaceae, Scrophu- lariaceae, Gesneriaceae, Loganiaceae, Rubiaceae, and Caprifoliaceae. Al- though most recent taxonomists (following De Candolle, 1839) treat Columellia as a monogeneric family placed near Gesneriaceae, Wettstein 1935) stressed the uncertainty of its position. On the other hand, Sole- reder (1899), taking into consideration the anatomy of the genus, thought Collumelliaceae to be closely allied to Saxifragaceae, rather than to Ges- neriaceae. It is evident that gross morphology alone is inconclusive in solving this problem and that more attention must be paid to the anatomical, as well as embryological and cytological, characters of the taxa involved. In consideration of this, a comparative anatomical and taxonomic study has been undertaken by Dr. William L. Stern, Division of Woods, U. S. National Museum, Smithsonian Institution, and the author in the hope of shedding more light on the problem of the systematic position of Columelliaceae. The present study is then to be considered as a taxonomic introduction to this work, which is now in progress. Columellia Ruiz et Pavon, Fl. Per. Chil. Prodr. 3. pl. 1. 1794, nom. cons. i Uluxia Jussieu, Dict. Nat. Sci. 10: 103. 1818. Ramose, evergreen shrubs or trees with shreddy bark; branchlets terete or obtusely quandrangular, usually more or less hairy, the internodes short. Leaves opposite, crowded, exstipulate simple, + inequilateral, en- tire or toothed (teeth and leaf apices gland-tipped), pergamentaceous to 1 Columellia Ruiz et Pavon, 1794, should be conserved against an earlier homcnym, Columella Loureiro, Fl. Cochinch. 85. 1760 (nom. rejic. vs. Cayratia Jussieu, 1818, nom. cons.). See A. A. Bullock, Taxon 9: 221. 1960. 364 JOURNAL OF THE ARNOLD ARBORETUM [ VOI. XLII coriaceous, with distinct or indistinct lateral nerves, the petiole usually short, margined (from the decurrent leaf blade), the opposite bases of petioles sheathlike, connate. Inflorescence axillary and/or terminal, basi- cally a solitary dichasial cyme (sometimes reduced to a single central flower) or a small panicle of dichasia. Flowers bisexual, somewhat irregu- lar, epigynous, short pediceled to subsessile, bibracteolate. Calyx tube narrowly to broadly obpyramidal or obconical, adnate to the ovary, the lobes 4—6 (8?), usually 5, subequal, each tipped with a gland, valvate or slightly imbricate in aestivation, accrescent. Corolla yellow, sympetalous, somewhat irregular, subrotate to subcampanulate, the tube short, the lobes 5 or 6 (8?), usually 5, slightly unequal, usually longer than the tube, imbricate in aestivation. Stamens 2, very rarely 3, attached at base of corolla tube, the filaments short and stout, adnate at base or for 1/3—1/2 of their length to the corolla tube, dilated at apex into a broad subcircular- reniform, sometimes inconspicuously 3—5-lobed, connective on which an extrorse, irregularly plicate and contorted anther is borne. Gynoecium syncarpous, 2-carpellate. Stigma broad, depressed-capitate, 2- or ob- scurely 4-lobed, ending a terminal, relatively short, stout style; ovary inferior, imperfectly 2-locular, with numerous anatropous, unitegmic and tenuinucellate ovules multiseriately attached to 2 much intruded parietal placentae. Fruit an imperfectly 4-locular, obpyramidal, obconical to sub- fusiform capsule, dehiscing septicidally into bifid valves; seeds numerous, minute, laterally compressed, + oblong, smooth, the endosperm fleshy, the embryo minute, straight, axile. Lectotype species: C. oblonga Ruiz et Pavon; see Bullock, Kew Bull. 14: 41. 1960. The genus includes three or four species occurring in the Andean region of Ecuador, Peru and western Bolivia at altitudes of from about 1600 to 3600 meters. orphological notes. In addition to the characteristics given in the formal generic description above, a number of other morphological details, some of which have been used in the delimitation of species, should be given. The young branchlets, leaves (especially the lower surface), axes of the inflorescences, hypanthia, calvx lobes (at least the outer surface), and the exterior surface of the corolla tube usually are sericeous, appressed- puberulous, or more or less woolly. There is sometimes a tendency toward glabrescence, and rarely these parts are glabrous or subglabrous (Columel- lia lucida). The leaf apices and teeth, the bracts, and the calyx lobes usually are tipped with a mucro-like gland. The data in the literature on the number of perianth parts show some discrepancies: calyx and corolla 5-merous (Ruiz & Pavon, 1794, 1798: Fritsch, 1894; Schlechter, 1920); or calyx 5-lobed, but corolla 5—8-lobed (De Candolle, 1839; Lindley, 1853; Bentham & Hooker, 1876); or calyx and corolla 5—8-merous (Engler & Diels, 1936). In this study only 5- and occasionally 6-lobed calyces and corollas and a single 4-lobed calyx have — 1961] BRIZICKY, SYNOPSIS OF COLUMELLIA 365 been observed. The calyx lobes themselves are triangular-ovate to oblong, 0.25-0.5 cm. long, and 0.15—0.3 cm. broad, acutish or obtuse. The tube of the corolla varies from 0.15 to 0.55 cm. long, and the lobes are usually longer than the tube, more or less semicircular in outline, and 0.3—0.7 cm. long and 0.4-0.9 cm. broad. The diameter of the expanded corolla limb varies from 0.8 to 2 cm. The stamens are usually two (the only number recorded for the genus), and these are alternate with the lateral and posterior (adaxial) corolla lobes. In a single flower of Columellia oblonga subsp. oblonga and in one of C. obovata, three stamens were present. The position of the extra stamen was different in both cases. In a single flower of C. obovata one subsessile and one sessile anther (anomaly?) were observed. The stamens, with their peculiar plicate and contorted anther, have been regarded by some taxonomists as complex structures arising from the connation of two or three stamens. However, Van Tieghem (1903), on the basis of the anatomy of the stamens, affirmed their simple character and assumed a basic number of five stamens and the abortion of three. The pollen, as described by Erdtman (1952), is “3-colporate, (oblate- ) suboblate. Sexine thicker than nexine, rather finely reticulate (muri simplibaculate). Ora + circular.’ The size of pollen grains varies from 25 & 34 p in Columellia oblonga (Peru, Weberbauer 7791), to 24 & 31 pin C. obovata (Peru, Ruiz & Pavoén, 1788), to 22.5 * 29 » in C. sericea (Peru, Ruiz & Pavén), according to Erdtman (1952). Although some authors have mentioned the presence of a nectariferous disc (Lindley, 1853; Rusby, 1896), the author could not detect this in the dried herbarium material. The presence of a nectary on the top of the ovary is probable, however. As mentioned in the generic description, the shape of the capsules in Columellia varies from obpyramidal or obconical to subfusiform. The subfusiform shape is due to the growth of the top of the ovary after anthesis. In Columellia oblonga subsp. sericea, the length of the inferior portion of the fusiform capsule (that united with the calyx tube) varies from 0.3—0.6 cm., and the length of the superior free part of the fruit varies from 0.15—0.7 cm. Since fruits, in general, are very scantily repre- sented in the herbarium material, possible Sena differences in the shape and size of fruits and seeds could not be studiec The general anatomy of Columellia, which seems to be fairly uniform within the genus, was studied by Solereder (1899) and Van Tieghem (1903). The wood anatomy of C. oblonga subsp. sericea was reported on by Record and Hess (1943). Uses. The very bitter leaves of Columellia oblonga subsp. oblonga anc d C. obovata are used by natives in Peru as an infusion or decoction, some- times as a powder, against tertian fever, and also serve as a stomachic (Ruiz & Pavén, 1798). The very hard wood of C. oblonga is (or was) utilized for manufacturing handles and various utensils in Peru (Ruiz & Pavon, loc. cit.) but only for fuel in Ecuador (Record & Hess, 1943). 366 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XLII KEY TO THE SPECIES A. Leaves pergamentaceous to subcoriaceous, with 5-10 pairs of distinct, as- cending lateral nerves, mostly 3-6 cm. long, conspicuously inequilateral, sericeous or appressed-puberulous on the lower surface; flowers usually in small panicles; corolla subrotate to subcampanulate; shrubs or trees. . 1. C. oblonga. A. Leaves coriaceous, with indistinct lateral nerves, mostly a 3 cm. long, in- conspicuously inequilateral, glabrous or + woolly pubescent on the lower surface; flowers in 1—3-flowered cymes; corolla subrotate; shrubs, sometimes —_ > olay od WwW. . Leaves glabrous on the lower surface, oblanceolate and/or ov oe cuneate, entire or rarely 1- or 2-dentate at apex, 1-3 cm. long. ..... . C. lucida. Leaves pubescent on the lower surface, elliptic- oblong to we obovate. C. Leaves all entire, pena yi on the lower surface, elliptic- caries to elliptic-obovate, 0.9-1.3 cm. long; flowers ca. 0.8 cm. in diam low shrubs. .....000..0.00000000......02....., a OF porseoe ri . Some or most leaves remotely dentate, densely or sparsely sublanugi- nous or subtomentulose on the lower surface, oval, elliptic- to broad- obovate, 1-2.7 cm. long; flowers 0.9-2 cm. in me Lak baa es -Q? Columellia oblonga Ruiz et Pavén, Fl. Per. Chil. 1: 28. pl. 8, figs. a & 1-8. 1798 C. arborescens Pers. Syn. Fl. 1: 13. 1805. DIsTRIBUTION: Peru, Ecuador, and western Bolivia. KEY TO THE SUBSPECIES . Leaves glandular-serrulate to -serrate, at least toward apex, appressed- dg nsig on the lower surface; corolla subrotate, the tube (0.15)0.2-0.3 m. long, as long as or oe than calyx lobes; staminal filaments adnate eee tube only at oo ise) Leaves glandular- dats toward the apex (rarely entire), oblong, ellip- tic or oblanceolate-oblong, 3—4.5 cm. long, obtuse, rarely acute; corolla tube appressed-puberulous externally. .............. la. subsp. oblonga. . Leaves jp emgh aden from near middle to apex, elliptic to elliptic- oblong, 4-6 cm. long, acute or obtuse; corolla tube nearly glabrous ex- TOPOS ooo e882 4, Lacie 2 x9 Ganev Red oR Oh Re ES 1b. subsp. serrata. . Leaves entire, rarely with a few teeth near apex, usually densely sericeous on the lower surface (very rarely appressed- s aetilarvion oblanceolate to elliptic- obovate, (1.6)3-5.2 cm. long and 0.5-2.3 . broad, usually acute; corolla usually subcampanulate, the tube (0304-055 cm. long, appressed-pilose externally, mostly longer than the calyx lobes; staminal filaments adnate to corolla tube for 1/3-1/2 of their length. .............. lc. subsp. sericea. la. Columellia oblonga subsp. oblonga (Typr: Pillao, Ruiz & Pavén). Chachapoyas, Mathews 1319 (Ny). HvuAnuco: Pillao, Peru. AMAZONAS: Ruiz & Pavoén (F¥, isotype) Pillao on the side of river Chaglia-monte, alt. 2700 1961] BRIZICKY, SYNOPSIS OF COLUMELLIA 367 m., Woytkowski 34067 (r, Mo). AyacucHo: La-Mar, between Tambo and Apurimac, way of Osno, alt. 2900-3000 m., Weberbauer 5584 (¥, GH, . Cuzco: Urubamba, Machu Picchu, alt. ca. 2100 m., Cook & Gilbert 850 (us), alt. 2000 m., id See 7408 (us), Machu Picchu y Huayana Picchu, 2400 m., Scolnik 830 (NY) pete eas Marcapata, Weberbauer 7791 (A, F, MO, US); San Miguel, Urubambs Valley, alt. ca. 1800 m., Cook & Gilbert 1749 (us). Ecuador. Azuay: Eastern Cordillera, 1-8 km. north of the village of Sevilla de Oro, alt. 2400-2700 m., Camp E-4461 (GH, NY, Y). This subspecies has usually been described as a shrub or small tree up to 6 m. high. Camp, however, collected his herbarium specimen from a tree 20 m. tall. In a field note with this specimen he says, ‘““This species flowers when quite young, specimens seen in the region in full flower and only 1 meter high; it is fairly abundant on cut-over and burned areas which are regenerating (probably from root-sprouts) where it might be taken for a shrub; also seen in forested areas, the tree easily detected by the irregular contour of the trunk and shreddy bark, one patriarch seen along this same quebrada at about 8000 ft. elev. with a trunk diam. of nearly 2 meters (the top broken out of this tree and so no estimate can be given of the maximum height).’” Camp’s collection, the first of subsp. oblonga from Ecuador, is also interesting as it represents a form of the subspecies with predominantly entire leaves intermixed with two- to-few-dentate leaves. In this character the form approaches subsp. sericea. There are no grounds, however, for a supposition of hybridization (at least recently), for no specimens of the latter subspecies have been seen or re- corded from Prov. Azuay. Furthermore, our knowledge of variations of subsp. oblonga (even on individual trees) is very limited. Camp’s field note suggests a further source of variation, regeneration from the root sprouts, which could result (at least on younger plants) in larger, perhaps more dentate and more densely pubescent leaves and larger flowers than in plants grown from seeds or in adult specimens. In addition, Mathews 1319 seems to be intermediate between this subspecies and subsp. sericea. In Ruiz’s Travels (1940) this subspecies is mentioned as Columellia corymbosa. 1b. Columellia oblonga subsp. serrata (Rusby) Brizicky, stat. nov. C. serrata Rusby, Mem. Torrey Bot. Club 6: 93. 1896. (Type: Miguel Bang 1172). Bolivia. CocHABAMBA: Espirito Santo, Miguel Bang 1172 (ny-holotype; A, F, MO, US-isotypes). The type locality, Espirito Santo, is in the ‘“yungas” floristic region at an altitude of ca. 1600 m. (Herzog, 1923). Unfortunately, the specimen Herzog 2240, (‘‘Baumstrauch im Gebiisch tiber Incacorral, ca. 2700-2800 m.”) reported by Herzog (1915) as Columellia serrata has not been seen. Schlechter (1920) says regarding the specific differences between C. serrata and C. oblonga, “Bei naherem Vergleich der Art mit C. oblonga Ruiz & Pav. zeigt sich, dass beide recht nahe verwandt sind. Ich halte sie hier 368 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XLII getrennt, weil bei C. serrata Rusby die Blatter meist deutlich grésser und die Bliten fast kahl sind, wahrend die letzteren bei C. oblonga Ruiz & Pav. aussen nach dem Grunde zu ziemlich dicht, wenn auch sehr kurz behaart sind. Ausserdem scheinen bei C. serrata Rusby die Friichte mehr denen der C. sericea H. B. & Kth. zu ahneln.” These differences by them- selves can hardly be considered of specific significance. Furthermore, the density of pubescence of the corolla tube in C. oblonga subsp. oblonga varies to a considerable degree; for instance, the corolla tube in Weber- bauer 5584 is almost glabrous outside, as is that of C. serrata. The latter is therefore reduced to subspecific rank under C. oblonga. Ic. Columellia oblonga subsp. sericea (HBK.) Brizicky, stat. nov. C. sericea HBK. Nov. Gen. Sp. 2: 388. 1818 (Type: Humboldt & Bonpland 3204). C. Mathewsii Briquet, Ann. Conserv. Jard. Bot. Genéve 20: 367. 1919 (Type: Chachapoyas, Mathews). VERNACULAR NAMES: quinoa blanca (Rimbach), pantza (Acosta Solis). Ecuador. ImBABURA: Lake Cuicocha, alt. 3200 m., Penland & Summers 738 (coco, F), Lake Cuicocha, Islote Chica, alt. 3150 m., Aspland 7152 (us); small wooded quebrada 5 km. north of Hacienda Pinnon, Wiggins 10351 (ps). PicHincua: M. Pichincha, Ed. André K-588 (¥); Malchingui to Pomasqui, alt. 3000-3600 m., A. S. Vache 20846 (GH, NY, US); Valle Seco del Pedregal, Cord. Oriental, Acosta Solis 8433 (r); Pifo, Mille 409 (us); Pululahua, Ed. André 3798 (¥, GH); Calacali, Western Andes of Quito, 3000-3300 m., Lehmann 4685 (F, US); Bajade de Lache a Condor Machay, Cord. Occ., alt. 2700-3100 m., Acosta Solis 7004 | ; San Juan, Cord. Occidental, Camino a Saloya, alt. 3200 Acosta Solis one (Ff). TUNGURAGUA: Tunguragua, Humboldt & Bonpland 3204 (F-photograph of type; F, US-isotypes); vicinity of Patate, hacienda Leito, alt. ca. 2900 m., Asplund 8003 (us). BoLivar: Pajonal de Gualicon Loma, Cord. Occ. alt. 3240 m., Acosta Solis 6264 (¥). CHImMBorAzo: Ad pagum Penive, in reg. arborea superiori, alt. 2500-3200 m., Mille 44 (a, GH, Ny, US). GUAYAS: Guayaquil, in 1804, “Ruiz & Pavén” (F*). Loco speciali haud indicato: Andes of S. America, U. S. Dep. Agr. (us); in Andibus Ecuadorensibus, Spruce 5468 (GH); Jameson (US 534966, US 534793, US 534794); inner slope of Eastern Cordillera, alt. 3400 m., in cae Rimbach 30 (F, GH, Y), 3200 m., Rimbach 81° (A), Rimbach 122 (A, F, Y), Rimbach 227° (us). Peru. Amazonas: Chacha- poyas, Mathews s.n. (G- pens of C. Mathewsit). Subspecies sericea seems to be very distinctive, not only in its usually entire, very rarely toothed leaves, which are sericeous on the lower surface * The label on this specimen bears the heading: “Ex antiquo elaine pgeoles Her- barium Horti Botanici Matritensis Plantae a ‘Ruiz et Pavon’ in vice-r » Peruviano et Chilensi lectae. (1778-1788). On the label is typed: “Guay sank te in 1804.” This specimen was apparently collected by Juan Tafalla, a disciple of Ruiz and Pavon, who at about that time collected cae for Ruiz in the mountains of Guayaquil, Loja, and Quito, Ecuador. (See Ruiz, 1801.) *Most probably Rimbach 81 and ae 227, cited here, are misnumbered and are, in fact, Rimbach 122, since in Rimbach’s ena list of field numbers in the Record Memorial Collection, Yale University, Rimbach 81 is Nicotiana Pavonii Dunal, and Rimbach 227 represents Siparuna echinata (HBK.) A. DC 1961] BRIZICKY, SYNOPSIS OF COLUMELLIA 369 (e.g., Hitchcock 20846, and specimens cultivated in the 1870’s at Kew, according to Schlechter, 1920), but also in its usually subcampanulate corollas, the tube of which usually is longer than the calyx lobes Two exceptions, however, are Mathews s.n. (type of C. Mathewsii) and Asplund 5003, which have subrotate corollas with tubes shorter than the calyx lobes. (The former specimen also has leaves appressed short-pilose on the lower surface.) The diameter of the expanded corolla limb varies from one to two centimeters. The occurrence of such individuals transitional in these respects between Columellia oblonga and C. sericea is ample reason to regard both as components of a single species. 2. Columellia lucida Danguy et Chermezon, Bull. Mus. Hist. Nat. Paris. 28: 436. 1922 (Typr: Rivet 740). C. Andrei Macbride, Publ. Field Mus. Bot. 4: 92. 1925 (Type: Ed. André K-1 Ecuador. Loya: Chuquiribamba, André K-1444 (r-holotype of C. Andrei; GH-isotype). Cerro de St. Barbara,’ André 4500 (¥, GH, Ny-paratypes of C. Andrei). The type of Columellia lucida was collected in Cerro Narihuifia (‘‘Nairi- huina’’), alt. 3900 m., in the Cordillera Chanchan, Prov. Azuay, Ecuador. Although the type has not been examined, the author has no doubts of the conspecificity of C. lucida and C. Andrei. The descriptions of the two are almost identical. The only differences noted are ovate-cuneate leaves, 1—1.5 cm. long, sometimes one- or two-dentate at the apex, in C. lucida, as against oblanceolate leaves, 1.5-3 cm. long and completely entire, in C. Andrei, If one takes into consideration that many smaller (younger) leaves in C. Andrei may be described as ovate-cuneate and that the pres- ence of one or two teeth can hardly be an essential character, especially in this genus, the differences between these taxa appear to be insufficient for the recognition of two separate species. Their occurrence in adjacent prov- inces of Ecuador should also be noted. Leaves of this species (as repre- sented by C. Andrei) strikingly resemble those of some specimens of C. oblonga subsp. sericea (e.g., Acosta Solis 8433, Asplund 7152, Penland & Summers 738), but are smaller than in the latter (as are the flowers) and are glabrous on the lower surface. When more material is available C. lucida may Aas prove to be treated best as a variety of C. oblonga subsp. serice 3. Columellia subsessilis agi Notizbl. Bot. Gart. Berlin 7: 356. 1920 (Type: A. Stubel 4 Bolivia. Yungas, mas ariba de Taca, 3300 m., Dec. 10, 1876, A. Stubel 45a (r-photograph). “The exact geographic position of this locality has not been determined. It i possible, however, that Cerro St. Barbara is in Prov. Loja, since André 4446 (Passi. flora sanguinolenta Mast.) was collected near Chuquiribamba. 370 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII This species is known only from the type collection. According to Schlechter (1920), Columellia subsessilis differs from C. obovata in the habit, being a low dense shrub; in the relatively narrow, apparently always entire leaves subsericeous on the lower surface; in the subsessile inflores- cences and flowers; in the smaller corollas; and in the remarkably short filaments of the stamens. These differences do not seem to be sharp ones and are hardly of specific significance, with the possible exception of the small flowers and very short filaments of the stamens. However, the con- stancy of the two latter characters as well as that of the entire leaves, in C. subsessilis has to be proved on more abundant herbarium material of this species. The available flowers of the taxa concerned are very few, but it may be noted that one subsessile anther (with a very short filament) and one completely sessile anther were observed in a flower of C. obovata, suggesting that this may not be a constant feature. At present, Columellia subsessilis appears to be a variant of C. obovata, but for lack of conclusive data it is here maintained at specific rank. 4. Columellia obovata Ruiz et Pavon, Fl. Per. Chil. 1: 28. pl. 12, figs. b & 1-7. 1798 (Type: Tarma, Ruiz & Pavoén). C. frutescens Pers. Syn. Pl. C. Weberbaueri ina pears "Bot. Gart. Mus. Berlin 7: 358. 1920 (Type: Weberbauer 3241). VERNACULAR NAMES: ulux, ullus, usluss (Columellia obovata); zoce- tullu, tsoquetullu (C. Weberbaueri). Peru. JUNIN: Tarma, Ruiz & Pavoén (F-photograph of type; isotype). AN- CASH: Prov. Huaylas, encima de Caraz, Cordillera Blanca, alt. 3200-3400 m., Weberbauer 3241 (¥-photograph and small fragment of type of C. Weber- baueri; G-isotype; UsM-small fragment of isotype); Prov. Huaras, circa 4 La Portada (Pacharuri), Cordillera Blanca alt. 3500 m., Velardo Nunez 3309 (us). AyaAcucHOo: Prov. Huamanga, between Totorabamba and Ayacucho, alt. 3300- 3400 m., Weberbauer 5482 (¥F, GH, US). Cuzco: Valle del Urubamba, Ollantai- tambo, alt. 2800 m., Herrera 3451 (¥). URuUBAMBA: Pumahuanca, alt. 3100— 3500 m., Vargas 7693 (mo, Us). Without indication of locality, but probably in the vicinity of Tarma: Dombey (F, us). According to Schlechter (1920), Columellia Weberbaueri differs from C. obovata in the more vigorous growth; larger, dull leaves; more villous pubescence; larger flowers, ca. 2 cm. in diameter; and narrower calyx tube. None of these characters seems to be of specific significance. The dullness of leaves in C. Weberbaueri, as a characteristic of the living plant (rather than of a dried herbarium specimen) should be questioned. Weber- bauer 3241, identified by Schlechter as C. obovata, has mostly dull (but some lustrous) leaves, but in all other respects it looks very much like a “small-leaved” C. Weberbaueri. The difference between “sublanuginous” pubescence in C. Weberbaueri and “‘subtomentulose” in C. obovata is not a sharp one, and both kinds are certainly only small variations of the same 1961] BRIZICKY, SYNOPSIS OF COLUMELLIA 371 basic type. Although the type specimen of C. Weberbaueri possesses some- what larger flowers than C. obovata, another specimen from the same geographic and floristic province (Cordillera Blanca) which agrees with C. Weberbaueri in its vegetative organs, has flowers about 1 cm. in diam- eter. Thus, with these indications of intergradation, C. Weberbaueri is re- duced here to the synonymy of C. obovata. . The presence of at least a few dentate leaves has been given as a diag- nostic character of Columellia obovata, but the species certainly is very variable, especially in dentation and pubescence of leaves. Although the type specimen of C. obovata shows a few leaves with dentate margins, there are other herbarium specimens in which it is hard to find a single leaf with even one or two teeth. On the other hand, specimens in which most of the leaves are markedly toothed have been seen. Thus the species either includes populations with entire, intermediate, and toothed leaves, or the dentation of leaves is an inconstant character reflecting the local ecological conditions. It seems likely that variation in size of the flowers is connected, not only with local ecological conditions, but also with the season. Perhaps in the future the necessity of subdividing C. obovata into two or more morphological-ecological types will be clear, but on the basis of the present data the recognition of any intraspecific categories within this species would be groundless. This species is referred to as Columellia ovalis in the translation of Ruiz’s Travels (1940). CONCLUSIONS The genus Columellia does not appear to be polytypic. Variations, although not uncommon, are usually limited to the vegetative organs, especially the leaves (size, degree of dentation, and pubescence). When variations involve the flower, as in C. oblonga subsp. sericea (subcam- panulate corollas, stamens adnate to the corolla tube for 1/3-1/2 their length), they do not seem to be entirely fixed. Only four of the nine species which have been described are recognized in the present paper, and it is likely that still further reductions will be necessary as more adequate material is obtained. Field studies accompanied by extensive collecting, experimental cultures, and cytogenetic investigations are desirable to check the outlines of the taxonomy of the genus as proposed in this study, which has, of necessity, been based entirely upon scanty herbarium ma- terial. The author is grateful to the curators of the herbaria cited in this paper (with the abbreviations of Lanjouw and Stafleu, Jndex Herbariorum, ed. 3) for their kindness in lending specimens which have been essential to this study. He is indebted to Dr. W. L. Stern, curator of the Division of Woods of the Smithsonian Institution, Washington, D. C., formerly curator of the S. J. Record Memorial Collection, Yale University, New Haven, 372 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Connecticut; to Dr. R. A. Howard, director of the Arnold Arboretum; and to Dr. C. E. Kobuski, curator of the Arnold Arboretum and Gray herbaria, for arranging loans of specimens from herbaria in this country and abroad. ARNOLD ARBORETUM GRAY HERBARIUM ARVARD UNIVERSITY LITERATURE CITED BENTHAM, G., and J. D. Hooker. 1876. Genera plantarum 1. Londor Briguet, J. 1919. Decades plantarum novarum vel minus Benen Ann. Conserv. Jard. Bot. Genéve 20: 342-42 CANDOLLE, A. P. DE. 1839. Pointenea. Prodr. Syst. Nat. Reg. Veg. 7: 549. Danouy, P., and H. CHERMEZzON. 1922. Sur quelques espéces et variétés nou- velles a la République de l|’Equateur. Bull. Mus. Hist. Nat. Paris 28: 432-439 ENGLER, A., and L. Diets. 1936. Syllabus der Pflanzenfamilien. ed. 11. Berlin. ErpTMaANn, G. 1952. Pollen morphology and plant taxonomy. Angiosperms. Stockholm. Fritscu, K. 1894. Columelliaceae. Nat. Pflanzenfam. IV. 3b: 186-188. Herzoc, T. 1915. Die von Th. Herzog auf seiner zweiten Reise durch Bolivien gesammelten Pflanzen. II. Meded. Rijks Herb. Leiden 27: 1-90. 1 fl. . 1923. Die Sai der bolivischen Anden. /n: Engler & Drude, Veg. der Erde XV. Leip Humpo tpt, A., A. Borla, andl C. S. KuntH. 1817-1818. Nova genera et species plantarum 2. Jussieu, A. 1818. Columelle. In: Dictionnaire des Sciences Naturelles (Lev- rault, ed. ) i 103. Paris LINDLEY, J. 1853. The vegetable kingdom. ed. 3. Lon MacprIibE, J. : 1925. South American plants. Publ. Field Mus. Bot. 4: 79-93. Recorp, S$. J., and R. W. Hess. 1943. Timbers of the New World. Yale Uni- versity Press. New Haven. Ruiz, H. 1801. Supplemento a la Quinologia. 154 pp. Madrid. ———. 1940. Travels of Ruiz, Pavon, and Dombey in Peru and Chile (1777- 1788). Transl. by B. E. DAHLGREN from the Spanish edition of 1931 ‘“Relacion del viaje etc.” Publ. Field Mus. Bot. 21. ——— and J. Pavon. 1794. Florae Peruvianae, et Chilensis prodromus. Madrid. . 1798. Flora Peruviana, et Chilensis 1. Madrid. Russy, H. H. 1896. On the collections of Mr. Miguel Bang in Bolivia. Part III. Mem. Torrey Bot. Club 6: 1-130. SCHLECHTER, R. 1920. Die Columelliaceae. Notizbl. Bot. Gart. Berlin 7: 352- 358. SOLEREDER, H. 1899. Systematische Anatomie der Dicotyledonen. Stuttgart. IEGHEM, P. vAn. 1903. Sur les Columelliacées. Bull. Mus. Hist. Nat. Paris 9: 233-239. WetTsTEIN, R. 1935. Handbuch der systematischen Botanik. ed. 4. Leipzig & Wien JOURNAL OF THE ARNOLD ARBORETUM VoL. XLII OcToBER 1961 NUMBER 4 A REVISION OF THE ASIATIC GENUS ENKLEIA (THYMELAEACEAE) LorIN I. NEVLING, JR. TuHIs stupy OF ENKLEIA is the second of three revisions treating the three genera which comprise the subtribe Linostomatinae of the Thy- melaeaceae. It completes the revision of the two Asiatic members, Linostoma Wall. ex Endl. and Enkleia Griff., and prepares the way for a more enlightened appraisal of the Amazonian genus Lophostoma Meissn. The members of the Linostomatinae have been combined taxonomically in various ways in the past so that a taxonomic re-evaluation is of primary importance. The revision of Linostoma and Enkleia has resulted in the reduction of the number of species and the alignment of several taxa with Enkleia, rather than with Linostoma. There is little doubt that Lophostoma is related to both Linostoma and Enklieia, as is evidenced by the common possession of a number of diagnostically important morphological and anatomical characteristics. An effort has been made to include, as a sec- ondary study, certain anatomical information which may prove to be taxonomically significant. The large geographic disjunction that now exists between the South American and the Asian genera adds to the fascination of the interrelationships to be found among members of the subtribe. SPECIAL MORPHOLOGY Some of the more interesting morphological features found within the genus Enkieia are presented in this section. In addition, a deliberate and conscientious attempt has been made to parallel closely the discussion of morphological characteristics of Linostoma (Nevling, 1961). This has been done to simplify comparisons between the two genera. Additional morphological information can be found in the description and in the dis- cussion following each species. Gross and microscopic structure of the leaf was studied by clearing and staining with safranin or with ferric chloride and tannic acid. In addition, it was studied from safranin-fast green-stained cross and longitudinal thin sections, as was the petiole. This material was tested with phloroglucin and 374 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII hydrochloric acid, as well as with potassium iodide and aig acid, to estimate chemically the degree of lignification of certain tissue The stem and inflorescence axes were studied by means of free-hand sections which were stained with either safranin or phloroglucin and hydro- chloric acid. This technique was also employed with inflorescences and young shoots. Flower structure was studied from material that was cleared and stained with safranin. In this instance, the study was limited to whole mounts and did not include sectioned specimens. The source material was from herbarium specimens in all instances, and the details of the techniques outlined above can be found in a previous paper (Nevling, 1961). Vegetative Morphology. The vegetative axis is monopodial, Leaf position is opposite or subopposite except in extremely vigorous young shoots. In slow-growing shoots a cycle composed of an elongate internode and an extremely shortened internode accounts for the opposite to sub- opposite position, the trace for one leaf departing from the stele prior to the initiation of the second trace. In vigorous shoots, all internodes undergo similar elongation, thus accounting for the alternate leaf position. A single trace per node is formed in all cases, and the resultant gap is unilacunar. Unequal bifurcations and trifurcations of the shoot are not unusual in Enkleia but do not occur with the frequency found in Linostoma. They are induced more generally by damage to the terminal bud which allows the development of subtending axillary buds in a haphazard manner, but, in some cases, the trichotomies occur without apical damage (see Put 2305 and the flowering sheet of Van Royen 3592). In the Van Royen collection the opposite axillary branches are gracefully curved (not recurved) and the subtending leaves are borne a considerable distance above the point of external branching. The lowest leaf on these branches is unpaired and is the displaced leaf of the subtending node; the trace for this leaf has its origin below the point of external branching. It seems that the leaf is borne upon the axillary branch which it subtends. In reference to the “misplaced” leaf one should not expect to find a bud in its axil, but an axillary bud is almost invariably present. It is impossible to account for this unless it is a supernumerary bud. In several specimens of Enkleia malaccensis (Toroes 4099 and Ridley 6427) development of multiple axillary buds has occurred. In the Toroes specimen two axillary branches develop from the axil of a single leaf and are positioned one above the other. A specimen of EF. siamensis ssp. sia- mensis (Kerr 16136) has the terminal bud injured, and supernumerary axillary buds have developed (three from one leaf axil, two from another). In the subgenus Nervira of the genus Daphnopsis supernumerary axillary buds are the rule, rather than the exception, and allow the plants to be- come cauliflorous. After examining the beautiful photograph of a Phaleria in Hou’s treatment of the Thymelaeaceae (1960), I suspect that the same phenomenon is acting in that genus. 1961 | NEVLING, REVISION OF ENKLEIA 375 Axillary branches sometimes become gently to strongly uncinate, fa- cilitating climbing. Such branches are quite thickened towards the base and bear very reduced leaves. The frequency of these modified branches varies rather widely, but some generalizations about their occurrence among the species may be made. Enkleia thorelii lacks them altogether; in EF. paniculata and E. siamensis they are found infrequently but may be well developed; and in E. malaccensis they are found generally, sometimes very well developed (Anta 1713, for example). The extraxylary fiber sheath surrounding the stele, by which the family is often identified, is well developed in this genus. It becomes well ligni- fied after a number of years. In addition, in the pith are found scattered fibers which also appear to be well lignified. In general, the fibers and xylem of the Enkleia species are lignified to a greater extent than the cor- responding structures in Linostoma. The leaves are dorsiventral. The upper epidermis is composed of tab- ular cells, is unbroken by stomata, and is very poorly cutinized. In Enkleia siamensis and E. paniculata the upper epidermis is a single cell thick, with the periclinal walls much thicker than the anticlinal walls. There is a possibility that these differentially thickened cells may be slime cells. A single epidermal layer is found also in E. malaccensis, but the cell walls are not differentially thickened. Enkleia thorelii is unusual because the epidermis is one or two cells thick, varying irregularly with neither number predominant. When double, the two cells occupy the same volume as that occupied by the single cell components. The palisade tissue is com- posed of compact columnar cells and is two or three cells thick, except in E. paniculata in which it is only one or two cells thick. In all species, the palisade cells are smaller in diameter than the cells of the upper epidermis. A spongy parenchyma layer underlies the palisade and is composed of loosely knit branched cells interspersed with many large air spaces. This layer is developed to the fullest in F. paniculata in which it occupies fully three-quarters of the entire depth of the leaf. The lower epidermis is com- posed of cells smaller than those of the upper epidermis, and, in addition, it is interrupted frequently by stomata. In E. siamensis, E. paniculata and E. malaccensis it is uniseriate, the individual cells with the outer periclinal wall slightly bullate. The periclinal cell walls are differentially thickened in FE. siamensis and E. paniculata but not in E. malaccensis. As might be expected, the lower epidermis in £. thorelii occasionally is double, and the individual cells are short-columnar with conspicuously bullate outer peri- clinal walls. The stomata of all species of this genus are of the pit-type previously described for Linostoma decandrum (Roxb.) Wall. ex Endl. These stomata, when examined microscopically in cleared material stained with ferric chloride, resemble in face view many small doughnuts scattered on the lower leaf surface. A slight shift in focus is necessary to observe the guard cells. As in L. decandrum, a rosette of accessory cells surrounds each stoma, the number being rather variable (generally eight, nine, or ten, but as few as six and as many as thirteen). Each accessory cell is 376 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII somewhat awl-shaped and extends beyond and slightly over the guard cells, forming an urceolate or campanulate structure. Occasionally an ac- cessory cell can be found being shared by two adjoining stomata. The guard cells are borne at the base of the accessory cells and are situated at most only slightly above the lower epidermis. Differences in the structure of the stomata can be found among the species (Fics. 1—5), but the amount of variation has not been determined, so their taxonomic worth cannot be evaluated. 5 —5. Cross sections of lower leaf epidermis of species of Enkleia in- cluding the stomatal apparatus. Guard cells (borne above the lower epidermis) are shown in cross section while the accessory cells (projecting below the lower epidermis) are shown in longitudinal section. Mesophyll and other leaf tissues are not illustrated. Note variation in the differentially thickened walls of epidermal cells. 1, E. thorelit (Kerr ie? 0) 2, E. paniculata (Van Royen 3592) ; 3, E. malaccensis (Kostermans E. siamensis ssp. siamensis (Poilane 14229); 5, E. siamensis ssp. andamanicc a (Parkinson 384) The pinnate venation is similar throughout the genus. The primary lateral veins number from nine to thirty-five per leaf. They are separated by a distance of several millimeters to a centimeter and a half (the dis- tance between major veins appearing to be determined by leaf size). These veins are somewhat arcuate-ascending and end in a variously developed submarginal vein. Each vein is supported above and below by an extra- xylary fiber sheath which is more extensive than the vein. The secondary veins are numerous and generally form direct connec- tions between adjacent primary veins. In the larger leaves, at least, these cross-connecting veins are almost at right angles to the costa. The pat- 1961 | NEVLING, REVISION OF ENKLEIA 377 tern thus formed is quite distinctive (Fics. 6, 8). The secondary veins have a fiber sheath associated with them in the same manner as the larger veins. The xylem is very poorly lignified, and, in Enkleia malaccensis, I could perceive no color reaction when testing with phloroglucin and hydrochloric acid. The xylary elements, with the exception of the proto- xylem, often have very thickened walls and reduced lumina. This pe- culiarity is discussed further in the paragraph on petiole structure. in 6-9. ieee of cleared leaves showing veins and ga eeeey fibers o species of Enkleia, * 3. Fics. 6, in E. thorelii (Thorel 2823); 6, with transmitted light showing venation pattern; 7, with eee light oe vena- n pattern as outlined by the Cotes fibers. FIG — 9, BE. malaccensis (Gr iffith transmitted light showing vena atio pa attern; 9, with polarized light showing venation pattern and fibers not peearer with the vena- 75) ) The extraxylary fibers accompanying the veins have very thick sec- ondary walls. In prepared material, the secondary wall is usually pulle away from the thin primary wall. Concentric lamellations appear to be present in the secondary walls but should be studied further. The sec- ondary walls react quite strongly with potassium iodide and sulphuric acid but give no reaction with phloroglucin and hydrochloric acid, thus indi- cating their cellulosic nature. These fibers, which appear superficially similar to the primary phloem fibers of Cannabis figured by Esau (f A), are intimately associated only with the veins in Enkleia thoreli (Fic. 7) and E. siamensis ssp. siamensis. In E. siamensis ssp. andamanica and E. paniculata a few of the fibers wander from the vein and intrude into adjacent tissues. The wandering of fibers reaches its zenith in E&. 378 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII malaccensis where they are found in great profusion (Fic. 9). They ap- parently grow apically, often striking out vertically until reaching the upper or lower epidermis and then running horizontally. They can be ob- served readily in leaf sections and appear to be similar to the fibers de- scribed by Thoday (1921) in the African Passerina filiformis L. Near the veinlet endings in E. thorelit and E. siamensis ssp. siamensis the fibers are replaced by irregularly shaped sclereids. Sclereids also are found asso- ciated with the vein endings in £. siamensis ssp. andamanica and E. pani- culata but are not as reduced as in the preceding two species. The situa- tion in FE. malaccensis could not be determined because of the too numerous wandering fibers. Cross sections of the petiole were studied to check information given by Metcalfe and Chalk (1950) indicating the presence of intraxylary phloem in petioles of Enkleia. In all species of the genus the xylem, at midpetiole, is arc shaped. The arc often has strongly recurved ends and appears somewhat involute. The tracheary elements, except for the protoxylem but including the parenchyma, have curiously thickened walls and very reduced lumina. When stained with safranin and fast green, only the middle lamella appears to stain well with safranin but the entire wall can be weakly stained with phloroglucin and hydrochloric acid. No color reaction is formed with potassium iodide and sulphuric acid, probably indicating very weak lignification of the wall. Visually the walls appear to be rather similar to those of the extraxylary fibers. The greatest degree of thickening has taken place in E. thorelit. Phloem is found on both sides of the arc, confirming the statement of Metcalfe and Chalk, but the taxo- nomic significance of this is doubtful. Few to many scattered extraxylary fibers are found surrounding the trace, but the fibers are not nearly as prevalent as in the stem and leaf. The fibers have thick nonlignified cellulose walls. Reproductive Morphology. The same terminology which I applied to the inflorescence in previous papers concerning the Thymelaeaceae (1959, 1961) is used here. The inflorescence is composed of the primary peduncle with a pair of bracts, a bracteole, a rachis, the secondary pedun- cles, and the flowers with their pedicels. The floriferous branches of Enkleia are composed of many inflorescences which collectively appear to form a single, large, terminal inflorescence. The species of Enkleia (with the exception of E. malaccensis) bear their inflorescences in an axillary position on the terminal portions of young shoots. Each axillary inflorescence is simple, with the individual members cymosely arranged, i.e., opposite, except in very vigorous plants in which they become alternate by elongation of the internodes. The floriferous shoots of Enkleia malaccensis are characterized by the development of several to many axillary branches which bear the indi- vidual inflorescences. The floriferous branches thus appear to be pani- culiform. A few specimens of E. siamensis subsp. siamensis, such as Poilane 14229, also branch in this manner. 1961] NEVLING, REVISION OF ENKLEIA 379 The reduced leaves subtending the axillary inflorescences are borne on the primary peduncle which they subtend; the trace initiation is, of course, below the point of external branching. Occasionally, this displacement is quite striking, as in Zippelius 148a. On one of the flowering shoots of this specimen the reduced leaves have been “carried out” onto the axillary shoots for about eight millimeters. The mechanics and meaning of this displacement are not entirely clear. It is interesting that the separation layer of the petiole appears to be forming at the point where the petiole and axillary branch fork. The entire pattern is essentially the same as that found in Linostoma. Vegetative growth is resumed by the development of dormant axillar buds subtending the floriferous region of the shoot. The old inflorescences die shortly after fruiting and are soon broken off. A single bracteole is borne at the apex of the primary peduncle. It is linear, to 8 mm. in length, and almost invariably persistent in fruit. In addition to the bracteole, the primary peduncle bears a pair of bracts. In this genus the bracts are small and very strongly keeled until anthesis. At anthesis they are opposite or subopposite and are borne near or below the middle of the primary peduncle (Enkleia paniculata and E. siamensis ssp. siamensis), below the middle of the primary peduncle (£. malaccensis and E. thorelii), or above the middle of the primary peduncle (E. siamensis ssp. andamanica). Subsequently, they rapidly expand and become plane. A change in orientation generally accompanies the expan- sion, for the bracts are suberect prior to anthesis and then become de- flexed, sometimes strongly so, in fruit. In addition, their position on the primary peduncle is radically shifted sometimes because of differential elongation of the axis. Elongation, during fruit maturation, takes place in all taxa except FE. siamensis ssp. siamensis and is in striking contrast to the situation in Linostoma in which elongation is negligible. The function of the bracts is still an open question, as I suggested previously (1961), although Ridley (1930, p. 93) says, “When ripe the fruit on its peduncle, with the 2 bract leaves attached, separates from the plant and, rotating rapidly, drifts away in the wind across the forest to some distance, very much in the same way as the fruit of a Lime tree does.” As in Linostoma, the separation layer occurs between the secondary peduncle and pedicel so that the chances of the drupe separating with the primary peduncle and its bracts seem slim. Mature bracts are elliptic to obovate in shape and are quite unlike the foliage leaves. In addition to shape, the bracts are dissimilar in other respects, namely, smaller size, lighter color, thinner texture, and more reticulate venation. The venation pattern is essentially the same as that of the leaf (particularly with respect to the orientation of secondary veins). The individual veins, however, tend to be larger, giving a more thickly veined aspect to the bracts. The submarginal vein is well de- veloped in Enkleia siamensis but poorly developed in EF. paniculata and E. malaccensis (E. thorelii not determined because of insufficient material). A study of the anatomical characteristics of the bract, based on cross sec- 380 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII tions, was not pursued, except in Enkleia malaccensis, because of the dif- ficulty in obtaining satisfactory results. The bracts from herbarium speci- mens re-expand poorly because of the large number of thin-walled cells. The epidermal cells lack the extreme thickening found in the leaves, but the stomatal distribution is the same (i.e., on the abaxial surface only). The stomatal structure, however, is modified and is discussed below. Since extreme crushing and distortion were found in the mesophyll, no observa- tions could be made on this tissue. The veins have the usual fibers asso- ciated with them. In E. siamensis and E. paniculata the outermost fibers of the fiber sheath are modified into shorter sclereids which have one wall conspicuously enated as in Linostoma, but in FE. malaccensis this does not occur. As might be expected the fibers in the bracts of E. malaccensis and E. paniculata wander from the veins just as in the leaves of these species. e bract petiole is enervated by a single arc-shaped trace with strongly inrolled margins. All cells with the exception of phloem cells embraced by the arc are heavily sclerified. Extraxylary fibers do not appear to be present. The stomatal structure is superficially dissimilar in leaf and bract. In the bract, the numerous accessory cells are fewer in number (generally four to eight), and they retain the tabular form which characterizes the sur- rounding epidermal cells. Thus, the stoma is not borne within a sub- stantial pit. I believe that this indicates the pit-type (as found in leaves of all Enkleia species and Linostoma decandrum) to be much more closely related to the ranunculaceous type (as in L. persimile and L. pauciflorum) than might be expected. It indicates also the need for ontogenetic studies of the stomatal differentiation to determine the sequence of divisions in the formation of the pit-type. Unfortunately, small leaves from herbarium specimens are too difficult to handle because of brittleness and iain to permit such a study, even if sufficient material was availab The flowers are arranged on the rachis in an indeterminate manner (the een flower being the first to bloom). The flower clusters are um- belliform to subracemiform, depending on the length of the rachis. The form of the flower cluster is somewhat static within the species, but it is not considered to be a dependable taxonomic character. The number of flowers per inflorescence varies from three to sixteen in Enkleia siamensis, with the range of variability of other species falling within this number. The flowers are bisexual, regular, pentamerous, perigynous and_ pedi- cellate. The terminology applied to the parts is the same as used in my previous papers concerning the family (1959, 1961). Further gross mor- phological information concerning the flower, beyond that presented in the following discussion, may be found in the specific descriptions. The calyx tube, as in other members of the family, is composed of the fused bases of calyx, corolla, and androecial members. It is tubular in all species except Enkleia malaccensis in which it varies to suburceolate. The exterior is variously pubescent; the pubescence in all cases is minute. The interior is glabrous in all species, except /. malaccensis in which the tube interior is completely glabrous, or villous above and glabrous below, or 1961] NEVLING, REVISION OF ENKLEIA 381 completely villous. The trichomes are exclusively unicellular and un- branched. The tube is vascularized by ten distinct veins which are marked ex- ternally by inconspicuous ridges. Five of the traces are in the antisepalous position (median traces) and five are in the alternisepalous position (each consisting of a fused pair of traces designated as commissural traces). Occasional branching of a main trace occurs, but these branches usually reunite with the parent trace. Approximately half way up the tube the traces to the alternisepalous stamens depart from the commissural traces and shortly thereafter the traces to the antisepalous stamens depart from the median traces. In Enkleta stamensis each commissural trace and median trace gives rise to two lateral traces slightly above the insertion of the alternisepalous stamens. These lateral traces fuse (i.e., one of commissural origin with one of median origin) forming an inv erted “V” pattern lying between the main traces. At the apex of this inverted ‘V” a trace is formed which contributes eventually to the vascularization of the calyx lobes. The commissural traces then dichotomize giving rise to the two lateral calyx lobe traces, and the median traces continue undivided as the calyx lobe median traces. In F. thorelit and E. paniculata a minor de- viation from this pattern is found. The median trace gives off a pair of lateral traces in the same position and manner as in E. siamensis, but the corresponding lateral traces from the commissural traces are not formed. The laterals (from the median trace only) thus do not form the inverted “V" pattern but fuse directly with the commissural traces. This pattern appears similar to that demonstrated by Heinig (1951, Fig. 56) in Pimelea prostrata, except that in Enk/leia the commissural traces are fully de- veloped for the entire length of the tube. The size and shape of the five calyx lobes vary insignificantly among the species. The lobes are vascularized by three or five main veins which send off numerous branches which anastomose repeatedly. As in Linostoma, the aestivation of the lobes is quincuncial. The petals are inserted at the orifice of the calyx tube in an Loan sepalous position and are erect and glabrous in all species. They are di- vided medially either completely or incompletely. The median division is incomplete in Enkleta paniculata and E. malaccensis but complete in E. siamensis, and EF. thorelti. The shape of the lobes is ensiform (EF. pani- culata), linguiform (£. malaccensis), or liguliform (EF. siamensis and E. thorelii). The petals can be interpreted in two conflicting ways depending on the supposed origin of these structures. Heinig supports the hypothesis that the structures are stipular in nature (see her 1951 publication for com- prehensive discussion), while I support a petalaceous origin (1960, 1961). It is, admittedly, easier to explain the various types found within the family on the basis of her hypothesis but I feel that an attempt to gain information in support of the other view must be made. I interpret the in- complete division of the lobes as primitive and the complete division of the lobes as advanced (Heinig’s position would be quite the opposite). 382 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Only the petals of Enkleia thorelit have any hint of vascularization. I found one petal lobe which was vascularized for about one-third its length, the vascularization consisting of a single trace at most three tracheids in diameter and with a few accompanying fibers. Several other lobes bore a few fibers in the position that vascularization would be expected. I in- terpret the presence of these fibers as indicating a degeneration of a previously existing vascular system. The vascular supply of the single lobe and the fibers of several additional lobes appear to arise from the lateral calyx lobe trace. The androecium consists of ten stamens inserted on the calyx tube. They are in two whorls, the upper whorl in the antisepalous position and the lower whorl in the alternisepalous position. The traces which supply the stamens reflect the external position of the stamens in that they also depart from their parent traces in two distinct whorls. The filaments are short (to 1 mm. long), stout, and slightly inflated just below the anther. They are glabrous except in some specimens of Enkleia paniculata in which a few trichomes may be present at the base of the anther. Vascularization is by a single bundle which either ends at the base of the connective (EZ. siamensis and E. malaccensis) or extends to the very distal portion of the connective (EF. paniculata and E. thorelii). The antisepalous anthers are exserted or subexserted, except in Enkleia siamensis ssp. andamanica where they are included. Alternisepalous anthers are included except in E. paniculata in which they are subexserted. They are 4-lobed and 4-loculed in cross section. Dehiscence is longi- tudinal. There is some basal inflation of the connective but not nearly so pronounced as in Linostoma. The connective is minutely produced beyond the pollen sacs and, on the basis of cell shape and size, may be glandular in function. The pollen of the four species is, for all practical purposes, uniform. It is polyporate and highly sculptured. It cannot be distinguished from the pollen of Linostoma. The disc is borne at the base of the calyx tube. It consists of minute free lobes in Enkleia siamensis and E. thorelii, while in E. paniculata it is basally adnate to the calyx tube. No vascularization is present. The disc in FE. malaccensis is completely adnate and amounts to only a few additional cells in the thickness of the calyx tube wall. It is regularly lobed with the apex of each lobe coinciding with one of the main veins of the tube. The gynoecium is composed of a single pistil, presumably of the pseudomonomeric type. It is sessile or nearly so. I have found it im- possible to clear the gynoecium satisfactorily, even when placed in clearing agents for greatly extended periods of time, so that vascular patterns could not be established. The style is borne terminally on the ovary. It is filiform, glabrous, and sometimes variable in length. Style length in comparison to ovary length has been used in the past as a character to separate Enkleia and Lino- stoma. An evaluation of this character is made in the discussion of E. 1961] NEVLING, REVISION OF ENKLEIA 383 siamensis. The stigma is subcapitate in FE. paniculata and E. stamensis but clavate in EF. malaccensis and E. thorelii. As in Linostoma, the drupes of Enkleia are preserved poorly and I do not feel that I can comment upon them. The calyx tube is persistent but not accrescent. It is longitudinally ruptured by the developing drupe in E. malaccensis but transversely ruptured in E. paniculata and E. siamensis (see also discussion following EF. paniculata). The fruit of E. thoreli is unknown at this writing. Distribution of the species of Emkleia: squares, E. paniculata; circles, E. pee large solid dots, E. siamensis ssp. siamensis; small solid dot, E. Siamensis ssp. andamanica; stars, E. thorelit. GEOGRAPHY The four species of Enkleia have a composite range extending from Burma, Thailand, and Cambodia, on the north, through the Malay Pen- insula, Andaman Islands, and Sumatra to Bangka, on the south, eastward through Borneo to western New Guinea and the Philippine island of Luzon. The distribution is strongly correlated with specific delimitation with the exception of E. paniculata. The disjunct distribution of the latter species (Luzon and western New Guinea) is puzzling indeed, but the species is represented so poorly in herbaria that additional collections may show this disjunction to be an artifact. The distribution of each species is shown on the accompanying map. The symbols indicate place of collection only and are in no way indicative of the relative frequency of the plants. 384 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII MATERIALS This revision is based on specimens from the following herbaria, the abbreviations for which are taken from Lanjouw & Stafleu’s /ndex Her- bariorum, Part I, Ed. 4 (Regnum Vegetabile, 15. 1959). A Arnold Arboretum of Harvard University, Cambridge BM British Museum (Natural History), London re Botanical Museum and Herbarium, Copenhagen bp Forest Research Institute & Colleges, Dehra Dun F Chicago Natural History Museum, Chicago c Conservatoire et Jardin botaniques, Genéve GH Gray Herbarium of Harvard University, Cambridge K Herbarium, Royal Botanic Gardens, Kew 5 Rijksherbarium, Leiden P Muséum National d’Histoire Naturelle, Paris us U.S. National Museum, Smithsonian Institution, Washington I wish to take this opportunity to thank the directors and curators of the above institutions for allowing me to examine the specimens in their care. I also wish to thank, in particular, Drs. C. E. Kobuski, C. E. Wood, Jr., and O. T. Solbrig for unfailing interest, constructive advice and tech- nical assistance. TAXONOMY Enkleia Griffith, Calcutta Jour. Nat. Hist. 4: 234. 1844 (Type: E. malaccensis Griffith) . — — Linostoma Kurz, Jour. As. Soc. Bengal 39(2): 83. 1870, as to specie : Macgre corianths Merrill, Philippine Jour. Sci. 7: 312. 1912 (Type: M. pani- culatus Merri Erect or scandent shrubs or lianas, the axillary branches sometimes modified for climbing, the bark containing many fibers. Leaves opposite, subopposite, or rarely alternate, simple, pinnately veined, the primary lateral veins arcuate, with numerous oblique secondary veins; entire, petiolate, estipulate. Inflorescences laxly clustered towards the apex of the young shoots, the bracts small and keeled at anthesis, becoming plane and much enlarged in fruit, the primary peduncle generally elongating fol- lowing anthesis, Flowers bisexual, pentamerous, perigynous; calyx tube tubular, nonarticulated, ribbed, variously pubescent; calyx lobes 5, quin- cuncial, spreading; petals 5, medially cleft to the base or nearly to the base, inserted alternisepalously at the orifice, the lobes ensiform, linguiform or liguliform, glabrous; stamens 10, inserted in 2 whorls, the upper whorl antisepalous, the lower whorl alternisepalous, the anthers short fila- mented, basifixed, longitudinally dehiscent, the connective minutely pro- duced beyond the pollen sacs; disc of irregular lobes, minute, free to adnate to the calyx tube; gynoecium single, pseudomonomeric, superior, 1961 | NEVLING, REVISION OF ENKLEIA 385 unilocular with a single anatropous ovule, the style terminal, the stigma clavate or subcapitate. Fruit drupaceous, the calyx tube variously rup- tured by the developing drupe. KEY TO THE SPECIES a. Leaf surface at most minutely puberulent, often glabrescent beneath, the trichomes short, straight, at most suberect; calyx tube becoming spirally twisted in the upper half before being transversely ruptured by the developing drupe; petals cleft nearly to the base; alternisepalous ce subexserted. eT ee te ee eee eer ee 1. E. paniculata. a. Leaf ‘surface tomentulose, consiades glabrescent or glabrous Beach. the trichomes of medium length, curly or at least curved, suberect to erect; calyx tube never twisting before are transversely or longitudinally eriased by the developing drupe; petals cleft to the base or nearly to the base; alter- nisepalous stamens include b. Erect shrubs. scramblers or lianas; leaves tomentulose, velutinous, gla- brescent or glabrous beneath; inflorescence umbelliform; calyx tube 5-8 mm. long; ovary densely sericeous. Primary lateral veins 10-15 pairs per leaf; calyx tube glabrous, par- tially or completely villous within, longitudinally ruptured by the de- veloping drupe; petals cleft nearly to the base, about 1 mm. long; antisepalous stamens exserted; disc completely adnate to the calyx 2. E. malaccensis. Primary lateral veins 20-35 pairs per r leaf: calyx tube glabrous within, transversely Sik see by the developing drupe; petals cleft to the base, 2-2 m. long; antisepalous stamens included to subexserted; siamensis. a disc free. ke Erect shrubs, usually low; leaves glabrous beneath; inflorescence subra- cemiform; calyx tube 6-11 mm. long; ovary sparsely sericeous. ........ 4. E. thorelit. — 1. Enkleia paniculata (Merrill) Hallier f. Med. Rijksherb. 44: 26. 1922 Macegregorianthus paniculatus Merrill, Philippine Jour. Sci. 7: 312. 1912 (Type: McGregor B.S. 12360!). Enkleia zippeliana Hallier f. Med. Rijksherb. 44: 26. 1922 (Type: Zippelius 148a!) Climbing shrubs or lianas; young stems minutely ochraceous-velutinous and glabrescent, reddish brown, the lenticels conspicuous, horizontally elongate. Leaves opposite, subopposite or alternate; leaf blades ovate to elliptic or oblong-elliptic, 4-11 cm. long, 2.5—6 cm. broad, acute to obtuse at the apex, obtuse at the base, subcoriaceous, minutely and_ sparsely puberulent, often glabrescent, slightly darker above than beneath, the costa plane to immersed above, elevated beneath, the primary lateral veins 9-20 pairs, the submarginal vein coinciding with the margin; petiole terete or nearly so, 3-9 mm. long, minutely puberulent or velutinous. Inflor- escences borne terminally-on the young stems, usually 15-45 cm. long, many nodes involved, minutely velutinous throughout, each inflorescence 386 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII 5—12-flowered, subumbelliform, the primary peduncle 2—3.5 cm. long, en- larging to 6 cm. in length with most of the elongation above the insertion of the bracts, usually bearing the reduced leaf of the subtending node near the base, the rachis 1-2 mm. long, the secondary peduncles to 2 mm. long, the bracts at anthesis opposite or subopposite, borne near the middle of the primary peduncle or below, keeled, small, minutely velutinous, suberect, the bracts in fruit becoming plane, narrowly elliptic, 5—6.5 cm. long, blunt at the apex, obtuse to subcordate at the base, membranaceous, deflexed, the bracteole borne near the summit of the primary peduncle, linear, 2-5 mm. long, persistent. Calyx tube tubular, 5.5-6.5 mm. long, 1—1.5 mm. in diameter at the orifice, minutely puberulent without, glabrous within; calyx lobes ovate, 2—2.5 mm. long, 1—-1.25 mm. broad, minutely but densely puberulent within, erect; petals cleft nearly to the base, the lobes ensiform or when immature erosely truncate, ca. 1 mm. long, glabrous, erect; stamens inserted in two whorls, the antisepalous whorl inserted just below the petals, exserted as far as the petal apices, the alternisepalous whorl inserted about one anther’s length below the anti- sepalous one, subexserted, the anthers oblong, 0.5—1 mm. long, 0.25 mm. broad, the filaments 0.25—0.75 mm. long, thick, glabrous or with a few minute trichomes near the base of the anther; disc of minute, irregular lobes, adnate to the calyx tube; ovary ellipsoid, densely short-sericeous, the style terminal, filiform, ca. 2 mm. long, glabrous, the stigma sub- capitate, papillose; pedicel 2.5-5.5 mm. long, greatly thickening in fruit. Only one fruit per inflorescence developing; calyx tube spirally twisting before being ruptured by the developing drupe, torn portions persistent at the base; drupe ovoid to ellipsoid, ca. 1.5 cm. long, 1 cm. in diameter, sericeous, becoming glabrescent. ILLUSTRATION. Thymelaeaceae. Jn: Flora Malesiana I. 6: 1-48, fig. 10, hi. 1960 DISTRIBUTION. Collected only three times, once on Luzon and twice in western New Guinea. The McGregor collection from Luzon was found “on forested slopes at low or medium altitudes.” This collection was made in December when the plant was beginning to flower. The Van Royen collection from New Guinea was made in December in flower and fruit. Van Royen wrote of the locality, “The area in which the species is found consists of elevated limestone cliffs, alternating with brown muddy swamps.”! He says, !n addition, ‘I am quite surprised that my collection of this species is only the second one as this species was really common where I found it. I had to scale a relatively steep slope covered with an open forest of Quercus sp. and Pometia as the dominating trees. This Enkleia was climbing on an Artocarpus vrieseanus Miquel, but was found climbing also over Paratrophis sp., Cyathocalyx papuanus Diels and over both Quercus and Pometia. It is a rather striking species owing to its relatively large flowers.” ' Personal communication. 1961] NEVLING, REVISION OF ENKLEIA 387 Philippine Islands. Luzon: Calauan, Prov. of Laguna, McGregor B.S. 12360 (x-lectotype, BM, L, US). Netherlands New Guinea. Steenkool, road to Tem- boeni, km. 1.5, hills east of the road, opposite police barracks, Van Royen 3592 (L); without locality, Zippelius 148a (L-holotype and isotoypes of E. zip- peliana). Presumably, the holotype of this species was destroyed with the Phil- ippine National Herbarium. As lectotype I choose the specimen of McGregor 12360 deposited in the herbarium at Kew. It is the most complete specimen of this collection and consists of two immature florif- erous branches and a packet containing a nearly mature flower. Merrill allied his new genus Macgregorianthus with Wikstroemia, a large and wide-ranging genus of the Eastern Hemisphere. He placed it near this genus, undoubtedly, because of his familiarity with it rather than because of any morphological similarities. There is no question that Macgregorianthus (= Enkleia) is not closely related to Wikstroemia, al- though they are members of the same subfamily. The calyx tube, immediately following anthesis in the New Guinea specimens, becomes spirally twisted in the upper half. It twists, up to a half dozen times, quite tightly before being ruptured by the developing drupe. The twist is invariably counter clockwise. The significance of this peculiar behavior has not been ascertained. Enkleia paniculata was first united with E. zippeliana shortly after Hallier’s description of the latter species. Hou recently (1960) pointed out that the distinguishing features mentioned by Hallier in the original description of E. zippeliana are of little taxonomic consequence. These features include leaf pubescence, pubescence color and prominence of leaf venation. There is, however, a discrepancy in one major floral character which requires explanation. The flowers of the New Guinea specimens collected by Zippelius and Van Royen have petal lobes which are ensiform, whereas, the McGregor collection consisting of immature flowers has petal lobes which are broadly squamelliform with the apices erosely truncate. The petals have not fully developed in the McGregor specimens. Fortunately, the superb Van Royen collection consisting of flowering and fruiting ma- terial is available so that immature flowers can be compared with those Gregor collection, the petal lobes have already undergone major elongation and they have assumed the mature form, but in somewhat younger buds the erosely truncated petal apices so characteristic of the Luzon collection can be observed. Cleared leaves of the New Guinea specimens have a remarkable number of druse-like crystals which can be observed readily under low magnifica- tion. The leaf specimen of McGregor 12360 which I cleared was ab- solutely devoid of crystals. The applicability of this character is very doubtful as crystal formation may be due to the physiological state of the plant at the time of collection and not directly genetically controlled. Indeed, in Enkleia malaccensis crystals can be found in the leaves of most 388 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII of the specimens but are entirely lacking in others, Griffith 4375, for ex- ample. Since all of the leaves were handled in a similar manner there is little chance that crystal formation was induced by the clearing process in some instances and not in others. 2. Enkleia malaccensis Griffith, Calcutta Jour. Nat. Hist. 4: 234. 1844 (Type: Griffith ? Lasiosiphon scandens Endl. Gen. Suppl. 4(2): 67. 1847, based on Enckleia | Enkleia | malaccensis Griffith. Enkleia malayana Griffith, Notul. As. 4: 363. 1854. Linostoma scandens (Endl.) Kurz, Jour. As. Soc. Bengal 39(2): 83. 187 Enkleia riouwensis Hallier f. Med. Rijksherb. 44: 25. 1922 (Type: 7 ssamane s.n,! Enkleia coriacea Hallier f. Jbid. (TypE: Hallier B.2 264!). Climber; young stems terete, densely golden-brown velutinous, gradually glabrescent and conspicuously lenticellate, reddish brown: axillary branches often strongly recurved uncinate, thickening at the base with age. Leaves subopposite, the blade ovate to elliptic (sometimes broadly so), 3.5-14(-15) cm. long, 2-7(-10) cm. broad, acute to obtuse or emar- ginate or rarely short-acuminate at the apex, cuneate to obtuse at the base, coriaceous, both surfaces golden-brown velutinous and glabrescent, the trichomes slightly curved, the costa immersed above, elevated beneath: the primary lateral veins 10-15 pairs, the submarginal vein coinciding with the margin, conspicuously thickened; petiole terete, densely puberulent al- though somewhat glabrescent in very old leaves, 6—-10(—12) mm. long. Inflorescences borne terminally on the young shoots, involving several nodes, generally compound and forming a paniculate structure often 15 cm. long but sometimes to 30 cm. in length, golden-velutinous throughout; each inflorescence 3—9-flowered, umbelliform, the primary peduncle 3—5 cm. long, enlarging to 9 cm. in fruit with most of the elongation above the insertion of the bracts, usually bearing the very reduced leat of the sub- tending node near the base, the rachis to 2 mm. long but sometimes elongating in fruit, the secondary peduncles obsolete to 2 mm. long, the bracts at anthesis subopposite, borne slightly below the middle of the primary peduncle, keeled, 3-5 mm. long, golden-brown velutinous. erect, the bracts in fruit becoming plane, elliptic to obovate 4—7 cm. long, 1—2.5 cm. broad, obtuse at the apex, subtruncate at the base, puberulent, con- spicuously veined, strongly deflexed, the bracteole borne at the summit of the primary peduncle. linear, 1-3 mm. long, often persistent. Calyx tube tubular to suburceolate, 5--6 mm. long, green to yellow and densely golden- brown velutinous without, glabrous, villous above and glabrous below or completely villous within; calyx lobes more or less oblong but obtuse at the apex, 1-2 mm. long, 0.75 mm. broad, erect, white puberulent within; petals Cleit almost to the base, the lobes somewhat linguiform, ca. 1 mm. long. somewhat carnose, glabrous, exserted: stamens inserted in two whorls, the antisepalous stamens inserted just below the petals, exserted 1961 | NEVLING, REVISION OF ENKLEIA 389 but not as far as the petals, the alternisepalous stamens inserted one to two anther-lengths below the antisepalous whorl, included, the anthers oblong, 0.5-1 mm. long, 0.25-0.5 mm. broad, the filaments obsolete to 0.5 mm. long, slightly inflated just below the anthers, glabrous; disc com- pletely adnate to the calyx tube, ca. 0.5 mm. tall, somewhat fleshy; stigma included, clavate, papillose; style filiform, 0.5—1 mm. long, the ovary fusiform, ca. 2 mm. long, densely short-sericeous; pedicel 1-3 mm. long. Usually only one or rarely two fruits per inflorescence developing; calyx tube persistent, usually longitudinally ruptured and persistent at the base of the drupe; drupe ovoid, 10-15 mm. long, 6-8 mm. in diameter, glabrescent but usually puberulent at least at the apex, prominently ribbed; pedicel thickening in fruit. ItLustRAtiIons. Ridley, H. N., Dispersal of plants throughout the world. tab. 6, fig. 13, L. Reeve & Co., Ltd., 1930; Thymelaeaceae. In: Flora Malesiana I. 6: 1-48. fig. 10, a—g. 1960. DistriBuTION. Primary forest in Malaya, south to Sumatra and Bangka, east to North Borneo. All specimens which I have seen indicate that this species is never found above 50 meters altitude. It has been col- lected flowering in April, September and October; fruiting in April, May, August and September. Merrill (1929, p. 212) says, “A woody vine forming tangled masses in forests, the stems a few inches in diameter.” Keith reports that the bast fibers can be used for tying purposes, and Burkill (1935, p. 925) says that it yields an inferior aromatic wood. K, L): without collector (possibly Maingay?) 1256 “Sept. 1867” (Gc). Indonesia. Sumatra: Laboehan Batoe; Dist. Kota Pinang, Si Mandi Angin (on the Soengel Kanan). Toroes 4099 (a). BINTAN: Tandjung-Pinang, Teysmann sn. (L- holotype and isotoype of Enkleia riouwensis). BANGKA: G. Maroe, Anta 1373 (a). Borneo: Tandjong Bangko region, near mouth of Mahakam river, Koster- mans 7034 (L); Loa Djanan, west of Samarinda, Kostermans 6672 (L); Pleihari [Pelaihari], V.J.F.S. bb. 19256 (A, L); zwischen dem Sungai Djéméla und dem Ganiing K(e)lamm, Hallier B.2264 (-holotype and isotype of Enkleia coriacea). British North Borneo. East Coast: Selangan, F.R., Semporna, Keith N.B.F D.N. 9234 (kK, L); Tawao [Tawau], Elphinstone Province, Elmer 20834 (BM, ae i It appears that a type specimen has never been selected for this species, although I have seen several specimens which have been considered by others to be type material. This species was described by Griffith (1844) in an article entitled, “On some remarkable Plants in the H. C. Botanic Gardens. Calcutta.” However, immediately preceding the description of Enkleia (p. 234) he says, “The other plants of this family that I have met with on the N. E. frontier and Straits of Malacca, localities that, how- ever distant, present remarkable affinities in vegetable forms, are-.”’ 390 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Enkleia malaccensis is the last of nine species in a list and it is marked clearly from “Malacca.’”’ The actual description is contained in an exten- sive footnote in which no specimen was selected as a type. Thus, it would seem that a selection of a type would be restricted to specimens collected by Griffith (or one of his collectors) in Malacca prior to 1844 but, as I intend to show, the selection of a type is relatively complicated. Griffith first set foot in Malacca in August, 1841 (see Van Steenis- Kruseman, pp. 201, 202. 1950), and served as a civil surgeon until mid- 1842. During this period he made collections and trained native collectors. In August of 1842 (see Griffith 1847, introductory notes) he left to as- sume the directorship of the Botanic Gardens in Calcutta, a post which he occupied until December, 1844. In January of 1845 he returned to Malacca. Of no little consequence is that during his directorship at Cal- cutta he remained in contact with his native collectors in Malacca and they sent him a steady stream of dried specimens, seeds, and living plants. The living plants were grown in the gardens (see Griffith, 1847, Extract of Letters). One would assume from the title of Griffith’s paper that the plants which he intends to discuss are growing in the botanic garden until he adds the qualification that he has met some of them elsewhere. In the case of the locality of this particular species I believe that it can be interpreted in two ways: one, the description is based on a plant which he collected in Malacca between August, 1841, and August, 1842; or two, the description is based upon a specimen growing in the botanic garden which was sent from Malacca by his collectors, but which may have been a plant with which he had familiarity prior to leaving Malacca. If the description is based upon one of his Malaccan collections there is a possibility that the specimen did not go through his normal channels of distribution, for, in a letter to R. Wight, postmarked Malacca, April 15, 1842 (Griffith 2: xxvii) he says, “I intend adopting an entirely new system of tactics, and when I pounce on any thing interesting, to keep it until it is in print.” On the other hand, if the description is based upon a living specimen there is the distinct possibility that no herbarium specimen was made. As an example, in the same letter quoted above, he says, of three new genera, “which I suppose no body will adopt at home without specimens.” I have seen three specimens which have been considered to be authentic collections. Two of these are deposited at Kew. The one bears the stamp of “Herbarium Hockerianum, 1867,” and in longhand is written, ‘“Ma- lacca, Griffith, 1845. Enkleia Malaccensis Griffith;” the other bears the same stamp of Hocker, plus the printed label of the East India Company with the identification number 4375 (see Hooker, 1865), and is identi- fied as “‘Lasiosiphon scandens, Endl.” (a substitute name for Enkleia malaccensis), the place of collection is given as “Birma and Malay Pen- insula.” In addition, it bears the longhand identification “Enkleia Malayana Gr.” The third sheet, desposited at the Gray Herbarium, bears the identical East India Company label and identification as the Kew specimens but has no other marks. There is little question that all of 1961 | NEVLING, REVISION OF ENKLEIA 391 these sheets are Griffith collections but the sheet at Kew bearing the date 1845 is automatically eliminated as authentic type material. The remain- ing two sheets cannot be either confirmed or denied as type material. Griffith’s description of the new genus is lengthy, well drawn, and easily interpreted so that I believe that the choice of a type should be left an open matter at this time. Enkleia coriacea, on first examination, appears to be a distinct species. The aspect of the two Leiden specimens which seem to set them apart from others is due to a decidedly reddish cast to the leaves which is coupled to an almost complete lack of pubescence. It is surprising how this com- bination of characters is so readily noticeable. On very immature leaves the characteristic pubescence of E. malaccensis can be observed but the trichomes are shorter than usual. It is possible that these two specimens show one extreme of the total variation in trichome length of the species. The trichomes of other Bornean specimens in general appear to be shorter than those of peninsular specimens but none reach the extreme attained in Hallier B.2264. It is of interest that these plants were growing in asso- ciation with Linostoma pauciflorum Griffith, a species of the genus most closely related to Enkleia. Enkleia riouwensis was described by Hans Hallier from sterile collec- tions of Teysmann from the Riouw island of Bintan. Hou (1960), in his treatment of Thymelaeaceae for the Flora Malesiana, relegated this species to the synonymy of E. malaccensis, an action with which I agree. I have examined the holotype and an isotype from the Leiden herbarium and can find no characteristics to distinguish it from £. malaccensis, except a more pronounced undulation of the leaf margin which may be due to the method of drying. In the description, Hallier refers to the plants as shrubs but I can find no such information on the two specimens which I ex- amined, so I regard this character as questionable. 3. Enkleia siamensis (Kurz) Nevling, comb. nov. Linostoma siamense Kurz, Jour. As. Soc. Bengal 39(2): 82. 1870 (TYPE: Teysmann 5986!). Erect to scrambling shrubs or large lianas; young stems terete, puber- ulent or golden to light-brown velutinous, glabrescent; axillary branches usually normal, only rarely becoming uncinate. Leaves opposite, sub- opposite or rarely alternate on vigorous young shoots; leaf biade ovate, lanceolate, oblong-elliptic, broadly elliptic or rarely orbicular, 4-13 cm. long, 2-9 cm. broad, acute, attenuate, or obtuse to truncate and minutely mucronulate at the apex, cuneate to obtuse at the base, coriaceous, gla- brous or densely to sparsely tomentulose and often glabrescent, the tri- chomes curly, the costa immersed above, elevated beneath, the primary lateral veins 20-35 pairs, the submarginal vein poorly or well developed, coinciding with the margin; petiole scarcely canaliculate, glabrous or velutinous, 5-8 mm. long. Inflorescences borne on the modified portions of the young shoots, involving several nodes (to as many as 8), the main 392 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII axis seemingly leafless, to 15(—30) cm. long, minutely golden-velutinous throughout; each inflorescence 3—16-flowered, umbelliform, the primary peduncle 2.5—5 cm. long, sometimes elongating to 9 cm. in fruit, with the reduced leaf of the subtending node near the base, the rachis to 2 mm. long, the secondary peduncles to 1 mm. long, the bracts opposite to sub- opposite, keeled at anthesis, becoming plane and more or less oblong, 2—9 cm. long, 1-2.5 cm. broad, obtuse at the apex, obtuse to subcordate at the base, membranaceous, tomentulose, deflexed, the bracteole borne at the summit of the primary peduncle, linear, 1-8 mm. long, often persist- ent. Calyx tube tubular, 5-8 mm. long, 1-1.5 mm. in diameter at the ori- fice, golden-tomentulose without, glabrous within: calyx lobes spathulate, 2.5—4.5 mm. long, 1.5—2.5 mm. broad, minutely puberulent within, spread- ing; petals cleft to the base, the lobes liguliform, ca. 2.5 mm. long, glabrous, fleshy, long-exserted; stamens inserted in two whorls, the antisepalous whorl inserted just below the petals to 1 anther-length below, subexserted or included, the alternisepalous whorl inserted 2-3 anther-lengths below the antisepalous whorl, included, the anthers oblong, 0.5—1 mm. long, ca. 0.25 mm. broad, the filaments 0.25—1 mm. long, somewhat inflated. gla- brous; disc of minute free lobes, glabrous: ovary ellipsoid, 1-2.5 mm. long, densely sericeous, the style of varying length, filiform, glabrous, the stigma subcapitate, included to exserted (see discussion); pedicel 3—6.5 mm. long. Only a single fruit per inflorescence developing; calyx tube transversely ruptured but persistent as a short ruffle at the base of the drupe; drupe ovoid to almost ellipsoid, 10-18 mm. long, 6-8 mm. in di- ameter, puberulent at least at the apex; pedicel thickening in fruit. The proper taxonomic treatment of this species and the preceding, Enkleia mataccensis, is somewhat perplexing but it has been my decision to recognize two species, one with two subspecies. There is little doubt that these taxa are derived from common ancestral stock as indicated by morphological similarities and present day distribution. Each of the taxa is allopatric and the possibility of interbreeding between them seems re- mote because of distance alone. Enkleia mataccensis is most certainly distinct from E. siamensis ssp. siamersis as can be demonstrated with any number of morphological or anatomical characters, many of which are apparent to the unaided eye. t is with the introduction of specimens from the Andaman. Islands (originally described as Linostoma andamanica and here treated as 1 siamensis ssp. andamanica) that difficulties begin. Unfortunately, only two collections from the Andamans are available, and this seriously im- pairs knowledge of the variability of these plants; but even with these few it is possible to see that they combine characteristics of E. malaccensis and EF. siamensis ssp. siamensis. Further, they entirely lack distinguishing characters of their own; otherwise I would recognize three species. Of the characters which are combined, the great majority are from £, siamensis ssp. siamensis. It is not surprising to find the Andaman plants most closely related geographically with those of ssp. siamensis. Char- 1961 | NEVLING, REVISION OF ENKLEIA 393 phological characteristics which now delimit the species were still in a state of flux and not yet geographically segregated, or they may be the result of considerable convergent evolution. The former case seems most probable because of the lack of distinguishing characters not represented in other taxa. Whether E. malaccensis or E. siamensis is primitive or ad- vanced, generalized or specialized, cannot be determined, for both have elements which I consider primitive and advanced. KEY TO THE SUBSPECIES a. Leaves densely to sparsely tomentulose, often glabrescent, the submarginal vein well developed; primary peduncle to 5 cm. long, racts at anthesis borne near the middle of the primary peduncle or below, 2-4(-5) cm. long, 1-1.5 cm. broad; calyx tube 6-8 mm. long; antisepalous stamens subex- CELE, ch Geena ne Bakken eRe ET Sees 3a. E. siamensis ssp. siamensis. a. Leaves glabrous, the submarginal vein poorly developed; primary peduncle to 9 cm. long; bracts at anthesis borne above the middle of the primary peduncle, 4-8 cm. long, 1.5-2.5 cm. broad; calyx tube about 5 mm. long; antisepalous stamens included. ........ 3b. E. siamensis ssp. andamanica. 3a. Enkleia siamensis ssp. siamensis Linostoma siamense Kurz, Jour. As. Soc. Bengal 39(2): 82. 1870 (TYPE: Teysmann 5986!). Linostoma scandens var. cambodianum Lecomte, Not. Syst. Paris 3: 127. 1915. ‘“cambodiana” (Type: Pierre 511!). DistripuTion. Open deciduous forest or scrub in Burma, Thailand and Cambodia. Collected from 100 to 1200 meters altitude. Flowering in January, although buds apparently forming as early as November; fruit- ing in January, February and March. Burma. Pecu: Pegu Yomak, e. and w. slopes, Kurz 3122 (kx). Thailand. CualvapHum: Chaeyapuin |Chaiyaphum], Kerr 19959 (BM). CHIANGMAT: Lampang Mae Saikham, Wanadorn 1590 (L). CHonpurt: Kanboerie | Chon- buri], Teysmann 5986 (L-isotype of Linostoma siamense), Kerr 10136 (BM), 10620 (sm). Nakon PHANoM: Nawkawn Panom [Nakhon Phanom], Kerr 8214A (pM). PHeETBURI: Tung Quang, Petchaburi, Kerr 20601 (pM). PRACHUP: Kan Kradai, Prachuap, Put 2305 (pm); Hua Hin, Kerr 16136 (BM). Nang Rawng (?) Kerr 8214 (BM). Cambodia. In montibus Krewand Prov. Tpong [Thpong], Pierre 511"'* (A, BM, G, GH, K, P); in prov. Pei Lover (see discus- sion), Pierre 511 (isotypes of Linostoma scandens var. cambodianum: A, BM, F, K, L, P); Stung Treng, Poilane 14229 (Pp); without locality, Julien s.n. (Pp). 394 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XLII In the original description of Linostoma siamense, Kurz gave the place of collection as, “Siam, Bookit Kathay near Kanburi.” Attempts to locate Bookit Kathay were in vain. Re-examination of the handwritten label on the isotype showed the collector’s name and number plus, ‘‘Daph- noideae. Boekiet Kethaay, Kanboerie, Siam.” Therefore, it appears as though Kurz removed the comma following “Kethaay” and inserted the word “near.’”’ Teysmann’s account of his travels (1863, p. 201) shows that “Boekiet Kethaay” is not the place of collection but the common name! The place of collection is Kanboerie (= Bang Pla Soi, Chanburi, Cholburi, Chonburi, Jolburi or Kanburi) which Teysmann visited for several days toward the end of March, 1862. Pierre 511 is a collection consisting of two parts. The first, collected in March, 1870, is the more important as it serves as the basis of Lecomte’s Linostoma scandens var. cambodianum. In the original description of this varietas Lecomte gives the location as “Prov. de Pan Louvea.” Examina- tion of the numerous sheets of this collection has added these possibilities: Pen Saver, Peti Saver, Pei Sover and Ped Lover, Unfortunately, I have not been able to locate, to my satisfaction, the place of collection. I have the following possibilities: Kompong Svay (ca. 13° N, 195° E), Prey Lovea (ca. 11° N, 105° E) and Lovea (ca. 13° N, 103° E) and therefore have not plotted this collection on the map. The second portion of Pierre 511 was collected June, 1870, “in montibus Krewand prov. Tpong.”” The locality is probably Thpong and I have plotted it so. This portion of the collection has been designated as 511 is. The displacement of the reduced leaves borne near the base of the primary peduncles has been discussed previously (Nevling, 1961). These leaves generally are only several millimeters in length, although in Kerr 16136 they are fully developed, normal leaves. This evidence helps se- cure the hypothesis that these reduced leaves are homologous with the leaves of the vegetative shoot. Only a very small number of the specimens of this subspecies which I examined were in flower. The description and measurements of the flower are, of necessity, based on few specimens and further variability should be expected. The final position of the stigma with respect to the calyx tube orifice is dependent upon the amount of elongation of the style. The mechanics of stylar elongation in Pimelea has been discussed by Burrows (1960) and a similar mechanism is suspected in Linostoma (Nevling, 1961). It is almost certainly in operation in this species of Enkleia. The elongation ceases at the time of pollination or shortly thereafter, and the length of the style thus is determined by the pollination time. In Wanadorn 1590 two floral dissections by Dr. Ding Hou show one with a very short style and the other with an extremely long style. Other collections which I have observed have the style of varying length. Unfortunately, the length of style has been used as a key character to separate Enkleia from Linostoma, and now the character appears to be useless. 1961] NEVLING, REVISION OF ENKLEIA 395 3b. Enkleia siamensis ssp. andamanica (Hutchinson ex C. E. Parkin- son) Nevling, comb. & stat. nov. Linostoma andamanicum Hutchinson ex C. E. Parkinson, Fl. Andaman Is. 229. 1923, “andamanica” (Type: Parkinson 384!). According to Parkinson, the plants are found on Baratang and Have- lock islands. They are dispersed but very uncommon. They flower and fruit during the hot season. The habit is described by Parkinson as “a woody climber as thick as a man’s arm and climbing over the tallest trees, stems hooked.” Andaman Islands. Ali Masfid Reserve, Parkinson 384 (pp-holotype of Lino- stoma andamanicum, «K); without locality, Parkinson 309 (DD). See discussion following specific description above. 4. Enkleia thorelii (Lecomte) Nevling, comb. nov. Linostoma thorelii Lecomte, Not. Syst. Paris 3: 127. 1915 (Type: Thorel 2823!). Erect shrubs to 0.5 m. tall; young stems terete, strigose and glabrescent, becoming reddish brown, the axillary branches normal. Leaves suboppo- site to alternate; leaf blades elliptic, 2.5-4.5 cm. long, 1-2 cm. broad, obtuse and mucronate at the apex, cuneate at the base, coriaceous, glabrous, the costa plane above, slightly elevated beneath, the primary lateral veins in- conspicuous, ca. 15-20 pairs, the submarginal vein coinciding with the margin, conspicuously thickened; petiole nearly terete, glabrous, 1-3 mm. long. Inflorescences borne from the young shoots, generally simple; each inflorescence 6—12-flowered, subracemiform, the primary peduncle 2—4 cm. long, elongating to 6.5 cm., with most of the elongation above the inser- tion of the bracts, bearing the reduced leaf of the subtending node near the base, the rachis 1-7 mm. long, the secondary peduncles 1-2 mm. long, the bracts at anthesis opposite to subopposite, borne below the middle of the primary peduncle, keeled to plane, S—25 mm. long, to 20 mm. broad, minutely strigose, the bracteole borne at the summit of the primary pe- duncle, linear, to 5 mm. long, minutely strigose, persistent. Calyx tube tub- ular, 6-11 mm. long, ca. 1.5 mm. in diameter at the orifice, minutely strigose to hirsute without, glabrous within; calyx lobes oblong, ca. 3.5 mm. long, 1.5 mm. broad, spreading, minutely puberulent within; petals cleft to the base, the lobes liguliform, ca. 4.5 mm. long, greatly exserted, erect to spreading, glabrous; stamens inserted in two whorls, the antisepalous ones inserted about 1 anther-length below the petals, subexserted, the alter- nisepalous whorl inserted about 3 anther-lengths below the antisepalous whorl, included, the anthers oblong, ca. 1 mm. long, 0.5 mm. broad, the filaments thick, 0.25—0.5 mm. long, glabrous; disc of many minute lobes; ovary fusiform, ca. 2 mm. long, very sparsely sericeous, the style filiform, ca. 1 mm. long, glabrous, the stigma clavate; pedicel 2-3 mm. long. Fruit unknown. 396 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII DisTRIBUTION. The Kerr specimen was collected in open scrub at an altitude of about 200 meters, flowering in January. Thus far known only from Thailand. Thailand. NAkoN RATCHASIMA: Dun Kuntot, Korat, Kerr 19930 (BM), Upon: Kemarath | Khemmarat ], Thorel 2823 (p-holotype) This species is notable for its erect habit, small leaves and large flowers with sparsely sericeous ovary. It is unfortunate that it is so poorly rep- resented in our herbaria LITERATURE CITED Burrows, C. J. Studies in Pimelea I. The breeding system. Trans. Roy. Soc. New Zealand 88: 29-45. BurkILL, I. H. A dictionary of the economic products of the Malay Peninsula. 1(A-H): 1220 pp. Crown Agents for the Colonies, 4 Millbank, London. Esau, K. Plant Anatomy. 735 pp. John Wiley & Sons, Inc., New York. 1953. GRIFFITH, W. On some remarkable plants in the H. C. Botanic Gardens, Cal- cutta. Cal. Jour. Nat. Hist. 4: 231-256. 1844. . (Arranged by J. M’Clelland.) Journals of Travels. 2 vols. Bishop's College Press. Calcutta. Hernic, K. H. Studies in ne floral morphology of the Thymelaeaceae. Am. Jour. Bot. 38: 113-132. 1951. Hooker, J. D. Catalogue of - plants distributed at the Royal Gardens, Kew. John E. Taylor. Lincoln’s Inn Fields. 1865. Hou, D. Thymelaeaceae. /n: Flora Malesiana I. 6: 1-48. 1960. MerriLt, E. D. An enumeration of Philippine flowering plants. 3: 132. 1923. . Plantae Elmerianae Borneenses. Univ. Calif. Publ. Bot. 15: 1-316. 1929. TEYSMANN, J. E. Verslag eener Reis naar Siam. Nat. Tijkschr. Nederl. v Indie 25: 149-208. 1863. Tuopay, D. On the behaviour during drought of leaves of two Cape species of Passerina, with some notes on their anatomy. Ann. Bot. 35: 585-601. 1921. VAN STEENIS-KRUSEMAN, M. J. Malaysian plant collectors and collections, being a cyclopedia of botanical exploration in Malaysia. Flora Malesiana 1: 1-639. 1950. 1961 | FOSTER, KINGDONIA UNIFLORA 397 THE FLORAL MORPHOLOGY AND RELATIONSHIPS OF KINGDONIA UNIFLORA ADRIANCE S. FOSTER THE BROAD COMPARATIVE STUDIES of Bailey (1) and his associates have revealed many examples of the combination of primitive and advanced levels of specialization in the vegetative and reproductive structures of woody members of the Ranales. Indeed, certain ranalian genera are the only known dicotyledons which exhibit such primitive characters as mono- colpate pollen, vesselless xylem, and “open,” styleless, conduplicate carpels. Despite the significant advances in our understanding of the trends of evolutionary specialization in the xylem and in floral organs, the nature of the truly primitive leaf type in angiosperms remains an unsolved problem. The predominance in the woody Ranales of simple leaves with pinnate-reticulate venation is regarded by Eames (8, p. 421) as “strong evidence that this is a primitive leaf type in the Angiosperms.” Bailey (2), however, takes a more conservative or skeptical viewpoint and states that the modern angiospermic leaf may have arisen “from one of several diversified forms of potentially ancestral appendages.” In full accord with this cautious approach to the problem, I have already raised the question of whether the “closed” reticulate type of venation, so character- istic of living angiosperms, may not have evolved from open dichotomous vasculature (9). This possibility is clearly indicated by the remarkable dichotomously veined leaf of Kingdonia uniflora, an herbaceous perennial which is classified at present as a member of the Ranunculaceae (10, 11, 12) The occurrence of dichotomous foliar venation in a plant of presumed ranalian affinities deserves full consideration in any theory of foliar evo- lution and emphasizes the need for a comprehensive study of ad aspects of structure and morphology in Kingdonia. For this reason an investi- gation of the morphology and vasculature of the floral organs of this genus has been undertaken and the results are presented in this paper. MATERIALS AND METHODS The flowering specimens used in my study were secured, through the kindness of Professor Yu-Wen Tsui, from the following Shensi collections of Kingdonia deposited in the herbarium of the Northwestern Institute of Biology at Wukung, Shensi Province, China, viz.: Liu é& Chun 1307, collected September 28, 1937, on Taipaishan; Ho, s.n., Taipaishan, Fan Yang Szu (Pasturing Sheep Temple); and Fu 12296, collected June 7, 1959, on Taipaishan, Grateful acknowledgment is made to Dr. Shiu-ying 398 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Hu, of the Arnold Arboretum, Harvard University, for translating into English the Chinese data included with the Ho collection of Kingdonia. Despite the fragile and desiccated condition of the material, the com- bined use of cleared specimens and serial microtomed sections made it possible to study many features of floral morphology in considerable a tail. All flower buds and flowers were first cleared in 2.5% NaOH a after thorough washing in water, transferred to cold concentrated chloral hydrate. Specimens cleared in this way proved suitable for drawings of the general organography of the flower and were subsequently dehydrated and lightly stained with safranin. After staining, the individual organs of some of the flowers were dissected from the receptacle and mounted directly in “Piccolyte.” Other flowers, as well as the buds, were de- hydrated by the tertiary butyl alcohol method, infiltrated with ‘Histo- wax,” sectioned 7—8» in thickness, and stained with a combination of safranin and light green. I wish to thank Mr. Roy L. Taylor for his skillful assistance jn all phases of the embedding, sectioning, and staining of the floral material. I am also greatly indebted to my wife who prepared the drawings repro- duced in this paper as Fics. 2—7, assembled the photomicrographs and assisted me in proofreading the manuscript. Thanks are also due Mr. Victor Duran for his skill and patience in making the photomicrographs. GENERAL ORGANOGRAPHY OF THE FLOWER According to Balfour and Smith’s (5) description of the type material collected by F. Kingdon Ward in northwestern Yunnan Province, China, the solitary, apetalous flower of Kingdonia is borne on a naked scape 7-10 cm. long and consists of 5 sepals, 10-15 stamens, and 5—7 carpels. My observations, restricted to a study of flowering specimens from a montane area in Shensi Province, differ in several important respects from this description. In the first place, it seems more appropriate to designate the spirally arranged perianth members as ‘‘tepals” rather than as “sepals” (Fic. 1). The latter term implies the former existence of petals in the flower of Kingdonia, but no evidence of vestigial petals or their traces was found in either cleared or sectioned flowers. Furthermore, it is obvious even from the limited material examined that the number of perianth mem- bers is inconstant. In some flowers five tepals are present (Fic. 1), but flowers with six or seven tepals have also been observed. These fluctua- tions clearly suggest that a well-defined ‘calyx,’ in the formal sense, is not present. A more serious divergence between Balfour and Smith’s description and the present study concerns the nature of the androecium. In all the flowers I have examined, the androecium consists of an outer series of 8-12 spirally arranged staminodia and a more centrally located series of 3—6 stamens, each of which is demarcated into anther and filament (Fics. 1961] FOSTER, KINGDONIA UNIFLORA 399 1, 8, 10, 11). Although no organs morphologically intermediate in char- acter between staminodia and stamens were observed, the study of a wide range of flowers might reveal such transitions. Whether the larger num- ber of stamens (10-15) cited in Balfour and Smith’s description repre- sents a condition peculiar to the material from Yunnan or whether they failed to discriminate between staminodia and stamens must remain open questions for the present. | Fic. 1. Organography of a mature flower with five tepals, numerous stam- inodia, six stamens, and seven carpels, X 14. (Drawn from a cleared specimen, Fu 12296, by Mrs. Emily R. Reid.) The apocarpous gynoecium of the flower of Kingdonia is composed of 5—8 spirally arranged carpels (Fics. 1, 2, 18). Following pollination, the tepals, staminodia, and stamens fall from the receptacle, and the stylar portion of each of the enlarging carpels gradually becomes bent or re- curved over the dorsal edge of the ovary (Fics. 2, 21-24). This beaked character of the developing carpel and fruit is highly distinctive of King- donia and was noted in the brief descriptions of Balfour and Smith (5), Diels (7), and Wang (15). 400 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII COMPARATIVE MORPHOLOGY OF FLORAL ORGANS One of the interesting results of the present study was the discovery of the general similarity in the nodal anatomy of all the appendages of the flower. The vascular system of most of the receptacle is a typical eustele in which, as the result of interfascicular cambial activity, a single well-defined gap is associated with the divergence of the vascular supply of each of the tepals, staminodia, stamens, and lower carpels. In con- trast, the vasculature of the upper portion of the receptacle consists en- tirely of the traces of the inner carpels, and no “residual” xylem or phloem strands appear to exist (Fic. 18). ae 2. Form and spiral arrangement of a series of eight carpels of Kingdonia Ho, s.n.) at various stages of enlargement, * 8. Except for a single persistent S inoaiy m (at lower right) all other floral organs have fallen from the prom- inent ener Note dorsally reflexed styles of three largest carpels at the left. The following description and interpretation of the form, general struc- ture, and venation of the floral organs of Kingdonia is based largely upon the examination of mature flowers. Although much care was devoted to restoring the structure of flower buds, the results in some respects were quite disappointing. As a consequence, many important aspects of floral 1961 | FOSTER, KINGDONIA UNIFLORA 401 morphology, such as the early differentiation of the anther and the struc- ture and vasculature of the ovule, cannot be described until adequate fixed material of all stages in floral development becomes available. Tepals. A comparison of the cleared and stained tepals revealed in- teresting variations in both form and venation (Fics. 3-7). Only a single example of a lobed tepal was observed in the present study (Fic. 6). The lamina of this remarkable appendage terminates in two short lobes sep- arated by a deep and clearly defined sinus. Whether or not this form of tepal should be regarded as a rare ‘“‘anomaly” can, of course, only be de- termined by a survey of many flowers. It is entirely possible that three- or even five-lobed tepals may occasionally develop. Most commonly, and in agreement with Balfour and Smith’s description of the “sepals,” the tepal of Kingdonia is a simple foliar structure consisting of a short petiole and an elliptical-oval, prominently acuminate lamina (Fics. 3-5). Before describing the venation pattern of the tepals, it is essential to comment briefly on the structure of the tepal traces near their point of divergence from the eustele of the receptacle. A study of thin serial transections of one of the flowers showed that each trace consists of two more or less discrete strands of primary xylem which appear to be inde- pendently connected with the receptacle bundles bordering the single gap. However, the limitations of the material prevented a fully satis- factory reconstruction of the exact nature of the subnodal connections and origin of the tepal traces. The basal portion of each tepal is traversed by the unbranched exten- sion of its double trace, which, in cleared specimens, usually consists of two separate but closely spaced strands of xylem. In the lower region of the tepal lamina, this double bundle gives rise to a midvein and two lateral veins, one or both of which may branch dichotomously (Fics. 4a—6). No evidence of anastomoses or commissural veinlets was found, and hence the tepal venation in Kingdonia is simple and “open” in type. However, the exact mode of origin of the midvein and the two main lateral veins from the double bundle fluctuates even among the tepals of one perianth and may be conveniently described under the two following Tyrer 1. In this very common type, the midvein and one of the lateral veins of the tepal lamina represent respectively the central and outer branches of one of the main xylem strands of the double trace (Fics. 3, , 6). However, the level at which divergence and dichotomy of the main strands of the trace occur varies considerably. In many tepals, the single dichotomy which yields the midvein is near or at the level of divergence of both of the lateral veins. But Fic. 3 shows that the dichotomy may take place well above the point of divergence of the other main lateral vein. Type 2. The distinctive feature of this less common type is the dichot- omous branching, at approximately the same level, of each of the two xvlem strands of the double bundle and the union of the central strands 402 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII to form the midvein of the tepal (Fic. 4a—b). The two outer strands re- sulting from these dichotomies constitute the main diverging lateral veins of the lamina. A detailed study, with the aid of dark field illumination, showed that the two central strands which join to form the midvein are relatively slender and often unequal in size (Fics. 4b, 7). As Fics. 3—6 clearly indicate, there is no apparent correlation between these two types of midvein origin and the degree of branching of the main lateral veins. Type 1 may be associated with tepals in which both lateral veins remain unbranched and terminate blindly in the upper half of the lamina, or one or both of the laterals may dichotomize. In the latter case, the inner branch very commonly is longer and more highly developed than its sister branch. A striking example of this is shown in the bilobate tepal in which the lateral lobe is vascularized by the stronger inner branch of one of the lateral veins (Fic. 6). A few comments are appropriate regarding the interesting variations in the amount of xylem differentiated at various levels in the midvein. Near its point of origin, the midvein is relatively slender and comparable in this respect to the main lateral veins. Quite commonly, however, the upper portion of the midvein is conspicuously thicker as the result of the increase in number of tracheary elements. At its point of termination in the apex of the tepal, the midvein is usually dilated, suggesting the possi- bility that it is associated with a hydathode. Staminodia and Stamens. As mentioned earlier, the outer mem- bers of the androecium are clearly defined staminodia (Fic. 1). Each staminodium consists of an elongated filamentous stalk which terminates abruptly in an apically dilated “knob.” The vasculature is represented by a single unbranched vein which extends vertically through the stalk and then curves adaxially into the dilated end of the staminodium (Fic. 8). Although serial transections revealed no traces of rudimentary or abortive microsporangia, the dilated portion of the staminodium is quite com- parable in general form and venation with the corresponding sterile region of the anther (compare Fics. 8 and 11). On the other hand, a definitive character of the staminodial apex is the presence on its adaxial side of a median groove or furrow. As seen in transectional view, the epidermal cells which line the groove occur directly over a tissue comprised of cells which are much smaller and more deeply stained than the bulk of the parenchyma cells of the staminodium (Fic. 9). On the basis of this his- tological evidence, the groove may be glandular or “nectariferous” and may secrete substances which attract insect visitors to the flowers. Each of the stamens, like the staminodia, is vascularized by a single unbranched bundle which diverges at the juncture of anther and filament into the broad area of sterile tissue separating the two pairs of micro- sporangia (Fics. 10, 11). Although the sporangia are protuberant and superficially may appear “terminal,” their position with reference to the anther as a whole and to the plane of dehiscence is more accurately de- scribed as “‘latrorse-extrorse.” This is shown by Fics. 10, 11 which re- 1961] FOSTER, KINGDONIA UNIFLORA 403 spectively depict abaxial and adaxial views of mature, cleared stamens, and illustrate the oblique, more or less abaxial (i.e., extrorse) line of de- hiscence between the members of each pair of sporangia. The tissues of the young stamens of the flower buds which were sectioned were so badly collapsed and distorted that the following description and interpretation of the structure of the microsporangia is necessarily based on the study of mature, dehiscent anthers. FIGURE 12 represents a tran- section of an anther at the time of dehiscence and shows the remains of the partitions which separated the members of each pair of sporangia. The confluent “pollen sacs” contain “normal” pollen grains (see also Fics. 13-16), small tetrads, and apparently many examples of shrunken abortive cells. At this late stage in maturation it is impossible to describe with any accuracy the cellular structure of the sporangium walls. The dark “lining” of the pollen sacs may correspond to the remains of the tapetum and middle layers of the wall but there is no certainty on this point. In contrast, the endothecium or “fibrous layer” is well preserved and consists of two sharply delimited semicircular layers of cells with bandlike thickenings which jacket the protuberant portions of the confluent sporangia. These strips of endothecium are not continuous across the sterile portion of the anther, nor have cells with similar structure been observed bordering the internal edges of the sporangial cavities. The taxonomic and phylogenetic interest of information on pollen mor- phology in Kingdonia is emphasized by the present classification of this genus in the Ranunculaceae, a family characterized, according to Bailey and Nast (3), by tricolpate or derived forms of pollen. Because of my extremely limited material, a somewhat “unconventional” technique was employed in securing permanent whole mounts of pollen grains. Follow- ing the usual clearing in NaOH and chloral hydrate, the stamens were dehydrated, stained in dilute safranin and the mature anthers “squashed” directly in the ‘“Piccolyte” before adding the cover-glass. Selected ex- amples of pollen processed in this way are shown in Fics. 13-16. A comparison of polar and side views clearly indicates that the pollen grain is provided with three longitudinal furrows or colpi (Fics. 13, 15, 16). Although it has not been possible to study germinating pollen, there is some suggestion that a definable pore or aperture is located in each of the colpi. In the grain shown in median optical view in Fic. 13 the delicate “intine” appears to have evaginated as a short tubular process from the center of each of the furrows. On this basis it would appear justifiable to classify the pollen of Kingdonia as tricolporate in type. The exine of mature grains is moderately thick and when seen in surface view exhibits a complex foveate-reticulate structural pattern (Fic. 14 Carpels. Although the general form and uniovulate nature of the carpels of Kingdonia have been described by Balfour and Smith (5) and Diels (7), their accounts are very brief and include no information on early ontogeny or vasculature of these organs. On the basis of the present investigation it is believed that a relatively full description may now be 404 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII given of some of the early and critical stages of development and vascu- larization of the carpel. The most serious gaps in my description and interpretation concern the ontogeny and vasculature of the single pendu- lous ovule developed in each carpel (Fic. 20). Despite the greatest pos- sible care, the tiSsues of the ovule appeared collapsed and disorganized at all stages, and the problem can only be solved by the use of adequately fixed and preserved material (see also Fic. 19c). Before presenting the details of carpel ontogeny it is necessary to de- scribe briefly the general pattern of growth which leads to the adult form of the carpels. At anthesis, the carpel consists of a very short basal stipe, a laterally flattened ovary, and an erect or slightly recurved style with a terminal papillate stigma (Fics. 1, 21). When the carpels are approximately 2—3 mm. in total length, they contain a single ovule at- tached to the ventral side of the carpel wall (Fig. 20). As growth con- tinues, the basal stipe elongates, the ovarian region increases conspicu- ously in length and width and the style becomes reflexed over the upper dorsal edge of the carpel (Fics. 21-24). According to Wang (15), the mature fruit of Kingdonia is a one-seeded achene and exclusive of the recurved “beak” is 8 mm. long and 2.2 mm. wide. A carpel of approxi- mately similar dimensions is shown in Fic. 24. A transection of the central region of a developing flower bud showing three young carpel primordia is depicted in Fic. 17. At this early stage, the primordia are 200-275 y in height and throughout most of their length are prominently conduplicate structures. However, because of their spiral arrangement on the receptacle these primordia are not cut at the same level, and the lower primordium in Fic. 17 is represented by a section through the short basal stipe. Comparison of serial sections indicates that the post-genital fusion of the carpel margins to form the closed locule occurs in an acropetal direction. The carpel primordium at the upper right in Fic. 17 shows an early stage in the fusion of the margins, while the primordium at the left has been sectioned at a slightly higher level and exhibits an open conduplicate structure. In both of these carpels a single ovule primordium has arisen distal to the margin of the carpel. Thus the placentation in Kingdonia is obviously submarginal in type. The continued differentiation and enlargement of the carpel is accom- panied by the progressive acropetal fusion of the carpel margins and the development of a simple but highly distinctive type of vasculature. A relatively advanced stage in the closure of the carpel is shown in Fics. 19a—e which represent transections cut at successively higher levels through a carpel approximately 1 mm. in height. Although complete union be- tween the margins and outer adaxial surfaces of the carpel wings has produced a closed locule in the lower portion of the ovary (Fic. 19b, c), the upper third of the carpel, including the style is still unsealed and con- duplicate in organization (Fic. 19e). Limitations of material prevented a study of the later and final phases of closure of the carpel, and serial transections of a carpel 6.75 mm. in height showed no evidence, except in the lower adaxial portion of the style, of conduplicate structure. 1961 | FOSTER, KINGDONIA UNIFLORA 405 Each of the spirally arranged carpels of Kingdonia is vascularized by a single trace which, in young carpels, consists of a strand of xylem flanked and partially enclosed by two well-defined strands of phloem (Fic. 18). Soon after its extension into the basal region of a carpel, the trace divides tangentially into a dorsal (abaxial) and a ventral (adaxial) bundle. This aspect of carpellary vasculature is shown in the bases of three of the outer carpels in Fic. 18 and is further illustrated in the pho- tographs of the cleared carpels represented by Fics. 22-24. The struc- ture and course of the dorsal and ventral bundles in a young, apically conduplicate carpel are shown in the transectional views depicted by Fic. 19a—e. The relatively weak development of xylem in the ventral bundle is evident even in sections of the stipe (Fic. 19a). At successively higher levels the ventral bundle appears to consist wholly of phloem, and the strand terminates its development near the point where the carpellary margins are still open and conduplicate (Fic. 19d). In contrast, the dor- sal bundle is more robust, and the existence of both phloem and xylem is evident at high levels in the carpel (Fic. 19e). During the further enlargement of the carpel, the dorsal bundle continues its more rapid development and ultimately extends without branching into the lower region of the style. Ficures 21-24 clearly suggest that xylem develop- ment in the dorsal bundle is precocious, acropetal, and continuous, in contrast to the discontinuous and bidirectional pattern of early xylem formation within the ventral bundle. During the late phases of closure of the upper part of the carpel, the ventral bundle dichotomizes into two slender branches which terminate blindly near the top of the ovary but do not extend into the base of the style (Fic. 24). The point at which the bifurcation of the ventral bundle occurs appears to be below or in the immediate vicinity of the attach- ment of the ovule to the placenta. The ventral bundle shown in Fic. 19b, for example, although devoid of xylem, appears somewhat “double” in structure. Transections of carpels 6-7 mm. long show that each of the branches of the ventral bundle consists of well-defined phloem and xylem. Unfortunately, however, the ovule in each of these carpels appears abortive and collapsed, and there was no convincing evidence of the derivation of an ovule trace from either of the ventral strands. A few observations were made of the structure and vasculature of appar- ently mature fruits. No evidence of amy venation was seen in the carpel “walls” and it seems clear that the type of vasculature represented in Fic. 24 is typical of both the carpel and the mature achene of Kingdonia. During the development of the carpel into a fruit, the dorsal and ventral veins become more or less completely jacketed by sclerenchyma, and the surface cells which line the locule develop remarkable reticulate bands of secondary wall thickening. The prominent beak of the achene is like- wise composed largely of sclerenchyma, the initiation of which occurs in the upper region of the styles of comparatively young carpels and pro- ceeds basipetally into the upper portion of the ovary. 406 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII DISCUSSION In the light of current phylogenetic interpretation of floral organs in both woody and herbaceous ranalian plants, the flower of Kingdonia appears to show a combination of primitive and advanced levels of specialization. For example, the marked fluctuation in the number of tepals, staminodia, stamens, and carpels, their spiral arrangement on the receptacle, and the absence of fusion between any of these organs point to a rather primitive level of floral organization. On the other hand, the form and vasculature of the stamens and carpels seem decidedly advanced as compared with the foliaceous reproductive structures in the flowers of such woody genera as Degeneria, Austrobaileya, and Himantandra. The following discussion attempts to summarize the anatomical evidence se- cured by the present study and to evaluate it with particular reference to (a) the phylogenetic significance of the remarkable dichotomously veined leaf of Kingdonia and (b) the problem of the systematic affinities of this genus. The tepals, despite their small size and typically simple form, are the only floral organs in which the vasculature offers points for direct com- parison with the foliage leaf (Fics. 3-7). The nodal anatomy of the tepal is unilacunar, as in the leaf, and the double trace seems to repre- sent merely a simplification of the system of four traces which vascu- larizes the foliage leaf. Indirect evidence in support of this interpreta- tion is provided by the cataphylls of Kingdonia. These scale leaves occa- sionally develop a rudimentary lamina, and in such cases they are vascu- larized by two well-defined traces. Further indication of the fundamental resemblance between the vasculature of tepal and leaf is shown by one of the modes of origin of the tepal midvein. As illustrated in Fics. 4 and 7, this unbranched vein may arise from the union of the two central branches formed by the dichotomy of each of the tepal traces. A similar transition from an even to an odd number of main veins occurs in the base of the median lamina segment of the foliage leaf but results in this case in the formation of three systems of dichotomizing veins (cf. Foster and Arnott 11, Figs. 5, 7). In connection with these comparisons it is in- teresting to note the marked resemblance between tepal vasculature in Kingdonia and certain forms of cotyledonary venation in the angiosperms recently discussed by Bailey (2). The distinctly asymmetrical derivation of the tepal midvein shown in Fic. 3 is matched in certain types of cotyledons which develop an odd number of veins and the venation pat- tern represented in Fic. 4a, b corresponds very closely to another wide- spread type of cotyledonary vasculature (cf. Bailey 2, Figs. 2 E, F). Bailey’s (2) comprehensive surveys of the nodal anatomy of cotyledons, foliage leaves, and floral organs provide excellent evidence that the two- trace unilacunar node represents the primitive condition in angiosperms. On the basis of this interpretation, the probable nature of the complete vasculature of the primitive leaf type in angiosperms requires careful consideration. Bailey (2) has advanced the suggestion that the “mid-vein 1961] FOSTER, KINGDONIA UNIFLORA 407 of angiospermic appendages may have evolved by the approximation and fusion of two independent systems of vasculature.” This idea implies that the venation of primitive leaves was not “pinnate” in the strict sense of the term but may have resembled certain “pseudo-palmate” or transi- tional forms of cotyledonary vasculature (Bailey 2, Fig. 24, C, G). In this connection, is it not possible that the leaf of Kingdonia — notable for its unilacunar node, its four leaf traces and its palmate system of open dichotomous venation — may typify one of the ancient forms of foliar vasculature in the angiosperms? In my opinion, the occurrence of a two- trace and open dichotomous type of vasculature in the tepals of this genus lends additional weight to this theory. Although it may be argued that the persistence of a primitive leaf type is improbable in an herbaceous plant, the family Nymphaeaceae appears to have retained primitive forms of stamens (14), and vessel members with scalariform perforations occur in the xylem of Paeonia and Hydrastis (8). From these examples there seems to be no a priori reason to assume that the dichotomous vasculature of Kingdonia must be “specialized” or the result of “reversion” merely because this genus is herbaceous, rather than woody, in ha it. In contrast to the double-trace pattern of vasculature of tepals, each of the stamens of Kingdonia is vascularized by a single unbranched bundle (Fics. 10, 11). This fact, coupled with the clear demarcation of anther and filament and the latrorse-extrorse dehiscence of the paired sporangia, indicates a comparatively advanced type of stamen. The morphological and phylogenetic significance of the staminodia, which are likewise uni- fasciculate appendages, is problematical. However, their inconstant num- ber and their position below the stamens suggest an evolutionary origin from fertile microsporophylls. Like the stamen, the carpel of Kingdonia also seems relatively special- ized in both form and vasculature, especially when compared with the cialization of the stipitate carpel of Kingdonia is shown by the clear de- marcation between ovary, style and stigma, by the development of a single ovule and by the simple and sparse pattern of vasculature (Fics. 20, 24). Although the conduplicate and open structure of the carpel primordia in Kingdonia suggests the persistence of a primitive type of ontogeny, the closure of the carpel occurs relatively early and results in a sealed ovary without a definable ventral suture. The mature vasculature consists of an unbranched dorsal and a forked ventral vein, both derived from the single trace which enters the stipe region of the carpel. As mentioned earlier in this paper, it proved impossible because of technical difficulties, to determine the nature of the “ovule trace” in Kingdonia. The bifurcation of the ventral strand occurs near the point of attachment of the ovule to the inner ventral surface of the carpel and 408 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII it is probable that the vasculature of the ovule is derived from one or possibly both of these traces. In this connection, the venation of the carpels of certain members of the Ranunculaceae offers certain points for comparison with the situation in Kingdonia. According to Chute (6), the carpel of various species of Anemone, Clematis, and Hepatica is vascular- ized by a single trace which divides into a dorsal and ventral vein. Near the top of the ovarian cavity, the ‘fused ventrals’’ divide into three strands; the median represents the ovule trace, and the two laterals con- tinue into the style. In Kingdonia, however, the two branches of the ven- tral bundle terminate near the base of the style which is vascularized ex- clusively by the unbranched dorsal bundle (Fic. 24). In conclusion, it seems appropriate to discuss briefly the bearing of the evidence from leaf and flower on the systematic affinities of Kingdonia. As I have recently pointed out (10), the original classification of King- donia in the Ranunculaceae by Balfour and Smith (5) was accepted with- out question by Diels (7) and Janchen (13), and the same taxonomic assignment of this genus was later adopted by Wang (15) and Hutchinson 2). Admittedly the herbaceous habit, polymerous flowers and one- seeded achenes of Kingdonia are duplicated, at least in a superficial way, in such ranunculaceous genera as Thalictrum and Anemone. But in my opinion, the totality of morphological and anatomical evidence derived from my studies on the leaf and flower strongly indicate that Kingdonia is a “relict” genus, without obvious affinity to any ranalian family, in- cluding the Ranunculaceae. Prominent among the “unique” and isolated features of Kingdonia are the dichotomously veined foliage leaf and the two-trace vascular system of the tepals. Comparable types of foliar vas- culature do not occur in any of the genera of Ranunculaceae which I have studied or which have been described anatomically in the literature (10). A similar but less decisive ‘‘remoteness” is shown by the pollen grains of Kingdonia (Fics. 13-16). These structures appear to be tricolporate in type and when more thoroughly studied and compared with the pollen of other ranalian plants may support the removal of Kingdonia from the Ranunculaceae. Many important morphological aspects of Kingdonia remain for future investigation. These include the determination of the chromosome num- ber, the morphology of the tracheary elements, the study of the seedling (with particular reference to cotyledonary vasculature), and information on the development and structure of the embryo sac, endosperm, and embryo. When these gaps in our present knowledge have been bridged, I suspect that Kingdonia will ultimately be treated as the representative of an independent and “relict” family within the Ranales. LITERATURE CITED 1. BatLey, I. W. Contributions to plant anatomy. Chronica Botanica Co., Waltham, Mass. 1954 ———.. Nodal anatomy in retrospect. Jour. Arnold Arb. 37: 269-287. 1956. 1961 | FOSTER, KINGDONIA UNIFLORA 409 ——_ and C. G. Nast. The comparative morphology of the Winteraceae. I. Pollen snd ae Jour. Arnold Arb. 24: 340-346. 1943. _G. L. Swamy. The conduplicate carpel of dicotyledons and its initial ae of specialization. Am. Jour. Bot. 38: 373-379. 195 5. Batrour, I. B., and W. W. SmitH. Kingdonia uniflora. In: Didenoses spe- cierum novarum LI-CII (species Chinenses). Notes Royal Bot. Gard. Edinb. 8: 191-192. 1914. 6. Cuute, H. M. The morphology and anatomy of the achene. Am. Jour. Bot. 17: 703-723. 1930. ELs. L. Circaeaster eine hochgradig reduzierte Ranunculacee. Beih. Bot. Centralbl. 49 (Erg. Bd.) : 55-60. 1932. 8. Eames, A. J. Morphology of the angiosperms. McGraw-Hill Book Gs aap ha New York. 1961. 9. Foster, A. S. The phylogenetic significance of dichotomous venation in angiosperms. Proc. IX Int. Bot. Congr. 2: 119-120. 1959. —. The morphological and taxonomic significance of dichotomous vena- ation in Aingdonia uniflora Balfour f. et W. W. Smith. Notes Royal Bot. Gard. eee 23: 1-12 ~I 11. ——— and H. J. ARNOTT. Morphology and dichotomous vasculature of the leaf of Kingdonia uniflora. Am. Jour. Bot. 47: 684-698. 1 12. Hutcurnson, J. The families of flowering plants. Vol. 1 Dicotyledons. Ed. 2. Oxford. 1959. . JANCHEN, E. Die systematische Gliederung der Ranunculaceen und Berberi- daceen. Denkschr. Osterr, Akad. Wiss. Math.-Naturwiss. Klasse 108: 1-82. 1948. 14. Mosetey, M. F., Jr. Morphological studies in the Nymphaeaceae. I. The nature of the stamens. Phytomorphology 8: 1-29. 1958. 5. Wanc, W. T. Notulae de Reais sinensibus. Acta Phytotax. Sinica slay: 361-391. 1957. — DEPARTMENT OF BOTANY UNIVERSITY OF CALIFORNIA, BERKELEY EXPLANATION OF PLATES PLATE I Fics. 3-7. FoRM AND VENATION PATTERNS OF TEPALS OF Kingdonia | Fu a ee 3. Origin of midvein and lateral vein from dichotomy of one strand of double |X 7. 4a, Origin of midvein by union of central branches of both ek strands, s< 6. 4b, Enlargement of region of vein-union repre- sented in 4a, X 30. 5. Tepal showing prominent dichotomous oo of e each lateral vein, X 6. 6, Form and vasculature of a bilobed tepal. further example of dichotomy of both strands of pan trace and union of the two short central bundles to form midvein, X 30 — note close approximation of central bundles as compared with 4b. For further SS ications see text PLATE: if nee 8-12. STAMENS AND STAMINODIA OF Kingdonia. 8, Cle ared staminodium [Ho, sm.| showing general form and vasculature, 3 bo 32.5. 9, Transection of eee region of staminodium | fu 12296] showing adaxial groove and single vein, 410 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII x 135. 10, 11, Cleared stamens [from Ho, s.n.], showing vasculature and the orientation and plane of dehiscence of the microsporangia, X 32.5; 10, abaxial view; 11, adaxial view. 12, Transection of dehiscent anther, * 135 — note pol- len grains and clearly delimited endothecium PLATE III Fics. 13-17. POLLEN AND CARPELS OF Kingdonia. 13-16, Mature pollen grains [from Fu 12296], X 1480. 13, Polar view, showing tricolporate structure. 14, Same grain at a higher focal plane, illustrating structural pattern of exine. 15, 16, Side views of pollen grains with longitudinal colpi. 17, Transection of central region of flower bud [Liu & Chun 1307] to illustrate young stages of carpel development, X 295. Note conduplicate nature of carpel primordium at left and fused margins of young carpel at upper right. PLATE IV Fics. 18-20. VASCULATURE AND DEVELOPMENT OF CARPELS. 18, Transection of flower [from Fu 12296] with seven carpels, the outer five of which are at- tached to the receptacle, < 138. Note particularly that each of the traces of the strands. 19a-e, Transections, at successively higher levels, of one of the carpels shown in Fig. 18, X 138. 20, Longitudinal section of a carpel about 2 mm. high (exclusive of reflexed style) showing the attachment of the single pendulous ovule to the ventral wall [ Ho, s.n.], X 22. PLATE V Fics. 21-24. CLEARED CARPELS, selected from gynoecium [from Ho, s.n.] shown in Fig. 2 and photographed with dark-field illumination. 21, Carpel, ae a 2 mm. in height, showing ovary, style, and papillate stigma, « 32.5 ote precocious and acropetal development of xylem in dorsal bundle. 22, Older stage than preceding figure, showing dorsal bundle extending into base of reflexed style and an early phase of xylem development in ventral bundle, x 19. 23, Still later stage illustrating apparent bidirectional and interrupted pattern of xylem differentiation in ventral bundle, < 19. 24, Carpel approximately 7.5 mm. high (exclusive of reflexed style) showing complete vasculature, * 19 —note dichotomy of ventral bundle near top of ova Jour. ARNotp Ars. VoL. XLII ; PLaTE I i 4a 4b ( FOSTER, KINGDONIA UNIFLORA Jour. ARNOLD Ars. VoL. XLII PLaTE II FOSTER, KINGDONIA UNIFLORA Jour. ARNoLD Ars. VoL, XLII PiaTeE III FOSTER, KINGDONIA UNIFLORA Jour. ARNOLD Ars. VoL. XLII \ tae \ | ; te WY Wy : +g ef @, i 7 & - 4 ¢ “8 : ne (Vi) é 4 gue a —— FOSTER, KINGDONIA UNIFLORA PLATE Jour. ARNOLD Ars. VoL. XLII PLATE V FOSTER, KINGDONIA UNIFLORA 416 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII CONIFEROUS POLLEN TYPES OF THE SOUTHERN HEMISPHERE, I. ABERRATION IN ACMOPYLE AND PODOCARPUS DACRY DIOIDES Lucy M. CRANWELL IN COMPARING THE POLLEN GRAINS of the Podocarpaceae with those of the Pinaceae, Mme. van Campo (1950) has recognized their common and very ancient source. In this view she has followed Prof. R. Florin (1936), who has shown how very closely aberrant grains of members of both families may approach ones normal to the Cordaitales. It is his opinion that the morphological affinity with the extinct group is strongest in the grains of certain genera of the Pinaceae. Van Campo considers that the Podocarpaceae may prove to be older, however, and points to the di- vergent groups persisting today. Further, she admits difficulty in deciding to which family certain types belong. Sir Albert C. Seward (1914) ran into the same difficulty in identifying the first gymnospermous grains discovered in Antarctic deposits. He set up and defended his organ-genus Pityosporites in the belief that it rep- resented abietineous grains, but soon decided that it was most probably based on podocarp material. Unfortunately, the name must be upheld as a catch-all for older bisaccate grains, of both hemispheres, which have characters common to the two families. However, as pollen studies pro- ceed, closer identification of fossil with living types is made possible, so that a dwindling remnant of southern hemisphere “incertae” is likely to lodge, or remain, in Pityosporites. Thus, despite evidence for relationship in the two major families, diag- nostic differences do occur and could be exploited much more than at present for ape arer distributional, and climatological studies. The work of Dr. Isabel ‘ookson, Dr. R. A. Couper, and Mr. Basil E. Balme, to mention only some who have worked in both hemispheres, is outstand- ing in this respect Particular weakness still lies in our lack of detailed knowledge of Pinus pollen, as contrasted with that of Podocarpus, or of Cedrus, as opposed to some of the Dacrydia. Records of one or the other will crop up some- what erratically in fossil work from either hemisphere until the position is clarified. For instance, a report on Lower Miocene deposits of Silesia by Prof. Stefan Macko (1957) presses identity of his conifer types with living podocarps such as Podocarpus dacrydioides A. Rich. and Dacry- dium cupressinum Soland. ex Forst. f. While I am willing to admit that genera now represented in New Zealand may also have occurred in Eurasia, I feel sure that his photomicrographs and descriptions do not clinch the matter for two reasons: the reticulate sculpture rules out a Dacrydium, and the presence of Crossotheca spores in the same deposit 1961 | CRANWELL, CONIFEROUS POLLEN TYPES, I 417 suggests redeposition from older beds for the “Podocarpus,” if closer affinity with known three-sac types cannot be proved. Neither set of microfossils exemplified by Macko comes any closer to Miocene anteced- ents of these species as we know them from New Zealand deposits. I mention these examples in order to stress the difficulties involved in comparing fossil podocarps of the two hemispheres, rather than to detract from Prof. Macko’s main theme, which is the very real affinity between the Tethyan (or Poltavian) part of his microflora and that of Indo- Malayan vegetation of today. Few still believe that Pinus and its close allies ever reached far into the Southern Hemisphere: Pinus today stops short in the West Indies and Guatemala, on one side of the world, and in the Sunda Islands, less than six degrees south of the Equator, on the other. On the other hand, some podocarps are found far above the Equator, mainly in the wet highlands of Central America, of Nepal, the Philippines, China, the Ryukuyu Islands, and Japan. We are indebted to Dr. R. Florin, particularly, for purging the records, living and fossil. For South America, Antarctica, and New Zealand he has shown, for instance, that all the “Sequoia” records are false, and he has usually been able to place the macrofossils on which they were based in Podocar pus or Acmopyle. Microfossil records usually lag behind those for macrofossils, and this is true for Acmopyle, one of the rarest of living conifers. Two living spe- cies, 4. pancheri Pilg. and the very closely allied A. alba Buchholz, are known from the high serpentine areas of eastern New Caledonia (see Chevalier, 1957), and a third, A. sakniana Buchholz & N. E. Gray, from Fiji. Florin’s ancient species, based on leaves and twigs, are A. antarctica from Seymour Island, Graham Land (Oligocene or Lower Miocene) * and A. engelhardtii (Eocene) from Rio Pichileufu in Argentina. An excellent note on the habitat and rich assemblage of conifers asso- ciated with Acmopyle is given for New Caledonia by Compton (1922): There, at 3000 ft. above sea level, the generic mixture and the ecological conditions must be very similar to those at a slightly lower altitude on Mount Te Moehau, an isolated “island,” untouched by glaciation, on the Coromandel Peninsula about 40 miles east of Auckland, New Zealand. In both areas the relict forest seems to survive, as Compton puts it, through “stubborn resistance,” as it is out of equilibrium with its environment. ABERRATION IN CONIFER POLLEN GRAINS Various authors have dealt with normal pollen grains in the Podo- carpaceae, but few have described the range of aberrance in this im- portant group, or have tried to assess its importance. I shall, therefore, confine myself in this note to its occurrence in Acmopyle pancheri and. * The age of the Seymour Island sediments was given by Florin (1940) and others as Eocene: recent work by the Falkland Islands Dependencies Survey indicates, how- ever, that there are no Eocene sediments left in the region. 418 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Podocarpus dacrydioides and shall mention the normal form only where comparison is necessary. Hybridism and unknown factors lead to many deviations in pollen grain morphology, disturbing the familiar face, or physiognomy, of a grain. It is as though a cheek were turned in the Coniferales, in par- ticular, when grains gain or lose bladders, reveal ancestral markings, or remain fused back-to-back with four bladders instead of two. The change of profile, so to speak, now offers new clues to cell behavior and, per- haps, to evolutionary trends. In some, the differences seem to represent a recapitulation difficult to trace except in very young or aborted grains. Among irregularities noted for these southern genera are trisaccate grains in the normally disaccate Acmopyle and grains with more or less confluent sacs (bladders) or with triradiate or slashlike scars on the cap (the proximal side of the grain) in Podocarpus dacrydioides. Sac fusions or abortions may also occur in Dacrydium cupressinum and in the New Zealand species of Phyllocladus. In Acmopyle and Podocarpus dacrydioides, at least, hybridism does not seem to have been involved, as the species are geographically and genet- ically isolated. No hybrids have been reported from Acmopyle or from the Dacrycarpus section of Podocarpus. Further, the frequency of ir- regularity is quite low, about five per cent (of 300 grains studied; Cran- well, 1940) being noted in Podocarpus dacrydioides and under one per cent in Acmopyle (with 600 grains studied). Van Campo (1950), on the other hand, reports 100 per cent aberrance in certain hybrid Tsuga material she investigated. It is worth noting that few conifer hybrids have been accepted for the southern hemisphere, the late Dr. H. H. Allan (1961) giving only two groups in Flora of New Zealand. Both were between species with disaccate grains. Hair (1958) and Hair and Beuzenberg (1958a,b) have confirmed one of these groups in their wide surveys of chromosome numbers in the Podocarpaceae. Their material was a cross of Podocarpus hallii Kirk P. nivalis Hook. from the closely related “hallii-totara-nivalis” complex, and it yielded counts of 2n = 36, while the parents yielded 2n = 34 and 38 respectively. The authors stress the importance of the evidence for chromosome inconstancy discovered in many of the 52 species, and the one hybrid, they have studied. They state (1958b) that the Podocarpaceae, instead of displaying “the unvarying basic numbers of most families of slow-growing, long-lived” gymnosperms, shows “surprising variation . . . which points both to the past and to the future.” Hair (1958) adds that this “chromosomal polymorphism” is “a surprising discovery amongst the long-lived woody plants generally.” Their lists show that few species with high basic chromosome numbers have trisaccate grains either regularly or occasionally. Apart from Microcachrys and Pherosphaera this condition is strong only in the groups in which Hair and Beuzenberg have dem- vastrated a low basic number, e.g., 10 for age pancheri, Podocarpus dacrydioides and Dacrydium cupressinum (2n = Only a few of the trisaccate species cross the are one running north 1961 | CRANWELL, CONIFEROUS POLLEN TYPES, I 419 to the Philippines. All of these species belong to section Dacrycarpus, which has also a Chilean species described by Florin (1940) from fossil twigs. The discovery of more fossil material, as claimed by Macko (1957), would therefore represent a great extension of range for species with this pollen type. According to European reports, coniferous grains with more than two sacs are by no means common in northern deposits. TRISACCATE GRAINS IN ACMOPYLE The only illustrations of Acmopyle pollen grains to which I have had access are those of Erdtman and of Ueno. Dr. Erdtman illustrated the pollen of A. pancheri in 1957, promising to give the necessary text in his second volume on the gymnosperms, while Dr. Ueno gave an extremely small sketch for A. alba in 1960. Thanks to Dr. Erdtman, I was permitted to see the slide on which his sketches were based; later I received fresh pollen cones from Dr. David de Laubenfels of Syracuse University, New York (P 130, collected on Mount Mou, New Caledonia, Oct. 9, 1957), and it was in this second collection that I noted the aberrant forms, which comprised under one per cent with three sacs and a small number only that were dwarfed, doubled, or slightly irregular in size of the sacs. The accompanying plate illustrates a trisaccate grain, together with two typical disaccate grains, seen in distal and lateral views (Fics. 1-3). It should be noted that the three sacs are in perfect balance and condi- tion (Fic. 1), but the exine of the corpus (body of the grain) is very thick and is embayed between the sacs. As in normal grains, the puffed bladders are attached by narrow bases and lie wholly on the distal surface of the grain. Most of the irregular grains were smaller than the normal ones, but the overall measurement, with sacs included, was little smaller, being about 62. No grains were broken, but several (normal) had shed a sac very cleanly, possibly due to the rigors of acetolysis during the preparation of the sample. The tendency to shed seems stronger than in other podocarp pollen types: it may thus be significant that sacless grains of this type occur rather freely in Upper Cretaceous Antarctic sediments which I am studying at present. No bladders like those of Acmopyle have been rec- ognized, however, although many fragments have persisted. SOME FOSSIL OCCURRENCES OF TRISACCATE TYPES Occurrences in older floras will not be discussed in detail here, but it is worth noting a second trisaccate type also found by the author in the Antarctic samples. Deviations from the coniferous types suggest that it may have been shed by a pteridosperm. The trisaccate condition is certainly very old, though not as old, it seems, as some of the types with enveloping or cingulate sacs. In Podo- cartus dacrydioides, at least among the podocarps, a deviant takes the 420 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII form of this more ancient fused sac, comparable with those sometimes developed in Pinus sylvestris, according to reports by Eneroth (see Florin, 193 Such grains occurring in podocarps, pines, and their allies, even if rarely, seem to give good support 7 Florin’s views on the common ancestry of these groups in the Cordaitale Trisaccate grains occur rather anus among the Abietineae in both fossil and living material. They occurred also in extinct groups the af- finities of which are little understood. Couper (1958) has reported them, for instance, in frequencies of one in a thousand for the normally bisaccate Caytonanthus arberi (Thomas) Harris, a member of the Caytoniales which lived in the Mesozoic period in Britain. The nomenclature for these types is especially difficult: some are lumped in Trisaccites, or, if a comparatively recent origin is suspected, i an organ-genus suggesting affinity, e.g., Podosporites. Where there is a continuous record for a species, as for Podocarpus dacrydioides in New Zealand, the affinity can well be claimed, but even for this distinctive type the record does not yet go back beyond the Tertiary. TRIRADIATE SCARS IN PODOCARPUS DACRYDIOIDES The triradiate scar, the mark of contact with other grains of a de- veloping tetrad, occurring always on the proximal or cap side, is rarely noted in conifers, and has never been confirmed for any angiosperm, al- though Coprosma macrocarpa Cheesem., for one, has a polar marking which mimics a scar. Wodehouse (1935) has illustrated a scar for Abies nobilis Lindl., for instance, and has stated that it is a Pcie char: acter when found but it is generally very faint and difficult to see.’ Florin (1936) reports much the same thing for (¢ plug and illus- trates a grain with a small, compact scar Ueno (1957) has given the only reports of a scar in Podocarpus dacrydioides, as far as I know. His illustration I noted with admiration and a mild chagrin, as I had never observed one of these scars in many grains I have worked with, living or fossil: the only deviations I had seen were frillings and fusions of the sacs. I was fortunate, therefore, on October 25, 1958, to be able to collect young cones, whic took from 4 small tree, not long out of its juvenile stage, by the Waitakere Stream not far from Auckland. In this material I found an abundance of scars, as well as broad slash-like markings not mentioned by Dr. Ueno. In Podocarpus decrydioides, the scar lies in the middle of the body of the grain, sending ‘ts long, untapered arms out almost to the equator where each comes symmetrically and abruptly to a point. As will be seen (Fics. 6, 7), the arms are broad. both margins and surfaces being clearly delimited. By comparison with such ancient types as //linites Kosanke and Fuldaesporites Leschik (as illustrated in Potonié, 1958. Ta f. 7, 62 and Taj. 9, 91, respectively), or with modern fern types, such as the spores of Ad inline, this is a large and magnificent scar. It does not run into and fuse with the ridge encircling the equator, as Ueno reports for Tsuga. I 1961] CRANWELL, CONIFEROUS POLLEN TYPES, I 421 agree with him, however, in his suggestion that only young grains are likely to show the triradiate marking to advantage. IRREGULARLY SCARRED GRAINS The chief aberrations in this species, and the ones most likely to be of importance, are the slash-like strips running singly on in pairs across the proximal surface. Their “selvage” margins are wide and calloused (Fics. 9, 10) and thus suggest rigidity, rather than distensibility, of the wide and delicate membrane of the distal surface. This latter is usually spoken of as the “furrow,” although it is not homologous with the furrow of most of the angiosperms. Several fossil pollen genera are characterized by rather similar rifts (see Potonié, 1958, Taf. 7, 63, 64, and especially Balme and Hennelly, 1955). These are disaccate grains, the gashes tending as a rule to run straight and at right-angles to the side furrows on the other side of the grain. In Podo- carpus dacrydioides, likewise, the seemingly haphazard pattern of rifts may be determined by the geography of a triangular, rather than an elongate, furrow area. Ina “normal” young grain, the arms of the perfect triradiate scar will run toward the apex of each subtending air sac, as it would to the three angles formed on the amb of trilete fern spores. The disturbance of the abnormal grain, whatever its cause, may thus lead to gross thickenings and erratic trends in the trisaccate types. In “Lueckisporites” of Balme and Hennelly (1955) a straight gash occurs together with lesser parallel striation. Such grains are especially important in members of the Gondwana Flora (see also Pant & Nautiyal, 1960) but have not yet been reported from Gondwana deposits in New Zealand. Their association with delicate triradiate crests in some sacless microspores described by the Australian workers suggests that their early function was separate from that of the scar. It will be interesting to dis- cover for modern conifers what relationship the heavy gashes have to both the triradiate scar and the ancient striations. ACKNOWLEDGMENTS I offer my thanks to Dr. Gunnar Erdtman, of Stockholm, and to Dr. David de Laubenfels, of Syracuse University, New York, for sharing their Acmopyle material with me, and to my nieces, Miss Eleanor Ann Cranwell and Miss Norma L. Cranwell, of Henderson, New Zealand, for helping collect the young Podocarpus dacrydioides cones near their home in 1958. I also have had helpful correspondence concerning gymnosperm pollen morphology with Dr. Vishnu-Mittre, of the Birbal Sahni Institute of Palaeobotany, Lucknow, India, and with Dr, Jitsuro Ueno, of Osaka City University, Osaka, Japan. Finally, [am grateful to Dr. Elso S. Barghoorn, of Harvard University, for allowing me to browse in his pollen library dur- ing the summer of 1960. 422 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII LITERATURE CITED ALLAN, H. H. Flora of New Zealand. Vol. I. Indigenous Tracheophyta, Psilopsida, Lycopsida, Filicopsida, Gymnospermae, Dicotyledones. Govt. Printer, Wellington, New Zealand. 1961. BAUME, B. E., AND J. P. F. HENNELLY. Bisaccate sporomorphs from Australian Permian coals. Austral. Jour. Bot. 3: 89-98. 1955. CHEVALIER, L. Les coniféres actuellement connue en Nouvelle-Calédonie. Etudes Mélanésiennes II. 10-11: 105-118. 1957. Compton, R. H. A systematic account of the plants collected in New Caledonia and the Isle of Pines by Mr. R. H. Compton, M.A., in 1914, Part II. Gymnosperms. Jour. Linn. Soc. Bot. 45: 421-434, 1922. Couper, R. A. British Mesozoic microspores and pollen grains. A systematic and stratigraphic study. Palaeontographica 103B: 75-179. 1958. CRANWELL, L. M. Pollen grains of New Zealand Conifers. New Zealand Jour. Sci. & Tech. 22: 1B-17B. 1940. ERDTMAN, G. Pollen and spore morphology and plant taxonomy. Gymnospermae, Pteridophyta, Bryophyta (An introduction to palynology II). Almqvist & Wiksell, Stockholm; Ronald Press, New York. 1957. FLorin, R. On the structure of the pollen-grains in the Cordaitales. Sv. Bot. Tidskr. 30: 624-651. 1936. . The Tertiary fossil conifers of South Chile and their phytogeographical significance. Sv. Vet-akad. Handl. 19(2): 1-107. 1940. Harr, J. B. The chromosomes of the Podocarpaceae. Proc. X Int. Congr. Genetics 2. 1958. AND E. J. BEUZENBERG. Contributions to a chromosome atlas of the New Zealand flora—1. New Zealand Jour. Sci. 1: 617-628. 1958a. ———.. Chromosomal evolution in the Podocarpaceae. Nature 181: 1584-1586. 8b 1958b. LAKHANPAL, R. N., anp P. K. K. Narr, Some abnormal pollen grains of Picea smithiana Boiss. Jour. Indian Bot. Soc. 35: 426-429. 1956. Macko, S. Lower Miocene pollen flora from the valley of Klodnica near Gliwice (Upper Silesia). Prace Wroclawskiego Towarazystwa Naukowego B88: 1-313. 1957 Pant, D. D., anp D. D. Nautiyat. Some seeds and sporangia of Glossopteris flora from Raniganj Coalfield, India. Palaeontographica 101-107B: 41-64. 6) PotoniE, R. Synopsis der Gattungen der Sporae dispersae II. Teil: Sporites (Nachtrage), Saccites, Aletes, Praecolpates, Polyplicates, Monocolpates. Beih. Geol. Jahrb. 31: 1-114. 58. SEwarp, A. C. Antarctic fossil plants. Brit. Ant. Exped. 1910. Nat. Hist. Rep. (Geology) 1(13. 1914, Ueno, J. Relationships of genus Tsuga from pollen morphology. Jour. Inst. Polytech. Osaka City Univ. D8: 191-196. 1957. . Palynological notes of Podocarpaceae. Acta Phytotax. Geobot. 18: 198-207. 1960. Van Campo-Duptan, M. Recherches sur la Phylogénie des Abiétinées d’aprés leurs grains de pollen. Trav. Lab. Forest. Toulouse Tome II, Prem. Dect. Pollen. 4(1): 9-183. 1950. WobenousE, R. P. Pollen grains. McGraw-Hill, New York and London. 1935. GEOCHRONOLOGY LABORATORIES, UNIVERSITY OF ARIZONA, TUCSON 1961] CRANWELL, CONIFEROUS POLLEN TYPES, I 423 EXPLANATION OF PLATES PLATE I Fics. 1-5. Acmopyle and Podocarpus. Fics. 1-3. Photomicrographs of pollen grains of Acmopyle pancheri. 1, Trisaccate grain, 62x, showing the symmetric and widely spaced sacs, the lowest being turned inward slightly toward the germinal area or “furrow.” 2, Disaccate grain, 694, showing distal or ‘‘furrow” side of grain. 3, Disaccate grain, 69u, in lateral view showing the puffed sacs their open reticulation of the sacs [Material: De Laubenfels P130; pollen a antads p5 (acetolysed)]. Fic. 4. Photomicrograph of pollen grain of Podocarpus dacrydioides. Normal grain in slightly oblique view, the sacs crowd oe together above the rounded “furrow” area; 65x. [ Material : L. M. & _ A. Cranwell; pollen preparation S768 (acetolysed)]. Fic. 5. Very old trees Podocarpus dacrydioides (‘‘Kahikatea”) isolated after felling of forest around them; King Country, New Zealand. Photographs by author. PLATE II Fics. 6-11. Photomicrographs of pollen grains of Podocarpus dacrydioides. 6, Distal view of normal grain Sopra | aed mature), 65x, showing attach- ment of the wide bases of the sacs on t ge triangular ‘ ‘furrow’ area. Young grains, 46 », showing the ae eae ae scars on their cap (proximal) side — bladders have been trimmed away in fig. 7 leaving only body or corpus of grain. 9, 10, Aberrant grains with open wound-like scars. 11, Normal grain overlapping an aberrant one (10), both about 70 I PLATE Jour. ARNOLD Ars. VoL. XLII ONIFEROUS POLLEN Types, I LL, C CRANWE PLATE II Jour. ARNOLD Ars. VoL. XLII s. 1 CRANWELL, CONIFEROUS POLLEN TYPE 426 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII STUDIES IN THE THEACEAE, XXXIII VARIATION IN THE FRUIT OF TERNSTROEMIA KWANGTUNGENSIS CLARENCE E. KopuskI COMPLETE DIAGNOSES of taxa in the genus Ternstroemia, like those of any other genus, should include the descriptions of both flowers and fruit. Considering that the flowers of this genus are either dioecious, andro- dioecious, or occasionally hermaphroditic and that flowers seldom are col- lected at the same time as the fruit, one may see how inadequate a de- scription based on a single herbarium specimen must be. In the species discussed below the flowers are still unknown. The fruiting material avail- able to the original authors (two names are involved) was insufficient for a complete diagnosis of the fruit. However, it so happens that in Tern- stroemia kwangtungensis the leaf characters seemed outstanding enough at the time of its description to set it apart from all other Chinese mem- bers of the genus. The fruit is widely variable, and it is with this extreme and unusual variation, in contrast to the uniformly 2-loculate condition of most other Asiatic species, that the present paper is concerned. In 1918, E. D. Merrill described an interesting new species of Tern- stroemia from Kwangtung Province, China, as T. kwangtungensis. He obviously never realized at the time —or later — how distinct and un- usual the species actually would turn out to be. The type collection was made by Merrill himself, and, although it was a rather wretched specimen, Merrill with his keen eye for novelties recognized it as different from any Ternstroemia previously described from China. What attracted Merrill, I am sure, was the heavy, subrotund, shiny, coriaceous leaves and the broad, very thick petioles, with the overall appearance of a stiff fan or paddle. The fruit was immature, and as such he dismissed this character. Since then several workers have identified Kwangtung material in the Arnold Arboretum herbarium as belonging to this species, but, upon close observation, none of the specimens truly resembles Merrill's isotype which had been deposited at the Arnold Arboretum. Since the isotype available (Merrill 11013) was so sparse (one attached and one unattached immature fruit) I had hesitated to dissect the material until absolutely necessary. However, when a dissection was finally made of the unattached fruit I discovered that it was so very young in develop- ment that it was still in the advanced ovulate stage. There were five locules. However, of these, only three were fertile and contained respectively three, two, and two ovules. The inside surface at the apex of the locules where the ovules were attached was dark red, smooth, and shiny. In the two intervening spaces (or sterile locules) such was not the case, these being filled with spongy tissue. Two of the 1961] KOBUSKI, STUDIES IN THEACEAE, XXXIII 427 fertile locules were adjacent to each other and were separated by a single septum. On each side of these two adjacent fertile locules was a sterile locule with the third fertile locule between the two sterile locules. It is quite likely that in the maturation of the fruit the fully developed seeds may fill the fruit completely, thereby eliminating the sterile locules, for in this genus it is not unusual to find that the septa separating the locules have been broken down by the developing seeds. A second specimen which appeared to belong to this species, Tsang 25486, was found among our unidentified material. This specimen also had only two fruits. A single dissection of a fruit of this second specimen showed it to be 3-loculate with the three locules clearly taking up the entire fruit, with no intervening spaces. The seeds numbered seven and were arranged four, two, and one in the three locules. A short while ago, I was surprised to find that H. T. Chang (Zhongsan Daxue Xuebao, Ziran Kexue [Bull. Sun Yatsen Univ. Nat. Sci.] 2: 25. pl. 7. 1959) had described a new species, Ternstroemia subrotundifolia, from Kwangtung and had cited Tsang 25486, already mentioned above, as the type of his species. What startled me even more was that Chang men- tioned the fruit as 5-locular with a single seed in each locule. Since the Tsang collection was made under the sponsorship of the Arnold Arboretum and is not yet distributed, four duplicates of Tsang 25486 have been available for my study, thus affording an opportunity to make sev- eral dissections of the fruit without destroying the value of the specimens, which, after all, are isotypes of Chang’s species. My findings are recorded in the table which follows: MarTuvRE FRUIT OF Tsang 25486 No. oF LOCULES No. OF SEEDS SEEDS IN LOCULES 3 7 —2-1 3 6 2-2-2 3 5 2-2-1 3 5 2-2-1 4 7 1—2-—2-2 4 5 2-1-1-2 5 7 1—2-—1-2-1 In the seven dissections made from this single collection, the fruits were found to have either three, four, or five locules, with three the dominant number. The total number of seeds for each fruit varied very little, usu- ally five or seven with only one fruit showing six seeds, two to each of the three locules. Chang in his description (both Latin and Chinese) gives the fruit as 5-locular, with a single seed in each locule, an arrangement which I did not find in any of my observations. Since Chang was working with a duplicate of the same collection (Tsang 25486) with which I have worked, we can only assume that his finding was made from a single dissection and that his finding, although quite ideal, is mere chance. The style, being quite short, may persist unbroken into the fruit. It is 428 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII quite thickened at the base and then divides into three, four, or five short branches. In the dried fruit, the branches are usually pressed down, brittle, and difficult to separate. I do feel, however, from the examination of the material available that the number of branches of the style is the same as the number of locules in the ovary or fruit, a conclusion which is logical rather than observationally accurate. This occurrence of varying numbers of locules in the fruit obviously shows that an evolutionary change is taking place. In all of the material of Ternstroemia which I have studied from the Eastern Hemisphere I have found only one other variant from the 2-loculate ovary and fruit. Ternstroemia patens (Korthals) Choisy, a very rare plant of Sumatra, is 1-loculate in both flower and fruit. Furthermore, a single pendulous ovule is found in the ovary, and a single seed in the fruit. One may con- sider this the extreme in development. It is possible that Ternstroemia kwangtungensis represents a more primitive taxon of the genus in transition from a 5-loculate to a 2-loculate state. The two sterile locules in the undeveloped fruit of the type might be considered further evidence of this transition. As far as I know, this taxon has been collected only three times and only in Kwangtung. The type was collected by Merrill at Lofaushan, an area > known botanically, about 60 miles east-northeast of Canton (23° 15° N, 114° 8’ E). Tsang (25486) collected the type of Ternstroemia eran a at Linfashan (22° 59’ N, 115° 20’ E), which is about 75 miles southeast of the Lofaushan. Both of these specimens I have exam- ined. Chang cited a third specimen (H. Tse 608) collected at Yingtak (24° 10’ N, 113° 5’ E), located about 80 miles west-northwest of Lofau- shan. These three collections extend over a distance of 150 miles. In well-collected Kwangtung, it seems strange that it has not been encoun- tered more often, since Tsang commented in his field notes that it was “fairly common.’ Below is recorded a nomenclatural treatment with an amplified descrip- tion. Ternstroemia kwangtungensis Merrill, Philipp. Jour. Sci. Bot. 13: 148. 1918.— Melchior, Nat. Pflanzenfam. ed. 2. 21: 141. 1925. Ternstroemia subrotundifolia Chang, Zhongsan Daxue nena Ziran Kexue (Bull. Sun Yatsen Univ. Nat. Sci.) 2: 25. pl. 7. 1959 — syn. DISTRIBUTION: China (Kwangtung). KWANGTUNG: Loh Fau Mountain (Lofaushan), in damp, forested gorge, alt. ca. 300 m., E. D. Merrill 11013 (a, lectotype* of Ternstroemia kwangtungensis ), Aug. 29, 1917; Hai-fung Distr., Lin Fa Shan, Pak Shui Chai, W. 7. Tsang 25486 (A, isotype of Ternstroemia subrotundifolia), Aug. 15-24, 1935 (scattered shrubs, fairly common, 3 ft. high; fruit red) * The original holotype, deposited in the herbarium of the Philippine Bureau of Science, was destroyed during the Second World War, hence this designation of a lectotype 1961 | KOBUSKI, STUDIES IN THEACEAE, XXXIII 429 Shrub or small tree, 1-4 m. high. Branches and branchlets stout, 5-7 mm. diameter, rugose, reddish brown, the petiolar scars prominent. Leaves thick-coriaceous, elliptic-ovate to suborbicular, 7-9 cm. long, 3-5 mm. wide, rounded or obtusely acuminate at the apex, obtuse or broadly cuneate at the base, dark green above, tawny or reddish brown and dark-punctate beneath, the margin glandular-denticulate, especially along the upper half on younger leaves, less obvious on the older leaves, the midrib ie above, plane or lightly elevated below, the primary veins 5—7 eA quite indistinct, curving upward near the margin, the petiole robust, —2 cm. long. Flowers not seen. Fruit axillary, solitary, ieee eee 1—-1.6 m. long, 1.2-1.8 cm. diameter, 3—locular, occasionally 4- or 5-locular, 5—7-seeded, the seeds conduplicate, 7-8 mm. long, ca. 6 mm. diameter, covered with a tawny, mealy aril; fruiting style short, thick, 1-2 mm. long, branching near the apex, the style branches apparently equalling the locules in number, each branch topped by a small capitate stigma; fruiting pedicel stout, 1-1.5 cm. long, ca. 3 mm. diameter; persistent bracteoles 2, opposite, immediately below the calyx, broadly ovate or deltoid-ovate, ca. 3 mm. long, 4-5 mm. wide at the base, the margin scarious, persistent calyx-lobes rounded, ca. 6-7 mm. long and wide. 430 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII THE CORRECT NAMES FOR “DIOSPYROS EBENASTER” RicHArp A. Howarp THE NAME Diospyros ebenaster Retzius has been commonly applied to trees from the West Indies, Central America, South America, and many areas of tropical Asia, The tree and its fruit are known as black apple or sapotte negro. Collectors in both the New and the Old World have sug- gested that certain of their collections may represent cultivated plants, but some have considered the tree to be indigenous in all of the area cited. The occurrence of early historical records of this plant in botanical gardens suggests that its use by man may have been the basis for its introduction into new areas. It is reported to have an edible fruit, but either the quality of the fruit varies, or else individuals vary in their acceptancé of its palatability, if one is to believe the conflicting reports. In both hemi- spheres, parts of the plant are used by the native peoples as a fish poison. The peculiarly widespread distribution of this species has made it diffi- cult to determine the country of origin. Scott (Kew Bull. 1915: 65. 1915), who reviewed thoroughly the existing literature regarding this species, pointed out that “the bulk of the evidence, then, points to Diospyros Ebenaster being introduced into the East Indies.” He also pointed out that Urban (Symb. Ant. 4: 485. 1910) “is apparently convinced that the tree is a native of the West Indies” and that Duss’ (Fl. Phan. Ant. Fr. 390, 1896) “discoveries leave us no doubt whatever, for he found the tree in various wild places” in Guadeloupe. Scott concluded that all records from Brazil seemed to indicate that the tree was cultivated, and he summarized his findings by saying that “from the evidence we have collected, there- fore, we must conclude that the natural habitat of Diospyros Ebenaster, Retz., is in the West Indies. With regard to Mexico, it is evident from the references quoted . . . that the tree must have existed there for at least three centuries . . . It is very probable that the tree was introduced into Mexico by the Spaniards, but if this be correct, the introduction must have taken place at a very early date.” In Montserrat, I have been able to examine trees, called the black apple, which would be referred to Diospyros ebenaster and have also seen her- barium specimens of this plant from Cuba, Hispaniola, Puerto Rico, Guadeloupe, Marie Galante, and Dominica in the West Indies. Excluding the collections from Cuba and Haiti, not one of the sheets suggests a plant under cultivation, or anything other than a component of the natural vegetation. The trees in Montserrat were in native woodland in areas which have never been heavily cultivated, having been preserved histor- ically as watershed areas. However, a total of seventeen collections from Mexico, British Honduras, and Costa Rica has been studied, of which three specimens are clearly indicated as being from cultivated plants. From these recent collections one might draw evidence in support of Scott’s 1961] HOWARD, “DIOSPYROS EBENASTER” 431 conclusion that D. ebenaster is a native of the West Indies which has been introduced into Mexico and Central America. However, a study of the material from these two geographic areas indicates that two species are represented, and that it is the Mexican-Central American population which has been introduced into the Pacific area, Cuba, Haiti, and Brazil. The Mexican-Central American plants have oblong-lanceolate leaf blades evenly tapered at both ends or rounded at the base and bluntly acute at the apex. The venation is ascending and arcuate, anastomosing near the margin. The fruit is borne on a stout pedicel averaging 5 mm. in thick- ness and the fruit is 5-6 cm. in diameter. The fruiting calyx is 4—5 cm. in diameter with large, broadly ovate lobes, the whole calyx appearing undulate or wavy. In contrast to this, the ample material obtained for the purpose of study- ing variation in Montserrat, supported by additional collections for the other islands of the Lesser Antilles, has, on the average, smaller leaves which are elliptic-oblong, narrowed at the base, and rounded, obtuse, or often emarginate at the apex. The blades appear to have had a heavier texture than those of plants from Central America. The venation is ascending, but is not arcuate until the veins bifurcate and anastomose near the margin of the blade. The fruit is smaller, averaging 3 cm. in diameter, although it is noted on Hodge 2091 that the “fruits grow to 3 >< the size of this specimen.” More significant, however, is the shape and size of the fruiting calyx which, in the largest specimen available, is 1.5 cm. in diameter and almost square in surface view, with the calyx lobes scarcely indented and the whole calyx flat and coriaceous. Adequate flowering material is lacking from both geographic areas. The evidence thus indicates that there are in the New World two species which have passed as Diospyros ebenaster. Herbarium specimens available from the Philippines and Malaysia may be clearly assigned to the taxon from Central America. Both species require names. Bakhuizen van den Brink, in a revision of the Malaysian Ebenaceae (Bull. Jard. Bot. Buitenzorg III. 15: 316-320. 1938), considered Dio- spyros ebenaster Retzius to be a probable synonym of D. ebenum and renamed the material generally called D. ebenaster as D. nigra (Gmel.) Perrottet. Bakhuizen considered this species to be indigenous to the Antilles, Brazil, Mexico, Texas, and Florida, and both spontaneous and cultivated in the Philippines, Celebes, Mauritius, Réunion, Ceylon, and Malacca. In taking up D. nigra, Bakhuizen gave an extensive citation of both pre-Linnaean as well as post-Linnaean literature. The basic species cited and the supporting literature references, here partly rearranged, are the following: Diospyros nigra (J. F. Gmel.) Perrottet, Mém. Soc. Linn. Paris 3: 113. 125, Sapota nigra J. F. Gmelin, Syst. Nat. ed. 13. 2: 750. 1791. D. digyna Jacq. Hort. Schoen br3: 35). 313. 1798. D. revoluta Poir. in Lam. Encycl. Méth. Bot. 5: 435, no. 18. 1804. 432 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII D. “i ag Humb. & Bonpl. ex Willd. Sp. Pl. 4: 1112, no. 17. 1805; HBK. Nov. Gen. et Spec. Pl. 3: 253. pl. 247. 1818 D. pe eeeaiceey, DC. Ess. Prop. Méd. PI. ed. . 200. 1816. D. sapota Roxb. Flor. Ind. ed. 2. 2: 535. 18 Sapota nigra Blanco, Fl. Filip. 409. ee D. nigra Blanco, FI. Filip. hag 2.211. 18 D. membranacea A. DC. in DC. Prodr. a Nat. 8: 227, no. 20. 1844. D. laurifolia A. Rich. FI. Cub. in Ramon de la Sagra, Hist. Cuba 11: 86. 1853; Ramon de la Sagra, Icon. Fl. Cub. Pl. Vasc. pl. 55. 1863. D. brasiliensis Mart. ex Miq. in Mart. Flor. Bras. 7: 5. pl. 2, fig. D. Ebenaster Hiern, ane r. Eben. in Trans. eae Phil. Soc. ein 2) : 244. no. 125. 1873, non ack Although two species are involved in the description and in the ref- erences given by Bakhuizen, the name he has chosen cannot be used for either one. Apparently Bakhuizen did not see the reference he used for Perrottet’s eipthet, for Perrottet did not transfer Gmelin’s name, but de- scribed a new species. The article by Perrottet is short and reads ‘*DIO- SPYROS nigra Perr. Espéce nouvelle des Philippines que les créoles de Mascareigne appellent Sapot negro; son fruit est trés-gros, assez semblable, pour la forme, au melon cantaloup galeux” (Mém. Soc. Linn. Paris 3: 113. 1825). Perrottet further indicated that the plant was cultivated in the Mascarene Islands and at Cayenne. Gmelin (Syst. nat. ed. 13. 2: 750. 1791) in contrast, had based his de- scription on the phrase ‘““Sapotte negro” used as a common name, although given with a description and illustration, by Sonnerat in his Vovage a la Nouvelle Guinée (p. 45. pl. 14-16. 1776) for a plant from the Philippines. Although Perrottet, Gmelin, and Sonnerat were all describing the same species, Gmelin’s Sapota nigra cannot be transferred to Diospyros be- cause of Perrottet’s publication. It is also of interest to note that the same plant from the Philippines was described as new by Blanco as Sapota nigra (FI. Filip. 409. 1837) and later transferred to Diospyros as D. nigra (Blanco) Blanco (FI. Filip. ed. 2. 211. 1845). The earliest available name for the Mexican-Central American plant introduced in the Pacific Islands is Diospyros digyna Jacquin, based on a plant grown in Austria from seeds collected in the Celebes. No type specimen of this species exists at Vienna, but the original ink drawing used in preparing the published illustration (Hortus Schoenbrunnensis 3: 35. pl. 313. 1798) is preserved in the Vienna herbarium. The illustration must suffice to typify the species. The drawing of the branch showing the leaves, their venation, and the position of the flowers, matches fairly well the Pacific Island material of “‘Sapote negro,” as well as material from Mexico and Central America. The calyx, illustrated by Jacquin, appears to be that of a staminate flower. The pistil has a pubescent ovary typical of that found in recent collections examined from the Old World as well as Central America. The Jacquin iJlustration shows two styles, as indicated in the specific epithet, but a comparable abnormality has not been located in the staminate flowers examined. Certainly the majority of the char- 1961 | HOWARD, “DIOSPYROS EBENASTER” 433 mee tht ale Cot OR om wan. * Ko LF 6 aj — ares £o5 4 HOLOTYPE OF DiospyRos- MEMBRANACEA A. DC. FrRoM Mauritius. (Pro- dromus Aa haw. Geneva acteristics to be seen in the illustration are pertinent to the Central American species. The common name of “‘false mangosteen” may also be considered as support. The correct name and the synonymy of this species appears to be: 434 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Diospyros digyna Jacq. Hort. Schoenbr. 3: 35. pl. 313. 1798. Sapota nigra J. F. Gmelin, Syst. Nat. ed. 13. 2: 750. 1791. Diospyros obtusifolia Humb. & Bonpl. ex Willd. Sp. Pl. 4: 1112. 1805; HBK. Nov. Gen. Sp. Pl. 3: 253. pl. 247. 1818. Diospyros sapota Roxb. FI. Ind. ed. 2. 2: 535. 1832. | Diospyros sapotanigra A. DC. Ess. Prop. Méd. Pl. 200. 1816. Diospyros nigra Perrottet, Mém. Soc. Linn. Paris. 3: 113. 1825. Sapota nigra Blanco, FI. Filip. 409. 1837, non Gmelin. Diospyros membranacea A. DC. Prodr. 8: 227. 1844. Diospyros nigra (Blanco) Blanco, FI. Filip. ed. 2. 211. 1845, non Perrottet. ie — A. Rich. Fl. Cub. in Ramon de la Sagra, Hist. Cuba 11: ; Ramon de la Sagra, Icon. Fl. Cub. Pl. Vasc. pl. 55. 1863. Several of these synonyms have not been discussed above. Diospyros obtustfolia was based on a plant from Mexico and clearly is to be referred here. Diospyros sapota Roxb. has the same basionym as Sapota nigra Gmelin in a Sonnerat reference to a plant from the Philippines. Dio- Spyros sapotanigra A. DC., attributed to Mexico, is a nomen nudum. Diospyros laurifolia A. Rich. is well illustrated and, although from Cuba, clearly refers to the Central American species and must represent an intro- duced plant. Diospyros membranacea A. DC. was based on a collection from Mauritius. A photograph of the holotype which was made available through the kindness of Dr. Charles Baehni indicates that this species is also to be referred here. Dr. R. E. Vaughan, of Reduit, Mauritius, has written me, “I have seen a specimen in the Paris Herbarium collected by Commerson in Mauritius C. 1769 so it has been here for a long time. It appears in a list of plants growing in the Botanical Gardens, -Pample- mousses in 1785 under the name ‘Sapotte negro des Molucques.’ I have seen no specimens determined as Diospyros brasiliensis which Miquel attributed to Martius (Flor. Bras. 7: 5. pl. 2, fig. 2. 1856), but the species was collected by Pohl in Brazil and, judging from the illustra- tion, may well be assigned to the synonymy of D. digyna. Bakhuizen’s reference to Diospyros nigra occurring in Texas and Florida was not supported by cited specimens, and I have seen no cultivated specimens from either state. The date or the path of the introduction and the distribution of the Mexican-Central American sapote negro to the Pacific area cannot be de- termined. It is clear, however, that the plants of the Pacific area called Diospyros ebenaster Retzius by many authors and D. nigra by Bakhuizen originated in Mexico-Central America and are correctly called D. digyna Jacq. The native plant of the Lesser Antilles (with one record from Puerto Rico) is distinct, although previous authors have placed it with the species just discussed. The correct name for this plant is Diospyros revoluta Poiret (Lamarck, Encycl. Méth. Bot. 5: 435. 1806), and it was origi- nally described as follows: 1961] HOWARD, “DIOSPYROS EBENASTER” 435 Diospyros foliis obovatis, obtusis, glaberrimis; pedunculis solitarius, reflexts ; calicibus profunde lobatis. (N). Les tiges des jeune rameaux sont striées, presqu’anguleuses, munies de feuilles alternes, pétiolées, coriaces, trés-glabres, luisantes, ovales, rétrécies, aigués a leur base, arrondies 4 leur sommet, longues de quatre a cinq pouces, larges au moins des deux, supportées par des pétioles trés courts, marquées de nervures doncules courts, fortement recourbés: leur calice est glabre, persistant, coriace divisé en quatre grands lobes profonds, arrondis. Les quatre divisions du tube de la corolle sont équalement rréfléchie es en sea sur le jeune fruit. Celui-ci est un baie glabre, arrondie, d’une médiocre grosseur. Cette plant est originaire de l’Amérique mnentoals Elle a été communiquée par M. Vahl a M. Jussieu. [symbol for a shrub] (V. S. fruct. in herb. Juss.) I have been unable to secure information regarding the existence of a specimen in the Jussieu Herbarium in Paris. In the Copenhagen herbarium there is a specimen studied by Vahl which was collected by Ryan in Montserrat. Associated with the specimen is a complete diagnosis in Latin, with the name Diospyros revoluta. The specimen was originally identified as D. tetrasperma Sw. and then considered to be a different and new species. Vahl sent the information, and perhaps a specimen, to Jussieu. The specimen and both the Latin and the French diagnoses agree with the Lesser Antillean plants commonly known as D. ebenaster Retzius. Diospyros revoluta apparently has not been transferred by man to other areas, and the following records appear to represent native plants of the West Indies. Dominica: Calibishie, W. H. & B. T. Hodge 3177 (cu); Laudat, W. H. Hodge 2091 (cH); between Salybia and Concorde Valley in Carib Reserve, W.H. & B.T. Hodge 3282 (GH); Pointe Rond to Milton Estate, W. 8 Hodge 2676 (GH). Marte GALANTE: Pointe de Folle Anse, Proctor 21068 (A). MonTSERRAT: Without specific location, Ryan (type); Salem, Shafer 559 (us), Proctor 18897 (a), Howard 15104, 15117 (A). Puerto Rico: La Sardinera, Britton 8508 (GH); Toa-alta, Stahl 872 (cH). This species is known as “babara’’ in Dominica and as the “black apple in Montserrat. It is used as a fish poison in both areas. 436 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII IsoTYPE OF DIOSPYROS REVOLUTA PorRET. Specimen from the Botanical Mu- seum, University of Copenhagen, collected by Ryan in Montserrat. 1961] ANDRESEN & BEAMAN, NEW SPECIES OF PINUS 437 A NEW SPECIES OF PINUS FROM MEXICO! JoHN W. ANDRESEN AND JOHN H. BEAMAN THE summit of Cerro Potosi, possibly the highest mountain in the state of Nuevo Leon, Mexico, is encircled by a timberline made up of two species of pines. One of these is Pinus hartwegit Lindl., the dominant timberline species on the high volcanoes of southern Mexico. The other is a dwarf pifion not previously described. The pine which Muller (1937, 1939) called P. flexilis James is this species. The localities indicated by Martinez (1948, p. 108) under P. flexilis may refer to the new species, but the specimen he illustrated (p. 105, fig. 72) includes a cone which is definitely that of P. flexilis. We therefore cannot be certain that P. flexilis does not also occur on Cerro Potosi, although we did not find it there. The appearance of timberline on Cerro Potosi (Fic. 1) is similar to areas where P. flexilis occurs in the Rocky Mountains, but the two species re- semble each other only in superficial habit and leaf characters. Their cones are very different. Pinus flexilis is a member of the subsection Cembra, group Flexiles (sensu Shaw, 1914). The new species belongs to subsection Paracembra, group Cembroides. Specimens in the Herbarium of the Chicago Natural History Museum were properly considered to be- long to the pifion group by Dr. J. A. Steyermark, who, in 1940, annotated them as P. cembroides Zucc. The Schneider specimens in that Herbarium have also been previously determined as P. quadrifolia Parry ex Parl. Pinus culminicola Andresen & Beaman, sp. nov. Fics. 2-4. DESCRIPTION OF THE SPECIES. Widely branching shrub 1—5 m. high with dense crown; leaves 5 per fascicle (rarely 4 or 6), 3-4.9 cm. long, 1-1.3 mm. thick, remotely serrulate, incurved in dense clusters at the ends of thick branchlets, glaucous ventrally, blue-green dorsally, stomata on ventral surfaces only, apices rounded and conspicuously cutinized; fascicle sheathes of young leaves 6-8 mm. long, stramineous, later becoming curled into persistent rosettes; conelets ovate, subsessile, single or in pairs, 8 mm. wide, 11 mm. long, rufous-brown, with short-mucronate scales; cones sub- globose, 3.2—4.5 cm. long, 3.5-4 em. wide; cone scales 50—60, inner and outer surfaces (exclusive of apophyses and seed cavities) stramineous; apophyses slightly raised, subrhomboidal, stramineous- to chestnut-brown, laterally keeled, convex above and below the umbo, those of the largest scales 1.8 cm. wide, 1.3 cm. high, 3 mm. thick; umbo dorsal, rhomboidal, minutely appendaged; seed-bearing cavity 9 mm. long, 6 mm. wide; seeds not seen, apparently unwinged, | or 2 per scale. — by grants G-9045 and G-15879 from the National Science Foundation and by a grant from the ee Council of the American Museum of Natural His- tory. 438 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII DIAGNOSIS TYPICI EXEMPLARIS. Frutex, 1 m. altus, coma effusa; folia 3—4.9 cm. longa, 1—-1.3 mm. crassa, quinque in fasciculo, ad margines remote serrulata, rigida, incurva densaque, ad terminos aggregata, ventrale glauca, dorsale caeruleo-viridia, cum stomatis solum ventralibus; vaginae 6—8 mm. longae, in brevi tempore crispae-rosulatae; strobilus subsessilis, sub- globosus, 4.5 cm. longus, 4 cm. latus; squamae stramineae; apophyses subrhomboideae, fulvo-spadices, a latere carinatae, supra et subter con- vexae, maxima earum 1.8 cm. lata, 1.3 cm. alta, circa 3 mm. crassa; umbo dorsalis, rhomboideus, cum appendiculis minutis; cavum ubi semen posi- tum 9 mm. longum, 6 mm. latum; semina ignota, evidenter sine alis, 1 vel 2 per squamam.” Fic. 1. Aspect of timberline on Cerro Potosi. The dominant low shrub is Pinus culminicola. The erect, longer- leaved pine in the foreground and the tree on the horizon are P. hartwegii. Type. Mexico. Nuevo LEON: Cerro Potosi, near top of mountain, ca. 3600 m. alt., in dense stand at timberline in soil weathered from hard, gr ay limestone, Beaman 2675 (MSC 164008, holotype; A, US, isotypes) OTHER SPECIMENS EXAMINED. Nuevo LEON: Northeast summit of Cerro Potosi, ca. 3650 m. alt., Andresen 666 (msc); Sierra Madre Oriental. ascent of Sierra Potosi by north hogback about 20 miles NE. of Galeana, alt. 11.600 ft.. * We are grateful to Dr. Mladen Kabalin for editing the Latin diagnosis. 1961 } ANDRESEN & BEAMAN, NEW SPECIES OF PINUS 439 Mueller & Mueller 1241 (a, ¥F, MICH, MO); the peak of Cerro Potosi, Municipio de Galeana, Mueller 2261 (a, F, MICH, MO); at timberline on Cerro Potosi, Mu- nicipio de Galeana, 12,100 ft. alt., Schneider 957 (¥, 2 sheets).* Only three species of pions (Pinus cembroides, P. nelsoni Shaw, and P. pinceana Gord.) were recognized by Shaw (1914). Pinus nelsoni and P. pinceana are Mexican species with a limited distribution and conspicuous cone characters. Pinus cembroides (sensu Shaw) is widely distributed in the western United States and Mexico. Its variability is such that four ies Ns WE Rife SIRS iy SIA A 4 Soars CRE. Oar. a J ste FA CACO SA ae. Nee = x (S GLEE Z Z SS = YB ————— SS = i = —_ = = a So G => ra im 3 - 5 5 OE ves — 2 = > rT 7 << “| Go = Re, eS = I, +t x iy) TY Spa ~ a. 7 Ax eh ae oe lo _ ~ = SU : = Ss ~ — - = cs SS q : S ~ 07: bie ; Weecomeseees a PS eae Fics. 2—4. Branch, cone, and leaf fascicle of Pinus culminicola, 2, Branch showing dense, clustered fascicles of incurved leaves and enlarging candle (x 2/3). 3, Cone with details of apophyses and umbos (X 2/3). 4, Portion of five- leaved fascicle with rosulate, persistent sheath (X 4 2/3). Drawings by Nancy R. Halliday. * We appreciate the privilege of examining the specimens in the herbaria cited. 440 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII segregate species (P. cembroides, sensu stricto; P. monophylla Torr. & Frem.; P. edulis Engelm. in Wisliz.; and P. quadrifolia) have been main- tained by several other authors. Shaw did not attribute much importance to the leaf characters by which these species can be distinguished, but to us they appear to have considerable value. The leaf differences are also correlated with cone characters of diagnostic value (cf. Little, 1950). Likewise, Mirov (1953) noted that the three species of the P. cembroides complex which he examined are not the same chemically. Differences be- tween the species are evident in areas where their ranges overlap, suggest- ing the presence of genetic isolating barriers. The recognition of these taxa as species therefore seems preferable to the alternative of placing them at an infraspecific level. Pinus culminicola and the four other species of the P. cembroides com- plex are separated from each other by characters of about the same magni- tude. Probably the closest relationships of the new species are with P. edulis and P. quadrifolia. The most conspicuous character for distinguish- ing P. culminicola from its relatives is the five-leaved fascicle. Its cones are larger than those of P. cembroides, smaller than those of P. monophvila, and about the same size as those of P. edulis and P. quadrifolia. In hab- itat P. culminicola differs greatly from the other pifons which are found mostly in semiarid regions at the lower limit of forest vegetation. This species occurs in less arid circumstances at the upper forest limit. Prob- ably no other species of pinon occurs as high as 3000 meters, while the lower altitudinal limit of P. culdminicola is around 3400 meters. Pinus culminicola and P. cembroides are essentially sympatric. The latter occurs abundantly near the base of Cerro Potosi (cf. Martinez, 1948, p. 85). We have compared the collection Andresen 675 (msc) of P. cembroides from 8 km, south of Galeana, Nuevo Leon, with P. cul- minicola and found the two species very different in characters of the cones, foliage, branches, and habit. The ranges of P. culminicola and P. cem- broides are altitudinally separated but are close enough that pollen ex- change is theoretically possible. Their distinctness is nevertheless main- tained. Since pines are conspicuous plants of considerable economic importance they have been frequently collected and studied. Probably almost all of the species of the genus are already known, but it is not entirely surprising that a new species should be found on Cerro Potosi where endemics are common. For example, Muller (1939) noted that six of 16 species he aster from the alpine meadow were collected for the first time in 1934 and 1935. The high elevation of the mountain (probably ca. 3650 aeal. its geological age, and its wide geographical separation from the Rocky Moun- tains and the Mexican volcanoes are responsible for the high rate of endemism. As the flora of Cerro Potosi becomes better studied probably still more new species will be found there. 1961] ANDRESEN & BEAMAN, NEW SPECIES OF PINUS 441 LITERATURE CITED LittLe, E. L., Jr. 1950. Southwestern trees: A guide to the native species of New Mexico and Arizona. U.S. Dep. Agr. Handb. 9. 109 pp. ee M. 1948. Los pinos ee Ed. 2. Ediciones Botas, Mexico. ee Nv. T 1953. Taxonomy and chemistry of the white pines. Madrono 12: MULLER, c H. 1937. Plants as indicators of climate in northeast Mexico. Am. Midl. Nat. 18: 986-1000 . 1939. Relations of the ices and climatic types in Nuevo Leon, Mexico. Am. Midl. Nat. 21: 687- SHaAw, G. R. 1914. The genus Pinus. ned Arnold Arb. 5. 96 pp. DEPARTMENT OF FORESTRY AND DEPARTMENT OF BOTANY AND PLANT PATHOLOGY, MICHIGAN STATE UNIVERSITY, EAST LANSING 442 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII FIBERS IN THE SECONDARY XYLEM OF AESCHYNOMENE U. PRAKASH WHILE WorRKING on the petrified woods from the Deccan Intertrappean beds of India, I encountered a fossil wood which proved to belong to the genus Aeschynomene, of the family Leguminosae. In the course of this identification, doubts arose as to the proper interpretation of the ground- mass of the secondary xylem in Aeschynomene. Solereder (1908, pp. 275— 276, fig. 117a; Taubert, 1894, p. 319) and Metcalfe and Chalk (1950, p. 523, figs. 117a—d) report that the parenchyma constitutes the greater part of the tissues in this genus and forms a ground-mass in which groups of fibers and vessels are distributed. Indeed, the elements of the ground-mass very much resemble those of fusiform parenchyma cells. However, my own investigation shows a different interpretation. Thin sections of the woods of Aeschynomene hispida Willd. (Yw * 2169), A. cf. aspera L. (Yw 37202), A. sensitiva Sw. (Yw 43509), and an unidentified species (Yw 16552) were examined. These preparations were made from dried material in which, unfortunately the protoplasmic contents and the starch grains were no longer visible. Professor I. W. Bailey sug- gested that the ground-mass appeared to be composed of fibers. He further suggested that this could be confirmed from fresh material of Aeschynomene and by examining it without hydrofluoric acid treatment. Through the courtesy of Dr. P. T. Richards, of the Imperial College of Tropical Agri- culture, Trinidad, West Indies, I secured fresh material of Aeschynomene sensitiva Sw. collected in Trinidad and preserved in formalin-acetic- alcohol. The present study is based on this material which was sectioned at 20-25 micra and stained in Haidenhain’s iron-alum haematoxylin and safranin. A part of the wood was also macerated in chromic-nitric acid. As the objective of this study was to ascertain the nature of the ground- mass of Aeschynomene, i.e., whether it is parenchymatous or fibrous, the observations given here are confined to the secondary wood. The micro- scopic examination of the cells forming the ground-mass shows that they are devoid of protoplasmic contents and do not possess starch grains (Fics. 2, 5). This indicates their nonliving condition in the mature wood and, hence, their fibrous nature. It is interesting to note that this tissue shows a variety of cell structures, from libriform fibers to intermediate forms which are difficult to classify (Fic. 1b-f), culminating in a type of cell similar in shape to the fusiform cambial initials. In addition to these prosenchymatous cells, but excluding cells of ray parenchyma, there are other types in the secondary xylem which possess protoplasmic con- tents with nuclei, as well as starch grains (Fic. 6). These are the true * Record Memorial Wood Collection, Yale ger citation of wood specimens as proposed by Stern and Chambers, Taxon 9: 1961 | PRAKASH, FIBERS IN AESCHYNOMENE 443 parenchymatous cells forming the longitudinal parenchyma of the wood. Their parenchymatous nature is also shown in some cases by the presence of septa, forming strands of two to four cells (Fic. 5). In the present species the longitudinal parenchyma is usually associated with the vessels in a scanty, abaxial and aliform to confluent distribution (Fics. 2, 4). A study of different forms of fiber and longitudinal parenchyma cells shows that the fusiform cambial initials differentiate two types of cells: (1) fusiform parenchyma cells, which may divide into parenchyma strands of two to four cells (Fics. 5, 1b ', b*), and (2) libriform fibers, with long, attenuated ends, which are associated with a series of cell types of various shapes, including those possessing a form like that of fusiform cambial initials (Fic. 1b—f). ® 2S DAMapoTceee 2eGne® Ir CAMBIAL INITIAL > 83a 2228 9599000200 20g <2, FIBER CELLS b PARENCHYMA CELLS Fic. 1. Semidiagrammatic camera lucida sketch showing various types of fiber and parenchyma cells in Aeschynomene sensitiva, X 235. Parenchyma cells contain starch grains (s). See explanation in tex From the above considerations it is quite evident that the fibers con- stitute the greater part of the tissues in the secondary wood of Aeschyno- mene and form a ground-mass in which the parenchyma and vessels are distributed. In view of these interpretations, the wood of Aeschynomene sensitiva Sw. is described below. Growth rings: Not seen. Vessels: Small, 20-65 » in tangential diameter 444 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII (mean 42 ,), chiefly solitary, occasionally in multiples of 2-4 cells or rarely in clusters (Fic. 2), quite uniform in size and evenly distributed, thin walled; solitary vessels mostly polygonal to oval in cross section. Vessel elements 90-210 » long, with truncate to oblique end walls. Perfora- tion plates exclusively simple, horizontal to slightly oblique. Intervascular pit-pairs vestured, alternate to opposite or subopposite, orbicular to oval in cross section, 7-10 uw in diameter, with linear to lenticular apertures, sometimes coalescent (Fic. 7). Tyloses absent. arenchyma: Paratracheal, scanty abaxial or vasicentric, sometimes forming irregular wings which rarely unite to form confluent parenchyma (Fics. 2, 4); rarely isolated diffuse parenchyma cells also seen among the fibers. Parenchyma cells storied, thin walled, angular in cross section, 10-18 ,» in diameter and containing nucleus, starch grains, cytoplasm, and protoplasmic contents (Fics. 4-6). Parenchyma cells fusiform or in 2—4- celled strands (Fic. 5), sometimes chambered with solitary crystals in each chamber. Xylem rays: Homogeneous (Kribs Homogeneous Type 3) composed of procumbent cells, storied, mostly uniseriate, sometimes biseriate (Fic. 5 rarely quite broad and containing vascular strands; uniseriate rays 1-24 (often 8-10) cells high; up to 10-15 rays per mm. Ray cells variously shaped, square, rectangular, oval to oblong and rarely pointed at one end, containing starch grains (Fic. 5). Imperforate tracheary elements: Mostly thin walled, sometimes slightly thicker walled, especially near the vessels, and extending tangentially from them (Fie. 3), usually polygonal (mostly 5 or 6 sided) in cross sec- tion with large lumina, about 10-25 » in diameter (Fics. 2, 3, 4); non- septate; storied (Fic. 5); pits simple, slightly more numerous on the radial than on the tangential walls, often appearing to be arranged in groups (Fic. 1b-f); pitting often markedly more abundant on and near the pyramidal walls. Cells variously shaped in longitudinal section (Fic. 1b-f), varying from libriform fibers to a fusiform type. The author wishes to express his sincere appreciation to Professor Elso S. Barghoorn for his kind advice and for his critical reading of the manu- script. To Professor I. W. Bailey are extended many thanks for his help in cutting microtome sections of the wood of Aeschynomene sensitiva and for his many valuable suggestions. Dr. P. T. Richards, of the Imperial College of Tropical Agriculture, Trinidad, and Dr. William L. Stern (then curator of the S. J. Record Memorial Collection, Yale University), curator, Division of Woods, National Museum, Smithsonian Institution, Washing- ton, D. C., have kindly made available woods of extant species of Aeschynomene, for which the author is greatly indebted. This study was completed while the author was a participant in the Visiting Research Scientists Program of the National Academy of Sciences, Washington, D. C., to which he is grateful for financial assistance. DEPARTMENT OF BIOLoGy, HARVARD UNIVERSITY 1961] PRAKASH, FIBERS IN AESCHYNOMENE 445 LITERATURE CITED Metcatre, C. R., and L. CHALK. 1950. Anatomy of the dicotyledons. 2 vols. Oxford xford. SoLEREDER, H. 1908. Systematic anatomy of the dicotyledons. vol. 1. (Trans- ted I,, A. Boodle and R. E. Fritsch, and revised by D. H. Scott.) Oxford. TauBERT, P. 1894. Leguminosae. Jn: Engler and Prantl, Nat. Pflanzenfam. III. 3: 319-320 EXPLANATION OF PLATE Fics. 2-7. Modern wood of Aeschynomene sensitiva. 2, Transverse section, x 50 — note vessel and parenchyma distribution ; parenchyma cells appear black due to presence of starch grains in them. 3, Transverse section magnified to show the distribution of thick-walled fibers * 140. 4, Transverse section magnified, < 140—note paratracheal parenchyma with starch grains (shaded black). 5, Tangential section, X 140—note storied rays, fibers, and paren- chyma. 6, Longitudinal section of a parenchyma and a fiber cell, X 455 — note starch grains and nucleus in the parenchyma cell. 7, Intervascular pitting, X 180. Jour. ARNOLD Ars, VoL. XLII PLATE I ae a me vy . ® © a Bwvraae = + . Uy TA eee... 8 #: e fe Ree ao Sor. r | Ri@ete a@ @ 18,018 2fe. 7 baz ween} qinessaah tae @a ean @e e@ a a ¢ \ | @e- et. | r “| > aay sai 7 8 -¢@ J: ad a op eguen? te @@ et fre o'2e %e- Be-ges 509),9 @-g%2 styattrees © 565 ts 5.08 07, onltume © PRAKASH, FIBERS IN AESCHYNOMENE 1961] THE DIRECTOR’S REPORT 447 THE DIRECTOR’S REPORT Tue ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JuNE 30, 1961 The Staff: It is with regret that we record here the death of Joseph Horace Faull, professor of forest pathology, emeritus. Professor Faull joined the staff of the Arnold Arboretum on July 1, 1928. A special laboratory was built for this area of research on the grounds of the Bussey Institution, and within a few years a greenhouse was added. Professor Faull, his students, and associates were among the first to recognize the significance of an arboretum with its massed collections of related species as a locale for the study of diseases of ornamental plants. To him and his team must go credit for the early work involving a service to the public in the identification and control of plant diseases now commonly carried on by state and federal experiment stations. From his laboratory came the early identification of a disease first found in this country in Ohio and now known as the Dutch elm disease. He warned prophetically of the danger of this fungus. In the course of his work, he and his students pub- lished studies of the fungal diseases of many genera of ornamental plants, including Gleditsia, Cornus, Abies, Pinus, Picea, and Fagus. Dr. Faull taught a class in forest pathology in Harvard College and guided the pro- grams of nearly a dozen graduate students, many of whom have also achieved success in their field. Dr. Faull retired in 1940 but remained active in his laboratory for several additional years. Regrettably his posi- tion vacated during a shortage of skilled personnel concurrent with World War II could not be immediately filled and after the war was not con- tinued due to a lack of funds. Dr. Faull’s extensive herbarium of fungus diseases of cultivated plants was transferred to the Farlow Herbarium in Cambridge. Dr. Faull died on June 30, 1961, at the age of 91. biographic report of his life will be prepared by his daughter, Dr. Anna Faull, for publication in a later issue of the Journal of the Arnold Arbo- retum., On January 1, 1961, Mr. Peter S. Green joined the staff of the Arboretum. Formerly scientific officer at the Royal Botanic Garden, Edinburgh, Scotland, Mr. Green has been studying various genera of the Oleaceae, his special interest being in the cultivated genera Osmanthus and Jasminum. Mr. Green will continue his studies in this family, while at the same time serving as a horticultural taxonomist. 448 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Dr. George K. Brizicky was appointed jointly with the Gray Herbarium to work with Dr. Wood on the flora of the Southeastern States. Dr. Brizicky replaces Dr. Kenneth A. Wilson, who resigned to accept a posi- tion at San are Valley State College, Northridge, California. Mr. Don M. Jayaweera, director of the Royal Botanic Garden, Peradeniya, oar and the holder of a Rockefeller Foundation Fellow- ship, was appointed as a Research Fellow at the Arboretum for the fiscal year. Mr. Jayaweera has worked with staff members in the herbarium on taxonomic problems of some cultivated and native plants of Ceylon and adjacent areas. He is also revising some of his work on the orchids of Ceylon through the use of the Orchid Herbarium of Oakes Ames, housed in the Harvard University Herbarium building. Mr. Peter Tigerstedt, co-director of the Mustilla Arboretum, Mustilla, Finland, was appointed as the first Mercer Fellow at the Arnold Arboretum for the spring semester. Mr. Tigerstedt is the grandson of the founder of the Mustilla Arboretum who in the early days of the Arnold Arboretum was a correspondent of Charles Sargent. The Mustilla Arboretum, which is the northernmost in Europe, has received many plants from the Arnold Arboretum in past years. Mr. Tigerstedt took part in a variety of the activities of the Arboretum staff to acquaint himself with our methods and procedures. The promotion of Dr. Lorin I. Nevling, Jr., to the position of associate curator of the herbarium was recommended and approved. Dr. Nevling also serves as assistant editor of the Journal of the Arnold Arboretum. It is a special pleasure to record the honors awarded to several staff members. At the fiftieth anniversary of the College of Forestry, at Syracuse University, Irving W. Bailey, professor of plant anatomy, emeritus, was honored with the degree of Doctor of Science, honoris causa, with the following citation: IrvING WIDMER BAILEY ur name is synonymous with the role of re- search in the development of forestry. Your dedication to truth, zeal in research, and capacity to see the unexpected are gifts you have used to enhance and ennoble all the lives you have touched. Your solid and creative scholarship still guides and challenges forestry research. We honor you as forestry’s pre- eminent wood anatomist and botanist and a truly inspired teacher.” Professor Karl Sax, professor of botany, emeritus, was honored on November 11, 1960, with a citation from the American Horticultural Council in annual meeting at Pasadena, California. Special notice was given to the many contributions in cytology and genetics and to the breeding of ornamental plants completed by Dr. Sax while an active staff member of the Arboretum. Dr. Donald Wyman, horticulturist, was elected president of the American Horticultural Society for the coming year. Two other resignations were accepted during the year. Miss Ann Close, business secretary, resigned and was replaced by Miss Anne Woolf. Miss Ann Waterman, assistant librarian, resigned to accept a position as librarian at the ‘Pennsylvania Genie daad Society. a . X yee ; yer = es % ~~ ‘ ae ake i Ses ee: eS. eee sees i Rs 00k 3 = : Architect’s drawing of the proposed headhouse, greenhouse and associated construction to be built at 1050 Centre Street, Jamaica Plain, Massachusetts. 450 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Horticulture: Throughout the history of the Arnold Arboretum various directors have pointed out that in its location in the vicinity of Boston its plants are subjected to one of the most arduous and variable environments of any of the world’s arboreta. Few areas of the world can show the vagaries of climatic conditions which our plants experienced during the past year. summer of subnormal heat and rainfall was followed by the effects of hurricane ‘‘Donna”’ which hit the collections of the Arnold Arboretum on September 12, 1960. Although in comparison with hurricanes of pre- vious years the loss was slight, our plants experienced severe branch damage which kept two tree pruners busy throughout the fall and winter months repairing and protecting the trees and shrubs affected by the strong winds. Although only thirteen trees were broken to a degree that required their removal, as frequently happens, three of these were trees familiar to the general public in conspicuous places on the grounds. The dishgurement of trees through the loss of branches, however, was severe. There are lasting effects of hurricanes in injury to root systems, and the growth of damaged specimens is generally poor in succeeding years. The superintendent and his staff succeeded in pulling to an erect position several dozen trees blown over by the storm. Although we were fortunate in having a mild winter at first, allowing these plants to become re-established, our New England winter demon- strated once again the severity of environmental conditions in the area after the first of the new vear. During February and March the Boston area experienced a_ record-tying cold spell when the temperature in Jamaica Plain did not rise above the freezing mark for twenty-six succes- sive days. Subzero temperatures at night were frequent both during the period and in the days that followed. Fortunately, a heavy snowfall was on the ground during this period, and this, coupled with the heavy mulch which the staff deems desirable around the plants, protected the root systems. As a consequence this spring, irregular flowering or a complete lack of flowering in many plants that form flower buds in late summer and fall, the death of branch systems which necessitated heavy pruning, and the complete loss of many of the more tender species, were found. Throughout the “spring season” the Arboretum collections ap- peared to be behind what is considered the normal sequence of develop- ment. Much remains to be learned from continuing observation and re- cording of environmental conditions and subsequent flowering of plants in the Arboretum collections. The development of new greenhouse facili- ties for the Arboretum should allow not only the experimental duplication of some of these unusual environmental effects but a complete testing of the variations in the handling of such plants as an aid to their survival The approval in principle by the President and Fellows of Harvard University as trustees of the Arnold Arboretum for the construction of a headhouse and new greenhouses for the Arboretum was reported last 1961] THE DIRECTOR’S REPORT 451 year. During the current fiscal year much time and effort have been spent on this development, and it is a pleasure to report that construction is now under way. The firm of Griswold, Boyden, Wylde and Ames, of Boston, was selected as the architects and approved by the Corporation. The plans developed were discussed and checked with the staff members involved in each field, for the new greenhouse will house the propagation staff, a laboratory suitable for morphological, cytological, and genetical investigations, and a conference room for staff meetings and discussions. The main building (the headhouse) is designed to facilitate work in the area of propagation of woody plants, equipped as it will be with cold chambers for experimentation for hardiness, seed dormancy, flower initia- tion, and general problems of plant growth as these are related to the Arboretum’s interests. A small apartment will be located on a second floor to enable a staff member to live on the grounds and be associated with the continual care which nursery and greenhouse collections require. Three greenhouses (with space for a fourth) are provided with automatic controls, insofar as possible, offering several environmental areas for experimentation. A pit house with a mechanism for heating and cooling is capable of extending either a warm fall or a cold spring season and will offer numerous research possibilities. A hexagonal display house for bonsai and a shade house for nursery stock complete the construction plans. The entire area will be surrounded by a chain link fence which hopefully will afford a security lacking in our present nursery area. The new location, to be known as 1050 Centre Street, represents land acquired as three purchases in 1924, 1926, and 1927 from the Adams Nervine Hospital. Purchase at that time was by means of special contri- butions for the purpose. Regrettably, all of the top soil had been removed from one area prior to its sale, and in recent years this land has been of little value. Currently a program is under way to replace the top soil and to build up the fertility of the area for its future use for nursery or display plants. Fortunately, an abundance of top soil, which can be used for regrading the greenhouse area, has been encountered in the excavations for the basement of the headhouse and for the pit house. A new road will be constructed from the Arboretum proper, departing from the existing road system near the upper part of the lilac collection. This entrance will reduce the need to enter or depart from the busy Centre Street section. The new greenhouses will allow a further development of many aspects of the work of the Arboretum staff. The greenhouses will not contain display collections and will not normally be open for public inspection but will function in the scope of the indenture which established the Arnold Arboretum for the growth and study of plants hardy to New England and for the research of the staff. The construction of the green- houses at this time without a special financial drive for construction funds is possible by the use of the income from the bequest of Mrs. William Dana Mercer. A suitable plaque is planned for the interior of the building. A ground-breaking ceremony was held on May 12th during 452 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII a meeting of the Board of Overseers’ Committee to Visit the Arnold Arboretum. Token shovels of soil were lifted by Dr. Bradford Washburn, chairman of the committee; Dr. Nathan M. Pusey, president of Harvard University; Dr. George M. Taylor, director of the Royal Botanic Gardens, Kew, both as a member of the Visiting Committee and in behalf of other arboreta; Mr. Martin Walsh, chairman of the Board of Park Commissioners of the City of Boston; and by Dr. Howard, director, for the Arboretum staff. Construction and occupancy are expected to be completed during the next fiscal year, Breaking ground for the new greenhouses of the Arnold Arboretum, May 12, 1961. Front, left to right: President Nathan M. Pusey, Mr. Martin F. Walsh, Dr. Bradford Washburn, Dr. George Taylor, and Dr. Richard How ee Rear, left to right: Mr. Lincoln Boyden, Jr. (architect), Dr. Donald Wyman, Mr. _ A. Javaweera, Mr. Chester E. Bond (contractor), and me pei aea of the Committee to Visit the Arnold Arboretum, including far right, Mr. Sidney Shurcliff. In a continuation of the annual maintenance program for the Arboretum grounds the Department of Parks and Recreation of the City of Boston rebuilt or replaced all of the concrete and wood-slatted benches in the Arboretum. These reconditioned benches freshly painted a dark green make the grounds more attractive for the visitors. The department was also responsible for painting the gates to the Arboretum and _ for replacement of defective fire hydrants within the grounds It is hoped in future years to continue to resurface many of the roads within the Arboretum and to complete the fencing of the property. Regrettably the Arboretum has never been completely fenced, and theft 1961] THE DIRECTOR’S REPORT 453 and vandalism continue. In spite of increased protection supplied by the Boston Police Department twenty-five choice evergreens were stolen during the year, some within twenty-four hours of being planted. Mis- cellaneous acts of vandalism continue and are annoying and expensive in the demands on staff time. Various steps for better protection are under consideration. Currently the roadway gates at the Arboretum are open from sunrise to sunset and the pedestrian gates continuously, al- though the staff is present only during the usual working hours. A staff police force or a restriction in the visiting hours may be required within the next few years, but a sizeable expenditure for fencing must be made first. During the year about 280 species and cultivars were added to the Jamaica Plain collection from the nursery area in Weston. Included were approximately 60 Rhododendron plants which were added to the steep bank-planting on Hemlock Hill near the South Street gate. Additional selected lilac plants were placed near the walk in the lilac collection to continue a program of revising this collection to display all of the culti- vars considered in the Lilac Committee’s list of 100 best varieties. Addi- tional plantings were made on the outcrop of Roxbury conglomerate known as the “rockery.” It is hoped that dwarf plants and a heavier mulch will aid plants to survive in this difficult yet intriguing area. We were fortunate to continue to receive horse manure for the trucking and have added each week about six truck loads obtained from local stables. Smaller amounts of wood shavings, approximately one truck load per week, are do- nated for the hauling by a local lumber yard, and a truck load of cocoa shells was donated by the Hershey Chocolate Company with delivery sup- lied. The hedge collection located near the Bussey Building has received special attention during the year. It is planned to relocate this collection near the new greenhouses. Ground has been prepared for these plants and the old collection has been root pruned in preparation for moving. In October 1960, the Arboretum staff was notified that legislation filed by Representative James J. Craven, Jr., of Ward 19, had been approved by the General Court. This act provided for a swap of land involving a 17-acre tract of property owned by Harvard but adjacent to Arboretum property, between South Street and the New Haven Railroad for land owned by the Metropolitan District Commission in Randolph, Massachu- setts, and a cash settlement. The Randolph property would then be used by the Medical School with another cash settlement to the Arnold Arboretum. There was considerable publicity during the election cam- paign of the benefits to the public of the playground and picnic area planned adjacent to the Arboretum. Although this transfer of land title may occur eventually, legal considerations have delayed positive action. All of the Salix and Populus species which grow in the marshy tract have been propagated for replanting elsewhere on the grounds. Many of the activities of the propagator and his assistants were con- cerned with the design and the future relocation of the greenhouse. Thus, 454 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII . ae a f Ca Asove: A portion of the collections of on Bussey Hill. 3ELOW: Placing <¢ ate. ke Rhododendron and Cytisus species barrier to protect small plants against sledding. during the year 363 taxa in our general collections were propagated, these representing plants too large to be moved, plants needing replacement due to age or condition, or plants represented by insufficient individuals in our general plantings. In addition, another 103 taxa were propagated 1961] THE DIRECTOR’S REPORT 455 from our general collections in response to specific requests from other gardens or individual scientists. The Arboretum staff members made requests for propagating material of 116 taxa for use in experimental or taxonomic or cytological studies. The propagator has also collaborated with the taxonomists in supplying information on propagation to be pub- lished with taxonomic studies of cultivated plants. For experimental projects involving the hardiness of cuttings or young plants 76 taxa were processed, and 104 taxa as seeds and 49 taxa as cuttings were subjected to experimental procedures for information relative to improving tech- niques of propagation of these taxa. During the fiscal year a total of 270 shipments of plant material comprising 1105 taxa was sent to co- operating nurseries and other arboreta in the United States and twelve other countries. Of these, 190 were sent as seeds and 915 as cuttings or plants, indicating the growing need and desire for reliable propagating material to represent the taxon and the lessening interest in seeds with their possibilities for hybrid contamination. From other sources repre- senting the United States and 21 other countries, the Arboretum re- ceived 227 shipments of living plants or propagating materials comprising 705 taxa and 47 shipments of seeds of 118 taxa. Experimental work conducted in the greenhouses concerned the germina- tion of various kinds of seeds, the rooting of cuttings or plants usually propagated by grafting or considered difficult to root, and additional work on the length of storage time that cuttings of different taxa may be held under different conditions. In co-operation with the work of Mrs. Claude Weber on Chaenomeles, Mr. Fordham determined that a stratification of two months at 41°F. was the best method for producing a uniform stand of seedlings most quickly. Seeds of the fringe tree (Chionanthus retusus) responded most satisfactorily to three months of warm stratification, followed by three months of cold stratification at 41°F. A similar treatment applied to seeds of Hamamelis mollis proved equally successful to the five month warm period usually recom- mended. Special attention has been given to several members of the Hamameli- daceae. Experimental work with the rooting of cuttings of Parrotia showed that grafting these plants as most nurseries do traditionally is not necessary. Cuttings taken from a 79-year-old specimen of Parrotia persica in the Arboretum collection on June 23, treated with Hormodin No. 2 or No. 3, and maintained with bottom heat of 73°F. under plastic had rooted 100% by September 10. A control lot without hormone rooted only 14%. Rooting at this time was heavy and the plants could have been removed and potted earlier. The potted plants produced addi- tional growth before going dormant, and all survived the winter, an unusual condition for Parrotia which has a reputation for poor survival of rooted cuttings. The house at 383 South Street formerly occupied by Professor Johnston was renovated, including modernization of the plumbing, painting of the exterior, and roofing, as a portion of our program of building maintenance. 456 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Case Estates: During the year an increase in the number of visitors to the Case Estates in Weston has become evident. The annual “open house” held in early May was well publicized to the “Friends of the Arboretum” and in the papers of adjacent communities. The attendance was gratifying to the staff members on duty there to answer questions. Many of the visitors have made repeated visits for subsequent study. The ground cover plots and the perennial garden vie for principal interest. The Case Estates remain primarily a nursery area for the Arboretum, how- ever. To increase the beauty of the grounds, surplus plants have been placed where they will be seen better by the visitors or those driving by. The spectacular flowering of Malus ‘Henrietta Crosby’ this year created traffic tie-ups on Wellesley Street as drivers slowed to admire this hand- some row of ornamental apples. We anticipate that the developing plant- ings near the High School and along Ash Street will soon draw comparable interest. This spring saw the first flowering of the test plots of Narcissus taxa planted by Dr. Helen Scorgie and other members of the New England section of the American Daffodil Society. Approximately a hundred varieties were in flower during the “open house,’ and Dr. Scorgie was kind enough to answer the many questions posed by the admirers of these plants. An exhibit of Christmas decorations and the plant materials used in making them. Staged at Horticultural Hall, Boston, Massachusetts, November 28- December 2, 19 1961] THE DIRECTOR’S REPORT 457 It was from the nursery areas that plants were sent to the ‘‘co-operating nurserymen” during the year. Young plants of Buxus sempervirens ‘Vardar Valley, Cornus pumila, Euonymus europaeus ‘Red Cascade,’ Rhododendron ‘Mandarin Red,’ Sorbus cashmiriana, and Viburnum pli- catum ‘Lanarth’ were made available to 90 nurseries and arboreta which requested plants from a circulated list. Other surplus plants were again made available to the Department of Buildings and Grounds of Harvard University for plantings around university buildings in Cambridge. A large general collection was also given to Union College, Schenectady, New York, to increase the small arboretum maintained on that campus. It has been pointed out previously that the Case Estates in Weston experience more severe cold than does the Arboretum proper in Jamaica Plain. During the past winter, specimens of Deutzia showed comparable die-back in each area due to the cold, but many clones of Rhododendron fortunei suffered more in Weston than did comparable plants in Jamaica Plain. The wildlife of the area is again on the increase and the destructive effects of rabbits and woodchucks were evident after the winter snow disappeared. Deer have been seen on the property on two occasions; pheasants are numerous; and a covey of quail appeared again for the first time in several years. Land was again made available to staff members of the Bussey Institu- tion and the Cabot Foundation of Harvard for growing experimental plants in Weston. Education: Only one formal course was offered during the spring semester when Dr. Howard taught a class in horticultural taxonomy in Harvard College. Dr. Thomas offered special work in cytotaxonomy to a group of graduate students working both in Cambridge and with the living collections in Jamaica Plain. The staff participated in the regular series of weekly seminars on taxonomic subjects. Four classes were offered to the general public at Jamaica Plain or Weston in the fall, two field classes in the spring, and a class in propaga- tion which met at appropriate intervals during the year. During the spring semester a lecture series was offered evenings in the Administra- tion Building and was well attended. These lectures, designed to be semitechnical in nature, were presented by various staff members and research fellows. The usual service of guided tours of the Arboretum plantings both in Jamaica Plain and Weston was accepted by many garden clubs, women’s clubs, and civic groups during the fall and spring seasons. The spring tour offered jointly by the Massachusetts Horticultural Society and the Arboretum drew a large crowd which was accommodated in eight busses. A staff member in each bus led the groups through some of the more floriferous areas of the Arboretum while explaining the plants and 458 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII answering questions. These trips seem to be particularly worth while, for a large percentage of the people return for more leisurely walks alone or with smaller groups. Arboretum staff members spoke to many garden clubs, horticultural groups, and similar organizations during the year. In addition, a set of kodachrome slides of the plants and activities at the Arboretum with a printed commentary was made available to other groups for their own programming. Dr. Wyman was the principal speaker at the thirtieth anniversary luncheon of the Holden Arboretum in Cleveland, Ohio. Dr. Wagenknecht described registration procedures at the organizational meet- ing of the American Boxwood Society in Boyce, Virginia, and at the annual meeting of the New England section of the American Society of Horticultural Science. Mr. Fordham presented a paper on double-dormant seeds at the Plant Propagators’ annual meeting at Cleveland. Dr. Wood presented a series of lectures in a program in plant evolution in the De- partment of Biology at Vanderbilt University, Nashville, Tennessee. Drs. Perry and Hu presented papers at the annual meeting of the Society for Economic Botany held in Boston. Dr. Howard spoke on the campuses of Union College, in Schenectady, New York, and Mount Mercy College, in Pittsburgh, Pennsylvania, under the sponsorship of the American Insti- tute of Biological Sciences. He also opened the lecture series at Longwood Gardens, was a speaker at the annual meeting of the New England Nurseryman’s Association, and presented the commencement address at the graduation exercises at the State College, Bridgewater, Massachusetts. Exhibits and Displays: Five flower-show exhibits were prepared by the Arboretum staff during the winter and spring. A display of Christmas decorations and the plants comprising them, formerly held in the Administration Building in Jamaica Plain, was staged at Horticultural Hall, Boston, in co-operation with the Massachusetts Horticultural Society. When held previously at Jamaica Plain the show had had only a moderate attendance due to our more isolated location and the inadequate local transportation. Horticultural Hall proved to be an excellent location, however, and the exhibit was well attended. Members of several garden clubs under the leadership of Mrs. Donald Wyman prepared swags, wreaths, decorations, and arrangements, while the Arboretum staff prepared specimens and descrip- tive data of the plant materials involved. The Arboretum exhibit at the Massachusetts Horticultural Society’s Spring Flower Show showed numerous cultivated plants in their area of origin. A Mercator projection map of the world, 40 feet by 10 feet allowed small plants to be placed in that section of the world where each plant was native. One hundred and seventy-eight plants could be ac- commodated in this educational exhibit. It was awarded a first prize and a cold medal. The Arboretum’s tree pruning exhibit was set up at the International Flower Show of the New York Horticultural Society held in the Colli- 1961] THE DIRECTOR’S REPORT 459 THE ARNOLD ARBORETUM SOME CULTIVATED PLANTS IN THEIR PLACES OF ORIGIN Apove: The Arnold Arboretum exhibit at the Spring Flower show of the ‘awe neue Horticultural Society, Boston, Massachusetts, March 11-19, 196 BeLow: The Arnold Arboretum exhibit at the International Flower Show, New York City, March 4-12, ite 460 JOURNAL OF THE ARNOLD ARBORETUM [| VOL. XLII seum. Mr. Robert Williams, superintendent of buildings and grounds, receives full credit for this informative exhibit which won a first prize, a gold medal, and the T. A. Weston trophy for the most educational exhibit in the show. The Arnold Arboretum bonsai collection was displayed in Detroit, Michigan, for the flower show of the Michigan Horticultural Society. The dwarf plants displayed in an oriental setting won a gold rosette, the highest award of the show. Finally, in late spring, the New World portion of the exhibit used for the Massachusetts Horticultural Society show was reconstituted using later flowering materials and was exhibited by request at the flower show sponsored by the Federated Garden Clubs of Massachusetts at the Jordan Marsh Company, Boston. Comparative Morphology: Professor I. W. Bailey served again this year as curator of the wood collection. His continuing care and interest are greatly appreciated. Dur- ing the year a sizeable number of wood samples were made available to the Quartermaster Research and Engineering Command, U.S. Army, for a specific chemical analysis of cellulase inhibitors, a program of research conducted at their headquarters in Natick, Massachusetts. In addition, requests for wood samples of various genera and families were received from fourteen individuals and institutions and all were filled as completely as possible. The research program of Professor Bailey, supported in part by a grant from the National Science Foundation, continues to concern the putatively primitive leaf-bearing genera of the Cactaceae. Papers summarizing his research will continue to appear in the Journal of the Arnold Arboretum. Dr. Uttam Prakash, of the Birbal Sahni Institute of Palaeobotany, Lucknow, India, has been a visiting scholar for a second year and has continued to use the wood and slide collections of the Arboretum in connection with his investigations of fossil floras of the state of Wash- ington. Library: The attention of the librarians was devoted in considerable extent to a reorganization of the horticultural section of the library located in Jamaica Plain. The books formerly shelved on the second floor of the herbarium were moved to the third floor and rearranged on the shelves in the main reading room and in the alcoves, thereby freeing the second floor space for an increase in the herbarium of cultivated plants. Three hundred and twenty-three volumes were added to the library, in- creasing the total to 51,106 on June 30, 1961. Nine hundred and seventy- two pamphlets were added to the monographic collection, making a total of 18,302. A total of 2,182 catalogue cards was added to the main file. Two hundred and seven books were bound or rebound, and fifty old 1961] THE DIRECTOR’S REPORT 461 books given special restoration service. Miscellaneous indices also received their regular accessions. Three thousand cards were added to the Gray Herbarium Card Index of American Plants, 4,000 cards to the Torrey Index of American Botanical Literature, and approximately 700 cards to the Rehder index of cultivated plants. Sets 12-14 of the Index Nomi- num Genericorum were also filed. Japanese bonsai of the Lars Anderson Collection of the Arnold Arboretum on exhibition at the Flower Show of the Michigan Horticultural Society, Detroit, Michigan, February 25—March 5, 1961. Although “Xerox” reproduction service became available through the Widener Library and was recommended whenever possible, 120 books nevertheless were sent out on Interlibrary loan. Twenty-five books were borrowed from other libraries. Herbarium: During the year 12,208 specimens were mounted and added to the herbarium, bringing the total collection to 726,344 specimens on June 30, 1961. During the same period 17,232 specimens were received through exchange, gift, subsidy, or for identification. This year the largest number represented a subsidized collection made by Dr. Hugh M. Raup in the Mackenzie Basin of Canada in 1938. A sizeable collection of native and cultivated plants of Japan was purchased from Mr. Miyoshi Furuse. A total of 3,345 specimens was distributed to other institutions in exchange. The staff filled 113 requests for herbarium specimens on loan, repre- senting 14,614 specimens sent to 65 institutions, forty-three of these in the United States. The staff requested 65 loans totaling 3,399 specimens 462 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII for study from 38 institutions, including twelve in the United States. The incoming loans averaged 52 specimens, while the outgoing loans averaged 129 specimens. The additions to the herbarium of cultivated plants housed in Jamaica Plain have now reached the point where many of the cases are crowded. Fortunately, space for expansion is available, and new steel herbarium cases have been ordered. A general shift of this herbarium is being planned. Much of the material added to the cultivated herbarium repre- sents cultivated herbaceous plants or exotic greenhouse plants. We appre- ciate very much a large exchange of specimens from the comprehensive collections of the Longwood Gardens, nearly all of which represent new genera or species for our herbarium. This material will assist materially in the routine identifications of many of the specimens sent to us for naming. The publications of the taxonomists cited in the bibliography of the staff clearly indicate the diversity of area and plant groups represented in their research. Travel and exploration: The Arboretum was represented at various annual science meetings in several parts of the country by staff members. These meetings included those of the American Association of Botanical Gardens and Arboretums, the American Institute of Biological Sciences, the American Horticultural Council and American Horticultural Society, the Plant Propagators So- ciety, the Society for Economic Botany, the National Shade Tree Con- ference and the American Society of Horticultural Science. Field work was undertaken by Dr. Howard, in Montserrat, St. Croix, and Redonda; by Dr. Wood, in central Florida with Dr. George R. Cooley; by Dr. Wagenknecht, in Virginia and North Carolina; and by Dr. Thomas, in Florida and Alabama. In all cases, herbarium specimens and living plant materials were brought back for addition to the Ar- boretum collections. Dr. Kobuski spent two months in Europe visiting herbaria in Scotland, England, Switzerland, and Sweden in connection with his principal re- search interest in the Theaceae and was engaged especially in the examina- tion of types and other herbarium materials of the Old World species of Ternstroemia which he is revising. The Arboretum has also contributed to the support of several collectors working in Japan; to an expedition of Mr. James Keenan, of the Royal Botanic Garden, Edinburgh, to Burma; and to the work of Mr. Theodore Dudley, in Turkey. These areas may yield plants of horticultural value to the Arboretum and specimens for the herbarium. Gifts and grants: The annual appeal to the Friends of the Arnold Arboretum received its usual generous response. To all of these contributors we extend the 1961] THE DIRECTOR’S REPORT 463 appreciation of the staff for the support which makes possible extra services to the area of horticulture. A special gift from Dr. George R. Cooley supports the work of Dr. Wood and his collaborators on the generic flora of the southeastern states. During the year we have received an unusually large number of books, pamphlets, and research materials for the library from the estate of Dr. Franklin P. Metcalfe, and from Mrs. Susan D. McKelvey, Mr. P. Bernat, Dr. H. Field, and Dr. A. F. Hill. These contributions materially increase the value of our library and pamphlet collections. Plant materials have been received in exchange from many sources, but the gift of living plants for the education classes received from Mr. Robert Pirie, of Hamilton, Massachusetts, is particularly appreciated. I also wish to acknowledge the assistance of the many individuals who have supplied propagating material, photographs and information in re- sponse to specific requests for research projects of the staff. The largest number of requests concerned the monograph of Chaenomeles being pre- pared by Mrs. Claude Weber. Publications: The regular publications of the Arnold Arboretum include the quarterly Journal of the Arnold Arboretum and Arnoldia, which is issued at irregular intervals. Both appeared as planned. A double number of Arnoldia issued during December, 1960, comprised a revision by Dr. Wyman of his earlier publication, “How to Establish an Arboretum or Botanical Garden.” The first edition of this article is out of print, and the demand for reprints of the revised edition exceeded all expectations. Five numbers of Arnoldia dealt with the registration lists of cultivar names and the problems of preparing such lists. The council of the American Association of Botanical Gardens and Arboretums requested 175 extra copies of each list for distribution to its membership, and many additional requests for reprints from nurserymen indicate the interest in these bibliographic listings. To the present, registration lists for cultivar names in Cornus, Gleditsia, Forsythia, Liquidambar, and Pieris have been published. The librarian was also responsible for the photoprint reproduction and the distribution of Shaw’s Pines of Mexico. Published originally by the Arboretum in a limited edition, this work has long been out of print. To date, the sales of the photoprint reproduction have exceeded the distribu- tion of the original printing. This report would not be complete without acknowledging the efforts of Miss Ethel Upham, of East Woodstock, Connecticut, who for many years has painstakingly prepared the index for the Journal of the Arnold Arboretum. Miss Upham has now relinquished this arduous but invaluable task which passes from her capable hands to others. 464 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XLII Bibliography of the Published Writings of the Staff and Students July 1, 1960 — June 30, 1961 Bartey, Irvinc W. Comparative anatomy of the leaf- bearing Cactaceae, I. oliar vasculature of Pereskia, Pereskiopsis and Quiabentia. Jour. Arnold Arb. 41: 341-356. pl. 1-7. 1960. Comparative anatomy of the leaf-bearing Cactaceae, II. Structure and distribution of sclerenchyma in the phloem of rae Pereskiopsis and Quiabentia. Jour, Arnold Arb. 42: 144-156. pl. BrizicKy, GEORGE K. The genera of Turneraceae and Sao in the a eae States. Jour. Arnold Arb. 42: 204-218. ForDHAM, ALFRED J. Germination of ae julibrissin rosea seeds, Rhode Island Sianital s Newsletter 5: 6. 1961 Germination of aeiligdtce seeds. Plant Propagator 7(1): 4, 6. 1961. Metasequoia glyptostroboides (dawn redwood) and its propagation. Plant Propagator 6(4): 7, 8. 1960. Propagation of Liguidambar styraciflua. Arnoldia 21: 66. 1961. . Propagation of Parrotia persica from cuttings. Plant Propagator 6(4): 8,9. 1960. ———.. Propagation of woody plants by seed. Arnoldia 20: 33-40. pl. 9. 1960. GREEN, PETER S. Further notes on Malaysian Osmanthus, Notes Bot. Gard. Edinb. 23; 175-178. 60 Studies in the genus Jasminum, I. Section Alternifolia. Notes Bot. Gard. Edinb. 23: 355-384. Howarp, RICHARD A. Bibliographic data on the Hillcrest Gardens Books, 1911-1941. Jour. Arnold Arb. 41: 318, 319. The botanical results of the U. S. Commission of Inquiry to Santo Domingo in 1871. Jour. Arnold Arb. 42: 115-143. 1961. ———.. Coccoloba P. Br. ex L. In: Woodson, R. E., et al. Flora of Panama. Ann. Missouri Bot. Gard. 47: 340-353. 1961. ———. Concerning the registration of cultivar names. Arnoldia 21: 1-8. 1961. . The Director’s report. The Arnold Arboretum during the fiscal year ended June 30, 1960. Jour. Arnold Arb. 41: 429-445. 1960. The Hillcrest Gardens, Weston, Massachusetts. Arnoldia 20: 53-67. 60. pl. 13, 14. 19 . How to read the label. Newsletter (Garden Club. Fed. Mass.) 12: 22-24. 1961. Ivan Murray Johnston, 1898-1960. Jour. Arnold Arb. 42: 1-9. port. 1961. et al. Ivan Murray Johnston. Harvard Univ. Gaz. 56: 216, 217. 1961. pore In: Gray, P., ed. Encyclopedia of Biological Sciences, pp. 583, 584. 1961. OA aides lists of cultivar names in Cornus L. Arnoldia 21: 9-18. l. Studies in the genus Coccoloba, IX. A critique on the South American species. ‘Jour. Arnold Arb. 41: 231-258, 357-390. 1960. ———. Studies in the genus Coccoloba, X. New species and a summary of distribution in South America. Jour. Arnold Arb. 42: 87-95. 1961 Studies in the genus Coccoloba, XI. Notes on the species in Asia. Jour. Arnold Arb. 42: 107-109. 1961. 1961] THE DIRECTOR’S REPORT 465 Thrinax rex —a regal palm of Jamaica. Principes 4: 133- 137. 1960. _ Why Montserrat? Jn: Brown, John, ed. Leewards, pp. 12-16. Advocate Go. Barbados. 61. Hu, Suru-yinc. The economic botany of the Paulownias. Econ. Bot. 15: 11-27. 1961. Jayvaweera, D. M. A. Palms in the Royal Botanic Gardens, Peradeniya, Ceylon. Principes 5: 53-59. 1 + JOHNSTON, Ivan M. Notes on some Texan Borages. (Studies in the Bo- raginaceae 31.) Wrightia 2: 158-162. 1961. Kopuskt, CLARENCE E. Studies in the Theaceae, XXX. The African species of Ternstroemia. Jour. Arnold Arb. 42: 81-86. 1 - Studies in the Theaceae, XXXI. A new species of Adinandra from the Celebes. Jour. Arnold Arb. 42: 112, 113. — Lorin I., Jr. Aizoaceae. Jn: Woodson, R. E., ‘of al, Flora of Panama. _ Missouri Bot. Gard. 48: 80-85. 1961. ace In: Woodson, R. E., et al. Flora of Panama. Ann. Missouri Bot. Gard. 47; 303-308. 1960. Chloranthaceae. Jn: Woodson, R. E., ef al. Flora of Panama. Ann. Missouri Bot. Gard. 47: 81-83. 19€0. -. Corylaceae. Jn: Woodson, R. E., et al. Flora of Panama. Ann. Missouri Bot. Gard. 47: 93, 94. 1960. Lacistemaceae. Jn: Woodson, R. E., et al. Flora of Panama. Ann. . Myricaceae. Jn: oe R. E., et al. Flora of Panama. Ann. Missouri Bot. Gard. 47: 88, 960. _ Nomenclatural changes in ag ere (Thymelaeaceae). Jour. Arnold Arb. 41: 412, 413 O -— Olacaceae. Jn: Woodson, R. E., ef al. Flora of Panama. Ann. Missouri Bot. Gard. 47: 293-302. 1960. _ Opiliaceae. Zn: Woodson, R. E., e¢ al, Flora of Panama. Ann. Missouri Bot. Gard. 47: 291, 292. 1960. — Portulacaceae. Jn: Woodson, R. E., et al. Flora of Panama. Ann. Missouri Bot. Gard. 48: 85-89. 1961. _s proteaceae. Jn: Woodson, R. E., e¢ al. Flora of Panama. Ann. Missouri Bot. Gard. 47: 199-203. 1960. _ Ulmaceae. Jn: Woodson, R. E., ef al. Flora of Panama. Ann. Missouri Bot. Gard. 47: 105-113. 1960 Sax. Kary. Radiation sensitivity of Tradescantia chromosomes to a second exposure of X-rays. Radiation Research 14: 667-673. 1961. ScHWARTEN, LazELLA. Bibliography. Jn: Howard, Richard A., Ivan Murray Johnston, 1898-1960. Jour. Arnold Arb. 42: 4-9. 1961. Index to American botanical literature. Bull. Torrey Bot. Club 87: age 361-374. 430-440. 1960; 88: 66-76. 1961. with Harotp W. RICKETT). Abbreviations of titles of serials cited by eee Supplement 1. Bull. Torrey Bot. Club 88: 1-10. 61. Tuomas, Joan L. The cytology of some cultivated species of Viburnum. Jour. Arnold Arb. 42: 157-164. _ Forsythia ‘Karl Sax’. peer 20: 49, 50. pl. 12. 1960. ____. The genera of the Cyrillaceae and Clethraceae of the southeastern United States. Jour. Arnold Arb. 42: 96-106. 1961 _ Liguidambar. Arnoldia 21: 59-65. pl. 7, 8. 1961. 466 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII . Schizocardia belizensis: A species “a Pbelveae ene from ee America. Jour. Arnold Arb. 42: 19 WAGENKNECHT, Burbertte L. ana i Siar names in Gleditsia L. Arnoldia 21: 31-34. 196 . Registration lists of cites names in the genus Pieris D. Don. Arnoldia 21: 47-50. 1961. . The tree legumes in the Arnold Arboretum. Arnoldia 21: 19-30. pl. 1-3. 1961. ——. uae is “The Yellow Rose of Texas?” Am. Rose Annual 46: 196 WEBER, fen (with SHirteEy C. Tucker). Flowers and botanical subjects on stamps, including fruits, trees, vegetables, seeds, ferns, etc., all forms of plant life on stamps. Topical Handb. 30: 1-162. illus. Milwaukee, 1960. Eduard Riibel, 1876-1960. Pl. Sci. Bull. 7(1): 6. 1961 Witson, KenNnetH A. The genera of Convolvulaceae_ in the southeastern United States. Jour. Arnold Arb. 41: 298-317. . The genera of Myrtaceae in the southeastern United States. Jour. Arnold Arb. 41; 270-278. 1960. Woop, Carroit E., Jr. The genera of Ericaceae in the southeastern United States. Jour. Arnold Arb. 42; 10-80, 1961. . A study of hybridization in Downingia (Campanulaceae). Jour. Arnold Arb, 42: 219-262. 61. Wyman, Donatp. The best ornamental spireas. Arnoldia 21: 51-58. pl. 5, 6. 1961. ‘sda forms of Potentilla. Am. Nurseryman 113(5): 16, 17; 93-96. Sinuadar Am. Nurseryman 112(5): 12, 13, 50-56. 1960. The Cotoneasters. Bull. Hort. Soc. New York 11(2); 6-8. 1961. ———. Forsythias. Am. Hort. Mag. 40: 190-197. 19 fe . The Forsythia story. Arnoldia 21: 35-38. Graceful birches popular. Am. Nurseryman eee 20, 21, 65-72. 61 Hardy honeysuckles. Am. Nurseryman 112(3): 16, 17, 93-100. 1960. Hemlock species. Am. Nurseryman 112(11): 12, 13, 48, 49. 1960. How to establish an arboretum or botanical garden. Arnoldia 20: 69-83. 1961. ———. Hurricane “Donna”. Arnoldia 20: 47, 48. 1960. Ilex crenata and its varieties. Arnoldia 20: 41-46. pl. 10, 11. 1960. Interest in hollies. Am. Nurseryman 112(9): 12. 13; 116-124. 1960. ———. Japanese hollies can give a boxwood effect. Horticulture 39: 141, 142. 961. . Lindens important as shade trees. Am. Nurseryman 112(1): 22, 23, 76-82. 1961. Low junipers. Flower Grower 48: 38-41. 1961. Majestic beech requires space. Am. Nurseryman 113(1): 16, 17, 112, 1961. 1 ———. Many forms of four arborvitae species. Am. Nurseryman 113(7): 1 11, 67-79. 1961. : ape aes of flowering quince. Am. Nurseryman 113(9): 11, 12, 89-99. eae fruits, 1960. Arnoldia 20: 52. 1960. THE DIRECTOR’S REPORT 467 Popular hedges in list of barberries. Am. Nurseryman 112(7): 10, 11, 107-110. 1960. Registration lists of cultivar names of forsythias. Arnoldia 21: 39-42. 1961. . Showy wisteria still a problem vine. Am. Nurseryman 113(11): 10,11, 68-72. 1. Winter injury not severe. Arnoldia 21: 43-46. pl. 4. 1961. RicHuarp A. Howarp, Director 468 JOURNAL OF THE ARNOLD ARBORETUM | VOL. XLII Staff of the Arnold Arboretum 1960-1961 RicHARD ALDEN Howarp, Ph.D., Arnold Professor of Botany, Professor of Dendrology, and Director. IRVING WipMER Baltey, S.D., Professor of Plant Anatomy, Emeritus. JosepH Horace FAutt, Ph.D., Professor of Forest Pathology, Emeritus. Kar Sax, S.D., Professor of Botany, Emeritus. GEORGE KONSTANTINE BrizicKy, R.N. Dr., Botanist, Southeastern Flora Project.* MicHaAEL ANTHONY Canoso, M.S., Curatorial Assistant.* ANN Crose, A.B., Business Secretary.+ Henry Draper, Superintendent, Case Estates. ALFRED JAMES ForpHAM, Propagator. PETER SHAW GREEN, B.S., Horticultural Taxonomist. HEMAN ARTHUR Howarp, Assistant Horticulturist. SHIU-YING Hu, Ph.D., Botanist. Don MartTIN ARTHUR JAYAWEERA, M.A., Research Fellow. CLARENCE EMMEREN KosuskI, Ph.D., Curator.* MARGARET CATHERINE LEFAVouR, Herbarium Secretary. SusAN DELANO McKELvey, A.B., Research Associate. Lorin Ives NEVLING, Jr., Ph.D., Assistant Curator. Lity May Perry, Ph.D., Botanist. LAZELLA SCHWARTEN, Librarian.* JoaB LANGSTON Tuomas, Ph.D., Cytotaxonomist. PETER MAGNUS ALEXANDER TIGERSTEDT, A.B., Mercer Fellow. BuRDETTE LEWIs WAGENKNECHT, Ph.D., Horticultural Taxonomist. ANN H. WATERMAN, M.S., Assistant Librarian. ROBERT GEROW WILIAMS, B.S., Superintendent. KENNETH ALLEN WILSON, Ph.D., Botanist, Southeastern Flora Project.*+ CARROLL Emory Woop, Jr., Ph.D., Associate Curator and Editor. ANNE TUCKER Woo tr, A.B., Business Secretary. DonaLtp Wyman, Ph.D., Horticulturist. * Appointed jointly with the Gray Herbarium. + Resigned during the year. 1961] INDEX INDEX Abies nobilis, 420 Acer beckianum, 169-170 — olearyi, 170-171 — puratanum, 166-169 Acmopyle cae Podocarpus dacrydioides, Aberration in. Coniferous Pollen Types of the Southern Hemisphere, I, 416 Acmopyle pancheri, 417 Adinandra from the Celebes, A New Studies in the Theaceae, 270 — urdanatensis, 268 Adinandrella congolensis, 82 sa Lid Fibers in the Secondary Xylem Ae schynomene cf. aspera, 442 442 443 eae beanie 349 Agarista, 39 Agave interm , 133 Albizzia manner 178-180 Albizzioxylon, 180 5 Z Ampelohamnus 52 Amphani rea hy isan, 137 — paucifloscula, 1 ANDRESEN, JOHN 7 and Joun H. BEa- A New Species of Pinus from — sect. Aare 48 eaten deus. 48 Anatomy of Ay cophions dnciptienln: 282 zalea, 30 Azaleastrum, 30 cian 436 Bal , I. W. Comparative 7 o of ke Tee bearing Cactaceae, II. Struc- Cry tals | in Pereskia, Pereskiopsis and Guiana. 334 7 and U. PRAKASH. e Co- lumbia Basalts of Central Washington, 1653-11, Batodendron, 69 Batohan-no-mahingtig, 2 BEeEAMAN, JOHN H. and Joun W. ANDRE- New Species of Pinus from 8 - Racemosae, 19 a. oy Betula, nilotaphy, Ivan Murray Johnston, 4 Bi 30 Black apple, 436 Blue ae 9 Boalau au Botanical pee of the U. S. Comm nqu iry to Santo eer in Ze e, Bae: GeEoRGE K. Genera of d Passifloraceae in the Brizicky, GrEorGE K. The Genera of Violaceae in the Southeastern United E Synopsis of the Genus Columellia (C ne ae 363 BruMMEL, H., 115 Buckwheat tree, 101 Buxella, 77 Cactaceae, Comparative Anatomy of the Leaf-bearing, IJ. Structure and Distri- bution of Sclerenchyma in the Phloem 470 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII of Pereskia, ea yn and Quiaben- Coccoloba conduplicata, 95 Ill. For 4 tia, 144; and Distribution — confusa of Crystals in Pereski, Pereskiopsis — cordata, 94 nd Quiabentia, 33 — coronata, 95 Campanulaceae, 2 —crispata, 108 Cannabis, — cruegeri, 9 pein: erie 297 — cujabensis, 94 Car cymosa, 108 —- tho 176-178 — declinata, 94 Cassandra, 56 — densifrons, 94 Caytonanthus arberi, 420 — dubia, Celtis, — duckei, Celto 187 — dugandiana, 95 Cercidiphyllum, 176 — excelsa, Cestru arian 138 — fagifolia, 95 Chamaedaphne, — fallax, 95 Chaptalia membranace, 137 — fastigiata, 95 — primulacea, 13 — filipes, 95 Chops ee 282 — gardneri, 95 — fue — glaziovii, 94 Chilictrichiorsis keidelii, 282 — grandiflora, 94 — ledifoli — gymnorrhachis, 95 Chiliotrichium diffusum, 282 — ilheensis, 94 — rosamarinifolium, 2 — indica, 10. Chimaphila, 61-62 —] Chiogenes, 42 — lanceolata, 94 Cladothamnus, 22 — latifolia, 94 Clethra, 20, 103-106 — laurifolia, 95 — sect. Clethra, 104 — lehmannii, 95 — sect. Cuellaria, 104 — lepidota, 9 — sect. Euclethra, 104 — llewelynii, 89-90 — ser. Alnifoliae, 104 — longipes, 94 Clethraceae of the Southeastern United — lucidula, 95 States, The Genera of the Cyrillaceae — marginata, 94 and, 96 — meissneriana, 94 yera japonica montana, 270 — mollis, 94 Cliftonia, 101-102 — mosenii, 95 Coccoloba, Studies in the Genus, X. — nitida, 95 New Species and A Summary of Dis- — ob vata, 95 tribution in South America, 87; XI. — obtusifolia, 95 Notes on the Species in ers 107 — ochreolata, 94 Coccoloba achrostichoides, 9 — orinocana, 90 — acuminata, 9 — ovata, 94 —adpressa, 107 — padi — alagoensis, 94 — paraguariensis, 94 — alnifolia, 94 — parime — arboresce — peltata, — argentinensis, 94 — persicaria, 94 — asiatica, — peruviana, — australis, 107 — pipericarpa, 94 —b , 89 — plantaginea, 94 — brasiliensis, 94 — platyclada, 108-109 Se 95, 109 — portuguesana, 90-91 — cerifera, 9 — ramosissima, 94 — chartstahya, 95 — rigida, 9 — chita, — rosea, 94 — a. 88-89 — ruiziana, 95 1961] Coccoloba salicifolia, 95 Columbia ‘ae of Central Washington, sea Fossil Woods from the, 165; 13 Colinas ae A Synopsis of the Gen Columellia er 369 — arborescen 67 — subsessilis, 369-370 weberbaueri, 370 Galuncliiacese 363 Commission of Inquiry to Santo Dcemingo 871, The Botanical Results of the S:, 115 Comocladia acuminata, 136 — cuneata, 136 Comparative Anatomy of the Leaf- bear- of Crystals in Pereskia, Pereskiopsis and Quiabentia, 334 Compositae—Astereae, Leaf and Nodal A y of Some Andean, 276 Coniferous Pollen Types of the Southern misphere, I. Aberration in Acmopyle and Podocarpus Kner 416 Coprosma macrocarpa, INDEX 471 Cordaianthus, 4 Correct Names rae “Diospyros ebenaster,” . Coniferous Pollen Types of the Southern Hemisphere, I. Aberration in Acmopyle and Podocar- us dacrydioides, 416 Crossotheca, 416 Cryptophila, 64 Crystals in Pereskia, Pereskiopsis and Quiabentia, Form and Distribution of. Comparative sap Ri the Leaf- bearing Cactaceae, III, Cultivated Species of "Viburnum, Cytology of Some, 1 Cuphea micrantha, me The Cyrilla, 98-100 Cyrillaceae and Clethraceae of the South- eastern United States, The Genera of the, 96 Cyrilaceae, bat Cyt f Some “Culdivated Species of a The, 157 Dacrydium cupressinum, 418 4 s, 48 ebenaster,” The Correct er he brasiliensis, 432 — digy 4 —- ei 430, 432 — laurifolia, 434 — membranacea, 434 — nigra, — obtusifolia, 434 —revoluta, 431, 434 4 — washingtoniana, 171-172 Director’s Report, The. Arnold Arbore- tum During the Fiscal Year Ended June 30, 1961, 447 Downingia (Campanulaceae), A Study of Hybridization in, Downingia bicornuta bicornuta, 226 — picta, ——var., 230 472 Downingia concolor brevior, 232 — concolor , — cuspidata — elegans brachypetal, 234 — insignis, ——x mee 246 — pusilla, 245 — willamettensis, 236 —yina major, 23 Drypetes i incurva, 141 Dupinia, 81 Eaton, D. C., 125 23 Enkleia (Thymelaeaceae), A Revision of the Asiatic Genus, 373 Enkleia, 295, 320, 384~385 — coriacea, 388 — malaccensis, 388-391 88 Epigaea, 58-60 Erica, Reiareac in the aerial United States, The Genera of Ericaceae, 1 bfam. Arbutoideae, 12 — subfam. Ericoideae, 18 — subfam. car agit 64 . Pyroloidea he a ae 12 , 6 ro = Erythrochiton, 81 Eubotrys, 39 Euptelea, 183 Fagopyrum cymosum, 108 Fagus, 183 False mangosteen, 433 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Fibers in es ream Xylem of Aes- Floral Mercksiok and ae of Kingdonia uniflora, The, Fossil Woods from the Baim Basalts of Central Washington, Miocene, 165; I, 34 Foster, ADRIANC ee gy E S. The Floral Mor- and Relationships of King- uniflora, 397 Eee 420 Garamansatai, 271 ,7 Genera of the Cyrillacéas and Clethraceae oe the Southeastern United States, The, Pat era of Ericaceae in the Southeastern 10 Genera of Turneraceae and Passifloraceae in the Southeastern United States, The, 204 Genera of Violaceae in the Southeastern United States, The, 321 Gleditsia columbiana, 180-182 Gnidia sericea, 310 Gochnatia oligantha, 137 Gray, Asa, 123 Guettarda tae 141 Gymnocladus, Heath Family, 10 Homalocladium, 109 Howarp, Ricuarp A. The Botanical Re- sults of the U.S. Commission of Inquiry to Santo Domingo in 1871, 115 Howarp, Ricuarp A. The Correct Names for “Diospyros ebenaster,” 430 Howarp, Ricuarp A. The Directcr’s Re- por 7 ’ 1961 | INDEX 473 Howarp, RicHarp A. Ivan Murray Leucothoé sect. Eubotrys, 40 Johnston, 1898-1960, 1 — sect. Leucothoé, 3 Howarp, RicHarp A. Studies in the —sect. Maria Genus Coccoloba, X. cies and —sect. Oligarista, 40 ummary of Distribution in Sout — sect Paraleucothoé, America, 87; XI. in Asia, 10 oe 77 Huge Phos as 323-324 Hybridization in Downingia (Campanu- laceae), A Study of, 219 Hymenanthes, 30 Hypopitys, 67 Notes on the Species Illinites, 420 Indian pipe, 67 Isodorea pungens, 138 Isotoma axillaris, oes ete bodies, 33 an Murray Saath 1898-1960, 1 Johnston, Ivan Murray, 1898-1960, 1 Juglandinium, 35 Juglandoxylon, 351 Juglans fryxellii, 350 Juglexylon, 351 Kalmia, 23-26 Kalmiella, oe Kalmiopsis, Kingdonia ae The pete Morphol- Species of eras from the Celebes, 112; XXXII. A Review of the Genus Teenstroemia in the Philippine | ea 263; XXXIII. Variation in the of aPeraeinoerain kwangtungensis, 426° ae 77 iosiphon scandens, 388 ave and Nodal Nnaleas of Some Andean ompositae—Astereae, 276 Leatherwood, Ledum, 23, — sect. Leiophyllum, 27 80 Pero Nat: € — sect. eed 40 of the Asiatic Genus, 295 — subg. ieee 320, 384 — subg. Nectandra, 307 —albifolium, 320 andamanicum, 320; 395 — siamense, 320, — thorelii, 320, 395 Liquidambar at ages: 175-176 Liquidambaroxylon, 1 81 8 — sect. Pend 48 — subg. Maria, 48 Tanai paniculatus, 385 Malapuya Malpighia ideas, 135 Miocene Fossil Woods from the Colum- bia Basalts of Central Washington, 165; II, 347 Mitchella, 54 , 66-69 t, Eumonotropa. 67 474 Monotropa sect. Hypopitys, 67 sect notropa, 67 sabre vg adpressa, 107 — australis, corre tes lesb eevee armatum, 282 Se — bryoides — clothes 282 A Revision of the ” Enkleia (Thymelaea- e), 373 NEVLING, Lorin I., Jr. A Revision of the Asiatic Genus Linostoma (Thymelaea- ceae), 295 New Species of Pinus from Mexico, A, 3 Nyssa eydei, 353 Nyssoxylon, 354 Oak, red, 173 pcb eps tig 56-60 Palilag, 273 emventie% 266 Pantza Paratephia lepidophylla, 282 lucida, _ ade 282 — quadrangularis, 282 — teretiuscula, 282 Parry, C. C., Passerina iliformis, 378 Passiflora, pe seria Southeastern United s, The Genera of Turneraceae and, Passion- flower, 4 Pereskia, Pereskiopsis and Quiabentia, JOURNAL OF THE ARNOLD ARBORETUM [VOL. XLII Structure and Distribution of Scleren- chyma in the Phloem of. Comparative Anatomy of the Leaf-bearing Cacta- ceae, II, 144 ian Pereskiopsis and Quiabentia, or ee a of Crystals in. rei atomy of the Leaf- bearing Fuad III, 334 Pernettya, Phaleria, Pherosphaera, 418 Phloem of Pereskia, Pereskiopsis and e and ee enchyma in _ e. mparative the Leaf- mee te Cacta- Pine-sap, sweet, 65 Pinus from ae. A New Species of, 7 Pinus culminicola, 437-440 0 op 5 Pityosporites, 416 Plataninium, 184 Platanoxylon, 184 Platanus americana, 182-185 Podocarpus — Aberration in Acmopyle Coniferous Pollen ta of the Suan Hemisphere, I, Bac, Types of the Southern Hemis- phere, Coniferous, I. Aberration in ane and Podocarpus dacrydioi- des Boicodincs Polygonum See 107 valifolium, 108 — —thunbergiana, 108 — cymos 108 —molle, 109 — paniculatum, 109 Porterella carnulosa, 223, H, in te Secondary felis of Aeschynomene, 44 PRAKAS U. and Exso S. BarcHoorn. “pies Fossil Woods from the Co- 1961] lumbia Basalts of Central Washington, 165; II, 34 Psilaea, 318 — dalbergioides, 318 Pterocarya, Pterospora, Purdiaea (Cyrillaceae) from Central America, Schizocardia_ belizensis: