mo JOURNAL = x7es5 Vr +f OF THE 1467 ARNOLD ARBORETUM HARVARD UNIVERSITY B. G. SCHUBERT EDITOR T. G. HARTLEY C. E. WOOD, JR. LAZELLA SCHWARTEN CIRCULATION VOLUME 48 f 1 ARI ( as ke Ly JNM e oa or PUBLISHED BY THE ARNOLD ARBORETUM OF HARVARD UNIVERSITY CAMBRIDGE, MASSACHUSETTS 1967 DATES OF ISSUE No. 1 (pp. 1-104) issued January 13, 1967. No. 2 (pp. 105-202) issued April 16, 1967. No. 3 (pp. 203-361) issued July 18, 1967. No. 4 (pp. 363-588) issued October 14, 1967. TABLE OF CONTENTS CoMPARATIVE MorPHOLOGICAL STUDIES IN DILLENIACEAE, I: Woop Anatomy. William C. Dickison A Synopsis oF AFRICAN SPECIES OF DELPHINIUM AND CONSOLIDA. Philip A. Munz FioraL Brotocy AND SysTEMATICS OF EucNnipE (LOASACEAE). Henry J. Thompson and Wallace R. Ernst Tue Genus Duasanca. Don M. A. Jayaweera Srupies IN SwIETENIA (MELIACEAE): OBSERVATIONS ON THE SEX- UALITY OF THE Fiowers. Hsueh-yung Lee Some ProBLEMS oF TropicaAL PLant Econocy. Arturo Gémez- Pompa ANATOMY OF THE Patm Ruapis EXCELSA, IV. VascuLaR DE- VELOPMENT IN APEX OF VEGETATIVE AERIAL AXIS AND Ruzome. Martin H. Zimmermann and P. B. Tomlinson .... Two New American Pats. Harold E. Moore, Jr. NoMENCLATURAL Notes on Gossypium (MAtvacEaE). George K. Brizicky A Synopsis oF THE ASIAN SPECIES OF CoNSOLIDA (RANUNCULA- cEAE). Philip A. Munz ........ PRELIMINARY INVESTIGATIONS OF HYMENAEA COURBARIL AS A Resin Propucer. Jean H. Langenheim CoMPARATIVE MorPHOLOGICAL Stupigs IN DILLEeNiAcEAg, II. THE Potten. William C. Dickison .. Tue CHROMOSOMES oF AUSTROBAILEYA. Lily Riidenberg .............. CoRONANTHERA GRANDIS (GESNERIACEAE), A New SPECIES FROM THE SOLOMON IsLaNps. George W. Gillett A Synopsis or THE ASIAN SpEcIES OF DELPHINIUM, SENSU Srricto. Philip A. Munz .. Tur GENERA oF EUPHORBIACEAE IN THE SOUTHEASTERN UNITED Srates. Grady L. Webster Tue GENERA OF EUPHORBIACEAE IN THE SOUTHEASTERN UNITED Srates (Concluded). Grady L. Webster 20.0.0... ces Freeman AND Custis’ AccouNT OF THE RED River EXPEDITION oF 1806, AN OVERLOOKED PUBLICATION OF BoTANICAL INTER- a Eien ae. A Revision or THE Genus LunasiA (RUTACEAE). Thomas G. ME scat ecss a csens pon eanae nat eaenaeevae etna gees shee eke nehimet teen A capetiate or THE ASIAN Species OF DELPHINIUM, SENSU Srrictro (Continued). Philip A. Mund 0.0.0... VoLuME 48 NuMBER 1 JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY B. G. SCHUBERT EDITOR T. G. HARTLEY C. E. WOOD, JR. LAZELLA SCHWARTEN CIRCULATION SANUARY, 1967 Qo alee Gen THE JOURNAL OF THE ARNOLD ARBORETUM Published quarterly by the Arnold Arboretum of Harvard University. Subscription price $10.00 per year. Volumes I-XX, reprinted, and back issues of volumes XXI-XLV are avelable: ‘Son the Kraus Reprint Corporation, 16 East 46TH Srreet, New York, N.Y. 10017 Subscriptions and remittances should be addressed to Mrs. Lazenua ScHwarTeN, ARNOLD ARBORETUM, 22 Diyinrry AveNUE, CamMBrRipGcE, Massa- CHUSETTS 02138. CONTENTS OF NO. 1 Comparative MorpHonocican Stupres In DILLENIACEAE, I: Woop Anatomy. William C. Dickison 1 A Synopsis or Arrican Species or DELPHINIUM AND CoNnsoLmpA. Philip A. Munz 30+ Frioran Brotocy anp Systematics or Evcnme (LossAcEas). Henry J. Thompson and Wallace R. Ernst 56 Tus Genus Duasanea. Don M. A. Jayaweera 89 Srupres IN Swierenra (MeE.IAcEAE): OBSERVATIONS ON THE SEXUALITY OF THE FLowrrs. Hsueh-yung Lee 101 Volume 47, No. ‘ including pages 273-356, was issued October a: 1966. s Po a | NSO an eae es oe gee ath yen ev? __ sé itech % Eg : JOURNAL OF THE ARNOLD ARBORETUM VoL. 48 JANUARY 1967 NUMBER | COMPARATIVE MORPHOLOGICAL STUDIES IN DILLENIACEAE, I. WOOD ANATOMY Wiriiam C. DickIson ! THE TAXONOMIC TREATMENT of the Dilleniaceae has been based, for the most part, on a limited number of characteristics which are not well defined nor thoroughly understood. As a result, a survey of the taxonomic litera- ture reveals that there is little agreement on the proper phylogenetic posi- tion of the family (Taste 1). This taxon has been included in as many as six orders and allied to no less than fifteen different families. Not only is there considerable controversy regarding its position in relation to other families, but there is still no satisfactory treatment to show intrafamilial relationships. The early placement of the Dilleniaceae was in association with the ranalian complex, e.g., by Bentham and Hooker (1862) who divided the seventeen genera recognized by them as comprising the taxon into three tribes, based mainly on the form of the stamens. It is also interesting to note that these workers recognized the putative affinity of Crossosoma to ‘the Dilleniaceae. Such an affinity was to be argued for many years. On the basis of the supposed parietal placentation, Gilg (1893) in Engler and Prantl’s Die natiirlichen Pflanzenfamilien, placed the family between the Sterculiaceae and Eucryphiaceae in his large order Parietales. The artificiality of this order has been noted by Lawrence (1951) and by Standley and Williams (1961); their arguments have been supported by anatomical evidence described by T. K. Wilson (1960, 1964, 1965, 1966) in connection with his investigations on the Canellaceae. Within the family Gilg formed three subfamilies and six tribes and included the genera Actinidia and Saurauia. In a later edition of the Syllabus, Gilg and Werdermann (1925) removed the two above-mentioned genera to form a separate family Actinidiaceae. The four tribes within the Dilleniaceae proper were retained. This study represents a portion of a thesis submitted to the Graduate School, Indiana University, in partial fulfillment of the requirements for the degree of Doctor of Philosophy. This investigation was supported by Grants GB 3820 and GB 4127 from the National Science Foundation to Dr. J. E. Canright. 2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 TABLE 1.— Systematic Treatment of the Dilleniaceae AUTHOR PLACEMENT De CANDOLLE, 1824 LINDLEY, 1830 BENTHAM & Hooker, 1862 BAILLON, 1871 Le Maout & DECcAISNE, 1873 KUNTZE, 1891 Giiec, 1893 HALtier, 1905 WaRMING, 1911 BEsseEy, 1915 Gitc & WERDERMANN, 1925 RENDLE, 1925 HutTcHINson, 1926 Borvin, BENSON, 1957 CopELAND, 1957 CRONQUIST, 1957 CHADEFAUD & EMBERGER, 1960 EAMEs, 196 MELcHIOR, 1964 Ranales; nba Ranunculaceae and Magnolia Ranales; between Magnoliaceae and Winte ae Ranales; yy reat Ranunculaceae and and Calycanthaceae peters between Sterculiaceae and Eucryphiaceae Rosales; near maa Cistiflorae; near Bix Ranales; between Gee cic and Canellac Parietales; near gee a eon Guttiferales; near Ochnaceae Dilleniales; near Connaraceae Guttiferales; near Actinidiaceae Theales; near Actinidiaceae Dilleniales; near Crossosomataceae Ranales; between Annonaceae and Cercidiphyllaceae Guttiferales; near Theaceae Dilleniales; near eet olen aie Parietales; near Actinidiac Dilleniales: near Paeoniaceae Guttiferales- near Paeoniaceae Hoogland (1952, 1953) suggested the family should be divided into two subfamilies, the Tetraceroideae and Dillenioideae, again based primarily on stamen morphology. According to this author, the genera within the Dillenioideae show reticulate relationships with their arrangement depend- ing on the principles employed. At the present time, the most widely accepted treatment places the Dil- leniaceae, Paeoniaceae, and Crossosomataceae in close alliance under the Dilleniales (Eames, 1961). Cronquist (1957, 1965) noted, however, that this association was not firmly established and suggested, furthermore, that the centrifugal stamen development of the Dilleniales is the only essential character by which the Dilleniales differed from the Ranales. Since the Dilleniaceae possess a combination of characters particularly 1967 | DICKISON, DILLENIACEAE, I 3 interesting to those concerned with the discipline of comparative mor- phology and phylogeny (see, e.g., C. Wilson, 1965), it appeared highly advisable to undertake a comprehensive comparative morphological and anatomical study of the family; not only to help clarify those characters already known to exist within the family, but also to attempt to uncover additional evidence which might help elucidate the phylogenetic relation- ships of the taxa in question. This initial paper concerns the features of phylogenetic value in the woods of the Dilleniaceae. There are numerous descriptive works dealing with the wood anatomy of the Dilleniaceae, the most notable of which are: Moll and Janssonius (1906); Pearson and Brown (1932); Record and Hess (1943); Metcalfe and Chalk (1950); and Chowdhury and Ghosh (1958). Nevertheless, the wood of this family has neither been adequately described nor fully exploited in attempts to interpret phylogenetic relationships. Vestal (1937) studied twenty-nine species representing seven genera of Dilleniaceae in reaching his conclusion, based entirely on wood anatomy, that the Guttiferales have a natural alliance with such families as Ochna- ceae, Eucryphiaceae, Theaceae, Actinidiaceae, and Saurauiaceae, being derived through the dillenias. Bausch (1938), on the other hand, found no similarity in the wood of the Dilleniaceae and that of the Eucryphia- ceae. The only other reference to the secondary xylem of this family from a phylogenetic viewpoint is that of Wilson (1960) who found some agreement with the Canellaceae. The author respectfully acknowledges the assistance of Dr. James E. Canright, Arizona State University, for initially defining the scope of this research and for his continuing interest and guidance throughout. MATERIALS AND METHODS A total of seventy-one wood samples representing forty species in all ten genera of Dilleniaceae was examined. The writer is extremely thankful to Dr. William L. Stern for his generous personal gift of dilleniaceous wood slides. This collection included samples from the great majority of wood specimens of Dilleniaceae housed in the Samuel James Record Memorial Collection at Yale University. Special thanks are also given to Dr. P. S. Ashton, Kuching, Sarawak, for his contribution of wood from the mono- typic Didesmandra aspera Stapf, without which this study would have been incomplete. Additional wood samples were obtained from the follow- ing sources: the Arnold Arboretum of Harvard University (Aw); Chicago Yatural History Museum (Fw); Imperial Forestry Institute, Oxford University (FHOw); Division of Forest Products, Commonwealth Scien- tific and Industrial Research Organization, Melbourne (F PAw); Forest Research Institute, Kepong, Malaya (KEPw); personal collection of Dr. . L. Wilson (Witson); Division of Plant Anatomy, Department of Botany, Smithsonian Institution, Washington, D.C. (USw); Yale Uni- versity, School of Forestry (Yw). Woods of all ten genera considered by Hoogland (1951) to belong to 4 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 the family were examined. However, the semi-herbaceous genera Acro- trema and Pachynema were represented only by twig material, and were, therefore, used simply for comparative purposes and not studied statistical- ly. As suggested by Stern and Chambers (1960), a complete list of the woods studied, together with pertinent collecting information, is given in APPENDIX 1. Transverse, radial, and tangential wood sections were cut on a sliding microtome at a thickness ranging from 15 to 30u. The sections were stained by first being placed in a mordant of 2 per cent ferric ammonium sulphate (iron alum), after which they were transferred to 0.5 per cent aqueous Heidenhain’s haematoxylin and then to a 1 per cent solution of safranin in 50 per cent alcohol. Slides of macerated wood were prepared by putting small slivers into vials containing Jeffrey’s macerating fluid (equal parts 10 per cent aqueous nitric and chromic acids) and placing them in an oven at 60°C. The individual xylem elements resulting were washed and stained with a 1 per cent solution of light green in 75 parts clove oil and 25 parts absolute alcohol. Wood features of phylogenetic value were selected from the compilations of Tippo (1941, 1946). Measurements of cell length were obtained ex- clusively from macerated material. All statistical data were derived by measuring twenty-five cells at random from each sample. Tippo (1941) and Rendle and Clark (1934) recommend that one hundred measurements be taken from each sample in order to get a truly accurate mean length. Because of the small number of wood samples available for study of Tetracera, Schumacheria, Didesmandra, and Davilla, the data given for these genera is indicative only of general size ranges. The statistical in- formation on the remaining genera, however, should be considered as truly representative. Rays are described in accordance with the system of Kribs (1935), as well as by giving a brief written description as advised by the Committee on Nomenclature of the International Association of Wood Anatomists in the International Glossary of Terms used in Wood Anatomy (1957). This classification was employed because of its ease of application and wide usage in the literature. The descriptive terms for the wood parenchyma are those suggested by Hess (1950). In discussing relative cell size, the suggestions of Chattaway (1932), Chalk (1938), and the Committee on the Standardization of Terms of Cell Size (1937) are used. Those adopted in this investigation are given below: PorE NUMBER Numerous - more than 40 per mm 2 Many . 20-40 per mm 2 . 10-20 per mm 2 . 5-10 per mm 2 . 2-5 per mm 2 . fewer than 2 per mm 2 Rather many . Rather few es Very few 1967 | DICKISON, DILLENIACEAE, I £ VESSEL DIAMETER Extremely email i. ac ghanw eo sie ees pb MOE SUNN 5 se we casei re a asian nee Moderately... small «25:1: 954 pcg aoe xg Sk OO Medium PS ME pe re ics Oe aes 100-200p, WHOECHIONY TAP ne as a ee ee ee ope ee 200-300, MODY URIGG cn “a ee ek ee ae ee PcOMeN TRIG es ee ek ae a 400p, VessEL MEMBER LENGTH Extremely SRO 0 0 aa oe 8 oe) op a ee CI WECy SWORE; 2. oo ace eee ee ee ee oe Moderately short 2 20. ee ke a 2I0-SOOy RE gk ee a eg ee eee ee Moderately long. 2 ue PP SOOT IOOS iy in ie eeee 1100-1900p. Batremely. long iy .~ perce a over 1900n LENGTH OF IMPERFORATE TRACHEARY ELEMENTS Very hort oo ee ee a ee ee SHMt icy ee ES ae i OP ee oe iaoen LOGE hos: we Se ae oe es MORY TORR see) iyrsa PB cece wp geek ee | The remainder of the terminology is generally that advocated by the Committee on Nomenclature of the International Association of Wood Anatomists in the Multilingual Glossary of Terms used in Wood Anatomy (1964). OBSERVATIONS Description of the Wood of Dillenia. Growth rings absent. Wood diffuse-porous; vessels (Fics. 5, 7, 23, 24) solitary, or occasionally in pairs or small clusters, rarely as many as 4 in radial rows, very few to many (1-21, mostly 7-13 per mm 2), circular to angular in outline, very small to very large, usually medium-sized (diam. 45-330, mostly 113-228p, mean 147); vessel member length medium to extremely long, usually very long (625-3062, mostly 687—2500u, mean 1476); tyloses infre- quently present. Intervascular pitting (Fic. 8) opposite to transitional, pits circular to elongate (diam. 7-35); pitting to rays similar to scalari- form. Perforation plates exclusively scalariform with some branched bars, occasionally nearly reticulate; completely bordered, with 5-120 (mostly 15-40) bars. End wall inclination (Fic. 6) 25—72°. Imperforate tracheary elements (Fic. 7) of the tracheid and fiber tracheid type, thin- to thick- walled; fiber tracheids with distinctly bordered pits and included or ex- tended, slit-like, often crossed apertures; very short to very long, usually very long (625-4375, mostly 1875-3750, mean 2520n). Rays (Fic. 6) heterogeneous, Type I, of two sizes; uniseriate rays 1-25 (mostly 4-12) cells high; multiseriate rays 1-18 (mostly 1-10) cells wide, 23 to more than 200 (mostly 52-183) cells high. Multiseriate rays often with long 6 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 48 uniseriate wings composed of upright cells. Elongated upright cells in long series (2-19 cells) often running through multiseriate rays. Crystals in the form of raphides often numerous in ray parenchyma, as are enlarged secretory cells. Ray parenchyma sometimes containing dark, amorphous, gummy deposits. Wood parenchyma (Fic. 5) apotracheal diffuse and aggregate diffuse, often paratracheal diffuse to incomplete vasicentric. The very close similarity of the wood anatomy in the two genera supports Hoogland (1952) in reducing Wormia Rottb. to Dillenia L. Description of the Wood of Schumacheria. Growth rings absent. Wood diffuse-porous; vessels (Fics. 1, 22) solitary, apparent pairs due to overlapping ends, rather many to numerous (16-51, mostly 21-41 per mm *); circular in outline, tending to angular, very small to medium, usually moderately small (diam. 30-140u, mostly 70-120, mean 90,); vessel member length moderately long to very long to extremely long, usually extremely long (1044-3074, mostly 1740-2610u, mean 2047,). Intervascular pitting scalariform to opposite to transitional, pits circular to elongate (diam. 7-35); pitting to rays opposite to transitional to scalariform. Perforation plates exclusively scalariform, bars commonly branched, completely bordered, with 34-135 (avg. 40-60) bars. End wall inclination 35—75°. Imperforate tracheary elements of the tracheid and fiber tracheid type, thin- to thick-walled; fiber tracheids with distinct- ly bordered pits and extended, slit-like, often crossed pit apertures; long to very long, usually very long (1160-3480p, mostly 2030-2900, mean 2380). Rays (Fic. 2) heterogeneous, Type I, of two sizes; uniseriate rays 5-28 (mostly 10-15) cells high; multiseriate rays 1-5 cells wide, 35-68 (mostly 50-60) cells high; multiseriate rays often with extremely long uniseriate wings composed of upright cells. Wood parenchyma apotracheal diffuse. Description of the Wood of Hibbertia. Growth rings present or absent, when present not well defined. Wood diffuse-porous; vessels (Fic. 3) mostly solitary, with some apparent pairs due to overlapping ends, very small clusters infrequently present, rather few to numerous (5—40, mostly 10-20 per mm 2); circular to angular in outline, very small to medium sized (diam. 30-165p, mostly 60-135); vessel member length medium to very long, usually moderately long (625-1562y, mostly 812— 1437n, mean 1074). Intervascular pitting opposite to transitional to scalariform, pits circular to scalariform (diam. 8-40), pitting to rays similar. Perforation plates exclusively scalariform, completely bordered to only bordered at the ends, with 15-88 (mostly 35-50) bars. Bars com- monly branched to nearly reticulate. End wall inclination 45-85°. Im- perforate tracheary elements of the tracheid and fiber tracheid type, thin- to thick-walled; fiber tracheids with distinctly bordered pits with extended or included, slit-like, often crossed, pit apertures; very short to very long, usually long (625-2150y, mostly 1250-1875, mean 1551). Rays (Fic. 1967] DICKISON, DILLENIACEAE, I 7 4) heterogeneous, Type I, of two sizes; uniseriate rays 1-8 (mostly 2-7) cells high; multiseriate rays 1-5 (mostly 1-3) cells wide, 7-34 (mostly 15-30) cells high. Multiseriate rays often with long uniseriate wings. Ray parenchyma frequently contains dark, amorphous, gummy deposits. Wood parenchyma apotracheal diffuse; crystals in the form of raphides some- times present in enlarged wood parenchyma cells. Description of the Wood of Didesmandra. Growth rings absent. Wood diffuse-porous; vessels solitary, rarely in pairs, apparent pairs due to vessel overlap, rather many to numerous (13-65, mostly 25-45 per mm”); angular in outline, very small to moderately small, usually mod- erately small (diam. 30-90, mostly 45—70u, mean 60p); vessel member length medium to extremely long, usually very long (780-2600, mostly 1150—2250u, mean 1574). Intervascular pitting scalariform to opposite; pitting to rays similar. Perforation plates exclusively scalariform. Bars commonly anastomosing to reticulate; completely bordered, with 25-160 (mostly 57-120) bars. End wall inclination 30-75°. Imperforate tra- cheary elements of the tracheid and fiber tracheid type, mostly thin-walled; fiber tracheids with distinctly bordered pits and included or extended, slit-like, often crossed, pit apertures; short to very long, mostly very long (1400-3050, mostly 1800-2870», mean 2292y). Rays heterogeneous, Type I, of two sizes; uniseriate rays 2-33 (mostly 6-21) cells high; multi- seriate rays 1-15 (mostly 3-10) cells wide, 14—83 cells high. Main body of multiseriate rays composed of both procumbent and upright cells, often with long uniseriate wings composed of upright cells. Crystals in the form of raphides infrequently present in ray tissue. Ray parenchyma sometimes containing dark, amorphous, gummy deposits. Wood paren- chyma apotracheal diffuse. Description of the Secondary Xylem of Acrotrema and Pachy- nema. Growth rings absent. Wood diffuse-porous; vessels (Fics. 9-12) solitary to rarely in pairs, apparent pairs due to vessel overlap; circular to angular in outline, extremely small to very small (diam. 11-34, mostly 20-27). Intervascular pitting (Fic. 13) scalariform to opposite to rarely transitional, pits circular to elongate (diam. 7-18»). Perforation plates exclusively scalariform, with 16-48, completely bordered, bars. Bars sometimes anastomosing. Imperforate tracheary elements of the fiber tracheid type, thin- to thick-walled; fiber tracheids with distinctly bor- dered pits with included, or extended, slit-like, often crossed, apertures. In Acrotrema, elements were found with greatly extended pit apertures and much reduced pit borders. Rays reduced to mostly uniseriate with occa- sional biseriate rays. Rays composed of both procumbent and upright cells, 1-14 cells high. Wood parenchyma apotracheal diffuse. Crystals in the form of raphides observable in the periderm of Acrotrema. Description of the Wood of Doliocarpus. Growth rings absent. Two samples (D. coriaceus (Mart. & Zucc.) Gilg and D. sp., Yw 39903) 8 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 exhibit successive cambial activity which resulted in concentric zones of included phloem, conjunctive tissue, and sclerotic cells. Wood diffuse- porous; vessels (Fic. 26) solitary, very few to rather many (2—20, mostly 4-9 per mm *); circular in outline, moderately small to extremely large, usually moderately large (diam. 50-555, mostly 150-450, mean 256n); vessel member length medium to very long, usually moderately long (362-1812, mostly 687-1250, mean 917). Intervascular pitting (Fic. 21) opposite to transitional, pits circular to elongate (diam. 7—40,) ; pitting to rays similar. Perforation plates mostly simple, scalariform in the smaller vessels. Scalariform plates with 1-15 bars, sometimes branched, completely bordered to bordered only at the ends. End wall inclination 35-90°. Imperforate tracheary elements of the tracheid and fiber tracheid type, thin- to thick-walled; fiber tracheids with distinctly bordered pits and extended, slit-like, often crossed, pit apertures, very short to very long, usually short (687—4000n, mostly 1125-1562u, mean 1471 ,). Rays heterogenous, Type I and IIA, of two sizes; uniseriate rays 1-9 (mostly — 3-7) cells high; multiseriate rays 1-41 (mostly 1-35) cells wide, 36 to over 500 (mostly 45-200) cells high. Crystals in the form of raphides often very numerous, as are enlarged secretory cells. Wood parenchyma mostly apotracheal diffuse, some paratracheal diffuse. One sample studied, identified with an unpublished name (Yw 39903), displayed features inconsistent with those of the other species examined. The radial pore chains and pore clusters, vasicentric parenchyma, and significantly different vessel member and fiber length suggest that it is not a member of this genus. The accompanying data stated it was taken from Fw 11189. However, the two wood samples were clearly not taken from the same specimen. Description of the Wood of Davilla. Growth rings absent. Wood diffuse-porous; vessels (Fics. 15, 23) solitary, rather few to rather many (3-16, mostly 4-8 per mm 2); circular in outline, very small to extremely large, usually moderately large (diam. 30-450, mostly 222-382y, mean 284); vessel member length medium to moderately long, usually medium- sized (437-1062y, mostly 625-937, mean 777»). Intervascular pitting opposite to transitional, pits circular to elongate (10-25,); pitting to rays similar. Perforation plates mostly simple, scalariform in the smaller vessels. Scalariform perforation plates with 5—10 bars, completely bordered to bordered only at the ends. End wall inclination 55—90°. Imperforate tracheary elements of the tracheid and fiber tracheid type, thin-walled; fiber tracheids with distinctly bordered pits with extended or included, slit-like, often crossed, pit apertures: very short to very long, usually long (938-2625, mostly 1250-2375, mean 1613). Rays (Fic. 15) hetero- geneous, Type I and IIA, of two sizes: uniseriate rays 1-31 (mostly 15— 30) cells high; multiseriate rays 1-38 (mostly 1-31) cells wide, 75 to over 500 (mostly 200-450) cells high. Large sclerotic cells infrequently present in ray tissue, quite common in bark. Ray cells often containing large 1967 | DICKISON, DILLENIACEAE, I 9 secretory cells and dark, amorphous, gummy deposits. Wood parenchyma apotracheal diffuse to scattered paratracheal diffuse. Description of the Wood of Curatella. Growth rings absent. Wood diffuse-porous; vessels (Fic. 14) solitary, or rarely in clusters of up to three vessels, very few to rather many (1-15, mostly 2-7 per mm 2), circular to rarely angular in outline, very small to very large, usually medium-sized (diam. 45-360u, mostly 145-200u, mean 180); vessel member length very short to very long, usually medium-sized (250-1312y, mostly 625—900u, mean 725u). Intervascular pitting opposite to transi- tional, pits circular to elongate (diam. 8-25); pitting to rays similar. Perforation plates mostly simple (Fic. 17), scalariform in the smaller vessels. Scalariform perforation plates completely bordered, with 1-12 (mostly 1-5) bars, branching of bars common, occasionally reticulate perforations are present. Inclination of end walls 55-90°. Imperforate tracheary elements of the tracheid and fiber tracheid type, thin- to thick- walled; fiber tracheids with distinctly bordered pits with extended, slit-like, often crossed, pit apertures; very short to very long, usually long (562- 2750p, mostly 1354-2208u, mean 1805). Rays (Fic. 16) heterogeneous, Type I and IIA, of two sizes; uniseriate rays 1-20 (mostly 1-11) cells high; multiseriate rays 1-20 (mostly 1-15) cells wide, 32-279 (mostly 54-119) cells high. Crystals in the form of raphides often numerous, as are enlarged secretory cells. Ray parenchyma often containing dark, amorphous, gummy deposits. Wood parenchyma (Fic. 14) aggregate to narrow-banded apotracheal, some paratracheal diffuse. Description of the Wood of Tetracera. Growth rings absent or present; if present, not well defined. One species (T. volubilis L.) ex- hibited successive cambial activity which resulted in concentric zones of included phloem, conjunctive parenchyma with scattered fibers, and sclero- tic cells. Wood diffuse-porous; very weak ring-porosity in one species (T. boiviniana Baill.) as shown by larger vessel diameter in the spring wood. Vessels (Fics. 18, 27) solitary, rarely in pairs, few to rather many (3-11, mostly 5—8 per mm ”); circular in outline, very small to very large (diam. 50-350); vessel member length very short to moderately long, usually medium-sized (220-1062, mostly 370-937, mean 565). Inter- vascular pitting opposite to transitional, pits circular to elongate (diam. 7-35); pitting to rays similar. Perforation plates mostly simple, scalari- form in the smaller vessels. Scalariform plates completely bordered, with 1-20 (mostly 1-10) bars. Vessel members occasionally with reticulate or “multiperforate” perforation plates (Fic. 19). Inclination of end walls 35-90°. Imperforate tracheary elements of the tracheid, vasicentric tracheid, and fiber tracheid type, thin- to thick-walled; fiber tracheids with distinctly bordered pits and extended or included, slit-like, often crossed, pit apertures; very short to long, usually short (500-1625p, mostly 625-1250, mean 1003). Rays (Fic. 20) heterogeneous, Type I and ITA, of two sizes; uniseriate rays 1-18 (mostly 3-11) cells high; 10 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 multiseriate rays 1-13 (mostly 1-10) cells wide, 20-152 (mostly 45-110) cells high. Multiseriate rays often containing large secretory cells. Crystals in the form of raphides numerous along with dark, amorphous, gummy deposits. Series of upright cells sometimes running through rays. Wood parenchyma apotracheal diffuse. The small number of wood samples studied in this genus revealed a relatively large variability in anatomical structure. The extent and sig- nificance of this variation can only be determined by the examination of additional samples and species when they become available. DISCUSSION The woods of the Dilleniaceae reveal a rather low level of specialization. This is supported by the occurrence in all woods of the following primitive features: (1) wood diffuse-porous; (2) predominantly solitary pores; (3) opposite-transitional intervascular pitting; (4) fiber tracheids; (5) hetero- geneous Type I or ITA rays; and (6) wood parenchyma mostly apotracheal diffuse. Despite this apparent homogeneity, it is quite obvious that con- siderable variation exists between genera in vessel member structure. Since the trends of vessel specialization were initially discovered for dicotyledons by Bailey and Thompson (1918), Bailey and Tupper (1918), and Frost (1930, 1931), and for monocotyledons by Cheadle (1942, 1943a, 1943b, 1944), the vessel has been an invaluable tool for interpreting phy- logenetic relationships. The importance ascribed to the vessel member in this regard results not only from a thorough understanding of vessel phylog- eny, but more importantly from the fact that these clearly defined trends are both undirectional and irreversible. It is of great interest, therefore, to find vessels within a single family which reveal salient irreversible trends of specialization. On the basis of vessel member structure, as well as more subtle wood characters, the Dilleniaceae are most conveniently segregated into the two subfamilies as suggested by Hoogland (1952, 1953). The anatomical features of the two subfamilies are summarized in TABLE 2. I. CoMPARISON OF THE Two SUBFAMILIES OF DILLENIACEAE The most primitive forms of vessel members are found in the subfamily Dillenioideae. In this group the vessels are angular to circular in outline, of very great length (mean 1532), possess exclusively scalariform per- foration plates, with up to 160 bars, are mostly completely bordered, have very oblique end walls, and have scalariform to opposite to transitional intervascular pitting. The extremely tracheid-like vessel members of Schumacheria and Didesmandra are the least specialized in the family. The vessel members in the genera Dillenia and Hibbertia are, in general, only slightly more advanced. The Tetraceroideae are characterized by certain salient trends of vessel 1967 | DICKISON, DILLENIACEAE, I 11 TABLE 2. — Comparison of Averages of Characters of Wood Anatomy of the Two Subfamilies of Dilleniaceae CHARACTER DILLENIOIDEAE TETRACEROIDEAE VESSEL MEMBERS 1. Distibulion «os. aes mostly solitary mostly solitary . Shape in outline . . . circular to angular circular 3. Type of perforation DLC. «yw eee hee scalariform only scalariform and simple 4. Bars in perforation plate Lowest average... . 18 2 Mean averages 40 5 Highest average. . . 123 9 5. End wall inclination . 30-80° 50-90° 6. Length PBSC se ee, beak 5 625-3074, 220-1812y Most frequent range . . 687-2610u 370-1250u 2 aoa caer te 766p 7. Width ROPER ca sg ee 30-330n 30-555 py, Most frequent range . . 60-228, 145-450u jaan a en EEE 124 240u 8. Intervascular pitting . . scalariform-transitional _ opposite-transitional Woop FIBERS Bo: WR. ae erica a fiber tracheid fiber tracheid 10. Wall thickness . thin to thick mostly thin 11. Length eae: 625-4375 500-4000. Most frequent range 1250-3750p 625-2375 DEPA s ee s 2150u 1473, Woop Ray RAEN TG ais Be eee Type I Type I and IIA 13. Average width of multi- seriate rays (no. of cells) 1-18 (avg. 9) 1-41 (avg. 22) 14. Average height of multi- seriate rays (no. of cells) | 7-200 (avg. 39-91) 20-500 (avg. 86-219) Woop PARENCHYMA 15. Distribution. . . . . mostly apotracheal mostly apotracheal diffuse diffuse member specialization. The four genera included in this subfamily all possess vessel members with both simple and scalariform perforations. In all, simple perforations are considerably more numerous than the scalari- form type, which are confined strictly to the smaller vessels of the second- ary xylem. When scalariform plates do occur, they seldom have more than ten bars. A small number of vessel members were observed which bore scalariform and simple perforation plates at opposite ends. All gradations 12 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 from porous-oblique to transverse-porous perforation plates are present. Correlated with the change in type of perforation plate is a significant decrease in length (mean 766.) with a considerable increase in width. The transectional outline is predominantly circular. Whereas Doliocar pus, Davilla, and Curatella are intermediate in the above features, the genus Tetracera appears to be the most advanced of all Dilleniaceae in vessel structure. The last-named genus is also unique in its possession of so called ‘‘multiperforate” or reticulate perforations. The significance of this type of perforation plates is still a matter of controversy. Thompson (1923) suggested that they may represent modifications of scalariform types or retrogressions from simple plates. Gottwald and Parameswaran (1964), in describing their occurrence in the Dipterocarpaceae, believe they are intermediate forms between scalariform and simple perforations. That all known trends of vessel specialization do not always occur simultaneously (Bailey and Howard, 1941; Bailey, 1944) is well exempli- fied by the existence of opposite-transitional, lateral-wall pitting in both subfamilies. Bailey (1957) pointed out that in certain specialized cases, such as lianas, vessel members may become quite transformed as a result of their habit, and do not, therefore, indicate true phylogenetic advancement. Ayensu and Stern (1964) encountered such modification in their study of the scandent Passifloraceae, and so considered the presence of shorter, broader, solitary vessel members in this group to be related to the physi- ology of their habit rather than to their phylogenetic position. They in- dicate, however, that habit-related anatomical modification may be super- imposed upon phylogenetic specialization and the separation of these facets is often difficult. Since the Dilleniaceae range from large trees to scandent lianas, with all intermediate forms, it is quite possible that the advancement exhibited by certain species of Davilla, Doliocarpus, and Tetracera is actually co- ordinated with habit. However, the occurrence of scalariform perforation plates in these genera would also seem to indicate that they are really in an intermediate stage of advancement. Furthermore, similar advanced vessel types were observed in Curatella, which may be a small tree up to eight meters in height (Standley, 1928). It is clear, therefore, that it would be erroneous to attempt to interpret phylogenetic position of the woods from vessel structure alone. It is necessary to correlate this evidence with that obtained from other features of the xylem. The unspecialized nature of the Dillenioideae is indicated by the pres- ence of the longest imperforate tracheary elements and by the exclusively heterogeneous Type I rays (except in the reduced Acrotrema and Pachy- nema). Both uniseriate and multiseriate rays are of extreme height. In addition, the multiseriate rays are often conspicuous by their long uni- seriate wings composed of upright cells. Although the tremendously high and broad rays of the Tetraceroideae are the largest in the family, there is a marked tendency for reduction in the uniseriate wings and an increased 1967 | DICKISON, DILLENIACEAE, I 13 uniformity in the ray cells to a point where more advanced Type IIA rays are present. Both subfamilies have predominantly apotracheal diffuse parenchyma. Within the Dillenioideae, only Dillenia is characterized by paratracheal diffuse parenchyma, whereas in the Tetraceroideae both Davilla and Doliocarpus show this condition. Also, Curatella exhibits fairly specialized aggregate to narrow-banded apotracheal with some paratracheal diffuse parenchyma. Crystals in the form of raphides are often quite numerous in the wood of the Dilleniaceae. Although Chattaway (1955) states that crystals have been reported in approximately 1000 genera of 160 families, the presence of certain types of crystals, such as raphides, is much less common and may, therefore, be of more taxonomic value (Metcalfe and Chalk, 1950; Chattaway, 1955). Raphides are undoubtedly diagnostic for the majority of genera within the family, but are of questionable taxonomic importance for the family as a whole. The only two genera in which raphides were not confirmed are Davilla, where they have been reported (cited by Chattaway, loc. cit.) and Schumacheria. In addition, any one genus may not possess raphides uniformly in all species (e.g., Dillenia). When present, these were observed only in certain species of Hibbertia. However, the occur- rence of crystals in such a large number of unrelated plant families seems to negate any phylogenetic significance in this anatomical character. Secretory cells were similarly encountered in all dilleniaceous genera except Schumacheria and Didesmandra. Stern (1954a) advised the term “secretory intercellular cavity” be used when the chemical composition of the secretion is not known. The darkly staining amorphous mucilage may be deposited either in ray parenchyma, axile parenchyma, or vessels (Fic. 18). The presence of this mucilage is apparently not uniform within genera or even in species. Canright (1955) considered oil cells to be of no phylogenetic significance in the wood of the Magnoliaceae. Stern (1954b), however, suggested they may represent a particular kind of specialization in lauraceous wood. The occurrence of secretory cells in the Dilleniaceae offers little evidence to support either concept. However, it is of importance to mention that secretory cells are common features of all organs of most families generally regarded as ranalian; on the other hand, this feature is relatively rare in families in the higher orders, e.g., Guttiferae. II. Comparison oF Woop OF THE DILLENIACEAE I PUTATIVELY RELATED FAMILIES The following discussion will attempt to evaluate, on the basis of wood anatomy, the relationships of the Dilleniaceae to many putatively related families. In the majority of cases, woods of the various families were actually examined. In addition to these personal observations, descrip- 14 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 tions were obtained from the following sources: Canright (1955), Chatta- way (1937), Keng (1962), Metcalfe and Chalk (1950), Swamy and Bailey (1949), Vander Wyk and Canright (1956), and Vestal (1937). The Dilleniaceae have been placed, at one time or another, mainly on the basis of gross floral morphology, with the following ranalian families: Ranunculaceae, Magnoliaceae, Winteraceae, Calycanthaceae, Annonaceae, Canellaceae, Cercidiphyllaceae, and Ceratophyllaceae. Evidence from wood anatomy, for the most part, does not support such alliances. The vesselless Winteraceae and aquatic Ceratophyllaceae may be readily dismissed as showing little similarity to the Dilleniaceae. The presence of simple perforation plates, alternate intervascular pitting, libriform wood fibers and paratracheal parenchyma in the Ranunculaceae (excluding Paeonia) indicates that the xylem of this family is on a higher level of advancement than that of the Dilleniaceae. The Calycanthaceae do possess fiber tracheids, heterogeneous Type I rays, and apotracheal diffuse to vasicentric scanty parenchyma, but, on the other hand, show pores mostly in radial multiples with exclusively simple perforation plates, alternate intervascular pitting, as well as uni- formly occurring vasicentric tracheids. The Magnoliaceae are characterized by wood which shows transitions from comparatively primitive to rather advanced. Although the most primitive genera do have some similarities with the Dilleniaceae, the family appears to be well placed near the Degeneriaceae and Himantandraceae (Bailey, Nast and Smith, 1943; Canright, 1955). The Annonaceae, likewise, is anatomically advanced over the Dillenia- ceae and accordingly, was considered by Vander Wyk and Canright (1956) to share more common features with the Myristicaceae. Although Wilson (1960) did find several points of agreement between the Dilleniaceae and Canellaceae, he found the wood features in the latter family had the greatest degree of similarity with the Illiciaceae and Eupteleaceae. The wood of the Cercidiphyllaceae approaches the Dilleniaceae in primi- tive vessel structure and the presence of fiber tracheids, but differs in the terminal wood parenchyma and heterogeneous Type IIB rays (Swamy and Bailey, 1949). Due to the large degree of family overlap in ordinal concepts, it will be more meaningful for the remaining discussion to consider relationships of individual families irrespective of specific orders. There is little similarity between the wood of the Sterculiaceae and that of the Dilleniaceae. Such features as exclusively simple perforation plates, alternate intervascular pitting, abundant parenchyma in uniseriate lines and in some cases together with distinct vasicentric sheaths, Type IIA or IIB rays, and libriform wood fibers, are much more advanced than those of the Dilleniaceae. The Ochnaceae and Connaraceae are also anatomically on a higher level of advancement than the Dilleniaceae as evidenced by the occurrence in both families of the following features: typically simple perforation plates, 1967 | DICKISON, DILLENIACEAE, I 15 alternate intervascular pitting, and libriform wood fibers (exclusively in the Connaraceae and partially in the Ochnaceae. The Connaraceae is further characterized by highly advanced heterogeneous Type III or homo- geneous Type III rays. Although the Eucryphiaceae and Brunelliaceae are not as advanced anatomically as the previous families, they are still more highly developed than the Dilleniaceae. Woods of both families contain both simple and scalariform perforation plates. The parenchyma of the Eucryphiaceae is diffuse or slightly banded to terminally banded, whereas wood parenchyma is absent from the Brunelliaceae. Both families have genera with occa- sional heterogeneous Type III rays. However, other lines of evidence, viz., floral morphology and chemical data, support placing these two families near the Cunoniaceae (Bausch, 1938). The genera Saurauia and Actinidia have often been placed near, if not included in, the Dilleniaceae. Such wood features in Saurauia as mostly solitary pores, vessel members of extreme length and exclusively scalari- form perforation plates (with up to 90 bars), scalariform to opposite to transitional intervascular pitting, long fiber tracheids, exclusively hetero- geneous Type I rays, and apotracheal diffuse parenchyma, are very primi- tive; comparable features are found only in the most unspecialized mem- bers of the Dillenioideae. Actinidia shows features more in common with the Tetraceroideae, such as solitary pores, both simple and scalariform perforation plates, fiber tracheids of similar mean length, and heterogeneous Type I or IIA rays. At the same time, however, Actinidia also has some wood characters not shared with any genera in the afore-mentioned subfamily. These include scalariform perforation plates which may have over 75 bars, and much reduced ray structure. In contrast to the above, the alliance of the Dilleniaceae to the Thea- ceae is supported by anatomical considerations. Keng (1962) compared the wood of members of this family and found the wood of the Camel- lioideae to be, in general, more advanced than that of the Ternstroemioi- deae. Furthermore, although the wood of the Theaceae is considered to be relatively primitive, there are specific cases of high advancement. The greatest agreement between the Dilleniaceae and Theaceae exists between the Dillenioideae and tribes Adinandrieae and Ternstroemieae. This sim- ilarity is evident in the mostly solitary pores, exclusively scalariform per- foration plates with many bars, similar mean vessel member length (Adin- andrieae), scalariform to opposite intervascular pitting, uniform occur- rence of fiber tracheids of similar mean length, heterogeneous Type I rays (significantly broader and higher in the Dillenioideae), and wood paren- chyma mostly apotracheal diffuse. ‘ The occurrence of spiral thickenings in the vessels of some T heaceae is a character of questionable phylogenetic significance (Canright, 1955; Keng, 1962; Stern, 1954b). This character was not encountered in any of the Dilleniaceae observed but does show up in the vessels of Saurauia and Actinidia. 16 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 Recent systems have repeatedly linked the monotypic family Crossoso- mataceae to the Dilleniaceae by the presence in the genus Crossosoma of free follicular carpels and strongly arillate seeds However, the semi-ring porosity, exclusively simple perforation plates, mostly alternate inter- vascular pitting, very short imperforate tracheary elements, heterogeneous Type IB rays, very sparse to absent parenchyma are features clearly more advanced than those of the Dilleniaceae. If Crossosoma were closely related to the Dilleniaceae, it would seem that its secondary xylem should retain more of the primitive features found in that family. Since being removed from the Ranunculaceae, the genus Paeonia is now most commonly placed as a monotypic family in close association with the Dilleniaceae. Examination of the wood reveals that it is more highly evolved than the Dilleniaceae in a number of characters: reduced ray structure (uniseriate or biseriate rays commonly occur), occasional alter- nate intervascular pitting, reduction in the number of scalariform bars in the perforation plates to 1 to 5, vessels occasionally in small clusters with semi-ring porosity. In spite of these features, wood anatomy does not negate the possible position of Paeonia as allied to, yet more highly advanced than, the Dilleniaceae. Since a taxon which retains primitive wood characters cannot be derived from one which is uniformly advanced, Bailey (1957) has pointed out that the evidence obtained from wood studies are often more helpful in nega- tions than in positive assertions of close alliance. Furthermore, structures often appear similar in quite distantly related groups owing to the fre- quent occurrence of parallel evolution (Bailey, loc. cit.). It would, there- fore, be very unwise to speculate concerning phyletic relationships solely on the basis of wood anatomy. Bailey (1951, 1953, 1957) and Canright (1955, 1963) have stressed the need for comprehensive studies which take into account evidence from all organs and parts of the plant. Only when all the information has been gathered and carefully evaluated can mean- ingful phylogenies be constructed. APPENDIX 1. — Dilleniaceous Wood Examined COLLECTOR GEOGRAPHICAL HERBARIUM LOCATION AND SPECIES AND NUMBER ORIGIN VOUCHER CATALOG NUMBER Acrotrema uniflorum Hook. Thwaites 265 Ceylon — Aw 26224 Curatella americana L. A. C. Smith 2185 British Guiana y Yw 35484 C. americana L Zanderij 209 Surinam Y Yw 41257 C. americana L L. Williams 9842 Mexico x Yw 35006 C. americana L A. J. Fors 96 Cuba — USw 21446 C. americana L H. Pittier x USw 2773 C. americana L Forest Dept. 2185 British Guiana — FHOw 14636 C. americana L. . A. Krukoff — — Fw 20 Davilla aspera (Aubl.) Benoist Cuatrecasas 17302 Colombia F, VALLE Yw 43186 D. aspera (Aubl.) Benoist Cuatrecasas 17302 Colombia F, VALLE Fw 17302 D. aspera (Aubl.) Benoist Woy en 5461 Peru USw 15527 D. rugosa Poir. H. Leo Cuba — Yw 40712 Didesmandra aspera Stapf Banyi ing anak Ludong Sarawak = SN. Dillenia aurea Smith Java — Yw 30030 D. biflora (A. — Martelli as Fiji — Yw 25835 ex Dur D. biflora = Gray) Martelli A. C. Smith 3471 Fiji NY Yw 28327 ex Dur D. biflora Ke Gry) Martelli Fiji _— Yw 25674 ex Dur. & Jac D. excelsa ae Gilg — Java — Yw 30025 D. excelsa (Jack) Gilg ~- Java — Yw 30026 D. excelsa (Jack) Gilg —_ Java _— Yw 30027 D. excelsa (Jack) Gilg B. A. Krukoff 271 Sumatra NY USw 7127 D. excelsa (Jack) Gilg B. A. Krukoff 271 Sumatra NY FHOw 8368 [L961 AVAOVINATTIC ‘NOSTMOIG ‘ I COLLECTOR GEOGRAPHICAL HERBARIUM LOCATION AND SPECIES AND NUMBER ORIGIN VOUCHER CATALOG NUMBER D, excelsa (Jack) Gilg D. E. Carver 20 Sarawak — FHOw 6655 D. eximia Miq. — Malaya KEPw 7302 D. indica L — Java a Yw 30028 D. indica _— Malaya — KEPw 5968 D. ingens (B. L. Burtt) A. C. Se Waterhouse 25 Bougainville Y Yw 21156 D. ingens (B. L. Burtt) A.C. S F. S. Walker 146 Solomons ~~ FHOw 17958 D. luzoniensis (Vidal) Martelli — Philippines — Fw 272982 ex Dur. & Jacks. D. pentagyna Roxb oo a — Yw 30032 D. pentagyna b. J. F. Kools 33 Indonesia — FHOw 18866 D. philippinensis Rolfe — Philippines — USw 5213 D. pulchella (Jack) Gilg H. E. Desch. 41887 alaya KEP Yw 38502 D. pulchella (Jack) Gilg H. E. Desch. 41887 Malaya KEP KEPw 3675 D. pulchella (Jack) Gilg oo Malaya — KEPw 6544 D. quercifolia (White ‘ Francis Holland New Britain —_ USw 22228 ex Lane-Poole) Hoo D. quercifolia (White : F rancis — — — FPAw 6437 x Lane-Poole) Hoogl. D. reticulata Kin Forest Dept. 422 Malaya FHO FHOw 7602 D. reticulata King _ Malaya — KEPw 7261 D. reticulata King — Malaya — KEPw 5779 D. retusa Thunb. — Ceylon — Yw 9781 D. scabrella i Don) Forest Dept. 607 Burma — FHOw 1676 Roxb. ex D. sss (White) Hoogl. — Solomon Is, — FPAw 8226 D. solomonensis em Hoogl. F.S. Walker 145 Solomon Is. — FHOw 17957 D. sumatrana — — — KEPw 403 8T WOLAYOdAY GIONYV AHL AO TVNUAOL Sb “10A | D. sumatrana Miq. D. triquetra (Rottb.) Gilg D. triquetra (Rottb.) Gilg D. turbinata Finet & Gagnep. D. turbinata Finet & Gagnep. Doliocarpus sp. dD. casein & cc.) G 2. oat ha Standl. D. lasiogyne Benoist D. major Gmel D. olivaceus Sprague & Wms. ex Stand. Doliocarpus sp. Hibbertia coriacea sees f.) Gilg H. cuneiformis S H. lucens Brongn. & Griseb. H. lucens Brongn. & Griseb. H. lucens Brongn. & Griseb. H. saligna R. Br. ex DC. H. tetrandra (Lindl.) Gilg Pachynema complanatum R. Br. ex DC. Schumacheria castaneifolia Vahl Tetracera boiviniana Baill. T. scandens (L.) Merr T. volubilis L. Sun Yatsen Univ. 0835 Tang S. J. Record A. Ducke 327 B. A. Krukoff 7020 Cuatrecasas 16556 Stern et al. 549 Reitz & Klein 5776 B. A. - Krukof 11189 A. C. Smith 3546 Aa — 3045 Wilson 848 Von Mueller H. J. Schlieben 5187 H. Leén Ceylon Ceylon Hainan China British Honduras Brazil Brazil Colombia Panama Bolivia New Caledonia Pemberton, W. Australia Fiji Fiji New Caledonia New South Wales New Sout Wales lon Ceylo Tanganyika Philippines Cuba SYS Fan Mem. Inst. Chungking y US F, VALLE US, MO KEPw 382 Yw 9807 USw 21889 Yw 29564 USw 8597 Yw 8831 Yw 34091 Yw 37027 Yw 43052 USw 16329 USw 15034 USw 16485 USw 16042 Fw 11189 FPAw 23183 FPAw 11684 Yw 28402 Yw 27903 FPAw 12037 FPAw 13034 WILSON 848 Aw 26225 Aw 2632 Yw 29962 S.M. Yw 40713 [£961 I ‘AVAOVINATIIC ‘NOSIMOIG 20 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 LITERATURE CITED Ayensu, E. S., & W. L. STERN. 1964. Systematic rer he = ontogeny of the stem in Passifloraceae. Contr. U. S. Nat. Herb. 34: BatLey, I. W. 1944. The development of vessels in eae. and its signifi- cance in morphological research. Am. Jour. Bot. 31: 421-428. . 1951. The use and abuse of anatomical data in the study of phylogeny and classification. Phytomorphology 1: 67-69. - 1953. Evolution of tracheary tissue in land plants. Am. Jour. Bot. 40: 1957. The potentialities and limitations of wood anatomy in the phy- logeny and classification of angiosperms. Jour. Arnold Arb. 38: 243-254. .& - Howarb. 1941. The comparative morphology of the Icacina- ceae. I. Anatomy of the node and internode. Jour. Arnold Arb. 22: 125-132. » oa; Nast, & A. me SMITH. 1943. The family Himantandraceae. Jour. Arnold Arb. 4 190-20 . P. THompson, aa Additional notes upon the angiosperms Tetra- centron, Trochodendron, and Drimys. Ann. Bot. 37: 503-512. & W. . Tupper. 1918. Size variations in tracheary cells. I. A compari- son between the secondary xylems of Meee cryptogams, gymnosperms, and angiosperms. Proc. Am. Acad. 54: BarLton, H. 1871. The Natural History f ae Vol. I. (Transl. by M. M. Harrec.) L. Reeve & Co., London. Bauscu, J. 1938. A revision of the Eucryphiaceae. Kew Bull. — 317-349. BENSON, L. 1957. Plant Classification. D. C. Heath & Co., Bos BENTHAM, G. & J. D. Hooker. 1862. Genera Plantarum. 1. s pore & Co., London. Bessey, C. E. 1915. The phylogenetic taxonomy of flowering plants. Ann. Missouri Bot. Gard. 2: 109-164. Borvin, B. 1956. Les familles de trachéophytes. Soc. Bot. oo 103: 490-505. CANDOLLE, A. P. pe. 1824. Prodr. Syst. Nat. Reg. Veg. 1: Par Canricut, J. E. 1955. The comparative morphology and a of the Magnoliaceae IV. Wood and nodal anatomy. Jour. Arnold Arb. 36: 119- oo Contributions of pollen iad to the phylogeny of some ranalean families. Grana Palyn. 4: 64-7 CHADEFAUD,, M., L. EmBercer. 1960. eds de Botanique Systématique. ris CHALK, L. 1938. Standardization of terms for vessel diam Trop. Woods 55: CuHaTtaway, M. M. a ‘Proposed standards for numerical values used in describing woods. Trop. Woods 29: 20-28. . 1937. The wood aged the family Sterculiaceae. Philos. Trans. Roy. Soc. London B. 228: 366. - 1955. Crystals in sige ee I. Trop. Woods 102: 55-74. CueEapLe, V. L. 1942. The occurrence and types of vessels in the various organs of the plant in the Monocotyledoneae. Am. Jour. Bot. 29: 441-450. The origin and certain trends of pa aeons of the vessel in the Monocotyledoneae. Am. Jour. Bot. 30: 11- - 1943b. Vessel specialization in the late elec ile of the various organs in the Monocotyledoneae. Am. Jour. Bot. 30: 484490, eter and ray width. 1967 | DICKISON, DILLENIACEAE, I 21 ——. 1944, cami of vessels within the xylem of each organ in the Monocotyledoneae. Am. Jour. Bot. 31: 92. CHowpuHoury, K. A., & S. S. GHosH. 1958. Indian Woods, their identification, properties and uses. Vol. 1. Northern Circle Survey of India. Dehra Dun. 304 pp. COMMITTEE ON NOMENCLATURE, INTERNATIONAL ASSOCIATION OF Woop ANAT- OMIsTS 1957. oe glossary of terms used in wood anatomy. Trop. Woods 107: 1 1964. wae Glossary of Terms used in Wood Anatomy. Zurich. ComMrrtee OF THE STANDARDIZATION OF TERMS OF CELL Size, International saecbiaei of Wood Anatomists. 1937. oe terms of length of vessel mbers and wood fibers. Trop. Woods 51: fhesetncs H. F. 1957. Forecast of a system of ri dicotyledons. Madronio 14: 1-9, Cronguist, A. 1957, Outline of a new system of families and orders of dicoty- ledons. Bull. Jard. Bot. Bruxelles 27: 13-40. . 1965. The status of the general system of classification of flowering plants. Ann. Missouri Bot. Gard. 52: 281-303. Eames, A. J. 1961. Morphology of the Angiosperms. McGraw-Hill. New York. Frost, F. H. 1930a. Specialization in secondary xylem of dicotyledons. I. The origin of vessels. Bot. Gaz. 89: 67-94. . Specialization in secondary xylem of ai eon III. Special- ization of the lateral walls of vessel segments. Bot. Gaz. 91: 88-96. Gre, E. Bek! Dilleniaceae. Nat. Pflanzenfam. III. 6: 100-128. E. WERDERMANN. 1925. Dilleniaceae. Nat. Pflanzenfam. Band 21: 7. GOTTWALD, H & N. PARAMESWARAN. 1964. Vielfache Seep ieninbeenpsies in der Familie Dipterocarpaceae. Zeitschr. Bot. 52: 321-334. GuNDERSEN, A. 1950. Families of Dicotyledons. kl Botanica. Waltham, Mass HALLER, H. 1905. Provisional scheme of the natural (phylogenetic) system of flowering plants. New Phytol. 4: 151-162. Hess, R. W. ee Classification of wood parenchyma in dicotyledons. Trop. Woods 96: 1- Hoocranp, R. D. , et Dilleniaceae. Jn: VAN STEENIS, Fl. Males. I. 4: 141-174. 52. A revision of the genus Dillenia. Blumea 7: 1-145. 1953. The genus Tetracera (Dilleniaceae) in the eastern Old World. Reinwardtia 2: 185-225. Hutcuinson, J. 1926. The Families of Flowering Plants. I. Dicotyledons. MacMillan & Co., London Kenc, H. 1962. Comparative morphological studies in Theaceae. Univ. Calif. Publ. Bot. 33: 269-384. Kriss, D. A. 1935. Salient lines of structural specialization in wood rays of dicotyledons. Bot. Gaz 547-557. Kuntze, 0. 1891. Revisio generum plantarum. H. Stiirtz. Wiirzbur LAWRENCE, G. H. M. 1951. Taxonomy of Vascular Plants. MacMillan & Co., ew Y rk. LE Maovt, es & J. Decatsne. 1873. System of Botany. Longmans & Green. London Linb-ey, J. "1830. An Introduction to the Natural — of Botany. Vol. L. Ed. enero Rees, Orme, Brown & Green. Lon MELcutor, H. . Dilleniaceae. Jn: ENGLER, ak der Pflanzenfamilien. lin Zwilfte hates. Band II. Gebriider Borntraeger. Ber 22 JOURNAL OF THE ARNOLD ARBORETUM [voL, 48 MetcatreE, C. R., & L. CHALK. 1950. Anatomy of the Dicotyledons. 2 Vols. The Clarendon Press. Oxford. Mott, J. W., . H. Janssonius, 1906. Mikrographie des Holzes der auf Tava Sa Wee 50 Baumarten, I. Leiden Pearson, R. S., & H. P. Brown. 1932. Conmercial Timbers of India. 2 Vols. Govt. of India Central Publ. Branch. Calcutta. Recorp, S. J., & R. Hess. 1943. Timbers of the New World. Yale University idge. RENDLE, B. J., & S. H. CLarKe,. 1934. st diagnostic value of measurements in wood anatomy. Trop. Woods 40: 27-4 STANDLEY, P. C. 1928. Flora of the ies Canal Zone. Dilleniaceae. Contr. U.S. Nat. Herb. 27: 264-266 & O. WiittaMs. 1961. Flora of Guatemala. Dilleniaceae. Fieldiana 24: 1-12 Stern, W. L. 1954a. A suggested classification for intercellular spaces. Bull. Torrey Club 81: 234-235. . 1954b. eri oenets anatomy of xylem and phylogeny of Lauraceae. if & rop. Woods 100: 1-72 K.L _ CHAMBERS. 1960. The citation of wood specimens and herbari- um vouchers in anatomical research. Taxon 9: 7-13. Swamy, B. G. L., & I. W. Battey. 1949. The morphology and relationships of Cercidiphyllum. Jour. Arnold Arb. 30: 187-210. cigar naa A. L. 1958. Origins of oe plants. (Transl. by O. H. GANKIN.) Am. Inst. Biol. Sci., Washingto he saw, W. P. 1923. The relationships of 6h different types of angiospermic vessels. Ann. Bot. 37: 183-192. Trepo, O. 1941. A list of diagnostic characters for descriptions of dicotyledonous woods. Ill. Acad. Sci. Trans. 34: 105-106. 1946. The role of wood anatomy in phylogeny. Am. Midl. Nat. 36: 362— VANDER WYK, R. W., & J. E. CanricHT. 1956. The anatomy and relationships of the Annonaceae. Trop. Woods 104: 1-24. VesTAL, P. A. 1937. The significance of comparative anatomy in establishing the relationship of ed Hypericaceae to the Guttiferae and their allies. Philippine Jour. Sci. 64: 199-256. Warmine, E. 1911. Handbuch systematischen Botanik. Gebriider Borntraeger. Berlin WErTsTEIN, R. 1935. Handbuch der systematischen Botanik. Leipzig & Wien. Witson, C. L. 1965. The floral a of the Dilleniaceae. I. Hibbertia Andr. Phytomorphology 15: 248 Witson, T. K. 1960. The ied morphology of the oo I. Synop- sis of genera and wood anatomy. Trop. Woods 112: . 1964. ata morphology of the ncaa “TIL. Pollen. Bot. 197, Gaz. 125: 192— . 1965. The comparative ama a of the pe gay II. Anatomy of pe bbe stem and node. Am. Jour. Bot. 52: 369-3 The comparative id of the sie IV. Floral oes and conclusions. Am. Jour. Bot. 53: 336-343 DEPARTMENT OF BroLocy VIRGINIA POLYTECHNIC INSTITUTE BLACKBURG, VIRGINIA 24601 1967 | DICKISON, DILLENIACEAE, I 23 EXPLANATION OF PLATES PLATE I Fics. 1-4. 1, Schumacheria castaneifolia Vahl, cross section of he sapideten sues angular vessels and sparse apotracheal parenchyma, * 130. 2, end wall of vessel member, 130. 3, Hibbertia lucens ee & Griseb., cross section of wood illustrating mostly solitary pores, & 130. 4, the same, ‘radial section, 70. PLATE II 5-8. 5, Dillenia ingens (B. L. Burtt) A. C. Smith, cross section of wood downs solitary vessels, apotracheal eee ms parenchyma, 60. 6, the same, tangential section showing heterogeneous rays and oblique angle of nei i lent ‘a on Martelli ex Dur. cks., cross section of wood illustrating solitary pore: S, thick- walled fiber tracheids, apotrac cheal parenchyma, and dark-staining deposits in ray parenchyma, * 100. 8, Dillenia bint Finet agne ongitudinal section of wood showing eotse opposite to transitional intervascular pitting, « PLATE III Fics. 9-13. 9, Acrotrema ve Hook., cross section of stem with limited secondary xylem; note capt ige xX . 10, Pach omplanatum R. Br. ex DC., cross section of stem with limited secondary xylem, 1a ES Pires uniflorum. dca. section of stem illustrating oblique angle of many-barred scalariform perforation plate, 2, the s longitudinal section of ste i calariform perforation plate, chynema showing a complanatum, eA Be ce section of stem showing scalariform to opposite inter- vascular pitting, « 400 E IV Fics. 14-17. 14, Curatella americana L., cross section of wood showing solitary vessels, broad ys, aggregate to narrow-banded and pyciaciag parenchyma, Me Bay TS; rugosa Poir., cross section of wood showi very large cir- oO b Q ee = cular vessels and broad rays, X uratella americana, a ngential section of wood, 1 the same, radial section flenesting obliquely oriented, simple perforation ee < 400. Fics. 18-21. 18, Tetracera volubilis L., cross section of wood showing solitary pores and broad rays containing dark gummy deposits, 60. 19, T. boiviniana intervascular pitting with es elongate pits, X 4 isis VI Fics. 22-27, Vessel Member n Dilleniaceae. 22, Schumacheria cas- taneifotia Vahl, * 400. 23, Dillons peste yg Gilg, « 400. 24, Dil- lenia ingens (B. L. Burtt) A. C. Smith, < 400. 25, Davilla aspera (Aubl.) Benoist, < 400. 26, Doliocarpus dentatus (Aubl.) Soa, x 400. 27, Tetra- cera volubilis L., x 4 Jour. ARNOLD Ars. VoL. 48 PLATE I SOR Ware 160enttes eae woh by = @2@ S600," . *, é eae Tid asst @raen ee . mantel 1 * eeek NAN Se Seem ns Ty VL 4964 mains oat hp PK 08 re ‘. . tt r6'5'8 4 f ft ANY oo ie t ee; eS irje SD eapite Sars eee te nage re samen” ~ | —— a » 5 @. at & ; Crest: € ces Eo satis +9 ‘ ah! if i f Dickison, DILLENIACEAE. I Jour. ARNOLD Ars: VoL. 48 PLATE II Ss “ Na Wy bite My * * $ é : jas AO) ef @ ke SRY f i BRS tee, t4 ie TIA SS AR 6g =. a Oe, 6: a, hak OM y DO, eS) ') lags a eabe ty. Ware ce FRR bey "8 His ; ’ re HI No Setat ae Seal ets: Ooh, liege ‘ By ‘a j! | . 4 ie A ‘ig UA areal Sewn Oyen: u u A oo) g = _ e DICKISON, DILLENIACEAE, I Jour. ARNOLD Ars. VoL. 48 Piate IIT DICKISON, DILLENIACEAE. I PLATE IV Jour. ARNOLD Ars. VoL. 48 « See iia oan 2 Z ap ° Sere . 4 sere rt LOPE DIOR yy a Lap es a Se oe AE Op a, A < * care 8: : . s as SS aa DICKISON, DILLENIACEAE, I le ee " : sial " i LRP Y y rg Oe ee : Pe. «tAd PLATE V Jour. ARNOLD Ars. VOL. 48 DIcKISON, DILLENIACEAE, I Jour. ARNOLD Ars. VoL. 48 PLATE V1 ie X % DicKk1son, DILLENIACEAE, I 30 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 A SYNOPSIS OF AFRICAN SPECIES OF DELPHINIUM AND CONSOLIDA Puitie A. Munz LIKE SEVERAL OTHER HERBACEOUS GENERA of the Ranunculaceae, no- tably Isopyrum, Coptis, Actaea, Aquilegia, and Aconitum as examples, Delphinium is of the North Temperate Zone, but it differs remarkably in its far southern range on to the higher mountains, even, of central Africa. My interest in making some attempt to understand the species of this widespread group began when preparing a treatment of cultivated forms for Hortus III, since I had considerable misgiving as to the correct names and limits of some of the species involved. With the American species fairly well understood (cf. Ewan, J. A., A synopsis of the North American species of Delphinium. Univ. Colo. Stud. D. 2(2): 55-244. 1945; and Lewis, H. & C. Epiine, A taxonomic study of Californian delphiniums. Brittonia 8: 1-22. 1954), and with European species well summarized by A. O. CHaTER & B. PAWLOWSKI (cf. Flora Europaea 1: 213-217. 1964), one turns naturally to Asiatic and African species as most needing study. The latter continent has the fewer and it seems well to treat these first, hence the present paper. The much more complex and larger Asiatic representation will be treated in subsequent articles. In both the Flora Europaea cited above and in P. H. Davis, Flora of Turkey 1: 108-134. 1965, two genera are recognized: Delphinium, sensu stricto, and Consolida. The former is characterized by four separate petals (nectaries, honey-leaves), the two upper being spurred and secreting nectar, the spurs inserted into the sepal spur, the two lower petals not nectariferous, merely with a narrow claw and dilated lamina. Stamens in 8 spiralled series. Carpels 3, rarely 4 or 5. An almost entirely perennial group which includes the plants cultivated as “delphiniums.” Consolida, on the the other hand, has a single ‘“‘petal” (nectary or honey- leaf), said to be formed by the union of the two superior petals of Del- phinium, and one spur. The two lower petals are said to be lacking. Stamens in 5 series. Carpel solitary. Plants annual. Here fall the species known in gardens as “arkspurs.” I am following the above usage in my series of proposed papers. In his monograph of Delphinium, sensu lato, Huth (Bot. Jahrb. 20: 322-499. 1895) has rather an elaborate scheme of sections and tribes, which in many cases do not seem to me to represent natural groups and which separate widely quite closely related species. I am not prepared to enter into any such scheme of subgeneric categories and can do no more than offer artifi- cial keys for identification of species. Under each species treated synonymy 1967 | MUNZ, DELPHINIUM AND CONSOLIDA Si is given at least for the geographic area involved, with a description and general distribution, and with mention of some representative illustrations in the literature. Where material has been available, there is shown a line drawing of the petals or “petal.” Some of the principal exsiccatae are listed with the name of the collector and his collection number, by country, but without stating the exact locality or herbaria in which deposited. To the curators of the herbaria at the following American institutions, from which material was borrowed, I express herewith my very great appreciation: Arnold Arboretum (a) Bailey Hortorium (BH) California Academy of Sciences (cas) Field Museum, Chicago (F Gray Herbarium Missouri Botanical Garden (Mo) New York Botanical Garden (Ny) Academy of Natural Sciences, Philadelphia (PH) United States National Arboretum (NA United States National Herbarium (vs) University of California, Berkeley (uc) University of Michigan (micH) I was given, also, the privilege of working for some months at the British Museum of Natural History (pm), at the Royal Botanic Gardens at Kew (K), and the Muséum National d’Histoire Naturelle at Paris (Pp). In addition, I have had loans from the Royal Botanic Garden at Edinburgh (E), the Komarov Botanical Institute at Leningrad (LE), the Conserva- toire et Jardin botaniques at Geneva (G), the Naturhistorisches Museum at Wien (w), and the Botanisches Institut und Botanischer Garten der Universitit Wien (wu). I am greatly indebted to all the officers in charge of these collections. It is a pleasure also to acknowledge my obligation to the National Science Foundation for the financial aid received under Grant GB-2716. The present series of papers cannot be monographic in nature, but mere- ly synoptic, with the hope of presenting keys and means of identification as a useful step in our present knowledge. True monographic work must come later, when there is access to the regions inhabited by many of the species and observations can be made of them as living organisms. 1. Delphinium L. Delphinium, as a genus, L. Sp. Pl. 1: 530. 1753; Gen. Pl. ed. 5. 236. 1754. Huth, Bot. Jahrb. 20: 322-499. pis. 6, 7. 1895. As a subgenus, Eudel- phinium (Huth) Dalla Torre & Harms, Genera Siphonogamarum 165. 1901. Mostly perennial, more rarely biennial or annual, herbs. Leaves alter- nate, palmately divided. Flowers zygomorphic, mostly in racemes or 32 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 panicles. Sepals 5, petaloid, the posterior one normally spurred. Petals four, in two unlike pairs, the upper pair, sometimes called nectaries or honey-leaves, produced into nectariferous spurs inserted into the sepal spur; the lower pair with a basal claw and expanded lamina. Stamens in eight spiralled series. Carpels 3 (—5), free, becoming follicles. TYPE sPEciEs, D. peregrinum L. KEY TO THE AFRICAN SPECIES OF DELPHINIUM, sensu stricto a. Plants perennial. b. Aaa glabrous, the upper with lateral wings on the limb. Morocco to OMe Ge otk shh el EER ete ee eee 2. D. balansae. b. eee bearded and/or ciliate. c. Spur 5-8 mm. long, ca. 4 mm. wide at base; upper Dae short-pubes- cent toward the tips. Ethiopia southward. .... D. dasycaulon. c. Spur longer; ee petals glabrous. The spur 12-15 mm. long; flowers bluish. é. nied leaves 10-15 cm. in diameter; racemes many-flowere d; aban 3; seeds blackish, 2-3 mm. long. E. Algeria and Radical leaves 3-7 cm. in diameter; racemes 4-10-fowered: e. carpels 5 (3, 4, or 6); seeds dark, 15-2 mm. long. Morocco to Tunis. f. Carpels nearly always 5, divergent; sepals 12-16 mm. BO ese Gthatnens Coie cee 7. D. pentagynum. ng. f. Carpels 3, erect; sepals 10-13 mm, lon g. Sepals bright blue, distinctly shorter than Soe spur. as eB Peeing tay ais Medi genes, eee ees od vadense. g. phi bluish lilac, about as long as or shorter se the Terese Cine ace tne eT. | emarginatum. d. The spur poy 7a m. long; flowers white to aN pinkish or bluish. Tropical Africa at higher elevations h. Spur slender, 3-4 mm. wide at the base, 30-45 mm. long; upper petals glabrous; sepals mostly 25-40 mm. long. ............ A is ROP ae On INS Chir DLE ooo oA mare eels D. leroyi. h. Spur thick, 5-7 mm. wide at base, 20-26 mm. ne upper petals pubescent; sepals mostly 10-15 i. Ee MOR ed Ne antucuite) eeee aes eon ae ge ty Ue 5. D. macrocentron. a. Plants annual or biennial. i. Middle cauline leaf blades 10-20 cm. long, 5—9-parted; blade of lower petal asymmetrically oblanceolate, gradually meee into the claw. ee a ORE Nanak (nigh ae eT he Oe SE D. staphisagria. Middle cauline leaf blades mostly 2-3 cm. long; blade i the lower petals asymmetrically oblong to rounded. j. Carpels 3; flowers few to several, the receptacle not inflated. k. Limb of lower petals ovate to es cuneate at base, gradually narrowed into claw. Eastern Mediterranean. 1. The limb of lower petals more or fe elliptic, often about as long as the claw; plant subglabrous to on mad pubescent, 2-7 dm. tall; racemes several- to many-flowered. ............ se 1967] MUNZ, DELPHINIUM AND CONSOLIDA 38 1. The limb of lower petals almost round, shorter than the claw; plant grayish pubescent, 0.3-1.5 dm. tall; racemes mostly d. 12 oa eS as 3 o ° eh — ° = oO ta 42 oO ine i co wa rst} = tnt i= uo) or aa mS a ° =| = a ry a -2§ o Qo. oa aa ° nae io oO 5, i} as i=] jo} et cuneate at the base. Western Mediterranean. m. The limb of the lower petals not or little shorter than the claw, usually distinctly exserted, not cordate at the base. .... Pe eenrare mee ltr en ent hye, 13. D. obcordatum. m. The limb of the lower petals 12—*4 as long as the claw, cordate or subcordate at the base. n. Limb of lower petals oblong, 144-2 times as long as wide, WROTE 55 ea Se en ee 14. D. gracile. n. Limb of lower petals suborbicular, included. o. The limb of the lower petals strongly cordate at the base. Uncommon. 15: Be o. The limb of the lower petals weakly cordate at the 16 1. Delphinium staphisagria L. Sp. Pl. 1: 531. 1753. Fic. 1. Staphisagria macrosperma Spach, Hist. Nat. Vég. Phan. 7: 348. 1839. Sibth. & Sm. Fl. Graeca 6: pl. 508. 1826. Annual or biennial, usually 1-stemmed, simple or few-branched above, 3-12 dm. tall, leafy, soft hairy with fine, often long, spreading hairs and densely covered also with short retrorse hairs; leaves large, the blade rounded or broader, to 2 dm. in diameter, more or less finely strigose and with longer hairs, 5—9-parted, the segments entire or 2-3-lobed, these ultimate divisions broadly lanceolate, sharply acute, 3-veined, often entire or with few large teeth; petioles mostly longer than the blades, dilated at base, with both kinds of pubescence; racemes terminal, dense or lax, usual- ly many flowered on main stem, few flowered on branches; lower bracts foliose, 3-parted or entire, sessile or petioled, from half as long to as long as pedicel; pedicels 0.5—5 cm. long, densely hairy and strigose, bibracteo- late near the base; flowers mostly deep blue, 2—3 cm. in diameter; calyx fine-hairy without, the upper sepal oblong-ovate, 1-2 cm. long, the saccate spur 2-3 mm. long and wide; lateral sepals ovate, ca. 1.5 cm. long, rounded at apex, lower sepals similar but slightly shorter; upper petals white or bluish, pubescent, obtuse, simple or more or less bilobed at summit, 10— 12 mm. long, short-spurred at base; lower petals glabrous or pubescent, more or less spatulate or obovate-cuneiform, somewhat asymmetrical, ca. 14-16 mm. long; stamens 8-10 mm. long, filaments ciliate above, glabrous and dilated below; carpels 3, long-pubescent, inflated and 12-20 mm. long at maturity; seeds few, 5—6 mm. long, blackish, obpyramidal, 3-angled, reticulate-foveolate. Type not given; described as from Istria, Dalmatia, Apulia, and Cala- bria. 34 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 DIsTRIBUTION. Mostly in woods and rocky places about the base of mountains; Canary Islands; Mediterranean region in Morocco, Algeria, southern Europe; Syria, Palestine, Turkey. ILLUSTRATIONS. SIBTHORP & SMITH, FI. Graeca 6: pl. 508. 1826; Coste, Fl. France 1: 51. 1901; Marre, FI. de l’Afrique du Nord 11: 84. 1964. REPRESENTATIVE SPECIMENS. Can Islands: Bornmiiller 2013; Bourgeau 1231, 428. Morocco: Font Quer 107. Algeria: Alleizetti, 1953; Le Cesve, 1938; Cosson, 1854; G. Munby, 1844, 1850; Durando, 1855. Delphinium staphisagria is a species quite unrelated to any others. To begin with, there are few annual species in Delphinium and they differ widely from this one in appearance and technical characters. The leaves and terminal spikes of flowers give the general appearance of a lupine until they are examined more closely. 2. Delphinium balansae Boiss. & Reut. in Boiss., Diagn., II. 5: 12. 1856. Fic. 2. D. balansae formae pallidum and caeruleum Maire, Fl. de l’Afrique du Nord 11: 83. 1964. Perennial from a somewhat woody crown above a taproot system, but sometimes flowering in its first year; stems slender, several suberect, more or less strigulose to subglabrous, 2—7 dm. tall, simple or with some branch- ing almost throughout, slightly angular; basal and lower leaves long- petioled, tripartite, the cuneate parts trifid, the ultimate segments linear to oblong, subglabrous, sometimes more or less ciliate, callous-tipped, the whole blade 2-5 cm. wide, somewhat shorter; cauline leaves simpler, shorter-petioled, quite persistent, the uppermost linear, entire; racemes rather lax, 2-10-flowered; bracts linear, 5-12 mm. long; pedicels strigose, 5-25 mm. long, the bracteoles ca. 4.5 mm. long; flowers whitish with a bluish or greenish tinge, or blue; the upper sepal ca. 6-7 mm. long, some- what strigulose and with an almost straight or slightly curved horizontal strigulose spur 3-4 mm. wide at the base, 12-16 mm. long; lateral sepals oblong, ca. 8 mm. long, strigose along the midrib, the lower pair strigose over the entire outer surface and more pointed; upper petals glabrous, 2-lobed, slightly shorter than the sepals and ca. 7 mm. wide; lower petals ca. 8 mm. long, oblong-oval, sometimes with a cordate base, abruptly nar- rowed to a narrow claw; stamens somewhat shorter than the sepals, the filaments dilated below, glabrous; carpels 3, pubescent, 8-9 mm. long in fruit; seeds dark, subglobose, ca. 2 mm. in diameter, with rings of scales. Type: near Lambese, North Africa, Balansa 1853, as D. junceum (photo. BH). DIsTRIBUTION. Calcareous areas below 2500 m., Morocco to Tunis. ILLustRATIoNS, Marre, Fl. de l’Afrique du Nord 11: 82. 1964. REPRESENTATIVE SPECIMENS. Morocco: O. Polunin 2138; Gattefossé 220b; 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 35 Font Quer 209; Jahandiez 742, 858; E. K. Balls 3133. Algeria: Kralik, June 2, 1858; Cosson, July 26, 1854, July 22, 1861; Faure, Aug. 3, 1933; Chevallier 255. Material at British Museum has the pale-flowered plants more erect and with stouter stems, the blue-flowered plants more spreading and diffuse, and with slender stems. 3. Delphinium dasycaulon Fresen, Mus. Senckenb. 2: 272. 1837. Fic. ? D. dasycaulon var. minor Cufodontis, Bot. Inst. Univ. Wien 39: 103-126. 1958. Perennial from a woody elongate root, the stem largely solitary, weakly erect, 5-15 dm. tall, slender, scattered villous, more or less openly branched above; basal and lower cauline leaves 1—1.5 dm. wide, with slender petioles 1-4 dm. long, the blades 3—5-parted into broad incised lobes, the upper leaves remote, reduced, 3-parted into simpler divisions, often ca. 1 cm. wide; inflorescence of few open subnaked branches with remote bracts ca. 1 cm. long; flowers often more or less crowded at ends of branches, bright blue to paler; pedicels 5-20 mm. long, densely pubescent, the 2 bracteoles at the middle or above, entire, linear or nearly so, 5-10 mm. long; sepals oblong to obovate, 1.3—1.8 cm. long, the upper with an outer brownish spot, crisp-pubescent, the spur 5-8 mm. long, ca. 4 mm. wide at the base, pubescent, more or less decurved, lateral sepals strigose along the middle, lower pair strigose along upper half from base to tip; upper petals cream, more or less lanceolate, 8—14 mm. long, 2-toothed at apex, short-pubescent upward, the spur 3-6 mm. long; lower petals blue, 10-14 mm. long, the limb oval, unequally bifid, bearded, narrowed into a claw 5—6 mm. long; stamens 5-8 mm. long, somewhat pubescent, dilated at the base; follicles 3-5, crisp or spreading pubescent, 12-20 mm. long, 3 mm. wide, the beak 2 mm. long; seeds covered with numerous imbricate scales. Type: Simen, Ethiopia, Riippell. DIstRIBUTION. Open woodland and grassland, 1400-3000 m., Ethiopia, Karara Hills of eastern Sudan, eastern Belgian Congo, Kenya, Tanganyika, Kirk Range of Mozambique and eastern Nigeria. REPRESENTATIVE SPECIMENS. Ethiopia: Schimper 397, 752; Scott 231; Curle 140, 380; Archer 8657; F. G. Meyer 7465; Schweinfurth & Riva 1472; Pappi 390, 4811; Mooney 5820, 6264; Bally 7035, 10033; Massey 62. Northern Rho- desia: Burtt 6114, 6214; Nash 274; Milne-Redhead 444; F. White 2742. Sudan: Robbie 6. Kenya: Mooney 7528. Belgian Congo: Dubois 1049; Kassner 2748; de Witte 343; Robyns 2155. Tanganyika: Chapman 321; Brass 16990, 17150; Jackson 1860. Nigeria: Migeod 344; Lightbody 26257; Tamajong 22210; Hep- ber 2847; Unwin 214. Mozambique: Mendonga & Wild 994. Delphinium dasycaulon has a remarkable distribution from Ethiopia to the Congo, Kenya, Mozambique and Nigeria. It is quite distinct from 36 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 48 Fics. 1-6. Petals in Delphinium (length given for upper petal includes spur). Fic. 1. D. staphisagria. Left-hand figure, upper petal (16 mm. long) with spur to left; right-hand drawing, of lower petal, 10 mm. high, from Alleizetti, May 1, 1953 (mo). Fic. 2. D. ‘ , , lo petal 7.5 mm. long, from Chevallier 255 (¥F). Fic. 3. D. dasycaulon. Spurred petal 14 mm. long, lower petal 13 mm., from Pappi 4811 (mo). Fic. 4. D. leroyt. Spurred petal 48 mm. long, from Greenway 9711 (Ny); lower petal 6 mm. high, from MacInnes 206 (mo). Fic. 5. D. macrocentron. Spurred petal 38 mm. long, lower petal 14 mm., from Piemeisel & Kephart 553 (us). Fic. 6. D. sylvaticum. Spurred petal 22 mm. long, lower petal 10 mm., from St. Lager, 1909 (Ny). the other similarly distributed species (D. leroyi and D. macrocentron) in its much smaller flowers, especially in its shorter spurs. It would, there- fore, be expected to have quite different pollinators. In spite of going so far into the tropics, these three species inhabit areas with temperate climates. 4. Delphinium leroyi Franch. ex Huth, Bot. Jahrb. 20: 474, pl. 5.1895. D. goetzeanum Engl. Bot. Jahrb. 30: 308. 1902. Type, from the Ruhudge River, at 1900 m., Ubena, March 25, 1899. At Kew in a type cover is Goetze 804 from Ubena. 6, fig. Fie. 4. 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 37 D. wellbyi Hemsl. Kew Bull. 1907: 360. 1907. Based on a type from be- tween Harrar and Addis Ababa, Wellby, 1898 (x!). D. macrocephalum Engl. Bot. Jahrb. 45: 267. 1910. Based on Jaeger 434 from Wanegehochland, between Olmoti and Ossirwa. Not seen. D. gommingeri Volkens ex Engl. Bot. Jahrb. 45: 266, 269. 1910. The type, Gomminger from Kilimanjaro, not seen. D. ruspolianum Engl. Bot. Jahrb. 45: 267. 1910. Riva on the Ruspoli Expedi- tion, Dscharibule, Gallahochland. Not seen. D. candidum Hemsl. Bot. Mag. 133: pl. 8170. 1907. From between Mt. Elgon and Kisumu, Uganda, grown in the garden of Lady Hindslip in 1904 (x!). Perennial from a woody rhizome, the stem mostly one, erect, with scattered slender hairs below, retrorsely crisp pubescent above, 4-15 (-20) dm. high, slender, branched in the inflorescence; basal and lower leaves with petioles 12-35 cm. long, sparsely pilose; the midcauline petioles shorter, dilated at the base, retrorsely pubescent; upper remote, passing into sessile bracts; blade of lower leaves to ca. 1 dm. wide, more or less reniform-rounded in outline, palmatifid into 5 main cuneate-obovate divisions, these in turn 3-lobed and ending in ultimate broad teeth; main cauline leaves more narrowly parted and sharply toothed; inflorescence paniculate with few-flowered lax racemes; bracts lance-linear, mostly en- tire, 1-3 cm. long; pedicels subérect, 4-12 cm. long, with 2 bracteoles near middle or in upper half, these largely 5-15 mm. long; flowers fragrant, white, sometimes with a purplish, pinkish or bluish tinge, or deep blue; sepals with a brownish spot outside, oblong to obovate, 2.5—4 cm. long, the lateral subtruncate, the others more or less pointed, all somewhat crisp pubescent without, the spur horizontal, curved upward, 3-4.5 cm. long; upper petals lance-obovate, the blade 2-3 cm. long, 5-9 mm. wide, gla- rous, more or less deeply 2-toothed or -lobed at apex, the spur 3—4.5 cm. long; lower petals clawed, the limb lanceolate or lance-obovate, entire to unequally bilobed to 3-lobed, mostly 0.6-1.5 cm. long, 0.5-1 mm. broad; stamens 6-9 mm. long, the filaments dilated below, more or less ciliate, anthers black or dark brown, sparsely pubescent; follicles 3, more or less pubescent, ca. 2—2.5 cm. long, 3-4 mm. thick, the beak 3 mm. long; seeds dark, ca. 1.5 mm. in diameter, covered with numerous small imbricate scales. Type: Kilimanjaro, Leroy (P!). DistrrpuTion. At about 1800-3000 m., Ethiopia, southern Sudan, Congo, Tanganyika, Kenya, and eastern Northern Rhodesia. ILLustrations. Bot. Mag. 133: pl. 8170. 1907; Gard. Chron. III. 74: 217. 1923, 88: 193. 1930. REPRESENTATIVE SPECIMENS. Ethiopia: Hiller & Lithgoe 516; F. G. Meyer 7339; Scott 30; Archer 8619; Mooney 6705, 7012; Gillett 14559, 14936. Sudan: Jackson 405: J. G. Myers 11688; H. B. Johnston 13531; Chipp 70. Kenya: Greenway & Doughty 8532; Rogers 412; Anita Curtis 674, 679, 693; Snowden 501; Chandler 996; Maitland 1263; Tothill 2608. Tanganyika: Schlieben 838, 5084; Greenway 9711; Brass 17172, 17269; Duncan 48117. Northern Rhodesia: P. White 2741. 38 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 48 Dr. W. A. Archer, formerly of the National Arboretum at Washington, D.C., told me some years ago that the natives in the region of Addis Ababa, Ethiopia, collect large quantities of Delphinium leroyi for the cut- flower market, partly because of the size of the flowers and partly for their fragrance. As can be seen from the synonymy, the species varies greatly in flower color, but the floral morphology seems quite consistent, as do foliage and habit. 5. Delphinium macrocentron Oliver, Jour. Linn. Soc. Bot. 21: 397. 1885. Fic. 5. Perennial from a woody rhizome, mostly one-stemmed, the stem erect, coarse, hollow, 6—20 dm. tall, few-branched near summit only, more or less retrorse-pilose throughout, more densely spreading-pubescent above; lower petioles 0.3—3 dm. long, the blades 3-12 cm. broad, 3—5-parted into rather narrow divisions, these in turn laciniate into lance-linear sharply toothed segments; cauline leaves gradually reduced and simpler upward, the uppermost linear and entire to trifid, 3-4 cm. long; inflorescence openly few-branched, the racemes to ca. 10-flowered, rather dense; bracts lance- linear, commonly 1-2 cm. long; pedicels 2-5 cm. long, stout, erect with recurved apex; bracteoles opposite, 5-12 mm. long, in upper part of pedicel; flowers blue and greenish, especially the latter when dry; upper sepal spreading, round-ovate, 12-15 mm. long, densely pubescent without, the spur straight, erect, 2.2-2.6 cm. long, 5—7 mm. wide at base and almost as wide for its entire length; lateral sepals rounded, 10-12 mm. long, pubescent mostly along median part; lower sepals much the same, but more pointed and more generally pubescent; upper petals pale with greenish tinge, the blades 12-16 mm. long, 2-lobed and pubescent at tip, the spurs 2 or more cm. long; lower petals ca. as long, gradually expanded from the broad claw to a shallowly bilobed pubescent tip; stamens ca. as long as sepals, expanded and somewhat ciliate below, the anthers dark; carpels 3, pubescent, erect or recurved when mature, 2—2.8 cm. long, the beak 7-8 mm. long; seeds dark, rounded, ca. 1.5 mm. in diameter, covered with small imbricate scales. Type: Likipia, Kenya, J. Thomson (K!). DisTRIBUTION. Between 1500 and 3500 m., Kenya. ILLustTRATIONS. Hooker, Ic. 16: pl. 1501. 1886; Bot. Mag. 133: pl. 8151. 1907. REPRESENTATIVE SPECIMENS. Kenya: G. Taylor 1277, 1367, 1309; Anita Curtis 870; Greenway 7854; Pole Evans & Erens 1435; Greenway & Doughty 8519, 8521; C. W. Eliot 1545; Bogdan 547; Snowden 486, 817; Lugard 26; Hardy & Bally 10777. Delphinium macrocentron Oliver seems amply distinct from its closest relative, D. leroyi, as a rather local species of Kenya. The peculiar green shade of the flowers, especially as they dry, the thicker shorter spurs, the 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 39 pubescence of the upper petals, the mostly much shorter sepals set it quite apart. 6. Delphinium sylvaticum Pomel, Nouv. Mat. Fl. Atl. 382. 1875. Fic. 6. Perennial from a base covered with fibrous leaf vestiges, the stem usually solitary, erect, 4-8 dm. tall, more or less flexuous, striate, branched above, retrorse-hairy below, more spreading-hairy above: radical leaves gone by time of anthesis, the petioles to ca. 3 dm. long, ‘pubescent, the blades 10-15 cm. wide, pubescent, roundish in outline, 3—5-palmately-parted into obovate-cuneiform segments which are more or less deeply incised into lanceolate pointed ultimate divisions; cauline leaves scattered, gradually reduced up the stem, the upper tripartite with trifid segments and 3—6 cm. long; inflorescence an open panicle of many-flowered lax racemes to 2 dm. long; bracts simple to trifid with linear lobes and commonly 5-15 mm. long; pedicels densely pubescent, more or less glandular, 1—2.5 cm. long, the green linear bractlets above the middle, 3-6 mm. long; flowers a lively blue, ca. 2 cm. in diameter; upper sepal strigose, ovate-oblong, ca. 10 mm. long, the spur ca. 15 mm. long, mostly straight, horizontal or ascending, strigose; other sepals oblong-ovate, 10-12 mm. long, the upper pair strigose along the middle, the lower on the entire outer surface; upper petals almost as long, glabrous, cream, 2-toothed; lower petals bearded, bifid, ca. 10 mm. long, gradually attenuate to the broad claw; stamens subglabrous, 5-6 mm. long, bluish except for the dilated membranous lower half; follicles 3, oblong, 12-15 mm. long, more or less loosely pubes- cent; seeds blackish, 2-3 mm. long, oblong, subtrigonous, densely covered with elongate scales. TypeE: Guerouch, Beni-Foughal, Algeria, Pomel. Not seen. ; DistriBuTion. In fairly well watered places, largely in eastern Algeria and in TunIs!a. ILLustRATION. Marre, Fl. de l’Afrique du Nord 11: 71. 1964. Rep fevruer tae SPECIMEN. Algeria: D’Azazga a Takouren, Kabylie, Saint- Lager, July 18, 1909. Delphinium sylvaticum Pomel is apparently an endemic of rather local distribution and not well represented in herbaria. Of the perennial African Species of the genus it agrees with D. dasycaulon, D. leroyi and D. mac- rocentron in its rather large leaves and mostly three carpels, but is more northern in distribution. 7. Delphinium pentagynum Lam. Encycl. 2: 264. 1786. Fic. 7. D. pentagynum vars. homeotrichum, glabratum, phialotrichum, and heterotri- chum Maire, Bull. Soc. Hist. Nat. Afr. Nord 20: 209-210. 1929. A series based on pubescence, whether the hairs are bulbous at base or not, gland- tipped or not, appressed or not. 40 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 D. hybridum Willd. subsp. ae a & K.) Huth var. pentagynum (Lam.) Huth, Bot. Jahrb. 20: 341. 18 Perennial from a tuberous root, base of plant covered with fibrous re- mains of old leaves; stems one to few, 2—7 dm. tall, more or less pubescent below with rather long retrorse hairs, glabrous or pubescent above, mostly somewhat branched in the inflorescence; basal leaves in a rosette, often withered at anthesis, long-petioled, the blades rounded to subreniform in outline, 3-7 cm. broad, subglabrous to pubescent, palmately 3—5-parted Fics. 7-13. Petals in Delphinium (length given for bagi eet includes — al 12 long, 7, entagynum. Spurred petal 22 mm., low ure, May 31, 1908 (us). Fic. 8. D. cossonianum. Sted petal 10 Sat lower petal 5.5 mm. long, from diel araiacteat June 15, 1936 (mo). oi. ti coage Spurred petal 13 , lower petal 9 mm. long, from Balansa 728 (A G. 10. D. nanum. Senned: petal 25 mm., lower petal 13 mm. long, from penaiicothe 1877 (mo). Fic. 11. D. obcordatum. Spurred petal 22 mm., other 1 : , from Gattefossé, gee 18, 193 Fic D. gracile. Spurred petal 22 mm., other 11 mm. long, from Reverchon 249 (cu). Fic. 13. ean Spurred petal 20 tn mm., other 10 mm. long, from Faure, April 8, 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 41 into cuneiform-obovate segments, these more or less lobed into narrow or rather broad lobes or teeth, acute to mucronate; cauline leaves remote, short-petioled to subsessile, the segments commonly divided into linear lobes; inflorescence open, simple to several-branched, laxly 4—10-flowered on each branch; bracts green, the lowermost sometimes dissected, most are simple, linear, 3-10 mm. long; pedicels pubescent to glabrous, 3—20 mm. long, bibracteolate near the middle or below; flowers pale or dirty violet-blue, 2—2.5 cm. in diameter; upper sepal appressed pubescent with- out, ovate, 9-12 mm. long, the spur horizontal or ascending, straight or somewhat curved, pubescent, 12-16 mm. long, 3-4 mm. wide at base; lateral and lower sepals more or less oblong, 12-16 mm. long, the upper petals strigose along middle line, the others over the whole outer surface; upper petals glabrous, yellowish usually tinged with blue, 20-22 mm. long including the spur, 2-toothed at tip; lower petals with the limb oval, bearded, blue, bifid, 9-12 mm. long; stamens 7-8 mm. long, glabrous or sometimes ciliate, broadly dilated in lower half; follicles 5 (rarely 3, 4, or 6), usually pubescent, divergent, 12-14 mm. long including styles; seeds obpyramidal, dark, 1.2-1.5 mm. long, with long straight scales. Type: ‘on trouve cette plante dans le Portugal’; in Lamarck Herba- rium, Paris (p-LA). DIstr1BuTIon. In semiarid and wetter regions, below 1500 m., Morocco, Algeria, Iberian Peninsula. ILLUSTRATIONS. DESFONTAINES, FI. Atlant. 1: pl. 117. 1798; Marre, Fl. de l'Afrique du Nord 11: 68. 1964. REPRESENTATIVE SPECIMENS. Morocco: Font Quer 108, 153; Faure, May 1 1933; Fairchild 59. Algeria: Faure, May 31, 1908, May 19, 1936; Allard 2739: Durando 402; Balansa 575; M unby 35; Jamin 4 Among the small-flowered species along the northern fringe of Africa is the D. pentagynum group including that species and D. nevadense and D. emarginatum. They constitute the Series “Pentagyna” of Pawlowski (Fragm. Fl. & Geobot. 9: 436. 1963), whose key and treatment I am following. They are a western Mediterranean group of southern Europe and northern Africa. They are perennial, rather small-leaved and few- flowered and have from three to five carpels. 8. Delphinium emarginatum Presl, Delic. Prag. 6. 1822. D. emarginatum var. africanum Maire, Fl. de | Afrique du Nord 11: 66. 1964. No type cited. Maire says he cannot distinguish a noticeable difference from the Sicilian plant. Near to D. pentagynum Lam., the petioles dilated at the base; stem to 8 dm. tall, pubescent throughout or glabrous above; stem-hairs retrorse- appressed, not more than 0.6 mm. long; sepals 10 mm. long, bluish-lilac, glabrous without, not longer than the spur; petals bilobed into attenuate obtusish lobes; filaments glabrous; follicles 3, erect, glabrous. 42 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 TyPE: Sicily. DIsTRIBUTION. Occasional in Algeria and northern Tunisia; Sicily. ILLUSTRATION. Marre, FI. de l’Afrique du Nord 11: 65. 1964. REPRESENTATIVE SPECIMENS. Algeria: Mustapha near Algiers, Gandoger 614; Bone, en Barbarie, Decaisne dedit 1835. Tunisia: Hamman-el-Lif, Mt. Dj. Bou- Kournein, S. Murbeck 2. 9. Delphinium nevadense Kunze, Flora (Regensb.) 29: 647. 1846. D. emarginatum Presl var. nevadense (Kunze) Maire, Fl. de l’Afrique du Nord 11: 66. 1964. ear to D. emarginatum Presl, the stems glabrous above, distinctly angled, to 12 dm. tall; stem-hairs retrorse, but not appressed, up to 0. mm. long; sepals and spur glabrous; sepals 10-13 mm. long, bright blue, yellow-spotted above, distinctly shorter than the spur; petals simply emarginate; filaments more or less ciliate at summit of dilated part; follicles 3, glabrous. Type: “Sierra Nevada, ad fluvium Monachil, inter frutices, prope San Geronimo,” Spain, probably Boissier. Not seen. DiIsTRIBUTION. Southwestern Spain, northwestern Africa. REPRESENTATIVE SPECIMENS. Tunisia: Djebel Zoghouan, July 2, 1854, Kralik 22. Morocco: Montes des Kebdana, June 12, 1934, Sennen & Mauricio. 10. Delphinium cossonianum Batt. Bull. Soc. Hist. Nat. Afr, Nord 8:.215. 1917; Fic. 8 D. cossonianum var. laxiflorum Gattefossé & Maire in Maire, Bull. Soc. Hist. Nat. Afr. Nord 29: 404. 1938. Annual, more or less finely strigulose, or subglabrous below, 2—7 dm. tall, erect, simple or few branched, the stems angled; basal leaves gone by flowering time; cauline leaves short-petioled to subsessile, usually present at anthesis, closely placed, commonly 1.5—4 cm. long, tripartite into seg- ments 2-3 times divided into linear parts, glabrous, subcoriaceous; racemes densely and usually many flowered, up to 4 dm. long; bracts green, exceeding pedicels, 2-3.5 cm. long, strigulose; pedicels strigulose, ridged, 4-10 mm. long, the 2 bracteoles linear, 7-15 mm. long; flowers deep lively blue; upper sepal ca. 1 cm. long, somewhat strigulose, the spur horizontal or ascending, slightly curved, 10-13 mm. long; lateral sepals oblong, 8-10 mm. long, slightly strigulose along the middle; lower sepals similar, but more generally strigulose; upper petals white at base, violet at apex, retuse or emarginate, ca. 8 mm. wide at summit; lower petals ca. 9 mm. long, the rounded violet limb more or less cordate at base and narrowed abruptly into the claw; stamens shorter than the petals, the filaments whitish, glabrous; carpels 5 (3, or 4), usually glabrous, becoming almost 1 cm. long; seeds dark, ca. 1 mm. in diameter, with rings of scales. 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 43 Type not designated. Battandier had three specimens collected by Grant near Fez and Meknés in central Morocco. These I have not seen, except for a specimen from Fez, Abdul Grant, com. Aug., 1887 (x). DistriBuTIOn. Argillaceous soils of rather arid plains and hills, partic- ularly in central and eastern Morocco. ILLUSTRATION. Marre, FI. de l’Afrique du Nord 11: 74. 1964. REPRESENTATIVE SPECIMENS. Morocco: Pitard 2925, 2926, 2929; Jahandiez 94; Gattefossé, June 15, 1936; Maire, June 22, 1936; C. G. Garnett, May 7, 1936. Among the annual species of Delphinium in Africa the only one with five carpels is D. cossonianum. It is a Moroccan endemic noteworthy be- cause of its rather remarkable dense, many-flowered racemes of deep but lively blue flowers. 11. Delphinium peregrinum L. Sp. Pl. 1: 531. 1753. Fic. 9. Synonymy to be in a paper on Asian species. Annual, more or less gray-pubescent on stems with short stiff, often retrorse hairs, 2—7 dm. tall, usually virgately branched from below, the branches stiff, straight or sometimes spreading, ridged; lower leaves 3—5- parted, the parts cut into 3 or more linear-lanceolate lobes, blades 1-3 cm. long, often wider, longer than the petioles, glabrous, coriaceous; basal and middle leaves deciduous by anthesis; upper leaves entire, linear, much reduced, sessile; racemes terminal, often rather dense, many flowered, bracteate, the bracts subulate, 3-6 mm. long; pedicels stiff-pubescent, to almost 1 cm. long; bracteoles 2, to ca. 3 mm. long; flowers violet; upper sepal 7-8 mm. long, apeaiae prolonged into a straight ascending spur 1.5—2 cm. long and 3—5 mm. wide at base; lateral sepals lance-oblong, pubescent without, 5-7 mm. long; lower pair more pointed and with pubescence more median: upper petals bilobed, ca. 7 mm. long, yellowish at base, violet toward apex, the upper lobe linear, erect, the lower broad, rounded; lower petals 8-10 mm. long, the blade longer than wide, gradually narrowed into the slender claw; stamens 5-8 mm. long, with more or less lavender filaments; follicles 3, pubescent to glabrous, 7-8 mm. long, reread the style; seeds ca. 1 mm. long, rounded, scaly, the scales united into ro Type: “Habitat in Italia, Sicilia, Melita”; in Clifford Herbarium (BM). DistriBpuTIon. Eastern Mediterranean region from ne Sicily, Malta and Dalmatia to Egypt, Turkey, Asia Minor and Ira ILLUSTRATION. SipTHORP & SmiTH, FI. Graeca 6: ry 506. 1827. REPRESENTATIVE SPECIMENS. Libya: Attag, Dena Amon, Keith 731; Zuwetini Terraces, Tripolitania, at 2000’, B. C. Par Of the annual species of North Africa with three carpels, two are eastern Mediterranean and have the laminae of the lower petals cuneate at the 44 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 base and pass gradually into the claw. Of these, D. peregrinum is the better known and ranges from Libya and Sicily into Asia where it is quite polymorphous. I am grateful to Dr. R. K. Brummitt for sending me a mounted dissected flower from a Libyan collection at Kew to verify iden- tification. 12. Delphinium nanum DC. Syst. 1: 349. 1817. Fic. 10. D. peregrinum L. subsp. nanum (DC.) Graebner in Asch. & Graebn. Syn. 5. 2: 701. 1929. Annual, widely and stiffly branched from base, to subsimple, the branches simple or nearly so, more or less hirtellous or spreading-pubescent through- out, the hairs sometimes grayish, more or less interwoven, plants mostly 3-15 cm. high; lower leaves obovate in outline, broadly 3-lobed, long- cuneate at base, 4-6 mm. long, the lobes broad, undivided; main cauline leaves largely simple, narrowly oblanceolate, 1—-2.5 cm. long, narrowed at base into a broad petiole, gradually reduced up the stem, the upper bractlike, sessile, 2-6 mm. long; racemes short, loose, few-flowered; pedicels 1-5 mm. long, bibracteolate, the bracteoles lance-linear, 3—4 mm. long; flowers bluish-purple to brownish-violet, with yellow base; upper sepal 6—7 mm. long, densely pubescent as is the horizontal or ascend- ing spur which is 3-4 mm. wide at base, 15-20 mm. long; lateral sepals subovate, 6-7 mm. long, with median strip of pubescence; lower sepals oblong-lanceolate, 6-7 mm. long, tomentellous-pubescent; petals pale, the upper broadly bilobed, glabrous, 7-8 mm. long, 8-10 mm. wide; lower to ca. 1 cm. long, the blade roundish-oblong, ca. 4 mm. wide, narrowed to a claw 4 mm. long; stamens 5—8 mm. long, the filaments apparently bluish; carpels 3, sparsely pubescent, to ca. 8 mm. long in fruit and 2 mm. wide, the beak 2-2.5 mm. long; seeds ca. 1 mm. long, almost round, dark brown, with very closely appressed overlapping scales. Type: near Alexandria, Egypt, Delile; probable isotype (sm!). DIstrIBUTION. Lower Egypt from Bengasi to Adjedabia, especially in Maramarica. a ILLUSTRATION. TACKHOoLM, Students’ FI. Egypt, pl. 57. 1956. REPRESENTATIVE SPECIMENS. Egypt: Delile, 1803; Parquet 49; Letourneux, Gaillardet 121; Ehrenberg, 1820; Muschler, Feb. 1903; Mrs. Palmer, 1932. Delphinium nanum is very close to D. peregrinum, but is a much more pubescent dwarfish species of the very arid deserts of Egypt. It has rounder shorter laminae in the lower petals and fewer-flowered racemes. 13. Delphinium obcordatum DC. Syst. 1: 350. 1817. Fic. 11. D. halteratum Sibth. & Sm. var. obcordatum (DC.) Huth, Bot. Jahrb. 20: 477. 1895. D. macropetalum DC. Syst. 1: 350. 1817. 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 45 D. halteratum oe & Sm. var. macropetalum (DC.) Ball, Spicil. FI. arocc. 310. 1878. Dp; ike L. var. macropetalum (DC.) Maire, Fl. de l'Afrique du Nord 11: 64. Annual or biennial, 1-3(—6) dm. tall, with slender stems and many long, divaricate branches, usually strigulose: racemes 2~10(—15)-flowered, usually lax; pedicels and sepals strigulose; sepals deep blue-violet, 8-10 mm. long, the spur ca. twice as long; upper petals with a lamina 1 cm. long, winged from near the base, the spur ca. 15 mm. long; lower petals with the lamina exserted, roundish, more or less obcordate, not or little shorter than the claw, not cordate at base; follicles 7-11 mm. long, 14—-% as wide, slender, densely strigulose, often divaricate. Type: “in Barbaria’”’, North Africa, P. Durand. Type not seen. DISTRIBUTION. Sandy and fallow or rocky places below 1100 m., Mo- rocco, Algeria; Spain. ILLUSTRATION. DELEsSERT, Ic. 1: pl. 57. 1820. REPRESENTATIVE SPECIMENS. Morocco: Murdochee, 1875; Balansa, May, 1867; Hooker, 1871; Gattefossé, June 18, 1936; Font Quer 201. Alpaca: Morsiy, 1847; Reverchon, May, 1896; Faure, June 17, 1934; Gandoger, 1906-7 In the group of annual species with three carpels there seem to be four species along the Mediterranean coast of north Africa very closely related to Delphinium halteratum and by Maire (FI. de Afrique du Nord 11: 76-81. 1964) even included under D. peregrinum, but the much more abrupt contraction of the lower petals into the claw seems to me to set them quite apart and I am following Pawlowski (Fragm. Fl. & Geobot. 9: 438-444. 1963) in his concept of this complex. He proposes a new series Halterata as contrasted to the series Peregrina. As shown in the key at the beginning of the present paper, the distinction between the constituent species is based largely on the shape of the laminae of the lower petals and is quite technical. One certainly feels that D. halteratum is more than subspecifically distinct from D. peregrinum as in the Maire treatment. Pawlowski has carefully worked out synonymy and character- ized the species and I am happy to follow his treatment for D. obcordatum, D. gracile, D. verdunense and D. halteratum. 14. Delphinium gracile DC. Syst. 1: 347. 1817. Fic. 12. D. cardiopetalum DC. var. gracile (DC.) Willk. in Willk. & Lange, Prodr. Fl. Hisp. 3: 970. 1880. D. halteratum Sibth. & Sm. var. gracile (DC.) Huth, Bot. Jahrb. 4: 477. 1895. Stems 2-9 dm. tall, with long slender virgate branches, strigulose to subglabrous; upper leaves mostly entire; racemes usually lax, 2~15(—30)- flowered ; sepals often pale lilac-violet to whitish, 7-9 mm. long, strigulose- puberulent: spur mostly 2-214 times as long as sepals; limb of upper 46 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 petals ca. 12 mm. long, the spur ca. 17 mm. long; lower petals with a limb ca. 5 mm. long, exserted, cordate or subcordate at base, somewhat oblong, 114-2 times as long as wide, the claw ca. 7 mm. long; follicles 6-11 mm. long, 214-4 times as long as wide, glabrous or sparsely strigulose. Type: Andalusia, Spain, Lagasca. Not seen. DisTRIBUTION. Occasional in fallow and waste places below 1200 m., Morocco and Algeria; Spain and Portugal. REPRESENTATIVE SPECIMENS. Morocco: brahim, June 23, 1888; Font Quer 200. Algeria: Reverchon 249; Faure, June 7, 1906; Garrigues, 1887; Le Cesve, 1923, 1933; C. G. Abbot, 1911. 15. Delphinium verdunense Balbis, Cat. Stirp. Hort. Bot. Taur. App. 3231. 1813. D. cardiopetalum DC. Syst. 1: 347. 1817 D. halteratum Sibth. & Sm. var. cardiopetalum (DC.) Huth, Bot. Jahrb. 20: 477. 1895. D. tribracteolatum DC. Syst. 1: 350. 1817 D. halteratum Sibth. & Sm. var. tribracteolatum (DC.) Huth, Bot. Jahrb. 20: 477 Much like D. halteratum; leaves usually all divided into linear-oblong segments; racemes short and dense, mostly 5- to 15-flowered; sepals deep blue-violet, 8-12 mm. long, strigulose; spur 14-19 mm. long; lower petals with the limb up to %4 as long as the claw, included, strongly cordate at the base; follicles 6-10 mm. long, 2 4-414 times as long as wide, almost glabrous. Type: Verdun, southern France. DIsTRIBUTION. Apparently occasional in fallow fields, clearings, etc., Morocco and Algeria; southwestern Europe. ILLUSTRATION. PawLowskI, Fragm. Fl. & Geobot. 9: 442. 1963. REPRESENTATIVE SPECIMENS. Morocco: Gattefossé, 1933. Algeria: Garrigues, 1887. I am following Pawlowski (Fragm. Fl. & Geobot. 9: 441. 1963) in taking up the name D. verdunense for what has generally been called D. cardiopetalum, His description and excellent illustration have greatly aided me in interpreting the few North African collections I have seen. 16. Delphinium halteratum Sibth. & Sm. Prodr. Fl. istueca. 12 371. 1806 and FI. Graeca 6: pl. 507. 1826. Fic. 13. D. peregrinum L. ssp. halteratum (Sibth. & Sm.) Batt. Fl. Alg. 16. 1888. D. peregrinum sensu many auth., not L. D. peregrinum L. var. confertum Boiss. Voy. Espagne 2: 12. 1839 —45. D. oa Sibth. & Sm. var. confertum Fenzl ex Huth, Bot. Jahrb. 20: 4 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 47 D. cardiopetalum DC. var. oranense Debeaux, Rev. Bot.. 8: 264. 1890. Annual, the stems glabrous or mostly retrorsely strigulose, erect, simple or more usually branched, at least above, 1-8 dm. tall; lower leaves 3—5- parted, the parts cuneate, with lance-linear mucronate divisions; blades commonly 1.5—4 cm. wide and long, glabrous to pubescent, on petioles to the same length; upper cauline leaves shorter petioled to subsessile, com- monly 3-parted into lance-linear lobes or simple, acuminate, the upper and middle leaves persisting until anthesis; racemes few to many flowered; bracts subulate, 3-10 mm. long; pedicels strigulose to pubescent, to ca. 1 cm. long, with 2 bracteoles; flowers blue-violet; upper sepal 7-9 mm. high, mostly strigulose without, the spur ascending, strigulose, ca. 3—4 mm. wide at base, 15 mm. long; lateral sepals lance-oblong, ca. 7 mm. long, with a median strigulose band; lower sepals more pointed and more gen- erally strigulose; upper petals ca. 1 cm. long, greenish yellow at base, violet in upper half, bilobed, the upper lobe narrow, erect, the lower rounded; lower petals scarcely exserted, the lamina orbicular, truncate or rounded and weakly cordate at base, ca. 34.5 mm. in diameter, rather abruptly narrowed into a longer claw; stamens 5—7 mm. long, with some- what lavender filaments; carpels 3, usually strigose, rarely glabrous, 7-10 mm. long in fruit; seeds subglobose, the scales coalesced in ca. 3 horizon- tal rows. Type: “In Sicilia, et, ni fallor, in monte Athé, legit Sibthorp” (oxr). DistripuTion. Sandy and disturbed places, in the western Mediterra- nean of southern Europe and northern Africa, especially about Oran, Algeria. ILLustraTIoNns. SrstHoRP & SMITH, FI. Graeca 6: ph. 507. 1826: PAwLowskI, Fragm. Fl. & Geobot. 9: 440, 441. 1963. REPRESENTATIVE SPECIMENS. Morocco: Font Quer 200, 201, 202; Ibrahim, June 19, 1898; Mauricio 7759, 9215 ; Pitard 2927, 2928; Jahandiez 632. Algeria: Balansa 650, 651; Le Cesve 4363; Munby 22; Reverchon 249; Faure, 1934; Duffour 2317. 2. Consolida (DC.) S. F. Gray Consolida, as a genus, S. F. Gray, Nat. Arr. Brit. Pl. 2: 711. 1821. As a section, DC. Syst. 1: 341. 1817. As a subgenus, Huth, Bot. Jahrb. 20: 337. 1895; Dalla Torre & Harms, Genera Siphonogamarum 165. 1901, Plants annual. Leaves laciniate. Inflorescence racemose or paniculate, the flowers zygomorphic. Sepals 5, petaloid, the posterior one spurred. “Petal” one, apparently representing the fused superior petals of Del- phinium, sometimes called “nectary” or “honey-leaf”; subentire to 3—5- lobed, bearing a single nectariferous spur extending into the sepal spur. Stamens in 5 spirally arranged series. Carpel 1, forming a follicle. 48 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Type species, C. regalis S. F. Gray (Delphinium Consolida L.). A highly specialized genus ranging from the western Mediterranean to central Asia, but with several species widely cultivated and escaping from gardens in many parts of the world. Key To AFRICAN SPECIES OF CONSOLIDA a. pee 6-15 mm. long, more or less ea: plants mostly branched om the base and with widespread open ha C Spur 10-22 mm. long, exceeding the blade of the upper sepal in length. c. Sepals 7-8 mm. long, lilac; middle lobe of “petal” scarcely evident, much exceeded by the forward-projecting lateral lobes. Morocco, ys: ee Me er MC IER ay TRE: SEES RHO ea C. mauritanica. Sepals mostly 10-16 mm. long; middle lobe of ‘ ‘petal” evident, not exceeded by the lateral lobes d. Middle lobe of ‘‘petal” ‘not divided at apex. Algeria, Tunisia. OQ ai RR Ceara hah ts hens AONE ie ip yen iM i ubescens. d. sate lobe of “petal” with two small terminal lobes at least m. long. Occasional garden escape. .......... regalis. b. Spur 5- a mm. long, ca. equal to blade of upper sepal. ee Pe. ne ia RAG hal Ate Auta any PRU ee oaks Sa eee Ree, eet 4. C. deserti. a. Follicle 15-25 mm. long, — main stem forming an axis into the inflorescence; side branches fe e. Bracteoles usually below ni middle = the pedicel; gee 12-14 mm. long; spur ca. 15 mm. long. Garden e ~ Pera eres Rp) ambigua. Bracteoles near the summit of the ie sepals 10-12 mm. ae spur ca. 10 mm. long. Morocco, Algeria. ............... 6. C. orientalis. ® 1. Consolida mauritanica (Coss.) Munz, comb. nov. Fic. 14. Delphinium mauritanicum Cosson, Bull. Soc. Bot. Fr. 27: 68. 1880. D. Loscosii Costa var. irécirosvitam Pau, Bull. Soc. Iber. eines Nat. 22; 91 s.:1923; Annual, more or less densely cinereous-pubescent with spreading and retrorsely appressed hairs (many in the upper part stiff and with bulbous base), usually 1-stemmed, 2—5 dm. tall, branched from base and/or above, the branches stiff, more or less divaricate; lower leaves not or scarcely rosulate, mostly withered at anthesis, with petioles 2—3 cm. long, largely tripartite, then again divided into numerous linear lobes, the blades com- monly 1—2.5 cm. in diameter; cauline leaves rather crowded on main stem, gradually reduced upward and shorter-petioled, the uppermost sessile, a few mm. long, remote; upper branches with few remote flowers; bracts mostly with few linear lobes and 2—5 mm. long; pedicels 3—10(—20) mm. long, the bractlets near the middle, linear, entire, 2~4 mm. long; flowers lilac, 12-15 mm. in diameter; sepals lance-oblong, obtuse, pubescent without, 7-8 mm. long, each with a greenish median pubescent band; spur slightly curved, horizontal or ascending, ca. 15 mm. long; the “petal” whitish with colored border, almost as long as the sepals, 12-14 mm. wide, with 2 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 49 rounded lobes at the outer corners and 2 short median teeth; stamens somewhat shorter than the “petal”, ciliolate on the dilated lower part; anthers ca. 0.8 mm. long; follicle oblong, more or less compressed laterally, 10-12 mm. long, strigose, muricate-tubercled near ventral suture; seeds ca. 2 mm. long, obpyramidal, subtetragonous, blackish, covered with horizontal rows of overlapping scales TYPE: apalliasiiiaies Oran, Algeria, May 24, 1856, Bourgeau (P!; isotype, i paeenle Clearing in woodland, rocky places, etc. below 1500 m., Morocco, Algeria ILLUSTRATIONS. Cosson, Ill. Fl. Atlant. 1: pl. 3. 1882; Marre, Fl. de l'Afrique du Nord 11: 60. 1964 REPRESENTATIVE SPECIMENS. Morocco: Benoist, he 2, 1918; Font Quer 202 bis; Sennen & Mauricio 7754; Jahandiez 333, 802b. Algeria: Faure, June 30, 1933, July 11, 1923; Lefranc 501; Warion 2, 670; Battandier, May, 1922. 2. oe pubescens (DC.) Sod, Osterr. Bot. Zeitschr. 71: 241. 1922. Fie. 15. eerie pubescens DC, Fl. France 6: 641. 1815. D. pubescens DC. var. dissitiflorum Cosson, ee, 2: 46. 1887. Retrorse-strigulose, with some upper hairs bulbous at the base, mostly 1-stemmed, 1—5 dm. tall, more or less branched, especially above, rather ote Lok ttl See eee . a —. 26 Pm Fics. 14-19. “Petal” in species of Consolida (some shown with spur, some without). Fic. 14. C. mauritanica, Blade 13 mm. wide, from Faure, June 30, 3 1933 (cH). Fic C. pubescens. Blade 18 mm wee from Toulouse, June 19, 1841 (us). Fic. 16. C. regalis — 15 mm from a ae rite Fic. 17. ts deserti. Blade 8 m e, from Mies 1932 50 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 leafy in lower half; lower leaves in a basal rosette, petioled, gradually reduced upward and becoming sessile, all multifid into more or less mucronate linear lobes, the blades pubescent, commonly 2-3 cm. in di- ameter; flowers few to several at end of main stem and branches, lilac, pale blue or rose or whitish, 1-2 cm. in diameter; bracts mostly simple, sometimes palmatifid, commonly 1-2 cm. long; pedicels 1-3 cm. long; bracteoles entire, linear, 3-5 mm. long, opposite or alternate; sepals ovate to ovate-oblong, obtuse, strigose, short-unguiculate, commonly 10- 15 mm. long, each with a greenish spot toward the tip; spur straight or nearly so, horizontal, 18-20 mm. long, more or less strigose; “petal” 15- 18 mm. wide, glabrous, with 2 large rounded lateral lobes and a median emarginate to slightly divided lobe 3-4 mm. long; stamens 4—6 mm. long, the filaments dilated 34 their length and ciliolate; anthers yellow, 1 mm. long; follicle strigose, 10-15 mm. long, the style glabrous; seeds obpy- ramidal, dark, 1.5—2 mm. long, covered with transverse rows of scales. Type: “in agris et segetibus regionis Gallo-Mediterraneae.” DisTRIBUTION. Stony steppes and open places, high plateaus and Saharan Atlas; Algeria, Tunisia; southern Europe. ILLusTRATIONS. CosTE, Fl. France 1: 50. 1901; Marre, FI. de l’Afrique du Nord 11: 57. 1964. REPRESENTATIVE SPECIMENS. Algeria: Chevallier 535b; Kralik 23; Lefranc 501. Tunisia: Letourneux, April 29, 1884; Pitard 651; Murbeck, May 25, 1896. Libya: K. M. Guichard, March 4, 1955. 3. Consolida regalis S. F. Gray, Nat. Arr. Brit. Pl. 2: 711. 1821. Fic. 16. Delphinium consolida L. Sp. Pl. 1: 530. 1753. Plant 2-12 dm. tall, more or less divaricately, freely and rather finely branched, subglabrous to retrorse-strigulose; radical leaves petioled, early deciduous, the main cauline leaves trifid, then again partite into ultimate linear mucronate, strigulose lobes 1-5 cm. long; flowers few, in rather open terminal racemes, deep blue, or more seldom, pink to white; bracts mostly simple, linear, 3-15 mm. long; pedicels very slender, 1-5 cm. long; bracteoles linear, alternate to opposite, often well below the flower, 1.5-4 mm. long; sepals 9-16 mm. long, more or less ovate, narrowed at both ends, scarcely- or short-unguiculate, strigose, mostly along the midrib in the lateral pair; spur 17-22 mm. long; “petal” blue to cream, 12-15 mm. wide, the spur 16-17 mm. long, middle lobe of limb 3.5—5 mm. high, shortly 2-lobed at apex, each of these lobules often emarginate; lateral lobes rounded, 4-5 mm. long; stamens 4-7 mm. long, the filaments dilated on lower portion, anthers 0.6—0.7 mm. long, yellow; follicle to 2 cm. long, including the 2-3 mm. style; seeds dark, with transverse rows of subcontiguous scales. Type: “Habitat in agris Europae restibilibus.” 1967 | MUNZ, DELPHINIUM AND CONSOLIDA 51 DistripuTION. Europe, Caucasia, Turkey. Occasional escape from gardens in northern Africa. ILLUSTRATION. REICHENBACH, Ic. Fl. Germ. 4: pl. 66. 1840; Hect, FI. Mittel.-Eur. 3: 487. pl. 114. 1909; Marte, Fl. de l’Afrique du Nord 11: 62. 1964 4. Consolida deserti (Boiss.) Munz, comb. nov. Fic. 17. Delphinium deserti Boiss. F 1). Orient. 1: 83. 1867. Plant 5—30 cm. tall, diffusely branched from below, short viscid-hairy; leaves 5-25 mm. long, short-hairy, palmately parted into oblong-linear segments, these again laciniate in the main leaves, the tips sharp; bracts 5-8 mm. long, oblong-linear; pedicels 1.5-3 cm. long; bracteoles near middle of pedicel, 3-5 mm. long; flowers small, yellow; sepals hairy, oblong, 4-6 mm. long; spur straight, 5-6 mm. long; “petal” obcordate, ca. 6 mm. wide, 4-lobed, the outer lobes longer, semi-ovate, the middle ones smaller, a row of brownish red spots across the base of the lobes; spur horizontal, equalling the blade of the “petal”, ca. 5-6 mm. long; filaments sparsely hairy; anthers yellow, rounded, 0.5 mm. long; follicle oblong, viscid-hairy, 6-10 mm. long, 2-2.5 mm. wide, the beak slender, ca. 2 mm. long; seeds oblong, with densely imbricate scales about the cupuliform center. ae DistriBuTIoN. Egypt and Asia Minor, in desert conditions. REPRESENTATIVE SPECIMEN, Egypt: “Luksor”, R. Muschler, March 1904. Isotyprs: Bové 131 from desert between Suez and Gaza, July 1832 28 5. Consolida ambigua (L.) Ball & Heywood, Repert. Sp. Nov. 66: 151. 1962. Fic. 18. Delphinium ambiguum L. Sp. Pl. ed. 2. 749. 1763. D. ajacis sensu auth., not L. Sp. Pl. 1: 531. 1753 : : Consolida ajacis sensu auth., not (L.) Schur, Verh. Siebenburg. Ver. Naturf. 4: 47. 1853. Subglabrous, especially below, to more or less pubescent, especially above, with short fine appressed or spreading hairs, mostly 1-stemmed, 3-10 dm. tall and with ascending-horizontal branches above; lower leaves long petioled, the blades 2-5 cm. in diameter, multifid into ultimate mucronate linear segments largely 1-1.5 mm. wide; cauline leaves grad- ually reduced upward, the upper sessile, often 1-2 cm. long; bracts entire to multifid, the lower often at least as long as the pedicels, the uppermost 2-3 mm. long; inflorescence rather lax, each branch ending in a few- to many-flowered raceme; pedicels commonly 1-3 cm. long; bractlets mostly 1.54 mm. long, usually below the middle of the pedicel; flowers blue to rose to white, 2—2.5 cm. across; sepals ovate, attenuate at base, obtuse $2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 at apex, more or less pubescent, each with a greenish subapical spot, ca 12-14 mm. long; spur ca. 15 mm. long, straight or slightly curved; “‘petal’ 3-lobed, the central lobe oblong, ca. 6 mm. long, with 2 short rounded lobules, the 2 lateral lobes rounded, ca. as long, the spur 15-16 mm. long; stamens 6—7 mm. long, the filaments abruptly dilated at base, glabrous, an- thers yellow; follicle pubescent, cylindric, 16-22 mm. long, the style ca. 2 mm. long; seeds black, ca. 2.5 mm. long, oblong-trigonous, with undulat- ing transverse rows of scales. ? Type: Mauritania, northwest Africa, the type (LINN) a double garden form. DisTRIBUTION. Mediterranean region, but African material seems to be garden escapes. ILLUSTRATIONS. REICHENBACH, Ic. Fl. Germ. 4: pl. 67. 1840; Marre, Fl. de Afrique du Nord 11: 54. 1964. 6. Consolida orientalis (Gay) Schréd. Abhandl. Zool.-Bot. Ges. Wien 4(5): 25. 1909. Fic. 19. Delphinium orientale Gay in Desm. Cat. Dordogne 12. 1840, nomen; Boiss. Fl. Orient. 1: 79. 1867. D. hispanicum Willk. Prodr. Fl. Hisp. 3: 969. 1880. Type not cited. D. orientale Gay ssp. hispanicum (Willk.) Batt. in Batt. & Trab. Fl. Alg. 15. 1888. D. orientale var. hispanicum (Willk.) Huth, Bot. Jahrb. 20: 376. 1895. D. ajacis L. var. orientale (Gay) Finet & Gagnep. Bull. Soc. Bot. Fr. 51: 467. 1904. Mostly 1-stemmed, simple or branched, especially above, erect, 2-6 dm. tall, the lower parts retrorse-strigose, upper with more or less crisped pubescence, some hairs gland-tipped, some swollen at base; leaves mostly subglabrous, the lower petioled and early withered, the principal leaves multifid, crowded on stems, the blades 2-4 cm. long, the segments linear or lance-linear, 1-2.5 mm. wide, more or less ciliolate; uppermost leaves sessile; flowers few to many, in racemes that are rather dense in anthesis, more open in fruit; bracts green, the lower with linear parts, the upper entire, linear, 5-20 mm. long; pedicels 0.5—4 cm. long, bibracteolate near the summit; flowers 1.5—2.5 cm. in diameter, dark purple or violet to rose or white, sepals round-ovate to oval or somewhat narrower, 10-12 mm. long, abruptly contracted into a pubescent basal claw, almost as long as the blade which is rounded to obtuse at the summit, more or less pubescent; spur ca. 1 cm. long; “petal” usually of same color as sepals, 3-lobed, 8-10 mm. long, the median lobe ovate to oblong-ovate, 2-lobed, somewhat longer than the rounded lateral lobes; stamens ca. 6-7 mm. long, the filaments dilated at the base, with some gland-tipped hairs at upper end; anthers yellow, ca. 1.5 mm. long; follicle strigose, 14-22 mm. long, oblong-cylindric, with some hairs gland-tipped, the style 1.5-2 mm. long; seeds brownis black, obpyramidal, angular, 1.5-2 mm. long, with several undulating transverse rows of short scales. 1967] MUNZ, DELPHINIUM AND CONSOLIDA Type: described from cultivated material from the Orient, isotype (K!). DistrisuTIon. Grain fields and disturbed places, Morocco and Algeria; Spain, Hungary to Asia ILLUSTRATIONS. Bot. Mag. 169: (n.s.) pl. 186. 1952; Davis, Fl. Turkey 1; 129, fig. 2. 1965. REPRESENTATIVE SPECIMENS. Algeria: Dukerley 3805; Balansa 1009; Maire 4887; Reboud 203; Choulette 203 bis; E. G. Paris 402. A number of color forms has been proposed by Maire, Fl. de l’Afrique du Nord 11: 55-56. 1964 SYNONYMY OF AFRICAN SPECIES OF CONSOLIDA AND DELPHINIUM ConsoLipaA — genus #2 aaqaaaan ajacis = C. ambigua ambigua = species #5 deserti = species #4 mauritanica = species #1 orientalis = species #6 pubescens = species #2 regalis = swat #3 DELPHINIUM — genu ee oe Ye Poo fs Se ye & bo ajacis = Consolida ambigua ajacis var. orientale = Consolida orientalis ambiguum = Consolida ambigua balansae = species #2 balansae formae caeruleum cardiopetalum sie na cardiopetalum va gracile = D. mene cardiopetalum var. oranense = D. halteratum consolida = Consolida regalis cossonianum = species #10 cossonianum forma laxi- florum = D. cossonianum dasycaulon = species #3 dasycaulon var. minor = D. reeuieae erti - emarginatum = species #8 emarginatum var. africanum = D. emarginatum D. Sb eb 6 vv yyy Soy b&b boy oo . obcordatum = emarginatum var. nevadense = D. nevadense leroyi gracile = specie halteratum = species #16 halteratum var. cardio- petalum = D. verdunense pag leciae ke ka svelte: = D. halte as Sane var. ale = D. gracile halteratum var. macropeta- lum = D. obcordatum pega var. steal halter var. ygniion hs re um = D. verdunense ae Consolida orientalis hybridum ssp. fissum var. pentagynum = D. penta- num leroyi = specie loscosii var. rachel = Consolida mauritanica macrocentron = species #5 macrocephalum = D. leroyi D. tu mauritanicum = Consolida mauritanica nanum = species #12 nevadense = species species 4 13 53 54 JOURNAL OF THE ARNOLD ARBORETUM D. orientale = Consolida orientalis D. orientale ssp. and var. hispanicum = Consolida orientalis . pentagynum = species #7 pentagynum vars. glabratum, heterotrichum, homeotri- chum, phialotrichum = D. pentagynum peregrinum = species #11 sp merouna - otis ae = D. haltera gaps pi haltera = D. halte date” ; ae var. macropeta- lum = D. obcordatum uy Oo yb YD [voL. 48 D. peregrinum ssp. nanum = D. nanum . pubescens = Consolida pubescens D. pubescens var. dissiti- florum = Consolida pubes- S ruspolianum = D. leroyi staphisagria = species #1 sylvaticum = species #6 tribracteolatum = D. verdunense verdunense = species #15 by SeUd wellbyi = D. leroyi taphisagria macrosperma = D. staphisagria ~H INDEX TO EXSICCATAE Abbot, C. G., 1911, D. gracile Allard 2739, D. pentagynum Alleizetti, 1953, D. staphisagria Archer 8657, D. dasycaulon; 8619, D. leroyi Balansa 575, D. pentagynum; 650, 651, D. halteratum; 1009, C. orientalis: May, 1867, D. ‘obcordatum Bally 7035, 10033, D. dasycaulon Battandier, May, 1922, C. mauritanica Benoist, July 2, 1918, C. mauritanica Bogdan 547, D. macrocentron Bornmiiller 2013, D. staphisagria Bourgeau 428, 1231, D. staphisagria Brass 16990, 17150, D. dasycaulon; 17172, 17269, D. leroyi Burtt 6114, 6214, D. dasycaulon Cesve, Le, 4363, D. halteratum; 1923, 1933, D. gracile; 1938, D. staphi- ria Chandler 996, D. reek Chapman 321, D. dasycaulon Chevallier 255, D. hae: 5a50, ©. pubescens Chipp 70, D. leroyi Choulette 203 bis, C. orientalis Cosson, 1854, D. peter July 26, 1954, July 22, 1961, D. balansae Curle 140, 380, D. dasycaulon Curtis, Anita, 674, 679, 693, D. leroyi; 0, D. macrocentron Delile, 1803, D. nanum pi 1049, D. dasycaulon Duffour 2317, D. halteratum Dukerley 3805, C. orientalis uncan 48117, D. leroyi Durando 402, D. pentagynum; D staphisagria Ehrenberg, o D. nan Eliot, C. W., 1545, D. macrocenteon Fairchild 59, . pentagynu Faure, July 11, 1923, a 30, 1933, 1855, 1906, D. gracile: June 17, 1934, D. obcordatum. Font Quer 107, 209, D. staphisagria; 108, 153, D. pentagynum; 201, D. obcordatum: 200, 202, D. haltera- um; 202 bis, eC; anaes Goitlet te 121, D. nan Gandoger 614, D. PEE EH ea 1906- 7, D. obcordatum Garnett, C. G., May 7, 1936, D. cos- sonianu Garrigues, 1887, D. gracile; 1887, D. verdunense Gattefossé 220b, D. balansae; 1933, . verdunense; June 18, 1936, D. obcordatum 1967] Gillett 14559, 14936, D. leroyi Greenway 7854, D. macrocentron; 9511, D. leroyi Greenway & Doughty, 8519, 8521, D. macrocentron; 8532, D. leroyi Guichard, March 4, 1955, C. pubescens Hardy & Bally 10777, D. macrocen- ron Hepper 2847, D. dasycaulon Hiller & Lithgoe 516, D. leroyi Hooker, 1871, D. obcordatum Ibrahim, June 22, 1888, D. gracile; June 19, 1898, D. halteratum Jackson 405, D. leroyi; 1860, D. dasy- caulon Jahandiez 94, June 15, 1936, D. cosso- nianum; 333, 802b, C. mauritanica; 632, D. halteratum; 742, 858, D balansae Jamin 4, D. pentagynum Johnston, H. B., 1531, D. leroyi Kabylie, July 18, 1909, D. sylvaticum Kassner 2748, D. dasycaulon Keith 731, D. peregrinum Kralik 22, D. nevadense; 23, C. pubes- cens; June 2, 1858, D. balansae Lefranc 501, C. mauritanica Letourneux, April 29, 1884, C. pubes- s cen Lightbody 26257, D. dasycaulon Lugard 26, D. macrocentron Maire 4887, C. orientalis; June 22, 1936, D. cossonianum Maitland 1263, D. leroyi Marsilly, 1847, D. obcordatum Massey 62, D. dasycaulon Mauricio 7759, 9215, D. halteratum Mendonca & Wild 994, D. dasycaulon Meyer, F. G. 7465, D. dasycaulon, 7339, D. leroyi M geod 344, D. dasycaulon Milne-Redhead 444, D. dasycaulon Mooney 5820, 6264, 7528, D. dasy- caulon; 6705, 7012, D. leroyi Munby 22, D. halteratum; 35, D. RANcHOo SANTA ANA BoTANIC GARDEN 1500 N, CoLtecE AVENUE CLAREMONT, CALIFORNIA 91711 MUNZ, DELPHINIUM AND CONSOLIDA 55 pentagynum; 1844, 1850, D. sta- phisagria urbeck 2, D. emarginatum; May 25, 1896, C. pubescens Murdochee, 1875, D. obcordatum Muschler, March 1904, C. deserti; Feb., 1903, D. nanum Myers, J. G. 11688, D. leroyi Nash 274, D. dasycaulon Mrs. Palmer, 1932, D. nanum Pappi 390, 4811, D. dasycaulon Paris, E. G. 402, C. orientalis Park, B. C. 362, D. peregrinum Parquet 49, D. nanum Pitard 651, C. pubescens; 2925, 2926, 2929, D. cossonianum; 2927, 2928, D. halteratum Pole Evans & Erens 1435, D. macro- entron Polunin, O. 2138, D. balansae Reboud 203, C. orientalis Reverchon 249, D. gracile; May, 1896, D. obcordatum Robbie 6, D. dasycaulon Robyns 2155, D. dasycaulon Rogers 412, D. leroyi Schimper 397, 752, D. dasycaulon Schlieben 838, 5084, D. leroyi Schweinfurth & Riva 1472, D. dasy- caulon Scott 30, D. leroyi; 231, D. dasy- caulon Sennen & Mauricio 7754, D. mauri- tanica; June 12, 1934, D. nevadense Snowden 486, 817, D. macrocentron; 501, D. leroyi Tamajong 22210, D. dasycaulon Taylor, G. 1277, 1309, 1367, D. mac- rocentron Tothill 2608, D. leroyi Unwin 214, D. dasycaulon Warion 2, C. mauritanica White, F. 2741, D. leroyi; 2742, D. dasycaulon Witte, de, 373, D. dasycaulon 56 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 FLORAL BIOLOGY AND SYSTEMATICS OF EUCNIDE (LOASACEAE) Henry J. THOMPSON AND WALLACE R. ERNST THE GENUS Eucnide presents a pattern of species differences somewhat unusual among flowering plants. The eleven species are very similar in all respects except floral morphology, in which they differ widely. For example, the corollas in several species are apopetalous while in others they are sympetalous, forming a long tube. Characters such as apopetaly and sympetaly are so well established as criteria for separating major families of flowering plants that it is difficult to determine their taxonomic importance when they occur as differences between otherwise similar species. We shall evaluate the taxonomic significance of floral morphology in Eucnide by two approaches. The gross features of floral morphology will be analyzed by determining their anatomical basis so that more refined comparisons between the species can be made. The biological importance of floral structures will be evaluated by considering their role in repro- duction of the local populations. The assessment of the taxonomic im- portance of these characters will be independent of the correlations that they ultimately give in our classification. Other Loasaceae either have only five stamens, as in subfamily Grono- vioideae, or numerous stamens arranged in discrete fascicles opposite the 1967 | THOMPSON & ERNST, EUCNIDE 57 petals and accompanied by elaborate petaloid staminodia, as in subfamily Loasoideae. Within subfamily Mentzelioideae the genus Mentzelia is dis- tinguished by its three parietal placentae (five only in M. decapetala) with the ovules inserted in two distinct vertical rows on each placenta and by the seeds, which are variously shaped but always large and not longitudinally grooved. The remaining species of this subfamily, which we consider to comprise the genus Eucnide, are characterized by five placentae, each with many rows of minute, longitudinally grooved seeds (Fic. 1). We recognize 11 species in Eucnide and arrange them in three sections as follows: bartonioides grandiflora obata sect. EUCNIDE series Eucnide cordata hypomalaca irta xylinea series Tubiflorae sect, MENTZELIOPSIS urens aurea rupestris enella sect. SYMPETALEIA Sas = ~ MATERIALS AND METHODS Much of our information about Eucnide comes from material of 17 populations, representing nine species, that we have been able to study in detail. Taste 1 lists these collections and indicates how each was studied. Some have been observed as natural populations and some have been grown in the greenhouse at Los Angeles. All of these collections have been available in great quantity so that detailed morphological and an- atomical studies could be made. Material was fixed in formalin-acetic- alcohol and then either cleared in lactic acid or embedded in paraffin, sectioned, and stained. Additional material from herbarium specimens of other species was restored by gentle boiling and cleared or sectioned, So that all species have been studied from both sectioned and cleared material. Meiosis has been examined with the usual method of fixing buds in 1:3 acetic-ethanol, squashing anthers in acetocarmine, and observ- ing with a phase microscope. Meiosis was observed in at least three different individuals from most of the populations studied. Controlled Pollinations were made in all populations grown in the greenhouse. The Stamens were first removed and then the style and stigma were covered with a segment of plastic soda straw to prevent accidental pollinations. When the stigma became receptive the soda straw was removed momen- tarily while the desired pollen was applied to the stigma. In addition to our own collections we have studied the specimens in the following herbaria: A, BM, BR, CAS, CU, DS, E, F, GH, K, LA, M, MO, NY, POM, RSA, SMU, SRSC, TEX, UC, US. 58 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 TABLE 1. Collections of Eucnide of particular importance in this investigation. CoLLector & LocaLity HaPLoip Curomo- GARDEN FI£LD some No. CuLTuRES STUDY Sect. Eucnide E. bartonioides Thompson & Ernst 3283; Big Bend National Park, Texas E. cordata Chambers 1004; Catavina, Baja Calif., México Wiggins Puertecitos, Baja Calif., Méx E. grandiflora Ernst & Sais 2475; NW of Nacaltepec, Oaxaca, México. Garden voucher, Thompson 3386 E. hirta McVaugh sad Near Lake Chapala, Jalisco, Méxic E, lobata King 4505; Ciudad Victoria, Nuevo Leon, Méxic co Waterfall oe Monterrey, Nuevo Leon, Méxic Sect. Mentzeliopsis E. urens esi ne 3184; Panamint Valley, Inyo Co., Calif. abe 3182; Death Valley Buttes, Inyo Co., Calif. Thompson 3339, Raven 11784; Hoover Dam, Clark Co., Nevada Sect. Sympetaleia E. aurea Carter & Sharsmith 4160; Sierra Giganta, Baja Calif., México Carter & Sharsmith 4260; Isla Danzante, Baja Calif., México Carter & Ferris 387 oa W of Loreto, Baja Calif., Méxic Wiggins & Ernst 543; Santo dicen Sierra Giganta, Baja Calif, Méxi E. rupestris Raven 14802; N of San Felipe, Baja Calif., México Thompson 3316; S of San Felipe, Baja Calif., México E. tenella Wiggins, Carter & Ernst 280; Mision Los Dolores, Baja Calif., éxico 21 + + 21 + " 21 - - 21 + + 21 - “ 21 + ao 21 + ee 21 - + 21 _ 3 21 + + 21 + = 21 + 5 21 ~ — 21 ~ 5 i 21 _ - 21 —- + — = + 1967] THOMPSON & ERNST, EUCNIDE 59 Fic. 1, Characteristic fruit, placentation, and seed in Eucnide. a, capsule, E. tenella, showing position of ovules, valves, and major veins, X 10; b, ovary, . urens, transverse section showing ovules and cordate placentae, 4; c, see apex, E. aurea, X 70; d, seed, E. aurea, longitudinal view, X 75. COMPARATIVE MORPHOLOGY Morphological variation throughout Eucnide is somewhat unusual in that particular structures show either striking diversity or great uniformity. We have chosen to emphasize this aspect of the pattern of variation by discussing comparative morphology under two headings, considering first the morphological diversity, which occurs primarily in the flowers, and following this with the morphological similarities that occur throughout the remainder of the plant body. Morphological diversity. Our three taxonomic sections are founded on morphological innovations in the conformation of the corolla and the spatial relationships of the anthers and stigma. The corollas are apopetal- Ous or sympetalous anatomically, but this difference is bridged and made ambiguous by a peculiar condition of false sympetaly. Within either the apopetalous or sympetalous categories there are species with open, spread- ing corollas and species with narrow, tubular corollas. Among the species with spreading corollas and those with tubular corollas there are examples with included or exserted anthers and stigma. In our first two taxonomic sections, false sympetaly (i.e., apopetaly) is correlated with anthers that become two-locular at anthesis. In our third section, true sympetaly is Correlated with anthers that become unilocular. In section Eucnipe, the corollas are white to yellow and are falsely Sympetalous at the base. The staminal filaments are more or less filiform above but fused basally into a relatively very short tube (0.5—3 mm. long) to which the petals always are united. It is their union with this tube that holds the petals together when both stamens and petals abscise from the top of the ovary. We consider these corollas anatomically apopetalous but functionally slightly sympetalous by virtue of their union with the Staminal tube. Two modifications are evident in this flower type. In either case, the anthers and the stigma are conspicuously exserted from the corolla and the anthers are two-locular at anthesis. On the one hand, 60 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 48 series Eucnide, with three species, has more or less open flowers in which the petals and the stamens are loosely spreading (Fic. 2a, b). On the other hand, series Tubiflorae, with four species, has relatively long, cylindrical flowers in which the petals remain erect and broadly imbricate, even though the falsely sympetalous portion is only 2-3 mm. long (Fic. 2c). In the latter series, the position of the petals compresses the stamens into a tight, narrow, brush-like cluster. In section MENTZELIOPSIS, with only one species, Eucnide urens (Fic. 2d), the corolla is pale yellow to nearly white, and the broad, rigid lobes are narrowly imbricate and spread to form an open funnel. We consider the corolla apopetalous, as in section Eucnipe. Although the staminal tube to which the petals are united is 5-8 mm. long, it is relatively short, being about a fifth the length of the corolla. The anthers and stigma are distinctly included and only a little more than half as long as the corolla. Above the staminal tube, the stout free filaments are slanted inward form- ing a kind of chamber around the style. The anthers are two-locular at anthesis. A few of the filaments, the outermost and longest, curve outward at the top, holding their anthers away from the stigma. The proximity of most of the anthers to the style, however, and their inwardly nodding position causes most of the pollen to be shed into the interstices of the filaments and into the chamber the filaments create around the style. The deposition of the pollen into this chamber is peculiar to this species of Eucnide but it recalls a very similar situation found in the genus Mentzelia section BIcUSPIDARIA. In section SYMPETALEIA, with three species, the corollas clearly are sympetalous anatomically, the stamens are epipetalous, and the anthers (to be discussed shortly) are unique. In Eucnide tenella (Fic. 2f), the whitish corolla (ca. 8 mm. long) with spreading narrow lobes, is united for about a third of its length into a narrow tube. The anthers are ex- posed on long filaments that equal or slightly exceed the stigma and corolla. In E. aurea (Fic. 2g), the yellow or brilliant reddish orange corolla (ca. 15-20 mm. long) is distinctly salverform, with broad, some- times imbricate lobes, and is united for about two-thirds its length into a narrow tube. The nearly sessile anthers are included within the tube, and below them is a ring of upwardly curving hairs. In E. rupestris (Fic. 2e), the greenish corolla (to ca. 15 mm. long) with shadings of yellow or brown, is more or less cylindrical, with rigidly erect, valvate lobes, and is united into a tube for about three-fourths of its length. The anthers are included within the tube, the lower ones subsessile and the upper ones on short, stout, erect filaments. In the lower portion of the corolla beneath the stamens, as in E. aurea, there is a ring of upwardly curving hairs. The ring of hairs is absent in E. tenella. In this section the anthers are unilocular at anthesis. The interesting differences in the structure of the anthers require ex- planation. The anthers vary in general external conformation, but more important, the original masses of sporogenous tissue, as well as the final pollen chambers (locules), differ both in shape and in number (Fic. 1967 | THOMPSON & ERNST, EUCNIDE 61 3a-q). All of the anthers are basifixed, and the microsporangia are lined with a well-defined tapetal-endothecial region. The anthers of section EucnivE and section MENtTzELIopsis are divided medially by the con- nective, which sometimes projects slightly beyond the microsporangia; otherwise, the anthers are slightly emarginate apically (Fic. 3a, b, f). Transverse and longitudinal sections of these anthers (with the plane of section indicated in Fic. 3a, b, and the sections shown in Fic. 3c, d, e) clearly demonstrate that the sporogenous tissue of each occurs in four separate, straight, vertical masses. The two masses of sporogenous tissue on the same side of an anther are marked externally by a longitudinal groove (Fic. 3b, c) and are separated internally from each other only by the tapetal-endothecial layer. The paired masses of sporogenous tissue on one side of the anther are separated from those on the other side by the tissues of the connective. As the pollen matures, the tapetal-endothecial tissue breaks down joining the adjacent lateral pollen masses in a single locule on either side of the connective so that just before dehiscence, the anther becomes two-locular. The dehiscence is longitudinal at the lateral grooves. In the ovate to elliptical anthers of section EucNnipg, the wall of the locule frequently becomes reflexed so that after dehiscence, the ex- posed portion of the anthers is the lining of the locule (Fic. 3f). In the linear-sagittate anthers of section MENTzELIoPsis the wall of the locule usually does not reflex and the lining is less conspicuously displayed. In section SympETALetA, the anthers are reniform (Fic. 3g) and the connective never reaches the apex of the anther. Transverse and longi- tudinal sections (Fic. 3i-1; m—p) show the unusual nature of these micro- sporangia. In Eucnide tenella and E. aurea the sporogenous tissue occurs in two horseshoe shaped masses, for, although Fic. 3i shows four masses of Sporogenous tissue, these are continuous over the apical portion of the anther (Fic. 3k, 1). The nature of the sporogenous tissue in E. rupestris is particularly illuminating. In the portion of the anther facing the style (i.e., adaxially) the sporogenous tissue of the two sides of the anther is continuous apically (Fic. 30) just as in EZ. tenella and E. aurea. In the portion of the anther toward the corolla (i.e., abaxially), however, the Sporogenous tissue is not continuous, but occurs in two lateral masses separated apically by a thin septum (Fic. 3p). hus, in Eucnide tenella and E. aurea there are two masses of sporo- genous tissue, while in E. rupestris there are three. When the pollen ma- tures, the tapetal-endothecial tissues break down joining the pollen masses laterally so that in all three species the anthers have a single locule just Prior to dehiscence. The anthers dehisce at the lateral groove (Fic. 3h), which, in section SYMPETALEIA, is continuous over the apex of the anther. After dehiscence, the anthers are peltate with the anther wall slightly Cupped or opened out flat (Fic. 3q). Morphological unity. The diversity in morphology and anatomy among the species of Eucnide, which permit the recognition of the species sroups, is restricted to the flowers, i.e., corolla, stamens, and style. Under- lying these striking differences, however, are many similarities in basic 62 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 § MENTZELIOPSIS § SYMPETALEIA Ae . 2. Flower vidas! nide reduced t cat eh ame length, showing shape of corolla and po sitio oF eign cat cue ae ber ides, X 1; b, E. 1967 | THOMPSON & ERNST, EUCNIDE 63 floral and vegetative architecture which are unique to the genus and point out its morphological unity. The plants are herbaceous, short lived perennials, often with the lower stems stout and appearing woody. Plants of Eucnide aurea, E. bartoni- oides, E. cordata, E. hirta, E. lobata, and E. urens all flowered about three months after the seeds were sown in a warm greenhouse in Los Angeles and seemed able to persist as perennials. Some species, such as E. aurea, E. bartonioides, and E. rupestris, also may be successful facultative an- nuals. The leaves are alternate, simple, and of more or less characteristic appearance throughout the genus. They are petiolate, with the blades mostly cordate, orbicular to ovate with crenate, lobed or incised margins. The venation varies from nearly palmate in some species to quite pinnate in others; however, the general outline and appearance of the leaves remain very similar. In some species the upper surface of the leaves ap- pears to be varnished. The plants usually are densely pubescent with several kinds of trichomes that are characteristic of the family and often stinging to the touch. The trichomes are unicellular and either smooth or variously reflexly barbed; some are swollen at the base or rest upon multicellular epidermal plat- e ovary is inferior and unilocular, with each of the five parietal placentae alternate with the sepals. In transverse section the placentae are cordate in outline, intrude deeply into the locule of the ovary, and are densely covered with minute, horizontally radiating ovules (Fic. 1b). The dry, dehiscent fruits are globular to obovoid and are surmounted by the persistent sepals (Fic. 1a). In contrast to the usually fragile walls of the fruit, the apex is hard and durable and divides into five valves which are attached opposite and interior to the sepals (i.e., alternate with the placentae). The valves open wide, permitting the sifting out of the seeds. Frequently, the style remains attached to the tips of the valves and is torn as they open. he numerous seeds (Fic. 1c, d) are pale or straw colored, 0.5-1.25 mm. long, and usually acute at the attached end and rounded at the free end. They are conspicuously ribbed or grooved longitudinally with the Parallel crests either straight or spirally twisted, both forms often oc- curring in the same capsule. In some instances, the seeds are slightly con- Stricted near the middle, At the time of their dispersal, the seeds are filled by the embryo. The pollen is oblate (shorter on the polar axis), tricolporate (each of the three narrow furrows with a median, round pore), and without prom- nent surface sculpturing. The pedicels (or peduncles) in all species curve as the capsules mature lobata, sepals removed, X 3; c, E. cordata, X 1%; d, E. urens, forward sta- mens, one petal, and two sepals removed, X 1; e, E. rupestris, corolla shown as though transparent, forward two sepals removed, X 2; f, E. tenella, X 512; 8, E. aurea, corolla shown as though transparent, X 24. 64 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 48 and, in some species, elongate from a few millimeters at anthesis to as much as 33 cm. by the time the seeds are shed. In the plants grown in the greenhouse, the capsules always pointed north; from this we assume that the pedicels are negatively phototropic. Elongating pedicels are developed best towards the base of the plants, and, since these portions often are dry and broken in larger plants, they usually are not preserved on herbarium specimens and may not be noted by collectors. he chromosome number has been determined for eight species repre- senting all three sections of the genus. In each cytological examination, 21 pairs of chromosomes were observed in the meiosis of the pollen mother cells, and there were no cytological anomalies. The voucher specimens and collection data for these determinations are given in TABLE 1. Another expression of the basic similarity of the flowers is found in PETALEIA. The vascularization of the most diverse flower types is funda- mentally similar. The divergence of the stamen traces (c) from the petal trace (b) occurs at relatively the same level below the articulation of the corolla in all of the flowers regardless of the degree of petal-fila- ment union. The petal trace (b) always is more complicated than the 1967 | THOMPSON & ERNST, EUCNIDE 65 sepal trace (a), and the vascular supplies to the placenta, valves, and style (e) all arise in the same way. The differences between the sympetalous and apopetalous flowers con- cern the degree of relatively simple lateral fusion of nonvascular tissue without alteration of the tracheary tissue. It is noteworthy that in fresh material of Eucnide aurea, one of the most sympetalous examples in the genus with long salverform corollas with essentially sessile epipetalous anthers, the corolla tube readily can be stripped away leaving behind the anthers supported by the staminal tube. The corolla also can be torn from the staminal tube along a kind of cleavage plane in the apopetalous species with long filiform stamens. The differences, therefore, between the falsely sympetalous corollas (those actually apopetalous) and the truly Sympetalous corollas seem to be relatively minor anatomically. In respect to habit, leaves, hairs, fruits, pedicels, placentae, seeds, pollen, chromosome number, and pattern of floral vascularization, the species of Eucnide are remarkably similar. Quite the opposite is shown, however, with the corollas and stamens where there is great diversity in form, the differences being correlated to delimit three species groups, each with its own distinctive flower type. BREEDING SYSTEM The self-pollinations made in all of the greenhouse cultures listed in TABLE 1, representing six of the eleven species of Ewcnide, resulted in full sets of viable seeds, while unpollinated flowers formed neither seeds hor mature capsules. In the absence of any indication of apomixis we take this to indicate that the plants tested were self-compatible. In these plants, therefore, the pollination system determines the inbreeding-out- breeding balance by controlling the amount of self- and cross-pollination. The main factors of the pollination system are the natural pollinators and floral morphology, in particular the position and developmental timing of the anthers and stigma. In Eucnide the morphology of the flowers is not only a convenient marker of the reproductively isolated units but is an important element in the operation of the populations, serving to adjust the populations to specific pollinators and to regulate the amount of in- breeding and outbreeding. Section Eucnide. In all of the species of section EUCNIDE the flowers have yellow to pale yellow or nearly white petals and exserted stamens, and they open in the late afternoon. The species differ, however, in flower size, the degree of spreading of the petals, and in the timing of anther de- hiscence so that there are three different floral forms. We have studied flowering in greenhouse cultures of Eucnide bartonioides, E. cordata, sa srandiflora, E. hirta, and E. lobata, representing all of the three types of flowers (Fic. 2). In E. bartonioides the flowers are large, with the yellow petals usually 3 cm. long and spreading away from the exserted Stamens. The flowers first open in the late afternoon, at which time the already receptive stigma exceeds the anthers by a few millimeters 66 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ' bartonioides er 3) nd SJE) SYMPETALEIA OE ) Fic. 3. Morphology of anthers in Bride be! sect. EUCNIDE: a, adaxial view of anther, X 6; b, same, lateral view transverse nt and reece” tudinal sections of « a”-and “py? at Gedicnh dadiaiia d by numerals 1-3; f, lateral view of tice sire dehiscence. g-—q, sect. Sy : ad xial view of anther; h, lateral view; i-l, trans erse and longitudinal sections of “g” and “h” at positions marked by numerals 4-7; m—p ie view after dehiscence. Anthers of sect. EucNipE and sect. MENTZELIOPsIsS (not shown e four microsporangia and become two-locular at anthesis een pee IA are unilocular at anthesis; FE. aurea and E. tenella have tw microsporangia, but E. rupestris has three. a-f, X 6; g-q, X 19 (Fic. 2a), so that in completely undisturbed flowers there is no pollina- tion and no seeds are formed. The anthers are sufficiently close to the stigma, however, so that even a slight jarring of a flower at anthesis re- sults in some self-pollination. These observations on plants growing in the greenhouse indicate that in natural populations outcrossing would be encouraged by large insect visitors that would first touch the extended stigma as they approached the flower. The same pollinator would cause selfing by bending the stigma back into the anthers, as well as by visiting different flowers of the same plant. Thus, in natural populations the in- breeding-outbreeding balance must lie well toward the inbreeding ex- treme. Although we have not observed pollination of E. bartonioides in natural populations, we can make a reasonable prediction about its polli- nator, since a model is available in the flowers of Mentzelia laevicaulis (sect. BartontA) which are very similar in general conformation, size, color, and opening time. We know that M. Jaevicaulis is pollinated by hawkmoths (Sphingidae), and the similarity of the flowers suggest that E. bartonioides also is hawkmoth pollinated. In Eucnide lobata the flowers are small, with the deep yellow, rotate petals only 1 cm. long (Fic. 2b). The fully developed flowers have the stamens arranged in two distinct tiers: the shorter, outer stamens and the longer, inner stamens. The flower first opens in the late afternoon, at which time the outer stamens are fully expanded and release their pollen. The inner stamens, however, remain convolute around the base 1967 ] THOMPSON & ERNST, EUCNIDE 67 of the style. The stigma is receptive at this time, but, because the outer stamens are shorter than the style and held away from it, no pollination can occur unless an insect visits the flower. On the following morning, the inner, longer stamens straighten out and release their pollen more or less directly onto the stigma, so that even completely undisturbed flowers in the greenhouse set a full complement of seeds. Thus the manner in which the flower matures presents an initial opportunity for insect visitors to effect some outcrossing, while automatic self-pollination follows the following morning. Populations of E. lobata probably have a some- what greater range of outbreeding than the populations of its near relative, E. bartonioides. If pollinators are abundant, seed production and out- breeding would be very high; however, in the absence or shortage of pollinators seed production would remain high but the population would shift to maximal inbreeding. The breeding system is highly labile, re- sponsive to conditions of the environment, and integrated by the natural pollinators. While the natural pollinators of E. Jobata are unknown, the flowers of this size, form, and color in Mentzelia and other groups usually are bee pollinated. A third line of flower development occurs in Eucnide cordata. The petals are very pale yellow, nearly white, about 2 cm. long, and tightly imbricated about the filaments of the exserted stamens (Fic. 2c). The style elongates first, while the petals are still tightly enclosing the anthers. The filaments then elongate, forcing the anthers out of the corolla, but they remain well below the stigma. The petals never spread, and flowers with stamens in various stages of emergence from the corolla can be seen throughout the day, making it difficult to say when the flowers open; the pollen, however, is released in the late afternoon. The stamens still are held together by the imbricated petals, and the pollen is shed well below the stigma. Some of the flowers in an inflorescence, particularly the first to mature, are held more vertically and in these self-pollination does not occur in undisturbed flowers. Other flowers in an inflorescence, particu- the stigma, and some of these flowers, even though undisturbed, form some viable seed. Self-pollination in undisturbed inflorescences is not very effective, since only about 25 per cent of the possible seed production is realized. Although no information about natural pollinators is available for E. cordata, the structure of the flower, along with the well exserted stigma, indicates dependence on a pollinator. Flower and inflorescence Structure in E. cordata function to regulate the amount of inbreeding and outbreeding, the opportunity for inbreeding — more restricted in this species than in either E. bartonioides or E. lobat The flowers in Eucnide hirta have the same pene structure as those of E. cordata, with one important exception. In E. hirta the style elongates somewhat less in proportion to the rest of the flower, and the stigma is Surrounded by the anthers at the time pollen is shed. Thus, self-pollina- tion always occurs, and every flower in undisturbed greenhouse plants 68 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 produces a capsule full of seeds. The natural pollinators of E. hirta are unknown but even with an active pollinator the populations must be primarily inbreeding. The species of section EucNme thus show striking differences in gross floral morphology. Without any modification of the basic floral structure, variation in relative lengths of stamens, styles, and corolla and position of the petals are harmoniously recombined to give three flower types the function of which is explicable in terms of the pollination system. All are probably insect pollinated. Within each flower type there are ac- commodations for flexibility in the amount of self- and cross-pollination, with this feature being most highly developed in Eucnide lobata. Section Mentzeliopsis. In Eucnide urens, the only species of this section, the flowers are large, with the pale cream-colored petals about 5 cm. long. The stigma exceeds the anthers but does not extend beyond the petals. The stamens press inward around the style, the sturdy filaments forming a pollen chamber well below the stigma (Fic. 2d). Undisturbed flowers do not set seed, and natural populations must be highly dependent on insect visitors. We have studied pollination in two natural populations of this species, one on either side of the Panamint Mountains, Inyo Co., California. The population near Death Valley Buttes (Thompson 3182) was studied on 16 April, 1961, from 8:00 a.m. until 2:00 p.m. The pop- ulation in Panamint Valley (Thompson 3184 and 3187) was studied on 6 May, 1961, from before sunrise until sunset. In addition, several other populations have been spot checked, and these observations confirm those made in more detail at the above localities. The flowers of E. wrens open in bright sunshine at about 9:00 a.m., and pollen is released immediately into the pollen chamber. The flowers soon are visited by a melittid bee, Hesperapis laticeps Crawford, that remains active until early afternoon, when the desert winds usually increase in force. These pollinators alight on the exposed stigma and style and force their way head first into the pollen chamber. On backing out from the pollen chamber, they again touch the stigma. Although other insects occasionally are found in the corolla of E. urens, only Hesperapis laticeps enters the pollen chamber and affects pollination. The behavior of Hesperapis laticeps in the E. urens flowers results in both cross- and self-pollination. This species is the only Eucnide with this type of flower and pollina- tion system, but, in spite of its unique gross morphology, the basic architecture of the flower is the same as all of the other species of the genus. Flowers with the same gross morphology and the same pollinators occur in Mentzelia sect. BicuspipariA, but in Mentzelia this flower type is developed from an entirely different basic flower structure. For example, the seeds, placentae, and floral vasculature in Mentzelia are entirely differ- ent from those in Eucnide. Section Sympetaleia. The three species of this section show the greatest diversity in the features of gross floral morphology and corolla color that are associated with the pollination system. In all species, how- ever, the flowers have tubular corollas and open in the morning. In 1967 | THOMPSON & ERNST, EUCNIDE 69 Sr a ccigit Fic. 4. Diagram of the major vascular traces and ven of the flower, in Egencia cordata, transverse view from above. a, sepa a’, median sepal vein; b, petal trace; b, Aides vein; c, staminal traces; : ak al seal vein; e, traces to valves and sty ———— Eucnide aurea the corolla is salverform and the anthers are sessile on the corolla tube (Fic. 2g), but the two populations represented in our cul- tures differ in flower color and size. The plants from Danzante Island (Carter & Sharsmith 4260) have strong reddish-orange corollas about 20 mm. long. When the anthers release their pollen, the stigma is at or just above the ring of hairs in the lower portion of the corolla tube below the Stamens. About two days later, the style elongates until the now-recep- tive stigma is slightly exserted from the sympetalous corolla. In the plants from the Sierra Giganta (Carter & Sharsmith 4160) the flowers are smaller, with yellow corollas about 15 mm. long. Although the style elongates as it does in the Danzante Island plants, the elongation is relatively less, the stigma never being exserted at all and usually not exceed- ing the middle of the mass of anthers. Both the red and the yellow forms of E. aurea are self-compatible, and undisturbed flowers set seed in the 70 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 greenhouse. It should be pointed out that if pollinator activity is high the pollen of an individual flower might be depleted before the style elongates and the stigma becomes receptive. This would tend to increase outbreeding in environments where pollinators were abundant. Although we have no observations on the natural pollinators of E. aurea, the color and form of the flowers suggest that Lepidoptera and humming birds could be effective. In Eucnide rupestris, the flowers are very inconspicuous to the human eye. The corolla is green, somewhat darker than the leaves, and usually a little more than one centimeter long. While others have recorded a metallic luster on the corolla lobes we could not detect such an appearance in the population that we studied on an overcast day. The anthers are more or less sessile in the corolla tube, the stigma is located among the anthers, and the style does not elongate as it does in E. aurea; pollen thus falls directly on the receptive stigma (Fic. 2e). No pollinators were seen during a mid-day, three-hour observation period in one population, but it was noted that every capsule on every plant was forming a full comple- ment of seeds. This is evidence that the plants are self-compatible as well as self-pollinating. The information that we have about E. rupestris suggests that it is the most autogamous species of Eucnide, a feature cor- related with its great range and morphological uniformity. In Eucnide tenella, the corolla is white or cream colored and less than one centimeter long, the anthers are exserted, and the stigma is located among the anthers (Fic. 2f). We have not grown plants of this species and have made only very brief observations on one natural population. From the structure of the flowers we can determine that self-pollination occurs readily, and the prevalence of self-compatibility throughout the genus would indicate that EF. tenella also is self-compatible. We know nothing of the natural pollinators, but the information on flower structure suggests a rather high degree of inbreeding for E. tenella. Enough information is available about the breeding system of the species of Eucnide to allow us to outline the general pattern for the genus. The system is based on self-compatibility, with the inbreeding-outbreeding balance determined by the amounts of self- and cross-pollination permitted by the pollination system. The features of floral morphology involved in the pollination system are then important parameters of the entire breeding system. Each of the sections of Eucnide has distinctive features of floral morphology that are developed without alteration of the basic structure of the flower. The flowers in section Eucnipe have yellow, cream or nearly white petals, exserted anthers, and open in the late after- noon. We speculate that this section is largely insect pollinated. In sec- tion MENTZELIopsis the flowers open in the morning, and the stamens form a pollen chamber that limits effective pollination to bees with a very special behavior. In section SympeTALetA the corollas are tubular, but otherwise the flowers are diverse, with corollas varying from inconspicuous green to highly conspicuous vermilion. We speculate that the latter is butterfly or hummingbird pollinated. Throughout the entire genus, and 1967 | THOMPSON & ERNST, EUCNIDE 71 especially in section SYMPETALEIA, the basic floral structure has permitted morphological adjustments to a wide range of pollinators. It is important to note that species pairs, that is, species that are more similar to each other than to any other species, may differ in their breeding system. Thus £. hirta is a self-pollinating version of E. cordata. Eucnide rupestris, the most autogamous species, is most like the outcrossing E. aurea. Even within E. aurea there are two flower forms that we have not seen fit to recognize in formal taxonomy but which differ in their floral biology, one being more self-pollinated than the other. It is also important to emphasize that species as dissimilar as Eucnide urens and the species of Mentzelia (sect. BicuspIpARIA) can have nearly identical pollinating systems, although the basic morphology of the two groups is entirely different. HYBRIDIZATION Numerous pollinations were made between plants of different sections, but all of these failed to produce viable seeds and no hybrids were ob- tained. The crosses attempted were: Eucnide aurea, of section SyMPE- TALEIA, with E. bartonioides, E. cordata, and E. lobata, of section EUCNIDE, and reciprocals; E. urens, of section MENTZzELIOpSIS, with E. bartonioides and £. lobata, of section Eucnipe. Interspecific pollination within sec- tion EucNnine usually resulted in capsules full of viable seeds; plants of two hybrid combinations were grown and their fertility determined. Eucnide lobata (3298-2) 2 E. bartonioides (3283-6) 6. The Fi plants grew vigorously and flowered profusely. The parents differ greatly in floral morphology (see Fic. 2b, a), and the hybrids were recognized by their intermediate flowers. Two of the hybrid individuals were tested for fertility and both produced less than two per cent good pollen. The large number of chromosomes made exact determination of chromosome pairing in most microsporocytes impossible, but 23 cells at metaphase Showed 14 to 16 bivalents with the remainder of the chromosomes in- volved in multivalents or present as univalents. Heteromorphic bivalents were noted and unequal segregations at anaphase I of 20 to 22 and 19 to 23 were common. Although several flowers were self-pollinated, no viable seeds were obtained. Eucnide cordata (3300-2) < E. lobata (3298-2) and reciprocal. The F, plants were vigorous and flowered profusely. The morphology of the flowers was intermediate between the very different flowers of the parents (see Fic. 2c, b). Five hybrid individuals were tested for fertility. All had less than five per cent good pollen. Meiosis in all five plants was irregular. Chromosome pairing was analyzed more or less completely in a total of 43 cells from all five of the F, plants. Most cells had about 15 bivalents, with the remainder of the chromosomes involved in multivalents Or present as univalents. Heteromorphic bivalents were noted, and ana- Phase I cells with 21-21 segregations were less frequent than those with V2 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 unequal segregations. Self-pollinations on all five plants failed to pro- duce seeds. There were no differences noted between the reciprocals. The data from these cross-pollinations indicate a strong incompati- bility between plants of different sections of Eucnide that operates at the level of crossability. Within section Eucnipe hybrids are readily obtained between the most morphologically diverse species, but the F; plants, although easily obtained, are sterile. DISTRIBUTION AND ECOLOGY The genus Eucnide is distributed from the southwestern United States, through Mexico to Guatemala. Section MENTZzELIOPsIS, with only one species, E. urens, occurs in the Death Valley region of California and in the lower Colorado River basin, with an outlying station in Baja California (Map 1). Eucnide urens is the only species of the Mojave Desert area, and its distribution at elevations usually below 4000 feet does not over- lap that of any of the other species of Eucnide. The three species of section SYMPETALEIA occur in Baja California, with E. rupestris extending into California, Arizona, Sonora, and Sinaloa (Map 2). All are restricted to the Sonoran Desert region and do not occur above 3000 feet elevation. The species of this section apparently are allopatric — at least there are no data available to indicate that they ever occur as adjacent populations. Section Eucning, with seven species, has the widest range of the three sections. Eucnide cordata and E. hypomalaca occur in the Sonoran Desert area of Baja California and Sonora (Map 4). The remaining species of this section inhabit southern Texas, eastern and southern Mexico, and Guatemala (Maps 3, 4). In the northern portions of this range, about latitude 29° north, the populations usually occur below 3000 feet. In southern Mexico, at latitude 18° north and in Guatemala, at latitude 15° north, the populations occur at elevations from 2000 feet to 8000 feet above sea level. Some interesting relationships between the breeding system and dis- tribution are found in Eucnide. The four species with long, narrow flowers, series Tubiflorae, replace each other geographically (Map 4). In series Eucnide, the two species with large flowers, E. bartonioides and E. grandiflora, also are allopatric, but the third species, E. lobata, with small, rotate flowers, is at least geographically sympatric with E. bartonioides in part of its range (Map 3). Species with different flower types, hence placed in different sections or series, often are sympatric. For example, E. cordata, of section Eucninr, is sympatric with each of the three species of section SyMPETALEIA, and, likewise, E. xylinea, of series Tubiflorae, is sympatric with both £. dobata and E. bartonioides of series Eucnide. Dis- tribution and floral morphology in Eucnide are related through the general- ization that species with similar flower types, and thus probably with similar pollination systems, are always allopatric; and conversely, sym- patric species always have different flower types and presumably different ‘= ~ Hoag A / cordata tenella rupestris aurea Fie. 5: Comparison of vasculature in two sepals and one petal in Eucnide, showing insertion of stamens and ovules for four species, adaxial view. a, sepal trace; a’, sepal vein; b, petal trace Sicee ovules) : b’, petal vein; c, stamen traces; d, lateral sepal vein; e, traces to valves and style; f, peripheral veins of ovary wa [L961 AGINOOAA “LSNYA ® NOSdWOHL 74 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 pollination systems. Eucnide rupestris with its small, greenish flowers is the strictest inbreeder of section SyMPETALEIA. It is also the most wide- spread and least variable of the three species in the section. Although the genus Eucnide ranges over a large segment of the southern portion of North America, all eleven species occupy the same general habitat. They all occur primarily on cliffs or rocky slopes, although scattered individuals or transient populations occasionally occur in washes and on lower, more stable slopes below the more permanent populations. Cliffs and rocky slopes are themselves a discontinuous phase of the topography of an area, and, as a result of their restriction to these habitats, the populations of Eucnide are relatively small and discontinuous. The negatively phototropic and elongating pedicels found in many of the species of Eucnide function to maintain the populations on steep cliffs. When the pedicels turn away from the light they assure that the opening of the capsule is directed toward the face of the cliff when the seeds are shed, increasing the chance for lodging of the seeds in a crack in the rocks. In the species with greatly elongating pedicels, such as E. bartonioides, the entire capsule may be forced deep into a crevice before the seeds are released. TAXONOMIC RATIONALE To any student of variation and evolution in plants the outstanding feature of Eucnide is the great diversity in the morphology of corolla and stamens which contrasts so sharply with the uniformity in all other characters. We submit that a systematic interpretation of such a group should reflect a prior interpretation of flower structure derived from a con- sideration of the floral biology of natural populations. Clearly, pollination is the process that gives meaning to floral morphology. In Eucnide, with its breeding system based on self-compatibility, the pollination system, with floral morphology as its important component, is not only an es- sential feature of reproduction but also regulates the amount of inbreeding. Even slight differences in floral morphology can make some individuals better inbreeders and others more effective outbreeders. Any situation that favors inbreeding will select for the flower forms that promote self-polli- nation. Thus the catastrophy of the loss of a pollinator might lead to the rapid evolution of the flower type of a highly autogamous form. This might account for the origin of such autogamous species as E. hirta and E. rupestris. Conversely, a population that is established as highly autog- amous with but a minimum of outcrossing by casual insect visitors might provide the stable base for the selection of an outbreeding form with a very specific adjustment of floral morphology and pollinator. We feel that the origin of very distinct floral types and pollination systems in Eucnide may be facilitated by autogamy and more readily visualized if an autog- amous stage is imposed between the two very specialized forms. For example, the direct origin of a bee-pollinated flower, such as that of E. 1967 | THOMPSON & ERNST, EUCNIDE 75 urens, from the hawk-moth flower of E. bartonioides by the occurrence of a “hopeful monster” in an E. bartonioides population is less probable than the transition through an autogamous stage. The striking similarity be- tween the flowers and pollination system of E. urens and Mentzelia in- volucrata must be the result of evolutionary convergence. It is difficult for us to imagine either a single step or a series of steps, each with a specific pollinator, that could accomplish this convergence without any resort to autogamy. The autogamous species, therefore, may be phylo- genetically very important in Eucnide and not merely evolutionary “dead ends.” We speculate further that the architecture of the Eucnide flower with its fusion of petals and numerous filaments provides for a plasticity that allows for modification to different flower types without alteration of the basic anatomical structure and without incurring any modification of seeds, leaves, or fruits. It is important for us to present our specula- tions on these matters because they are the compelling thoughts we hold as we subjectively consider variation in Eucnide and make the taxonomic decisions required for our classification Our comparative study of the wvatoniical details of floral morphology has pointed out the great overall similarities among the species of Eucnide which previously have been overshadowed by the gross differences. Fur- thermore, we give less taxonomic weight than previous authors to these striking differences in morphology of corolla and stamens because we feel that as components of the pollinating system they are evolutionarily labile. We conclude that there are three coordinate groups in Eucnide rather than two as proposed by the previous classification which grouped the taxa into two genera, Eucnide and Sympetaleia. Our conclusion that these three groups are best considered sections rather than genera rests on our evaluation of their degree of affinity viewed against our understanding of generic levels in the remainder of the family. The differences and sim- ilarities between the three sections of Eucnide that we propose are equiva- lent to those between the sections of M entzelia, the group of species most similar to Eucnide. TAXONOMY The revision of Eucnide by Waterfall (1959) has greatly simplified our task of presenting a formal taxonomy, for, although he excluded Sym- petaleia, his treatment of the species that he did place in Eucnide basically is the same as ours. Accordingly, our comments on the classification and nomenclature of these species are limited primarily to presentation of additional information or to minor points of differences in opinion. We present the distribution of each species with detailed dot maps and cite only critical specimens. More extensive citations of specimens are given in Waterfall’s revision, and we can provide on request a complete list of specimens. Our discussion of the species of section Sympetaleia are some- what more detailed because these have not been monographed previously. 76 JOURNAL OF THE ARNOLD ARBORETUM [VvoL, 48 Eucnide Zucc. Del. Sem. Hort. Monac. 4th unnumbered page. 1844.1 Microsperma Hook. Icon. Pl. pl. 234. 1839; not Microspermum Lag. Gen. et Sp. Nov. 25, 1816. Sympetaleia A. Gray, Proc. Amer. Acad. 12: 161. 1877. Loaselia Baill. Bull. Soc. Linn. Paris. 1: 650. 1887. Plants herbaceous, pubescent, potentially perennial (or annual), some- times + woody at base, the stems usually less than 1 m. long. Leaves alternate, exstipulate, petiolate, the blades pinnately to palmately veined, broadly ovate (rarely obovate) to suborbicular, often cordate at base Hairs smooth, needlelike, or variously reflexly barbed. Inflorescences -+ racemose, terminal or axillary, usually bracteate, of 1 to many flowers. Flowers pedicellate (or pedunculate), 5-merous, + regular, bisexual, epigynous. Sepals 5, free, persistent. Petals 5, yellow, + white, greenish or reddish orange, apopetalous, but united at base to staminal tube, or corolla sympetalous. Stamens few to many, maturing centripetally; fila- ments free above and + filiform but connate basally into a short tube (and united to base of free petals), or filaments epipetalous; anthers basifixed, exserted or included, oblong to ovoid and 2-locular at anthesis, with longi- tudinal dehiscence, or reniform peltate and 1-locular at anthesis, with apically confluent dehiscence; pollen oblate, 3-colporate. Gynoecium 5- carpellate; stigma exserted or included and equaling or exceeding the anthers, globular to linear and sometimes twisted, of 5 appressed (some- times spreading) lobes (or stigma a small terminal tuft of hairs); style usually filiform (sometimes thick), usually elongating and somewhat per- sistent; ovary inferior, + spherical, obovoid to clavate, 1-locular, with 5 intruded, transversely cordate parietal placentae; ovules numerous on each placenta, anatropous. Fruits globular to narrowly obovoid, often nodding or reflexed on conspicuous pedicels (or peduncles), dehiscent apically by 5 centrifugally opening triangular valves interior to and op- posite the persistent sepals. Seeds minute (usually less than 2 mm. leng), numerous, + oblong, grooved or ribbed longitudinally, sometimes spirally twisted, pointed at the micropylar end, + rounded at chalazal end. Seed- ling cotyledons subreniform to ovate, short-petiolate. Type sprectes: E. bartonioides Zucc. (Name from Greek, eu, pretty, and knide, nettle, “Schonnessel.”’) 1 The place of publication for this name is “Delectus Seminum in Horto R. Botanico vista meiner Collectorum Anno 1844,” a list of seeds offered by the Munich Botanical a rden name Eucnide bartonioides first appeared on the 2nd page with an asterisk which leads one to the 4th (last page where both the generic and the specific characters are se in Latin description closes with “Zuccarini in Act et Jardin Botaniques, Genéve. We are grateful to - B. Bonner and J. E. Dandy for locating this rare publication and for making a photocopy available to us. 1967 | THOMPSON & ERNST, EUCNIDE 77 The generic name Microsperma, first used in Loasaceae for M. lobata Hook. in 1839, is earlier than the generic name Eucnide Zuccarini, pub- lished at the end of 1844. In Loasaceae, Microsperma was used in original descriptions also for M. rudis Schauer (1847) and M. grandiflora Groen- land (1861). Since Microspermum had been used in Compositae by Lagasca in 1816 the name Microsperma in Loasaceae was rejected as a homonym in 1900 by Urban & Gilg in favor of Eucnide. If the view of Urban & Gilg is regarded as contrary to Art. 75 (a subjective decision) then it will be necessary to conserve Eucnide or, failing this, to adopt the name Microsperma for this genus. We have submitted a proposal for the conservation of the well known name Eucnide. The genus Eucnide, including 11 species, is divided into three sections. Two series are distinguished in section EucnineE on the basis of the form of the corolla. In series Eucnide, primarily of eastern Mexico, EZ. bar- tonioides and E. grandiflora are very similar, differing principally in size, color of the corolla, and in distribution. Eucnide bartonioides and E. lobata normally are strikingly dissimilar but a few specimens, ironically, almost seem to bridge the morphological gap, and the ranges of the two species overlap. Series Tubiflorae is widely distributed in Mexico. Three of the species, Eucnide cordata, E. hirta, and E. hypomalaca, are quite similar, differing mostly in corolla size, inflorescence, pubescence, elonga- tion of the fruiting pedicels, and position of the stigma in relation to the anthers. As in series Eucnide, these species are isolated geographically from one another and from the more distinctive E. xylinea, also of series Tubiflorae. The section MeEntTzeLiopsis of the far southwestern United States, comprised of a single species, E. urens, is very different morphologically from all other species of the genus Section SyMPETALEIA, almost confined to Baja California, is composed of three species that are strikingly dissimilar from one another in mor- phology and color of the corolla. KEY TO THE SPECIES 1. Corolla united for at least 1%4 its length into a narrow, relatively long tube (8-10 mm.) with a ring of hairs below the + sessile, included anthers 2. Corolla salverform, bright yellow or reddish orange; sepals half the length of the corolla tube or less; fruiting pedicels elongating greatly (to aT tm)? SOulpern eka CAUtOte, es ee te ce 9. ured. 2. Corolla + cylindrical, with erect lobes, dark green with shadings of yellow pedicels reflexed, elongating only a few centimeters; Baja California, local in southern California, southwestern Arizona, Sonora, and Sinaloa DOR ar eres Ci AUGN she te ed ed Richa tay ey yh RCN Sinead esa Uo Or 10. E. rupestris. Corolla united for % or less of its length into a relatively short, + incon- spicuous tube (1-8 mm.); ring of hairs lacking; all anthers on long filaments, exserted or included 3. Anthers reniform (sporogenous tissue continuous over the connective), — ww JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 exposed, 1-locular and peltate at anthesis, the pemeeene confluent api- cally; corolla white or pale yellowish, to 10 mm. long; stigma among the anthers; fruiting pedicels elongating at least several precise rare and loc al, eastern Baia California Sur... ..+........-...- 11. E. tenella. Anthers ovate to oblong (sporogenous tissue interrupted apically by the ar 2-locular at anthesis, the dehiscence not confluent; flowers mostly lar, Stigma a stamens both distinctly included, about % to *4 as long as the open, funnelform, rigid corolla; filaments linear, stout, com- pressed around the style; fruiting pedicels 1-2 cm., not elongating ; desert regions of California, Nevada, Utah, Arizona; ss in north- eC Sa Ry ew ees ae E. urens. 4. Stigma and stamens ouinetad or clearly exposed; rae Mexico, Guatemala = 5. Corolla open and stamens spreading (series Eucnide). Corolla rotate, to about 12 mm. long or 2 cm. wide; outer stamens shorter than stigma, inner stamens slightly exceeding the minute, subcapitate stigma; fruiting pedicels 1-2 cm., not elongating; northeastern Mexico to Puebla. .... 3. E. lobata. 6. Corolla open-funnelform and larger 7. Stigma + linear-truncate, to about 6 mm. long, less than 1 mm. broad, sometimes separating into 5 narrow lobes; sepals to about 2.2 cm. long; petals to about 4 cm. long; stamens to 5 cm. long; style to 5.5 cm. long; fruiting pedi- cels elongating up to 33 cm.; Texas, eastern Mexico south San Leis Pale Sis Os are se: 1. E. bartonioides. Stigma obconical, to 2.5 mm. broad, 2-3 mm. long, 5-lobed; sepals 2.5-5 cm. long; petals 5.8-8 cm. long; stamens 8-11 cm. long; style 10-12 cm. long; Oaxaca and Guatemala. RON EE ORS Fe et Ir, eae 2. E. grandiflora. 5. Corolla + cylindrical, the petals broadly imbricate and holding the stamens in a long, narrow cluster (series Tubiflorae). 8. Leaves small, to 15 mm. long, 10 mm. wide, bsnl eg mafia flowers solitary at ends of ceed branches; "corolla to 25 m long, fruiting pedicels to 2.5 cm. long; rare, western saad Coahuila and southwestern Pesaran — Nuevo Leon RIOUERE 55). OF res aa a ee ae E. xylinea. 8. crue larger, 3-12 cm. long, 2.5-10 cm. Lins pu beeen but villous-canescent; flowers usually clustered; corollas 2—4 igh ng. ~ cm 9. Inflorescence many flowered, usually on a distinct peduncle, hence above the leaves; petals 1.5-2.5 cm. long; pedicels not elongating in fruit; Raia California, aa of the Gulf of California, and near Guaymas, Sonora E. cordata. Inflorescence few flowered, not on a ais peduncle; petals 2-4 cm. long; pedicels sometimes elongating in fruit. 10. Pubescence on pedicels with simple hairs conspicuously longer than the reflexly barbed hairs; pedicels greatly elongate in fruit; south-central Sexicy and Guate- mala. 5. E. hirta. ye 1967 | THOMPSON & ERNST, EUCNIDE 79 10. Pubescence on pedicels with reflexly barbed hairs at least as long as the simple hairs; fruiting pedicels un- known; Sonora, southwestern Chihuahua, and northern SOURIS es convicies’s ba cea ne ee oe 6. E. hypomalaca. 1. Eucnide bartonioides Zucc. Del. Sem. Hort. Monac. 4th unnum- bered page. 1844.1 Mentzelia gronoviaefolia Fisch. & Mey. in Fisch., Mey. & Ave-Lallemant, Index Sem. Hort. Petrop. 10: 54. 1845 (not before 3 Feb.). Microsperma bartonioides Walp. Repert. Bot. 5: 777. 1845-46 (probably 1846). Inflorescences of a few flowers, terminal, or the flowers axillary. Corolla 1—4 cm. long, yellow, paler below, the petals spreading. Anthers exserted, slightly exceeded by the linear (to 6 mm. long) stigma. Pedicels to 3—4 cm. long at anthesis, usually elongating to 15 cm. or more in fruit. © OONGA Sect. MENTZELIOPSIS E. urens Map 1. Distribution of section MENTZELIOPSIS in the southwestern United States and Baja California, Mexico. 80 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 The size of the corolla (and the leaves) is quite variable. The small flowered forms can be confusingly similar to depauperate plants of Eucnide lobata even though these two species normally are quite distinct. (See discussion of E. watsonii under E. lobata.) DISTRIBUTION. United States: southwestern Texas; and Mexico: north- eastern states southward into San Luis Potosi (Map 3). Hototypr: The descriptions were based on plants grown in the Munich Botanical Garden from seeds probably collected by de Karwinski (perhaps number 285, gathered in Chihuahua, Mexico, in 1843; not seen by the present authors). Regarding the latter, Urban & Gilg (1900, p. 108) remarked: “Colitur in hortis Europaeis ab anno 1844 e seminibus Karwinskianus.” At least 170 specimens have been annotated. 2. Eucnide grandiflora (Groenl.) Rose, Contr. U. S. Natl. Herb. 3: 317. 1895 Microsperma grandiflora Groenl. Rev. Hort. 1861: 349. fig. 84. 1861. Eucnide guatemalensis Standl. & Steyerm. Field Mus. Publ. Bot. 23: 178. 1944. Eucnide pera var. guatemalensis (Standl. & Steyerm.) Waterfall, Rho- dora 61: 242. 1959. Inflorescences of a few large flowers, terminal, often each flower sub- tended by a small leaf. Corolla to 8 cm. long, nearly white, the petals spreading. Anthers exserted, the filaments to 11 cm. long. Stigma ex- ceeding the anthers, conspicuously enlarged, 2-3 mm. long, 2—2.5 mm. broad, the stout style to 10-12 cm. long. Pedicels to about 4 cm. long, not elongating in fruit. The closest affinity is with Eucnide bartonioides. Eucnide guatemalensts (holotype, Steyermark 50818, Guatemala, F), known only from the type collection, appears to be conspecific with E. grandiflora; varietal status seems unnecessary. This species first became known through the illustrated description of Microsperma grandiflora Groenland. The plant described was grown in Bruxelles from seeds obtained from the firm of Vilmorin-Andrieux, France. The original source of material may have been a collection by Roezl made in Mexico before the end of 1861. Rose based his new com- bination on Groenland’s description and plate of M. grandiflora on the grounds of the unusually large flowers and the color of the petals. The very distinctive character of the stigma, however, is not shown in Groen- land’s plate, which otherwise more or less resembles Eucnide bartonioides. In the apparent absence of holotype or isotype, the above plate should be taken as lectotype for both Rose’s and Groenland’s binominals. The material designated by Waterfall as “lectotype” C. G. Pringle 10077, 462126; see Rhodora 61: 242. 1959) would be a neotype and inap- propriate in view of the above plate. DIsTRIBUTION. Mexico: State of Oaxaca; and Guatemala: Huehue- tenango (Map 3). SuccEsTEpD LEcToTyPE: The Fig. 84, p. 350, in Groen- land, Rev. Horticole 1861: 349-351. 1861. At least 17 specimens have been annotated. 1967 | THOMPSON & ERNST, EUCNIDE 81 3. Eucnide lobata (Hook.) A. Gray, Boston Jour. Nat. Hist. 6: 192. 1857. Microsperma lobatum Hook. Icon. Pl. 3: pl. 234. 1839, “Microsperma lobata.” Mentzelia lobata (Hook.) Walp. Repert. Bot. 2: 224. 1843. Eucnide floribunda S. Wats. Proc. Amer. Acad. 17: 358. 1882, Inflorescences of few to many flowers, usually more or less terminal (or flowers sometimes solitary, axillary). Corolla usually 6-12 (or 15) mm long, rotate, the lobes broad. Anthers and stigma exserted and usually about equal. Stigma capitate (or sometimes about 1 mm. long, oblong). Pedicels short (rarely 2 cm. long) at anthesis, usually reflexed in fruit but not elongating conspicuously. The stamens developmentally are in two ranks. The inner ones are convolute around the base of the style when the flowers first open and the outer ones are extended but shorter than the stigmas. On the second day the inner stamens elongate past the outer ones and equal the length of the stigma. A very few collections have giant-sized corollas. Eucnide floribunda (holotype, E. Palmer 832, Coahuila, cH) was based upon some- what distinctive specimens with strict, elongated, more or less secund in- florescences. We have not seen the holotype of Microsperma rudis S. Schauer, Linnaea 20: 721. 1847 (presumably based upon Aschenborn 233, collected in Mexico); it was the opinion of Gray (1857), Urban & Gilg (1900), and Waterfall (1959) that this binomial belonged in syn- onymy under £. lobata. The northeast margin of distribution of E. lobata overlaps the range of EF. bartonioides, and there would be the possibility of some hybridization. Although these species normally are strikingly distinct, some depauperate specimens are difficult to determine and, in some instances, could be annotated as extreme variants of either species. Eucnide watsonii Urban & Gilg, Nova Acta Akad. Leop.-Carol. 7: 105. 1900, was based upon plants with few and rather small flowers (petals about 10 mm. long; stamens about equalling the petals) on relatively long pedicels (2-5 cm. long). The holotype, E. Palmer 1067 (K), was collected between Tampico and San Luis Potosi, Mexico. An isotype at the Gray Herbarium also agrees with the above description and confirms, more or less, the placement by the original authors of this material between E. lobata and E. bartonioides. Waterfall (1959) treated the binomial as a synonym of E. lobata. During our study we have placed it alternately in Synonymy under both E. lobata and E. bartonioides. Several other her- barium collections can be matched more or less with the holotype and isotype of E. watsonii but we remain in doubt regarding the biological status of this material. DistripuTIon. Mexico: eastern states from Coahuila to Puebla (Map 3). Hototyper: Berlandier, Santa Catarina, near Monterrey, Nuevo Leon, Mexico (kK). This sheet has on it pencil sketches for the plate that ap- peared with the original description in Jcones Plantarum. At least 142 Specimens have been annotated. 82 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 Series Tubiflorae Thompson & Ernst, ser. nov. Series in sect. Eucnide, sed differt: corolla longe cylindrica, petala late imbricata, quasi tubus angustus, stamina in fasce angusto. Circa 4 species inclusae. Typus: E£. cordata (Kell.) Kell. ex Curran. 4. Eucnide cordata (Kell.) Kell. ex Curran, Bull. Calif. Acad. 1: 137. 1885. Mentzelia cordata Kell. Proc. Calif. Acad. 2: 33. 1860. nflorescences of many flowers, crowded, terminal, and much branched (sometimes elongating in fruit). Corolla about 1.5-2.5 cm. long, white Sect. SYMPETALEIA *% E. tenella A E. rupesiris e —. aurea Map 2. Distribution of section SympeTatera in the region of the Gulf of California, Mexico, 1967 | THOMPSON & ERNST, EUCNIDE 83 (or yellowish), narrowly cylindrical, holding the stamens in a narrow cluster. Anthers exserted, well exceeded by the small, more or less clavate stigma. Pedicels short (rarely to 2.5 cm. long), not appreciably elongating in fruit. This species most closely resembles E. hirta, of south-central continental Mexico, but is distinguished by its almost white corolla, very long exserted style, and pedicels which do not elongate. DistriBuTIoN. Mexico: Baja California Norte and Sur (including Cedros Island and some islands in the Gulf of California) and near Guaymas, Sonora. (Map 4). Hoxorype: Dr. J. A. Veatch, “Cerros Island.” The species has been recollected on Cedros Island, its only insular station on the Pacific side of Baja California, by several botanists includ- ing Greene, Palmer, Anthony, and Mason. None of the original material seems to have survived at the California Academy of Sciences. An appar- ent isotype is preserved at the Gray Herbarium and probably this should be taken as the lectotype. At least 190 specimens have been examined. 5. Eucnide hirta (G. Don) Thompson & Ernst, comb. nov. Mentzelia hirta G. Don, Gen. Syst. Gard. Bot. 3: 66. 1834, Eucnide sinuata S. Wats. Proc. Amer. Acad. 17: 358. 1882. Eucnide nelsonii Rose, Contr. U.S. Natl. Herb. 12: 286. 1909. Eucnide pringlei Rose, Contr. U.S. Natl. Herb. 12: 287. 1909. Inflorescences of few to several flowers, more or less terminal or axillary. Corolla about 2.5 (4) cm. long, yellow, narrowly cylindrical, holding the stamens in a narrow cluster. Anthers exserted. Stigma linear (or narrowly oblong), to 2 mm. long, more or less equalling the anthers (or sometimes exceeding them). Pedicels well developed, often 2 cm. long at anthesis, elongating in fruit to as much as 8 cm. b The holotypes of Eucnide sinuata (Botteri 266, cu) and of E. nelsonii (E. W. Nelson 6926, Michoacan, us 399295) are conspecific with E. hirta. Eucnide pringlei var. pringlei (holotype, C. G. Pringle 10077, Guerrero, US 462195; H. E. Moore, Jr., 2122, Hidalgo, GH), known only from two collections, seems to be a robust form of E. Airta with larger flowers. Eucnide hirta, distinguished by its yellow corolla, shorter style, and elon- gating pedicels, appears to be a continental counterpart of E. cordata. (See also E. hypomalaca). DiIstr1BuTION. Mexico: south-central states, from San Luis Potosi to Guerrero; and discontinuously in Guatemala: Jalapa (Map 4). Hoto- TYPE: Sessé & Mocino, Mexico (BM). At least 57 specimens have been annotated, many of them as E. sinuata, but we now know that the correct name is E. hirta. 6. Eucnide hypomalaca Standl. Field Mus. Publ. Bot. 22: 41. 1940. Eucnide pringlei var. hypomalaca (Standl.) Waterfall, Rhodora 61: 239. 1959. Inflorescences of few flowers, terminal (or flowers axillary). Corolla 84 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 large, to about 4 cm. long, white, narrow, holding the stamens in a narrow cluster. Anthers and stigma exserted. It is not known whether the pedicels elongate in fruit. The flowers suggest a giant-sized form of Eucnide cordata, perhaps with the petals slightly spreading. The species is known from only five localities. We find this material distinct from E. hirta (and also from E. pringlei var. pringlei which we consider an unusually large form of E. hirta). The species is distinguished by its distribution, the large, rather coarse flowers, the whitish, heavily veined corolla, and the prominent barbed hairs of the pedicels. DISTRIBUTION. Mexico: northern and (discontinuously) southern Son- ora and near the boundaries of adjacent northern Sinaloa and southwest- ern Chihuahua (Map 4). Horotype: H. S. Gentry 1315, Sonora (F). OTHER SPECIMENS: Chihuahua: Hartman 1016 (cu, uc); Hewitt 272 (GH). Sinaloa: Breedlove 1507 (1a); Mason & Brewer 1855 (wc). Sonora: Gentry 3021 (GH, Mo); Gentry 14445 (us); Kaiser s.n., April 17, 1951 (cas). 7. Eucnide xylinea C. H. Muller, Amer. Midl. Nat. 27: 487. 1942. Flowers solitary, terminal on short leafy shoots. Corolla to 2.5 cm. long, yellow, narrowly cylindrical, holding the stamens in a narrow cluster. Anthers and stigma exserted. Pedicels 2-3 cm. long, not elongating in fruit. This very distinctive species is distinguished by its solitary flowers, especially matted habit, and villous-canescent small leaves (to about 1.5 cm. long). It is known only from three stations. DisTRIBUTION. Mexico: western central Coahuila and discontinuously in southwestern Tamaulipas near the southern Nuevo Leén boundary (Map 4). Horotype: C. H. Muller 3311, Coahuila (us 2109944). OTHER SPECIMENS: Coahuila: Johnston 9003 (cH); Muller 3311 (cH, UC); eniconae 1075 (GH). Tamaulipas: Stanford, Lauber, & Taylor 2449 (ps, RSA). Section Mentzeliopsis Thompson & Ernst, sect. nov. Flores apopetali, sed petalis basi cum tubo filamentarum breviter coalitis. Antherae in anthesi biloculares. A sectione Eucnide differt: stigma anthe- raeque in corolla inclusae, staminum filamenta crassa interne obliqua, receptaculum pollinare circumstylare formantia. Species unica. TyPus: E. urens Parry. 8. Eucnide urens Parry, Amer. Nat. 9: 144. 1875. Mentzelia urens Parry ex Gray, Proc. Amer. Acad. 10: 71. 1874 (non Vell. Fl. Flum. 5: 97, 1825), Eucnide synandra A. Nelson, Bot. Gaz. 47: 428. 909. Eucnide parryi House, Bull. N.Y. State Mus. 234: 67. 1922 (a substitute name for £. urens). 1967 | THOMPSON & ERNST, EUCNIDE 85 Inflorescences of several flowers, terminal (the flowers sometimes sub- tended by small leaves). Corollas about 5 cm. long, pale yellowish white, the petals rigid, spreading. Stamens included, the filaments somewhat thick, slanted inward around the stout style. Stigma included, slightly exceeding the anthers. Pedicels short, to about 1.5 cm. long, neither re- flexed nor elongated in fruit. DIsTRIBUTION. United States: eastern California, southern Nevada, southwestern Utah, western Arizona, and discontinuously o ecg eastern Baja California Norte (Map 1). Lectotype: C. C. y 79, Utah (cH); see U. T. Waterfall, Rhodora 61: 236. 1959. At sll 196 specimens have been annotated. Section Sympetaleia (A. Gray), Thompson & eae stat. nov. Basionym: Sympetaleia A. Gray, Proc. Amer. Acad. 12: 161. 1s Classes 2 morphologici: corollae tubus quam lobi longior, antherae inclusae, filamenta brevia (typo incluso); vel corollae tubus quam lobi brevior, antherae plus minusve exsertae, filamenta longa. Differt: corolla vere sympetala, antherae anthesin uniloculares, stamina quasi epipetala. Species 3 inclusae. Typus: Sympetaleia aurea A. Gray = Eucnide aurea. 9. Eucnide aurea (A. Gray) Thompson & Ernst, comb. nov. Sympetaleia aurea A. Gray, Proc. Amer. Acad. 12: 161. 1877. Inflorescences usually of many crowded flowers, terminal, sometimes im -” t pe. pete Sariane 100°wW a oulie ene s : ; + ) iy ee 4 Peal Pe \ Ne af ze 3 / * * . x a Supra \ \ Seay . ‘oe ‘ iy, aye ke SN ’ \ - coe 1 » ag hilear’ ? ‘ ' L ae a hd Pai oD ° eee. 1 pariah i Mead ie psy oh ree | Wyn et Pa ) : of Ly we ies “ eae: eM ke Dey Oe Nate as Bs 3 Sect. EUCNIDE ple gia INS, eC oe oe tee Og ro er. \.7 ‘ oe y oid fo Sry \ Ser. Eucnide ge Noe : a Ph ANN gat ’ * E. bartonioides eS ae Coes gs pate a YT - e E. lobata ee es i ; A E. grandiflora . 3. Distribution of section Eucnwe, series Eucnide, in Texas, Mexico and Guatemala. 86 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 elongating appreciably in fruit (or flowers sometimes solitary, axillary) Corolla 1-2 cm. long, bright yellow or strong reddish orange, salverform, the lobes broad and rounded. Stamens epipetalous, included, the anthers more or less sessile; a ring of upwardly curved hairs below the stamens. Style filiform, the stigma minute, situated among the anthers (or some- times barely exposed in the mouth of the corolla tube). Pedicels usually short at anthesis but elongating conspicuously, sometimes to 27-33 cm in fruit. This species probably includes two taxa: one has smaller yellow corollas with stigma included; the other larger reddish orange corollas with stigma becoming slightly exposed. DistrIBUTION. Mexico: Baja California Sur, including some islands in the Gulf of California (Map 2). Hototype: Dr. Thomas H. Streets, Pulpito Point (GH). At least 70 specimens have been annotated. 10. Eucnide rupestris (Baill.) Thompson & Ernst, comb. nov. Loasella rupestris Baill. Bull. Soc. Linn. Paris 1: 650. Sympetaleia rupestris (Baill.) S. Wats. Proc. Amer. wed he 50. 1889. Inflorescences usually of many crowded flowers, terminal, sometimes elongating appreciably in fruit (or flowers sometimes solitary, axillary). Corolla 1 100° Ww | 2a 6 Day eee aor Sect. EUCNIDE ind bot Ya npyl } Ser. Tubiflorae oH: ook : iss wi’ no ean OR e £. cordata ae ad . ahs. Pe igo «ae oe *£. hypomalaca ower ee ee: aw E. hirta 5 AE. xylinea Map. 4. Distribution of section Eucnine, series Tubiflorae, in Mexico and Guatemala. 1967 | THOMPSON & ERNST, EUCNIDE 87 hairs below the stamens. Style short and thick, the stigma not exceeding the lower anthers. Pedicels usually short at anthesis, sometimes reflexed in fruit and elongating up to 2 cm. DistRIBUTION. United States: Southern California and southwestern Arizona; and Mexico: Baja California Norte and Sur including some islands in the Gulf of California, also northwestern Sonora and discon- tinuously in northwestern Sinaloa (Map 2). Hotorype: Thiébault 1099 (p); photograph (a) examined, courtesy of R. A. Howard. At least 55 specimens have been annotated. 11. Eucnide tenella (I. M. Johnst.) Thompson & Ernst, comb. nov. Sympetaleia tenella I. M. Johnst. Proc. Calif. Acad, IV. 12: 1106. 1924. Flowers few, small, mostly axillary. Corolla about 8 mm. long, white (? or yellowish), the tube inconspicuous, the lobes narrow and loosely spread- ing. Anthers exserted on filiform filaments, spreading. Stigma minute, terminal on the filiform style and equalling the staminal filaments in length; ring of hairs below the stamens lacking. Pedicels filiform, about 1.5 cm. long at anthesis, elongating in fruit to 8-10 cm. The corolla while distinctly sympetalous and short tubular, very closely approaches in form the corollas of sect. EUcNIDE, series Eucnide, thus differing con- spicuously in form from the corollas of the other two species of sect. SyMPETALEIA. The species has been collected only three times. DistRIBUTION. Mexico: eastern Baja California Sur (Map 2). Hoto- type: J. M. Johnston 3091, Agua Verde Bay (cas 1239). Other speci- mens: Wiggins 15546 (ps); Wiggins, Carter & Ernst 280 (ps, LA, OKLA). ACKNOWLEDGEMENTS Many persons have helped us in the preparation of this paper. We are grateful to the curators of the herbaria where we have examined or bor- rowed specimens. The Latin descriptions were prepared by Robert C. Foster and additional scholarly assistance generously was provided by George K. Brizicky and Conrad V. Morton. In obtaining seeds for our use Annetta Carter, Roxanna Ferris, and Helen Sharsmith cheerfully risked life and limbs. Rogers McVaugh kindly led us to the lost name of Eucnide hirta. The insect pollinator of Eucnide urens was identified by Gerald I. Stage. Peter H. Raven corrected a final draft of the manuscript. A grant from the National Science Foundation provided financial support. It is an especial pleasure to acknowledge the faithful and competent assistance of Karen Bartholomew who has helped provide much of the technical information. LITERATURE CITED DarLincton, J. A monograph of the genus Mentzelia. Ann. Missouri Bot. Gard. 21: 103-226. pls. 4-6. 1934. 88 JOURNAL OF THE ARNOLD ARBORETUM Urspan, I., & E. Gmc. Monographia Loasacearum. Nova Acta Akad. sa 76: 1-370. pls. 1-8. WarTERFALL, U. T. A revision of Eucnide. Rhodora 61: 231-243. 1959. DEPARTMENT OF BOTANICAL SCIENCES UNIVERSITY OF CALIFORNIA, Los ANGELES AND DEPARTMENT OF BOTANY SMITHSONIAN INSTITUTION, WASHINGTON, D.C. AND THE UNIVERSITY OF KANSAS, LAWRENCE [ VOL. 48 Leop.- 1967] JAYAWEERA, DUABANGA 89 THE GENUS DUABANGA Don M. A. JAYAWEERA THE GENUS Duabanga was first suggested by Buchanan-Hamilton as a new genus with characters connecting the Lagerstroemias and the Son- neratias. Probably the name is derived from “Duyabangga” a vernac- ular name for the tree in Tripura. Roxburgh (1832) had collected and described as Lagerstroemia grandiflora a plant which was the same as that collected by Hamilton and sent to Sir Joseph Banks in 1798 and described by him as Duabanga sonneratioides in 1835. Roxburgh’s L. grandiflora was later transferred to the genus Duabanga by Walpers. Duabanga and Sonneratia compose the family Sonneratiaceae. The two genera have been placed under various other families by different authors. Lindley (1836) placed Sonneratia in the tribe Myrteae (Myrtaceae) and Duabanga in Lagerstroemieae (Lythraceae) while Mi- quel (1855) placed Sonneratia in the tribe Sonneratieae (Myrtaceae) and Duabanga in Lythreae (Lythraceae). Bentham and Hooker (1867) in- cluded both genera in the Lythreae (Lythraceae) and this was adopted by Koorders and Valeton (1894), while Koehne excluded both genera from his monograph on the Lythraceae. Niedenzu (1892) placed Dua- banga in the family Blattiaceae along with Sonneratia as a synonym to Blatti. Engler (1897) formed the family Sonneratiaceae to include the two genera. The family is distinguished from the Lythraceae by flower, pollen, and to the receptacle, forming a distinctive half-inferior fruit. The smooth pollen has two or three regularly placed germ pores, and fiber bundles are distributed in the pith. The small genus Duabanga consists of three species, D. moluccana B1., D. grandiflora (Roxb. ex DC.) Walp. and D. taylorii sp. nov., all large trees growing in the rain forest extending from the southeastern Himalaya to New Guinea and on the islands in between. A new species, Duabanga taylorii, described here, is named for Sir George Taylor, Director of the Royal Botanic Gardens, Kew. The seed of this species was received, probably from Java, in 1853, and four trees planted in the Royal Botanic Gardens, Peradeniya, have now reached an average height of over 128 feet, 26 feet in girth at breast height, and a spread of 79 feet (Fic. 6). KEY TO THE SPECIES Stamens uniseriate. Stamens 12; trunk columnar; flowers 4-merous; ovary dome- oe fruit ee: ois es hich i eee Fed eM D. moluccana. 90 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Stamens 24-45; trunk branched 1.8-6 m. from ground; flowers 4- or 5-merous; ovary ovate: frat 15-2.5 cm, long... 065.5... .. eo lee ke D. taylorit. Stamens biseriate: trunk straight and columnar; flowers 6-merous; stamens in- definite; ovary conical Set SF Be 2 ONES es ini lee as D. grandiflora. C “peta transverse section of the ovary; «, fruit, lateral view; H, dehiscing fruit; . ) transverse section of a fruit before dehiscence : K, seed. 1967 | JAYAWEERA, DUABANGA 91 Duabanga moluccana Blume, Mus. Bot. Lugd.-Bat. 1: 109. 1849. (Fic. 1) Duabanga borneensis R. Knuth, Repert. Sp. Nov. 38: 121. 1935. Large or medium-sized trees, 8—42.4 m. tall, trunk more or less colum- nar, branching at 23 m. from ground, unbuttressed or faintly buttressed, young branchlets quadrangular, winged, becoming terete when mature, with 9-15 pairs of leaves; young leaves and stems covered with short adpressed hair but becoming glabrescent, internodes 2—9.5 cm. long, with 2 pairs of projections corresponding to bases of wings at base, branches brittle; leaves simple, opposite, oblong, oval, ovate, lanceolate or ovate- oblong, 4—-32.6 cm. long, 2.1-12.5 cm. broad, more or less cordate with rounded lobes or rounded at base, acuminate, acute, coriaceous, glabrous, entire, costa prominent below with 5—28 pairs of widely patent, arcuate, lateral veins merging into an intramarginal vein, petiole 0.3—1.2 cm. long; flowers regular, bisexual, perigynous in simple or compound, dichasial, terminal, minutely pubescent corymbs, buds ovoid-oval, pubescent, 0.4—2 cm. long, 0.3-1.5 cm. broad, bracts lanceolate or spathulate, caducous, 0.65—1.1 cm. long, hairy, pedicels 0.2—2 cm. long, strongly quadrangular, ridges extending to fused margins of sepals; calyx funnel-shaped, 1—2.3 cm. in diameter, segments 4, oblong-ovate, 0.5—1.6 cm. long, 0.4-1 cm. broad, short acuminate, acute, thick, fleshy, green, persistent in the fruit; petals 4, greenish cream-white, with center pink, 1.6-1.8 cm. long, 0.8—1.1 cm. broad, short clawed, valvate, overlapping in bud, caducous; stamens 12, uniseriate on a narrow circular rim round the ovary, filaments greenish white, broad and flat at the base, filiform, recurved, caducous, about 3 cm. long, anthers brownish white; ovary half-inferior, dome-shaped, broader than tall, 0.5 cm. long, 4-locular, with numerous ovules on cushion-shaped placentae, style 2.5—4 cm. long, greenish yellow, stigma faintly 4-lobed, dark green with a brown center; fruit oblong-ovoid, loculicidally dehiscent, capsule 4-valved, 1—-3.7 cm. long, 0.8—2 cm. broad, slightly narrowed at base on a stout quadrangular pedicel with persistent calyx segments; seeds numerous, testa drawn out on either side, 5 mm. long from end to end. ILLUSTRATIONS. VipAL, Synopsis de Familias. Atlas pl. 52, fig. F, 1-5. 1883; Koorpers & VALETON, Atlas der Baumarten von Java 4: fig. 784. 1918; JAvAWEERA & Howarp, Baileya 10: fig. 5. 1962 DIstrIBUTION. The tree flourishes along streams and slopes in primary forests of Java, Borneo, Celebes, Moluccas, New Guinea, Talaud, Lesser Sunda and Philippine Islands at altitudes between 10 m. and 1250 m., attaining a height of 842.4 m. and a girth of 1.12-2.3 m. Borneo. SARAWAK: Haviland & Hose 3622 M; Gaat, Upper Rajan River, Clemens 21548. Britis NortH Borneo: Paitan, Maidin 2669; Pinta-san, Clemens 34202; Sandakan, Elmer 20262; Sapilok & Kabili s.n.; without locality, Wood s.n.; Kadir A531, A561. NETHERLANDS BorNEO: Berau, 'NIFS b.b. 18890, b.b. 19237, b.b. 18872, b.b. 18811; Sangkulirang, Aet (exp. E. Walsh) 695; Sambodja, de Voogd 1610: Klumpeng, N/FS b.b. 17272. Celebes. Poso, Kalaena, JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 Fic. 2. Duabanga taylorii, a drooping branchlet and a mature leaf in outline. 1967 | JAYAWEERA, DUABANGA 93 NIFS b.b. 28740; Malini, NIFS Cel/III-90; Manado, NIFS b.b. 28231. Halma- hera. Motuccan TernateE: Batjan Is., NJFS b.b. 16441; Weda Is., NIFS b.b. 24926. Philippine Islands. Luzon: Bulacan Prov., Ramos 21746; Angat, Llanos 233; Rizal Prov., Vidal 2883; Bosoboso, Ramos 4577; Ahern’s collector 3246; Ramos 1341; Sorsogon Prov., Irosin (Mt. Bulusan), Elmer 15326; Tabayas Prov., Guinayangan, Hagger 254; Laguna Prov., Los Banos (Mt. Maquiling), Elmer 18275; Mount Prov., Tanit, Lizardo 20386; Albay Prov., Vidal 2 without locality, Ahern US 446105; Garcia 1110; Loher 2149; Whitford oe Merritt 11409; Ahern 158, 106; Ahern’s collector 2003. NEGROS: Santa Cruz, William’s collector 2885. MINDANAO: Davao Prov., de Mesa 27477; Pagpawan Sitio, Edano 11601; Cotabato, Ferraris 23045; without locality, Ahern, US 445972, Miranda 20521, Quadras 336. New Guinea. NETHERLANDS NEw GUINEA: Idenburgh River, Bernhard Camp, Brass & Versteegh 13514, 14013; Palmer River, Brass 7289; Madang Dist., Hoogland 5224. British NEW GUINEA: Papua, Djamu, Schlechter 17582. The leaves of collections from Borneo, Elmer 20262, NIFS 6.6. 18811, 18872, and 19237, are rounded at base, while those from Celebes, V/FS Cel/III-90, b.b. 28740 and Ramos 4577 from Luzon are faintly cordate, and those from elsewhere are quite cordate at the base. The fruits of Brass & Versteegh 14015 from New Guinea are fusiform and, along with those from Sarawak, Clemens 7287, are much larger than those from col- lections elsewhere. The tree growing at the Royal Botanic Gardens, Peradeniya, Ceylon, is 42.4 m. tall, 2.3 m. girth at breast height, with a columnar trunk branch- ing at 23 m. from ground level. It flowers throughout the year This species is distinguished from the others by its columnar trunk and smaller, 4-merous flowers with 12, uniseriate stamens. Duabanga taylorii sp. nov. (Fics. 2, 3) Arbor ingens, late patens, ramunculis pendentibus; folia magna, op- posita, oblonga, integra, glabra, apice acuminata, basi cordata cum lobis inaequaliter orbiculatis, 5-25 paribus arcuatarum costarum infra conspicu- arum cum vena intramarginali. Flores flavidi-albi, 5 cm. diam., in cymis simplice vel dichotomice terminalibus compositi; sepali 4 aut 5 aut plus, triangularia, Seva carnosi, in gemma uniti; petali 4 aut 5 aut plus, liberi, flavidi-albi, ungulati, caduci; stamina 24-45, l-seriata, filamentis longis, antheris Doi versatilibus; ovarium semi-inferum, carpellis 4 aut 5 aut plus, compositum, 4- aut 5- aut plus locularibus, placenta axil- lari; capsula loculicidalis cum sepalis persistentibus; semina plurima, minima, filiformia, testa obvia. Large trees, 33—44.5 m. tall with a spread of 22.6-25.6 m., trunk 7.3-8.8 m. in circumference branching at 1.8-6 m. from ground level, buttressed, young branchlets 1.2—5.2 m. long, quadrangular, soon becoming terete when mature, drooping, brown and lenticelled, internodes 5—12.5 cm. long, 13— 44 pairs to a branchlet; leaves opposite, oblong, 2.7—-43.5 cm. long, 1.4—17.2 cm. broad, cordate at base, lobes unequal, rounded, abruptly acuminate, 94 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Pr 1-5 CM. : Fic. 3. Duabanga taylorii. a, branch with inflorescence; B, flower, lateral view; Cc, flower with the sepals and petals removed to show the stamens; D, petal, spread out; gE, longitudinal section of the half-inferior ovary, petals and stamens removed; F, transverse section of an ovary; G, stamen at an early stage showing the recurved filament; H, stamen showing the hardened resinous globules at the back of the anther; 1, seed; J, fruit with persistent calyx and base of style. acute, entire, glabrous, veins 5—25 pairs, arcuate, prominent on the under surface with an intramarginal vein, petiole 0.5-0.6 cm. long, young leaves reddish brown in color; flowers regular, bisexual, perigynous, 5 cm. diameter, yellowish white, in simple or dichasial terminal cymes, buds ovate, 2.2 cm. long and as broad, puberulent, later becoming glabrous, pedicel 1.1 cm. long, articulate, quadrangular or terete; sepals 4 or 5, 1967 | JAYAWEERA, DUABANGA 95 Leaf measurements of Duabanga taylorii Qo 10 20 30 40 50 Position of the leaf in the Branchlet Fic. 4. Duabanga taylorii, showing the variation of leaf size in relation to position of leaves on the branchlets. rarely 6 or more, valvate, triangular, fleshy, green, 1.8—2 cm. long, 1.2-1.8 cm. broad, fused in the bud; petals 4 or 5, rarely 6 or more, free, valvate, 2.5-2.7 cm. long, 2—2.1 cm. broad, obovate, short clawed, yellowish white, crinkled and plicate at the margin; stamens 24-45, uniseriate on a narrow circular rim round the ovary, filaments 4—4.7 cm. long, broad and flat at the base, recurved in bud, anther 1 cm. long, splitting longitudinally, versatile, introrse, glandular at the back, exuding a secretion which on drying hardens to resinous globules with pollen grains embedded within; ovary half-inferior, ovate, half-buried in the receptacle, upper portion dome shaped, 4—7-locular with axile placentation, style 5 cm. long, yellow- ish green, stigma capitate and dark green; fruit a broadly ovate, loculi- cidally dehiscent capsule, 1.5—2.5 cm. long, 1.7—2.5 cm. broad with persis- tent calyx segments; seeds numerous, very small, filiform, testa attenuate, 5.5 mm. long from end to end. DistriBuTIoN. This is probably a native of Java; seed was introduced into the Royal Botanic Gardens, Peradeniya, Ceylon, in about 1853. It flourishes along river banks in the moist mid-country at an elevation of 457 m. above sea level. Ceylon. PerapENtIvA: Royal Botanic Gardens, Jayaweera 2498, July 8, 1964, holotype (PDA) and 4 isotypes; Jayaweera 2496, 2497. There are four large trees of this species about 112 years old, growing at the Royal Botanic Gardens, Peradeniya, most probably the original introduction. The drooping branchlets bear 13-44 pairs of leaves. The 96 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 first pair and the last pair of leaves just before the production of the inflorescence are very small, while the intermediate ones are larger, the largest being in positions 16-18, 27 and 28, and 32-39 all over 30 cm. long. The largest leaf measured was the 34th which was 34.8 cm. long and 13.7 cm. broad (Fic. 4) Floral counts of over 500 flowers were made (TABLE 1), and they were predominantly 4- or 5-merous. Some 17.6 per cent of the flowers had 30 stamens each, 13.4-15 per cent had 27-29 stamens and 13 per cent had 31 stamens each. This species, Duabanga taylorii, differs from the others in the lower branching habit of the main trunk, the number of calyx and corolla seg- ments, and the number of stamens which are uniseriate. It is, however, related to D. moluccana by its 4-merous flowers but differs in the number of stamens. TABLE 1. Variation in Numbers of Floral Parts in Duabanga taylorii. STAMEN NUMBER 20 Or 20 ek ee 2) Oy a oe 60 oF foe 66 AD R O SEPALS, PETALS, AND CARPEL FLOWERS 4,4, 4 DAS 90. 57a a aE ee ae ee ee 4, 4, 5 eee | 0 90 ee ee 60 4, 4, 6 Se he, dom, Ee Re ee 53 RR a ee ee Uae a eae) ee bik 1 By ty Doc big ete etree ie te 1—- —- —- —- ~—~ ~—~ — — 1 By2hgn 4 —- -—-— SF FS Eee ee se lh lh 1— — 1 ee Fae rey eS ee a RO eae PO) Sie Stee vies 125 SSL eter ce a ea A oe ee ee ee 5 Gps Se ae res Pe ee AS ae 3 6, 6, 4 FS ee i ae i ee Di eS Z 6, 6, 5 cult eiiceatiieestic et Ge Coe ces ee cae 8 6,6,:6 a Sipe cee oe Re Ae ee Be on el ee — 10 1, 66% st onli “feat aca, pat gai AGM apy: an pel lek AE alee 1 an Pa: er ean SE Se Se 0 etl ge ee ee en eet aS 1 8, 8, 6 i ae Se ee i Se ee es 1 OT INDIVIDUALS LEE SE GP IS e715 S865 SAS 26 ie ee 1 500 Duabanga grandiflora (Roxb. ex DC.) Walp. Repert. 2: 114. 1843. (Fic. 5) pa gl isang rege pee Bengal. 39. 1814) ex DC. Mem. Soc. t. Nat. Genéve 32: eae anga sonneratioides ao ee Trans. Linn. Soc. 17: 177-178. 1835. Leptospartion grandiflorum Griff. Ic. Pl. As. 4: 591. 1854. Medium or tall tree, 6-40 m. tall, all parts glabrous, trunk straight, erect, with a smooth ash-colored, wrinkled and fissured bark; branches spiral or whorled on the trunk, drooping; branchlets quadrangular be- 1967 | JAYAWEERA, DUABANGA SAN Beir 5. Duabanga grandiflora, drawn from herbarium specimen Howard & IG. Wagenknecht 15058, January 19, 1960; stamens and petals have fallen from the flowers. a, inflorescence and a leaf; B, reconstructed flower; c, petal; D, dehiscing fruit, lateral view; E, seed; F and G, abnormal forms with 5-merous and 4-merous perianths, respectively. 98 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 coming terete when mature; leaves opposite, large, distichously arranged ; g, or ovate-oblong, cordate at base, lobes rounded, short-acuminate, entire, chartaceous, glabrous above, glaucous beneath, with 12 i ginal vein, petiole short, stout, 4-1.2 cm. long; flowers large, regular, bisexual, perigynous 2.2—3.3 cm. across or more, many in lax, drooping, dichasial corymbs at the ends of branches, pedicels articulate, 1.5—4 cm. long; flower buds conical, 6-ridged, 1.8—2.5 cm. long, 1.8—2.2 cm. a sepals 6, sometimes 5 and very Fic 6-8. pare: taylorii. 6, A fu Persdnaive 7, 8, Magnifications of the lenticelled bark of the trunk. ll-grown tree in the Botanic Gardens, 1967] JAYAWEERA, DUABANGA 99 rarely 4, 1.2-2.5 cm. long, 0.8-1.3 cm. broad, ovate, acute, valvate, fused at the base with the receptacle into a wide cup-shaped calyx tube, 1.7—2 cm. across; petals 6, free, valvate, clawed, + 4.4 cm. long, + 2.7 cm. broad, obovate, plicate along the margin; stamens numerous, perigynous, biseriate, filaments long, filiform, 4.7—5.7 cm. long, broad at base, anthers versatile and curved; ovary half-inferior, conical, 1—1.4 cm. long, 1.2 cm. broad, 6-locular with numerous ovules on axile placentae, style 3.2-6 cm. long, stigma slightly lobed; fruit an ovoid-globose, pendulous, glabrous capsule, 3 cm. long, 3.2-3.5 cm. broad with persistent calyx segments, dehiscing loculicidally into 6-9 valves from apex downwards, seeds numerous, minute, filiform, 4 mm. long from end to end of the attenuate ILLUSTRATIONS. Hooker f., + Nai Plants, pl. 11. 1855; JAYAWEERA & Howarp, Baileya 10: pl. 4. Distemurion. This species grows along banks of streams and in ra- vines from sea level up to 2280 m. elevation in Assam (India), Andaman Islands, Burma, Thailand, Cambodia, Malay Peninsula, Indo-China, and in Yunnan on the mainland of China. It is deciduous and flowers from January to April, bearing fruit from April to June. India. SrkKiIm: without locality, Hook. f. & Thomson ee ie eae Kowan 24480. Assam: Duars, Biswas 1943; Lakhimpur, Rock 960; S. Lushai Hills near Fort Lunglek, Gage 72; without locality, King’s satin 252; Cal- cutta Bot. Gard., Biswas 9400, cultivated. Andaman Islands. Long Island, ‘Kirat Ram 3666; S. Andamans, Prain’s collector 19. Burma. CHiN HILts: Kaupetlet, Dickason 8616; Rangoon, Myout Chaw, Dickason 6942, Thailand. Khwae Noi River Expedition 1946, Bloembergen 21; Ban Khai, den Hode & Kostermans 473; Doi Luang, Rock 1802; without locality, Rock 665, 992; Mrs. Collins 953, 1141. Malaya. PAHANG: Cameron’s Highlands, Henderson 23662. Cam- bodia. Without locality, Pierre 536. Indo-China. Tonck1nc: Hoo Bink, Petelot 6371, China. YuNNAN: Lang-tsang Hsien, Wang 76628; Tsang-Yuang, Wang 73303; Cheli Hsien, Wang 79292. U.S.A. FtormpA: Homestead Subtropical Exp. Station, Howard & Wagenknecht 15058, cultivated. This species of Duabanga is distinguished from the others by its straight trunk, large 6-merous flowers, and numerous biseriate stamens. Uses. The wood is used for building purposes; it is close grained and coarsely fibrous and takes on a mottled grayish polish. BIBLIOGRAPHY Backer, C. i rca Flora van Java. Em. ed. 4. Fam. 73. 1-3. 1942. G. J. vAN STEENTIS. Flora Malesiana. I. 4: 280-289. 1951. BENTHAM, o & 7 D. Hooker. Genera Plantarum 1: 783, 784. 1867. BLUME, K. - Museum Botanicum 1: 109. ; Bor, N. L. Manual of Indian Forest Botany 223, 361. 1953. Crarke, C. B. Jn: Hooker, Fl. Brit. Ind. 2: 578, 579. 1879. Corner, E. J. H. Wayside Trees of Malaya 1: 427, 428. 1940. ENGLER, A. Nat. Pflanzenfam. Nachtr. zu III (7). 261. 1897. Hamitton, F. (BucHanan). Trans. Linn. Soc. Lond. 17: 177, 178. 1835. 100 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Hevne, K. De Nuttige Planten van Nederlandsch Indié. Ed. 2. 2: 1157. 1927. Int. Cope Bot. Nomenct. App. II. Nomina Fam. Conserv. 200. 1961 Kino, G. Jour. Asiat. Soc. Bengal 67(2): 10. 1898. KoeEuneg, A. Bot. Jahrb. 1: 305-335. 1881. eae: S. ies & Th. Vateton. Bijdrage No. 1. Boomsorten van Java 186- 205. 189 Kurz, e Prelit Rept. on Forest and other Vegetation of Pegu. App. B. 54. 1875. — . Forest Flora of British Burma 1: 525, 526. 1877. LinpbLeEY, J. A Natural System of Botany. Ed. 2. 45, 101. 1836. MerriLL, E. D. Enum. Philip. Fl. Pl. 3: | fis . Philip. Isl. Bur. Forestry Bull. 1: MrqvEL, F. A. W. Flora van Nederlandsch ae 1a 495-498, 624, 625. 1855. Mott, J. W., & H. H. JANssontivs. Agar at as des Holzes der auf Java vorkommenden Baumarten. 3: 594-604. ee . In: Engl. & Abts ties i 3(7): 16-21. 1892. t. Jahrb. 15: 167. RENDLE, ‘i B. The i Mec of Flowering Plants 2: 377. 1925. Riwwiey, H. N. Flora of the Malay Peninsula 1: 824, 825. 1922. RoxBurcH, W. Hort. Bengalensis 38. 1814. . Flora Indica, ed. Carey. 2: 503-505. 1832 SOLEREDER, H. Systematische Anatomie der Dicotyledonen 417. 1899. VipaL Y SoLer, S. Phanerogamae Cumingianae Philippinarum 115. 1885. Watpers, G. G. Repert. Bot. Syst. 2: 114. 1843. a, Ann. Bot. Syst. 2: 541. 1852 Roya. Botanic GARDENS, PERADENIYA, CEYLON 1967] LEE, STUDIES IN SWIETENIA 101 STUDIES IN SWIETENIA (MELIACEAE): OBSERVATIONS ON THE SEXUALITY OF THE FLOWERS HsuEH-YUNG LEE THE FLOWERS OF MELIACEAE have usually been described as perfect, or rarely unisexual and polygamo-dioecious. Of all the references con- sulted, only C. S. Sargent, in his Silva of North America, discussed the sexuality of the flowers of Swietenia, which he described as perfect. After three years of observation and pollination tests I have come to the con- clusion that the flowers of Swietenia are, rather, unisexual, and that the trees, which bear both staminate and pistillate flowers, are monoecious. These studies were conducted on sixteen-year-old trees of both Swietenia mahagoni Jacquin and S. macrophylla King at the plantation of Chungpu Branch Station of the Forest Experiment Station of Taiwan. Some trees 30 to 40 years old were also examined. The factors considered in the observations were the distribution of staminate and pistillate flowers on the inflorescence; the structure of the flowers; the sex of fallen flowers; and the results obtained from selfing of staminate flowers by bagging, selfing pistillate flowers by bagging, and by controlled pollination. The flowering period of Swietenia in Taiwan is from about 20th to the 30th of May, with the flowers of S. macrophylla always opening about one week earlier than those of S. mahagoni. The work was carried on and repeated during the years 1963 to 1965. Of about 700 flowers of the two species of Swietenia examined no per- fect flower was found. Although the appearance of the flowers is similar, the structure of the staminate and pistillate flowers is quite distinct. STAMINATE FLOWERS. The ovary of the staminate flower is conical, with a long style and discoid stigma. The ovary, which is about 2 mm in diameter, bears many reduced ovules. The anthers are sessile and in- serted on the inner side of the staminal tube slightly lower than the clefts. The stigma is borne above the anthers so the shedding pollen adheres to the style instead of the stigma (Fics. 1 and 2). PISTILLATE FLOWERS. The ovary of the pistillate flower is globular and about 3 mm. in diameter. The style is very short, and the discoid stigma is thickened. The stigma is borne slightly below the clefts of the staminal tube and is encircled by the reduced anthers. The anthers become black when the flower opens and shed no pollen. In each of the five locules of the ovary are 12—14 well developed ovules (Fics. 1 and 2). INFLORESCENCE. The axillary paniculate inflorescence is composed of many small 3-flowered cymes in which the central flower opens first. These determinate cymes are borne in turn on an indeterminate panicle. The inflorescence should, therefore, properly be called a thyrse (Fics. 7 and 8). The inflorescence of Swietenia mahagoni is about 10 cm. long and is 102 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 a Fic. 1, flow ers of eet Serio a9 eee a petals and portion of staminal tube removed, X ca. 36; Fie. 2, of S. mahagoni, some petals and portion of staminal ie removed, X ca. re oa 3, a cat section of staminate 1967] LEE, STUDIES IN SWIETENIA 103 much shorter and more slender than that of S. macrophylla, which reaches 20 cm. or more and is stouter. Most of the inflorescences in both species bear only staminate flowers; a few bear only pistillate flowers, and some inflorescences on the tops of trees bear both staminate and pistillate flowers. The number of flowers in each inflorescence varies in the two species. In S. mahagoni the number of flowers in a single thyrse is from 10 to 50, while in S. macrophylla it varies from 40 to 100. On a small branch of S. mahagoni examined there were 360 pistillate flowers and 204 staminate flowers on 8 inflorescences. From our observations it seems that the number of flowers and the proportion of pistillate flowers to staminate will vary with the nutritional state of each tree. One tree under observation (no. L;) bore many pistillate flowers in 1963 and 1964 but only a few in 1965. On a whole tree, according to our observations, the number of staminate flowers is about ten times more than the pistillate ones. EXAMINATION OF FALLEN FLOWERS. Hundreds of flowers which had fallen at flowering time were examined. None of these flowers was pis- tillate. The flowers seem to have fallen naturally after the shedding of pollen. The few pistillate flowers which were occasionally found on the ground consisted only of the receptacle and swollen ovary, the fallen staminate flowers, on the other hand, were complete. It seems reasonable to conclude, therefore, that the fall of the pistillate flowers was due to nutritional deficiency, not to incomplete fertilization. T he small inflor- escence branch apparently can not supply enough nutriment for all the developing ovaries. This condition seems to make quite clear that the staminate and pistillate flowers of Swietenia are distinctly separate. In an effort to show that the pistillate flowers bear no functional anthers they were bagged before the opening of the corollas. All the bagged pistillate flowers faded in three days because no pollen grains were pro- duced. The staminodia of these pistillate flowers proved to be function- less. In the same way staminate flowers were also bagged before the opening of the flowers. The same result was observed. The reduced ovary of the staminate flowers did not develop, although the pollen was viable. CONTROLLED POLLINATION. The pollen shed from anthers of staminate flowers always forms masses with sticky fluid. Some of the pollen masses ovaries developed and formed rounded young fruits, many of which fell because of growth competition. Usually only one fruit on each inflor- escence developed to maturity. Since the establishment of the genus Swietenia by Jacquin in 1760 the flower bud of S. mahagoni showing pollen grains in anther, X ca. 100; Fic. 4, longitudinal section of pistillate flower bud of S. mahagoni, showing empty anthers, X ca. 100; Fic. 5, cymule of S. macrophylla, X ca. 36; Fic. 6, cymule of S. mahagoni, X ca. 36; Fic. 7, inflorescence of S. macrophylla, X ca. %4; Fic. 8, inflorescence of S. mahagoni, X ca. “A. 104 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 48 flowers have always been described as perfect (hermaphroditic). Tax- onomists based their conclusions on the morphological structure of the flowers neglecting to consider whether the organs were functional or not. From the relative position of anthers and stigma it is easy to determine the sex of the flower. In the references cited, Sargent’s illustration shows a staminate flower; Rendle’s illustration of Swietenia mahagoni a pistil- late flower; King’s figure of S. macrophylla a pistillate flower; and Liu’s monograph a staminate flower of S. mahagoni As a result of these investigations and pollination tests I conclude that the flowers of Swietenia are unisexual and trees of this genus are monoe- cious. BIBLIOGRAPHY BAILEY, H. Manual of Cultivated Plants. 612, 613. 194 Barciay, F. W. Swietenia in L. H. Bartey, The Standard eats of Horti- BentHaM, G. & J. D. Hooxer. Genera Plantarum 1. 1862. CANDOLLE, C. DE Meliaceae in A. & C. De CANDOLLE, Monographiae Phanero- gamarum 1. 1878. GARDENER’S CHRONICLE. III. 54: 437. 1913. Kine, G. Jn: J. D. Hooker, Icones Plantarum 16: ¢. 1550. Liv, T. S. Illustrations of Native and Introduced lee ae of Taiwan. Liv, Y. C. Dendrology. 1954 [in Chinese]. Renpver, A. Manual of Cultivated Trees and Shrubs. 1940. RENDLE, A. B. The Classification of Flowering Plants apes . 1952. SARGENT, C. S. Silva of North America 1: 99-102. pls. 43, 44. ———. Manual of the Trees of North America, Ed. 2. 648, a ae DEPARTMENT OF BOTANY NATIONAL TAIWAN UNIVERSITY VoLuME 48 NUMBER 2 i JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY B. G. SCHUBERT EDITOR T. G. HARTLEY C. E. WOOD, JR. LAZELLA SCHWARTEN CIRCULATION APRIL, 1967 Precour! Botanica. Alo MaY 1- 1967 ae LIBRA, BL BY THE ARNOLD ARBORETUM OF HARVARD UNI VER es _ CAMBRIDGE, MASSACHUSETTS THE JOURNAL OF THE ARNOLD ARBORETUM Published quarterly by the Arnold Arboretum of Harvard University. Subscription price $10.00 per year. Volum es I-XX, reprinted, and back issues of ee XXI-XLV are (—] tros No.2 VEGETATION PROFILE FROM LOS NARANJOS VER. TO TEMAZCAL OAX. Savanna of Crescen—|Subdeciduous selva] Selva of Ginoria Oak forest Selva of Terminalia Oak forest} Selva of |Selva of moan ieourn tia cujete | y nudiflora f Que amazonia of a ina og Terminalia with Achra Tavast aa " ee CaS 40g Bl ming i 4 y a | : I 3 as) . 1 a ae a v i a ite, 4 ~~ © Ohag 48 ROO eel 226446 Bs | =o) ‘A sain __ TR SR PRE ore OTT Gob ins = ‘drained teritic soils : i ve : Litipoenes soils 3 ignocol derived from volcanic ash ir Sievial Soils | |= harsh ora. || Altuviat soils | [Lateritic soils || of rendzina S| *ROFILE 1 (ABOVE). This profile includes the most tga primary plant communities in the area. big eee: : oe 3000 mm. per year and the average of monthly temperatures is 24.8° C. (Redrawn from Ma -Pompa, Hernandez P. & Sous 964b.) *ROFILE 2 (BELOW). This profile includes the most important primary plant communities in the area. — rai Aeon is aga 2300 mm. per year and the average of monthly temperatures is about 25° C. (Redrawn a Gémez-Pompa, Hernandez P. & Sousa, 1964b. ) — [L961 ADOTOOU LNVTd TVOIGOUL ‘VdNOd-ZaWOD 110 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 in some cases, strict affinities for certain ecological conditions. For ex- ample, in the region of Papaloapan (Gémez-Pompa ef al., 1964b) we find a marked contrast in the floristic composition of the limestone hills with the rest of the region occupied mainly by lateritic soils of various types. Even in the lateritic soils we find marked contrasts in the floristic composi- tion according to the topographic variations, internal drainage, and hydro- morphism. These contrasts may be drastic or gradual, based on the rela- tionship to the same changes in the local edaphic conditions. A synthetic and diagrammatic profile of a subregion (Tuxtepec-Jacatepec) has been reproduced from the study mentioned above as PROFILE 1. In the Brosimum alicastrum selvas of the limestone hills many species are found which grow only in this type of soil in this area, among them: Mirandaceltis monoica, Mastichodendron capiri var. tempisque, Protium copal, Astronium graveolens, Croton aff. reflexifolius, Sickingia rhodoclada and many more. This fact is confirmed by other observations, made in many parts of the world, on the importance of calcium as an element in the soil which restricts the entrance of calcifugous species and permits the success of calciphilous species (Jefferies & Willis 1964), although the ex- planation is obscure especially for warm humid tropical zones. In the case of the topographic variations within the lateritic soils, how- ever, it should be noted that the floristic differences are less evident among them. Nevertheless, there also exist groups of species with similar patterns of distribution in relation to these edaphic variations. 2. These kinds of affinities or preferences are found not only among “primary” species but also in secondary species (Sousa 1964). This is of great importance inasmuch as in these species there may be found ele- ments of great importance for experimentation on the problems of ? . Another fact of interest is that these local patterns of special dis- tribution occur in shrubby and arboreal species in a way perhaps more notable than in other biological forms (vines, herbs, etc.). It is possible that this fact may have become a little distorted because the major em- phasis of our studies has been on shrubby and tree species. The most surprising observation is that in a great number of cases the most re- stricted of such species in certain soil conditions are the most abundant, frequent, and in some cases dominant ones in the environmental condition in which they are found. This fact has an enormous usefulness since it has permitted the use of such species as indicators for cartographic work on a small scale. It is also notable that if a different region, not very remote, is studied, the whole panorama discussed in the preceding paragraphs may change. It is something of a problem to state exactly why it changes so much because the differences depend upon the area, and vary from region to region. The floristic composition of two regions may be very similar, the ecologically important species may continue to be important but they may “behave” either in a very distinct or in a similar manner. A typical example of this may be seen in the profile of the subregion of Los Naranjos- 1967 | GOMEZ-POMPA, TROPICAL PLANT ECOLOGY 111 Temascal (PRoFILE 2), where, in comparison with that of Tuxtepec- Jacatepec, we meet the same important species, as for example Terminalia amazonia, found in acid soils on the slopes of the limestone hills, as well as on the margins of arroyos, etc. On the other hand, Brosimum alicastrum continues to occupy an important place in limestone soils, so that we may continue analyzing successively the “behavior” of species in the two sub- regions and shall see the whole gamut of change. For example, Curatella americana, a species with affinities for the soils of oak woods of Tuxtepec, Oaxaca, is one of the most important arboreal species of the savannas of Los Naranjos, Veracruz, on very different soils. If the comparison is extended to a more distant region and other cae are used the method is similar. For example, in Jacatepec, Robinsonella mirandae is typical in calcareous soils and is one of the dominants, but in the region of the Tuxtlas, Veracruz, it is also one of the dominants, here a is found in British Honduras forming an important and typical part of the forests of Pinus caribaea on sandy soils (Fic. 2). On the other hand, in this same country Brosimum alicastrum is also found on calcareous soils as in our two profiles. With a basis in these facts, presented in condensed form, we are able Other e sicceal species in these savannas are: Curatella americana, Byrsonima a folia, Coccoloba barbadensis. Bromelia karatas is shown growing at t oe . pls These savannas are found in clay soils in flatland areas use mainly r cattle. a ee ‘ : Ce ‘ Savanna are Croscentie faceie in Vecacres ‘7 see PROFILE 2). 112 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Fic. 2. Between Orange Walk and Belize (British Honduras). Crescentia cujete can be seen at the edge of the savanna, growing adjacent to the Pinus caribaea forest. The palm is Paurotis wrightii. These savannas grow in white san ils. to follow the two tendencies explained above, and the decision to follow one or the other will be related to the information used. 1. If considerations are based on the study of only one region or a subregion it is probable that the presence of more or less well defined patterns of distribution which fall within the strictest limits of the defini- tions of “association” can be proven. may present very marked differences in their responses to the diverse AN HYPOTHESIS TO EXPLAIN THE FACTS I believe that in the study of tropical vegetation it is necessary to have more hypotheses to work with, because accumulations of facts as such, 1967 | GOMEZ-POMPA, TROPICAL PLANT ECOLOGY 113 have a limited importance if we do not have some links between them to form some generalized ideas for a better understanding of tropical environ- ments. When a species is cited from a region as, for example, Terminalia amazonia in Chontalpa, Tabasco (Fic. 3), we are dealing, in fact, with a small fragment of the total population of this species, or rather we are prise the species whose total range of distribution covers an area from Brazil to Mexico. This local population is very small considering not only its total distribution, but also the fact that it is a tree species. The gene pool of this population may be very different from that of another popula- tion in Panama or in Brazil. The differences between these populations are due to many factors which basically have the effect of discontinuities within the total population of the species. This phenomenon is very well known and has had a strong influence on taxonomic studies although some taxonomists ignore the concept of the Mayr (1963) defines the species as “groups of actually or potentially interbreeding natural populations which are reproductively isolated from other such groups.” In consideration of this idea and others having a xe Fic. 3. Terminalia amazonia selva with its typical stratified treetops, in the state of Tabasco. This selva grows on plains or small hills, on deep, well drained, lateritic soils. 114 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 similar focus (Dobzhanski 1941, Huxley 1940), authors tend to call it a biological species concept, and all agree that the species is commonly Ss i of systems of populations, a theory which is of much interest tou Of the gone explanations of discontinuities those of most interest to me are the ones which enable me to establish an explanation for the ecological een of certain species to local regions. The most im- portant of those which I have considered may be explained by one example. On many occasions I have had the opportunity to see isolated individuals of tree species which have survived the effects of disturbance in different regions. In the majority of cases these individuals have heen undisturbed because of existing shade or because they do not infringe on agricultural or grazing activity (Fic. 4), one of the principal causes of the destruction of the original vegetation. In other situations there are survivors of fires and other disturbances, both artificial and natural (Fic. 5). If such a disturbed zone is abandoned and left to recuperate its vegetation through a secondary succession, it will be noticed that the majority of the plantules of trees which develop will surely come from the mature tree which has persisted. If, in the model we $re considering, one (or few) of the trees left belong to the species Terminalia amazonia, the future small local population will come from a single tree (in the case of a self-compatible species) or from a few individuals (in the case of self-incompatible species). As a result there is a genetic change which may or may not be morpholog- ically evident depending on the genotype wf the parent tree. This phe- Fic. 4 (Lert). Isolated tree of Terminalia amazonia in a valley near Cordoba, Veracruz. This tree has survived in spite of human disturbance. At the bottom can be seen some selvas ctning © oo ee. This species is not found in such areas as it grows chiefly on acid s Fic. 5 (r1GHT). Isolated tree ns penne amazonia near Macal river in Brit- ish Honduras. Shis tree has survived a cyclone that threw down most of the trees in the area 1967 | GOMEZ-POMPA, TROPICAL PLANT ECOLOGY 115 nomenon may be included in some of the distinct types of “genetic drift.” This change may be very abrupt depending on the genetic constitution of the parent tree from which a local gene complex will be derived. It abling it to compete successfully. An example of this mechanism may possibly be seen in the oak scrub of Quercus oleoides (Fic? 6) in the Llanos de Alvarado, Veracruz, a population with an aspect very distinct from the encinares of Q. oleoides of the neighboring regions (Gomez- Pompa, 1965). These populations resemble the “founder populations” of Mayr (1963) of which he says, “this term (Founder Principle) designates the establishment of a new population by a few original founders which carry only a small fraction of the total genetic variation of the parental population. The descendant population contains only the relatively few genes that the founders have brought with them until replenished by sub- sequent mutation or by immigration,” to which I might add -(in the case of plants) that these populations may continue a certain level of gene flow with other populations of the same species, adjacent or distant. We do not know with certainty that this has happened in the past or that it is happening now; but we know that the majority of so-called virgin forests show evidence of human activity in the past, or at least evidence of fires, inundations, hurricanes, and other catastrophic events which, we may assume, have left some individuals as the basis of future regeneration of the vegetation. Clausen (1962), in reference to such a situation says, “Each local population of a species has probably been started by a limited number of individuals, and it may be effectively isolated from other colonies by spatial isolation.” Fic. : Oak ah of Fula dees near Al as Veracruz, This popula- tion grows on flat, clay soils, near the coast of the Gulf of Mexicc 116 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Another possible explanation of a similar phenomenon (with a different origin, however) is in the consideration of an area denuded by a catastrophe of nature such as volcanic activity. The material with which the vegetation is reconstructed will be provided in large part by the margins of surround- ing species populations. Of all the propagules only those will survive which can compete in and are adapted to the new environment. This could result in a possible change in the new populations. This genetic drift will be more obvious if the surrounding zone presents different edaphic char- acteristics such as limestone soils, or even different climatic ones. This type of situation may be imagined to have occurred in many zones of the lowlands of the Gulf of Mexico. As the basis for such a mechanism a quotation from Grant (1963, p. 286) seems pertinent: “In other cases the natural population may be large but has descended from a few migrant individuals. This is often the case in organisms which colonize new areas. A volcano may destroy all traces of pre-existing life on a mountain ... ; a fire may sweep through a forest; or a tree may simply die of old age, exposing a spot of bare ground. The new ground becomes colonized by migrant plants and animals from surrounding areas. It is not likely that colonizing individuals can bring with them a complete sample of the genetic diversity in the old ancestral population. If the founders of the new population are few in number they could, by chance, represent only one or a few of the genetic variants in the ancestral population from which they came.” These ideas may well explain in part the pronounced response of some shrubby and tree species to certain regional ecological conditions. This is so especially for tropical zones in which the following characteristics are present: (1) that low winter temperature is not the most important limit- ing ecological factor with which the species has to contend and adapt, as is the case in extra-tropical regions; (2) that natural selection works in these local regions fundamentally in relation to edaphic differences, or rather that the soil plays a very important part in the selection of such individuals in these regions. This fact has been mentioned for different regions in the tropics (Van Steenis 1956). In relation to the soil factor Mason (1946), from his experience in California, says “Of the various categories of environmental factors, the condition of any factor or combina- tion of factors may serve to restrict the range of some species of plants. Of those factors however, the edaphic factor i is most apt to occur in sharply defined patterns and often i in small areas. With this hypothesis as a basis, the problem of the distinct behavior of some species from region to region may then be explained. As noted previously, the possible origin of populations may be different if the parent sources have different genetic constitution. This is more out- standing if we note also that no two areas have the same environmental conditions. We may find environments which are very similar, but into which new factors have entered to produce very strong variation. The “environment” concept is very complex; nevertheless, in relation 1967 | GOMEZ-POMPA, TROPICAL PLANT ECOLOGY 117 to the behavior of plant species in the Mexican tropics, we can make the following artificial classification. 1. Highly selective environmental factors. These factors are rela- tively independent, such as soils derived from limestone, with different texture and structure; or, poorly drained savanna soils. Several species follow these soils in areas with differing climates and topography. Such is the case of: Brosimum alicastrum, Protium copal, Mirandaceltis monoica, Mastichodendron capiri var. tempisque, Trophis racemosa, Manilkara zapo- tilla, etc., for the limestone areas in many regions of the lowland tropics in Mexico; and of Curatella americana, Crescentia cujete, Byrsonima crassifolia of many savannas of tropical America, Finally, the most ex- treme type of selective environment is probably the mangrove swamp with a typical floristic composition in many different areas. 2. Poorly selective environmental factors. These factors do not imply a great selectivity of species from region to region. The behavior of the species is very variable and seems to depend on variation of in- dependent local factors, such is the case of many different deep, well drained, acid soils. In these soils the following selvas will serve as an example: selvas of Terminalia amazonia, Dialium guianense, Pseudolmedia oxyphyllaria, Vatairea lundellii, Bernoullia flammea Since these are not very selective, the floristic composition and the rela- tive abundance of a single species may vary remarkably from one region to another in relation to still other factors, the most interesting probably being the historical factor. The historical factor may provide the explana- tion of the origin of the local flora in time and space. Some of these acid soils bear forest which is commonly known as “mixed” rain forest and which seems to have no fixed pattern of species composition, but rather a random dispersal of species. The nutrient mosaics in some of these local areas must play a very important role in such “random” dispersal of species but unfortunately, little is known about these mosaics in the tropics of the New World. 3. Intermediate factors. In this category are factors which may be highly selective locally but which are not “followed” by species from region to region, for example, recent alluvial soils. In such soils, characteristic species may be found, but they may differ depending on the region. In this category also two other factors may be included, topography and slope ex- posure which provoke responses from the species which can be explained only locally. In some cases the variation is very gradual and the responses of the species are equally gradual forming a continuum. On the basis of this hypothesis we may cite an enormous number of problems needing study which will allow us to open new lines of investiga- tion for the development of plant ecology in the tropics. Among them are: a). The problem of the experimental initiation of the secondary suc- cessions, which will permit us to know something of the modern historical factors and of the behavior of the individuals in these stages. Work of this type has already begun in Mexico (Sarukhan, 1964). b). The detailed study of the behavior of specific local populations. 118 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 c). Work on transplants between populations of the same species com- ing from different regions, and the initiation of studies of ecotypic differ- entiation, as well as transplant experiments of species to different environ- mental conditions from those under which they live. These studies should provide fundamental information for the proposed hypothesis. Also, this type of study could provide the key for resolving the problem of tropical silviculture. d). The study of breeding systems of the ecologically important species from which may be obtained information about some of the most inter- esting species. These studies must be complemented, however, with cytogenetical studies permitting extraction of information basic to funda- mental problems in ecology and in tropical forest genetics. Such an approach to the study of tropical plant ecology will remove it from the isolation to which it has been subjected by other sciences, since its problems are related to similar problems in other fields as evolution, taxonomy, population genetics, and so forth. As Mayr (1947) noted “At first sight the concept of geographically and ecologically variable species seems full of contradictions. We see that in one locality a species is re- stricted to a very definite habitat niche, while in another locality it occurs in a different, sometimes very different niche. We know that through selec- tion in each of these populations a definite gene complex has developed, which permits the population to survive and thrive in spite of competition, predation and all sorts of other adversities.” All the information obtained by the development of genecology, sup- ported by the work of Turesson (1923, 1936), Clausen, Keck, and Hiesey (1940), and more recently by McMillan (1964), Kruckeberg (1951), and Critchfield (1957), is related to such an approach and the impact of the results becomes increasingly stronger because of the same basis of the ii. which in the revision of this field by Bennett (1964) is defined as . the study of the genetic mechanisms which operate, within organ- isms, and between organisms and their environment, at the level both of the individual and of the population, viewed as a process, which at the — level produce those changes known collectively as micro- evolutio I think that these studies have to be initiated for tropical plants if we want to get to the basis of the problems. Surprisingly, the definition of genecology seems to be closely related to one we proposed for autecology, in which it is limited to “the study and definition of those environmental factors affecting the survival, adapta- tion, and fertility of the individuals of one species.”’ Since the factors can be defined in broad zones only with difficulty, and since they vary from region to region, we might say that autecology deals with the study of those factors in the macro and microenvironment (as well as the plants) affecting the survival, adaptation and the dynamics of reproduction, of those in- dividuals belonging to one or several regional specific populations. I think it is important to make the comparison between the field of autecology 1967 | GOMEZ-POMPA, TROPICAL PLANT ECOLOGY 119 and genecology because at certain levels there is no great difference, only perhaps a difference of emphasis, as long as both are studied in detail. All the studies related to these fields are the true basis of synecology. Since in tropical areas such studies scarcely exist, any serious, purely theoretical controversy on tropical vegetation is without basis and must, therefore, await the collection, organization and study of facts and data. There is a tremendous task in front of us because of our lack of knowl- edge of tropical plants. We need more and better taxonomic studies of tropical plants, we need also more biological information of them. It is my belief that ecological studies can make a very important contribution to taxonomy if vouchers are mentioned and widely distributed to herbaria. There is great potentiality in the combination of the taxonomy and ecology of tropical plants; there is much to be learned about speciation problems in the tropics and it is my hope that botanists of tropical countries can contribute significantly in this respect. SUMMARY 1. Most of the studies in tropical plant ecology have been carried on using methods and ideas developed in temperate regions. 2. Controversies have arisen because of very poor field work in the tropics, and because of lack of extensive surveys to prove or disprove some of the ideas 3. The two apparently opposite ideas, of the existence or absence of clear cut associations in the warm humid tropics, are used to illustrate the confusion existing in tropical plant ecology. 4. An hypothesis to conciliate the two positions in accordance with the facts obtained in studies done in the Mexican tropical lowlands is presented, based on the following factors: a) the edaphic mosaics as a basis of natu- ral selection of local species populations; and b) on the fact that there are not two areas with identical environments; 9) on the characteristics of the origin and “behavior” of plant populations in local environments; d) on the natural selection of species populations in local environments as an explanation of the striking patterns of species in these local areas. 5. It is suggested that new methods are much needed in the study of tropical plant ecology and that the evolutionary genecological approach will make a significant contribution to the understanding of the phenomena responsible for what we are finding in vegetation studies in the tropics. BIBLIOGRAPHY Bearp, J. S. 1955. The classification of tropical American vegetation types. Ecology 36: 89-100. Beckine, R. W. er The Ziirich-Montpellier school of phytosociology. Bot. Rev. 23: 411-4 BENNETT, E. 1964, ie perspectives in genecology. Scottish Plant Breed- ing Station Record 1964: 49-115. 120 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Buarucwa, F. R. 1958. oe for the study of tropical vegetation. Proc. Kandy Symposium. 89-92. ESCO. Carn, S. A. 1947. a a natural areas and factors in their develop- ment. Ecol. Monogr. 17: 185-200 CHATTERJEE, D. 1958. Tropical vegetation of eastern India. Proc. Kandy Symposium. 61-67. UNESCO. CLAUSEN, J. 1962. Stages in the evolution of plant species. Hafner Publ. Co. . D. Keck & W. M. Hirsey. 1940. Experimental studies on the nature of species. I: Effect of varied eaivivomnients on western North American plants. Carnegie Inst. Publ. 520: 1-452; II: Plant evolution through am- seine and autoploidy with examples from the Madiinae. Jbid. 564: {pana F. E. 1936. Nature and structure of the climax. Jour. Ecol. 24: 252-284. ape rae W. B. 1957. Geographical variation in Pinus contorta. Maria rs Cabot Foundation Publ. 3. Cesmascnsis J. 1934. Observaciones geobotanicas en Colombia. Mus. Nac. Nat. (Madrid), Trab. Ser. Bot. 27: 1-144. pe re & A. J. G. H. Kostermans. 1958. Research on the vegetation of Indonesia. Proc. Kandy Symposium, 28-32. DoszHANSkKy, TH. 1941. Genetics and the origin of species. Columbia Univ. ress. GL eEason, H. A. 1939. 1 individualistic concept of the plant association. Am. Midl. Nat. 21: 92-10 & M. I. Coox. ie The plant ecology of Porto Rico. Sci. Surv. of Porto Rico and the Virgin Isl. GOMEz- srecigs A. 1965. La ‘vegetacién de México. Bol. Soc. Bot. México 29: 120. , J. VAzquez Soto & J. SARUKHAN K. 1964a. Estudios ecolégicos en las zonas calido hamedas de México. Publ. Esp. Inst. Nac. Inv. For. México 3: 1-36. —, L. HernAnpez P. & M. Sousa S. 1964b. Estudio fitoecolégico de la cuenca intermedia del Rio Papaloapan. Publ. Esp. Inst. Nac. Inv. For. México 3: 37-90. Grant, V. 1963. The origin of adaptations. Columbia Univ. Pre HEVLICERS, P. C. 1963. Vegetation and soil of white-sand savanna in Surinam. Meded. Bot. Mus. Utrecht 191: 1-148. Huxtey, J. S. igo Pay new systematics. Clarendon Pres Jerrertes, R. L., & A. J. Witits. 1964. Studies on the sal olen letune habit. . Methods of gar of soil and plant tissues and some results of investiga- tions on four species. Jour. Ecol. 52: 121-138. KrucKeEBErRG, A. R. 1951. Intraspecific variability in the Anges of certain native plant species to serpentine soil. Am. Jour. Bot. 38: 4 19; MANGENOT, G. 1950. ia sur les foréts denses de la Céte cine Bull. Soc. Bot. France 97: 159-16 . Les recher si sur la végetation dans les régions tropicales humides de l’Afrique Occidentale. Proc. Kandy Symposium. 115-1 126. CO. S Mason, H. L. 1946. The edaphic een in narrow endemism. I. The nature of environmental influences. Madro : Mayr, - 1947. Ecological factors in nate Evolution 1: egies —., 1963. Animal species and evolution. Harvard Univ. Pre 1967 | GOMEZ-POMPA, TROPICAL PLANT ECOLOGY 121 pees C. 1964. Survival of transplanted Cupressus and sa ie thirteen years in Mendocino County, California. Madrofio 17: 250- ree i‘ F. V. 1934-5. Succession, development, the climax, it the complex ism: An analysis of concepts. Parts I-III. Jour. Ecol. 22: 554-571; pa "210-246, 488-508. Ravp, H. M. 1942. Trends in the development of geographic botany. Ann. Assoc. American Geographers 32: 319-354. Rosayro, R. A. DE. 1958. Tropical ecological studies in Ceylon. Proc. Kandy Symposium. 33-39. UNESCO. SARUKHAN, K., J. 1964. Estudio sucesional de un area Talada en Tuxtepec, Oaxaca. Publ. Esp. Inst. Nac. Inv. For. México 3: 107-172. Sousa, S., M. 1964. Estudio de la vegetacién secundaria en la regidn de Tuxtepec, Oaracn. Publ. Esp. Inst. Nac. Inv. For. México 3: 91-105. emote C. G. G. J. van. 1956. Basic principles of rain forest sociology. Proc. Symposium. 159-165. UNESCO. Gesciecder G. 1923. The scope and import of genecology. Hereditas 4: 171-176. 1936. Rassenokologie und Pflanzengeographie. Bot. Not. 3-4: 420-437. WHITTAKER, R. H, 1962. Classification of natural communities. Bot. Rev. 28: 1-239 INSTITUTO DE BrIoLocia UnrIversipAp NACIONAL AUTONOMA DE MExIco CruDAD UNIVERSITARIA Mexico 20, D. F., MExico 122 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ANATOMY OF THE PALM RHAPIS EXCELSA, IV. VASCULAR DEVELOPMENT IN APEX OF VEGETATIVE AERIAL AXIS AND RHIZOME! MArtTIN H. ZIMMERMANN AND P. B. TOMLINSON PREVIOUS ARTICLES IN THIS SERIES have analyzed quantitatively the course of vascular bundles in the vegetative aérial stem and in the rhizome of Rhapis excelsa (Zimmermann & Tomlinson, 1965; Tomlinson & Zim- mermann, 1966a). The present article extends our descriptive analysis of Rhapis to the apex of both aérial and underground axes, tracing vascular strands, in their earliest recognizable condition as procambial strands.” From this we can make certain deductions about the way in which the vascular system develops. No attempt is made to describe the development of phloem and xylem tissues within the procambial strands, this important aspect being reserved for a future article. As Esau (1965, p. 33) has pointed out, information on leaf trace rela- tionships in monocotyledons, especially the perennial types, is incomplete; the apparent complexity of these plants has inhibited their investigation. Nevertheless our analysis of Rhapis has shown that the basic plan is quite simple, only the overall large number of bundles obscures this. Despite the rather artificial attempts by earlier anatomists to classify monocotyledonous vascular systems (e.g., by Falkenberg, 1876), only two main groups need be considered (Priestley & Scott, 1937), namely those with a nodal plexus of vascular tissue, exemplified by the grasses, and those without nodal plexi, exemplified by the palms. The former have been studied quite extensively (Kumazawa, 1961; Scott and Priestley. 1925; Sharman, 1942). However, in the non-nodal type of monocotyledons the only study which continuously relates development of the vascular strands to their distribution in the mature stem is that by Priestley et al. (1935) in Alstroemeria, an analysis which is worthy of wide attention. A notable observation of these workers is that vascular bundles are recognizable as acropetally growing strands in the developing stem long before the leaf which they ultimately supply is produced. Alstroemeria is, however, a relatively diminutive monocotyledon and its vascular system is much less elaborate than that of even the smallest palm. Because of the greater bulk and complexity of palms, workers have been content to describe development in general terms. The earliest literature * A further contribution towards a continuing study of oat anatomy of the palm stem by one of us (P.B.T.) supported by N.S.F. Grant GB 2 * The terminology used in this paper, as well as in the bine of this series, is a purely descriptive anatomical one (see the statement in Zimmermann & Tomlinson, 1965, p. 167 top). 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 123 discusses mostly the hypothetical “endogenous” growth of vascular bundles and is now largely of historical significance. Branner (1884) summarizes this history in his introductory pages. Branner’s own factual contribution is noteworthy because he grasped the principle governing the distribution of vascular bundles in the mature palm stem, but he obscured it in a developmental discussion which seems not to be based on observation. Falkenberg’s (1876) developmental account is also theoretical. His il- lustration of the palm is at best incomplete and, as we have shown, has been more a source of error than information for later writers (Tomlinson & Zimmermann, 1966b). Baranetzky’s (1897) account of vascular bundle development in monocotyledons describes simply a centrifugal or cen- tripetal, sometimes mixed centrifugal and centripetal, direction of develop- ment but adds nothing to our knowledge of the longitudinal course of differentiation. More recently Ball (1941) has verified the “centrifugal” development of vascular bundles recorded by Baranetzky in palms. Ball and other writers on the shoot apex of palms (Chouard, 1936; Eckardt, 1941; Helm, 1936) have simply described and illustrated the general topography of the meristematic regions in palms, showing that thickening growth of the stem is not a direct result of the shoot apical meristem proper but is associated with a primary thickening meristem developed below the expanding leaf primordia.? Otherwise no attempt seems to have been made to follow the distribution of leaf traces in the meristematic crown, although it is clear that an understanding of development is impossible without knowledge of the course of vascular bundles. In Rhapis, as in other small palms, activity of the primary thickening meristem is such that the overall outline of the meristematic crown is a shallow cone rather than a bowl. This is fortunate, because analysis is facilitated when vascular bundles can be followed continuously in one direction through serial sections. The crown is also of a size suitable for handling by orthodox microtechniques. In addition, our previous quantita- tive analysis of the mature R/apis stem has provided the information about vascular organization which, as Esau (1965) emphasizes, is an essential prerequisite for developmental understanding. For convenience this vas- cular organization in mature stems is very briefly outlined below, although the reader is referred to the first two papers of this series for details (Zimmermann & Tomlinson, 1965; Tomlinson & Zimmermann, 1966a). Familiarity with the facts in these two articles will greatly facilitate the understanding of the present paper. In Rhapis all bundles of the central cylinder behave essentially alike. Each axially running vascular bundle is linked, at regular intervals, to leaves via a branch, the leaf trace. We have called the distance between leaf contacts of a given bundle the “leaf-contact distance” or “‘leaf-contact interval.” This distance can be measured in number of internodes (cf. Zimmermann & Tomlinson, 1965, p. 169, Fig. 3). The departing leaf trace * Distinction must be made between the shoot apex proper, which in palms produces only leaf primordia, and the meristematic region 0 the whole apex. Subsequently, the former will be referred to as “apex”, the latter as the “crown.” 124 JOURNAL OF THE ARNOLD ARBORETUM [voL, 48 shows the familiar outward curve originally recognized by von Mohl (1824) as characteristic for all palms. Each leaf is supplied by relatively few major bundles, but by a larger number of intermediate and minor bundles. Major leaf traces originate from the more central bundles so their outward curvature is most pronounced. Major bundles also give off leaf traces at the longest intervals (about 15 internodes). Intermediate and minor bundles are progressively more restricted to the periphery of the central cylinder and consequently the outward curvature of their leaf trace branches is less pronounced. In addition the leaf-contact distance is progressively shorter for bundles farther towards the periphery. Each continuing bundle (vertical bundle) above the level of departure of the leaf trace initially follows the outward curve of the leaf trace to the crowded periphery of the central cylinder, whereupon it turns erect and then gradually approaches the stem center again to give off a further leaf trace at some higher level (cf. Zimmermann & Tomlinson, 1965, p. 169, Fig. 3). During this gradual inward movement the bundle describes a shallow helix in the direction of the phyllotactic spiral. In the rhizome bunZles deviate less from the axial direction and the helical path is scarce- ly evident (cf. Tomlinson & Zimmermann, 1966a, p. 254, Fig. 4 below). In addition, there is greater irregularity of leaf-contact distance and bundle continuity. In both axes lateral continuity between different bundles is achieved by short bridges between outgoing leaf traces and nearby vertical bundles. Cortical bundles occur in both types of axes as anastomosing strands con- tinuous from the leaves, independent of the vascular system in the central cylinder, and ultimately ending blindly below. PREPARATION OF MATERIAL Vegetative stem and rhizome apices of Rhapis excelsa, together with enveloping leaf bases, were isolated from plants cultivated at Fairchild Tropical Garden and fixed in FAA. During dissection a record was kept of the state of development of the youngest exposed leaves in such a way that the corresponding leaves could be identified in serial sections. Stem apices and surrounding leaves thus prepared were cylindrical blocks of the order of 1-114 «& 2-3 cm. Blocks were desilicified in hydrofluoric acid, dehydrated and embedded in paraffin wax by routine procedures and sectioned serially at thicknesses from 8 to 15 ». Shrinkage of material was inevitable but did not interfere with subsequent analysis. Sections were stained either in safranin and Delafield’s haematoxylin on the one hand, or in erythrosin and toluidin blue on the other. Most observations were made on serial transverse sections but longitudinal series were also used for comparison. The bottom slides of the series of sections show all the anatomical features of the mature axis except that tissues are immature, e.g., metaxylem elements and fibers have not yet developed secondary walls. 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 125 GENERAL TOPOGRAPHY OF THE CROWN For the sake of clarity we shall now introduce the reader to the general topography of the crown. This actually represents part of the results of the present study; but by giving this short initial description it will be easier for the reader to understand the next chapter of the paper describing the method of serial-section analysis wherein we break newer ground. Aérial stem (Fics. 1, 2, 6, and 7). We have not investigated the shoot apical meristem proper. Reference may be made to Ball’s (1941) work on this subject in those species he examined. In palms this apical meristem functions largely as a “leaf-initiating” meristem. In Rhapis it includes no procambial tissue, i.e., no vascular strands can be discerned above the youngest primordium. The overall outline of the crown of Rhapis is a shallow cone, without the bowl-shaped depression of larger palms. This difference between smaller and larger palms is related to the activity of the primary thickening meristem and is not a fundamental one. Leaves in Rhapis are arranged in a 2/5 phyllotactic spiral (Fic. 1). Each leaf has a closed tubular base tightly enclosing younger primordia, but somewhat thickened on the dorsal side. This dorsiventral difference in thickening is more pronounced distally in each leaf and is most easily seen in successively younger leaves cut at the same level, as towards the center of Fic. 1. Fic. 1 is drawn from a section cut too high to show the circular attachment of primordia 2—5. This leaf-base topography is the result of leaf growth by a pronounced basal meristem after the establish- ment of an encircling attachment. Development of this encircling attach- ment in the young primordia must be rapid because in the sections ex- amined it is only incomplete in the youngest primordium (P1). Procambial strands are evident even in the youngest primordium in- cluded in Fic. 1, their further rapid increase is indicated in the counts for total numbers of strands in successively older leaves plotted in Fic. 4. Immediately below the apical meristem proper, and continuous with it, is a dome-shaped or umbrella-like mass of meristematic cells. This un- differentiated meristematic “cap” (subsequently referred to simply as the cap) is pierced by leaf traces entering the youngest primordia. The cap is more or less easily visible on single sections of the rhizome crown, but it is much less easily recognizable in the crown of the aérial stem, except if one runs rapidly through transverse section series with the motion-picture analyzer. Towards the lower border of the cap individual procambial strands are distinguishable from vacuolating cells. These strands are the distal ex- tremities of the vertical bundles which, when traced upwards from below, end blindly in the cap. This cap-like meristem therefore is continuous with the crowded periphery of the central cylinder of the mature stem. The more central stem tissue, immediately below the shoot apex and en- closed by the cap, contains procambial strands already linked to leaves via the leaf traces which penetrate the cap. 126 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 FIRST FULLY EXPANDED LEAF TIP JUST (i) EXPOSED ®) ENCLOSED 2¢ UNEXPANDED LEAF 21M OF FOLDED “SPIKE” EXPOSED _ Fis. 1. Rhapis excelsa, Transverse section through the crown of the vegeta- tive aérial axis. Leaf primordia are num mbe red from within, number 1 (P1) is the : h major leaf trace from P3 (note the ‘sharp turn below the apex). The numbers along the spiral indicate millimeters below the apex. InseT: central portion of section at higher magnification. 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 127 It must be emphasized at this point that the cap is not identical with the primary thickening meristem. The primary thickening meristem must be located under (i.e. inside) the cap. In neither of the crowns can the primary thickening meristem be structurally recognized in single sections because cell divisions are not uniform in any one direction. Towards the base of the crown, however, at the level of insertion of recently expanded leaves, ground meristem cells divide in a predominantly transverse plane, producing the vertical files of cells which indicate uniform elongation of the axis without distinction between nodal and internodal regions (Fic. 7). Procambial strands in longitudinal section are strikingly distinct from the surrounding ground tissue as fascicles of cells with coincident end walls. The nuclei of each tier of cells at the same level produces a “tiger- tail” effect in longitudinal view. This indicates the origin of procambial strands from undifferentiated meristem by numerous longitudinal divisions in series of meristematic cells, the common end wall of each tier indicating the limits of each parent meristematic cell. This coincidence of end walls persists in the vascular tissues of mature vascular bundles (Tomlinson, 1961, p. 58). In transverse section procambial strands are less conspicuous, except where the sections include the nuclei. Otherwise they are distin- guished by their narrow cells with somewhat densely staining cytoplasm. Limits of procambial strands are not sharp in the more meristematic regions where they become progressively more difficult to recognize. Rhizome (Fics. 3, 8, and 9). The apical region of the rhizome (Fic. 8) is similar to that of the aérial stem in all essential details, except for being smaller (cf. Fics. 2 and 3). The 2/5 phyllotactic spiral is less obvious because the dorsiventral asymmetry of the leaves is less marked. The meristematic cap at the lower edge of which the vertical bundles terminate is more clearly visible in the rhizome than in the aérial stem (F 16. 9). Helical path of bundles. The helical path of the procambial strands in the uncrowded center of the crown is not evident in any single section. Only a short length of any one bundle is included in single sections because of the conical shape of the crown. However, the spiral path can be demon- Strated by plots of bundles. As an example, a major trace of P3 is given in Fic. 1. The trace is projected vertically into a transverse plane, the numbers along its path representing the vertical distance below the apex in millimeters. In flat or bowl-shaped crowns of larger palms the overall spiral path of a number of bundles can be included within a single section. This is shown, for example, in Chrysalidocarpus lutescens (Fic. 5). This demonstration of the helix as an inherent feature of the crown in palms is historically Significant because it finally settles an old controversy. Meneghini ( 1836) suggested that the helix is a mechanical consequence of torsions set up in the palm stem as it develops. This suggestion was debated at some length by early authors such as von Mohl (1849) but without examination of developing palm stems. Clearly the initiation of vascular bundles along a Spiral pathway requires a physiological explanation, not a mechanical one. 128 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 METHOD OF SERIAL-SECTION ANALYSIS This chapter describes how the serial transverse sections were analyzed. and how Fics. 2 and 3 were constructed. These two figures contain most of the information obtained during the = study; they form the basis for subsequent discussion of developme Two aérial and one rhizome crown were analyzed in full quantitative detail. Both aérial crowns showed an almost identical course of vascular strands, they will therefore not be discussed separately Initially the sections were surveyed so that the general topography of the crown could be established. Diameter measurements of the axis at each leaf insertion (in its dorsiventral plane) were plotted against section num- ber. The resulting diagram shows a radial longitudinal section of the apex, but with all the leaves in the same plane (Fics. 2 and 3). The vertical scale of these figures indicates the axial dimensions (i.e., the thick- ness of the sections when stacked up) and the horizontal scale is the radial distance of the plotted strands from the stem center. Measurements of peripheral strands were made from the strand to the stem periphery, more central strands were measured from the strand to the stem center, which was established by eye and marked with a spot of India ink on the cover glass. This visual determination of the center is of ample accuracy be- cause the transverse sections are far from exactly circular. It did mean that plotted positions do not fall precisely on the smooth curves drawn in Fics. 2 and 3; to this extent these plots are made diagrammatic. Of course, bundles do run along a more or less smooth path; if plotted curves come out a little irregular it is because there is no fixed point of reference in the axis. As a given vascular strand was followed, its distance from the stem center was measured about once every 300 w and its position was then same strand was achieved is described below. The largest major bundle from each leaf was followed in turn, from the leaf insertion to the base of the crown (the lowest section in the series). Plotting started with the oldest available leaf (P14) and continued with each progressively younger leaf. Each bundle having been followed in the downward direction as a leaf trace, plotting was then continued in the reverse direction but now along its diverging vertical bundle, as far towards the shoot apex as the strand could be discerned. For successively younger leaves this operation became in- creasingly difficult because their traces are correspondingly less well dif- ferentiated. Fics. 2 and 3 show, therefore (as solid lines), the course of the largest major leaf trace which enters the axis from the dorsal side of the leaf. A minor and a cortical bundle from P11 are also indicated in Fic. 2. It must be emphasized again that these figures are not drawings of actual longi- tudinal sections because all dorsal sides of the leaves have been rotated into a single radial plane. Furthermore, the spiral path of the strands is entirely ignored. This spiral path is shown for a major bundle from P3 in Fic. 1, °o X3dv MOT3E SY3SLIWITIIW 1+ 2- 3- 4- rm ¥ T YT TY T T 0 | | 4 5 RADIAL DISTANCE FROM SHOOT AXIS IN MILLIMETERS Fic. 2. Rhapis excelsa. Diagrammatic radial po ane section through the crown of the aérial vegetative axis based o measurements as described in the text. This is not a drawi ng bee a single longitudinal section. All dorsal sides of leaf tions are ed into a single radial plane so that major leat traces can be compared. The spiral path of the bundles is ignored. Leaf fcmuberink @ in ad 1. All solid lines are major bundles. A minor and a cortical (fibrous) bundle of P11 are included for com- parison (dotted lines). Broken line separates cortex from central cylinder. Traces to P2, P4, P6, P8, and P10 are omitted for the sake of clarity. [L961 VSTAOXA SIAVHU ‘NOSNITNOL ¥ NNVNUANWIZ 130 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 projected vertically onto a tranverse plane. The same trace is shown in Fic. 2. The reliability of these plots depends on the certainty with which a single bundle can be followed. Anyone who carefully looks at the photo- micrographs in this article will realize that it is not easy to follow an individual bundle throughout a series of sections, especially if he considers that the stem contains about a thousand central vascular, and another two thousand five hundred cortical bundles. Therefore, during the plotting a motion picture was prepared with the drawing method outlined earlier (Zimmermann & Tomlinson, 1965). This procedure provided the physical discipline needed to ensure that continuity of a vascular strand was not lost. The problem of loss of continuity became increasingly severe in the vascular strands associated with the youngest primordia. Here the pro- cambial strands are narrow, poorly differentiated and make sharp turns near the center of the stem. However, it was found that in critical areas (the top half millimeter), by making two superimposed drawings in con- trasting colors at two magnifications for each section, the procambial strands could be followed with certainty. The drawing at low magnifica- tion served for quick orientation, the one at high magnification for iden- tification of the strand under observation. The most important data thus obtained are the resulting plots (Fics. 2 and 3). Nevertheless, the motion pictures were quite useful to verify con- tinuity and to allow study of incidental anatomical features. Moreover, it demonstrates in a few minutes the discoveries resulting from several weeks of hard work. In the motion picture of the rhizome crown, for example, the cap can be seen as a closing diaphragm which “sweeps away” all vertical bundles; a dynamic demonstration of the way vertical bundles “fuse” into the cap. COURSE OF VASCULAR BUNDLES IN THE CROWNS Ficures 2 and 3 show the course of vascular traces in the top 3 to 4 millimeters of the crown of the aérial axis and the rhizome respectively. They are all dorsal major traces except those clearly labelled in the figure. These two figures show a very complex three-dimensional system reduced to a single radial plane, as described in the previous chapter. In order to reconstruct the three-dimensional system, the reader has to go through the following mental exercise. All points of entry of dorsal major bundles from leaf base to stem have to be rotated around the stem axis back to their five respective radial plans so that the 2/5 phyllotactic requirement is fullfilled (i.e., Pl, P6, P11, etc. belong to the same orthostichy). Then, looking along the stem axis, one has to twist the five radial planes into the spiral shape shown for the trace to P3 in Fic. 1. A comparison of Fics. 2 and 3 with the diagrams of the adult axes, shown in previous papers (cf. Tomlinson & Zimmermann, 1966a, Fig. 4, below) shows clearly that the vascular pattern of the adult axes is laid down in the crown before primary thickening growth occurs. In the crown, leaf traces run axially and vertical bundles run radially. Subsequent 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 131 APEX 2 X3dV MO1ZG SY3SLIWITIIW 0 i 2 as MILLIMETERS FROM CENTER dles. Broken d P9 are omitted 132 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 primary thickening growth reorientates the system through 90°, with the result that in the mature stem leaf traces run more or less radially, vertical bundles more or less axially. Let us now look a little more closely at the aérial stem and examine the major trace system of each successively younger leaf (Fic. 2). Vertical bundles diverging from the major leaf traces to P14 are released somewhat below the bottom of the section series, but are still readily identified be- cause they initially follow the leaf trace closely in its centrifugal path. The major traces to P13 and P12 are complete, but located proportionately higher. No vertical bundle could be recognized diverging from the traces to P11 or any younger primordium. Corresponding to the overall outline of the crown, all vertical bundles of the crowded peripheral region are approximately horizontal at the insertion of P13. In this region they are conspicuous as radially running procambial strands (“tiger tails’) in a single transverse section (Fic. 6). About 1/4 millimeter below the apex, approximately at the level of insertion of P8 they all fuse into an undif- ferentiated meristematic cap (but the cap is not uniform, it is penetrated axially by procambial leaf traces). The cap is, therefore, the meristematic region out of which the congested peripheral vertical bundles differentiate. Leaf traces, even those which supply the youngest primordium, are all continuous; below they consist of the vertical bundle, above of the leaf trace. The ultimate basal continuity cannot, of course, be demonstrated directly in the short length of axis examined, but since the basal region is fully developed (though not fully differentiated), extrapolation into the mature stem is obvious. With a leaf-contact distance of about 15 inter- nodes for major bundles P1 would connect with P16, P2 with P17, etc. The curvature of the trace is established by reorientation of central bundles, immediately below the apical meristem, through a right angle; the method of reorientation is described below. These distal leaf traces are more or less perpendicular to the crowded vertical bundles and they pierce the meristematic cap. This development of two procambial systems more or less at right angles to each other is the fundamental feature which ac- counts for the final configuration of bundles in the mature stem. The vascular anatomy of the rhizome crown (Fic. 3) is similar to that of the aérial stem, with quantitative differences which reflect distinguishing features of the mature rhizome. Inward and outward curvature of bundles is less pronounced, The meristematic cap is larger and more easily recog- nizable in single sections (Fics. 8 and 9). The youngest leaf trace with attached vertical bundle is that supplying P7. The greater plasticity (i.e., irregularity in construction) of the mature rhizome trace system is further verified in the crown where leaf traces may give off more than one vertical bundle (as the traces to P11 and P12, Fic. 3) or none. DEVELOPMENTAL INFERENCES From the information in the diagrams, Fics. 1-4, the essential features of vascular development in the crown of Rhapis become clear. By counting 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 133 1000 CORTEX La /CORTICAL Pg BUNDLES EXTENT OF CAP i... 3. MINOR BUNDLES 3Sv@ 4v371 LV S3STGNNG 43O Y3SEWNN AVLOL INTERMEDIATE BUNDLES CENTRAL CYLINDER MAJOR BUNDLES . 3s 4 kh Ge Fee eH se he LEAF NUMBER (i.e., the differential of the number of bundles) may be an artifact resulting from the difficulty of recognizing fibrous bundles in their early stage of development (see the discussion of this in the text). the total number of recognizable bundles at the leaf insertion at each developmental stage an idea can be gained of the period over which leaf trace development continues. Fic. 4 is a plot ona logarithmic scale of such * count. It shows that each leaf produces new traces continuously for 134 JOURNAL OF THE ARNOLD ARBORETUM [vor. 48 about 13 plastochrones. In P14 the total number of bundles is established. Quite obviously major bundles are produced first, intermediate ones later, minor ones still later, and cortical ones last. Of the total number of traces in each leaf (a little over 1,000) less than 100 in the stem are bundles of the central cylinder, the remainder are cortical bundles. Renewal of the meristematic cap in relation to the origin of dif- ferent bundle categories. A consideration of successively younger leaf traces in Fic. 2 shows that vertical bundles are differentiated basipetally from the cap, or expressed more correctly, the cap grows distally from the differentiating bundles. The cap must therefore be renewed from above by tissue (below about P1I—-P8) which at a later stage of development dif- ferentiates as vertical-bundle branches of the leaf traces (below about Pil). As a leaf is initiated on the apical meristem its first leaf traces each make contact with the distal extremity of a vertical bundle within the cap. Because early contacts with the leaf are made in the center of the stem the upper portion of the bundle will, after differentiation, still be located in the center. As the leaf grows older newly initiated leaf traces establish vascular contact with more and more centrifugal parts of the cap. These later vascular strands are intermediate and then minor ones, they reach less far into the stem center and their leaf-contact distances are shorter. Shorter leaf-contact distances for later vascular strands are simply the result of fewer plastochrones during which the vertical bundle differentiates as a distally unconnected strand. The rhizome crown is similar with an apparent exception that the vertical bundles in the cap keep growing even after leaf contact has been made. It must be remembered that nutrition of the rhizome crown takes place con- tinuously from proximal regions, since scale leaves do not assimilate. In the aérial crown newly matured leaves can supply assimilates to the meri- stematic region (note in Fic. 1, that leaves of the developmental stage P11 and older, are green and exposed to light). Only future experimental work on the translocation of nutrients in developing regions will allow a clearer understanding of these differences. Further mention may be made here of the relative irregularity of bundle development in the rhizome (Tomlinson & Zimmermann, 1966a). Vertical bundles occasionally (though rarely) split; vertical bundles commonly fuse to form the ‘long bridges’’ described in our earlier account. The number of vertical bundles diverging from a leaf trace is more variable than in the aérial axis and vertical bundles also seem able to continue unlimited growth in the cap without making contact distally with a leaf trace. In the mature rhizome we have followed numerous vertical bundles which had no leaf contact over distances in excess of 15 internodes. With further understand- ing of developmental processes in Rhapis it is likely that these “mistakes” can be seen to belong to the developmental norm. Leaf trace — vertical bundle linkage and the origin of the cor- tex. The way in which vertical bundles maintain a constant rate of dif- ferentiation in relation to the shoot apex and the meristematic cap is 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA erse section through the lowest le roi i gd y anda tran radially-running procam bi al Fiecaas (“tiger ta al ba tin region of the central cylinder in whic spira 136 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 48 evident from Fic. 2. Leaf traces are “‘sent out” by leaves for a period of about 13 plastochrones. The first leaf traces make contact with the central portions of the cap, subsequently initiated leaf traces contact more cen- trifugal parts of the cap. Still later traces ‘‘sent out” by a leaf when it is more than 7 plastochrones old appear outside the cap, i.e., outside the region of blind-ending vertical bundles. These traces must be located out- side the central cylinder, that is, in the area of the stem which we recognize later as the cortex. Having failed to make contact with vertical bundles these leaf traces must end blindly below, although they may anastomose among themselves as if still “seeking” vascular contact. This blind-ending cortical system is readily observed in the cortex of both axes; it is partic- ularly well developed in the rhizome. If we compare Fics. 2 and 4 we can see that the cap extends radially over approximately 7 leaf insertions. Thus over a period of about 7 plastochrones vascular contact between leaf and central cylinder is possible; subsequently all leaf traces are confined to the future cortex of the mature stem. Ficure 4 suggests that frequency of leaf-trace initiation occurs in two peaks. This, somehow, seems to be unlikely. It could be a reflection of a Fic. as ee excelsa. eos seage longitudinal section Rae the crown of the i aap érial axis. Because of their spiral path only short lengths of bundles can be 7, Compaction of this photograph with the diagram- matic Fic. 2 indicates Pa gfe 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 137 subjective phenomenon resulting from our inability to recognize early stages in differentiation of cortical compared with central bundles. In addition, the leaf base produces branching from existing leaf traces (vertical bundles, satellites, bridges) somewhat before stage 11; these are not included in the count and would “fill out” the curve. At any rate, it is reasonable to assume that only one peak of initiation frequency is in- volved. Fic. 4 should probably follow a simple sigmoid curve. Thus a developmental difference between early and late differentiating bundles manifests itself in mature parts as a clear morphological boundary between central cylinder and cortex. Cortex and central cylinder both remain essentially descriptive terms; both regions are axial parts and there is no need to enter into the futile philosophical discussion as to whether cortex may or may not be regarded as a “fundamental part of the stem” or merely as a downward continuation of the leaf base, a discus- sion to which earlier morphologists were prone (cf. for example, Baranet- zky, 1897). Origin of vertical bundles, bridges and satellites. The question now arises as to where in development the vertical bundle diverges from the leaf trace. This can only be estimated. It is quite evident that these branches must arise in continuity with the leaf trace but directed towards the cap because no vertical bundles ending blindly below have ever been observed in the central cylinder. Fics. 2 and 3 suggest that branching occurs when a leaf is about 11 plastochrones old in the aérial axis, 6 in the rhizome, since P11 and P6 respectively are the oldest primordia without an identifiable vertical bundle. We easily may have overlooked this vertical bundle but the configuration in Fic. 3 strongly suggests that the vertical bundle arises in continuity with the meristematic cap as soon as the leaf trace has been displaced to a position near the lower margin of the cap. Further consideration of Fics. 2 and 3 reveals that only the directly associated leaf can initiate development of vertical bundles, not the much higher one into which the vertical bundle ultimately proceeds. We know from our analysis of both types of axis that leaf-contact distances are of the order of a least 15 internodes for major traces. Fics. 2 and 3 further support this as they include 14 and 13 leaves respectively but no complete when a leaf is 6 plastochrones removed from the apex (it has already ap- peared in P7), then with a presumed leaf-contact distance of 15 internodes, the upper contact of this new vertical bundle would be to leaf 19,4 a leaf which does not appear for a further 9 plastochrones. Clearly it would be absurd to assume that I9 initiates the vertical bundle in P6. Even if we apex in an unidentifiable state, a simple computation by extrapolation from the number of cells visible in cross sections of the base of successively younger leaves gives a value for I9 of only a small fraction of a cell. It is ‘Leaf primordia are numbered P1 (the youngest), P2, P3, etc. Non-existing ones are numbered I1 (to appear next), 12, I3, etc. JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 8 (ABOVE), arse! a ee median se yporrages section ikea the crown of the rhizome. Bec ic cap (MC). traces : Fic. 9 (BELOW). Rhapis excelsa. Transverse section through the crown of the rhizome 0.3 seciiinias below the shoot apex. The undifferentiated meristematic 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 139 therefore quite impossible for vertical bundles to be initiated by non- existent leaves. This is precisely the conclusion of Priestley and his co- workers (1935) for Alstroemeria. The presence of numerous bridges and satellites (Zimmermann & Tom- linson, 1965) also lends support to the idea that vertical bundles are directed apically in their development. Bridges are usually short and always attached to neighboring vertical bundles in an upward direction. We regard bridges as “frustrated” vertical bundles, i.e., both are micro- scopically identical and seem to be branches sent out by the leaf trace. Both can now be regarded as developmentally identical, but positional (and probably also temporal) differences determine different fates. It can be envisaged that space for continued upward development of bridges is limited (they almost invariably diverge from leaf traces more distally than vertical bundles) ; as a result they fuse with other, nearby existing vertical bundles. Satellites may originate in the same manner and grow towards developing inflorescences (which may later abort). The inflorescence primordium, regardless of whether it aborts later or not, may be physi- ologically equivalent to a shoot apex in that it “attracts” some of the of branching is displaced downward du the traces to P12, 13, and 14 suggest in ! vertical bundles and bridges branch off the leaf trace in the mature stem is cap (MC) is seen as a ring penetrated by axially running leaf traces (LT). In the ee of the stem Pes ee right of the cap in the fi re and structurally proximal to the cap) are both leaf traces (LT) and vertical bundles (VB), in the periphery of the stem (to the left of the cap in the figure and morphologically distal to the cap) are only leaf traces. 140 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 incidental, for they are caused by displacement during primary thickening growth. RHAPIS COMPARED WITH ALSTROEMERIA Only in the work of Priestley, Scott and Gillet (1935) on Alstroemeria are we afforded a means of directly comparing our observations on Rhapis with those on another monocotyledon, although Alstroemeria is a great deal simpler than Rhapis. Nevertheless, these workers adopted an approach to a study of the developing vascular system in this non-nodal monocotyledon, similar to ours. It is noteworthy that Priestley and his associates also were forced to conclude that leaf traces in Alstroemeria developed acropetally long before the leaves they served became evident. On the other hand, Alstroemeria differs from Rhapis in that (1) there is no cortical system, (2) there are only 3 traces to each leaf, and (3) the leaf insertion occupies only a narrow sector of the stem circumference. This means that bundle linkage is dependent on leaf arrangement. In Rhapis, however, leaves encircle the stem completely at their insertion and phyllotaxis is incidental to bundle linkage. Nevertheless, a comparison of the diagrams illustrating the vascular system of Alstroemeria with ours for Rhapis suggests basic similarities in developmental terms. SUMMARY Vascular strands of the aérial axis and rhizome of Rhapis excelsa were followed in the meristematic crown with the microcinematographic method of analysis, and their course was plotted. From resulting plots vascular development is inferred. Vascular bundles of the crowded peripheral part of the central cylinder are continuous distally with a meristematic cap on the top center of which is the apical meristem proper. As each leaf develops it initiates leaf traces continuously during 14 plastochrones after inception. Leaf traces are oriented perpendicular to the vertical bundle and link with the blind ends of vertical bundles in the cap. Major vascular bundles are leaf traces initiated very early at a time when the leaf primordium is still located near the stem center. Leaf traces initiated later become inter- mediate and then minor bundles successively nearer the stem periphery, corresponding to the position of the leaf at the time of their initiation. The last traces, initiated outside the cap, end blindly below and differentiate into the cortical fibrous traces. The cortex, a structurally distinct region of the palm stem, is thus clearly defined by development. Determination of vascular bundles is a property of the meristematic cap effected in such a way that vascular bundles, as soon as they are continuous from a leaf primordium above to a leaf below, are displaced proximally (later centri- petally) and so lost from the cap. The cap must be renewed from above by tissue which at a later stage of development differentiates as vertical- bundle branches of the leaf traces. 1967 | ZIMMERMANN & TOMLINSON, RHAPIS EXCELSA 141 ACKNOWLEDGEMENTS We wish to express our further appreciation of the serial sectioning carried out by Miss Lesley Jackson and Miss Veronika Oswald (now Mrs. G. S. Walton) as a preliminary to this investigation. LITERATURE CITED BALL, E. 1941. The development of the shoot apex and of the primary thicken- ing meristem in Phoenix canariensis Chaub., with comparisons to Was hing- tonia filifera Wats. and Trachycarpus excelsa Wendl. Am. Jour. Bot. 28: 820-832. oe M. J. 1897. Sur le développement des points Mia be des tiges ez les monocotylédones. Ann. Sci. Nat. Bot. VIII. 3: 311-365. Sides J. C. 1884. The prio exowth of ee a in palms. Proc. Am. Phil. Soc. 21: CuHovarp, P. 1936. La fae # me des formations dites ‘secondaires’ dans Védification de la tige des monocotylédones. Bull. Soc. Bot. France 83: 819-836. Ecxarprt, T. 1941. Kritische Untersuchungen iiber das primare Dickenwachstum ei Monokotylen, mit Ausblick auf dessen Verhaltnis zur sekundaren Ver- dickung. Bot. Arch. 42: 289-334. Esau, K. 1965. Vascular differentiation in plants. 160 pp. Holt, Rinehart & inston, New Yor Pcp iciene P. 1876. Vergleichende Untersuchungen iiber den Bau der Vegeta- tionsorgane der Monocotyledonen. 220 pp. 3 pls. Ferdinand Enke, Stuttgart. Heim, J. 1936. Das Erstarkungswachstum der Palmen und einiger anderer Monokotylen, zugleich ein Beitrag zur Frage des Erstarkungswachstums der Monokotylen iiberhaupt. Planta 26: 319-364. Kumazawa, M. 1961. Studies on the vascular course in maize plant. Phyto- morphology 11: 128-139. MENEGHINI, G. 1836. Ricerche sulla ey del caule nelle piante mono- cotiledoni. 110 pp. 10 pls. Minerv. a. OHL, H. von. 1824. De palmarum alsa In: K. F. P. von Martius, Histovik Naturalis Palmarum 1: I-LII. 16 pls. ——. 1849. On the structure of the palm stem. Rep. Ray Soc. 1849: 1-92. Prirstiey, J. H., L. I. Scorr, & E. C. Grter. 1935. The development of the shoot i in Alstroemeria and the unit of shoot growth in monocotyledons. Ann, rae 49: 161-179. L. I. Scorr. 1937. Leaf venation and leaf trace in the monocotyledon. Proc. Leeds Phil. Lit. Soc. 3: 305-324. Scott, L. I., & J. H. Prrest.ey. 1925. Leaf “apd og anatomy of Tradescantia fluminensis Vell. Jour. Linn. Soc. Bot. 47: 1 SHARMAN, B. C. 1942. ia aan eeeneg ee the shoot of Zea mays L. Ann. Bot. II. 6: 245-2 Tomtinson, P. B. 1961. pale In: C. R. Metcatre, ed. Anatomy of the Monocotyledons. 2 : xv + 453 pp. Clarendon Press, Oxfor & M. ZIMMERMANN, 1966a. Anatomy of the palin Rhapis excelsa, II. Rhizome. Jour. Arnold Arb. 47: 248-261. 142 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 —————————.. 1966b. Vascular bundles in palm stems —their bibliographic evolution. ny Am. Phil. Soc. 110: 174-181. ZIMMERMANN, M. H., & P. B. Tomitnson, 1965. Anatomy of the 2% Rhapis excelsa, I. Mature vegetative axis. Jour. Arnold Arb. 46: 160-17 HARVARD UNIVERSITY Casot FOUNDATION PETERSHAM, MASSACHUSETTS 01366 AND FAIRCHILD TROPICAL GARDEN Mramt, Fiorina 33156 1967 | MOORE, NEW AMERICAN PALMS 143 TWO NEW AMERICAN PALMS Harorp E. Moore, JR. I. A SECOND SPECIES OF ARISTEYERA Shortly after the publication of the new genus and species Aristeyera spicata in this journal (46: 5. 1966), Dr. Julian A. Steyermark sent speci- mens for study which proved to represent a second species, here described as Aristeyera ramosa. Since there are major differences between the new A, ramosa and A. spicata, as listed below, an emendation of the generic circumscription is provided. The branched inflorescence is subtended by three bracts, a most unusual feature in a tribe where only two principal bracts have normally been known heretofore. These characteristics of the inflorescence suggest a less specialized species than A. spicata. The smaller number of stamens in the staminate flower and of staminodes in the pistil- late flower suggests the possibility that more than six stamens may be a derived rather than a primitive state in the Geonomeae, a possibility re- quiring further anatomical study before any conclusion can be drawn. Aristeyera H. E. Moore, char. amplif. Solitary, slender, unarmed, monoecious protandrous palms sometimes producing short lateral vegetative branches with small leaves and adventi- tious roots at upper nodes. Leaves normally undivided except for the bifid apex, pinnately nerved; sheath short, rupturing opposite the petiole. In- florescences interfoliar, erect, subtended by 2 or 3 closely sheathing bracts inserted at or near the base of the peduncle, the lower bract ancipitous, the upper 1 or 2 longer, with rounded margins, opening obliquely at the apex; peduncle elongate, terminating in one or in several subfasciculate spike-like flowering axes, the flowers borne in triads of 2 lateral staminate and a central pistillate (which sometimes aborts in apical triads) sunken in pits, these covered with a bract in bud and arranged in 6-8 vertical series. Flowers about 14 exserted at anthesis: staminate flowers with 3 gluma- ceous sepals, these distinct and more or less keeled above a short attach- ment to the receptacle, imbricate at least basally in bud; petals 3, adnate basally to the receptacle, connate above (sometimes unequally so) for \4—34 their length in a soft tube, the free, slightly asymmetric lobes del- toid, glumaceous, valvate above briefly imbricate bases; stamens (11-) 12-24, exserted at anthesis, the filaments adnate basally to the receptacle, then expanded and connate in a tube surrounding a fleshy, shallowly and acutely, often irregularly 3-lobed pistillode, the free portions of the filaments only slightly longer than the connate portions, erect, anthers in- flexed in bud, thecae separate, longitudinally dehiscent, terminal on arms 144 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 of a bifid connective, arms of the connective and thecae becoming erect or nearly so at anthesis: pistillate flowers with 3 glumaceous distinct sepals imbricate in bud; petals 3, connate basally in a soft fleshy tube with deltoid glumaceous lobes shortly imbricate basally, valvate above in bud, exserted and spreading at anthesis; staminodes (5—6—)7—18, fleshy, serted and more or less thickened above the throat of the staminodial tube at anthesis; stigmas 3, recurved, at anthesis; ovule hemitropous, axile, pendulous, attached near the top of the locule, only one normally maturing. Fruit ellipsoid-ovoid, slightly compressed laterally and slightly keeled apically when dry, the residual stigmas and abortive carpels basal; exocarp smooth; mesocarp rather dry, with longitudinal fibers closely placed around the endocarp, this thin, crustaceous, shining, not adherent to the seed; seed ellipsoid to obovoid and slightly compressed laterally; hilum small, excentrically basal; raphe encircling the seed from hilum to excentrically basal embryo, somewhat impressed and unbranched to furcate or with a few parallel branches; endosperm homogeneous. Aristeyera ramosa H. E. Moore, sp. n Fics. 1 and 2 Ab Aristeyera spicata inflorescentia ramosa bracteis 3 subtenta, rachillis 4—6 subfasciculatis glabris alveolas floriferas in seriebus 6—8 verticalibus gerentibus, florum masculorum staminibus (11—)12, florum femineorum staminodiis (5—6—)7—9 subulatis differt. Stems solitary, to 3 m. high, 3-8 cm. in diam., often producing short lateral vegetative branches with small leaves and adventitious roots at nodes above. Leaves rigid, subcoriaceous, paler green below (Steyermark) ; sheath short, ca. 11-17 cm. long at margin of petiole, 6-17 cm. long an frayed into irregular lobes opposite petiole, fibrous, densely brown-tomen- tose inside and outside; petiole 33-43 cm. long, brown-tomentose below, at least when young, becoming glabrous; blade cuneate in outline, undivided except at bifid apex, 58.5—66.5 cm. long along the rachis, 21-27 cm. wide at apex of rachis, 31 cm. or more long along inner margin of terminal lobes, the rachis glabrous above, more or less brown-tomentose below, primary nerves 34-38 on each side, with scurfy small brown or rusty appressed scales on these and the secondary and tertiary nerves below. Inflores- cence erect in flower, becoming pendulous in fruit, subtended by 3 bracts, the lowermost short, 15-20 cm. long, ancipitous, brown-tomentose, be- coming glabrous, the middle one inserted close to the lowest or to ca. 8.5 cm. above it, 25-49 cm. long, opening obliquely at apex, densely brown- tomentose, the uppermost similar to the middle bract and exceeding it by 11-18 cm.; peduncle densely brown- or rusty-tomentose, 46-178 cm. long; rachis very short, 1.3—2.5 cm. long; rachillae 4-6, subfasciculate, glabrous, 18-20 cm. long, 6—8 mm. in diam., with flower pits in 6-8 vertical series, becoming maroon-rose or witescoleted in fruit. Staminate flowers at anthesis ca. 6-6.5 mm. long; sepals 5 mm. long; corolla-tube ca. 3 mm. 1967] MOORE, NEW AMERICAN PALMS 145 MINSTERIO DE ACTUCULTURA ¥ CHIA HESDARIO NACIONAL DE VENEZOPIA %s. 96088 : lst sht- Aristeyera paste Trunk 2-3 m. tall; leaves paler green below; flowering peduncle erect STADO STAHE. PENNS he ania os ri a orks ‘eo TYPE eis wsttonily Firth. Ps Ce ee ar 01 bE Julias A. Steyermark y Marvin Rabo Fic. 1. Aristeyera ramosa. a, leaf; b, inflorescence. long, corolla-lobes ca. 3.4 mm. long; stamens (11—)12; pistillate flowers at anthesis (from preserved material) ca. 7 mm. long; sepals 5 mm. long; corolla-tube ca. 4 mm. long, corolla-lobes 2.6 mm. long, acute; staminodes (5-6—) 7-9, connate in a fleshy tube 4 mm. long and adnate to the corolla- 146 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ay ‘ is; iy oy. - Fic. 2. Aristeyera ramosa. c, habit, from kodachrome by J. A. Steyermark; d, portion of inflorescence with pistillate flowers at anthesis, from kodachrome by J. A. Steyermark. tube, then free, ca. 3.6 mm, long, fleshy, subulate, spreading at anthesis; style exserted and expanded above throat of staminodial tube; stigmas recurved. Fruit (not fully mature) ellipsoid, 15-16 mm. long, ca. 5 mm. in diam., dark purple-red. VERNACULAR NAME: anare Venezuela. EsTaDo Sucre: Peninsula de Paria, Cerro de humo, laderas pendi- entes de bosque nublado virgen que miran al norte, a lo largo de las cabeceras de uno de los afluentes del Rio Santa Isabel, arriba de Sipara, oeste de Santa Isabel, rercne! de 14 kms. al norte del pueblo de Rio Grande Arriba, arriba de Boca de Cumana y Punto Siparo, noroeste de Irapa, alt. 1060 m., 1 Mar. 1966, Julian A, Sacre 94803 (BH); Cerro Espejo, selva siempre verde en la cumbre, entre Manacal y Pauji, arr riba de Mundo Nuevo, arriba de Rio Seco de Irapa, alt. 0-850 m., 6 Au ug. 1966, Julian A. Steyermark & Marvin Rabe 96088 (BH, holotoes : The epithet for this second species of Aristeyera is used with reference to both the branched inflorescence and the unusual short lateral branches often produced along the upper portion of the stem. It is perhaps worth noting that in the two collections available, the flower pits are arranged in six vertical series in the three inflorescences of the type while in the single fruiting inflorescence of Steyermark 94803 rs are in eight vertical series on one axis, in six on the remaining five ax It may be useful to tabulate the diverging iipeatates of the two species in the following way: 1967 | MOORE, NEW AMERICAN PALMS 147 Stems not producing lateral branches Bracts subtending the inflorescence 2 Flowering axis simple, spike-like, densely ferrugineous-tomentose, with flower pits in 8 vertical series Stems often producing short lateral vegetative branches along the up- per portion Bracts subtending the inflorescence 3 Flowering axes 4-6, subfasciculate, glabrous, with flower pits in 6-8 vertical series Staminate flowers with 21-24 sta- Staminate flowers with (11—)12 sta- mens mens Pistillate flowers with (5-6—)7-9 staminodes, these subulate Pistillate flowers with 15-18 stami- nodes, these linear, angled, and sometimes thickened apically Fruit 15-16 mm. long when still im- mature Fruit 12-14 mm. long at maturity II. A NEW SPECIES OF PHOLIDOSTACHYS FROM COLOMBIA Several years ago, a description of a supposedly new palm was prepared and put aside pending the resolution of some generic problems in the tribe Geonomeae of the subfamily Arecoideae. At the time, not understanding distinctions among those genera with six stamens and sagittate anthers, I had thought this to be a new Calyptronoma. In the interim, Mr. R. W. Read has provided preserved material of Calyptronoma occidentalis col- lected in Jamaica, calling my attention to the peculiar cap formed by the petals, and I have been able to study old pistillate flowers of Pholidostachys pulchra, On re-studying specimens of the new palm to make them avail- able to the current monographer of the tribe, it is clear that the species does not belong in Calyptronoma but is referable to Pholidostachys. At present, only two species, P. pulchra from Costa Rica and P. Kal- breyeri from Colombia, are assigned to the genus Pholidostachys. Each has the inflorescence composed of a single spike-like flowering axis terminating the peduncle and the two are probably to be regarded as synonymous. ‘In the light of generic differences noted below, two other South American species now in Calyptronoma will be assigned to Pholidostachys (3. Ss, Wessels Boer, in correspondence). Pholidostachys dactyloides differs from P. pulchra and P. Kalbreyeri in having an inflorescence with several flower- ing axes. It differs from the other species assignable to the genus, each of which has a ramified inflorescence with an elongate rachis, in the few, subfasciculate rachillae borne on an extremely short rachis. It is these clustered, finger-like rachillae that suggest the epithet used. An inexplicable error describing the staminodes of Calyptronoma as united in a solid receptacle, in a published key (Journal of the Arnold Arboretum 46: 3. 1966), also calls for rectification. I would, today, sepa- rate Calyptronoma, Calyptrogyne, and Pholidostachys, all of which have very similar staminate flowers, as follows: 1. Petals of the pistillate corolla very briefly connate in a tube about as high 148 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 as the ovary, the lobes distinct, glumaceous, persistent, not forming a cir- cumscissile cap; staminodial tube gradually widened apically to 6 prominent fleshy lobes as long as or longer than the tube and spreading at anthesis; styles scarcely united except at the base, distinct and stout for most of their length, the stigmatic tips somewhat recurved. Costa Rica to Peru. ........ wh lair eric Rented train yi panei ei oh ce ERS Ai ae Bee ee Pholidostachys. Petals of the pistillate corolla connate basally in a thin membranous tube for more than half their length, the lobes valvate and thicker, not separating but forming a circumscissile conic cap caducous at anthesis; staminodes connate in a slender tube and adnate to the corolla-tube basally, the upper portion exserted at anthesis, inflated, urceolate, with 6 very short and narrow es- sentially erect lobes, caducous; styles loosely united for most of their length, sometimes thickened toward the apex, becoming long-exserted after the terminal portion of the staminodial tube has fallen; stigmas recurved. — 2. Upper bract of inflorescence inserted at or near the base of the peduncle, persistent or at length marcescent; inflorescence large, paniculately twice- ranched; stems moderately stout, to 20 meters high or more. Greater Pe Se ee er Calyptronoma. 2. Upper bract of inflorescence inserted at base of and enclosing the spike- like flowering axis in bud, caducous at anthesis or soon marcescent, leav- ing a ruff-like scar; stems short, seldom developed above ground. Mexico At SAR Oe oh he ree Calyptrogyne. Pholidostachys dactyloides H. E. Moore, sp. nov. Fics. 3 and 4 Palma mediocris caudice solitario ad 10 m. alto, 8 cm. in diam., pinnis 4—9-nervatis, inflorescentiae rachidi brevissima, rachillis 5-6 tomentosis erectis subfasciculatis alveolas floriferas in seriebus 10 verticalibus gerenti- bus, fructu obovoideo 14-15 mm. longo, 10-12 mm. in diam., semine late ellipsoideo 9 mm. alto, 8 mm. in diam. Stems solitary, to 10 m. high, 6.5-8 cm. in diam., grayish-brown. Leaves in a dense crown; sheath short, closed and swollen at the base, with red fibers; petiole 40-50 cm. long, narrow, shining, glabrous and slightly channelled above, rounded and densely ferrugineous-tomentose becoming glabrous below; blade 1.2-1.4 m. long, the pinnae 10-11 on each side of the rachis, this rounded and deciduous-ferrugineous-tomentose below, narrowly convex, elevated centrally with prominent flattish wing-like mar- gins and glabrous above, at least basally; pinnae subopposite to alternate, inserted on the winged margin, all 4—6—9-nerved, with prominently and acutely keeled nerves above, glabrous except for deciduous red-brown elongate scurfy scales along the predominant elevated and truncate nerves below, these alternating with red-purple-margined impressed nerves cor- responding to elevated nerves above, the surface densely and minutely pale-puncticulate, lowermost pinnae ca. 50 cm. long, 4.5 cm. wide, pinnae from mid-leaf ca. 70-75 cm. long, 7-9-13 cm. wide, falcate-acuminate, apical pinnae ca. 45 cm. long, 7 cm. wide, 16 cm. along the rachis. Inflores- cence interfoliar, subtended by two bracts, the lower ca. 33 cm. long, ancipitous, densely ferrugineous-tomentose, upper bract inserted 4—5.5 cm. above the lower, stouter, ferrugineous-tomentose, coarsely fibrous and 1967] MOORE, NEW AMERICAN PALMS 149 DUPLICADO de ly Coleechin de la Los de Potantce Fs de ba Secertaria 8 _ alters pase Eo) es Kawi Bog. de Crlombias COLOMBIA, cg ytanventa del Valle ie Yorumangai: Ei Papav: Fic. 3. Pholidostachys dactyloides. e, isotype (COL). fraying in age, not ancipitous, 54-63 cm. long, splitting abaxially and then 4.5-6 cm. wide, attenuate at apex; peduncle 15-22 cm. long, dorso-ventral- ly compressed and elliptic in section, densely ferrugineous-tomentose, with several acute triangular bractlets to 2 cm. long toward the apex, terminating in a very short rachis and 5-6 erect subfasciculate similarly tomentose rachillae 26.5-33 (—40 in fruit) cm. long, ca. 1 cm. in diam., the pits in 150 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 10 vertical series, the lips entire, about as broad as long. Staminate flowers ca. 5.5 mm. long; sepals distinct, 4.5 mm. long, slightly keeled, rounded to subcucullate apically; petals very briefly connate and united with the connate stamen-filaments basally but distinct and valvate above; stamen- y Heat bocde, HE AA (Ratiheetrelde Tseorr re y COLOMBIA, Departamento ded Valle. sasta del Parifiee; rie Yurnmangel: EEF fo. all, Ia “A ee Soke 1k . 4. Pholidostachys dactyloides. f, portion of leaf from isotype (COL); &. eatin of infructescence from isotype (COL). 1967] MOORE, NEW AMERICAN PALMS 151 filaments united in a 3- -angled solid fleshy obconic stipe, then distinct as 6 prominent narrowly triangular fleshy lobes ca. .1 mm. long tapered to very thin inflexed apices and sagittate anthers, these versatile and prob- ably spreading at anthesis; pistillode short, subulate; pistillate flowers with distinct acutish keeled sepals ca. 4.5 mm. long; petals similar to the staminate but united basally with the staminodial tube for about 1.5 mm.:; staminodes connate in a soft fleshy obconic tube with 6 fleshy lobes, these spreading at anthesis; pistil trilocular, triovulate, very deeply 3-lobed, the ovaries discrete except on the inner face; styles very briefly united basally, distinct and elongate above, with subclavate spreading exserted stigmas. Fruit black at maturity (17257), 14-15 mm. long, 10-12 mm. in diam., obovoid, the fibers of the mesocarp not prominent in dry mature fruit but very conspicuous in the dry immature fruit (15982); mesocarp with thick curved anastamosing fibers and an inner layer of slender fibers; endocarp thin, not adherent to the seed: seed broadly ellipsoid, ca. 9 mm. high, 8 mm. in diam., with raphe and chalaza forming a narrow continuous band from the subbasal point of attachment to the basal embryo; endosperm homogeneous. Colombia. DEPARTAMENTO DEL VALLE: Costa del Pacifico, rio Yurumangui, El Papayo, bosques, 10-20 m. alt., Feb. 5, 1944, J. Cuatrecasas 15982 (BH, type; COL, F, isotypes) ; rio Cajambre, Barco, 5-80 m. alt., Apr. 21-30, 1944, J. Cuatre- casas 17257 (COL, F). L. H. BarLey es CORNELL UNIVERS ITHaca, NEw Yorx er 152 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 NOMENCLATURAL NOTES ON GOSSYPIUM (MALVACEAE)? GEorGE K. Brizicky N THE COURSE OF SURVEYING the genera of Malvaceae in the south- eastern United States, I encountered in the genus Gossypium several nomenclatural problems which require further comments. These include the type species of Gossypium; the nomenclatural status of G. barbadense L.; and the sectional names in Hutchinson’s classification of Gossypium (1947). These problems are dealt with separately below. THE TYPE SPECIES OF GOSSYPIUM In 1920, Britton & Brown (Bahama FI. 273. 1920) designated Gossypi- um arboreum L. as the lectotype species of Gossypium L. This choice of lectotype, apparently the earliest for this genus, was also adopted by Britton and Wilson in their Botany of Porto Rico and the Virgin Islands (Sci. Surv. Porto Rico Virgin Is. 5(4): 566. 1924). In 1929, however, Hitchcock and Green (Int. Bot. Congr. Cambridge 1930. Nomencl. Propos. Brit. Bot. 173. 1929), apparently overlooking Britton & Millspaugh’s typification, proposed G. herbaceum L. as the type of Gossypium. This typification has been adopted by at least Prokhanov (Bot. Zhur. SSSR 32: 66. 1947), Phillips (Genera S. Afr. Flower. Pl. ed. 2. 501. 1951), and Hu (Malvaceae. Fl. China, Fam. 153. 61. 1955). Since Britton & Millspaugh’s choice of the lectotype species of Gossypium L. appears to be the earliest, antedates that of Hitchcock and Green, and does not contradict the Jnter- national Code of Botanical Nomenclature, it must be followed (see ICBN. p. 20. Art. 8 & p. 64. point 4f. 1961). THE NOMENCLATURAL STATUS OF GOSSYPIUM BARBADENSE L. It took botanists about a century and a half to form a notion of Gossypi- um barbadense L. and a further half-century to learn that their concept has been wrong and that the species so called must bear the name G. peruvianum Cav. (correctly, G. vitifolium Lam.), while G. barbadense L. has to be reduced to the synonymy of G. arboreum L. Since some taxono- mists have accepted the proposed change, while others have not, this eco- nomically important species is at present variously called G. barbadense L., G. peruvianum Cav., and G. vitifolium Lam., with consequent nomen- clatural and taxonomic confusion. Since rejection of a well-established * Continuing a series of miscellaneous notes and papers on the flora of the south- eastern United States made possible through the interest and support of George R. Cooley and grants from the National Science Foundation. 1967 | BRIZICKY, NOTES ON GOSSYPIUM 153 and generally used scientific name of an economically important plant usually is undesirable and regrettable, any proposal of such a change should be thoroughly discussed, and eventually the necessity of renaming should be reinvestigated. For this reason a review of the nomenclatural status of G. barbadense L. would seem to be useful. The protologue for Gossypium barbadense Linnaeus (Sp. Pl. 2: 693. 1753) consists of the following parts: 1) the Linnaean diagnosis, taken from his Hortus Upsaliensis (p. 204. 1748), “Gossypium foliis trilobis integerrimis”; 2) Plukenet’s phrase-name (Almagest. Bot. 172. 1696), “Gossypium frutescens annuum, folio trilobo [trilobato], barbadense,” and a reference to Plukenet’s illustration ‘‘|Phytogr.] ¢. 188. f. 1”; and 3) the habitat and growth form of the species, “Habitat in Barbados,” followed by the symbols for “biennial” and “shrub.” The Linnaean nomen specificum legitimum (phrase name or diagnosis) seems to refer to the plant grown in the Botanical Garden in Uppsala, because besides the diagnosis the following note and a brief description of this plant are given by Linnaeus in Hortus Upsaliensis (loc. cit.), ““Hos- pitantur in Caldario, biennis, altero anno fructificans. Desc. Folia cordata, triloba, integerrima, hinc diversa, licet fruticosa videatur, a Gossypio caule erecto, Hort. cliff. 350.” On the other hand, the circumstance that Lin- naeus applied to his species the epithet barbadense taken from Plukenet’s diagnostic name, seems to indicate that he based his species on that of Plukenet, which he knew only from the phrase-name and illustration. He apparently believed the plant grown in the botanical garden in Uppsala to be conspecific with that of Plukenet. Thus, G. barbadense L. was a mix- ture of two elements, the plant grown at Uppsala and the species estab- lished by Plukenet. Although the vast majority of post-Linnaean bota- nists considered both elements to be conspecific, Todaro (Osservaz. Talun. Spec. Cotone 90. 1863) regarded G. barbadense as based only on the plant cultivated in the botanical garden in Uppsala and described by Linnaeus for the first time in Hortus Upsaliensis. “The G. barbadense is to us a plant fairly uncertain; it was for the first time described from the Garden in Uppsala. . . The G. barbadense is not, as we have believed, a plant described from the illustration of Plukenet, but certainly from the living plant cultivated in the Botanical Garden in Uppsala.” (Translation sup- plied.) Consequently, Todaro removed Plukenet’s phrase-name and the reference to his illustration from the synonymy of G. barbadense L., but in his later work (Relaz. Cult. Cotoni 234. 1877) he returned both. How- ever, he then believed this species to be different from his G. maritimum, the Sea Island cotton, and to occur spontaneously on Barbados (loc. cit. 36). Contrary to Todaro, Watt (Wild Cultiv. Cotton 268, 269. 1907) typified Gossypium barbadense by Plukenet’s plant, saying: “There can, however, be little doubt as to the botanical type of the species. It was founded by Linnaeus on the description and plate given by Plukenet. - In the second edition of the ‘Species Plantarum’ Linnaeus added the information that the leaves below had three glands, but it seems doubtful whether this 154 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 supplementary feature is a constant characteristic or may not rather be an acquired one through hybridization or adaptation to insect visitors. Plukenet’s specimen is in the Sloane Herbarium of the British Museum (vol. 100, f. 105 and Ray 1064-1), and is therefore the absolute type of the species as originally conceived. It is reproduced here (see Plate 46 A). It was fairly accurately represented by Plukenet’s figure, also for con- venience reproduced by me (Plate 46 B), except that the fruit shown is not present on the specimen.” Taking into consideration the confusion (which goes back even to Plukenet himself) surrounding Plukenet’s species, Watt further stated that “while accepting his [Plukenet’s] t. 188, f. 1 (Plate 46 B) as the type of the species (as Linnaeus did), we must exclude all the synonyms cited by Plukenet.” Watt seems to be the first to identify the plant in the Linnaean Herbarium, which bears on the sheet a mark “HU” [Hortus Upsaliensis| and an inscription in Linnaeus’s handwriting “barbadense ?”, as Gossypium obtusifolium Roxb., which is now included in G. arboreum L. This presumably was the plant grown in Uppsala and described by Linnaeus in Hortus Upsaliensis as “Gossypium foliis trilobis integerrimis,” e.g., the type of G. barbadense, in Todaro’s opinion. Todaro’s (1863) typification was supported and adopted by Prokhanov (Bot. Zhur. SSSR 32: 67. 1947) who reviewed in detail (in Taxon 8: 41-46. 1959) the nomenclatural status of Gossypium barbadense L. and came to the following conclusions: “(1) Gossypium barbadense, in the original sense of Linnaeus, is merely a broad-leaved variety of G. arboreum L. It must bear the name G. arboreum L. var. nadam (Watt) Prokh. |G. obtusifolium Roxb. s. str.], thus the name G. barbadense L. becomes a synonym of G. arboreum L., and in practice has to be eliminated. (2) The specific epithet barbadense is due here to its arbitrary transference by Linnaeus from former collective species of Plukenet. No cotton of the South American group was ever described by Linnaeus. (3) The species of Gossypium, up till now erroneously named G. barbadense and repre- senting the South American group, must bear as its proper name, Gossypi- um peruvianum Cav. [correctly, the earlier G. vitifolium Lam.|.” Prok- hanov’s proposal, although adopted by several Eurasian and American botanists, has not met with general recognition, most authors following Watt’s typification, which I shall review later on. Although Todaro’s and Prokhanov’s assumption that Gossypium bar- badense L. was based only on the living plant grown in the Botanical Garden in Uppsala and described by Linnaeus in Hortus Upsaliensis does not appear to be quite correct, Todaro and Prokhanov were fully justified in typifying this species by that element. Nevertheless, this typification may be objected to because of the fact that Prokhanov’s identification of the above-mentioned plant (made solely on the basis of the Linnaean diagnosis, note, and a very incomplete description) as a broad-leaved variety of Gossypium arboreum seems to be fairly unreliable and uncon- vincing, being an educated guess rather than a determination. Certainly, the specimen from the Linnaean Herbarium, bearing a mark “HU” [Hortus Upsaliensis| and Linnaeus’s inscription “barbadense ?”? must be the basis 1967 | BRIZICKY, NOTES ON GOSSYPIUM $55 for identification of the type of G. barbadense L. in the sense of Todaro (1863) and Prokhanov (1959). This specimen, then, presumably repre- senting the plant grown in Uppsala and determined by Prokhanov as Gossypium arboreum L. var. nadam (Watt) Prokh., should be chosen as the lectotype of G. barbadense L. in the sense of Todaro and Prokhanov. However, “‘in choosing a lectotype, any indication of intent by the author of a name should be given preference unless such indication is contrary to the protologue. Such indications are manuscript notes, annotations on herbarium sheets,” etc. (ICBN. 65. 1961). The specimen of the Linnaean Herbarium (874.5), identified by Watt as G. obtusifolium Roxb. and by Prokhanov as G. arboreum L. var. nadam (Watt) Prokh., was annotated by Linnaeus himself as “barbadense ?”. Designation of this specimen, only questionably referred by Linnaeus to G. barbadense, as the lectotype of this species can hardly be justifiable from the standpoint of the Code. Consequently, Prokhanov’s typification of G. barbadense L. with the plant grown in the botanical garden in Uppsala and allegedly represented in the Herbarium Linnaeanum by the specimen of G. arboreum L. var. nadam (Watt) Prokh. can not be accepted. As mentioned before, Watt (loc. cit.) typified G. barbadense L. by Plukenet’s plate [Phytogr.] tab. 188, fig. 1, clearly saying that G. barba- dense L. “was founded by Linnaeus on the description and plate given by Plukenet” and that “accepting his [Plukenet’s] t. 188, f. 1 (Plate 46 B) as the type of the species (as Linnaeus did), we must exclude all the synonyms cited by Plukenet.” Despite the clarity of Watt’s typification, his further statement that Plukenet’s specimen in the Sloane Herbarium is “the absolute type of the species as originally conceived [by Plukenet, not Linnaeus],” was misunderstood as Watt’s choice of Plukenet’s speci- men as the lectotype of G. barbadense (e.g., Hutchinson in Hutchinson, Silow & Stephens, Evol. Gossypium 48. 1947). Recently, however, Wouters (Bull. Jard. Bot. Bruxelles 33: 516. 1963) really assumed Plukenet’s specimen to be the type of G. barbadense. “Tl est donc normal de con- sidérer le spécimen Plukenet vol. 100. folio 105 comme le modéle de la planche t. 188 fig. 1 de Plukenet, et par conséquent comme le type nomen- clatural de G. barbadense L. 1753. Il représente indiscutablement G. barba- dense sensu stricto.” Wouters’s assumption, however tempting, may be objected to on the following grounds. 1) The specimen Plukenet vol. 100. folio 105 was not that from which Plukenet’s illustration was drawn; this herbarium specimen, lacking any inscription by Plukenet, was arbitrarily designated by Watt (Joc. cit.) as “the absolute type of the species as originally conceived,” only on the basis of similarity of the leaf-shape of the specimen to that of the illustration. Regarding the true model for Plukenet’s illustration Watt (Joc. cit. 269) says the following: “In vol. 132, f. 18 (Duchess of Beaufort’s set of plants) there is a specimen, said to represent Plukenet’s ‘Phyt.’ t. 188, f. 1, but which is nearer to the leaves in his herbarium just mentioned [i.e., those of G. vitifolium Lam., which was regarded as a distinct species by Watt].” 2) Since the Plukenet speci- men in the Sloane Herbarium was not the model for Plukenet’s illustration, 156 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 there is no reason at all for choosing it as the lectotype of G. barbadense, because “A lectotype must be chosen from among elements that were definitely studied by the author up to the time the name of the taxon was published and included in the catalogue” (ICBN. 64. 1961). On the contrary, Watt’s choice of Plukenet’s illustration (Phytogr. tab. 188, fig. 1) as the lectotype of G. barbadense would be quite reasonable and admissable if there were no other objections from the standpoint of the Code. However, Todaro had already (Osservaz. Talun. Spec. Cotone 90. 1863, & Relaz. Cult. Cotoni 228. 1877) noticed that the plant illus- trated by Plukenet had three-lobed leaves similar to the upper leaves of Sea Island cotton (G. maritimum Tod.), but a fruit resembling that of some East Indian species of cotton in its obtuse, almost entire-margined, reflexed bracteoles. Prokhanov (Taxon 8: 42. 1959) corroborated Todaro’s observations and identified the capsule of Plukenet’s drawing as that of G. herbaceum. Thus, Plukenet’s drawing does not seem to portray a real plant, but rather presents a more or less imaginary composite plant with leaves approximately those of Sea Island cotton and capsule approximately that of G. herbaceum, thus a plant expressing Plukenet’s concept of his collective species (which included the species of the New and Old Worlds) rather than a real plant. Plukenet’s drawing presents heterogeneous materi- al consisting of two discordant elements. Since the type material of G. bar- badense L. is composed of two inseparable discordant elements, this Lin- naean binomial must be rejected in compliance with Article 70 of the International Code of Botanical Nomenclature (1961). Wouters’s pro- posal to regard the reflexion of the bracteoles of the involucel on the fruit of Plukenet’s figure as the result of an interpretation can not be accepted, because in their shape and margin characters (not solely in their reflexed position) the bracteoles of involucel on Plukenet’s figure certainly represent the cultivated Old World species. Thus, while disagreeing with Prokhanov in argument, I concur with his proposal that the binomial Gossypium barbadense L. must be replaced by G. vitifolium Lam. (G. peruvianum Cav.) as the correct name for the Sea Island cotton. Having reviewed the situation concerning the typification of Gossypium barbadense L., I have come to the following conclusions: 1. Gossypium barbadense L. is a heterogeneous taxon consisting of two main elements, a plant grown in the botanical garden in Uppsala and a species established by Plukenet (Almagest. Bot. 172. 1696, & Phytogr. tab. 188, fig. 1. 1691). 2. Todaro’s (1863) and Prokhanov’s (1959) typifications of Gossypium barbadense by the plant grown in the botanical garden in Uppsala can not be accepted since the specimen in the Herbarium Linnaeanum (874.5), allegedly representing this plant and identified by Prokhanov as G. ar- boreum var. nadam (Watt) Prokh., was only questionably referred by Linnaeus to G. barbadense and therefore can not be chosen as the lectotype. 3. Watt seems to be correct in his assumption that Gossypium barba- dense L. was based on Plukenet’s diagnostic name and figure in Phytogr. tab. 188, fig. 1, which is the type of this species. However, the Plukenet 1967 | BRIZICKY, NOTES ON GOSSYPIUM 157 figure, in turn, is heterogeneous, presenting a = ig more or less imaginary, plant with leaves resembling some of those in the Sea Island cotton and the capsule apparently that - G. jae L. Therefore, the choice of Plukenet’s figure as the lectotype of G. barbadense L. is not admissable under the Code (ICBN 1961. Art. 70). 4. Typification of Gossypium barbadense by the specimen Plukenet vol. 100, folio 105 (the Sloane Herbarium of the British Museum) is not justifiable in the light of the Code, because Linnaeus did not see this specimen or at least did not mention it with the original description. 5. Since Gossypium barbadense L. can not be typified, the name must be rejected as a nomen ambiguum and must be replaced by G. vitifolium Lam., the earliest correct name. THE SECTIONAL NAMES IN HUTCHINSON’S CLASSIFICATION (1947) OF GOSSYPIUM In 1947, Hutchinson established a classification of Gossypium (in Hutch- inson, Silow & Stephens, The Evolution of Gossypium, pp. 1-53), in which the genus was subdivided into eight sections. The sections, well- delimited morphologically and geographically, represented six main evolu- tionary, cytogenetically distinct groups, one of which was for convenience divided into three subsidiary groups also designated as sections. Apparent- ly the simplicity, clarity, and easy use of this system led to its general adoption by geneticists and cytotaxonomists. Unfortunately, descriptions of the sections in this classification were given only in English, and the names therefore are not validly published. In order to make Hutchinson’s classification nomenclaturally legitimate, I am listing below his sections, along with their correct names and the type species of each. Section I SrurtraNa Hutchinson in Hutchinson, : Stephens, Evolution of Gossypium 16. 1947, nom. illeg. = Sec Br.) Todaro, Relaz. Cult. Cotoni 117. 1877. TYPE SPECIES: G. bs one J. H. Willis (G. Sturtii F. Muell., nom. illeg.). Section II ERIOxyLA Hutchinson, Joc. cit. 18. 1947, nom. illeg. = Section Eri- ocshta (Rose & Standl.) Prokhanov, Bot. Zhur. SSSR 32: 71. 1947. Tyre ECIES: G. aridum (Rose & Standl.) Skovst. Seiticn III Kxorzscu1aNa Hutchinson, loc. cit. 22. 1947, nom. illeg. = Section Integrifolia Todaro, Osservaz. Talun. Spec. Cotone 19. 1863. TYPE SPECIES: G. Klotzschianum Anderss. Section IV TuurBerANA Hutchinson, Joc. cit, 24. 1947, nom. illeg. Thurberia (A. Gray) Prokhanov, Bot. Zhur. SSSR 32: 71. 1947 species: Thurberia thespesioides A. Gray (= G. Thurberi Tod.). eo V Anomata Hutchinson, loc. cit. 27. 1947, nom. illeg. = Section Anomo- ae Prokhanov, loc. cit. 66. 1947. TyPE SPECIES: G. anomalum Wawra = Section TYPE & . Section VE StocxstANa Hutchinson, loc. cit. 30. 1947, nom. illeg. = Section ibe aceon Prokhanov, Joc. cit. 65. 1947. TYPE sPpEcIES: G. Stocksii Mas Section VII Herpacea Hutchinson, Joc. cit. 32. 1947, nom. illeg. = Section Gossypium. LecToTyPE SPECIES: G. arboreum L. 158 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Section VIII Hrrsuta Hutchinson, Joc. cit. 36. 1947, nom. illeg. = Section Magnibracteolata Todaro, Osservaz. Talun. Spec. Cotone 64. 1863. Lrcrto- TYPE SPECIES: G. mexicanum Tod. = G. hirsutum L., fide Hutchinson, 1947; see Prokhanov, Bot. Zhur. SSSR 32: 72. 1947. The union of the sections Erioxylum, Integrifolia, and Thurberia into a single section including all the American diploid species with the genome D, with consequent reduction of the sections to the rank of subsections, as was done by Mauer (1950), seems to me to be an improvement of Hutchinson’s classification of Gossypium. These changes introduced by Mauer are given below. Section Integrifolia Todaro, Osservaz. Talun. Spec. Cotone 19. 1863, amplif. Mauer, Acta Univ. Asiae aon II. 18(Biol. 7): 20. 1950. TyPE SPECIES: G. Klotzschianum Anders Subsection Integrifolia Todaro, Relaz. Cult. Cotoni 188. 1877. (Sect. Klotz- schiana Hutchinson, Joc. cit., nom. illeg.) Subsection Ingenhousia (Moc. & Sessé ex DC.) Mauer, Joc. cit., excl. Erioxylum < Standl. Contr. U. S. Natl. Herb. 13: 307. 1911. (Sect. Ingenhousia (Moc. & Sessé ex DC.) Prokhanov, loc. cit. 72, in part, and sect. Thurberia (A. Gray) Prokhanov, Joc. cit. 71. 1947. Sect. Thurberana Hutchinson, tis cit. 24. 1947, nom. illeg.). Type species: G. trilobum (Moc. & Sessé x DC.) Kearney (/ngenhousia triloba Moc. & Sessé ex DC. Subsection Caducibracteolata Mauer, Joc. cit., emend. Brizicky (Sect. Erioxylum (Rose & Standl.) Prokhanov, loc. cit. 7 a and sect. Ingenhousia (Moc. & Sessé ex DC.) Prokhanov, Joc. cit. 72, in part. Sect. Erioxyla Hutchinson, loc. cit. 18. 1947, nom. illeg.). LECTOTYPE SPECIES: G. armourianum Kearney. Delimitation of the subsections Ingenhousia and Caducibracteolata has been slightly changed in order to bring their limits as close as possible to those of Hutchinson’s corresponding sections. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 159 A SYNOPSIS OF THE ASIAN SPECIES OF CONSOLIDA (RANUNCULACEAE) Puitip A. Munz AS USED IN THE TITLE of this paper, ‘‘Asian” refers to the mainland only and does not include the islands in the Aegean Sea and in the eastern Mediterranean, the plants of which are already treated in Tutin et al., Flora Europaea, vol. 1. 1964, and in Davis, Flora of Turkey, vol. 1. 1965. Material of Asian species has been studied as it exists in the principal herbaria of the United States. Considerable work was done in the library of the Arnold Arboretum and Gray Herbarium of Harvard University. Some months were spent in London working at the British Museum (Na- tural History) and at the Royal Botanic Gardens, Kew, in both of which institutions, as at Harvard, very great courtesy and help were given me. The British Museum secured for my use large loans of herbarium sheets from the Royal Botanic Garden, Edinburgh, and the Komarov Institute, Leningrad. I spent some time also in the Muséum National d'Histoire Naturelle, Paris, where again every courtesy was extended. Smaller loans, largely for individual species, have been obtained from various places and acknowledgment is made in discussion of the species concerned. To all the curators and keepers and to the librarians to whom I am indebted, I ex- press my very great appreciation. In citing herbarium material I employ the symbols given in Index Herbariorum (Lanjouw and Stafleu, Reg. Veg. 31. 1964) as follows: A Arnold Arboretum of Harvard University, Cambridge, Massachusetts, U.S.A BH Bailey Hortorium, Cornell University, Ithaca, New York, U. S. A. M British Museum (Natural History), London, Great Britain cas California Academy of Sciences, San Francisco, California, U. S. A. E Royal Botanic Garden, Edinburgh, Great Britain F Field Museum of Natural History, Chicago, Illinois, U. S. A. CH Gray Herbarium of Harvard University, Cambridge, Massachusetts, US. A K The pacity Royal Botanic Gardens, Kew, Richmond, Surrey, Great Brita LE Herbarium "a the Komarov Botanical Institute of the Academy of Sciences of the U. S. S. R., Leningrad, U. ae MICH University Herbarium, University of Michigan, Ann Arbor, Michigan, U.S.A. MO Missouri Botanical Garden, St. Louis, Missouri, U. S. A. NA United States National Arboretum, Washington, D. C., U.S. A. NY The New York Botanical Garden, New York, U. - P Muséum National d’Histoire Naturelle, Paris, Fra PH Academy of Natural Sciences, Philadelphia, fe eats U. 5..A. 160 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 uc Herbarium of the University of California, Berkeley, California, U. SoA, us Department of Botany, United States National Museum, Smithsonian Institution, Washington, D. C., U. S. A. Because of space limitations the large number of herbarium sheets ex- amined is not cited, but an appendix is given with names of collectors, collection numbers, and the taxa to which referred. It should be em- phasized that the present treatment is not revisional in nature, but rather synoptic. It is concerned with keying out and describing the Asian species, so that identification can be made of plants collected. Therefore, there is no attempt to set up sections within the genus, a thing which should be done only when the African and European species are also considered. It is a pleasure to acknowledge obligations to the National Science Foundation for grant number GB-—2716, which has been of very great assistance in assembling the data herewith presented. THE GENUS CONSOLIDA Lindley (Jour. Hort. Soc. London 6: 57. 1851), in writing about Con- solida aconiti, said “‘Forskahl regarded this curious plant, it is said, as an Aconitum; Linnaeus considered it a Delphinium. In reality it is neither the one nor the other. Its united petals and long sepaline spur, are at variance with the distinct hammerheaded petals and convex back sepal of Aconite. Its petals being reduced to two, and those completely combined into one, equally remove it from Delphinium. That the petaline body is really composed of 2 parts only seems to be proved by its origin, which looks as if opposite the back sepal in consequence of the union of the two contiguous edges of the lateral petals. But it is completely separated from the front sepals, with which it does not in any degree alternate. These considerations lead to the conclusion that the old genus Consolida should be re-established, and by no means confounded with Delphinium proper.” This statement from Lindley seems to me to be about as clear as any that can be made, even now. One wonders why it has taken so many years to recognize Consolida as a genus, but it may well be due to the fact (P. H. Davis, Notes Bot. Gard. Edinburgh 26: 172. 1965) that there was a “widespread belief that Delphinium should be typified by D. consolida L. Delphinium should, in fact, be typified by D. peregrinum L. and Consolida by D. consolida (ie. Consolida regalis S. F. Gray).” At any rate, some of the most recent floras dealing with Eurasian plants, as Tutin et al., Flora Europaea 1: 216-217. 1964, and Davis, Flora of Turkey 1: 119-134. 1965, give generic status to Consolida and I am following their procedure. DESCRIPTION OF THE GENUS ConsoLmpa (DC.) S. F. Gray! Consolida, as a genus, S. F. Gray, Nat. Arr. Brit. Pl. 2: 711. 1821. Asa section of Delphinium, DC. Syst. 1: 341. 1817. As a subgenus of ubescence. Pubescence is a conspicuous and often useful means of characterizing the various species of Consolida. The hairs may be white or yellow. White hairs may 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 161 Delphinium, Huth, Bot. Jahrb. 20: 337. 1895; Dalla Torre & Harms, Genera Siphonogamarum 165. 1901. Plants annual. Leaves palmately laciniate. Inflorescence racemose or paniculate, the pedicels subtended by bracts and each usually bearing 2 bracteoles. Flowers zygomorphic. Sepals 5, petaloid, the posterior one spurred, the lateral pair and lower pair not. “Petal” one, apparently representing the fused superior petals of Delphinium, sometimes called “honey-leaf” or “nectary,” subentire to 3—5-lobed, bearing a single nec- tariferous spur extending into the sepal spur. The lower petals of Del- phinium aborted and not represented. Stamens in 5 spirally arranged series, the filaments more or less broadened, the anthers short. Carpel 1, forming a follicle. Type species, C. regalis S. F. Gray (Delphinium consolida L.). A highly specialized genus of over 40 species, ranging from the western Mediterranean to central Asia in arid and semiarid regions. Several species are cultivated and escape from gardens in many other parts of the world. KEY TO THE SPECIES OF CONSOLIDA 1. Spur and “petal” ia i Saat ag MMe Mia re ea acs 29. C. rugulosa. A, sani and “petal” present “Petal” mostly 5-lobed (usually with 1 generally bifid upper lobe, 2 inter- mediate and 2 lower lateral lobes); spur more or less coiled or circinate at apex; main stem a continued axis through the whole plant with numer- ous, rather short, very slender lateral branches 33 Flowers ca. 23 mm, long, including the spur; sepals blue; pubescence of aerith retrorse-appressed; plant to 5 dm. tall; pedicels 2-8 cm. long “Shane teen A joins ats ee . C. teheranica. : Flowers er (-20) mm. long, including the spur Spur uncinate-saccate to almost straight; intermediate lobes of etal” sharply triangular Es ic mm. long; spur uncinate, 4-5 mm. 3 peat ccata ios) 5. Follcies = 10 mm. long; spur almost straight, 8 veh es aS wm SO c t ° ee mt fs) me i] i) cot ze a ate i=} = ,2 ae a) = o ~ = pre izZ r) a! =¥ =e 38 o Q. rt) 2 o wad o S o un o hh oo Ss Oo shortly bifid upper lobe; seat 8-10 mm. ge 3-4 mm. wide E. Turkey to Iran, Jordan. .......------- be appressed (usually retrorsely so) or spreading, and in the latter case, may be hee tipped or not. But the yellow hairs are quite different cg orig se I hav -_ sa mos sea a inarilv they are not gland-tipped, but are strongly en . apeingenrn namie Naked : in nature, since agen collection labels as I can see, then, e gland- toward the ste 162 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 6. Intermediate lobes of “petal” shorter than or subequal to the upper lobe. 7. The intermediate lobes of the “petal” acute. 8. Lower lobes of “petal” more or less approximate to the intermediate lobes, ovate; intermediate lobes broadly triangular; follicle semiovate, 6-9 mm. long, 2-3 m wide. Anatolia to Lebanon. ........ 2. C. scleroclada. 8. Lower lobes of “petal’’ separated from the intermediate lobes by a broad sinus; intermediate lobes narrowly triangular, sharply acute; follicle linear to linear-oblong. 9. Follicle linear-oblong with rounded abruptly rostrate apex, 10-13 mm. long; stem appressed-canescent. Central Pei i: ee 5. C. stenocarpa. Follicle linear, with gradually rostrate apex, 15-22 mm. long; stem with whitish spreading hairs. Dar- Ae Gea MMe eR ee eae Oa iia 6. D. aconiti. 7. The intermediate lobes of the “petal” obtuse. 10. Stem velvety with whitish spreading hairs mixed with glandular ones; sinus between lateral and intermediate lobes of “pet tal” ca. %4 as wide as deep. Thrace to bs Cee Pen oy on a ae e 3. C. thirkeana, 10. Stem strigulose: sinus between lateral and intermediate lobes of “pe ious ca. 4% as wide as deep. Transcaucasia, Arm a B21 aeans toy i a mae PE tr . hohenackeri. 2. “Petal” almost entire to 3 ieee (with 1 often bifid upper lobe and 2 lateral lobes); habit not as above, the stem either branched below or not diffuse above; “petal’-spur not circinate at apex. 11. Pedicels shorter than sepals at anthesis, sometimes elongating in fruit; bracteoles minute, often looking like additional lobes to the subtending ‘bract; stem ~epiee = flowers often from most axils; follicle nearly straight, subtere 12. The lateral lobes of the “petal” at right angles to the terminal lobe, longer than wide; spur 5-9 mm. long; flowers deep violet to blue or rose. 13. vig lateral lobes ca, 1.5 mm. long, the terminal lobe ca mm, long; sepals 13-18 mm. long. Turkey (Caria and tee through Phrygia and Cappadocia to ype Sg A EU A aati ont oils Gace ee rey Sa WE RUNES Tap Seale Eta 10. C. raveyi. 13. The lateral lobes of the “petal” 4-5 mm. ay about equal to the terminal lobe; sepals 10-13 mm. lon 14, “Petal” more or ‘Tess cruciate, the Gina lobes abruptly narrowed to the base. Turkey CO a Pema Se elas ur uterine Sethe de iMacs 0 11. C. cruciata. . “Petal” more kite-shaped, more or less triangular, since the lateral lobes os narrow to the base. South- eastern Anatolia to Syri 15. Bracteoles multifid; aon 10-13 mm. long; blade of “petal” 12 mm. ‘wide: Turkey to Syria. ...... eevee ee oe 12. C. axilliflora. +s. Bracteoles entire; sepals 6-8 mm. long; blade Of “petal” 10 sam: wide. Syria... 52 | os Say betel, Clee Coase 13. C. gombaultii. — > 1967 | 11. MUNZ, ASIAN SPECIES OF CONSOLIDA 163 — nN . The lateral lobes of the “petal” directed somewhat forward toward the apex of the “petal”; spur 16-25 mm. long; flowers whitish to pinkish. 16. Expanded “petal” 15-20 mm. wide. i sah 17. Sepals 9-11 mm. long; follicle 17~25 mm. long; main cauline leaves entire. Turkestan and pares Tran, wes RT ES otis Gein anit ea yA leptocarpa, 17. Sepals 6-8 mm. is main cauline eves divided, iran, Jean; Parkestan 20. 24. C. persica. Expanded “petal” 8-11 mm. 18. The expanded “petal” wider near the Popes somewhat narrowed toward the tip. aoe 8. C. halophila. The expanded ‘‘petal” HOE up ade 19. Lateral lobes of “petal” praca rounded, not at all triangular; ear “petal” 11 mm. wide. ‘Durkestan: to Iran: «5655655, 14. C. stocksiana, Lateral lobes of ‘ “petal” more or less triangular; expanded ‘petal’ 8 mm. wide. Transcaucasia to Pi gPAni sets hotles opi eas 27. C. rugulosa, — = rr = Pedicels usually longer than sepals. 20. Bracteoles high on the pedicel, hence ge above the base of the flower; leaf-lobes usually very narrow, ca. 0.5 mm. wide; stems widely and few-branched; eal: lobes of the “petal” more convex on upper edge than on lower. 21 | Peta ee and subcircular; flowers oo in a raceme. bdo _ Re ANDO Gore cats es . C. olopetala, . “Petal” more or less lobed; flowers ee oe 3-6 in a raceme. Flowers pole onciti terminal = of “petal” 3- toothed, Turkeg. 6 3 Akai 9. C. sulphurea. . Flowers violet to blue or rose. 23. Spur shorter than blade of sepal, or if equal to it, plant with spreading hairs 24. Follicle pendent on a recurved pedicel; plant strigose. Eastern Turkey, Pleas aes Pine Niketan mma eee ier . C. oligantha, 24, 4 poesia erect; plant with alee pubes- bo bo bo bo bo on . ieee rose; spur straight, 5-6 mm, long. Lebanon and Syria Sapa 3 Pige ey ealeae cs C, pusilla, Flowen violet. 26. Spur inate straight, 6-7 mm. long; “petal” terminal lobe 3 mm. ee Seok eed bea ey = ae O4 ee gira, tS _ pur more or less curve long; “petal” with 3 minute per lobes. Iran. .... 30. C. limarioides. 23. Spur equal to or longer than the blade of the sepal, if shorter, the plant with appressed hairs. bo de mM i=} a JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 27. Pubescence spreading; spur ca. twice as long as upper sepal. Iran. .. 29. C. trigonelloides. Pubescence appressed. 28. Spur 8-10 mm. long; sepals with a basal claw 3-4 mm. long. Lebanon, Jordan. me sro ea eat derepeas 36. C. tomentosa. Spur 12-24 mm. a sepals sessile or ong. at, ~I nN O° with a claw ca. 1 m “Petal” cee narrowed at tip into a beaklike upper lobe; rest at “petal” dome-shaped when spread open; spur 18-24 mm. long. ...... igh Dadi ie et 38. C. armeniaca. “Petal” either gradually narrowed toward the apex or the upper lobe subequal to the lateral lobes; spur 12-46 was. lang se es Pe Cee ns 37. C. hellespontica. 20. Bracteoles usually lower on the pedicel and not extending above the base of the flower 30. Lateral lobes of “petal” more convex on upper edge than on lower and with the subtriangular terminal lobe making ~) eo nN Ned lateral branches; follicle compressed, somewhat curved, ca. m. long, the style 4 mm. long; spur much longer than 34. C. camptocarpa. lobes of “petal” perpendicular to the terminal lobe or pointing somewhat toward the tip of the “petal’’ when oe oO ree Oo Boe. & o f a 8 2 es 31. Follicle mostly at least 15-25 mm. long, oblong-cylin- dric; style scarcely 1 mm. long; main stem forming an axis into the inflorescence with few side branches; leaf segments very narrow, ca. 0.5 mm. wide. 32. Bracteoles usually below the middle of the pedicel; sepals 12-14 mm. long; spur ca. 15 mm. long. 15. C. ambigua. . Bracteoles above the middle of the pedicel; sepals ww bdo 33. Spur ca. 18 mm. long. . 25. C. deserti-syriact. 33. Spur 6-10 mm. long. 34. The spur 6-7 mm. long; follicle 10 mm long. Kashmir. .. 31. C. schlagintweitit. 34. The spur 9-10 mm. long; follicle 15 mm. or more long. 35. Median lobe of “petal” longer than wide, the 2 apical lobules rounded; sepals 10-12 mm. sab UF pee eee Saas dete ete. C. orientalis. . Median lobe of poe about equal- ly long and wide, the 2 apical lobules wW un 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 165 pointed; sepals 6-9 mm. long. .... reins Soraemaeeer 32. C. phrygia. 31. Follicles 6-15 mm. long; style longer; plants mostly branched from the base with widespread open habit. 36. Spur 10 mm. or more long, exceeding the sepals. 37. Lateral lobes of ‘‘petal” at right angles to the middle lobe, not directed toward the tip of the “petal”; stems strigulose; lower bracts CL Sl ne ae 1 . regalis, 37. Lateral lobes of ‘petal’ directed forward somewhat toward the tip of the “petal”; stems with some spreading hairs in upper parts. 38. The lateral lobes of the “petal” some- what pointed and semilunate; spur 10- 13 mm. long. Armenia. .. 26. C. cornuta. The lateral lobes of the “petal” rounded; spur mostly longer. 39. Sepals 10-14 mm. long, narrowed into a claw-like base; bracts lacini- ate; pedicels glabrous except at the summit. Iraq. .... 20. C. oliveriana. Sepals a mm. long, scarcely nar- rowed at the base; bracts entire; i seid iene r- 19. C. glandulosa. 36. Spur 5-7 mm. Nee about as long as the sepals. ue oN) © long; lateral lobes of “petal” rounded, not projecting beyond ee middle lobe; filaments glabrous. Asia Minor. ........ 8. C. incana. 40. Flowers yellow, ae the ‘ ‘petal”’; sepals 4-5 mm. long; lateral lobes of the “petal” semiovate, as long as the median lobe; fila- ments hai 41. Stems strigose, at least below, little or not at all glandular. 21. C. flava . Stem with more or less ‘spreading and glandular hairs. ........ 22. C. deserti. aE _ 1. Consolida teheranica (Boiss.) Rech. f. Ann. Naturh. Mus. Wien 51: 376. 1941 Fic. 1, A; Delphinium teheranicum Boiss. Fl. Orient. 1: 85. 1867. ? Delphinium intricatum Pau, Trab. Mus. Ci. Nat. Bot. Madrid 14: 12. 1918. Annual, to 5 dm. tall, openly and widely branched, retrorse- strigulose throughout, the branches slender, subdivaricate; leaves largely gone by anthesis, multifid into acute linear segments; inflorescence open with scat tered flowers, strigulose and with some glandular spreading hairs with basal yellow pustules: lower bracts largely tripartite, to ca. 1 cm. long; the upper subulate, to ca. 5 mm. long; pedicels slender, 2-8 cm. long; 166 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 bracteoles 2 or more, subulate, scattered, well below the flower, ca. 2 mm long; calyx blue, pubescent without, 8-10 mm. long, not unguiculate, the upper sepal lance-ovate, with a circinate-involute pubescent spur ca. 13- 16 mm. long and 4-5 mm. wide at the base; lateral sepals oblong-ovate, blue-membranous except on the narrow central pubescent strip, obtuse; lower sepals narrow-ovate, obtuse, more pubescent; “petal” ca. 23 m long, including the spur, the terminal lobe subentire, broad, subcrenulate, longer than the intermediate triangular lobes (scarcely 1 mm. long), the membranous lower lobes grown to the intermediate lobes to near the rounded summit; stamens subequal, the filaments ca. 6 mm. long, dilated below, pubescent above; anthers dark, almost 1 mm. long; follicle sub- compressed, somewhat reticulate-veiny, rather straight, 12-15 mm. long, 3—3.5 mm. wide, the style an additional 2 mm.; seeds dark, asymmetrically subcylindrical, almost 2 mm. long, with several continuous transverse rows of short scales. Type: “prope Teheran Persiae ad pagum Gulkak”, Kotschy 884 (w). Not se pean Stony places at 1200 to 1300 m., Elburz Mts., n. Iran. REPRESENTATIVE SPECIMENS: Rechinger 2015 (x, ay Furse 2907 (X). 2. Consolida Sam Boca Schrod. Ann. K. K. Naturh. Hof- mus. Wien 27: 44. Delphinium ee Boiss. Diagn. I. 8: 8. 1849. Delphinium anthoroideum Boiss. var. sclerocladum (Boiss.) Boiss. Fl. Orient. 1; 85. 1867. One to 5 dm. tall, variously branched and pubescent, the ultimate branchlets very slender; leaves multifid into linear lobes, largely gone by anthesis, roundish in outline; bracts multifid, 2-4 mm. long; pedicels to ca. 1 cm. long; bracteoles lance- linear, 2 or more, largely remote from the flowers, to ca. 2 mm. long; sepals yellowish- white to lilac, 7-8 mm. long, not unguiculate, narrow-oblong, ca. 1.5 mm. wide, hyaline on margins, pubescent on median strip; sepal spur circinate-involute at apex, ca. 7 mm. long, 3.5 mm. wide at base; “petal” 5-lobed, 6—7 mm. long, exclusive of the spur, the terminal lobe ca. 2.5 mm. long, broad, entire; intermediate lobes triangular, acute, slightly shorter; lower lobes Sebrious more or less approximate to the intermediate lobes, ovate, ca. 3.5 mm. long; filaments —8 mm. long, inflated in lower part, more or less pubescent; anthers yellow, ca. 0.6 mm. long; follicle semiovate, compressed, glabrous, reticu- late-veined, 6—9 mm. long, 2-3 mm. wide, the style an additional 2.5—3 mm. Pedicels puberulous with interspersed spreading glandular hairs; plant more or ree lalate rigid up to the ultimate racemose bran chin ng. Turkey to DVO ee ee en ne Var. scleroclada. Pedicels strigulose, without glandular hairs, sometimes glabrescent. Branches ascending, the ultimate ones risid or slender. Turkey, Syria. ... ad ee ow es ay ee eee eos Var. rigidd. Branches, especially the basal ones, widely divergent, the ultimate branches slender. Syria; Lebanon 045) oo) ue hey ea eta Var. pumila. 1967] MUNZ, ASIAN SPECIES OF CONSOLIDA 167 Consolida scleroclada (Boiss.) Schréd. var. scleroclada Fic. 1, B. Stem puberulous, with spreading glandular hairs among the others: branching racemose, the contour more or less pyramidal, rigid to the ulti- mate branches “in montibus demissioribus Syriae borealis inter Latakieh et pra ” Boissier in July (c); not seen DIstRIBUTION. Dry slopes below 1700 m., northern Syria, southern Turkey (Cilicia). PECIMENS SEEN. Turkey: Balansa 730. Syria: Pinard, 1846; Boissier, June 1866; Sintenis 1416. Consolida scleroclada (Boiss.) Schréd. var. rigida (Freyn & Sint.) Davis, Fl. Turkey 1: 123. 1965. oo anthoroideum Boiss. var. rigidum Freyn & Sint. Ost. Bot. Zeitschr. 41: 1891. aes euphratica Schréd. Ann. K. K. Naturh. Hofmus. Wien ad: 43. 1913. Delphinium sclerocladum Boiss. var. pseudo-anthoroideum Davis & Hossain, Notes Bot. Gard. Edinb. 22: 415. 1958. Type, Davis 10030 (£) seen. Strigulose; branches ascending, the ultimate slender or rigid. Type: Armenia Turcica, Chama ad Euphratem, Ichtik prope Tuzla, Sintenis 2969 (w); not seen. DistriBuTION, Dry slopes and steppes, 900-1700 m., southern Turkey to western Syria and the Syrian desert. SPECIMENS EXAMINED. Turkey: Post, 1884; Siehe 387; Davis 30973, 31688, 31592, 35878; Bornmiiller 3221. Syria: Davis 6477, 10030, 9992. Consolida scleroclada (Boiss.) Schréd. var. pumila (Huth) Munz, mb. nov Delphinium anthoroideum Boiss. var. pumilum Huth, Bot. Jahrb, 20: 366. 1895. Delphinium sclerocladum Boiss. var. pumilum Hossain & Davis, Notes Bot. Gard. Edinb. 22: 414. 1958. Branches, especially the basal, widely divergent; stem 4-20 cm. high; lower leaves rather persistent. Lectotype: Syria, Zebdaine near samc at 5500 ft., June 6,.1858, Kotschy 103 (pm); other sheets (K, U DistriBuTION. At 1200 to 1800 m., pies Syria, Lebanon. SPECIMENS sEEN. Kotschy 103; Bornmiiller 32; MacDaniels, 1944; Born- miller 11324. 3. Consolida ae (Boiss.) Schréd. Abh. Zool.-Bot. Ges. Wien 4(5): 62. Fic. 1, C. Delphinium thirkeanum Boiss., Fl. Orient. 1: 84. 1867. Two to 3 dm. tall, with rather rigid lateral branches, white spreading Sahai ane interspersed with yellow glandular hairs with pustular bases; 168 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 FIGURE 1. Consolida species with “petal” of the 5-lobed type, as seen from the side, the lamina base at the bottom of the dotted line (main vein) and of min spur combined 13 mm. long; from Wiedemann (cH). D. C. hohenackeri, lamina plus spur 11 mm.; from Bourgeau 7 (uc). E. C. stenocarpa, lamina and spur 14 mm.; from Balansa 857 (uc). F. C. aconiti, lamina and spur 20 mm.; Sintenis 678 (pm). G. C. anthoroi i m.; fro 18498 (micH). H. C. saccata, lamina and spur 14 mm.; from Sintenis 1186 (pH). I. C. barbata, lamina only, front view with broad terminal lobe, sharp narrow intermediate and basal lobes; from Gontscharow & Grigoriev 871 (NY). leaves sessile, small, largely gone by anthesis, laciniate with slender crowded lobes; flowers few, scattered; sepals pale blue, 6-7 mm. long, not unguiculate, the lateral lance-oblong, ca. 2 mm. wide, membranous except for the narrow central strip; lower sepals much the same, asym- metrical; upper sepal, including the spur, ca. 13 mm. long, pubescent, the spur circinate-involute, ca. 6 mm. long, 4 mm, wide at base; “petal” 5- lobed, pale with violet tip, ca. 13 mm. long, including the spur, the terminal lobe shortly bifid with rounded parts; intermediate lobes shorter, obtuse; lower lobes membranous, rounded at tip, asymmetrical, ca. 3.5 mm. broad, separated from the intermediate by a very narrow sinus; stamens to ca. 6 mm. long, the filaments dilated below, pubescent above; anthers ca. 0.6 mm. long; follicle semiovate, more or less compressed, glabrous, somewhat reticulate, the rigid style 2 mm. long; seeds with short scales in continuous bands. SyntTyPEs: “in Bithynia (Thirke!)” and “c. Amasya et Tokat, Wiede- man.” Both the collections are at Geneva, but I have been unable to see them. I have seen the Wiedeman specimens (cH, Kk). DistriBuTION. Turkey, from Thrace to Galatia, growing at 500 to 1000 m. ILLustRaATion of “petal”. Davis, Fl. Turkey 1: 127. 1965. COLLECTIONS SEEN. Turkey: J. Ball, Oct. 13, 1867; Sintenis 4882; Uvarov 97; Bornmiiller 13696. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 169 4. Consolida hohenackeri (Boiss.) Grossheim, Fl. Kavkaza 2: 101. 1B; Delphinium hohenackeri Boiss., Fl. Orient. 1: 85. 1867. Aconitopsis hohenackeri (Boiss.) Kemularia-Nathadze, Trudy Inst. Bot. Tbilisi 7: 125. 1940. One to 6 dm. tall, freely branched more or less throughout, retrorse- strigulose except that the upper parts may be subglabrous or with a few hairs with glandular yellow base; ultimate branchlets slender; basal leaves largely gone by anthesis; cauline leaves mostly on lower stems only, lacini- ate into linear strigulose lobes, subsessile, the leaf-blades largely 1—2 cm. long; bracts mostly entire, subulate, 3-10 mm. long; pedicels 1-3 cm. long, mostly with 2 subulate bracteoles below the flower; inflorescence very open; flowers purple to violet or pale mauve, ca. 1.5 cm. across; sepals lanceolate, ca. 7 mm. long, the spur with a circinate-involute tip and ca. 7 mm. long and 3.5—4 mm. wide at the base, pubescent; lateral sepals lance- elliptic, 3.5-4 mm. wide, mostly membranous except for the strigulose median strip; lower sepals narrower, much like the lateral; “petal” 5- lobed, ca. 11 mm. long, including the spur, the terminal lobes subentire, broad, ca. 1.5 mm. long; the intermediate lobes obtuse, ca. 1 mm. long, the lower separated from the intermediate by a broad sinus, membranous, rounded-oblong and ca. 4 mm, long; filaments 4-5 mm. long, dilated only below, more or less pubescent above; anthers yellowish, ca. 0.7 mm. long; follicle glabrous, semiovate, reticulate-veiny, 6-7 mm. long, 3-4 mm. wide, the beak an additional 3 mm. long; seeds with longish separate scales. Lectotype: Turkey: Prov. Giimiisane, near Baibout in Armenia, July 12, 1862, Bourgeau 7 (xk); this number seen also (£, P, UC). DistriBuTion. Dry places at 900-1800 m., Turkish Armenia, Transcau- casia, Iran. : ILLUSTRATION. “petal”, Hossarn & Davis, Notes Bot. Gard. Edinb. 22: 416. fig. E. 1958; Davis, Fl. Turkey 1: 127. fig. 1. 1965. REPRESENTATIVE SPECIMENS, Transcaucasia: Wittman 93; Radde 307. Turkey: Sintenis 1336, 7110; E. K. Balls 1441A; Davis 30825, 32613, 30076; Holmberg 2408. Iran: Haussknecht, 1868; Stapf 2771. 5. Consolida stenocarpa (Hossain & Davis) Davis, Notes Bot. Gard. Edinb. 26: 173. 1965. Fic. 1, E. Delphinium stenocarpum Hossain & Davis, Notes Bot. Gard. Edinb. 22: 413. 1958. Slender-stemmed, divaricately branched throughout, 2.54 dm. tall, densely retrorse-strigose below and with some spreading hairs with swollen yellow base, sparsely pubescent above; leaves laciniate into linear lobes, strigulose, the lower with petioles ca. 1 cm. long and blades 1-1.5 cm. long, the lower cauline subsessile, rapidly reduced up the stem; bracts sessile, entire to lobed, 3—7 mm. long; pedicels strigulose, 4-6 (-15) mm. long; bracteoles subulate, 1.5-2 mm. long, alternate to subopposite; flowers widely scattered, bluish lilac to pale dirty mauve, sepals 6-7 mm. long, 170 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 elliptic, not unguiculate, ca. 3 mm. wide, acute, pubescent and greenish on midrib in the lateral sepals, more generally pubescent on the lower pair; sepal spur pubescent, 6-7 mm. long, 3.5 mm. wide, curved-oblong, circinately involute at the tip; “petal” 5-lobed, 7-8 mm. long, the spur an additional 5-6 mm. long, upper lobe 1.5—-1.75 mm. long, bluish, bifid half way into rounded lobes; intermediate lobes narrowly lance-deltoid, 1 mm. long; lower lobes membranous, rounded-oblong, ca. 2.5 mm. long; filaments sparsely pubescent on the inflated lower part; follicle glabrous, linear-oblong, 10-13 by 1.5—2 mm., the style an additional 3—3.5 mm.; seed with short scales. Type: Turkey: Prov. Konya 5 between Agabeyli and Korku- yu, Davis 16638 (£), isotype (K); DistTRIBUTION. Fallow fields and wane, 900-1300 m., central Anatolia. ILLUSTRATIONS. “‘petal’’; Hossatn & Davis, Notes Bot. Gard. Edinb. 22: 416, fig. B, C. 1958; Davis, Fl. Turkey 1: 127. fig. 1. 1965. SPECIMENS SEEN. Turkey: Balansa 857; Godfrey & Taysi SH-25; Davis 32810, 16643. 6. Consolida aconiti (L.) Lindl. Jour. Roy. Hort. Soc. 6: 55. 1851. Fic. 1, F. Delphinium aconiti L. Mantissa 1: 77. 1767. Aconitum monogynum Forsk. Fl. Aegypt-Arab., p. xxvii. 1775. Stiffly much branched, 2—5 dm. tall, with soft, white, spreading pubes- cence, subglabrate above; cauline leaves short-petioled, roundish in outline, to 2 or 3 cm. long, with many ultimate linear pubescent lobes; lower bracts leaflike, smaller, 3—5-partite, the upper reduced; pedicels very slender, to 3 or 4 cm. long; bracteoles linear, remote from fiower, alternate, 3-4 mm. long; flowers few on open branches; sepals ca. 5 mm. long, pale violet with darker median, more or less greenish stripe pubescent without; spur 14- 15 mm, long, circinately involute at the apex; “petal’’ 8-9 mm. long, pale yellow with violet tip, the upper lobe exceeding the acute, narrowly triangu- . lar intermediate lobes, the lower lobes separated from the intermediate by a broad sinus; stamens purplish, 5-6 mm. long; anthers dark, oblong, ca. 0.6 mm. long; follicle linear with gradually rostrate apex, 15-22 mm. long, 1.5-2 mm. wide, glabrous, veined-reticulate; seeds few, pyramidal, with elongate scales. Type: “Habitat in Dardanella, Forskdl; photograph of specimen in the Linnaean Herbarium seen. DIsTRIBUTION. Below 1100 m., region of the Dardanelles, Turkey. ILLUSTRATIONS. VAHL, Symb. 1: ¢. 13. 1790; Davis, Fl. Turkey 1: 217. 1965. SPECIMENS. Turkey: Aucher 66; Sintenis 678; Guichard, July 29, 1962. 7. Consolida anthoroidea (Boiss.) Schréd. Abh. Zool.-Bot. Ges. Wien 4(5): 62. 1909 Fic. 1, G. Delphinium anthoroideum Boiss. Ann. Sci. Nat. II. 16: 369. 1841. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 171 Delphinium acutilobum Turrill, Kew Bull. 1929: 223. 1929. Based on Gilliat- Smith 2086 («) from near Yam, Tabriz district, Iran; not seen. Erect, rigidly divaricate, to 8 dm. tall, subglabrous to strigulose below, short spreading-pubescent above and with many yellow-based hairs; leaves 3- to many-parted, the ultimate lobes linear to oblance-linear, the main cauline blades subsessile, ca. 3.cm. long and 5 cm. wide, strigulose; bracts mostly 3-lobed, 7-15 mm. long; pedicels very slender, 1.5—6 cm. long; bracteoles 2—3 mm. long, lance-linear, often more than 2; inflorescence very open, the flowers remote; sepals violet to pale lilac, with greenish tinge toward the apex, pubescent, 6—7 mm. long, not clawed; lateral sepals oblong, rounded apically, 2-3 mm. wide, pubescent along midrib, the 2 lower slightly wider; sepal spur circinately involute at apex, erect, pubes- cent, ca. 1 cm. long, 5 mm. wide at base; “petal” 5-lobed, more or less rose, the blade 6 mm. long, the upper lobe short, retuse, shorter than the elongate, triangular, divaricate intermediate lobes, the lower lobes mem- branous, elongate-rounded, almost 4 mm. long; filaments dilated below, slightly pubescent; anthers yellow to orange; follicle glabrous, subcom- pressed, oblong-obovoid, 8-10 mm. long, 3-4 mm. wide, abruptly nar- rowed into a style an additional 1 mm. long; seed pyramidal, densely squamate. Type (lectotype by Davis): Syria, Aucher 65 (G), not seen, but isolecto- types seen at K, P. DistRIBUTION. Fallow fields and stony slopes, 600-2000 m., Kurdistan in Turkey, Iraq, Iran, to Syria, Lebanon. InLustRATIoNs. Hooker, Icon. Pl. V. 2: pl. 3151. 1932; Davis, Fl. Tur- key 1: 127. 1965. REPRESENTATIVE SPECIMENS. Turkey: Aucher 4032; Krause 3828, Balansa 857: Davis 23802. Iran: Gilliat-Smith 2365, 2388; Koelz 18498; Furse 3282, 3283, 3254, Lebanon: Berton 217. Syria: Gaillardot 389; Dinsmore 13024. 8. Consolida saccata (Huth) Davis, Notes Bot. Gard. Edinb. ae 1965. » Ti. Delphinium saccatum Huth, Bull. Herb. Boiss. 1: 325. 1893. Rigid, much branched, several dm. tall, the branches divaricate, slender, largely retrorse-strigulose, but with more or less spreading, short, glandular hairs near the tips; leaves of stem palmately multipartite, to ca. 3 or 4 cm. long, subsessile, the segments linear, strigulose, rather few; inflores- cence open, with widely scattered, pale yellowish to blue or white flowers; bracts largely 3—5-parted, to ca. 5 mm. long; pedicels very slender, 0.5—3 cm. long; bracteoles entire, several, scattered, to ca. 2.5 mm. long; sepals ca. 10 mm. long, lance-oblong, ca. 3 mm. wide, with broad white hyaline margin, the lateral pair pubescent along the median strip, lower sepals less hyaline and more pubescent, the upper sepal and spur pubescent, the latter 4-5 mm. long, 2.5 mm. wide at base, hooked-saccate, not coiled, at apex; “petal” pale with bluish tip, 5-lobed, ca. 15 mm. long, including 172 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 spur, the terminal lobe ca. 3 mm. long, divided half its length, intermediate lobes 1.5—2 mm. long, sharply triangular, outer lateral lobes membranous, 4-5 mm. long, triangular, acute; filaments dilated at base; follicle com- pressed, glabrous, reticulate-veined, 12-15 mm. long, 3—4 mm. wide, the style an additional 3-4 mm. long. Type: Turkey, Mardin, rocky fields near Rischemil, Sintenis 1186 (B); the holotype now probably destroyed; isotypes seen at K, MO, P, PH. DisTRIBUTION. Stony fields, Turkey (E. Mesopotamia, Mardin) ; north- ern Iraq. ILLustrations. Hut, Bull. Herb. Boiss. 1: pl. 16, fig. 3. 1893; Davis, Fl. Turkey 1: 127. 1965. REPRESENTATIVE SPECIMENS, Iraq: Haines W 1218; Agnew, Sept. 1, 1961. 9. Consolida barbata (Bunge) Schriéd. Abh. Zool.-Bot. Ges. Wien 4(5): 16, 62. 1909. Fic. 1, I. Delphinium barbatum Bunge, Arbeit. Nat. Ver. Riga 1: 126. 1847. Delphinium barbatum Bunge var. hirsutum Huth, Bot. Jahrb. 20: 391. 1895. Apparently based on a collection at 800 m., Mt. Karatag, Regel in 1882, Turkestan Herb. (LE), which I have not seen. Two to 8 dm. tall, retrorse-strigulose, erect, the main stem with many short, slender, ascending, side branches throughout its length; leaves nu- merous along main stem, dissected into linear segments 0.3-0.4 mm. wide, the principal cauline leaves sessile, 1-2 cm. long; bracts divided into linear segments or the uppermost entire, 2-6 mm. long; pedicels 4-15 (-30) mm. long, very slender, subglabrous to glandular-pubescent with some hairs yellow and enlarged at the base; bracteoles several, lanceolate, 1—2 mm. long; sepals bearded without, blue, 6-8 mm. long, narrow-ovate, acute, ciliate-bearded at distal margin, the spur ca. 8 mm. long, blue, stri- gulose or glandular-pubescent, ca. 2.5 mm. wide at base; ‘petal’ whitish with bluish tip, 5-lobed, ca. 8-9 mm. wide, the middle lobe broadly rounded, ca. 4.5 mm. wide, the 2 lateral lobes lance-deltoid, slightly shorter than the median lobe, the outer lobes more membranous, lanceolate, acuminate, ca. 5 mm. long; filaments pale blue, 5-6 mm. long, glabrous; anthers blue, ca. 0.4 mm. long; follicle glabrous, oblong, 5-10 mm. long, reticulate-veined, the style an additional 2.5—3 mm.; seed ca. 1.5 mm. long, narrowly obpyramidal, with transverse rows of scales. Type: Karatau Mts. in upper Sarafschan near Samarkand, Turkestan, Oct., 1841, Bunge. The Bunge collection at LE is probably the type and has “Lehmann 38” on it, as does a specimen at P. I take it this means from the Lehmann herbarium. The Le specimen has a single flower and no leaves; the sepals are ciliate and hide the “petal”; the pedicels are glandu- lar-pubescent. DistTRIBUTION. Dry places at 700—2000 m., Turkestan. REPRESENTATIVE SPECIMENS. Capus 34; Regel 52, 1881; Komarov, Aug. 18, 1893; Gontscharow & Grigoriev 871. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 173 10. Consolida raveyi (Boiss.) Schréd. Abh. Zool.-Bot. Ges. Wien 4(5): 62. 1909. Fic: 2, A. Delphinium raveyi Boiss. Diagn. I. 1: 66. 1843. Erect, simple or more often with few upwardly curved branches from the base, 1-4 dm. tall, strigulose and with short spreading white hairs and many yellow pustular-based hairs, leafy throughout; basal leaves short- petioled, ternate, then further divided into linear pubescent segments; lower petioles to 2 or 3 cm. long, the upper ones gradually shorter; lower blades 1-3.5 cm. wide, upper smaller, passing into trifid leafy bracts ca. 1 cm. long or the uppermost linear, entire, green; flowers solitary in most leaf-axils, erect, sessile, each with 2 linear bracteoles at base so that there is a resemblance to a trifid bract; sepals deep violet, ca. 13—18 mm. long, without the spur; upper sepal linear, ca. 2 mm. wide, pubescent on most of outer surface, the pointed spur to ca. 5 mm. long; lateral sepals pubes- cent on the midrib and narrow claw, the limb ca. 4 mm. wide; lower sepals with broader claw, the limb 3 mm. wide; “petal” 3-lobed, 12-14 mm. long, 3-4 mm. wide, the terminal lobe divided into 2 small lobules 1 mm. long, the lateral lobes ca. 1.5 mm. long; stamens 5—7 mm. long, the filaments oblong-dilated below, narrow above, more or less pubescent; anthers yellow, ca. 0.8 mm. long; follicle reticulate-veiny, soft hairy, 10-12 mm. long, the beak an additional 1.5-2 mm., the body 3 mm. wide, sub- compressed, few seeded; seeds triquetrous, scaly. Type: “in cultis Cariae prope Aphrodisias” [Geyre], Bossier, June 1842 (c not seen); isotypes (K, NY, P) seen. DistTRIBUTION. Steppes, fields, etc., 700-800 m., Turkey from Caria and Lycia through Phrygia and Cappadocia to Paphlagonia. REPRESENTATIVE COLLECTIONS. Turkey: Pichler 22; Bourgeau 8; Balansa 599, 719, 856; Bornmiiller 1719, 4027, 13690; Siehe 175; Sintenis 4293; Davis 13010, 36622. 11. Consolida cruciata (Hossain & Davis) Davis, Notes Bot. Gard. Edinb. 26: 174. 1965. Fic. 2, B. Delphinium cruciatum Hossain & Davis, Notes Bot. Gard. Edinb. 22: 422. 1958. Erect, 1.5—4 dm. tall, simple or with few upwardly curved branches from base, strigulose; cauline leaves with petioles 0.5-1.5 cm. long, the blades ca. as long or longer, ternately laciniate into strigulose linear segments; flowers subsessile in most axils, lilac to pink, drying whitish, 10-15 mm. long; pedicels very short, thick; bracts foliaceous, partite, gradually re- duced up the stem; bracteoles 2, trisect, equaling flowers; sepals pubescent, the upper 10-12 mm. long, oblong-spatulate, narrowed into a broad basal claw; the 2 lateral ca. 2.5 mm. wide, the 2 lower rhombic-spatulate, obtuse, clawed; spur pubescent, 5—7 mm. long, ca. 2 mm. wide at base, slightly curved, pointed; “petal” 3-lobed, cruciate, 11.5-12.5 by 9.6-11 mm., purple-nerved, more or less abruptly attenuate at base; terminal lobe 174 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 oblong, 3-5 mm. long, with 2 small rounded lobes at tip; lateral lobes rounded-oblong, flaring; stamens unequal, filaments oblong-dilated below, linear and pubescent above; anthers purplish brown, ca. 1 mm. long; follicle erect, 12-17 by 2-3 mm., terete, strigulose, more or less furrowed longitudinally when mature, the style an additional 1.5—3 mm. Type: Turkey, Prov. Adana (Cilicia), Bozanti at 750 m., Siehe 362 (E); isotypes (GH, K, P) seen. ISTRIBUTION. Cilicia, Mersin, Turkey. ILLustraTions. Hossatn & Davis, Notes Bot. Gard. Edinb. 22: 422. fig. 4, C & D. 1958; Davis, Fl. Turkey 1: 127. 1965. 12. Consolida axilliflora (DC.) Schréd. Abh. Zool.-Bot. Ges. Wien 4(5): 62. 1909. Fide. 20. Delphinium axilliflorum DC. Syst. 1: 341. 1817. Delphinium axilliflorum DC. var. violaceum Boiss. Fl. Orient. 1: 75. 1867. With violet flowers; several collections cited; some seen by me. Delphinium axilliforum DC. var. coerulescens Boiss. loc. cit. Flowers pale gray-blue. Two collections cited; one seen by me. Delphinium axilliflorum DC. var. minus Huth, Bot. Jahrb. 20: 369. 1895. No collections cited. One to 6 dm. high, strigulose or glabrescent, the stem erect, mostly sub- simple, more or less equally leafy throughout; leaves twice trifid, the lower petioled, the upper subsessile, the ultimate segments linear, mostly 1—1.5 mm. wide, strigulose, acute; leaf blades largely 1-2 cm. long; flowers solitary in axils, often from near the base of the plant; pedicels very short; bracteoles subopposite, at base of flower; sepals blue, violet or rose, oblong, obtuse, the lateral and lower pairs 10-13 by 5-6 mm., unguiculate, the lateral with pubescence confined to the median strip; upper sepal narrower, pubescent, the spur slightly curved, pointed, 6-9 mm. long, ca. 2 mm. wide at base; “petal” 3-lobed, kite-shaped, almost 2 cm. long including the spur, the oblong terminal lobe ca. 5 mm. long, with 2 short rounded terminal lobules, the lateral lobes broadly triangular, obtuse, ca. 5 mm. long and wide; stamens unequal, the filaments oblong below, narrowed and more or less pubescent above; anthers brownish, ca. 1 mm. long; follicle more or less glabrescent in maturity, erect and appressed to the stem, ca. 16-20 by 3 mm., longitudinally furrowed when mature, the beak an additional 3 mm.; seeds triquetrous, the broad scales more or less distinct. Type: In Oriente, Labillardiére; not seen. DIsTRIBUTION. Fields, grassy places, etc., below 1800 m., Turkey (Cilicia to Mesopotamia) to Syria. ILLUSTRATIONS, DELESSERT, Icon. Pl. 1: ¢. 50. 1820; Hossain & Davis, Notes Bot. Gard. Edinb. 22: 423. 1958; Davis, Fl. Turkey 1: 127. 1965. REPRESENTATIVE COLLECTIONS. Turkey: Hennipman et al. 1393; Balansa 729; E. K. Balls 2243; Post 213; Davis 28753, 28680, 22135. Syria: Post 218; Kotschy, March 13, 1841; Letourneux 327: Dinsmore 20367 ; Haradjian 2266. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 175 13. Consolida gombaultii (Thiebaut) Munz, comb. nov. Fic. 2, D. Delphinium gombaultii Thiebaut, Bull. Soc. Bot. France 81: 114. 1934. Near to C. axilliflora, but glandular-pubescent throughout and also with very short non-glandular hairs; stem erect, simple or slightly branched, 3—5 dm. tall; leafy and flowering from the base; lower leaves 3-4 cm. long, multifid into linear subobtuse segments 1 cm. or so long, ca. 1.5 mm. wide and with very short stubby pubescence; median and upper leaves linear- lanceolate, entire, becoming bracts above, 1—3 cm. long and exceeding the pedicels; pedicels 2-3 mm. long; bracteoles subulate, 2-3 mm. long; flowers small, 8-12 mm. long, arranged in a long slender spike; sepals glandular-pubescent, intensely blue, 6-8 mm. long, oblanceolate, unguicu- late; spur ca. 6 mm. long, slightly curved; “petal” pale blue, 3-lobed, the upper lobe bifid, the lateral broader, widespread, pointing slightly forward, the whole “petal” ca. 1 cm. long and broad, its spur almost straight, some- what shorter than the blade; follicle erect, 10-12 mm. long, 3 mm. wide, short-beaked; seeds gray, squamate. Type: Syria, Djebel Druze, Gombault 1717, May 21, 1932 (P); seen. This species is near C. axilliflora, but with shorter pedicels, entire bracts, smaller flowers, shorter follicle, and a more general glandulosity. I have seen no collections other than the type. 14. Consolida stocksiana (Boiss.) Nevski in Komarov, Fl. U.S.S.R. 7: Li 1957, Fic, 2, ©, F. Delphinium stocksianum Boiss. Diagn. II. 1: 12. 1853. Delphinium stocksianum Boiss. var. glabrescens Boiss. loc. cit. 1853. Type, Griffiths 1376 (K); seen. : Consolida stocksiana var. glabrescens (Boiss.) Tamura in Kitamura, Fl. Afghan. 124. 1960. Delphinium kabulianum Akhtar, Kew Bull. 1938: 86. 1938. Based on S. A. Akhtar from near Kabul, Afghanistan (K); seen. Consolida stocksiana (Boiss.) Nevski var. kabuliana (Akhtar) Tamura in Kitamura, Fl. Afghan. 124. 1960. Plant 2—5 (—7) dm. tall, branched below, cinereous-strigulose and leafy throughout or with more spreading hairs above; lower leaves long-petioled, tripartite into narrowly lobed segments, the lobes oblong-linear, subobtuse; intermediate leaves with shorter petioles or sessile, the segments entire, linear-spatulate, canaliculate, 1.5-4 mm. wide, with mucronate recurved tips; leaves pretty much shed by anthesis; flowers creamy white, in long lax spike; pedicels to 6 mm. long, minutely bibracteolate at the base; sepals 7-11 mm. long, with violet median stripes, the lateral sepals pubes- cent on midrib only; spur vertical, 1.6-2.6 cm. long, 4 mm. wide at base, gradually attenuate, somewhat curved; “petal” 3-lobed, ca. 8 mm. high, 1 mm. wide, the middle lobe 2-2.2 mm. long, with 2 rounded terminal lobules; lateral lobes spreading, rounded, slightly turned upward, 2.5-3 mm. long, the spur 16—18 mm. long; stamens 4-8 mm. long; anthers apaek 0.7-1 mm. long; follicle cinereous, strigose, erect, short, cylindrical, sub- 176 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 arcuate, the slender style one-third as long as the body which is 8-10 mm. 2.7 mm.; seeds brown, subtriangular, regularly rugulose. Type: Fields in Baluchistan, Dr. Stocks 979 (x !, isotype; c, holo- type, not seen). DisTRIBUTION. Fields and open places, largely between te and 2000 m., Central Asia: Pamiro-Alai, Turkmannia, Iran, Afghanis REPRESENTATIVE COLLECTIONS. Afghanistan: Furse 6469, 5942, 5943; Herb. Griffith 42; Kay Beach 5029; Hay 210. Turkestan: Regel, May 1882, and Sept. , 1884. This species is confused with Consolida rugulosa and C. persica. From the former it should be separable by the more triangular and not rounded lateral lobes of the “petal”. From the latter by the narrower “‘petal’’, which is 8-11 mm. wide when expanded, not 15-20 mm. It varies in pubescence and flower size. So far as I can see, the plants described as C. kabuliana are small and with smaller flowers than in many populations and those called var. glabrescens may have a somewhat more glabrescent follicle in maturity. ye ambigua (L.) Ball & Heywood, Feddes Panne B6s-151, 962. PiG.2,;44, pee ambiguum L. Sp. Pl. ed. 2. 749. 1763. Consolida ajacis auct. non (L.) Schur, Verh. Siebenburg. Ver. Naturf. 4: 47. 1853 Ceratosanthes ajacis Schur, eh cit. 46. Plant subglabrous, especially below, to more or less pubescent, espe- cially above, with short fine appressed or spreading hairs, mostly one- stemmed, 3-10 dm. tall, generally branched above with ascending-horizon- tal branches; lower leaves long-petioled, the blades commonly 2—5 cm. in diameter, multifid into ultimate more or less linear mucronate segments largely 1-1.5 mm. wide; cauline leaves gradually reduced upward, the upper sessile, often 1-2 cm. long; bracts multifid to entire, the lower often at least as long as the pedicels, the uppermost 2—3 mm. long; inflorescence more or less lax, each branch ending in a few- to many-flowered raceme; pedicels commonly 1-3 cm. long, often about equal to the diameter of the flower; bractlets mostly 1.5—4 mm. long, usually below the middle of the pedicel; flowers blue to rose or white, 2—2.5 cm. across; sepals ovate, at- tenuate at base, more or less narrowed into an obtuse apex, 12-14 mm long, more or lets pubescent, each with a greenish subapical spot: spur Ca. 15 mm. long, straight or slightly curved; “petal” 3-lobed, the central lobe oblong, with 2 short rounded lobules and about 6 mm. long, the 2 lateral lobes rounded, about as long, spur 15-16 mm. long; stamens 6-7 mm long, the filaments abruptly dilated at the base, glabrous; anthers yellow; follicle pubescent, cylindric, 16-22 mm. long, the style ca. 2 mm. long; seeds black, ca. 2.5 mm. long, oblong-trigonous, with undulating transverse rows of scales. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 177 Type from Mauretania, ee Africa, a double form in the Linnaean Herbarium 694/4; photograph se DISTRIBUTION. Welieea ce. exon but much cultivated and widely escaped as various horticultural for ILLUSTRATIONS. REICHENBACH, So: Fl. Germ. 4: pl. 67, 1840; Marre, Fl. de l'Afrique du Nord 13: 54. fe. 24. 1964. REPRESENTATIVE COLLECTIONS. Transcaucasia: Hohenacker 1838; F. N. Meyer 541. Iraq: Rawi 10738; Guest 247. India: Giles, 1885; Wight 10, 18; Falconer 59; H. J. Walton, Aug. 14, 1904; G. Watt, March 1878, Honan Island: Levine 952, 824, 823. The question of the application of the name ambigua has not been en- tirely agreed upon. I am following Ball & Heywood (Feddes Repert. 66: 151. 1962), Chater (Tutin et al., Fl. Europaea 1: 217. 1964), and Davis (Fl. Turkey 1: 126. 1965) in using it, but E. Janchen (Feddes Repert. 72: 34. 1966) disagrees and believes the Linnaean epithet ajacis applies to this taxon. 16. ae Pett aia (Gay) Schrod. Abh. Zool.-Bot. Ges. Wien (59025, Fic. 2, H. Delphinium orientale Gay in Desmoul. Cat. Dordogne 12. 184 Delphinium pra i var. orientale (Gay) Finet & Gagnep. UL Soc. Bot. France 51: 467. 1904. Delphinium Pe ana Willk. Prodr. Fl. Hisp. 3: 969. Delphinium orientale ssp. hispanicum (Willk.) Batt. in ANG & Trab. Fl. Alg. 15. 1888. Delphinium orientale var. brevicalcaratum Huth, Bot. Jahrb. 20: 376. 1895. Not typified, but several collections cited seem to belong here, such as Aitchison, Falconer, Griffith. Delphinium bithynicum Griseb. Spic. Fl. Rumel. 1: 320. 1843. Type from near Bolu, Bithynia, pee not seen by me. Referred to D. orientale by Boiss., Fl. Orient. 1: 79. 1867. Mostly 1-stemmed, simple or branched, especially above, erect, 2—6 dm. tall, the lower parts retrorse-strigose, upper with more or less crisped pubescence, some gland-tipped hairs and some swollen at the base; leaves mostly subglabrous, the lower petioled and early withered, the upper sessile; principal leaves multifid, crowded on the stems, the blades 2—4 cm. long, the segments linear or lance-linear, 1-2.5 mm. wide, more or less ciliolate; flowers few to many, in racemes that are rather dense at anthesis, more open in fruit; bracts green, the lower with linear parts, the upper linear, entire, 5-20 mm. long; pedicels 0.5—4 cm. long, bibracteolate near the summit; flowers 1.52.5 cm. in diameter, dark purple to rose or white; the sepals round-ovate to oval or somewhat narrower, ca. 10-12 mm. long, abruptly contracted into a pubescent basal claw almost as long as the blade which is obtuse or rounded at the summit and more or less pubescent; spur ca. 1 cm. long; “petal” usually of same color as sepals, 3-lobed, 8-10 mm. long, the median lobe ovate to oblong-ovate, 2-lobed, somewhat longer than the rounded lateral lobes; stamens 6—7 mm. long, the filaments dilated 178 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 at base, with some gland-tipped hairs at upper end; anthers yellow, ca. 1.5 mm. long; follicle strigose, 14-22 mm. long, oblong-cylindric, with some gland-tipped hairs, the style 1.5-2 mm. long; seeds brownish black, obpyramidal, angular, 1.5—2 mm. long, with several undulating transverse rows of short scales. Type: Described from cultivated material originating from the Orient, apparently represented at K; seen. DistriBuTION. Grain fields and fallow places below 1500 m., Mediter- ranean region to Turkey, Caucasus, Iran, India, Tibet, central Siberia. ILLUSTRATIONS. CurTIS, Bot. Mag. 169: pl. 186. 1952-3; Davis, FI. Turkey 1: 129. fig. 2. 1965. REPRESENTATIVE COLLECTIONS. Caucasus: Brotherus 32; F. N. Meyer 541. Transcaucasia: Shevljakov, June 25, 1935; Gurvitsch, July 17, 1936. Turkestan: F. N. Meyer 572; Michelson 217-1911. Turkey: Sintenis 4137, 420; Bornmiiller 13692; Bourgeau 6; Balansa 1133; Krause 3397, 3723; Davis 30598, 28945, 30079, 31992. Iran: Bunge, 1858-9; Furse 2459, 1920; Lindsay 1151; Rechinger 4173, 5220, 383. Afghanistan: Herb. Griffith 47; Aitchison 444; Hay 239. Parsa in his Flore de l’Iran 1: 431. 1951, proposes Delphinium orientale Sims var. major, the description of which makes application of the name very uncertain. The terms sepals and petals are baffling. 17. Consolida regalis S. F. Gray, Nat. Arr. Brit. Pl. 2: 711. 1821. Delphinium consolida L. Sp. Pl. 1: 530. 1753. Plant 2-12 dm. tall, mostly divaricately and rather finely branched, subglabrous to retrorse-strigulose; radical leaves petioled, early deciduous, the main cauline ones trifid, then again parted into ultimate, linear, mu- cronate, strigulose lobes 1-5 cm. long; flowers few, in rather open terminal racemes, deep blue or more seldom pink or white; bracts mostly simple, linear, 3-15 mm. long; pedicels very slender, 1-5 cm. long; bracteoles linear, alternate to opposite, often well below the flower, 1.5—4 mm. long; flowers (excluding the spur) 9-16 mm. long; sepals more or less ovate, narrowed at both ends, scarcely or short-unguiculate, strigose mostly along the midrib in the lateral pair; spur 17-22 mm. long; “‘petal” blue to cream, 3-lobed, 12-15 mm. wide, the spur 16-17 mm. long, middle lobe of limb 3.5—5 mm. high, shortly 2-lobed at apex, each of these lobules often emar- ginate; lateral lobes rounded, 4-5 mm. long; stamens 4~7 mm. long; fila- ments dilated on lower portion; anthers 0.6—0.7 mm. long, yellow; follicle to 2 cm. long including the 2-3 mm. long style; seeds dark, with subcon- tiguous scales in transverse rows. Sepals 12-16 mm. long, usually light violet-blue; stems often relatively simple; follicle usually twice as long as broad. Central and northwestern Europe LO SURON ee ee Gs ee ubsp. regalis. Sepals 9-11 mm. long, usually dark violet-blue; stem usually repeatedly 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 179 Consolida regalis subsp. regalis Fic. 2, I. Stem usually once or twice branched; flowers 22-35 mm. long; sepals 11-15 mm. long, usually light violet- blue, rather close together in the in- florescence; follicle glabrous, usually twice as long as broad. TYPE LocaALity: Britain. DIstRIBUTION. Fallow fields, etc., largely below 1000 m., Europe, Cau- casia, rarely in Turkey. ILLUSTRATIONS. REICHENBACH, Icon. Fl. Germ. 4: pl. 66. 1840; Hect, Ill. Fl. Mittel-Eur. 3: 487. pl. 114. 1909. REPRESENTATIVE COLLECTION. Davis 29695. Consolida es subsp. paniculata (Host) Sod, Osterr. Bot. Zeitschr. 71: 243. 19 Poe: 2, me. Delphinium Rea Host, Fl. Austr. 2: 65. 1 Delphinium arg be subsp. paniculatum Hse Busch in Kuznetzow, FI. Cauc. Crit. 3: 4 Consolida regalis s - Gray subsp. paniculata Soo var. paniculata, Davis, Fl. Turkey 1: 128 5; Ceratosanthus Arabi Schur, Verh. Siebenburg. Ver. Naturwiss. 4: 47. 185 3. Delphinium consolida L. var. micranthum Boiss. Fl. Orient. 1: 78. 1867. A new name for D. paniculatum. Flowers woes oe spur 14-18 mm. long; sepals ca. 1 cm. long, to 6 mm. broad; “petal” c o mm. wide, the central lobe to,3.5 mm. long, the lateral 4.5— 5 mm.; hee s 4-6 mm. long; anthers 0.6 mm. long; fol- licle 7-11 mm. long, Le ing! the style, glabrous; seeds angular obovoid, 1.75 mm. long. Type: Jugoslavia, near Cattaro, Tomasini. Not se DIstRIBUTION. Cultivated and disturbed places ie 1000 m., south- eastern Europe to Turkey and Caucasia ILLUSTRATIONS. Curtis, Bot. Mag. 159: pl. 9435. 1936; Davis, FI. Turkey, 1: 129. 1965. REPRESENTATIVE COLLECTIONS. Caucasus: Czerniakowska, Aug. 28, 1936. Turkey: Sintenis 4294; Bornmiiller 13689; Bourgeau 10; Krause 3878; Balansa 595, 1134; Davis 32787, 13008, 13463, 38961. I have seen a number of cultivated specimens, some grown as Delphinium consolida some as D. paniculatum. Consolida regalis subsp. divaricata (Ledeb.) Munz, comb. gis aay Delphinium consolida L. erage 1 sateaaas (Ledeb.) Hossain & Davis, Notes Bot. Gard. Edinb. 22: 424. eu ee —— oS in Eichwald, Pl. Nov. Caspio-Cauc. ‘ pl. 1 Console ‘nis (Ledeb.) Schréd. Abh. Zool.-Bot. Ges. Wien 4(5): 25. 190 180 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 FicuRE 2. Consolida species. Drawings show lamina of “petal’’ of 3-lobed type with lateral lobes perpendicular to terminal lobe, the ae somewhat divided - bees saa lines, when shown, indicate principal vein . C. raveyt, amina m. wide; base of spur indicated. B. C. cruciata “aan 11 mm. wide; drawn rom sik 362 (cH). C. C. axilliflora, lamina 12 mm. wide; from Balansa 729 (us C. gombaulti, lamina 10 mm. wide; from Gombault 1717 (Pp). E. C. pike lamina 11 mm. wide; from Beach 5029 (us). Fe. — . : lana. G. C. bigua, lamina 7, mm, wide; from Levine 824 (cH). H. C. orientalis, mas 13 mm. wide. C. regalis subsp. regalis, lamina 15 mm. wide; from Taccan in 1926 (ny). J. é regalis subsp. divaricata, lamina 14 mm. wide: from — 1934 (cas). K. C. regalis subsp. paniculata, a 13 mm. wide; Sintenis 4294 (us). L. C. incana, lamina 8 mm. wide; from Post at Beirut ae Bed regalis S. F. Gray subsp. paniculata Ags So6 var. divaricata eb.) Davis, Notes Bot. Gard. Edinb. 26: Delhi divaricatum Ledeb. var. gilliati tui, Key ae bles according a, Fl. de l’Iran 1: 432. 1951, but I find no such referen ae like subsp. paniculata in habit and flowers, the ae lobe of the violet “petal” 3-4 mm. long; sepal spur 17-20 mm. long; follicle strigose. Type: Near Astrachan, at mouth of the Volga, Eichwald (LE); seen. DISTRIBUTION, Northern Balkans to Iran and Transcaspia, especially in the region surrounding the Caspian and on the north side of the Black ea. ILLustraTIoN. Eicuwatp, Pl. Nov. Caspio-Cauc. 2: pl, 16, 1831. REPRESENTATIVE COLLECTIONS. Transcaucasia: Hejdeman, May 22, 1930; Rubt- 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 181 zovd, Nov. 3, 1934; Hohenacker, 1834; Koenig 66b. Iran: Miss Lindsay 1424; Rechinger 6678; Harrinsion 203; Purse & Synge 799/16. I have seen cultivated specimens of subsp. divaricata grown as Delphin- ium parryi, D. paniculatum, and D. divaricatum. 18. Consolida incana (E. D. Clarke) Munz, comb. nov. Fic. 2, L. Delphinium incanum E. D. Clarke, Bb 2: 451. 1812-16. Delphinium rigidum DC. Syst. 1: 1817. Type “in Oriente,” piglet G?) not seen, but well Teed in n Delessert, Icon. Pl. 1: pl. § Consolida rigida Hayek, Prodr. Fl. Penins. Balcan. 1 : 315. 1924, ele Beih. Bot. Centralbl. 31(Abt. 2): 181. 1914, not valid. Delphinium exsertum DC. Syst. 1: 345. 1817. Based on a Labillardiére col- lection which I have not seen, but the 1820 plate (Delessert, Icon. Pl. 1: pl. 53) does not reveal any differences from D. rigidum except that it is more glabrous. Divaricately stiff-branched from the base, 2-6 dm. tall, pubescent throughout with short stiffish hairs, some of these in the upper parts gland- tipped, some with yellow swollen bases; lower leaves long-petioled, middle cauline subsessile; blades commonly 2-3 cm. long, multifid into linear- oblong segments, more or less mucronate and ca. 1 mm. wide, pubescent; racemes lax, rather few-flowered; bracts mostly 3—5-lobed, 3-10(—25) mm. long; pedicels commonly 1—2 cm. long, spreading to more or less decurved in fruit; bracteoles 1-3 mm. long, mostly remote from the flower; flowers 12-18 mm. across; sepals dirty white to pale violet to rose, elliptic-ovate, subobtuse, narrowed toward a clawlike base, ca. 5-7 mm. long, somewhat pubescent without, especially on the upper sepal and the lower pair; spur 5-7 mm. long, pubescent: “petal” ca. 8 mm. wide, 3-lobed, pale, the middle lobe 1.5 mm. long, divided three-fourths way to the base, the lateral lobes rounded, ca. 3 mm. in diameter; stamens to 5 mm. long, the filaments glabrous, dilated below; anthers 0.6 mm. long, dark; follicle flattened, more or less reflexed, pubescent, transversely ridged, 7-10 mm. long, 3—3.5 mm. wide, the style an additional 4 mm.; seeds 1.5 mm. long, obpyramidal, dark brownish, with transverse rows of broad scales. Type: The E. D. Clarke collections are at the British Museum (Natural History), but nothing of his material for this species seems to be extant. Huth (Bot. Jahrb, 20: 365. 1895) reduced the name Delphinium incanum to synonymy under D. aconiti L. However, that is a species endemic to the Dardanelles region of Turkey, while Clarke says he found his D. incanum at Cana. His description seems best to fit Consolida rigida of that region in the opinion of Miss Dorothy Hillcoat and Dr. N. K. B. Robson of the British Museum, to both of whom I am indebted and with whom I agree. It would seem, then, that a neotype must be chosen and I propose: Migdal, Lower Galilee, June 13, 1942, Peter H. Davis 4819 (neotype, BM; isoneotypes E, K), from the Cana region. DistripuTion, Below 1500 m. elevation; Israel and Lebanon, western Jordan 182 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 REPRESENTATIVE SPECIMENS of C. inmcana may be cited. Israel: Field & Lazar 314; F. T. Meyers 30. Lebanon: Ehrenberg 193; Blanche 348; Berton 222, Bornmiiller 30. Nov. Beih. 89: 13. 193 Delphinium glandulosum Boiss. & Huet in Boiss. Diagn. II. 5: 11, 1856. Delphinium glandulosum vars. lasiostemon and leiostemon Boiss. Fl. Orient. 1: 80. 1867, the former with sparsely ciliate filaments, the latter with them glabrous. Rather freely and divaricately branched, erect, 3-6 dm. high, finely retrorse-strigulose below, short yellowish-pubescent above, many of these hairs gland-tipped, some stiff and enlarged at the base; lower leaves largely gone by anthesis, petioled, multifid; cauline leaves sessile or subsessile, commonly 2-3 cm. long, the ultimate linear lobes ca. 1.5 mm. broad, pubes- cent with more or less crisped hairs; racemes rather many, lax, few- flowered; bracts mostly simple, lance-linear, to 5 mm. long; pedicels di- varicate, slender, 1-4 cm. long; the bracteoles 2-3 mm. long, near the middle of the pedicel; sepals oblong-spatulate to obovate, scarcely un- guiculate, blue-violet with light median strip, crisp-pubescent especially along the middle, 9-10 mm. long, the lateral pair broadest, rounded at the apex, to 7 mm. wide and least hairy; others more pointed and narrower; spur strigose, ca. 2 cm. long, 3 mm. wide at base, horizontal, flexuous, slender; “petal” whitish with violet tinge at apex, ca. 17-18 mm. wide, with 2 lateral rounded lobes ca. 5 mm. wide and curved slightly upward, the median lobe ca. 3 mm. high, bifid at the tip into lobules about 1 mm. long; stamens 5-8 mm. long, abruptly broadly dilated on lower half, whitish, glabrous or nearly so; anther yellow, ca. 1 mm. long; follicle oblong, with transverse elevated lines, appressed-pubescent, 9-11 mm. long, 3.5-4 mm. wide, the style an additional 4 mm. long; seeds grayish, narrow- ly obpyramidal, ca. 2 mm. long, with several transverse rows of contiguous scales. 19. Consolida glandulosa (Boiss. & Huet) Bornm. Fedde, Repert. Sp. 6. Fic..3, A. Type: Fields near Maimansour between Baibout and Erzerum, Turkish Armenia, at 6000 ft., Aug., 1853, Huet du Pavillon (G) not seen, but iso- types (BM, K, P, MO, UC) have been examined. DISTRIBUTION. Open places, 800-1600 m., central and eastern Anatolia excluding Mesopotamia and Kurdistan. ILLustRATION. Davis, Fl. Turkey 1: 129. 1965. REPRESENTATIVE SPECIMENS. Turkey: Balansa 854; Romia 510 and Born- miiller 1724 —all labeled var. leiostemon. Sintenis 1286 (labeled var. lasio- stemon). Other collections Davis 32616, 31983, 31968, 32715, 31732, 31860; Bourgeau 6. 4(5): 62. 1909. Delphinium oliverianum DC, Syst. 1: 341. 1817. 20. Consolida oliveriana (DC.) Schréd. Abh. Zool.-Bot. Ges. Wien Fig. 3,8 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 183 Delphinium cappadocicum Boiss. Ann. Sci. Nat. Il. 16: 367. 1841. Based on Aucher 74 (e) which I have seen. Delphinium oliverianum DC. var. cappadocicum (Boiss.) Huth, Bot. Jahrb. 1895 Delphinium hispidum Boiss. Fl. Orient. 1: 82, 1867, in synonymy. Branched throughout, 1-5 dm. high, glabrous to pubescent with white spreading hairs above, the branches divergent, somewhat angled, often with some hairs with yellow bases; lower leaves petioled, the blades 1-3 cm. long, trifid, the petiolulate divisions multifid into acute, lance-linear lobes 1—2.5 mm. wide; upper leaves sessile, passing gradually into 3—5- lobed bracts 8-20 mm. long; flowers loosely racemose with long intervals between; pedicels mostly 1—4 cm. long, glabrous except at the summit; bracteoles linear, subopposite, 2.5-3 mm. long, with a few hairs near the summit; flowers bluish purple, almost 2 cm. in diameter; sepals at first cream, later bluish purple, loosely white-pubescent, especially on the green- ish median strip, 8.5-14 mm. long, narrowed into broad clawlike bases, the upper sepal and lower pair ca. 4 mm. wide, narrowed apically, the lateral pair 5 mm, wide, rounded at the tip; spur straight, 13-17 mm. long, 2.5 mm. wide at the base, pubescent; “petal” cream with purplish tinge at the margin, 3-lobed, 15-18 mm. wide, 10-12 mm. high, the lateral lobes 4-5 mm. broad, upturned, the middle lobe ca. 3 mm. high, cleft almost to the base into 2 oblong lobules; stamens 5-8 mm. long, glabrous, somewhat dilated below; anthers yellow, ca. 0.8 mm. long; follicle appressed- pubescent, more or less compressed, 10-16 mm. long, 3-3.5 mm. wide, transversely nerved, the style an additional 2.5 mm., somewhat curved; seeds obpyramidal, ca. 1.7 mm. long, with transverse rows of scales. Type: Fields between Bagdad and Vermancha, Iraq, Olivier (P); seen. DISTRIBUTION. Hills and fields, 600-1400 m., Turkey, Iraq, and Iran. ILLUSTRATIONS, DELESSERT, Icon. Pl. 1: pl. 51. 1820; Davis, FI. Turkey 1: 129, 1965. REPRESENTATIVE SPECIMENS. Turkey: Sintenis 1187, 2732; Davis 280669, 28406, 31026; Kotschy 416 (labeled D. hispidum, sp. nov.) ; Strauss 3. Iraq: Field & Lazar 52, 770, 741; Gillett 8305; Guest 2757, 3742; Rawi 8540. Iran: Gilliat-Smith 1734; Koelz 18202. For some time I separated these as two species on the basis of a spur 13 mm. long and flowers at first cream in Consolida cappadocicum and spur 15-20 mm. long and flowers more purplish in C. oliveriana, but a reexam- ination of the situation seems to indicate that the two are scarcely separable taxa. 21. Consolida fava (DC.) Schréd. Ann. K. K. Naturh. Hofmus. Wien 2): 40. 1915, Fie, 3,.C. imi . Syst. 1: 346. 1817. sree ei pas erent velutinum DC. loc. cit. (pubescent throughout; bracteoles in middle of pedicel) and var. glaucum DC. loc. cit. (glabrous below; bracteoles at base of pedicel). Not typified. Plant 1.5—4 dm. high, branched from base, densely retrorse-strigose 184 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 below, more or less spreading-pubescent above; lower leaves petioled, the blades 2.5—5 cm. long, ternately divided into segments with linear lobes; cauline leaves palmately multipartite, the final divisions linear, very short; bracts entire, short-linear, 4-5 mm. long; pedicels 12—20 mm. long; brac- teoles 2, minute, linear-lanceolate; racemes lax, short; sepals 4-5 mm. long, oblong- elliptical, yellow to brown or pirblidh: spur 4-5 mm. long, straight, bent upward; “petal” wider than long, truncate, 3-lobed, the lateral lobes semiovate, longer than broad, the short- lobulate middle lobe with crimson spots; filaments densely hirsute, 3—4 mm. long; anthers rounded, yellow; follicle cylindrical, strict, viscid-hairy, linear-oblong, at- tenuate at both ends, many seeded, 10-12 mm. long; style 3—4 mm. long; seeds small, angulate, with separate scales. Type: “In Oriente, verosimiliter in Archipelago. Olivier” (Pp). Seen, the label reading “de Bagdad a Kermancha, Voyage d’Olivier et Bruguiére en Orient”; stem very retrorse-hairy below, glandular above only. DIsTRIBUTION. Apparently below 1000 m., Iraq and, less frequently in Syria. EXAMPLES SEEN. Iraq: Rechinger 148, 9907, 19357; Guest 787; Rawi — 21097, 21258, 21070, 209935, an Syria: Mayadine '(Meyadi ae ee May 26, 1946; Post, July 4,1 ee, ae deserti (Boiss.) Munz, Jour. Arnold Arb. 48: 51. fig. 17. 967. Fic. 3, D. cs deserti Boiss. Fl. Orient. 1: 83. Delphinium flavum DC. var. deserti a oe & Gagnep. Bull. Soc. Bot. France 51: 468. 1905. Plant 5—30 cm. tall, diffusely branched from below, short viscid-hairy ; leaves 5-25 mm. long, short hairy, palmately parted into oblong-linear segments, these again laciniate in the main leaves, the tips sharp; bracts 5-8 mm. long, oblong-linear; pedicels 1.5-3 cm. long; bracteoles near middle of pedicel, 3-5 mm. long; flowers small, yellow; sepals hairy, oblong, 4-6 mm. long; spur straight, 5-6 mm. long: “petal” obcordate, ca. 6 mm. wide, apparently 4-lobed, since the middle lobe is deeply divided, the outer lobes are longer, semi-ovate; “petal” with a row of brownish red spots across the base of the lobes; spur horizontal, equaling the blade, i.e., 5-6 mm. long; filaments sparsely hairy; anthers yellow, rounded, 0.5 mm. long; follicle oblong, viscid-hairy, 6-10 mm. lo ong, 2—2.5 mm. wide, the beak slender, ca. 2 mm. long; seeds oblong, with densely imbricate scales about the cupuliform center. Type: Bové 131, from desert between Suez and Gaza, July 1832; this number seen (kK, P), It is spreading-hairy throughout on both those sheets. At GH are two Bové specimens of July, 1832, from between Suez and Gaza: 131 is virgate, few-branched, ca. 3.5 dm. tall, strigulose and with a spur ca. 1 cm. long. It bears only one flower which seems to have the 2 lower petals unguiculate and with round limb. It is Delphinium peregrinum L. There is also at GH a sheet 130, much branched, spreading pubescent, spur 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 185 a . 4 ‘ H ‘ i ’ Ke! * - sf , . ’ . . . ° . ’ Pe ‘ ’ : ' RE 3. Consolida species. Drawings show lamina of “petal” of 3-lobed Bes ai lateral lobes projecting toward the tip of the “petal”. A. C. glandu- losa, lamina 18 mm. wide; drawn from Huet in 1853 (us). B. C. oliveriana, lamina 16 m ide; from Strauss in 1896 (us). C. C. flava, a 6 mm. 1 : C. persica, lamina 17 mm. wide; from poaee th 4 . G. C. deserti- Syriact, lamina — mm. wide; drawn from “V.C.R.”, from Jordan (xk ay cornuta, lamina wide; from Davis 30855 (pm). e C. rugulosa, lamina 8 mm. wide; emi Poa 143 (us). 4mm. long. At K no. 130 is strigulose, with 1-3 carpels. Apparently labels have been exchanged at cu and their Bové 130 should be 131. oe Arabian Desert west of the Gulf of Suez and Red Sea to ee oe SPECIMENS. Egypt: Sinai Desert, M. F. White, May, 1916. Israel: Meyers & Dinsmore 7238; P. H. Davis, 1942, Iraq: Aucher 64; Uvarov, May 26, 1932. This species is near to C. flava in its low stature and yellowish flowers, but differs in its spreading pubescence and glandulosity. 23. Consolida leptocarpa Nevski, Acta Inst. Bot. Acad. Sci. URSS, 4: 296. 1937. Fic. 3, E. Delphinium leptocarpum Nevski in Komarov, Fl. U.S.S.R. 7: 110. 1937. Plant 1-6 dm. high, with several erect or upcurved branches from above the base and below the middle, strigulose below, more spreading-pubescent and more or less glandular above, leafy throughout, the lower leaves lacin- iate into segments 2—2.5 mm. wide; main cauline leaves oblanceolate, entire, 2-4 cm. long, 5~7 mm. wide; lower leaves petiolate, cauline gradually re- 186 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 duced upward, all somewhat pubescent; flowers subsessile, solitary in all but the lower leaf-axils; bracts 1—-1.5 cm. long, trifid with obtuse, lance- oblong, lateral segments; pedicels 1-2 mm. long; bracteoles subulate; flowers pale cream-pink, the sepals not clawed, 9-11 mm. long, the upper lance-ovate, pubescent, the spur 22—24 mm. long, 3.5 mm. wide at the base, erect, flexuous; lateral sepals ovate, 5-6 mm. wide, hairy along the midrib; lower sepals lanceolate, more hairy, 3 mm. wide; “petal” trilobed, ca. 15 mm. wide, 10 mm. high, the terminal lobe ca. 2 mm. long, slightly divided at apex; lateral lobes round-truncate; stamens glabrous, 6-8 mm. long, dilated below; anthers yellow, 1 mm. long; follicle strict, scarcely curved bove, ca. 2 mm. wide, cylindric, rugulose-veined, 17-25 mm. long, pubes- cent, the style an additional 3-4 mm.; upper follicles longer than subtend- ing bracts; seeds 1.5 mm. long, obpyramidal, with pale overlapping scales. Type: “Ad declivia argillosa gypsacea prope trajectum Ak-Davan, June 21, no. 364”, collector’s name not given, possibly Nevski. At K was seen a specimen labeled “Plantae montium Kuhitang (Turcomania orientalis) ,” again without collector, but treated as type material. In his description in Fl. U.S.S.R. 7: 110, Nevski cites as representative of the species: Sintenis 283, which I have seen (BH, BM, GH, MO). DIsTRIBUTION. Turkestan, northern Iran. ILLUSTRATION. Komarov, FI. U.S.S.R. 7: 107. pl. 8, fig. 2. 1937. REPRESENTATIVE SPECIMENS. Transcaspia: Sintenis 283; Michelson, 1911. Turkestan: Fedtschenko at Tashkent; Regel, May 1882 at Samarkand; Gont- scharow, Grigoriev & Nikiton 114 near Bukhara [Bochara]. Iran: Donaldson 48. Formerly treated as Consolida persica and C. rugulosa, but with longer sepals than the former and broader “petal” than the latter. 24. Consolida persica (Boiss.) Schréd. Abh. Zool.-Bot. Ges. Wien 4(5): 17, 162. 1909; Grossheim, Fl. Kavkaza 2: 101. 1930. Fic. 3:8. Delphinium persicum Boiss. Ann. Sci, Nat. IT. 16: 362. 1841. Plant 1-3.5 dm. high, widely and divaricately branched, ashy-strigulose throughout, also with some spreading and glandular hairs, especially in the upper parts; leaves small, scattered, strigulose or pubescent, the blades mostly scarcely 1 cm. long, parted into narrow fleshy segments; lower leaves petioled, upper reduced and sessile; flowers largely in upper half of plant; bracts trifid or the uppermost entire, pubescent, commonly 2-8 mm. long; pedicels 2-4 mm. long; the bracteoles subulate, to ca. 3 mm. long; sepals pinkish or whitish, or with a blue tinge, with greenish midribs, not clawed, 6-8 mm. long, the upper sepal lance-ovate, with steeply ascend- ing spur 20-25 mm. long, 4 mm. wide at base, pubescent; lateral sepals oblong-obovate, 3.5 mm. wide, obtuse, pubescent on midrib; lower sepals broadly lance-oblong, 4 mm. wide, obtuse, rather generally pubescent; “petal” 3-lobed, ca. 17 mm. wide, the terminal lobe ca. 1 mm. high, with 2 short oblong lobules, the lateral main lobes rounded, reaching as high as 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 187 the terminal one; stamens 6—7 mm. long, glabrous; anthers yellow, 0.8 mm. long; follicle on a pedicel to 5 mm. long, hairy, slightly compressed, ca. 10 mm. long, 2.5 mm. wide, not strongly reticulate, the beak an addi- tional 3 mm. long. Type: Aucher 78, Persia circa Amadan; probably at Paris. No. 78 seen S 2 DiIstrIBUTION, At 1000-2000 m., Armenia to Turkestan, Iran, and Af- ghanistan. REPRESENTATIVE COLLECTIONS. Turkey: Demiriz 3412 (fide Davis). Russian Armenia: Szovitz 420. Turkestan: Herb. Fl. Ross. no. 2751; Regel, May, 1882. Iraq: Field & Lazar 532; Haines W 1543; Haussknecht, 1867; Gillett 11207. Iran: Kotschy 442; Herzfeld 4; Lindsay 268, 329; Furse & Synge 625. Afghan- istan: Clifford 14. West Pakistan: Crookshank 1353. 25. Consolida deserti-syriaci (Zohary) Munz, comb. nov. Fic. 3, G. Delphinium deserti-syriaci Zohary, Palest. Jour. Bot. Jerusalem Ser. 2: 155 1941. Stem solitary, erect, branched, ca. 15 cm. tall, with spreading hairs and dense appressed short ones; cauline leaves petioled, ternate or bifid, the lobes further divided into linear segments 1-1.5 mm. wide and 10 mm. long, pubescent; bracts simple, minute, oblong, glandular-pubescent; pedicels 16-18 mm. long, glandular-pubescent, spreading; bracteoles spreading, oblong, 2 mm. long; flowers remote, 23 mm. long; sepals whitish, the upper ca. 8 mm. long, obtuse, the spur horizontal, narrow, ca. 18 mm. long, saccate-uncinate at apex, white to pink, puberulent; lower sepals oblong-lanceolate, acutish, green, apically scarious, ca. 8 mm. long; “petal” white, ca. 2 cm. wide when spread open, 10 mm. high; lateral lobes broad, rounded, separated from the upper by a rounded sinus 2.5—3 mm. deep; stamens 5—6 mm. long, whitish in dilated lower half, yellowish above, glabrous; anthers yellow, 0.8 mm. long; follicle ca. 17 mm. long, 3 mm. wide, canescent-strigulose and with spreading gland-based hairs, beak 4 mm. lon Typr: Between Azra and Damascus, western Syria, Zohary 1931 ; The type was destroyed during the war in Israel and all that remains is a slide with the flower parts spread out. Through the kindness of Professor Naomi Feinbrun of the Hebrew University at Jerusalem I have been sent a pho- tograph of that slide. So far as I can tell, it and the original description agree with a collection from Jordan, without locality, made py PoC. Re Feb., 1954, Kew no. 47 (Kk). 26. Consolida cornuta (Hossain & Davis) Davis, Notes Bot. Gard. Edinb. 26: 174. 1965. Fic. 3, H. Delphinium cornutum Hossain & Davis, Notes Bot. Gard, Edinb. 22: 424. 1958. Erect, 1.5—4 dm, tall, sparsely branched, strigose below, with spreading 188 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 yellowish pubescence above; leaves puberulent, the lower linear-multifid, with blades 12-15 mm. long and petioles 7-9 mm. long; upper blades 20- 25 mm. long, the ultimate divisions linear, acute; raceme lax; pedicels 12-25 mm. long, densely pubescent; bracts linear, acute, the lower 9-12 mm. long and sometimes parted, the upper shorter, entire; bracteoles 2-3 mm. long, remote from the flowers; flowers bicolored when dry; sepals bluish, 8-10 mm. long, pubescent and greenish without along the median nerve, the lateral and lower alike, 8-9 mm. long, 44.5 mm. wide, ovate, obtuse; spur pubescent, 10-13 mm. long, with slightly decurved tip; “petal” 3-lobed, 11-12 mm. wide, 9-10 mm. long, the median lobe longer than the lateral, deeply emarginate at tip, the lateral lobes semilunate, 5—6 mm. long, rounded and wide at the base, attenuate into an upturned hornlike tip 2—2.5 mm. wide; filaments slightly pubescent, dilated below; anthers purplish-brown; follicle erect, densely strigose, obcompressed, semiovate-oblong, 9-10 mm. long, 3.5-4 mm. wide, with the style an addi- tional 2 mm. long. Type: Turkish Armenia, Calvert & Zohrab (holotype £). DIsTRIBUTION. Fields at 1800-1900 m., northern Armenia of Turkey. ILLustraATION. Davis, Fl. Turkey 1: 129. 1965. SPECIMEN SEEN. Erzurum: west of Ilica, Davis 30855. 27. Consolida rugulosa (Boiss.) Schréd. Ann. K. K. Naturh. Hof- mus. Wien 27: 43. 1913. Fie. 31: Delphinium rugulosum Boiss. Ann. Sci. Nat. II. 16: 361. 1841. Delphinium aucheri Boiss. loc. cit. 362. Based on Aucher 4030 from south Persia. I would interpret the specimen at Paris to be the type. I have seen it it. Delphinium persicum Boiss. var. aucheri (Boiss.) Boiss. Fl. Orient. 1: 77. 1867. Delphinium paradoxum Bunge, Arbeit. Nat. Ver. Riga 1: 124. 1847. See dis- cussion below Consolida paradoxa (Bunge) Nevski in Komaroy, Fl. U.S.S.R. 7: 113. 1937. Simple or with simple branches from the base, erect, largely 1-3 dm. high, with some appressed hairs and numerous spreading ones, especially in the upper part (many hairs glandular at the base) ; stems rather equally leafy throughout; lower leaves petioled, the cauline largely subsessile and to 4 cm. long or the uppermost sessile, all more or less fleshy, 4-3-partite, then again divided into narrow segments more or less dilated upward and to 2.5 (-4.5) mm. wide, subglabrous to somewhat pubescent, acute; flowers solitary in most leaf axils, almost sessile or the pedicels to 2 mm. long; bracteoles 2, entire, 2-3 mm. long; sepals pinkish or greenish white, 7-10 mm. long, the spur more or less flexuous, ascending to erect, pubescent, slender, 17-18 mm, long (rarely lacking); lateral sepals narrowly oblong- ovate, obtuse, scarcely clawed, 3.5-4 mm. wide, pubescent only along the median strip; lower sepals linear-oblong, ca. 2 mm. wide, rather generally pubescent; “petal” 3-lobed, ca. 9 mm. long, the broad more or less blunt- 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 189 deltoid terminal lobe more or less erose, the rounded-deltoid lateral lobes almost straight on the upper edge, obtuse, slightly rounded on the outer edges; stamens 6-9 mm. long, the filaments glabrous; anthers yellow, ca. 1 mm. long; follicle erect, shorter than the bracts, subcompressed, hairy, reticulate-veined, 10-15 mm. long, ca. 3 mm. wide, the beak an additional 3 mm.; seeds brownish, obovoid, somewhat angled, ca. 1.6 mm. long, with short pale scales not in very regular transverse rows. Type: Aucher 4028, from Lake Urmiah (Ourmiah), Iran; I would sup- pose the holotype is the specimen at P, isotypes BM, kK. All seen. DistrIBuTION. At 600-1400 m., Transcaucasia and Transcaspia to northern Iran and to Afghanistan. ILLUSTRATION. Komarov, FI. U.S.S.R. 7: pl. 8, fig. 4. 1937. REPRESENTATIVE SPECIMENS. Transcaspia: Sintenis 143, 283; Litwinow 440. Turkestan: Sovetkina 131; Kusckokewicz. Iran: Rechinger 1011, 789; Merton 3884, West Pakistan: Lace 3338. Afghanistan: Aitchison 310. In genera like Consolida, Delphinium, and Aquilegia spurlessness occurs apparently as a mutation, not only in cultivation, but also in nature. For example, in Aquilegia, refer to Munz, Gentes Herb. 7: 134. 1946, for its appearance in spurless garden plants of A. caerulea James. Sometimes in nature spurless forms have been given varietal status, as A. micrantha var. mancosana Eastw. In Delphinium I have found the spurless condition in a number of specimens, as in a collection by H. B. Morse (Ny) from Lung- chow, Kwangsi Prov., China and belonging to D. anthriscifolium Hance. It lacks both spur and petals. From Africa I have seen a collection of D. macrocentron Oliver, Bally 1205 (x), with 4 sepals, 2 upper petals without spurs, but with well developed stamens and carpels. In other cases I have noticed plants with both normal flowers and abnormal, some- times on separate branches. Since Consolida paradoxa resembles C. rugu- losa in its vegetative characters, so far as I can see, I agree with the pos- sibility expressed before (Huth, Bot. Jahrb. 20: 373. 1895) that it is simply a spurless form of the latter. Its sepals are a trifle smaller; it lacks petals altogether, but stamens and follicle and seeds are well developed. The type number is Lekmann 36 from Turkestan. I have seen specimens (GH, K, P, LE). Another collection, from northern Iran is Polak, 1882 («). 28. Consolida halophila (Huth) Munz, comb. nov. Fic. 4, A. Delphinium halophilum Huth, Bot. Jahrb. 20: 487. 1895. New name for D. freynit Hut Delphinium freynii Huth, loc. cit. 372. April, 1895; not Conrath, Bull. Herb. pre gtd Ee F reyn, Denkschr. Akad. Wien Math. Naturw. 51: 293. 1886; not DC. 1817 sos : : ? Delphinium freynii Huth var. hirtum Freyn ex Huth, loc. cit. 372. No speci- mens cited Stem flexuous, simple or openly branched below, 6-45 cm. tall, ashy- strigose, leafy throughout; lower leaves trifid, the parts petiolulate and 190 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 divided into narrow, obtuse, tripartite lobes; main cauline leaves 1-3 cm. long, the lobes to 2.5 mm. wide; branches floriferous from near the base, more or less glandular-pubescent in the upper parts; bracts sessile, 3- parted, ca. as long as the flowers, the lobes subacute; pedicels to ca. 5 mm. long; bracteoles lanceolate, 2-3 mm. long; sepals elliptic, glabrescent, pale violet-pink to creamy-yellow, 2 mm. wide, obscurely dotted, 5—7 mm. long; spur directed upward, 15-24 mm. long, 5-6 mm. wide at base, obtuse; “petal” 3-lobed, ca. 8 mm. wide, 10 mm. long, the middle lobe emarginate, slightly longer and much narrower than the lateral lobes which are directed slightly forward; follicle erect, 10-13 mm. long, straight, cylindric, veined, tomentose and spreading- hairy, the beak an additional 2-3 mm.; seeds scaly, transversely angular, 1.5 mm. long, 1 mm. thick. Tyrer: N. Iran, Gussediche on the Teheran road, Pichler 1882; holotype, Barbey Herbarium (G) not seen; isotype (K) seen. DIstRIBUTION. At about 1600-2000 m., northern Iran. EXAMPLES SEEN. Furse 3301; Bunge, March 1859; Gilliat-Smith 1724, 1914; L. S. Grown 1923. 29. Consolida trigonelloides (Boiss.) Munz, comb. nov. Fic. 4, B. Delphinium trigonelloides Boiss. Ann. Sci. Nat. II. 16: 366. 1841. Bushy, 3—15 cm. tall, branched at base, gray with short spreading pubes- cence; leaves canescent-pubescent, the cauline with 2-3 linear or linear- spatulate lobes 6-12 mm. long, obtuse or rarely bifid; upper leaves entire, oblong, 4-6 mm. long; bracts oblong, entire, 5 mm. long, 1-1.5 mm. wide; pedicels at tips of branches, remote, 10-12 mm. long at anthesis, 25-30 mm. in fruit; bracteoles oblong, opposite, entire, 4-5 mm. long; flowers bluish; sepals ca. 8 mm. long, oblong-spatulate, obtuse, the upper acute, VA? FicuRE 4. Consolida species. Drawings show lamina of “petal” of ; -lobed type (and in one case the spur) with patne lobes gale reduced. A. C. halophila, lamina 8 mm. wide; drawn from Pichler in 1882 BC. cater lamina 8 mm. wide; from ‘Aucher 4033 (pe). C.C. mien lamina 5 mm. wide; from Aucher 4029 (P). with a straight pubescent attenuate spur 16 mm. long; “petal’’ obtusely 3-lobed, the middle lobe subtruncate, slightly 3—5-lobulate, smaller than the short rounded lateral lobes; filaments glabrous; follicle cylindric, ap- pressed-canescent, incurved, 10-12 mm. long, 3 mm. thick, the beak ca. 1.5 mm. long; seeds ca. 1.6 mm. long, pyramido-pentagonal, basally trun- cate, with transverse white scales in rather irregular arrangement. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 191 Tyre: Southern Iran, Aucher 4033 presumably P, isotypes BM, G, K. All these seen. There are no flowers on the BM and K specimens. No other material seen. 30. Consolida linarioides (Boiss.) Munz, comb. nov. Fic. 4, C. Delphinium linarioides Boiss. Ann. Sci. Nat. II. 16: 368. 1841. Branched at the base, with spreading soft white pubescence, to 1.5 dm. tall; lower petioles ca. 1 cm. long, the laminae 3—4-parted, the segments with linear-obtuse lobes 0.5—-1.5 cm. long; upper leaves sessile, subentire or with 2 lateral lobes, 1—1.2 cm. long; racemes laxly 3—4-flowered; bracts broadly linear, 8-10 mm. long, 1.5—2 mm. wide; pedicels 1-2.5 cm. long; bracteoles 6-8 mm. long, linear, at base of flower; sepals violet, oblong, obtuse; upper sepal acute, 8 mm. long; spur subincurved, ca. 4 mm. long; “petal” scarcely longer than the sepals, tricrenate at apex, the 3 lobules rounded, equal; filaments glabrous, dilated at base, ca. 5 mm. long; anthers yellow, ca. 1 mm. long; follicle densely white-strigose. Type: Ispahan, Iran, Aucher 4029; holotype (P) seen; isotype (K) seen. Known only from the type collection; at least to me. T here is some question about how normal the plants of this collection are. The specimen at Kew seems to have one normal flower with oblong sepal and “petal” ca. 1 cm. long and with 3 minute lobes as in my figure, but the other flowers have an abortive spur (2 mm. long, crinkled) and the upper sepal more or less petal-like, but no “petal” developed. Obviously the species is near C. trigonelloides, but with quite a different “petal” and with the sepal spur shorter than the sepal-blade instead of twice as long. 31. Consolida schlagintweitii (Huth) Munz, comb, nov. Fic. 5, A. Delphinium schlagintweitii Huth, Bull. Herb. Boiss. 1: 329. 1893. Plant glabrescent below, pubescent above with some hairs having swollen yellow bases; stem slender, branched, apparently several dm. tall; cauline leaves largely 2—5 cm. long, subsessile, multipartite into linear lobes 0.5— 1.5 mm. wide, 1-3 cm. long, more or less strigulose; lower bracts pluri- partite, upper entire, subulate, ca. 5 mm. long; racemes rather few flowered, rather dense, the pubescence short, stiffish, much of it more or less retrorse; pedicels suberect, 1-1.5 cm. long at anthesis, to ca. 5 cm. in fruit: bracteoles 2.5-3 mm. long, almost or quite touching the flower : spur straightish, 6-7 mm. long, 1.5 mm. thick at the base, somewhat stri- gose; sepals 6-7 mm. long, deep violet, oblong to ovate, definitely unguicu- late, the lateral pair subglabrous except for the ciliate claw, ovate; lower pair narrowed, more generally pubescent, rather pointed; “petal” 3-lobed, the middle lobe ca. 3 mm. high, with 2 lobules, the lateral lobes spreading, the whole “petal” ca. 11 mm. wide; follicle oblong, more or less flattened, densely hairy, ca. 1 cm. long, 3.5 mm. wide; style ca. 1 mm. long. Type: Kashmir, prov. Balti, near Skardo, at 2300-2500 m., Schlagint- weit 821, 1856, William Barbey Herbarium (c); isotypes (BM, US) seen. 192 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 This is the only material I have seen of this species which is remote in its eastern range for the genus Consolida. It somewhat resembles C. orientalis, but is much smaller-flowered. 32: popes phrygia (Boiss.) Sod, Osterr. Bot. Zeitschr. 71: 245. Fic. 5, B. eae phrygium Boiss. Ann. Sci. Nat. II. 16: 363. Delphinium ajacis L. var. phrygium (Boiss.) Finet & par Bull. Soc. Bot. France 51: 467. 1904. Consolida orientalis oe) Schréd. ssp. phrygia (Boiss.) Chater, Feddes Repert. 68: 193. 1963 Plant 2—5 dm. tall, dil or few branched above, with short spreading, often gland-tipped hairs and rather densely leafy at least in the lower half; basal leaves withered at time of anthesis, with petioles to ca. 2 cm. long, blades 1-1.5 cm. across, multifid into linear, pubescent, mucronate seg- ments ca. 0.5 mm. wide; cauline leaves rather crowded, mostly sessile, with numerous narrow linear segments; racemes 2—10-flowered, to ca. 1 dm. long in fruit, many of the hairs papillose at the base; pedicels mostly 5—7 mm. long, erect; bracts largely linear, entire, 3-6 mm. long, the basal sometimes divided; bracteoles linear, 1-2.5 mm. long, near the base of the flower; flowers violet, rose or white, 10-13 mm. in diameter; sepals some- what unguiculate at the base, ovate, acutish, pubescent at least along a greenish median strip, ca. 6-9 mm. long; spur straight, pubescent, horizon- tal, 9-10 mm. long, slender, ca. 1.5 mm. wide at base; “petal” 3-lobed, the central lobes ca. 6 mm. high, rather deeply divided with a broad sinus, lateral lobes rounded-reflexed; spur ca. 9 mm. long; stamens ca. 5 mm. long, the filaments rather abruptly dilated, glabrous; follicles appressed- pubescent, cylindric, 10-18 mm. long, ca. 4 mm. thick, subterete, the style 2 mm. long; seeds with transverse rows of connate scales. Type: Phrygia (Turkey), Aucher 71, not see DistTRiBuTION. Fields in Phrygia and Anatille as well as Cyprus, Rhodes REPRESENTATIVE SPECIMENS. Turkey: Pinard, 1843; Sintenis 843. 33; ee be ryetien (Labill.) Schréd. Abh. Zool.-Bot. = Wien 4(5): g. 4,C. eas es Labill. Icon. Pl. Syr. Rar. Dec. IV. 5: pl. 2, a 2, 1812. Delphinium pygmaeum Poir. in Lam. Encycl. Bot. Suppl. 2: 458. 1812 Plant 2-8 cm. high, densely soft-pubescent, canescent, simple or divari- cately short-branched only near the top; leaves small, to 15 mm. long, 3—5-parted, the lobes lance-linear, obtuse; lower leaves short-petioled; bracts 3-parted, ca. 6 mm. long; pedicels ca. 1 cm. long; racemes dense, with 3—5 violet flowers; bracteoles linear, 6-7 mm. long, extending to above the base of the calyx: sepals elliptic- ‘oblong, obtuse, 8-9 mm. long, the lateral pair obtuse, 2 mm. wide, soft-hairy especially along the midrib; 1967] MUNZ, ASIAN SPECIES OF CONSOLIDA 193 upper sepals 9-10 mm. long, 4 mm. wide, the spur 5-8 mm. long, horizon- tal, pubescent, curved; lower sepals 3 mm. wide; ‘petal’? rhombic, ca. 9 mm. wide and high, the middle lobe ca. 3 mm. long, subtruncate and erose at tip, the lateral pair ca. 3 mm. wide; stamens 3—4 mm. long, the filaments broadly winged at base, glabrous; anthers yellow, 0.7 mm. long; follicle oblong, compressed, strigose, 13 mm. long, 2.5 mm. wide, the beak 2 mm. long. Type: Near Damascus at base of Mt. Djebel Cher; an isotype seen (Kk). ILLUSTRATION. LABILLARDIERE, Icon. Pl. Syr. Rar. Dec. IV. 5: pl. 2, fa: Bie. Distr1IBUTION. Rocky places, Lebanon and Syria to Jordan. REPRESENTATIVE MATERIAL. Syria: Post, May 1, 1900. Lebanon: Peyron 1453. Jordan: Kotschy 737; Boissier, June 1846. 34. Consolida camptocarpa (Fisch. & Meyer ex Ledeb.) Nevski in Komarov, FI. U.S.S.R. 7: 106. 1937. Fic,.5, 2, &. Rafotak, in 1884, was sent from Leningrad as representing this taxon; (BM, US) seen. Delphinium camptocarpum Fisch. & Meyer var. leiocarpum Ledeb. Joc. cit. Not typified. A specimen by A. Regel (p), collected much later, seems to be this. Delphinium camptocarpum Fisch. & Meyer B songoricum Kar. & Kir., Bull. c. Nat. Mosc. 15: 136. 1842. A collection “in arenosis Songoriae ad fl. Lepsa,” (se. of Lake Barkasch), Karelin & Kirilof’ 1165 (xy, «, P) seems to be the original one. Delphinium songoricum (Kar. & Kir.) Nevski in Komarov, FL USSR. 7: unculatum Boiss. FI. chahrud, Bunge G); not seen. ete stocksianum Boiss. var. obovatum Huth, Bot. Jahrb. 20: 373. 5 d on Aitchison from Afghanistan; may belong here. Delphinium persicum Boiss. var. regelanum Huth, op. cit. 370. Referred here by Nevski in Komarov, Fl. U.S.S.R. 7: 109. 1937. Based on a specimen by Regel from Bochara (Bukhara) ; not seen. Plant 2-6 dm. tall, erect, mostly with few ascending branches = above the base, with short retrorse-appressed or spreading hairs, not very densely leafy; leaves largely in lower half of plant, mostly petioled, trifid, the segments with 2-3 parts, each of which has 2-3 large coarse terminal 194 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 subacute teeth; blades mostly 1-2 cm. long, strigose, the lower petioles longer than the blades, the upper leaves passing into subsessile bracts, the uppermost 3—5 mm. long; flowers ca. 5-15 in terminal racemes; pedicels 2-20 mm. long, even longer in fruit; bracteoles setaceous, minute; flowers blue or blue-violet, ca. 17-19 mm. in diameter; upper sepal lanceolate, pubescent, 10-11 mm. long, 2.5—3.5 mm. wide, the spur ca. 18 mm. long, suberect, 4-4.5 mm. wide at the base; lateral sepals rhombic-ovate, 10 mm. long, not unguiculate; lower sepals asymmetrical, ca. 10 mm. by 5 mm., rather well covered with pubescence; ‘‘petal” 3-lobed, 10-14 mm. wide when expanded, 8—12 mm. high, the lobes broadly triangular, obtuse, the terminal ca. 3 mm. high, shallowly 2-lobed at apex, the two lateral ca. 4 mm. long, spreading; stamens 6-8 mm. long, glabrous; anthers yellow, 1 mm. long; follicle declined by the recurved tip of the pedicel, curved, more or less pubescent, reticulate-veined, compressed, ca. 10 mm. by 2-3 mm., the style an additional 2-3 mm. Type: Turcomania: “ad latus orientale maris caspii,” Karelin (LE); seen DIsTRIBUTION. Transcaspia and Turkestan to Afghanistan, at 1000 m. or thereabouts. ILLUSTRATIONS. NEVSKI in Komarov, Fl. U.S.S.R. 7: pl. 8, fig. 1a, 1b, 1c. 1937 REPRESENTATIVE COLLECTIONS. Transcaspia: Dimo 130; Litwinow 4380; Sin- tenis 1301. Turkestan: Androsow 2218; Schischkin, June 4, 1913. Afghanistan: Aitchison 1002. The recognition of C. songorica as distinct does not seem to hold. The chief difference seems to be in the more spreading pubescence and perhaps smaller flowers. It occupies the same region as C. camptocarpa and there is great variability in amount and mixture of appressed and spreading hairs. At Kew is a set of five sheets “Afghanistan, Griffith”, some with “Herb. Lehmann, 1852,” some with no. 1377 and 1377a written on them; all are labeled with a name apparently never published. One sheet has ‘No. 42”. The plants seem to be unusually robust camptocarpa, one with sort of a woody base (long tap root), branching new shoots with broken leaves. The blades are 3-4 cm. long. They are in a type cover, with the unpub- lished name on the outer folder. More material may reveal a distinct taxon. 35. Consolida oligantha a Schrod. Ann. K. K. sole Hof- mus. Wien 27: 43. 1913 Fic. 5, F. Delphinium oliganthum Boiss. Fl. Orient. 1: 80. 1867. Consolida tomentosa (Aucher) Schréd. subsp. oligantha (Boiss.) Davis, Notes Bot. Gard. Edinb. 26: 175. 1965. Delphinium oliganthum Boiss. var. brachycentrum Huth, Bot. Jahrb. 20: 381. 1895. Not typified, but the shortness of the spur would seem to place it here. Divaricately branched from the base, 1.5-3 dm. high, more or less densely canescent-strigose; lower leaves 1.5—3 cm. long, ternate then di- vided into mostly linear lobes scarcely 1 mm. broad; upper leaves less 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 195 divided; racemes dense, mostly 4—5-flowered; lower bracts trisect, 5-15 mm. long, the upper entire, linear; pedicels 1-3 cm. long; bracteoles 2, linear, 8-14 mm. long, exceeding the base of the flowers; sepals violet or blue, 10-12 mm, long, oblong-elliptical, the upper pointed, pubescent, the spur ca. 6 mm. long, pubescent, bent upward; lateral sepals ca. 12 mm. by 4 mm., obtusish, pubescent on midrib and part of upper edge, narrowed to a long lower part; lower sepals narrower, more pointed and pubescent; “petal” ca. 15 mm. long, purple-striped, rhombic, the terminal lobe tri- angular, the lateral lobes triangular, spreading; follicle oblong-linear, pendent, strigose, 10-14 mm. long, ca. 2.5 mm. wide, subterete, with some transverse veining, the style an additional 2 mm.; seeds about 1 mm. long, subglobose, with 4—5 rows of more or less contiguous scales. Lectotype: “Ad segetes prope Aintab, Syriae Bor.” Haussknecht (c, not seen; K, P, isolectotypes, seen). The Kew specimen is a mixture of Consolida oligantha and C. hellespontica. Aintab is, of course, now in Turkey. DistriBuTION. Grain and fallow fields, 600-1000 m., Mesopotamian region of Turkey and northern Iraq. ILLUSTRATION. HossAIn & Davis, Notes Bot. Gard. Edinb. 22: 419. fig. 2,C. 1958. REPRESENTATIVE COLLECTIONS. Turkey: Davis 22106; Sintenis 1131. Iraq: Field & Lazar 830, 586, 538; J. B. Gillett 8011; Wheeler & Haines W 1063; Rawi 21818 There is a question as to whether this species is distinct from C. tomen- tosa, of which it was made a subspecies (Davis, Fl. Turkey 1: 132. 1965). It is supposed to have the sepal blade about twice as long as the spur but, as Davis points out, this condition does not hold for some collections from Iraq. The more erose and plicately folded tip of the terminal lobe and the more spreading lateral lobes of the “petal” seem to me to distinguish it; indeed, to put it very close to C. hellespontica. 36. Consolida tomentosa (Aucher ex Boiss.) Schréd., Abh. Zool.-Bot. Ges. Wien 4(5) : 62. 1909. Fic. 5, G, H. Delphinium tomentosum Aucher ex Boiss. Ann. Sci. Nat. II. 16: 365. 1841. Simple or branched from the base, 5-15 cm. high, more or less appressed- canescent; lower leaves short-petioled, with 3 linear lobes; upper leaves sessile, 5-8 mm. long; racemes short, lax, 2—6-flowered; bracts 2-parted or entire, 5-6 mm. long; pedicels ca. 1 cm. long, apparently refracted in fruit; bracteoles 2, linear, 4-10 mm. long, at middle of pedicel or above; flowers deep violet to mauve; sepals oblong, acutish, grayish-strigose, ca. 1 cm. long, 3 mm. wide, narrowed rather abruptly to a claw 3—4 mm. long; spur usually 6-8 mm. or more long; “petal” 3-lobed, ca. 1 cm. long and wide, the upper lobe with dark stripes, truncate at apex, the lateral lobes round, somewhat reflexed-spreading. Type: Aleppo, Syria, Aucher 76 (P, presumable holotype; seen). 196 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 DistriBuTION. Dry fields and open places, 500-1200 m., Syria and Lebanon. ILLUSTRATION. Hossain & Davis, Notes Bot. Gard. Edinb. 22: 419. fig. 2, D. 1958. REPRESENTATIVE MATERIAL. Lebanon: Bornmiiller 11326; Fox, June, 1866; Trench, May 25, 1945. Syria: Haradjian 1849, 37. Consolida is rian (Boiss.) Chater in a Feddes Repert. 69: 55. HG Be a Delphinium aan Boiss. Ann. Sci. Nat. II. 16: 366. Delphinium hellesponticum Boiss. subsp. campylopodum Faye). igure & Davis, Notes Bot. Gard. Edinb. 22: 419. 1958. Delphinium campylopodum Freyn, Denkschr. Akad. Wien Math. Naturw. 51: 368. 1856. Consolida hellespontica (Boiss.) Chater subsp. Precoeytien: (Hal. & Charrel) Chater in Heywood, Feddes Repert. 69: 55. Delphinium hellesponticum Boiss. subsp. prenene (Hal. & Charrel) Hossain & Davis, Notes Bot. Gard. Edinb. 22: 419. 1958. Delphinium macedonicum Hal. & Charrel, in Abdur Rahman Hadji Effendi, eee Ottoman Géogr. Bot., Faites Nouv. Relatifs 4 la Prov. Salonique . 1892. Based on Frivaldsky, Macedonia; seen from NY oe paphlagonicum ants ane Herb. Boiss. 1: 328. 1893. Based on Sintenis 4547 (BH, K, NY, P, US pe oes holopetalum Baik, var. en oe (Huth) Huth, Bot. Jahrb. 20: 381. Daipiceaae oliganthum Boiss. var. ponticum Huth, loc. cit. Type from Pon- tus: Amasya, Bornmiiller, 1889. I do not know whether Bornmiiller 970, 1889, from Pontus (BM, K, PH; all seen) is the actual type number or not. Delphinium hellesponticum Boiss. subsp. aintabense Hossain & Davis, loc. cit. Based on Post, June, 1889, from Aintab (Gaziantep), Turkey (BM, MO; seen). ? Delphinium tomentosum Auch. in Boiss. var. longecalcaratum Post, Fl. Syr., Pal., Sinai, ed. 2, 1: 22. 1932. Apparently based on the same Post collection from Aintab, as was subsp. aintabense above Corymbosely divaricately branched, 1-4 dm. tall, sometimes simple, strigulose throughout; leaves divided into long linear lobes, the lower leaves petioled, withered by anthesis, the middle leaves gradually reduced and subsessile, the upper becoming bracts and the uppermost entire; flowers violet or violet-blue inside, more or less silvery-gray with appressed pubes- cence outside and with some green on the midribs of the sepals, rather few in number and in fairly lax to rather dense racemes; pedicels 1-4 cm. long, upwardly spreading-incurved; bracteoles 2, linear, 4-15 mm. long, placed just below the flower and growing up around it; sepals more or less lance-ovate, not unguiculate, 10-14 mm. long; spur 12-20 mm. long, slender, pubescent; “petal” light blue to yellowish and violet-nerved, 9-15 mm. long, triangular in outline, 3-lobed, either gradually narrowed toward the more or less erose apex or the upper lobe subequal to the lateral lobes; stamens glabrous; anthers yellow; follicles subterete, strigose, 10-13 mm. long. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 197 Type: Ad Hellespontum [Dardanelles], Aucher 67 (Pp, K); seen DistRIBUTION. Grain and fallow fields below 1200 m., Anatolia to north- eastern Greece and northwestern Syrian desert ILLUSTRATIONS. Hooker, Icon. Pl. 32: pl. 3152. 1932; Davis, FI. Turkey 1: 126. 1965. In 1958, Hossain and Davis (Notes Bot. Gard. Edinb. 22: 418) recog- nized four subspecies in Delphinium hellesponticum which they keyed out as follows: “Petal” 13-15 mm. long, 12-13 mm. wide; plants mostly 2.5-4 dm. a racemes rather lax. Northern Greece, Paphlagonia, Galatia, etc. in Turkey. ...... subsp. Neccedbabiew. a RE 5. Consolida species. Drawings show lamina of ag th in those 3- pra ina. ie in which the ier lobes turn toward the base of the la scMlaginbineine lamina 11 ve wide; rawn from Schlagintiecit 821 (us C. phrygia, a 12mm om Sintenis & Rigo (c C. pusilla, amina 9 mm e; ee Shoko in 1846 (GH). C. camptocarpa, lamina 4m ; from ibe 13 E. C. camptocarpa, the form often called songorica; lamin wide; Regel (us 2090341). F. C. oligantha, lami de; from Field & Lazar 586 (us) C. tomentosa, lamina 10 mm. wide; Iraq Petroleum Co. no C. tomentosa, lamina 9 .W. macedonica, lamin na 15 mm. wide; from ee 4547 (ny). K. C. armeniaca, amina 13 mm. wide; from Balls 1756 (us). L. C. sulphurea, lamina 8 mm. wide; rom Hiuesaicks at Maras (x). M. C. olopetala, lamina 15 mm. wide; aed Davis 30951 (3M 198 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 “Petal” 9-12 mm. long, ca. 10 mm. wide; plants mostly 1-2 dm. high; racemes rather dense Bracteoles 10-15 mm. long; lowest pedicel 25-40 mm. long; raceme of main stem pyramidal, prominent. Aintab [Gazianteb] (Turkey), Aleppo CIS fos ha as ie ie kw eae is subsp. aintabense. Bracteoles 4-8 mm. long; lowest pedicel ca. 15-20 mm. long; raceme of main stem seldom prominen “Petal” yellowish when dry, often with pinkish shades, 10-11 mm. long; sepals violet inside. Lycia, Caria, Phrygia (Turkey). .......... sco PO ET HLL Eley eePaPeD Ye One Nace IU TE a subsp. ee “Petal” usually light blue when dry, often pinkish white, 9-10 m long; sepals violet-blue inside. Mostly central Anatolia. ...... Phe re see ee ome in A erst Pelee 4 ees: subsp. hellesponticum. In the Flora of Turkey 1: 133. 1965, Davis says “Additional material breaks down the distinctions between the subspecies previously recognized by the author.” I am following him because of his great knowledge of the area involved, but the herbarium specimens that I have studied seemed to fit pretty well into his original four concepts. Of this herbarium ma- terial I give some examples using the Delphinium names, since some have never been transferred to Consolida: (1) Subsp. macedonicum. Paphlagonia: Sintenis 4547. Galatia: Ancyra, Born- miller 13693. Pontus: Amasya, Manisdjian 652; Sankiri Bornmiiller 13695, 970. (2) Subsp. aintabense. Turkey: Mt. Diiliik, n. of Aintab, Haradjian 1415. Syria: Aleppo, Kotschy, May 20, 1841. (3) Subsp. campylopodum. Phrygia: Balansa 1132. Caria: Pinard, 1843. Lycia: Bourgeau 7. Cappadocia: Balansa : (4) Subsp. hellesponticum. Central Anatolia: Uvarov 22; Davis 31010, 36538, 36430A; McNeill 336; Ratcliffe 414, 352; Sintenis 4979. 38. Consolida heir, (Stapf ex Huth) Schréd. Abh. Zool.-Bot. Ges. Wien 4(5) : 62. 1909. Fic. 5, K. Delphinium armeniacum Stapf ex Huth, Bot. Jahrb. 20: 380. pl. 7, fig. 9. 1895. Simple or with few, open branches from near the base, 1-3 dm. tall, strigulose, pale in upper parts; leaves multipartite into rather few narrowly linear lobes ca. 0.5 mm. wide, the laminae to 2 or more cm. broad; bracts tripartite or the upper entire, narrow; inflorescence racemose, few- and open-flowered, more or less cinereous, some of the retrorse-appressed hairs with glandular swollen bases; pedicels mostly 1-4 cm. long, curved-ascend- ing; bracteoles 2, linear, near and surpassing the base of the deep blue to violet-blue flowers; upper sepal ca. 14 mm. by 3 mm., lanceolate, not un- guiculate, acute, strigose without, the spur 17-24 mm. long, ca. 3 mm. wide at the base, curved, very slender at the tip; lateral sepals ca. 13 mm. by 6 mm., subobtuse, strigose along the midrib; lower sepals 14 mm. by 4.5 mm., acute, strigose; “petal” ca. 13 mm. wide, 14 mm. high, the tip a narrowed beaklike lobe 3.5-4 mm. long, 2-3 mm. wide, the lateral lobes rounded, the whole rather dome-shaped; stamens 5-10 mm. long, the filaments tinged bluish, glabrous; anthers bluish, 1 mm. long; follicle 12- 13 mm. long, appressed-pubescent with long hairs, beak 2.5 mm. long. 1967 | MUNZ, ASIAN SPECIES OF CONSOLIDA 199 TYPE COLLECTION: Sipikordagh toward Bendola, Turkish Armenia, Sintenis 3177, 1890 (xk, P), seen. DIsTRIBUTION. Dry igneous places, 1200-1800 m., Turkish Armenia. ItLustraTION. Davis, Fl. Turkey 1: 129. 1965. REPRESENTATIVE MATERIAL. Turkish Armenia: E. K. Balls 1756; Davis 31941; Sintenis 1320; H. de Lesse, July 26, 1956; Guichard T/50/60. 39. Consolida sulphurea (Boiss. & Hausskn.) Davis, Notes Bot. Gard. Edinb. 26: 175. 1965. Fic. 5, L. Delphinium sulphureum Boiss. & Hausskn. in Boiss. Fl. Orient. 1: 81. 1867. Strigose, slender, 1-2 dm. tall, divaricately few-branched at or near the middle; lowest leaves largely withered by anthesis, midcauline trifid into revolute, strigulose lobes 8-20 mm. long; racemes lax, 2—5-flowered, strigu- lose; bracts linear, mucronate, 5—6 mm. long; pedicels 7-15 mm. long, the bracteoles near the flower, 4-6 mm. long; flowers sulphur colored; sepals 9-10 mm. long, oblong, strigose, obtuse; spur subcylindric, 5—7 mm. long; “petal” 3-lobed, ca. 8 mm. wide when opened, the median lobe with 3 lobules at the tip, the whole longer and narrower than the lateral lobes; stamens 5—6 mm. long, very wide at the base of the apparently glabrous filaments; anthers 0.6 mm. long, apparently yellowish; follicle on a re- curved pedicel, pendulous, white appressed-hirsute, lance-oblong, 8 mm. long, 2.5 mm. wide; seeds with distinct scales. Type: Turkey, near Marasch (Maras) at 1300 m., Haussknecht, pre- sumably (c), not seen, but apparent isotypes (P as of July 11, 1865, at 4100 ft., and K) were seen. DistRIBUTION. Grassy places and clearings, 750-1300 m., western Meso- potamia, Turkey. , SPECIMEN SEEN. Turkey: Maras-Gaziantep road, June 18, 1960, K. M. Guich- ard T/25/60 (x). 40. Consolida olopetala (Boiss.) Hayek, Fedde, Repert. Sp. Nov. Beih. 30: 313. 1924. Fic. 5, M. Delphinium olopetalum Boiss. Ann. Sci. Nat. II. 16: 364. 1841. Delphinium holopetalum in Huth, Bot. Jahrb. 20: 381. 1965. : Delphinium holopetalum Boiss. var. Boissiert Huth, Bot. Jahrb. 20: 381. 1895. rom Armenia, for which were cited: Sintenis 1890 and Bourgeau 1862. Both are C. olopetala. Plant to 3 dm. tall, 3—4-branched above the base, strigulose; lower leaves short-petioled, the main cauline 1-3 cm. long, 3-parted to base, the divisions with numerous setaceous lobes scarcely 0.5 mm. wide; bracts 3-parted into bifid lobes, or entire, 8-12 mm. long; raceme 15—20-flowered, 10-12 cm. long; pedicels arched spreading, 2-4 cm. long; bracteoles 2, near the flower, linear-subulate, 7-10 mm. long; sepals oblong, obtuse, violet-blue, strigose without, 12-15 mm. long, the lateral sepals strigose only on the midrib; spur incurved, longer than the “petal”, ie. 15-20 200 JOURNAL OF THE ARNOLD ARBORETUM mm. long and 3 mm. wide at the base; [voL. 48 “petal” more or less entire and sub- circular, pale, ca. 15 mm. in diameter, the spur ca. 16 mm. long; filaments violet, to 8 mm. long; anthers rounded, pale, 1 mm. long; follicle strigose, oblong, straight, erect, to ca. 1 cm. long, 3 mm. wide, the beak 3 mm. long. Type: Aucher in herb. Mus. Paris, ‘probably from Persia” (Pp); seen. DistrIBUTION. Upper Euphrates, probably confined to Turke ILLUSTRATIONS. HutTH, Bot. Jahrb. 20: pl. 7, fig. 10. 1895; Turkey 1: 129, 1965. ee Fl. SPECIMENS SEEN. Turkey: Davis 30951, 31909; Bourgeau 3. REPRESENTATIVE EXSICCATAE OF ASIAN SPECIES OF CONSOLIDA Agnew Sept. 1, 1961, saccata Aitchison 145 & 310, rugulosa; 1002, camptocarpa Androsow 2218, camptocarpa Antonow April 25, 1889, leptocarpa Aucher 64, deserti; 65, hohenackeri; 66, aconiti; 67, hellespontica; 71, cata; 4032, anthoroidea; 4033, tri- gonelloides Balansa 595, 1134, regalis subsp. panic- ulata; 599, 719, 856, 1135, raveyi; 729, axilliflora; 730, scleroclada; 854, glandulosa; 855, 1132, helles- pontica; 857, stenocarpa; 1133, orientalis Balls, E. K. 1441, pence 1756, Boissier June 1842, near Aphrodisias, raveyi; June 1842, near Laodicea, regalis subsp. paniculata; June 1846, near Raschaya, incana; June 1846, in Antilibanus, pusilla; May 1863, oe May 1866, axilliflora § losa; 4029, 13692, 13698, orientalis; . 1719, 4027, 13690, raveyi; 3221, scleroclada var. rigida: 4031, 13696, thirkeana; 11326, tomentosa Bourgeau 3, olopetala; 6, from Elmalu, sa; 6, from Baiboul, orien- talis; 7 (July 4, 1860), hellespon- tica; 7 (July 12, 1862), hohenack- eri; 8, raveyi; 10, regalis subsp. Bové 131, deserti Bunge 372, camptocarpa; March 1859, halophila; April 1859, rugwulosa; 1858-9, orientalis; between Tehran and Tabris, regalis subsp. divaricata Capus 54, barbata; 55, 56, rugulosa; May 1886, camptocarpa Chung 1608, ambigua Crookshank, Helen 155, 414, orientalis; 153, persica Davis, P. H. 4819, incana; 3221, 6477, 9992, 10030, 30973, 31592, 31688, scleroclada var ] 13010, 36622, raveyi: stenocarpa; 37229, 38698, 39136, 39207, subsp. paniculata; 16678, 25080, S1I0I7, 31732, 31860, — a 31026, peepee 22106, oligantha; * Bed go. caenes axilliflora; 29695, regalis subsp. regalis; 28945, 30079, 30598, 31992, orientalis; 30855, , cornuta; 30076, 30825, 32613, hohenackeri; 30951, 31909, olopetala: 35960, am- Dinsmore 6030, incana 1967] Donaldson 48, leptocarpa Falconer 59, ambigua Fedtschenko July 5, 1897, barbata; at Tashkent, /eptocarpa Field & Lazar 52, 371, 741, 770, oliv- ertana; 314, incana; 532, 1020 3297, anthoroidea; 3301, halophila; 2907, teheranica; 5341, 5360, camp- tocarpa; 5942, 5943, 6469, stocksi- Furse & Synge 292, orientalis; 625, 777, persica; 799, regalis subsp. di- varicata; 911, glandu losa Geld 4268, camptocarpa Gillett, J. B. 8305, oliveriana; 8011, ati? i; oligantha; 11207, persica: 11210, flav Gillett & ve 11621, oliveriana Gontscharow & Grigoriev 114, lepto- carpa; 871, barbata Gorz 857, 866, regalis subsp. paniculata Griffith 42, stocksiana; 1376, rugulosa Grown, L. S. 1923, haloph ila Guest 787, flava; 2470, 2757, 3742, oliveriana Guichard 142/62, aconiti; T 50/60, armeniaca; 25/1960, sulphurea Haines, R. N. W 1543, persica Haradjian 1263, 2266, anlitoras 1415, hellespontica; 1524, sulphurea; 1849, tomen motos R Wheels 467, 2100, oliveri- ; 1218, saccata Sea ke in Luristan in 1868, ho- henackeri; in Kurdistan in 1867, anthoroidea; Aleppo in 1865, azxilli- i regalis May 1867, oligantha; Marasch in 1865, sulphurea Hay 210, 282, stocksiana Hohenacker, Helenendorf. i in 1838, am- chan, July 1837, hohenackeri MUNZ, ASIAN SPECIES OF CONSOLIDA 201 Horner, S. 155, oliveriana Maimansour in Jacobs, M. 6812, oliveriana Karelin in Turcomania, camptocarpa Karelin & Kiriloff 1165, camptocarpa Khan, Prance & Ratcliffe 352, 414, hellespontica: 759A, glandulosa Koelz, W. 18202, oliveriana Komarov, V. L., Varsaminor, Aug 1893, barbata; Pendjakent in rf rugulosa Kotschy 23, 34, 222, axilliflora; 103, 737, pusilla; Aleppo, May 20, 1841, hellespontica; Raschaya to Palmyra in 1855, imcana Kraschennikov 146, camptocarpa thoroidea; 5126, Lace 3338, rugulos LS S71, “oliveriana; begets June 13, Lehmann 32, 36, paradoxa; 38, bar- bata. I am not certain whether these were collected by Lehmann or are his herbarium numbers Letourneaux, June 22, 1881, oligantha Levine 823, 952, ambigua Lindsay, Miss, 260, 1151, orientalis; 268, 329, biiaas 1424, regalis subsp. divaric Litwinow 440, a. 442, orien- talis; 4380, camptocarpa Manisdjian 652, hellespontic McNeill 336A, iSawontice: 336B, raveyi Meade, E. 171, oliveriana Merton 3884, rugulosa Meyer, F. N. 541, 572, orientalis Meyer & Dinsmore 4030, incana; 7238, deserti Meyers, F. T. 30, incana Peyron, at Zaur, May 24, 1890, oligan- tha; 1453, pusilla; Tarsous, June 1896, axilliflora Pichler 22, raveyi; in 1882, halophila Pinard, Syria, in 1846, hohenackeri: 202 JOURNAL OF THE ARNOLD ARBORETUM Caria in 1843, phrygia; Caria in 1843, hellespontica June 1889, hellespontica; May 1, 1900, pusilla Radi 307, hohenackeri Radde in 1867, orientalis Rawi 8540, oliveriana; 21818, oligan- tha; 20427, 20708, 20995, 21070, 21097, 21258, 22859, flava Isriyeh, regalis subsp. divaricata; 1011, 789, rugulosa; 2015, teheranica Regel at Tashkent, May 1880 and 52 in June 1881, barbata; Samarkand, May 1882 and Tschardschuia, Sept. 21, 1884, ta 1884, rugulosa; Turkestan, campto- carpa Richter 263, ‘ea ay Schischkin, Jun 1913, carpa; campto- June 1 6, i 16, oliveriana eerie SANTA ANA Botanic GARDEN 00 N. CoLLEGE AVENUE PaO CALIFORNIA 91711 [voL. 48 Schlagintweit 821, schlagintweitii Siehe 135, orientalis; 362, 403, cruct- ate: 175; ae 550, raveyi; 387, _ rig. Sintenis 143, ru pa ts 301, 1301b, aconiti; 283 axilliflora; 843, 1320, 3177, armeniaca; 1336, hohenackeri; 1131, oligantha; 1186, Saccata; 961, i oliveriana; ~~ raveyi; 2969, scleroclada var. 4547, ean hellespontica; thirkea bee 131, rugulosa Strauss, Th., 3 and July 15, 1896, oliveriana Uvazov, B. P., Jezireb, June 2, 1932, oligantha; Kirkuk, May 27, 1932, rugulosa; Khanagin, May 26, 1932, deserti; 95, regalis subsp. paniculata Wiedemann, Anatolia, regalis subsp. paniculata; Zejradin, in 1932, rugu- pre osa Wight 18, ambigua VoLuME 48 NuMBER 3 JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY B. G. SCHUBERT EDITOR T. G. HARTLEY C. E. WOOD, JR. CIRCULATION JULY, 1967 THE JOURNAL OF THE ARNOLD ARBORETUM Published quarterly by the Arnold Arboretum of Harvard University. Subscription price $10.00 per year. Volumes I-XX, reprinted, and back issues of volumes XXI-XLV are available from the Kraus Reprint CorPoraTION, 16 East 46TH STREET, New York, NY. 10017. Subscriptions and remittances should be addressed to Miss Dutcm A. PowELL, otp ArBorRETUM, 22 Drivintry AvENUE, CamBripce, Massa- CHUSETTS 02138. CONTENTS OF NO. 3 PRELIMINARY INVESTIGATIONS OF HYMENAEA COURBARIL AS A Resin Propucer. Jean H. Langenheim Comparative MorpHouocicaL Stupies In Drnientaceag, I]. THe Potten. William C. Dickison Tus Curomosomes or Austroparteya. Lily Riidenberg .............. 241 , CoronaNTHERA GRANDIS (GESNERIACEAE), A New Species FROM THE Sotomon Istanps. George W. Gillett A Gripes oF sem Aare Bemus or Daz SIUM : Sraicto. a ee ees — 249 With affection and appreciation this number of the Journal of the Arnold Arboretum is dedicated to LAzELLA SCHWARTEN, friend and col- league, on the occasion of her retirement from the staff. The assistance and the friendship which she has extended to all in her role as Librarian of the Arnold Arboretum and her special efforts in the circulation of this Journal will be greatly missed. JOURNAL OF THE ARNOLD ARBORETUM VoL. 48 Jury 1967 NUMBER 3 PRELIMINARY INVESTIGATIONS OF HYMENAEA COURBARIL AS A RESIN PRODUCER Jean H. LANGENHEIM DATA HAVE RECENTLY BEEN PRESENTED to suggest that Oligo-Miocene amber from Chiapas, Mexico, was produced by an ancestral population of the leguminous genus Hymenaea (Langenheim, 1963; Langenheim & Beck, 1965; Langenheim, 1966). Evidence for this source was obtained through comparison of infrared spectra of the amber and resin from living populations of Hymenaea courbaril L., and supported by the presence of Hymenaea leaflets in the amber. The evidence was further corroborated by the discovery that the amber had been deposited at or near mangrove vegetation, a habitat in which H. courbaril occurs commonly today (Lan- genheim, Hackner, & Bartlett, 1967). This determination of a leguminous source is in contrast to the classic concept of amber as generally derived from pines or at least from conifers. The usual preconception of pine origin of amber results from relatively detailed studies of the extensive early Tertiary amber deposits along the Baltic coast (Conwentz, 1890; Schubert, 1953, 1961; Langenheim, 1964). Inclusions in the Baltic amber of needles and staminate cones of pines usually have been cited as a basis for origin, although other coniferous remains, as well as certain angiospermous ones, are more abundant than those of pines. The most convincing evidence for pine origin is droplets of amber found in resin canals in the xylem and around parenchymatous cells of the inner living bark (Schubert, 1953, 1961). Several species of pines were thought to have been involved in production of the Baltic amber but these are now all treated as a single species, Pinus succinifera (Conw.) Schubert. The great quantities of amber found in these Baltic deposits, sufficiently extensive to have been mined for centuries on the Samland Peninsula, opened questions as to how so much resin was produced. Because pines commonly do not exude this quantity of resin “naturally,” it was assumed that a pathological condition known as “succinosis” must have existed (Conwentz, 1890). Not only did Conwentz and his followers think that pathological conditions were necessary, but that the entire forest must have been diseased to have produced such deposits of amber, “Es gab 204 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 kaum einen gesunden Baum im ganzen Bernsteinwald — das Pathologische war die Regel, das Normale die Ausnahme!”’ Czeczott (1960) has pointed out that this picture of the “sick amber forest” was so masterfully presented by Conwentz that his view has “hypnotized” workers for almost 60 years. Schubert (1953, 1958, 1961) thought that the abnormally intensified resin production might have been caused by a “strong dis- turbance of physiological balance,’ some of which might possibly have resulted from trees being exceptionally susceptible to injury near the periphery of their distribution. Only a few investigators such as Klebs (Kunz, 1903) and Bachofen-Echt (1930, 1949) have indicated that the Baltic amber accumulations could have resulted from normal production of resin. They thought that pines, dominant in forests over extensive areas for millions of years, could account for the large quantities of amber found in the Baltic deposits. Although the flora occurring in the Baltic amber has been known to contain a large tropical element, except for Schubert’s (1958) reference to warm temperate pines, no one seems to have recognized that trees produce more copious amounts of resin in tropical environments than in temperate ones. Likewise it has not been pointed out that angiosperms belonging to such families as Leguminosae, Burseraceae, Guttiferae, Anacardiaceae and Dipterocarpaceae can produce as much or more resin than gymnosperms. Moreover, many of these angiosperms seem to pro- duce large quantities of resin under natural forest conditions. At cae resin accumulates in the soil today in sufficient amounts to justify collection for commercial purposes. Thus the Chiapas amber project has directed our attention to a tropical angiosperm which has produced sufficiently large quantities of resin in an environment where it could become incorporated into the geologic record. The Chiapas amber likewise has an advantage in being chemically stable enough to allow comparison with resin from living populations. The opportunity to compare the fossil and modern resins has brought up various questions, such as what genetic and environmental conditions control chemical composition as well as yield of resin. What are the conditions in tropical environments that result in active biosynthesis as well as exudation of resin to the surface of the plant where large quantities may accumulate? In what tissues does biosynthesis occur and what are the variations in production during the ontogeny of the plant? To answer these questions, resin production in populations of Hymenaea courbaril is being studied ecologically in the field and morphogenetically in the laboratory. This paper constitutes a preliminary report on field studies and initial morphogenetic considerations of a planned long-range investigation. TAXONOMIC AND ECOLOGIC PERSPECTIVE The genus Hymenaea, subfamily Caesalpinioideae, family Leguminosae, has the center of its distribution in the Brazilian Amazonian region (Record & Hess, 1943). East African and Asiatic species of Trachylobium 1967 | LANGENHEIM, HYMENAEA COURBARIL 205 resemble Hymenaea and sometimes have been included in the genus, as have some South American species of Peltogyne. Twenty-seven species are commonly recognized; the number can be greater if several varieties of H. courbaril are considered worthy of specific recognition. Twelve species (H. davisit Sandwith, H. microcarpa Huber, H. palustris Ducke, H. oblongifolia Huber, H. parvifolia Huber, H. pororoca Huber, H. velu- tina Ducke, H. reticulata Ducke, H. intermedia Ducke, H. floribunda HBK., H. venosa Vahl, H. rubriflora Ducke) have their distribution centered in northern South America. They generally occur in the Ama- zonian region and adjacent areas in northern Brazil, Peru, Colombia, Venezuela, and the Guianas. Five species (H. stigonocarpa Mart., H. stilbocarpa Hayne, H. martiana Hayne, H. splendida Vogel and H. eriogyne Benth.) generally occur in southeastern Brazil (from Bahia to Rio Grande do Sul) as well as in adjacent Paraguay and Bolivia. Six species (H. adenotricha Ducke, H. multiflora Kleinh., H. sellowiana Hayne, H. chapadensis Barb.-Rodr., H. correana Barb.-Rodr., H. olfersiana Hayne) are recorded only locally. Hymenaea courbaril L. has a wide range of distribution, occurring in all of the South American countries except Argentina, Uruguay and Chile, on most of the islands of the West Indies, and along the Pacific drainage throughout Central America to central Mexico. Although this species generally appears to be distinctive from the other described species, con- fusion apparently does exist in distinguishing it from H. oblongifolia which occurs in the Amazonian area. H. altissima Ducke is often recog- nized as a variety of H. courbaril in southeastern Brazil. Hymenaea candolliana HBK. in Central America is so close morphologically that it frequently is considered either synonymous with or a variety of H A courbaril. Ducke (1949) indicates that the typical form of H. courbaril occurs in Colombia, Venezuela, the Guianas, the Antilles, and Central America as well as in the lower Amazonian area to the tributary Rio Tapanjés. However, he described the variety subsessilis from the region of Manaus and the variety obtusifolia from northeastern Brazil, partic- stigonocarpa (Tschirch & Stock, 1936). Detailed statements regarding resin production for the other species are not known to the author. Hymenaea courbaril can reach a large size, i.e., 100 to 180 feet high, with a spreading crown supported by a large trunk sometimes more than six feet in diameter. The trees are taprooted when young but later develop a considerable ramifying root system. The tree reproduces extensively from root suckers and thus persists following cutting. In secondary vegetation in Central America it commonly attains only 40 to 60 feet in 206 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 height and a diameter of 2 to 3 feet. Rarely today can one see the few remaining relatively undisturbed areas where these trees attain the large sizes reported by Freise (1934), Record and Hess (1943), et al. This species, Hymenaea courbaril, seems to occupy a relatively wide range of habitats throughout its extensive distributional pattern. In Mexico and Central America it occurs in moist evergreen forests as well as seasonally dry ones (Miranda 1952-53; Miranda & Hernandez-X., 1963; Gomez-Pompa, 1965; Langenheim, 1966). The moist evergreen forest in Mexico (Selva Alta Perennifolia sensu Miranda) occurs where the median annual temperature is greater than 20° C., the annual precip- itation greater than 1500 mm., and where there is a short dry period. Where precipitation is greater than 2000 mm., the dry period can be longer. Variations in this type of forest are numerous and characterized by some of the components tending to dominate in certain zones. H ymenaea courbaril occurs only sporadically in this forest type except in flooded areas or where soil moisture is abundant. For example, it grows amid such populations as Ficus spp. and Andira spp. which occur along banks or lowlands with rapid movement of water during flooding, with Tabebuia pentaphylla (L.) Hemsl. in lowlands with little movement of water during flooding, and with Pachira aquatica Aubl. on banks of muddy rivers which are flooded frequently. Some of these same riparian species also occur in the semideciduous types of forest (Selva Alta or Mediana Subcaducifolia sensu Miranda & Hernandez-X., 1963). These forests are characterized as ones in which 50 to 75 per cent of the trees lose their leaves during the peak of the dry season. The climate, like that of the Selva Alta Peren- nifolia, is characterized by a median annual temperature greater than 20° C. The annual precipitation is above 1200 mm., but there is an accen- tuated dry season. On the Pacific slopes in Mexico H. courbaril is one of the most characteristic trees in this forest type. Other populations occurring frequently with it are Enterolobium cyclocarpum (Jacq.) Griseb., Cedrela mexicana Roem., Licania arborea Seem., Cybistax donnell-smithii Rose, and Hura polyandra Baill. (Miranda & Hernandez-X., 1963). Actually “selvas de guapinol’’ occur along rivers or arroyos where soil is deep in southern Mexico (Miranda, 1952-53). Commonly occurring with H. courbaril are Andira inermis (Sw.) HBK., Nectandra spp., Ficus glabrata HBK., Calophyllum brasiliense Camb., Brosimum costaricanum Liebm. and Tabebuia pentaphylla. Among smaller trees (6-15 m.) the following are most frequent: Trichilia cuneata Radlk., T. hirta L., Coccoloba cozumelensis Hemsl., Annona reticulata L., Swietenia humilis Zucc., Bursera simaruba (L.) Sarg., Psidium sartorianum (Berg.) Nied., Inga laurina (Sw.) Willd., Styrax argenteus Presl, Genipa americana L. and Karwinskia calderoni Stand. Hymenaea courbaril often occurs in Sabanas, ie., grassy areas with scattered trees, in either the Selva Subcaducifolia or Selva Baja Sub- perennifolia. Here the soil is usually characterized by deficient drainage, ie, muddy in the rainy season and parched in the dry season. Licania 1967 | LANGENHEIM, HYMENAEA COURBARIL 207 arborea and Bursera simaruba are common associates of Hymenaea in these habitats. In certain stands of Selva Mediana Subcaducifolia Hymenaea courbaril can be sufficiently dominant to form essentially a consociation. It occurs in this manner along the Guerrero coast in Mexico on sandy soils derived from granitic parent materials. It likewise can appear in coastal sites on sandy plains or beach ridges intermingling with Selva Baja Caducifolia, i.e., with Bursera spp., Pistacia mexicana HBK., Ceiba sp., et al. In coastal areas such as these in Mexico, H. courbaril commonly comes in contact with mangrove swamps (Manglares). Hymenaea apparently grew near mangrove swamps during the Oligo-Miocene also in the Chiapas, Mexico, area, as evidence from a palynological study indicates that the Chiapas amber was deposited in or adjacent to mangrove vegetation (Langenheim, Hackner & Bartlett, 1967). In Costa Rica H. courbaril also occurs in moist to relatively dry forest types. In the province of Guanacaste, it is an important element in the forest classified as Tropical Dry (Holdridge, 1964). The dry sites in Costa Rica, however, are considerably more mesic than those in Mexico. It generally grows on fertile, well-drained soils and is frequent along rivers with such associates as Anacardium excelsum (Bert. & Balb.) Skeels, Enterolobium cyclocarpum, Cedrela mexicana, Tabebuia chrysantha (Jacq.) Nichols., and Achras sp. Hymenaea courbaril is a common component of the Subtropical or Premontane Moist forests (sensu Holdridge) in Costa Rica. Plots were made in this forest type near Villa Col6n in the province of San José at approximately 900 m. altitude. Here the median temperature is above 20° C; the precipitation is about 2000 mm. annually, but there is a pronounced dry season from mid-November through April. The forest studied was near a river where the vegetation has been disturbed by long human occupation. The dominant trees are H. courbaril, Brosimum sp., Tabebuia pentaphylla, Luehea candida (DC.) Mart. and Cupania gua- temalensis Radlk. (Fournier, personal communication). n the Osa Peninsula of Costa Rica, H. courbaril occurs in a dense mixed forest (Tropical Wet sensu Holdridge or Lower Montane Rain Forest sensu Allen) where it reaches a height of over 120 feet and with a buttressed base 4 to 6 feet in diameter. Here it grows with Virola spp., Vantanea sp., Vochysia spp., Carapa sp., Calophyllum brasiliense, and Copaifera trapezifolia Hayne. This forest type develops where the median annual temperature is greater than 24° C and annual precipitation 1s greater than 4000 mm. ; It is reported that Hymenaea courbaril occurs in both moist and dry coastal forests in Puerto Rico, as well as on other islands of the West Indies (Little & Wadsworth, 1964). et Despite the abundance and economic importance of H. courbaril in Brazil, little ecologic information seems available. Freise (1934) does state that in Brazil it occurs in a variety of habitats, from sterile, sandy soils along the coast to well-drained soils in the midst of Amazonian 208 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 forests. In the Amazonian region difference in size of the trees apparently can be correlated with edaphic conditions (Rodrigues, personal com- munication). The trees attain a greater diameter, although not necessarily greater height, on fertile floodplain soils than on relatively more sterile upland soils. One would expect Hymenaea courbaril to be differentiated into ecological races, since it has occupied different climatic and edaphic ares over a wide distributional range probably for long periods of tim phological differences have been recognized by description of eu hs bik no evaluation of the variation pattern has been attempted, as is true for most genera from this part of the world. Nor are data available regarding genetic and physiological differences which might provide illumination concerning both chemical composition and yield of resin. Even though H. courbaril probably is ecotypically differentiated, what effect this might have on resin composition is not definitely known. In- vestigations present varying results for species populations within different genera. Baker and Smith (1901, 1920) in pioneer studies of the chemical composition of Eucalyptus oils found that generally, except for perhaps half a dozen, all species possess comparatively constant chemical charac- teristics throughout their geographical distribution. Penfold and Morrison (1927), however, discovered some forms of Eucalyptus dives Schauer in which the essential oils varied in composition, and they concluded that these variations were due to genetic rather than environmental differences. In a later study, Penfold and co-workers (1948) also showed differences n the composition of the oils in physiological races of Leptospermum citratrum Shallinor. Krestinsky et al. (1952), found that turpentine of Pinus sylvestris L. from different ecological sites in forests possessed the same physical properties and chemical composition. Mirov (1948), however, indicates that P. sylvestris has several physiological races and questions whether or not Krestinsky dealt with only one of these. Mirov further points out that when the physical characters (reflecting the chemical composition) of turpentine of the genus Pinus are studied, some species vary but little throughout their range, whereas others vary a great deal. Pinus merkusit De Vries, which grows over a wide range in Burma and Indonesia, varies but little in the physical characters of its turpentine. On the contrary, the chemical composition of the turpentine of Pinus ponderosa Laws. (Mirov, 1958, 1961) varies with the several physiological races which have been identified. Mirov also noted that variation in chemical com- position of the turpentine may be a result of hybridization in some cases. CHEMICAL COMPOSITION OF THE RESIN The pale yellow to reddish resin (sometimes referred to as a resin-like gum) produced by Hymenaea has been known in the commercial trade as “South American Copal” to distinguish it from various copals from other parts of the tropics (Tschirch & Stock, 1936). As a result of the varied 1967] LANGENHEIM, HYMENAEA COURBARIL 209 usage in literature two semantic confusions need to be clarified. First is the distinction between gums and resins. These products are synthesized through different metabolic pathways. Chemically gums are hydrophilic polysaccharides. Resins (often called oleoresins or balsams) are complex mixtures of mono-, sesqui-, and di- and triterpenoids. These terpenoids are built on a fundamental isoprenoid pattern regarded as evolutionary modifications of a primeval mevalonic acid pathway (Robinson, 1963). Gums are readily distinguished from resins by their ability to dissolve in water or to absorb water to form a jelly or gelatinous paste, and by their inability to dissolve in hydrocarbons which are solvents for resins. Although a small polysaccharide fraction is present in the product from Hymenaea, this material primarily is composed of terpenoid constituents. In order to indicate this polysaccharide component, the product from Hymenaea justifiably can be called a gum-resin, but not a gum per se. Part of the confusion with the term “gum” results from lax usage in the commercial literature, but it also may be due to the tendency of some botanists to refer to resinous secretions from angiosperms as “gums.” Confusion also exists with the term “copal,” as it is used to connote resins from different plant families. The word probably has a Mexican origin, being derived from “copalli,’’ meaning incense. In Mexico, how- ever, the natives do not use “copal” or ‘‘copalli,” to refer to resins from members of the Leguminosae but rather to those from the Burseraceae. In the commercial resin literature, nonetheless, ‘‘copal” certainly is more commonly used for resins of various leguminous genera, particularly the African Trachylobium, Copaifera (including Guibourtia), Daniellia et al., as well as Hymenaea in Central and South America. Copal also is used for resin from the gymnospermous Agathis alba (Lamb.) Foxw. in the East Indies and Philippines. : Throughout the wide distributional range of H. courbaril, natives use the resin for incense and generally as a cement. Extensive commercial usage, however, seems to have developed primarily in Brazil. Here it has been employed chiefly in the manufacture of special grades of varnish and to a limited extent for medicinal purposes. Howes (1949) reports that Brazilian copal is recovered frequently from water courses and varies in size from a pea to that of a man’s fist. When dug from the ground, it may be in large, irregularly shaped pieces. Record and Hess (1943) indicate that “the gum gatherers dig around the roots and sometimes obtain a barrelful of copal lumps in a place, while the former site on a big tree, long since decayed, may yield several barrels of so-called fossil copal.” Although resin apparently has been collected from various areas in Brazil, the great collection center seems to have been in the Amazonian region — the Ilha do Marajé near Belém and along Amazonian tributaries such as the Tapanjés and Xingu. Hymenaea resin also was collected and used by pre-Columbian aborigines in southern Costa Rica and Panama (Balser, 1960). Noriega (1918) likewise reports the economic value of Hymenaea resin in Oaxaca, Mexico, where large masses again are found in the soil. 210 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 Early chemical analyses of Hymenaea resin have been made in this commercial context. Freise (1934) analyzed resin from H. stilbocarpa in three conditions: 1) freshly flowing, 2) fresh resin from a recent cut in bark, 3) buried in the soil. The terms “fossil’’ and “subfossil’’ are used frequently in the commercial literature to refer to resins buried in the soil. This is a different connotation from that generally employed by geologists in which prehistoric conditions are at least implied. At our present state of knowledge, the processes involved in fossilization of resin appear to be progressive oxidation and polymerization. At what point a resin actually is considered to be fossilized has not yet been defined. Freise discovered that between the free-flowing and the buried states, the resin becomes harder, darker in color, and considerably more insoluble in alcohol. The melting point range increases from 180 to 192° C. for fresh resin from a bark wound to 265° C. plus, for an old, buried resin. The resin acid content varies from 8 to 16 per cent in the fresh, flowing resin, 32 to 38 per cent from the wound in the bark, 50 to 55 per cent in young, buried resin to 55 to 68 per cent in the old, buried resin. This increase in resin acid content may be due to volatilization of components initially present. Nakano and Djerassi (1961) have made the only modern analyses known to the author on the resin acid constituents of H. courbaril. They investigated a single sample from Brazil and did not indicate whether or not it came from the tree or was buried in the ground. They reported that the acidic fractions of this sample consist largely of a diterpene acid which they named copalic acid. Copalic acid has the same carbon skeleton of agathanedicarboxylic acid and related diterpene acids (cativic, eperuic and labdanolic) and appears to correspond in absolute configuration to dehydroeperuic acid. These latter resin acids occur in Prioria copaifera Griseb., Eperua falcata Aubl., and Cistus ladanum 1! Le Maout & Decne Corroborative evidence that an ancestral population of Hymenaea courbaril is the source of Oligo-Miocene amber from Chiapas, Mexico, was provided by comparing infrared spectra of the amber with spectra of resin from trees living today (Langenheim & Beck, 1965; Langenheim, 1966). Infrared spectrophotometry offers a simpler means of relating chemical composition to the botanical origin of amber than other tech- niques now recorded. Because amber is so highly polymerized, only small percentages of the material are readily soluble. The infrared technique provides a means whereby the whole amber can be analyzed in the solid state by dispersing it in potassium bromide pellets. Also the large num- ber of absorption bands that amber and modern resin show offer a multi- plicity of parameters from a single experiment. In the study of the Chiapas amber, spectra were made from living trees whose ancestors might have produced the amber, such as: Styrax, Amyris, M yroxylon, Guaiacum, Pistacia, Bursera, Hymenaea, Protium, Pinus, Taxodium, etc. The spectrum for each genus was clearly distinguishable “This name, supplied by a colleague in chemistry, was used by LeMaout and Decaisne in the legend to an illustration and is cited in Index Londinensis. It is probable that the name Cistus ladaniferus L. was intended. Fic. 4. Mexico; «om cog 00150 WAVE NUMBER (CM-!) 1000 CHIAPAS, MEXICO AMBER H. COURBARIL RESIN (GUYANA) GIRON, COLOMBIA AMBE L i | | i ! | i 4 | 7 8 ? 10 1 12 WAVELENGTH (MICRONS) Infrared spectra of Oligo-Miocene amber from Chiapas, Mexico, and resin from extant Hymenaea courbaril from amber from Giroén, Colombia, and resin from extant Hymenaea courbaril from Guyana. 13 14 15 [L961 TIMVAUNOD VAVNAWAH ‘WIAHNAONVT 212 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 (Langenheim & tere 1965; Langenheim, 1966). Populations of certain species, as well as closely related species within a genus, were found to others, Langenheim and Beck (1965) indicate that the differences may be explained by the progressive effects of both oxidation and polymerization. In order to try to assess the variability in chemical composition of the hardened present-day resin, as expressed in infrared spectra, samples of H. courbaril resin were obtained from various localities throughout its wide distribution. Problems exist in obtaining resin samples that have adequate information substantiating their determination. Samples and voucher specimens were collected in the field in Mexico, Costa Rica, Museum of Harvard University, from Guatemala, Guyana, Venezuela, Bra- zil, Ecuador, and Dominica. There is no absolute assurance that the resin specimens aon the museums were produced by H. courbaril, as no voucher collections are available. Within the risk of a possible error, however, these resins are assumed to be properly identified. Samples also have been analyzed from different trees in the same local populations in the states of Guerrero and Chiapas in Mexico, and the provinces of San José and Puntarenas in Costa Rica. Likewise spectra of samples of oxidized resins have been run from different parts of the tree, i.e., trunk, branches, and roots, in addition to pieces buried in the soil. FIGURE 2 presents a repre- are generally similar to one another, although certain differences are apparent. As Langenheim and Beck (1965) have stated, the identity of two infrared spectra of pure compounds can be taken as proof of the identity of the compounds. But for resins, which consist of mixtures of components of high average molecular weight, two restrictions must be made: 1) only major constituents can be expected to give strong absorp- tion bands, so that differences in minor constituents may not be apparent at all; 2) identity of spectra can never be expected even from samples from a single specimen, but the presence of similarities, particularly in the upper “fingerprint” region is evidence for structural similarity of major constituents. The variations displayed may possibly be a reflection of a slightly different chemical composition due to ecotypic differentia- tion of the various populations. Certainly from the indications of Mirov’s work, ecotypic differentiation and consequent variation in composition might be expected in a species with such a wide distributional range and occurrence in different environmental conditions. Fresh resin, however, might reflect ecotypic variation better than the resin changed by oxida- tion and polymerization. * pines, the mono- and sesquiterpene fractions (turpentine) are useful, in some cases, in indicating racial differences (Mirov, 1961), whereas a ple portion of the remaining resin fraction 1967 | LANGENHEIM, HYMENAEA COURBARIL 213 WAVE NUMBER (CM7!) 4000 209020001500 1000 MEXICO VENEZUELA 1 1 1 1 L 1 1 7 8 9 10 VW 12 13 14 15 WAVELENGTH — (MICRONS) Fic. 2. Infrared spectra of resin specimens from Hymenaea courbaril popula- tions in Mexico, Guatemala, Costa Rica, Dominica, Ecuador, Brazil, Guyana, and Venezuela. has converted to relatively stable diterpene resin acids of the abietic type (Lawrence, 1959). For possible use of the oxidized resin in re- flecting specific or subspecific biochemical variation, further chemical anal- yses are needed. Apart from the possible influence of ecotypic variation, Beck (per- sonal communication, 1967) indicates that the general similarity of the spectra are in accord with the expectations raised by the gross structure of copalic acid proposed by Nakano and Djerassi (1961). The strongest carbonyl absorption in all cases lies at a wave number (1700 cm.~') as would be expected for a, 8-unsaturated carboxylic acids, but there is 214 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 evidence of lesser amounts of saturated carboxylic acids in all the samples. In most of them, this evidence is limited to changes in slope at the edge of the carbonyl band toward higher wave numbers but variations exist in some of the samples. This is in agreement with the opinion of Nakano and Djerassi, that the resin acids of H. courbaril are mixtures not only of stereoisomers, but of double-bond isomers. Beck likewise states that “a remarkable and quite unexpected feature of all the H. courbaril spectra is the weak but unmistakable absorption at approximately 700 cm.~! and 750 cm.~—'. This pair of bands is associated with mono-substituted benzenes. Their low intensity might lead one to suspect that these resins contain minor amounts of relatively simple aromatic compounds, but to our knowledge none have been isolated from copals by conventional chemical means.” Considerable variation in in- tensity of these presumably aromatic absorption bands also occurred in all samples tested. These possible chemical differences in the oxidized resin of H. courbaril are currently being investigated in more detail by analyzing extracts of the soluble fractions Amber from three other localities in South America gives spectra that likewise are similar to resin from H. courbaril. Amber from Girén, Colom- bia (near Bucaramanga) occurs in goldbearing alluvium and hence its geologic age is unknown. The spectra from six samples fall into two types, one of which is presented in Ficure 1. Six spectra of amber from Medellin, Colombia, also of unknown geologic age, were run and five are similar to the pattern indicated for the Girén amber. In both of these cases the amber spectrum was more similar to the resin spectrum of H. courbaril from Guiana than to other spectra run. No spectra of resin of H. courbaril from Colombia are available. Another amber sample was run from Capanéma, Para, Brazil, from strata which are considered early Miocene in age. The spectrum is simi- lar to resin from extant H. courbaril, especially from Para, Brazil, as well as from Costa Rica (Fic. 3). Variation in spectra shown through the geographic distribution of H. courbaril is also reflected in the amber. The spectra of the Chiapas amber are related more closely to spectra of Hymenaea resin from populations in Mexico and Guatemala than to those resins from populations with a more southern distribution. Therefore, it was not surprising when the spectra of amber from Girén and Medellin, Colombia, compared closely with H. courbaril resin from Guiana, and that the amber from Para, Brazil, was more comparable to the resin from Para, Brazil, than from other areas. PRODUCTION OF RESIN The location of the cells and tissues in which the resin is synthesized and the anatomical mechanism for its internal collection, as well as the manner of exudation, have not been adequately investigated in H ymenaea courbaril. Because secretion of the resin occurs in a different manner in 4000 3000 WAVE NUMBER (CM-!) 1500 1000 vita 800 HYMENAEA COURBARIL RESIN (COSTA RICA) PARA, BRAZIL l | | AMBER H. COURBARIL RESIN (BRAZIL) rifoe 7 8 WAVELENGTH 1 ee (MICRONS) Fic. 3. Infrared spectra of Miocene amber from Para, Brazil, and resin from extant Hymenaea courbaril from Para, Brazil, and from Costa Rica \ 12 ] 5 [L961 TIMVEANOD VAVNAWAH WIATHNAONV'I 216 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 young plants and mature trees, the secretory system is being studied from time of germination to year-old plants in the laboratory, and in older trees in the field. MATERIALS AND METHODS In order to study the secretory system in the young plants, seeds from Mexico, Puerto Rico, Panama, and Costa Rica were scarified, planted in sand or vermiculite and put in growth chambers (16 hour day length) at temperatures of 25° C. during the day and 21° during the night. After three months of growth in the control chambers, the plants were removed to the greenhouse. All tissue specimens of seedlings grown in either the growth chamber or greenhouse were fixed in 10 per cent aqueous acrolein (Feder, 1960) or in 3 to 6 per cent glutaraldehyde. The fixed specimens were dehydrated in a series of methoxyethanol, ethanol, n-propanol, and n-butanol, and then embedded in a monomer mixture of glycol methacry- late according to the technique of Ashley and Feder (1966). Sections were stained in aqueous 0.05 per cent toluidine blue O; in some cases they were counterstained with 1 per cent acid fuchsin or periodic acid-Schiff re- agent. All sections were mounted in Permount. Samples of branches and root suckers taken in the field in 1965 were fixed in FAA, dehydrated in an alcohol series and embedded in the pre- viously mentioned monomer mixture of glycol methacrylate. Samples of the mature tree taken in 1966 were fixed in 10 per cent formalin in the field and post-fixed in 10 per cent acrolein upon return to the laboratory. They were taken through the dehydration schedule discussed above, ex- cept that the final changes were made in anhydrous ether. Extreme diffi- culty in obtaining sections of the cambial zone in Hymenaea had been encountered previously. Therefore a modification of the previously re- ported embedding mixture of glycol methacrylate was devised by Feder. The monomer mix consists of: 1) glycol methacrylate, 91.85 per cent, 2) polyethylene glycol 400, 5.0 per cent, 3) ethylene glycol dimethacry- late, 5.0 per cent and a-azodiisobutyronitrile, 0.15 per cent. The speci- mens then were polymerized in 34 fl. oz. aluminum weighing pans at 60° C. overnight, followed by a 12-hour period at 100° C. It was possible to section blocks as large as 2 cm. X 1 cm., containing both wood and bark. Sections were stained with 0.05 per cent toluidine blue O and mounted in indopol polybutene H-1900, thinned to about 25 per cent in xylol in order for the sections to adhere firmly to the slide. If allowed to air dry, as with smaller sections, considerable curling and wrinkling occurred. With this technique it was possible to obtain intact sections of the cambial zone between 1 to 6 » thick with metal knives on a Leitz rotary microtome. It also provided means of sectioning the heterogeneous cell types of phloem and xylem tissue adjacent to the cambium. OBSERVATIONS AND RESULTS Seedling plant: Seedlings are not commonly observed in the field, 1967 | LANGENHEIM, HYMENAEA COURBARIL 217 although seeds appear to be produced abundantly. Seeds often are re- ported to be slow to germinate in the forest, but seedlings grow rapidly (Record & Hess, 1943). The author has observed seeds germinating in semideciduous, seasonally dry forests in both Mexico and Costa Rica just after the rainy season had begun. Seedlings were observed neither in the tropical wet forest on the Osa Peninsula of Costa Rica nor in the Ama- zonian forests near Belém or Manaus in Brazil. In none of these sites had trees obviously developed from seedlings. All small trees excavated were attached to roots from a large tree; thus they appear to be root suckers. Root grafts may have occurred, but this has not been substantiated. Record and Hess suggest that, because the seedlings are bushy and suc- culent at first, they are attractive to browsing animals. Grazing could account for lack of survival of seedlings in some sites, but seems im- probable in others. Explanation for the striking lack of H. courbaril seedlings in the forest awaits further observation and study. Time required for germination of seeds in the growth chambers varied considerably from the same locality as well as from different localities from Mexico, Puerto Rico, Panama, and Costa Rica. No attempt had been made to keep seeds from individual trees separate, although seeds were segregated by local populations in Mexico and Costa Rica. For example, the time of germination varied from 9 to 18 days from seeds from the same population in Guerrero, Mexico, under the same controlled conditions. Seeds from a population in Panama which germinated in 12 days in the 16 hour chamber in January, 1965, took two months in No- vember, 1965. If germination studies per se were done, some care probably should be given to segregating seeds from individuals within a local popu- lation and to knowing the age of the pods which are collected. At germination, the hypocotyl emerges and fleshy cotyledons are pushed above the ground. The root begins to develop rapidly, constituting approxi- mately two-thirds of the growth for about two weeks. At this point the stem portion of the hypocotyl starts to elongate, and the epicotyl and juvenile leaves begin to develop. By 16 days the stem above the juvenile leaves has developed as well as the first pair of leaves. Although there are some differences in the length of time for development of stem and leaf, by the twenty-first day their pattern is generally developed. Schizogenous pockets of limited size, not canals, occur in the epicotyl and in the hypocotyl to the place where root tissue begins. The pockets appear as early as three days after germination or, since the growth rate may vary, when the hypocotyl is approximately three cm. in length. In the cortex they occur one to two cells under the epidermis. They vary in number, sometimes being scattered through the cortex (Fic. 4), at other times being closely packed next to each other at least in portions of the hypocotyl or epicotyl. One to two layers of epithelial cells are com- mon with another layer of “collapsed” cells often outside the functional layer. The pockets also occur commonly in the pith, apparently scattered at random. The pockets vary in size, but the variation in a single section may be due to cutting the pockets at different levels. Because they have 218 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 an ovoid shape, a cross-sectional diameter will vary in any given section. Petioles also present a similar morphological topography to that of the hypocotyl and epicotyl. Resin seems to be synthesized early and more of it actually appears to issue from cells in the petiole than from the hypo- cotyl and epicotyl. FicurE 8 shows the general topography of the petiole. In the petiole, as well as the hypocotyl and epicotyl, the actively secreting cells are densely cytoplasmic and appear to be metabolically active (Fics. 6 and 7). In some cells, an expansion of the cell wall is evident (Fic. 5). This expansion suggests that the resin has been secreted but not yet passed through the cell wall. Unfortunately, in these sections most of the terpenes may have been dissolved in the dehydration sequence. Also no specific stain has presently been developed which is unique for the lower terpenes, so that their presence can be substantiated. Considerably more cytological study is needed to determine the site of biosynthesis of resins in cells such as these. Year-old plant: Periderm is initiated in the cell layer adjacent to the epidermis and forms inward into the cortex. Therefore, the secretory pockets formed early in the cortical region are preserved after the peri- derm has developed (Fic. 12). The extent of cortical parenchyma has de- creased, however, with considerable development of phloem fibers. Pockets are not apparent either in phloem or xylem tissue, although some pockets have been maintained in the pith. Camargo (1960) noted a similar ana- tomical pattern for one- to two-year old stems of H. stilbocarpa. The epithelial cells surrounding the pockets in year or year-and-a-half old plants do not appear to have the dense cytoplasm characteristic of the epithelial cells of younger tissue. These epithelial cells possibly are no longer synthesizing resin, but have already emptied their product into the pocket. Resin exudes from the cortical region when the stem is cut or scraped. Wound experiments, in which pin pricks and razor slashes were made on stems of one-year old plants, were carried out. Care was necessary in any wounding experiment not to destroy completely the narrow cortical zone containing the secretory pockets. Four months following the wound- ing, samples of the stem were fixed, embedded, and sectioned. No evi- dence for an increased number of pockets was noted, nor was it possible to determine with existing techniques if additional resin synthesis was initiated. Resin pockets have not been observed in root tissue of year-old plants. Root suckers: As previously mentioned, all of the young trees seen in Mexico and Costa Rica appeared to have developed as root suckers. Sec- tions of these suckers present a general anatomical pattern similar to that in the one- to two-year old stems developed from seeds. The secretory pockets, however, appear more numerous, particularly in the pith, than in the seedlings. Also considerably more resin seems to be secreted from a sucker than from a seedling stem of the same diameter. This increased production may result from the stem being attached to an extensive root 1967 | LANGENHEIM, HYMENAEA COURBARIL 219 system of a mature tree, where more reserve carbohydrates might be available for biosynthesis than in the seedling. Branches: In the field, resin which had exuded in small globules along the branch was commonly observed (Fic. 15). Similar to the one-year old stem, the pockets which had their origin in the cortex have been preserved after periderm formation (Fic. 11). The resin appears superficially to be issuing through lenticels; however, in sections the pockets are not shown to have an obvious relationship to lenticels. Von Hohnel (Kisser, 1954) suggests that with increase in size of the pockets (which has been observed) the resin spontaneously bursts out of the container which is close to the periderm. As the branch enlarges, considerable cortical paren- chyma is replaced by sclerenchyma, and resin is produced from cells in the cambial zone. Globules of resin also frequently appear near the attachment of branches to the trunk. Larger masses occur associated with larger branches. These masses seem to have occurred where breaks, due to strain at the point of attachment, allow issuance of the resin which probably is produced in the cambial zone. Fruit: Young fruits have been observed to produce such large masses of resin that they often encase the pods. Sections of fruits have not been made by the author. Ribiere (Moens, 1955) reports that fruit of Hy- menaea courbaril contains abundant resin pockets which form in the vicinity of the epidermis in the young ovary. These pockets persist and often are engorged with resin in the wall of the older fruit. Mature tree: The schizogenous pockets which develop in the cortex progressively disappear with the secondary development of the bark. The ark in older H. courbaril trees can be one to two inches thick or more. For the greater part of its thickness, the bark is highly sclerenchymatous, has relatively narrow rays, and a thin periderm. The tree can be suffi- ciently large and the bark so “tough” in the Amazonian region that canoes with a carrying capacity of 25 to 30 men can be made from a single tree (Record & Hess, 1943). External bark character varies in individuals. vertical fissures which may be due to release of tension during rapid growth. In some cases, the smooth bark is interrupted by rugose patches (Fic. 14) and resin commonly is found in these ridges and furrows. If a cut is made into either the trunk or root of the mature tree, resin exudes from the cambial zone. Resin is exuded in different amounts from trees of essentially the same size in the same general region, but growing in habitats with different availability of moisture. For example, resin was viscous and scantily produced from machete cuts in a tree on a dry, upland site near Villa Colon, Costa Rica. On the other hand, it was produced more abundantly and was considerably less viscous from a 220 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 similar-sized tree which grew in more mesic conditions along the river near Villa Colén. Resin begins to issue soon after cutting occurs. Flow- age has been observed to occur for at least six weeks, and probably longer periods could be recorded if there were opportunity for observation. Quantity of exudation depends not only on the capacity of the tree to synthesize the resin, but upon the viscosity of the resin. Also humidity conditions in the forest, which determine the rate of drying of the outer surfaces of the resin masses, are critical. In cases where the resin con- tinues to flow for long periods despite drying conditions, sufficient pres- sure exists to force the mass with the oxidized surface outward. Concen- tric flow layers are formed within the resin mass. Often during the rainy season, the resin is beaten from the surface almost as soon as it is exuded, unless a well-oxidized crust is rapidly formed. Although resin pockets were not noted in sections of root in the young plant, production of resin in the cambial zone of the mature root was observed. Numerous instances of small lumps of resin associated with breaks in the bark of the root were found (Fic. 16). Large masses, how- ever, were not discovered by the author to be associated with the roots. Cuts made in excavated roots produced resin copiously and the mature root appears to have a capacity to synthesize the resin similar to that of the trunk. Questions have arisen as to the role of resin production from the roots in accounting for the large accumulations of resin in the soil around the tree. The location of resin three feet or more beneath the soil surface has been cited by Noriega (1918), Record and Hess (1943), and others as indicating that resin production underground by roots is likely. Yet it seems difficult to understand the mechanism whereby large masses of resin are produced underground, unless considerable pressure pushes the resin out into the surrounding soil. From both observation and wounding experiments no doubt exists that roots have the capacity for synthesis, and that they are highly susceptible to types of injury that would result in exudation. Adequate assessment, however, has not yet been made of the relative roles of resin production by trunk and root and subsequent accumulation of resin in the soil. Since extensive accumulations of resin occur primarily around large trees, it might be assumed that these are old trees. Thus time would have been provided for resin that had fallen from the tree to have become incorporated in the soil. In tropical areas where Hymenaea grows, however, the trees do not develop annual growth layers so that determination of the age of trees is difficult. Also, little is known about the rate of soil formation in these types of forest. Leaf litter is strikingly lacking and there is no basis for knowing how long it would take to incorporate lumps of resin that have fallen from the tree to any depth within the soil. In areas in Chiapas, Mexico, the resin was observed to have flowed into sand, cementing the grains into large lumps to a depth of at least six inches, On the Osa Peninsula of Costa Rica large pieces of resin had obviously fallen from long (up to 2 feet) resin stalactites hanging from the trunk and other masses incrusting the bark around fissures. Some resin lumps were incorporated in the surface 1967 | LANGENHEIM, HYMENAEA COURBARIL 221 of the soil above the level of the roots and hence were probably produced from the trunk. Sections made initially of trunk and root tissue where resin was issuing did not show evidence of pockets; probably as a result of separation of the samples at the delicate cambial zone. Sections using the methods developed by Feder show the development of pockets or cavities of vary- ing size in the cambial zone (Fics. 9 and 10). These pockets form be- tween the rays in a zone of undifferentiated xylem tissue. The cells in the zone where the cavities develop appear not to have become lignified as indicated by the red-violet color of the staining of the toluidine blue O. Lignified walls are blue-green in color and occur in tracheid and vessel cells just inward from the zone where the pockets form. The pockets apparently develop by cells breaking down and releasing resin into the cavity which is being enlarged, with subsequent breakdown of additional cells. The origin of these cavities would probably be considered lysigenous as compared with schizogenous separation of cells in the young plant (Esau, 1964). These pockets apparently can merge to form larger cavi- ties, a process shown beginning in Ficure 10. Longitudinal sections of the particular specimens examined thus far have not shown if these cavi- ties can attain any sizable vertical extent. The particular secreting layer of cells with characteristically dense cytoplasm that occurs in schizogenous pockets of the young plant, is not evident in the lysigenous pockets. These cavities apparently develop when a sufficient break in the bark has dis- turbed the cambium. As yet it is not known if resin production in the mature tree is restricted to this cambial zone or if individual parenchyma cells are capable of production in the newly differentiated phloem as well as in the rays. The relatively quick exudation of the resin after cutting seems to indicate that individual cells have already synthesized the resin or that precursors are present so that the final stages of synthesis are accomplished rapidly following injury. It seems unlikely that the large quantities of resin observed can be produced from a single, relatively narrow zone of pockets along the cambium as shown from the sections presented here. More extensive cavities probably are developed or at least biosynthesis must be kept in a very active state for a relatively long period to produce the large accumulations of resin found. These conditions need further investigation before an understanding of the anatomical organiza- tion for resin production in Hymenaea courbaril is reached. A pressing question raised at this point is whether or not pathological conditions are necessary for production of resin in the mature tree. As just pointed out, schizogenous resin pockets arise early as a part of the normal pattern of development in the young plant, but disruption of tissues in the cambial zone seems necessary for development of pockets in the mature plant. Even though disturbance of the cambial tissue occurs, diseased or physiologically weakened conditions are not indicated. On the contrary, the largest, most vigorous trees growing in what appear to be the most favorable environmental conditions (i.e., excellent water 222 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 relations and fertile soils) are clearly those observed to be producing the most resin. Insect damage has often been suggested to account for injury that might increase resin production. Stephen L. Wood, a specialist in the study of Central American bark beetles, reports that beetles of the Scoli- cidae are the most likely to produce damaging effects on a tree such as Hymenaea (personal communication, 1966). He, however, has not wit- nessed these beetles injuring H. courbaril in his experience in Central America. In fact, on the Osa Peninsula of Costa Rica he has observed a species of Platypus (a platypodid ambrosia beetle) that started tunnels it never completed, having been forced out by resin. He also noted that profuse quantities of resin were produced in large trees. In one case he states that deposits “more than an inch deep, four inches wide and a foot or two in length” were removed from an area of longitudinal furrows and ridges (Wood, personal communication, 1966). He further indicates that all tissues were healthy, with no hint of insect, fungal or other damage than forced separation of tissues from internal pressure. He offers the explanation that H. courbaril appears to grow slowly as a subdominant, but when an opening occurs, growth is extremely rapid until maturity is reached. During this period growth might be so rapid that the bark can- not accommodate the expansion in certain areas of stress and bark rup- tures. Thus Wood’s observations support those of the author’s in that the large quantities of resin appear to be produced by healthy trees growing rapidly, and that insect damage apparently does not play a role in inducing pathological conditions. The apparent individual variation in resin production in Hymenaea courbaril might be explained from evidence on comparative yields in cer- tain pine populations. Resin yields have been shown to be under genetic control in Pinus palustris Mill., P. caribaea Morelet and P. elliotti En- gelm. Snow (1949, 1954) reports that certain individuals of P. caribaea and P. elliottii yield two and one-half times as much resin as that of the average tree of the same size growing under similar conditions. No single or combined external characteristics, or environmental conditions, were discovered to set these trees apart. Squillace and Dorman (1961) and Squillace (1966), in breeding experiments on the variation and inheri- tance of resin yield in P. elliottii, have shown that this trait not only varies highly among individual trees, but is strongly inherited in com- parison with most biological traits. A strain of slash pine has been de- veloped that will yield about twice the normal amount of resin of aver- age trees. Their results also suggest that resin yield is positively corre- lated with growth rate. Squillace points out that, although they selected mainly for resin yield in their experiments, the high resin producing strain also yields about 12 per cent more volume growth than normal strains. Dyer (1963) states that the quantity of resin fremeseirees ei is directly proportional to the size of the crown of slash Boudreau and Schopmeyer (1958) have neh that physical factors re- lating to resin flow from wounds in Pinus elliottii are number and size of 1967 | LANGENHEIM, HYMENAEA COURBARIL 223 resin ducts, pressure, and viscosity. Both resin pressure and viscosity are under strong genetic control. Duct size and number vary with stem age. Pressure and viscosity seem unaffected by stem size above 5 cm. diameter but the ratio of pressure to viscosity accounts for a substantial amount of yield variation. They also indicate that resin pressure in trees varies with environmental conditions, particularly with water relations. Thus these studies on some pines indicate that resin yield is under genetic control, but likewise environmental conditions are important in producing favorable growth conditions that affect pressure and viscosity. These data also suggest that vigorously growing trees, rather than diseased, physiologically weakened ones, give the greatest resin yields. DISCUSSION AND SUMMARY Preliminary investigations of Hymenaea courbaril have provided evi- dence that at least some angiosperm resin producers living under tropical conditions do not need to be diseased or physiologically weakened by insect, fungal or other attacks to produce large quantities of resin. Resin is produced and can accumulate in sufficient quantity to account for large deposits of amber. Although these studies do not necessarily provide an- swers to the accumulation of the Baltic amber deposits, they do turn our attention to certain important aspects of resin production by angiosperms as well as gymnosperms. In order to assess the capacity of members of various species of Hymenaea to produce resin, as well as to account for differences in chemical composition of the resin, further systematic analy- sis of the group is needed. Particularly, more ecologic and genetic data in terms of the development of physiological races would be valuable. Studies of the unoxidized resin (i.e., the volatile terpene fractions) from populations of H. courbaril might show differences in composition related to ecotypic differentiation. Striking similarity of infrared spectra of oxi- dized resin from specimens throughout the wide range of distribution of H. courbaril may result from chemical stability of the diterpene copalic acid constituent. This copalic acid also may provide the basis for re- lating the present-day H. courbaril resin to amber of Oligo-Miocene age from Mexico, Miocene from Brazil, and unknown age from two localities in Colombia. In the anatomical adaptation for resin production in Hymenaea cour- baril, schizogenous pockets form in cortical and medullary tissue of the epicotyl and stem (but not the root) portion of the hypocotyl soon after germination. These pockets persist until bark formation finally obliterates them. In the older plant, lysigenously developed cavities appear in the cambial zone of the trunk or root. It seems unlikely that the narrow zone of pockets in the cambial zone can entirely account for the large quantities of resin that are produced, even allowing considerable time for production. Although not observed yet, more extensive cavities may be formed by enlargement of the existing ones. Whether or not parenchyma cells in the living phloem or in the xylem rays also synthesize resins is 224 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 still questionable. There also is question as to whether continual stress of some sort is needed to keep the enzymes active for continued biosyn- thesis when lysigenous breakdown is involved. Although not demonstrated as yet, it seems likely that resin yield in Hymenaea courbaril is under strong genetic control, as has been shown for several species of pine. Local populations of H. courbaril in particular geographic regions have been noted for outstanding resin yields by com- mercial resin collectors. This fact, however, does not sort out the genetic factors from the environmental ones. The largest, most vigorously grow- ing trees have been observed to produce the greatest amounts of resin, which possibly may be due to excess carbohydrate available for synthesis of a secondary product. Therefore, superimposed upon genetic capacity, all environmental factors that result in greater net photosynthesis, and hence possible higher growth rates, may increase synthesis. Partial ex- planation for the large number of different kinds of trees producing resin, which occur in tropical regions, as well as particularly high resin yields, may be provided by the favorable growth conditions throughout the year. Experimental work is needed to substantiate these possibilities. Likewise, vigorous growth activities of these tropical trees may help account for means of exudation which augment those resulting from injuries due to storms, insect and fungal infestations, etc. This leads then to the raison d’étre for biosynthesis of resin and its possible evolutionary significance. The metabolic role of these terpenoids which constitute resins still is ob- scure; the prevailing idea is that they simply represent ways of disposing of excess acetate. Thus the primary significance of these terpenoids may be ecologic, in providing protective adaptation against such damage as is inflicted by insects and fungi. Abnormal flows of resin are known to be stimulated by wounds and by the presence of fungal disease. Also terpe- noids can act as excellent fungicides and insecticides and their presence in a tropical environment, where both fungi and insects are abundant, could provide a selective advantage. With more studies of the chemistry of these terpenoids making up resins, their biogenesis, and their discharg- ing processes, however, a metabolic role may possibly be discovered that would augment the ecologic, protective role. ACKNOWLEDGMENTS Grateful acknowledgment is made for funds to support laboratory work for this study through National Science Foundation Grants GB—1312 and 2397 and the Maria Moors Cabot Foundation of Harvard University. Support of field work during the summer of 1965 in Mexico and Costa Rica was supplied by the Organization for Tropical Studies, and during the summer of 1966 in Brazil by the Maria Moors Cabot Foundation. Thanks are due many individuals who made this study possible. Professor Louis Fournier provided field assistance, as well as specimens and eco- logical data, from Villa Colén, Costa Rica. Professor Arturo Gomez- Pompa, Biol. Javier Valdes-G. and Mr. Rudolph Mesa-A. gave valuable 1967 | LANGENHEIM, HYMENAEA COURBARIL 225 assistance in the field in Guerrero and Chiapas, Mexico, as did Mr. Tim- othy Plowman in Manaus, Brazil. Mr. Gil Hillman, Mr. Anthony Pla- ceres, and Mr. Sumner Slavin prepared the infrared spectra. Mrs. Alex- andra Bartlett, Mrs. Ruth Freeman, and Mrs. Patty Shepard gave tech- nical assistance with the histological preparations. Dr. T. P. O’Brien and Dr. Margaret McCully offered helpful comments regarding histological techniques; special appreciation, however, is due Dr. Ned Feder for his aid in developing a technique to embed and section wood and bark from mature trees of Hymenaea. Appreciation is also expressed to Professor I. W. Bailey and Professor R. H. Wetmore whose stimulating discussions led to a pursuance of this project and plans for its continuance. To them and to Professor Curt Beck, Professor E. S. Barghoorn and Dr. Bernice Schubert thanks are also due for helpfully criticizing the manuscript. LITERATURE CITED Asuiey, C. A., & N. Feper. 1966. Glycol methacrylate in histopathology. Arch, Pathol. 81: 391-397. BACHOFEN-Ecut, A. 1930. co Bernstein und seine Einschliisse. Verh. Zool.- Bot. Ges. Wien 80: 35-4 . 1949. Der Bernstein ae seine Einschliisse. Wien. 204 pp. Baker, R. T., & H. G. SmirH. 1901. On relation of leaf venation and the presence of certain chemical constituents in the oils of Eucalypts. Jour. Proc, Roy. Soc. New S. Wales 35: 116-123. _ 1920. Research on the Eucalypts especially with regard to their essential oils. New S. Wales Tech. Mus. Ed. Ser. 13. Sydney. 471 pp. BALtseEr, C. 1960. Notes on resin in aboriginal Central America. Akt. 34. In- ternatl. Amerikanisten Kong. Wien. 374-380. Boupreav, P. F., & C. S. SCHOPMEYER. 1958. (Abs.) The inheritance of exuda- tion pressure and viscosity of resin in slash pine. PI. Physiol. 33 (suppl.): 50. 30s CAMARGO, PON, 1960. Contribucio para o conhecimento da area de “Hymenaca suibesaete” Hayne. Univ. S. Paulo Fac. Filos. Ci. Let. Bol. Bot. 17: 1-105. Conwentz, H. 1890. Monographie der baltischen Bernsteinbaume. Danzig. 51 Ciscae H. 1960. The flora of the Baltic amber and its age. Prace Mus. Ziemi 4: 119-145. Ducxe, A. 1949, As Leguminosas da Amaz6nia Brasileria. Bol. Téc. Inst. Agron. Belém. 18. Dyer, C. D. 1963. Naval stores products. Agr. Ext. Serv. Bull. Univ. Georgia Coll. Agr. 593. Athens. 28 Esau, K. 1965. Plant Anatomy. John Wiley & Sons. xx + 7 FEDER, N. 1960. Some modifications in conventional ee. ey tissue prepa- ration. Jour. Histochem. Cytochem. 8: 309 Fretse, F, W. 1934. Brasilianische Pflanzendrogen des Welthandels. Tropen- pflanzer 37: 469-486. Gomez-Pompa, A. 1965. La vegetacién de México. Bol. Soc. Bot. Méx. 29: 76-119 226 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Hotpripce, L. R. 1964. Life zone ecology. Tropical Science Center, San José, Costa Rica. 125 + vi pp. [57 pp. dl. Howes, F. N. 1949. Vegetable gums and resins. Chronica Botanica, Waltham, Mass. 188 pp. Kisser, J. 1958. Die ‘Acccheideus von Atherischen Olen und Harzen. Handb. der rhea 10: 91-13 Kunz, G. K. Precious stones. Min. Res. U. S. 911-975. KRESTINSKY, by "e MatevskatA, & F. SoLtopy. 1932. Influence of geographic factors and forest types on composition of gum turpentine. Zhur. Prikl. Khimii 5: 950-957. LANGENHEIM, J. H. 1963. Informe preliminar sobre los estudios botanicos del r de Chiapas. Resimenes, Segundo Cong. Mex. Bot. San Luis Potosi, aes : 1964, Peet status of botanical studies of ambers. Bot. Mus. Leafl. 20: 225-287 . 1966. Botanical source of amber from Chiapas, Mexico. Ciencia 24: 201-209. & C. W. Beck. 1965. Infrared spectra as a means of determining botan- ical source of amber. Science 149: 52- . HAacKner, & A. BARTLETT. 1967. Mangrove pollen at the deposi- tional og of Oligo- Miocene amber from Chiapas, Mexico. Bot. Mus. Leafl. n pres LAWRENCE, R. V. 1959. Oxidation of resin acids in wood chips. Tappi 42: 867-869. Littte, E. L., Jr., & F. H. WapswortuH. 1964. Common trees of Puerto Rico and the Virgin Islands. U. S. Dep. Agr. Handbook 249. 548 MriranpA, F, 1952-53. La vegetacién de Chiapas. Dep. Prensa y Turismo, Tuxtla Gutierrez, Chiapas, Mexico. 2 vols. pp. 334 & E. HERNANDEz-X. 1963. Los — de vegetacion de México y su clasificacién. Bol. Soc. Bot. Méx. 28: 29-179. sea N. T. 1948. The ean = relation to the biology) of genus Pinus. . Rev. Biochem. 17: 521 1958. Distribution of oe components among species of the genus Pinus. The physiology of forest trees. Ed. K. V. THIMANN. 251-268. 961. ape of gum turpentines of pines. Pacific Southwest ae Range Exp. Sta. U. S. Dep. Agr. Forest Serv. Tech. Bull. 1239. acean P. 1955. Les formations sécrétrices des copaliers congolais. Cellule 57: 5-59. NaKAno, T., & C. oe 1961. Terpenoids XLVI. Copalic acid. Jour. Org. Chem. 26: 167-1 Norteca, J. M. 1918. u cere Bol. Direc. Estud. Bidl. 2: 357-363. PENFOLD, A. R., & F. ORRISON. 1927. The occurrence of a number of varieties of Pais ese as determined by the chemical sn Ris of the essential oils. Jour. Proc. Roy. Soc. New S. Wales 61: 254- & H. H. G. McKern. 1948. Studies of Aces forms of the Myrtaceae. Pt. 1. Leptospermum citratum. Mus. Tech. Appl. Sci. Res. Essential Oils Australian Flora 1: i. ——— = J.. & R. W. Hess. 1943. Timbers of the New World. Yale Univ. s, New Haven. 640 pp. Minot T. 1963. The organic constituents of higher plants. Their chemistry and interrelationships. Burgess Publ. Co., Minneapolis, Minn. 306 pp. 1967 | LANGENHEIM, HYMENAEA COURBARIL 227 ScHusert, K. 1953. Mikroskopische Untersuchung pflanzlicher ama des Bernsteins. II. Rinden u. Borken. Paleontographica oe 119. 1958. Sobre el ambar baltico y los pinos de El Salvador. oo. Inst. Trop. Invest. Ci. Salvador 7: 51-55. 1961. Neue Untersuchungen iiber Bau und Leben der Bernsteinkiefern (Pinus succinifera (Conw.) emend.). Beih. Geol. Jahrb. H. 45. Nieder- sachsische Landesamt fiir Bodenforschung, Hannover. Snow, A. G., Jr. 1949. Research on the improvement of turpentine practices. Econ. Bot. 3: 375-394. . 1954, Progress in development of efficient coo methods. U. S. p. Agr. Forest Serv. Southeast. Forest Exp. Sta. P a A. E. 1966. Planning tree improvement Sena at Olustee. AT- FA [American turpentine farmers association cooperative] Jour. 28(5): 11-13. . W. DorMAN. 1961. Selective breeding of slash pine for high oleo- resin yield and ress characters. Recent advances in Botany, Univ. Toronto Press. 2: 1616-1 Tscuircy, A., & E. ee 1933-1936. Die Harze. Ed. 3. Berlin. 2 vols. in 4. pp. 418, 1858. Divison or NATURAL SCIENCES UNIVERSITY OF vpn SANTA CRUZ CALIFORNIA 9506 EXPLANATION OF PLATES PLATE I : AS. Tee sections of Hymenaea courbaril MS iopanet 4, topog- near epi is; 5, resin pocket from epicotyl (FIG. 4) n near ede th dis- tention of cell walls i in secreting cells; 6, resin ae in pith “Of petiole, (FIG. 8) showing densely cytoplasmic secreting cells; 7, resin pocket cortex of petiole; 8, topography of petiole showing extensive essen of resin pockets in cortex and pith, Line scale on all figures is 50u. PLATE II _ Fics. 9-12. Transverse section of Hymenaea courbaril. 9, pockets developed In cambial zone of of mature tree, Belém, Brazil; 10, canara aoa to FIG. 9 but showing coalescence to form cavi jes; 11, ; Brazil, ranc showing preservation of pockets in the cortex amid extensive develop: ent of sclerenchyma; 12, one-year old stem, Mexico, showing preserva ion of pockets after the periderm starts to develop. Line scale on all figures is 50. PLATE III Fics. 13-16. Field wigan or in which Hymenaea courbaril is found. 13, typical mature tree with s h bark, tropical dry forest, Guanacaste, Costa Sod fess tree ag rugo ie ial tches in bark, subdeciduous forest, Guerrero, Mexico; 15, branch showing exuded balls of resin, Guerrero, ao 16, injured root with beeciates small masses of resin, Guerrero, Mexic LANGENHEIM, HYMENAEA COURBARIL PLATE II > 48 Jour. ARNOLD Ars. VOL. & & eRe ygn's ¥ YL kom oety sheneins tee Sagas? 9 bh pe. : CFx. ee URBARIL MENAEA CO GENHEIM, Hy LAN Jour. ARNOLD Ars. VOL. 48 LANGENHEIM, HYMENAEA COURBARIL 1967 | DICKISON, DILLENIACEAE, II 231 COMPARATIVE MORPHOLOGICAL STUDIES IN DILLENIACEAE, II. THE POLLEN WILLIAM C. DICKISON IN A PREVIOUS PAPER (Dickison, 1967) it was indicated that since a comprehensive comparative morphological investigation of the Dilleni- aceae had never been undertaken, such a study might prove very reward- ing in the attempt to discover the phylogenetic relationships of this interesting tropical family. The lack of prior morphological work on the Dilleniaceae is particularly evident in the area of pollen morphology. This can be attributed, in part, to the comparatively recent application of palynological data in solving taxonomic problems. Erdtman (1952) gives a limited description of three genera and seven species of Dilleniaceae. From this treatment he concluded that “pollen morphology seems to support the assumption that Dilleniaceae should be related to Polycarpicae.” In the only other significant reference to dilleniaceous pollen, Barth (1962) described in detail the pollen of four species belonging to the genera Davilla, Doliocarpus, and Tetracera. It was this worker’s opinion that the family represented a parallel group to the Magnoliaceae from which it had previously been thought to be derived. The present study was carried out in order to provide evidence to be evaluated from all organs and parts of the plants in an effort to deter- mine the affinities of the Dilleniaceae. MATERIALS AND METHODS Pollen Collections (HPC), Pan American Oil Co., and the Rancho Santa Ana Botanic Garden (RSA) and (3) herbarium specimens obtained from the University of California, Berkeley (UC); Missouri Botanical Gar- den, St. Louis (MO); New York Botanical Garden (NY); and the United States National Herbarium, Washington (US) to whom I am especially grateful for providing material of the rare genus Acrotrema. Pollen was prepared by the standard acetolysis method outlined by Erdtman (1960). Material was subsequently mounted in glycerine jelly. This procedure generally left grains with a darkened exine which made staining unnecessary. In a few cases, however, a basic fuchsin dye was employed with good results. Difficulty was encountered in removing the 232 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 protoplasm from grains preserved in formalin-acetic acid-alcohol (FAA). I m The fact that pollen prepared by the acetolysis method is larger than pollen prepared by other means (e.g., lactic acid, KOH, etc.) is now well established (see Canright, 1953). Carlquist (1961) is of the opinion that not enough emphasis has been placed on the fact that pollen morphology is highly influenced by methods of preparation. In a recent discussion of this problem, Whitehead (1965) also stresses the need for a uniform mounting medium; however, he does not recommend glycerine jelly. Despite opposition to the acetolysis-glycerine jelly method, the ease of preparation, wide application in palynological research, and generally excellent results obtained, justify its employment in comparative mor- phological investigations. With the above discussion in mind, in addition to personal experience, I concur with Canright (1963) that pollen size and shape are the most unreliable of diagnostic features. In order to minimize usage of complex and confusing terminology, the suggestions of Faegri and Iversen (1964) are followed. All measurements and descriptions were made under oil immersion (> 1000). Size dimen- sions were determined by measuring at least twenty grains from each sample. No effort was made to treat the numerical data statistically. The dimensions are, accordingly, only intended to indicate relative size ranges. Pollen descriptions of putatively related families were obtained for the most part from Erdtman (1952) OBSERVATIONS 1. Dillenia L. (Fics. 1-3) The pollen grains of Dillenia are the most variable in the family with respect to both sculpturing and aperture type. SHAPE: Oblate, ea spheroidal or spheroidal; circular to semiangular in polar view. SIzE: thickness to ektexine. ScULPTURE: Scabrate to reticulate-rugulate to most frequently reticulate. APERTURE: Triporate in D. indica and D. philip- pinensis; tricolpate in D. turbinata, D. reifferscheidia, D. alata, D. excelsa, and D. papuana; tricolpate with rare or occasional tetracolpate grains in D. suffruticosa, D. ovata, and D. luzoniensis. T riporate grains have elon- gate pores and are provided with a well-defined annulus. Tricolpate and tetracolpate forms have granular furrows which may or may not have opercular membranes and margo. Furrows extend about three-fourths the length of the polar axis. *P refers to dimension of polar axis, E the length of equatorial axis. 1967 } DICKISON, DILLENIACEAE, II 233 2. Hibbertia Andr. (Fics. 4-6, 21) SHAPE: Oblate to spheroidal to prolate spheroidal; circular to sub- angular in polar view. Size: The smallest grains were recorded for H. salicifolia (18.9% X 17) and the largest for H. stricta (34 % 32.2,). Avg. size for all species examined, ca. 25.24 (P) X 27 (E). STRUCTURE: Tectate; endexine equal in thickness to ektexine. ScuLPpTURE: Foveolate (e.g., H. stellaris and H. salicifolia) to most frequently reticulate. APER- TURE: Tricolpate, furrows granular, often provided with opercular mem- brane and margo. Furrows either quite distinct and wide (10), in which case they extend the entire length of polar axis (e.g., H. cuneiformis and H. tetrandra), or else, not well defined and extending about two-thirds the length of polar axis (e.g., H. hexandra). Erdtman (1952) describes H. acicularis as tricolporate, a condition which was not observed in any of the samples examined in this study. 3. Pachynema R. Br. (Fic. 9) SHAPE: Spheroidal to prolate spheroidal; circular in polar view. Size: ca. 16.44 (P) & 16.1p (E). Structure: Tectate or semitectate; endexine equal in thickness to ektexine. ScuLPpTuRE: Foveolate to reticulate. APERTURE: Tricolpate, furrows granular and not well defined, extending entire length of polar axis. _ 4. Schumacheria Vahl (Fic. 8) SHAPE: Oblate spheroidal; mostly rectangular to spheroidal in polar view. Size: ca. 15y (P) 16.54 (E). Structure: Tectate; endexine equal in thickness to ektexine. APERTURE: Tetracolpate, rarely tricolpate; furrows poorly defined and extending about one-third the length of polar axis. Slight thickening of ektexine at margins of furrows. 5. Acrotrema Jack (Fic. 7) SHaPE: Oblate spheroidal; semiangular in polar view. Size: ca. 18y (P) X& 21 (E). SrructurE: Tectate; endexine equal in thickness to ektexine. ScutptureE: Finely reticulate. APERTURE: Tricolpate, furrows granular and provided with opercular membranes. 6. Didesmandra Stapf (Fic. 10) SHAPE: Oblate spheroidal; mostly rectangular to subangular in polar view. Size: ca. 25.74 (P) X 28» (E). Srructure: Tectate; endexine equal in thickness to ektexine. SCULPTURE: Reticulate. APERTURE: Tet- racolpate, occasionally tricolpate; furrows extend about one-third the length of polar axis. 7. Curatella Loefl. (Fics. 16, 17) SHapeE: Spheroidal to prolate spheroidal to occasionally prolate; circular in polar view. Size: ca. 234 (P) X 20 (E). Structure: Tectate; 234 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Bigs: 71-21. epee pollen (all ca. X 1000). 1, Dillenia bE (cult. SING s.n.), r view showing tectate structure. 2, Dillenia indica (cult. : I s.n.), polar view. 3, the same, equatorial view of elongate pore. 4, Hib- ertia pee ad 2981), al view de epicting tectate structure and prominent opercular membranes. 5, ertia scandens (cult. K s.m.), polar view. 6, thé Same, equatorial view. 7, ales trema bullatum (US 1576874), polar view in two cal levels, note granular ranes. 8, Schumacheria castaneifolia (Abey- em eh ta s.n.), polar view of tetracolpate grain. 9, Pachynema dilatatum (NT 9), polar view in two focal levels showing reticulate exine. 10, Didesmandra aspera (Burtt 2540), polar view showing reticulate exine. 11, Tetracera fagifolia ( 70), polar view showing tectate oo joe etracera asiatica ( HPC 1835), polar view, note opercular membra villa phe th) “(AP 2953), polar view of tricolporate grain. 14, Davilla *sunthit (UC 9 polar view in two focal levels showing nature of reticulum. 15; ASciebece nuit 1967 | DICKISON, DILLENIACEAE, II 235 endexine equal in thickness to ektexine. SCULPTURE: Finely reticulate. APERTURE: Tricolpate, furrows provided with opercular membranes and margo, extending entire length of polar axis. A limited number of grains observed were questionably tricolporate but the occurrence of this feature could never be satisfactorily proven. 8. Davilla Vandelli (Fics. 13, 14) SHAPE: Spheroidal to subspheroidal; circular to semiangular in n polar view. Size: Avg. size for all species examined, 26.2n (P) X 25p (E). STRUCTURE: Tectate; endexine equal in thickness to ektexine. SCULPTURE: Reticulate to coarsely reticulate (e.g., D. kunthii). APERTURE: tricol- porate; furrows provided with an operculum. Pores granular, spheroidal to slightly elongate. The pollen of D. rugosa has previously been described and diagrammed as tetraporate (Barth, 1962). My observations show that this species possesses three well-defined furrows and is, therefore, tricolporate. 9. Doliocarpus Roland. (Fics. 15, 19, 20) SHaPE: Spheroidal to prolate spheroidal to prolate; circular to semi- angular in polar view. Size: Avg. size for all species examined, 24» (P) < 20.64 (E). Structure: Tectate; endexine equal in thickness to ektexine. Scutpture: Finely reticulate to reticulate. APERTURE: Tri- colporate, furrows may or may not be provided with a margo extending three-fourths the length of polar axis. Pores indistinct due to opercular membrane, generally circular in outline. 10. Tetracera L. (Fies. 11, 12, 18) SHAPE: Spheroidal, prolate spheroidal or prolate; circular in polar view. Size: Avg. pollen size in this genus can be correlated with geo- graphical distribution. The largest grains occur in the New World species (ca. 26.5p (FP) & 227 (6)), the smallest in the Indo-Malayan species (P) & 20.4p (E)). Structure: Tectate; endexine equal in thickness to ektexine. Scutprure: Finely reticulate to reticulate. ApertTURE: Mostly tricolporate, tricolpate in T. alnifolia; furrows and pores granular, pro- vided with opercular membranes, pores generally circular in outline. SUMMARY OF POLLEN MORPHOLOGY OF DILLENIACEAE The pollen grains of the Dilleniaceae vary in shape from oblate to spheroidal to prolate. The smallest grains were recorded for Schumacheria castaneifolia (154 X 16.54), while the largest occurred in Hibbertia (HPC 2962), polar view in two focal —— ay, — americana (Irwin 5470), polar view. 17, the same, inepagea Ta , Tetracera alnifolia (MO 1598748), equatorial view of tri icolpate grai with lash margo. 19, Dolio- ral ded major (HPC 6272), ss erage Nt choeite well defined pore in granular furrow. 20, the same, polar view. , Hibbertia stellaris (HPC 1837), equa- patty view wining elongate furrows. 236 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 stricta (344 X 32.2u). The median size for all genera examined was ca. 22.4u (P) X 21.7p (E). Pollen was exclusively tectate or rarely semitectate, with no significant difference in thickness between endexine and ektexine. Sculpturing is predominantly reticulate or a modification of reticulate. Four basic aperture types can be recognized as existing within the family: triporate, tricolporate, tricolpate and _ tetracolpate. Segregation of tricolpate and tetracolpate genera is not always strict because of species in which both conditions exist. Likewise, separation of tricolpate and tricolporate aperture types is often difficult owing to the presence of granular opercular membranes. DISCUSSION Dilleniaceous pollen is characterized by a relatively large number of morphological types. Unfortunately, the family cannot be successfully divided into tribes or subfamilies on the basis of pollen morphology. This is due to the variation within and overlap between genera in diagnostic features. Briefly stated, morphologists now generally agree that the monosulcate pollen grain, with a distal solitary furrow, represents the primitive con- dition in dicotyledons (Eames, 1961, p. 161). Pollen of this nature is found only in dicot families which are known to have general ranalian affinities. In contrast, the more common tricolpate dicotyledonous grain has three meridional furrows or modifications thereof. It has long been difficult for botanists to explain the derivation of tricolpate pollen from the monosulcate type. Wodehouse (1936) suggested a derivation from a spore bearing a triradiate crest. Such an origin was originally sup- ported by the morphology of certain ranalian pollen grains (Schisandra). Wilson (1963) has offered an alternative explanation by theorizing that the tricolpate grain may have resulted from a trichotomosulcate condi- tion (as found in some Canellaceae) by loss of furrow contact at the distal pole. Meeuse (1965) believes that the distal extension of the fur- rows in some tricolpate dilleniaceous pollen supports the latter hypothesis. Since trichotomosulcate pollen was not observed in any Dilleniaceae, such presumptions can only be looked upon with skepticism. There are at present no widely accepted trends for pollen specialization above the tricolpate level, although some have been suggested for indi- vidual families (e.g., Dahl, 1952). Therefore, it is difficult to assess phylogenetic relationships within the Dilleniaceae on this basis. If it is assumed, however, that the tricolpate grains with very extended, wide furrows (as illustrated in the woody hibbertias) are the most primi- tive pollen types in the family, phylogenetic specializations could be hypothesized. Advancement in one line could possibly have led to the tricolporate grain, while in another direction a continuing reduction of furrow length would arrive at the triporate condition. It might also be suggested that there was a trend for the addition of an extra furrow which culminated in the tetracolpate forms. Species with both tricolpate and tetracolpate pollen stand as intermediate in this advancement. 1967 | DICKISON, DILLENIACEAE, II 237 Admittedly, there is no sound basis for this sequence. Since it would be dangerous to correlate pollen morphology with data derived from other parts of the plant, it should be emphasized that the data obtained from palynology alone does not provide evidence for any trends of pol- len specialization in the Dilleniaceae. Few similarities can be observed between the pollen of the Dilleniaceae and that of the ranalian complex of families. This is opposed to the opinion of Erdtman (1952) who advanced such an alliance on the basis of palynological evidence. Pollen grains of the Magnoliaceae and Canel- laceae are monosulcate or modifications of monosulcate types (trichoto- mosulcate and dichotomosulcate). Pollen of the Calycanthaceae varies from monosulcate to zonaperturate (belted) to bicolpate to infrequently tricolpate (Bailey, 1960) and is reticulated or baculate. Winteraceae pollen is shed in tetrahedral tetrads whereas that of Annonaceae is non- aperturate, monosulcate, or occasionally shed in coherent tetrads. A survey of other ranalian pollen forms reveals not a single family which closely approximates the Dilleniaceae in aperture type, size, and/or sculpturing. Within the Parietales (sensu Engler & Prantl, 1893), the eurypalynous Sterculiaceae and the Eucryphiaceae with bicolpate (syncolpate), very small grains offer no foundation for considering these families closely allied to the Dilleniaceae. The relationship of the Dilleniaceae to the Brunelliaceae, Ochnaceae, Connaraceae, and Theaceae is not as conclusive. These families all re- semble the Dilleniaceae in pollen morphology in having either tri (tetra) colpate or tricolporate pollen, although unusual bilateral, subisopolar, tetracolpate grains are encountered in one subfamily of Connaraceae. Pol- len exines in these families are psilate to reticulate with the exception of the baculate Brunelliaceae. Size dimensions are also in approximate agree- ment. Information gathered from pollen morphology in determining the affinities of Actinidia and Saurauia is likewise inconclusive. Actinidia is characterized by tricolporate pollen which has an indistinct psilate exine. Saurauia pollen is quite variable as evidenced by the occurrence of tricol- porate, triporate, and tetra (col) porate grains, as well as coherent tetra- hedral tetrads. Exines are all predominantly psilate. The pollen of Crossosoma is tricolporate or bicolporate, prolate, and the grains have reticulated exines. Paeonia pollen is tricolporate, with smooth furrows but pores covered by granular membranes, prolate spheroidal and reticulated. The size of the pollen in these two genera closely approxi- mates that of the pollen of Dilleniaceae. In conclusion, the data presented from pollen morphology does not provide a definite clue to the phylogenetic relationships of the Dillenia- ceae. Nevertheless, I believe the comparatively specialized tricolpate, tri- colporate, triporate, tetracolpate, and reticulated pollen of the Dilleniaceae clearly shows more resemblances to that of members of the Theales (or Guttiferales) than it does to ranalian families. 238 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ACKNOWLEDGMENTS I am extremely thankful to the following individuals who provided me either with preserved or dried specimens and thus made my study of the Dilleniaceae possible: Prof. B. A. Abeywickrama (Peradeniya, Cey- lon), Dr. P. S. Ashton (Kuching, Sarawak), Dr. O. M. Barth (Rio de Janeiro, Brazil), Mr. B. L. Burtt (Edinburgh, Scotland), Mr. G. Chip- pendale (Alice Springs, Australia), Dr. Hj. Eichler (Adelaide, Australia), Dr. S. L. Everist (Brisbane, Australia), Dr. G. W. Gillett (Honolulu, Hawaii), Dr. R. D. Hoogland (Canberra, Australia), Dr. H. S. Irwin (New York, U.S.A.), Dr. H. Keng (Singapore), Mr. M. R. Monsalud (Laguna, Philippines), Dr. J. V. Pancho (Laguna, Philippines), Dr. J. M. Pires (Brasilia, Brazil), Dr. R. L. N. Sastri (Waltair, India), Mr. T. Smitinand (Bangkok, Thailand), Dr. B. C. Stone (Kuala Lumpur, Malaya), Sir George Taylor (Kew, England), Dr. C. L. Wilson (Han- over, New Hampshire, U.S.A.), Mr. J. S. Womersley (Lae, New Guinea), also the Directors of the Singapore Botanical Gardens and of the Forest Research Institute, Kepong, Malaya. In addition, I am grateful to the various institutions who loaned the herbarium material used in this study. It is a pleasure to acknowledge the help of the Department of Botany, Arizona State University, for making facilities available during the course of this investigation. This study was partially supported by Grants GB 3820 and GB 4127 from the National Science Foundation to Dr. James E. Canright. The author wishes to thank Dr. Bernice G. Schubert for reading the manuscript and offering many helpful suggestions. MATERIAL EXAMINED ? Acrotrema bullatum Thw, Creyton: US 1576874. Curatella americana L. BraziL: Irwin 5470 (Ny). PANAMA: J. F. McBride 2825 (us). Mexico: UC 1249323. GuatemMALA: MO 1091094. Davilla aspera (Aubl.) Benoist. BritisH GuIANA: A. C. Smith 2184 ieee = 1836. Trintpap: W. E. Broadway eget aber Panama: MO 906 Davilla elliptica St. Hil. Botrvia: MO 181 Davilla kunthii St. Hil. sails RIca: ps i“ ‘Skutch 4091 (cH), HPC 2949. CoLomsia: UC 9635 Davilla multiflora (DC.) “5 Hil. Mexico: MO 127842 see cess Poir. Mexico: E. Matuda 0902 (cx), HPC 2953. Peru: MO 108213. *In additicn to these institutions previously aera maateria) studied was ob- tained from the Arnold Arboretum, Harva ministration, Alice Springs (NT ); Herbarium of Pom Sar awak Museum, Kuching (SAR); and fhe. Botanic Gardens, Singapore (SING) 1967 | DICKISON, DILLENIACEAE, II 239 Didesmandra aspera Stapf. SARAWAK: Burtt & Woods B.2540 (£). SARAWAK: SAR 18297 8 : Doliocarpus dentatus (Aubl.) Standl. Mexico. Yucatan: Gentle 1418 (GH), PC 2962. Costa Rica: Skutch 3829 (cH), HPC 6271. Cotompia: UC Doliocarpus guianensis (Aubl.) Gilg. Costa Rica: MO 1152277. , Doliocarpus lasiogyne Benoist. BraziL: Hatschbach 3580 (us). Brazit: UC 1088323. Doliocarpus major Gmel. PANAMA CANAL Zone: P. White 127 (cH), HPC 6272. Panama: MO 1189259; MO 1758246. BraziL: MO 1255920. Doliocarpus olivaceus Sprague & Wms. Panama: Yale 11. Dillenia alata (DC.) Martelli. Papua: Brass 7371 (a), HPC 2996. Hoogland 8521 (CANB). Dillenia excelsa (Jack) Gilg. Borneo: Castillo 619 (A), HPC 2954. NortH Borneo: Cuadra A1019 (us). KEP s.n. Dillenia indica L. AUSTRALIA: Cult. BRI s.n Dillenia luzoniensis (Vidal) Martelli ex Dur. & Jacks. PHILIPPINES: J. P. Rojé CLP) s.n. Dillenia megalantha Merr. Puitippines: Sulit PNH 6377 (A), HPC 2957. Dillenia ovata Wall. ex Hook.f. & Thoms. INpocHINA: Squires 775 (a), HPC . SINGAPORE: Cult. SING s.n. Dillenia papuana Martelli. NEw GUINEA: Darbyshire & Hoogland 8039 (CANB). Dillenia parviflora Griff. THAILAND: BKF s.n. Dillenia philippinensis Rolfe. Px1t1pprnes: Lambert & Brunson 39 (us); US 1861993; J. V. Pancho s.n. Dillenia reifferscheidia Villar. PHILIPPINES: Paniza PNH 9407 (A), HPC 2960. Dillenia suffruticosa (Griff.) Martelli, PHILIPPINES: Fenix 92 (cH), HPC 2961. NortH Borneo: Clemens 9504 (a), HPC 2999. SINGAPORE: Canright 978; Cult. SING s.n. Dillenia turbinata Finet & Gagnep. Hatnan: Ko 52211, HPC 1838. Hibbertia acicularis (Labill.) F. Muell. AustraL1a: Camfield (POM). Hibbertia baudouinii Brongn. & Gris. New Cateponta: US 2192565. Hibbertia cuneiformis (Labill.) Gilg. Cult. K s.n. Hibbertia dentata R.Br. Cult. K s.m. Hibbertia hexandra C. T. White. New SoutH Wates: Hoogland 8585 (CANB). Hibbertia linearis R.Br. ex DC. New SoutH WALES: Hoogland 7747 (CANB). Hibbertia salicifolia F. Muell. New Cateponia: Viellard 62 (a), HPC 2968. Hibbertia scandens (Willd.) Dryand. Austraia: C. T. White 8237 (a), HPC 2979. Cult. K s.n. Hibbertia stellaris Endl. AUSTRALIA: Pritzel 268 (cH), HPC 1837. Hibbertia stricta (DC.) R.Br. ex F. Muell. AUSTRALIA: Giblin H19 (a), HPC 2981 Hibbertia tetrandra (Lindl.) Gilg. Cult. E C3544. Hibbertia vestita A. Cunn. AUSTRALIA. New South Wales: Boorman (PoM). Pachynema junceum Benth. AustRALta. Northern Territory: NT 6750. Schumacheria castaneifolia Vahl. CEYLON. Waga: Abeywickrama s.n.; HPC 2983. Tetracera alnifolia Willd. Conco: MO 1598748. Tetracera euryandra Vahl. Matava: UC 390344. 240 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 igen fagifolia Willd. ex Schlecht. Panama: J. M. Johnston 624 (a), HPC 627 tek a ‘tadica (Houtt. ex Christm. & Panz.) Merr. SINGAPORE: H. Keng s.n. Tetracera korthalsii var. subrotunda (Elm.) Hoogl. Borneo: NY 21376. Tetracera macrophylla Wall. ex Hook.f. & Thoms. Matava: UC 243404. Tetracera portobellensis Buerl. Mexico: MO 1810305. Tetracera scandens — vin err. PHILIPPINES: Ahern’s Coll. 104 (a), HPC 2991. HaInan: UC 2785 Tetracera volubilis L. ee Schunke 168 (a), HPC 2995. British Honpuras: LITERATURE CITED Baitey, I. W. 1960. Some useful techniques i 2 he study and interpretation of pollen eee Jour. Arnold Arb, 41: —148. Bartu, O. M. 1962. Catalogo sistemAtico ae Nilens das plantas arboreas do Brasil ieridional II. Monimiaceae e Dilleniaceae. Mem. Inst. Oswaldo Cruz 60: 405-420. Canricut, J. E. 1953. The comparative morphology and relationships of the Magnoliaceae. II. Significance of the pollen. Phytomorphology 3: 355- 365. 1963. Contributions of — oe to the phylogeny of some nale ean families. Grana Palyn. 4 Gistocae. S. 1961. Pollen morphology of Ses Aliso 5: 39-66 Daut, A. O. 1952. The comparative morphology of the Icacinaceae. VI. The pollen. Jour. Arnold Arb. 33: 252-286. Dicxison, W. C. 1967. Comparative ov a studies in Dilleniaceae, I. Wood anatomy. Jour. Arnold Arb. 48: Eames, A. J. 1961. Morphology of the page, McGraw-Hill. New York. ENGLER, A., & K. Pranti. 1893. Nat. Pflanzenfam., Leipzig. ERpTMaN, G. 1952. Pollen morphology and plant taxonomy I. Almqvist & Wiksell. Stockholm 1960. The acetolysis method. A revised description. Sv. Bot. Tidskr. 54: 561-564. Farert, K., & J. Iversen. 1964. Textbook of pollen analysis. Hafner Publ. Co., New Yor MEeusg, A. D. J. 1965. Angiosperms — past and present. Advancing frontiers of plant nese 11, Institute for the Advancement of Science & Culture. New Thi WHITEHEAD, D. R. 1965. Pollen men ata in the Juglandaceae, II: Survey of the family. Jour. Arnold Arb. 46: 369-410. Witson, T. K. 1964. rier cach nae a of the Canellaceae. III. Pollen. Gaz. 125: 192-1 Bot Witkeasivi R. P. 1936. : Wiidliah of pollen grains. Bot. Rev. 2: 67-84. DEPARTMENT OF BIOLOGY VirGINIA POLYTECHNIC INSTITUTE BLACKSBURG, VIRGINIA 24061 1967 | RUDENBERG, AUSTROBAILEYA 241 THE CHROMOSOMES OF AUSTROBAILEYA Lity RUDENBERG THE PHYLOGENETIC RELATIONSHIPS AMONG the families of the primitive Ranales of Engler and Prantl have been of great interest to systematic botanists. A comprehensive knowledge of this group is vital for an under- standing of the evolution of early angiosperms. The relative taxonomic position of the families and their included genera within the Ranales has been difficult to determine primarily because the primitive characteristics of these plants are often paired with advanced or reduced specialized ones. To the present, a majority of studies on the group has been con- cerned with and based on macroscopic observations and on anatomical- morphological features. Now, it is important to supplement this infor- mation by an investigation of the chromosome numbers and karyotypes of some of these plants. No karyological studies exist for some of the families of this complex. No chromosome counts have been published for any members of the Amborellaceae, Austrobaileyaceae, Gomortegaceae, Gyrocarpaceae, and Trimeniaceae (Raven & Kyhos, 1965). The present note is a report on the chromosome number and karyotype of Austrobaileya. Plants of Austrobaileya were first collected in 1929 by Kajewski in Northern Queensland. White (1933) described A. scandens from this material, From the beginning, the taxonomic position of this plant posed difficult problems. It was studied by White (1933, 1948) and Croizat (1941, 1943) and its taxonomic position was changed several times. Bailey and Swamy (1949) discussed the merits of these changes on the basis of detailed investigations on the morphology and anatomy of the plant. Their material included specimens from the two species A. scandens and A. maculata C. T. White. The genus is composed of only these two species and is placed in a separate family, the Austrobaileyaceae, which is con- sidered an independent ranalian family with close affinity to the Moni- miaceae. MATERIALS AND METHODS Seeds of Austrobaileya sp. were collected near Ravenshoe, North Queensland, Australia, Webb & Tracy 6301. Four seeds of this collection were sown at the Arnold Arboretum in 1964. Only one seed germinated almost a whole year later. The seedling was potted in 1965 and cuttings made in 1966.1 It must be mentioned that there is some difficulty in The assistance of A. J. Fordham, propagator of the Arnold Arboretum, is grate- fully acknowledged. The author is greatly indebted to Dr. L. I. Nevling, Jr., for his advice and reading of the manuscript. 242 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 species determination and in Bailey and Swamy’s opinion there is a question “. . . . whether there actually are two distinct species of over- lapping ranges.” On the basis of vegetative characteristics, our plants seem to resemble most closely the isotype of A. scandens; up to now these plants have not flowered. Squashes were prepared from actively growing roots. The root-tips were pretreated for four hours in cold oxyquinoline (0.002 Mol.) or, for the purpose of comparison, in 0.05 per cent colchicine. They were fixed overnight in three parts absolute ethyl alcohol to one part glacial acetic acid, hydrolyzed in 1 N. HCl for 25 minutes, stained with leuco-basic fuchsin and squashed in diluted aceto-carmine. Slides were made perma- nent following the freezing technique of Conger and Fairchild (1953). Observations were made with a Zeiss phase microscope. CYTOLOGY A somatic chromosome number of 2n=44 for Austrobaileya is shown in Ficure 1. In the cell photographed most chromosomes are clearly became detached from the chromosomes at the secondary constrictions. Furthermore, there are two metacentric chromosome pairs in the somatic complement, each pair of a different length. The majority of the remain- ing chromosomes have subterminal centromeres. However, the proportion of short arm to whole chromosome length is not the same in all of them. A comparison of the individual lengths of some of the homologues may be slightly misleading because of the high magnification of the photo- graph by which small differences due to pressure and stretching are greatly multiplied. Nevertheless, based solely on observations of their similar poepeeey some of the subterminal chromosomes appear to be present in fours. DISCUSSION The chromosomes of Austrobaileya are highly asymmetrical and rep- resent a specialized karyotype in their unequal size. Most woody angio- sperms have small chromosomes with relatively few variations in size. The chromosomes of a number of ranalian families are generally larger 1967 | RUDENBERG, AUSTROBAILEYA 243 ie tay 4) 889333 CECT ALGO ES 68 8888 2 OR TREE 8a ge “IG, ABovE: Karyotype reconstructed from a cell of Austrobaileya. BELow: pt cell at mitotic metaphase with 2n 44 chromosomes, 50 Oo © (Stebbins, 1938). Still larger chromosomes, however, of more equal size are found in Gymnosperms, Gnetales, and Cycadales. Many specialized monocots have relatively large chromosomes of unequal morphology in the same gametic complement together with much smaller chromosomes. 244 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Such size differences are in some cases greater than those observed in Austrobaileya. A chromosome count of 27=44 is rare among families of the ranalian complex. For two genera of the Monimiaceae this chromosome number has been reported. In pollen mother cells at meiosis the chromosomes pair to form 22 bivalents in Laurelia novae-zelandiae A. Cunn. (Hair & Beu- zenberg, 1959). A species of the genus Kibara from East New Guinea has 2n=44 chromosomes (Borgmann, 1964). No karyotype of this species has been published and, therefore, it is not yet possible to determine whether the chromosome complements of Austrobaileya and Laurelja and Kibara show an affinity or whether their equal numbers represent an insignificant coincidence. A similarity of karyotype and number could reflect a taxonomic relationship of these genera which would be particularly meaningful in woody plants which are cytologically more stable than herbaceous plants, LITERATURE CITED Batrtey, I. W., & B. G. L. Swamy. 1949. The morphology and relationships of Austrobaileya. Jour. Arnold Arb. 30: 211-226. BorGMANN, E. 1964. Anteil der a eeaner in Flora des Bismarckgebirges von Ostneuguinea. Zeitschr. Bot. 52: pee A. D. & L. M. Fatrcuitp. 1953. a cee freeze method for making ar slides permanent. Stain Technology 28: 281-283. os, L. 1941. Notes on the Dilleniaceae and their allies: Austrobaileyeae 0 . 1943. New families. Cact. Succ. Jour, 15: 64. Har, J. B., & E. J. BeuzeNBerc. 1959. Contributions to a chromosome atlas of the New Zealand flora— 2. New Zealand Jour. Sci. 2: 148-15 Raven, P. H., & D. W. Kyuos. 1965. New evidence concerning the original basic chromosome number in angiosperms. Evolution 19: 244-248. STEBBINS, G. L., Jr. 1938. Cytological Sipe ge associated with the dif- ferent growth habits in the dicotyledons. Am. Jour. Bot. 25: 189-197. Wuite, C. T. 1933. Ligneous plants alleen for the Arnold Arboretum in North Queensland by S. F. Kajewski in 1929. Contr. Arnold Arb. 4: 1-29. . 1948. A new species of Austrobaileya (Austrobaileyaceae) from Austra- lia. Jour. Arnold Arb. 29: 255, 256. 74 Ponp STREET BELMONT MASSACHUSETTS 02178 1967 | GILLETT, CORONANTHERA GRANDIS 245 CORONANTHERA GRANDIS (GESNERIACEAE), A NEW SPECIES FROM THE SOLOMON ISLANDS GEORGE W. GILLETT THE’ GENUS Coronanthera C. B. Clarke includes eleven species, ten of which are restricted to New Caledonia. A single species, C. australiana C. T. White, was described from material collected in Queensland. How- ever, S. L. Everist (personal communication) indicates that White in a later collection (White 10548, Sept., 1936) confirmed that the Queens- land species has fleshy, indehiscent fruits, This character would exclude it from Coronanthera and also from the tribe Coronanthereae, as delineated by Burtt (1962). Therefore, the presence of the genus in Australia may be questioned. The tribe Coronanthereae is assigned by Burtt (op. cit.) to the sub- family Gesnerioideae, and is comprised of Coronanthera, Rhabdothamnus A. Cunn. (1 sp., New Zealand), Negria F. Muell. (1 sp., Lord Howe Island), and Depanthus S. Moore (based on Coronanthera glabra C. B. Clarke, New Care. & These four genera, in addition to the genus Fieldia F. Muell. (1 sp., S. E. Queensland) of the tribe Mitrarieae, con- genera) New World subfamily Gesnerioideae. It is perhaps noteworthy that this representation occurs in a somewhat limited phytogeographic area that includes New Zealand, Lord Howe Island, New Caledonia, and Queensland. The additional species of Coronanthera described in this paper is a first record for the Coronanthereae in the Solomon Islands. It is significant that the area of distribution presently excludes New Guinea, but it probably would not be excessively speculative to predict the rep- resentation of the Coronanthereae on that island. The Coronanthereae are characterized by dehiscent fruit and cymose inflorescences. In the genus Coronanthera, the fruit dehisces along the sutures of the four apically coherent valves (Fic. 1, H). Dispersal, apparently, is by means of the shaker action of the capsule and the sub- sequent transport, by wind, of the very light seeds. It is this highly dis- tinctive fruit that enabled me to recognize specimens of the new species in a survey of South Pacific material of the genus Cyrtandra. Coronanthera grandis G. W. Gillett, sp. nov. Fic. 1. Frutex procerus vel arbor ad 15 m. alta. Folia, ramuli et inflorescentia pilis velutinis septatis non Rohe 50-100u diametro, Ramulis teretibus, 3-6 mm. diametro. Folia opposita, in eodem pari aequalia, petiolus gracilis, ahaa 2-5 cm. longus, lamina elliptica 246 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 mm Fic C oronanthera tenon a W. Gillett. All drawings from Kajewski 1749. A, leaf ae ax illary i _ , upper leaf surface showing EEELOFESY ap- ; E, posterior view of flower; F, ‘dow er with corolla opened out on the posterior sinus; G, c So prior ie dehiscence; H, dehiscent capsule showing longitudinal fissures - I, seeds. 1967 | GILLETT, CORONANTHERA GRANDIS 247 usque ovata, 12-20 cm. longa, 7-12 cm. lata, basis rotundata usque acuta, apex acutus usque acuminatus, margines serrati usque serrulati (3-8 dentibus in quoque cm.), superne pubescentia appressa sericea instructa, inferne dense villosa, nervi secondarii in quoque latere 6-8, curvi sursum, subter prominulis, venulae reticulatae supra obscurae, subter prominulae. Cicatrices foliorum conspicuae in ramulos, obtriangulares, infra rotun- atae, 2-3 mm. latae, cicatrice fasce magna centrali. Inflorescentia axil- laris cyma, floribus 3-8, pedunculus gracilis, 1 mm. diametro, 3-10 cm. longus, dense pilosus, ad apicem bracteatus, bracteae duae, aequales, lanceolatae vel ovatae, 10-15 mm. longae, 2-5 mm. latae, pilosae, caducae, lobos lanceolatos, aequales, acuminatos, extus intusque sericeus. Corolla urceolata, flavovirens, 8-10 mm. longa, 4-5 mm. lata, fissa inaequalis 1-3 mm. in lobos rotundatos, extus sericea intus superne glabra, inferne pilosa rare, limbus corollae bilabiatus, recurvus, labio superiore lobis duo- bus aequalibus 1 mm. longis, 1 mm. latis instructo, labio inferiore lobis tribus inaequalibus 1.5-2 mm. longis, 2-4 mm. latis. Stamina 4, filamenta aequalia, 6 mm. longa, 2 mm. ad tubum corollae affixa, 4 mm. liber, curvata superne versus centrum floris, superne glabra, inferne indumento sparso pilorum capitatorum instructa, antherae aequales, sagittatae, cohaerentes, facientes figura lunata adversus basem labi inferni corollae. Staminodium 1, 4 mm. longum, 3 mm. ad tubum corollae affixa, 1 mm. liber, pars libra portata ad altitudinem stigmatis, adversus sinum superum corollae. Gy- noecium 5 mm, longum, ovarium ovoideum, 3 mm. longum, 1.5 mm. latum, piloso pilorum non capitatorum ascendentium instructo, uniloculare, structus, stigma peltatum integrum, in centro depressum. Fructus capsula fusca, ovoidea, subulata, tomentosa, 8-10 mm. longa, 3-5 mm. lata, quadrivalvis, valvae ad apices connatae, inferne fissuris tenuibus longi- tudinalibus separatae, fissurae 5-6 mm. longae, 0.5—1 mm. latae, seminibus per fissuras laterales elapsis. Semina numerosa, fulvo-brunnea, ovoidea, fusiformes, 0.75 mm. longa, 0.4 mm. lata, testa reticulato-foveolata. Holotypus Kajewski 1749 (a), insula Bougainville. Solomon Islands (Australian Territory of Papua and New Guinea). Boucatn- VILLE IsLanp, Crown Prince Range: Kupei Gold Fields, 900 m. Rupei, near Kieta), April 22, 1930, Kajewski 1749 (a, holotype; BISH, isotype) ; Koniguru, 950 m. (near Buin), Kajewski 2122 (a, BISH). SANTA YSABEL IsLAND (British Solomon Islands Protectorate): Kakatio, 900 m., Brass 3252 (A, BISH). The local name on Bougainville Island is “Coru-cokor” (Kajewski 2122), while on Santa Ysabel Island it is “Momona” (Brass 3252). Appreciation is extended to Drs. Richard A. Howard, Director of the Arnold Arboretum of Harvard University, and Roland Force, Director of the Bishop Museum, who permitted me to examine specimens under 248 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 their care. The financial assistance of the National Science Foundation (Grant GB-3336) is gratefully acknowledged. LITERATURE CITED Burtt, B. L. Studies on the Gesneriaceae of the Old World XXIV: tentative keys to the tribes and genera. Notes Bot. Gard. Edinb. 24: 205-220. 1962. Lyon ARBORETUM UNIVERSITY oF Hawalr HonoLutu, Hawatr 96822 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 249 A SYNOPSIS OF THE ASIAN SPECIES OF DELPHINIUM, SENSU STRICTO Puitie A. Munz As IN THE PAPER on the Asian species of Consolida (Jour. Arnold Arb. 48: 159-202. 1967), the present article deals with plants from the Asian - mainland and does not include the islands to the west in the Aegean and Mediterranean seas. The same herbaria, for the most part, have made loans and furnished material, as indicated in that paper (pp. 159, 160). I again acknowledge my indebtedness for and express my appreciation of the many courtesies and much help received. Additional aid is recorded from time to time in the body of this paper under individual species concerned. It is a pleasure, too, to acknowledge obligations to National Science Foundation funds for Grant Number GB-2716, which very greatly helped, especially in work in European and British centers, and in partly meeting the cost of publication. A discussion in the earlier paper presented the characters of the segre- gate genus Consolida (pp. 160, 161). There follow now a description of Delphinium in the more restricted sense and a synoptical treatment of the large number of Asian species. Unfortunately, herbarium material has not been available for all of those described, especially many first proposed by W. T. Wang (Acta Bot. Sinica 10: 59-89, 137-165, 264-284. 1962). Fortunately, the Royal Botanic Garden at Edinburgh received photo- graphs of types of some of these species and I have found isotypes and other cited material in various American herbaria, especially in the large collection of unidentified Chinese specimens at the Arnold Arboretum. I am very much indebted to Dr. Shiu-Ying Hu of the Arboretum staff for translating critical passages and otherwise helping me with the Wang paper. In the same way, I gladly acknowledge the aid received from Mrs. Myra White, librarian of the Rancho Santa Ana Botanic Garden, whose ability to read Russian has helped with many passages in books dealing with the plants of the U. S. S. R. and with deciphering herbarium labels. Since the publication of the monograph on Delphinium, sensu lato, by E. Huth (Bot. Jahrb. 20: 322-499. pls. 6-8. 1895), which has to be the basis for all subsequent study of the group, many students have expressed their dissatisfaction with two principal features. (1) The subdivisions of the genus as set up by Dr. Huth, The genus Delphinium has some quite natural and well characterized subgroups or botanical sections. Others, both of Huth and Wang, seem meaningless to me and separate related species rather widely. For my purpose, therefore, I have keyed out the Asian species into ten Groups which I shall treat, in sequence, as Group I, Group II, etc. Where these seem to be of botanical 250 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 significance I shall so indicate. In most cases they may be quite artificial assemblages. For the most part, to facilitate finding the descriptions, the species under the larger groups are treated oe In some cases the same species is keyed out in more than one (2) Typification of some of Huth’s varietal ie Quite often no specimen is cited under a new variety and there is doubt as to what the taxon was meant to include and where it should come in the synonymy. A paper, such as mine, is obviously a tentative treatment. Lack of more herbarium material, inability to do field work, recognition of the fact that the species of Delphinium hybridize freely, all such matters preclude the possibility of a truly monographic treatment. It is hoped, however, that the more than two years spent in this study and the examination of perhaps 6000 herbarium sheets make possible a more precise treatment for identi- fication of Asian collections than we have yet had. DELPHINIUM L. Delphinium, as a genus, L. Sp. Pl. 1: 530. 1753; Gen. Pl. ed. 5. 236. 1754. E. Huth, Bot. Jahrb. 20: 322-499. pls. 6-8. 1895. As a sub- genus, Hudelbhintan (Huth) Dalla Torre & Harms, Genera Sipho- nogamarum 165. 1901. Mostly perennial, more rarely biennial or annual, herbs; if perennial, with tuberiform compact roots or a woody deep-seated rootcrown or rhi- zome-like base with fibrous roots. Stems erect or ascending or much re- duced. Leaves basal or cauline or both, persistent or withering early (especially the basal ones), palmatifid or lobed, often further divided or dissected. Flowers mostly in terminal racemes or panicles, sometimes only 1 or 2 on a stem. Pedicels subtended by more or less foliose bracts and bearing bractlets, the latter usually reduced and often two in number. Flowers zygomorphic. Calyx colored, of 5 sepals, the upper or posterior one spurred, the 2 lateral alike and the 2 lower alike. Petals 4, in 2 unlike pairs, the upper pair produced into nectariferous spurs that extend into the sepal spur and the upper part or lamina more or less cleaver-shaped; the lower pair with a slender claw and an expanded lamina which is often vertical to the claw. Stamens usually included, in 8 spiralled series, with filaments somewhat broadened and anthers short. Carpels 3—5, usually free, becoming follicles. Seeds mostly obpyramidal, winged or wingless at the angles, the surface variously beset with overlapping scales in trans- verse rows or ridged. Type species, Delphinium peregrinum L. Key To Groups oF ASIAN SPECIES OF DELPHINIUM A. Plants annual or biennial; petals glabrous and not ciliate. B. Limb of upper petals with a lateral wing on one side and not clawed. C. Leaf-blades palmately divided; lower petals not deeply divided. ... - GROUP L. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 251 B. Limb of upper petals not winged on one side; upper petals with short claw. LDU aay ta ts sin asd Beacdciehd, Bemis ae as ky ee ete ea ae eee GROUP II. A. Plants perennial; lower petals bearded and/or ciliate. D. Roots tuberous, clustered, short, vertical; flowers small (the sepals to ca. 1 cm. long), usually rather crowded in short-pedicelled (mostly to 1 cm.) often dense racemes; at least the upper leaves dissected into linear or oblong laciniae. Plants of rather arid regions, Mediterranean Region to Turkestan, Afghanistan, West Pakistan. ................. GROUP IV. D. Roots not tuberous, but mostly slender or cylindrical, rhizome-like; flowers mostly larger and often on longer pedicels. E. Plants mostly very hairy; flowers mostly large, the sepals 2 cm. or more long, veiny, dry and papery in age, more or less persistent; plants low; spur thick, 4-10 mm. in diameter at its base, conical to decurved. Largely: central Asia... 35 in es ee le Re GROUP V. . Plants usually less hairy; sepals not so veiny and papery, caducous; spur more slender. F. Leaves dissected into linear, more or less oblong ultimate segments 1-3 mm. wide. G. Stems with some well developed leaves. ........ GROUP VI. G. Stems without leaves or with 1 or 2 reduced ones. GROUP VII. F. Leaves with broader, lanceolate to ovate ultimate divisions that are mostly over 5 mm. wide. H. Spur not or scarcely uncinate. ie) J. Petals not darker than the sepals. ......... GROUP VIII. J. Petals darker than sepals, almost black. .... GROUP IX. H. Spur uncinate, often forming a half circle. ...... GROUP X. It is my feeling that the present knowledge of the species of Delphinium growing in Asia is entirely inadequate in most cases to make any attempt whatsoever to recognize subspecies or varieties, at least by a person like myself whose work on the group has had to be entirely in the herbarium and not at all in the field. In the first place, I am confident that many of the taxa which I am recognizing as species will not retain that status. In the case of those, and of many others, no specimens were actually seen by me, although altogether I have probably had the opportunity of study- ing more than 6000 sheets. In some cases letters asking for loans or bits of material have not been answered up to now. In others, so few specimens were available that it was impossible to get the feel of the species, its variability or the geographical or zonal range for these variations. I am therefore, almost without exception, attempting only to differentiate what seem to me to be possible species, keying them and describing them as well as the scanty material allows, and not attempting to treat their sub- divisions. GROUP I Mostly annual or biennial. Petals glabrous, the upper winged on one side, not clawed. Limb of lower petal not deflexed. Seeds numerous, rounded, with transverse rows of closely overlapping scales. 252 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 This group happens to coincide with the section greene as recog- nized by Pawlowski, Fragm. Flor. et Geobot. 9(4) : . 1963, and with the section Delphinellum DC. Syst. 1: 346. 1817. ned type species is D. peregrinum L., the type of the genus Delphinium. Key TO SPECIES 1. Limb of lower petals longer than wide, cuneate at base, gradually narrowed into a claw 2. The limb of the lower petals obovate, ca. as long as ae claw; limb not or scarcely exserted. Eastern Mediterranean. ........ . peregrinum. 2. The limb of the lower petals elliptic. ovate, shorter =a the claw and more or less exserted. 3. Claw of lower petals ca. 6 mm. long; plant glabrous to ee Turkey to Jordan and Pita atonal nea ieee. 2. D. virgatum. 3. Claw of lower petals ca. 4 mm. .. long: plant hirtellous or spreading- pubescent, Sinai Peninadas 3.2 os)... . sedeeis exears nanum. 1. Limb of the lower petals more or less rounded os quadrate, abruptly nar- rowed into a longer claw, the limb more or less exserted. 4. Mature follicles sublinear, closely parallel, 10-12 mm. long; spur up to one and one-half times as long as the sepals. Turkey. .. 4. D. cinereum. 4. Mature follicles broadly oblong, 5-8 mm. long; spur ca. twice as long as sepals 5. Racemes lax, linear, several-flowered; median cauline leaves thick in texture, with cuneate lobulate lobes. Turke Me ca 5. D. venulosum. 5. Racemes dense, oblong, many-flowered; median cauline leaves thin, with lance-linear altineate divisions. Turke Oye eG 6. D. davisit. 1. Delphinium peregrinum L. Sp. Pl. 1: 531. 1753. Fig; 1, A. Delphinium forskolii Reichb. Ilustr. Sp. Acon. 5: 68, ¢. 5. 1823-27. Type material seen (BM). Delphinium junceum DC. FI. France 5: 641. 1815. Delphinium peregrinum subsp. junceum Batt. FI. Alg. 16. 1888. Delphinium peregrinum var. junceum Huth, Bot. Jahrb. 20: 475. Delphinium peregrinum var. subvelutinum DC. Syst. 1: 348. 1817. Pubescent form based on Italian materia Delphinium bovei Decaisne, Ann. Sci. Nat. II. 4: - 1835. Type Bové 130 from between Suez and Gaza. This number seen Delphinium peregrinum var. erio ocarpum Boiss. FI. Sil 1: 87. 1867. Based on plants from Caucasus with pubescent fruits. Delphinium eriocarpum (Boiss.) Halacsy, Consp. Fl. Graeca 1: 32. 1901. Delphinium junceum DC. var. subvelutinum Ledeb. Fl. Ross. 1: 59. 1841. oo beregrinum var. eriocalyx Post, FI. Syria, Palestine Sinai, 46. Delphinium subvelutinum Heldr. ex Boiss. Fl. Orient. Suppl. 19. 1888. Annual, more or less gray-pubescent on stems with short stiff, often retrorse hairs, plant 2—7 dm. tall, usually virgately branched from below, the branches stiff, straight, or sometimes spreading, ridged; lower leaves 3—5-parted, the parts cut into 3 or more lance-linear lobes; blades 1-3 cm. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 253 long, often wider, longer than the petioles, glabrous, coriaceous; basal and middle leaves deciduous at anthesis; upper entire, linear, much reduced on upper stems, sessile; racemes terminal, many flowered; bracts subulate, —6 mm. long; pedicels stiff-pubescent, to almost 1 cm. long; bracteoles 2, to ca. 3 mm. long; flowers violet; upper sepal 7-8 mm. long, pubescent, the spur straight, pubescent, ascending, 15-20 mm. long and 3.5 mm. wide at its base; lateral sepals lance-oblong, pubescent, 5—7 mm. long; lower pair more pointed and with pubescence more median; upper petals appear- ing bilobed because of the broad ovate to elliptical lateral wing, the lamina ca. 7 mm. long, yellowish at base, violet toward apex; lower petals 8-10 mm. long, the blade longer than wide and gradually narrowed into the slender claw: stamens 5-8 mm. long, with somewhat lavender filaments; follicles 3, pubescent to glabrous, 7-8 mm. long (including the style) ; seeds rounded, ca. 1 mm. long, the scales united into transverse rows. Tyrer: Specimens in Clifford Herb. (pm), probably originally from Italy, Sicily or Malta. DistrrBuTion. Largely at elevations below 2000 m., fallow fields, cal- careous places, etc., from type region to Egypt, Turkey, Jordan, and Iran. ILLUSTRATIONS. SIBTHORP & SMITH, FI. Graeca 6: ¢. 506. 1826; Davis, Fl. Turkey 1: 129. 1965. REPRESENTATIVE SPECIMENS of the species complex are: Turkey: Kotschy 1; Balansa 728; Sintenis 1207; Balls 1207; Davis 13567. Lebanon: Gaillardot 52 and 367; Davis 9913. Israel: F. S. Meyers 99; Davis 4801; Boissier, April, 1846. Iraq: Field & Lazar 730 and 850. Iran: Layard, ex Herb. J. Ball. Egypt: Bové 130. Plants from Asia tend to have pubescent follicles and constitute the var. eriocarpum Boiss., if this variety is tenable. Those with the inflorescence condensed have been called var. densum Post, Fl. Syria, Palestine Sinai 46. 1896 2. Delphinium virgatum Poir. in Lam. Encycl. 10: 458. 1811. Fic. 1, B Delphinium peregrinum var. virgatum (Poir.) Huth, Bot. Jahrb. 20: 475 . 1895. Delphinium peregrinum var. laxum Post, Fl. Syria, Palestine Sinai 46. 1896. Much like Delphinium peregrinum, but glaucous, glabrous to strigulose; lower leaves cuneate-obovate, short-petioled, the blades with 3-5 coarse mucronate teeth at apex; midcauline leaves subentire, subsessile, sharply acute, passing upward into lanceolate bractlike leaves ca. 1 cm. long; flowers deep blue, strigulose, 20-22 mm. long; sepals lanceolate, 8-10 mm. long; spur 14-18 mm. long, spreading-curved; upper petal with glabrous winged blade ca. 10 mm. long, the spur 14 mm.; lower petals with ovate, slightly exserted limb ca. 3 mm. long, broadly cuneate at base, gradually contracted into a claw 6 mm. long; carpels glabrous. Type: Syria, Labillardiére (1); not seen. ILLUSTRATION, DELESSERT, Icon. Pl. 1: ¢. 55. 1820. 254 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 UR Delphinium, Group I, Petals, the upper spurred and with wing of lamina on n si away from spur, the dotted line pemmpiirsindl posi ition of main vein toward lamina tip. Lower petal clawed. A. D. peregrinum; spur 13 mm. long, n fr : a 5 12 mm. long, lower petal 7.5 mm.; from Boissier in 1842 (cH). E. D. ven- ulosum; spur 18 mm. long, lower petal 10 mm.; from Bornmiiller 3093 (BH). F. D, davisii; spur 15 mm. long; lower petal 8. 5 mm.; from Sintenis 4979 (x). DistRipuTion. Limestone hills to ca. 1000 m., southern Anatolia to Jordan and Lebanon. REPRESENTATIVE COLLECTIONS. Turkey: Kara Da, Little; Davis 35701. Leb- anon: Kotschy 227. Jordan: Marsh in 1851; F. Meyers 99 3. Delphinium nanum DC. Syst. 1: 349. 1817. Fic. 1, C. Delphinium oo ssp. manum (DC.) Graebner in Asch. & Graebn. Syn. Ga? ily (1 ? pete eaitile var. boissieri Huth, Bot. Jahrb. 20: 475. 1895. Not typified. ? Delphinium peregrinum var. subsaccatum Huth, loc. cit. Not typified. Delphinium grandiflorum Forsk. Fl. Aegypt. Arab. 212. 1775, not L., 1753. Annual, widely and stiffly branched from base to subsimple, the branches simple or nearly so, more or less hirtellous or spreading-pubescent throughout or especially on the sepals, the hairs sometimes somewhat interwoven, plants 3-15 cm. high; lower leaves obovate in outline, broadly 3-lobed, long-cuneate at base, 4-6 mm. long, the lobes broad, undivided; main cauline leaves largely simple, narrowly oblanceolate, 1—2.5 cm. long, narrowed at base into a broad petiole, gradually reduced up the stems and Sessile, the upper lanceolate, sessile, bractlike, 2-6 mm. long; racemes short, loose; flowers few; pedicels 1—5 mm. long: bracteoles 2, lance-linear, 3~4 mm. long: flowers bluish purple to brownish violet with ‘yellow base; 1967] MUNZ, ASIAN SPECIES OF DELPHINIUM 255 lower sepals oblong-lanceolate, 6-7 mm. long, grayish tomentulose-pu- bescent; lateral pair subovate, 6-7 mm. long, with median strip of pubes- cence; dorsal sepal 6—7 mm. long, densely pubescent as is the horizontal or ascendant spur which is 3-4 mm. wide at base and 15-20 mm. long; petals pale, the upper pair glabrous, lamina winged on the side, 7-8 mm. long, 8-10 mm. wide; lower petals to ca. 1 cm. long, the blade roundish- oblong, ca. 4 mm. wide, narrowed to a linear claw 4 mm. long; stamens 5-8 mm. long, filaments apparently bluish; carpels 3, sparsely pubescent, to ca. 8 mm. long and 2 mm. wide, the beak an additional 2-2.5 mm.; seeds 1 mm. long, almost round, dark brown with very closely appressed overlapping scales. Type: Near Alexandria, Egypt, Delile. An 1803 collection seen (BM). DisTRIBUTION. From between Suez and Gaza, Egypt, Bové; otherwise in African Egypt. ILLUSTRATION. TACKHOLM, Students’ Flora Egypt Pl. 57. 1956. 4. Delphinium cinereum Boiss. Diagn. I. 1: 67. 1842. Fic. 1, D. Erect annual, divaricately branched from near the base, 1.5—2 dm. tall, densely cinereous-strigose with retrorse hairs, the branches several, rather coarse, simple or few branched, stiff, leafy to the racemes; lower leaves tripartite, 1—2.5 cm. long, short-petioled, the parts cuneate, 3—5-fid, more or less strigose, 2-5 mm. broad; upper subsessile, 3-lobed; uppermost simple, passing into bracts 3-20 mm. long; flowers 5-12, rather closely placed; pedicels 2-5 mm. long, pubescent; the bracteoles 2, linear, 5—7 mm. long; flowers pale blue, densely pubescent; upper sepal 7-8 mm, long, the spur straight, horizontal, pubescent, ca. 10 mm. long, 3 mm. wide at base; lateral sepals 6 mm. long, oblong, pubescent along midrib, the margins hyaline; lower sepals more pointed, pubescent over whole outer surface; upper petals about as long as sepals, 4-5 mm. wide, broadly emarginate- lower petals ca. 7 mm. long, the limb rounded, abruptly narrowed into a slender claw; stamens almost equaling sepals, dilated and glabrous at base; carpels 3, cylindrical, closely parallel, hirsute- -pubescent, 10-12 mm. long in maturity; seeds globose, umbilicate, the scales in rows. YPE: Turkey: “in collibus Lydia circa Laodiceam,” Boissier, June, 6. 1867, Boissier said “Fl. July.” The specimens seen (BM, GH, K, P), all give July, 1842, and I take them to be isotypes, the holotype being at Gene CoLLECTIONS SEEN. Turkey: Caria, Pinard in 1843; Cilicia, Persnin 170. A rare and rather local endemic. 5. Delphinium venulosum Boiss. Fl. Orient. 1: 86. 1867. Fic. 1, E. Delphinium halteratum Sibth. & Smith var. venulosum (Boiss.) Finet & Gagnep. Bull. Soc. Bot. France 51: 470. 256 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Robust annual, 4-10 dm. tall, strigulose below, subglabrous above, with open ascending branches; stem leaves tripartite into broad segments, or the upper entire, elliptic-oblong, 3-20 mm. broad, subglabrous, 1-3 cm. long; those below the inflorescence deciduous by flowering time; racemes rather open, mostly several flowered; bracts linear or wider, 3-10 mm long; pedicels strigulose, commonly 4-5 mm. long, bibracteolate; flowers violet-blue to deep blue-purple, the upper sepal ca. 7 mm. long, very finely strigulose, the spur straightish, subhorizontal, mostly ca. 1.5 cm. long, scarcely 3 mm. wide at base; lateral sepals narrowly oblong-ovate, a. s mm. long, with strigulose median band; lower sepals lance-ovate, rather generally strigulose; upper petals apically blue, ca. 7 mm. wide; lower petals well exserted, ca. 9 mm. long, the rounded lamina abruptly con- tracted to a longer slender claw; stamens 5—6 mm. long; follicles 5—6 mm. long, 2.5 mm. thick, with a few scattered hairs; seeds rounded, the apex narrowed, deeply umbilicate. Type: Turkey: “in cultis Cappadociae circa Caesaream,” in 1856, Balansa 858; isotype seen (Kk). DISTRIBUTION. Inner Anatolia in fallow places below 1200 m. IttustRATION. Davis, Fl. Turkey 1: 129. fig. 2 (12). 1965. REPRESENTATIVE SPECIMENS. Turkey: Bornmiiller 2674, 3093, and 1529; Sintenis 4627; Davis 32782, 32778, 32813; Uvarov 98; Khan, Prance & Rat- cliffe 759. An endemic Turkish species. 6. Delphinium davisii Munz, nomen novum. Pig. 1B he cocci sintenisii Freyn, Osterr. Bot, Zeitschr. 43: 374. 1893, not Uechtr. sopra oc Davis, Notes Bot. Gard. Edinb. 26: 172. 1965, not Grisebach, 183 Plant 1—4 dm. tall, divaricately branched from base or middle, glabrous except for the strigulose pedicels and calyx; leaves round in outline, palmately parted, the segments cuneate, parted into mucronate lanceolate lobules 1-2.5 cm. long, 1-3 mm. wide; racemes densely few to many flowered, 5-15 cm. long; bracts simple, entire, lance-linear, 3-12 mm. long, mucronate, to ca, 1.5 mm. wide; pedicels 2-3- bracteolate, strict, 1-2 cm. long; bracteoles lanceolate, near middle of pedicel, 2-3 mm. long; sepals blue-violet, minutely strigulose, the upper 9-11 mm. rane more or less oblong, rounded at apex; spur ascendin ed, ca. 15 mm. long, 3 g, curve mm. wide at base, subcylindric, rather blunt; awed ier oblong, 7 mm. long, 3 mm. wide, rounded at apex; lower sepals lance-ovate, 8 mm. long, obtuse; upper petals with blade 9.5 mm. long, bent at right angle to spur, the lateral wing broad, the apex entire, round-oblong, the spur 15 mm. long; laminae of lateral petals cordate- orbicular, 4 mm. in diameter, the glabrous claw 5 mm. long; stamens 5 mm. long, dilated at base, glabrous; anthers yellow, almost round, 1 mm. long; follicles erect, 3, short, oblong, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 257 parallel, torulose, glabrous, veined, 7-9 mm. long, the beak an additional 2 mm.; seeds black, almost 2 mm. long, depressed spherical, transversely lamellate-rugose, the lamellae imbricate. TYPE COLLECTION: Turkey: Paphlagonia near Kastemuni [Kastamboli | in valley Kara Dere near Kady Oghbi Chan, Aug. 15, 1892, Sintenis 4979; isotypes seen (BM, K). DisTRIBUTION. Disturbed places, 850-950 m., edge of scrub, Paph- lagonia and Bithynia. REPRESENTATIVE COLLECTIONS. Bithynia: Grisebach. Paphlagonia: Davis 39033. Related to D. halteratum of the western Mediterranean, but tending to be more glabrous vegetatively and on the follicles, the plant of lower stat- ure, etc. It is a pleasure to name this Turkish species for Dr. Peter H. Davis in recognition of the contribution he has made to an understanding of the genus Delphinium in Turkey. UNCERTAIN SPECIES 7. Delphinium syncarpum Freyn ex Stapf, Denkschr. Akad. Wien Math. Naturw. 51: 26. 1886. Fic. 2, E. Simple erect, apparently annual, to ca. 2 dm. high, leafy, cinereous- strigulose and with short stiff spreading glandular hairs; lower petioles to ca. 1 cm. long; blades 8-12 mm. long, trisect, then again lobed into oblong, mucronate, obtusish parts to ca. 1 mm. wide; flowers few, in a close simple raceme; bracts commonly trifid, 5-8 mm. long; pedicels 3-6 mm. long; bracteoles linear, 2-3 mm. long, at about the middle of the pedicel; sepals white with upper midrib greenish, mostly elliptic, 8-9 mm. long, attenuate into a claw, the spur descending, ca. 8 mm. long; upper petals sepal-like, but without the green midrib, clawed, ca. 11 mm. long, apparently spurless, hence nectarless; lower petals with blade ca. 8 mm. long and wide, shortly 2-lobed, glabrous, the claw 4 mm. long; stamens 3.5-4.5 mm. long, the filaments broadly dilated, glabrous; anthers pale, 1 mm. long; follicles 3, hairy, united at the base. Type: “ad portum Moschdiser,” northern Iran, legit Th. Pichler (wv); seen. A very doubtful species. The type consists of a single one-stemmed plant, with a few broken flowers; the carpels are very undeveloped. The flowers seem abnormal, the upper petals apparently not spurred, sepal- like, but lacking the dorsal green midrib of sepals. I believe it is annual and has lower petals glabrous, hence must be in my Group I, although I first placed it in Group IV. I Mave ea a coltection from Luristan, Iran, namely Koelz 185 19 (micH), which may represent this species. It has no flowers; the follicles are united about half way. I am very grateful to Professor Lothar Geitler, Botanical Institute, 258 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 University of Vienna, for his kindness in loaning me the type of D. syncarpum, GROUP II Plants biennial. Leaves palmately 5—7-partite into entire or incised segments. Petals glabrous, the upper short clawed, not winged on the side. Seeds few, reticulate-foveolate, not squamate. This group, in which three species are recognized in Europe (Tutin et al., Fl. Europaea 1: 216. 1964) is the same as Section Staphisagria DC. Syst. 1: 362. 1817, and Tribus Staphisagria Huth, Bot. Jahrb. 20: 481. 1895. In Asia there is one species. 8. Delphinium staphisagria L. Sp. Pl. 1: 531. 1753. Fic. 2, A. Staphisagria macrosperma Spach, Hist. Nat. Vég. Phan. 7: 348. 1839. Biennial, usually one stemmed, simple or few branched above, 3-12 dm. tall, leafy, soft-hairy with fine rather long spreading hairs and densely cov- ered also with short retrorse hairs; leaves large, the blades rounded or broader, to 2 dm. in diameter, more or less finely strigose and with longer hairs, the usually 5—7 segments entire or 2- or 3-lobed, these ultimate divi- sions broadly lanceolate, sharply acute, 3-veined, often entire or with few large teeth; petioles mostly longer than blades, dilated at base, with both kinds of pubescence; racemes terminal, dense or lax, usually many flowered on the main stem, few flowered on branches; lower bracts foliose, 3-parted or entire, sessile or petioled, from half as long to as long as the pedicels which are 0.5-5 cm. long, densely hairy and strigose, bibracteolate near the base; flowers mostly deep blue, 2-3 cm. in diameter; calyx fine-hairy, the upper sepal oblong-ovate, 1-2 cm. long, the spur saccate, 2-3 mm. long and wide; lateral sepals ovate, ca. 15 mm. long, rounded at apex; lower sepals similar but slightly shorter; upper petals white or bluish, pubescent, obtuse, simple or some bilobed at summit, 10-12 mm. long, short-spurred at base; lower petals glabrous, subspatulate or obovate- cuneiform, somewhat asymmetrical, 14-16 mm. long; stamens 8-10 mm. long, ciliate above, glabrous and dilated below; carpels 3, long-pubescent, inflated, 12-20 mm. long at maturity; seeds few, 5-6 mm. long, blackish, obpyramidal, 3-angled, reticulate-foveolate. TYPE LOCALITY: Described from Istria, Dalmatia, Calabria, Apulia, S. France (LINN. 694/12). DisTRIBUTION. Found mostly in woods and waste rocky places below 300 m., Mediterranean region from southern Europe and Morocco to Turkey, Syria, etc. ILLusTRATIoNS. SiptHorp & Situ, Fl. Graeca 6: f. 508. 1826; CosTE, ay ee 1: 51. 1901; Maire, Fl. de l'Afrique du Nord 11: 84. fig. 39.1 thes SPECIMENS. Turkey: Persnin 1; Sintenis 444; Little 52. Syria: Egger 68; Haradjian 2138. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 259 GROUP III Leafy annuals with leaves subpinnate into more or less ovate, cuneate segments. Petals glabrous, the upper sessile, with a wing opposite the spur so as to appear bilobulate. Lower petals more or less deeply divided. Seeds with a spiral band of 5-6 turns of confluent scales. This group is the section Anthriscifolium W.T. Wang, Acta Bot. Sinica 10: 277. 1962, and is quite distinct in its characters. It ranges widely over China. As I understand it, it consists of a single variable species, varying in length and coarseness of spur, flower size, size of leaf lobules, lobing of both pairs of petals, slenderness, and shape of follicles. Occasional plants lack spur and petals (H. B. Morse 533, Lungchow, Kwangsi (NY). Variations do not seem to be geographic and I am treating them simply as varieties. 9. Delphinium anthriscifolium Hance, Jour. Bot. 6: 207. 1868. Annual, the stems rather slender, more or less flexuous, grooved when dry, simple to openly branched, mostly 1~7 dm. tall, subglabrous below to short- or long-retrorse-strigulose, more densely so above, sometimes with some spreading hairs; stems leafy almost to the inflorescence, the basal leaves gone by anthesis, the cauline little reduced upward, petioled, ternate, the 3 main divisions short petiolulate, becoming pinnate into several more or less ovate cuneate segments 1-3 cm. long and in turn pin- natifid into few to several lanceolate lobules usually 2-3 mm. wide, the terminal acuminate; leaf blades commonly 3-10 cm. long, somewhat strigulose; inflorescence short- to long-pedunculate, rather few flowered, racemose; bracts from tripartite to simple and linear, mostly 3-10 mm. long; pedicels about as long as the flower; bracteoles linear, usually near the middle of the pedicel; flower violet to blue; upper sepal 5-11 mm. long, 1.5-3 mm. wide, usually acute, pubescent, usually spurred, the spur 5~22 mm. long, rather straight, very slender to somewhat coarse; lateral sepals asymmetrical, more or less unguiculate, 5-10 mm. long, 3-6 mm. wide; quite membranous near the edges; lower sepals more elliptic, 2.54.5 mm. wide; petals often purplish or veined with purple, glabrous, the upper with the lamina 7-11 mm. long, winged on one side and appearing bilobed, with a spur mostly 5-20 mm. long; lower petals somewhat asymmetrical, usually deeply 2-lobed, sometimes only emargin- ate, unguiculate, the lamina 4-5 mm. long, the claw 3.5-5 mm. long; stamens unequal, 3.5-7 mm. long, the filaments glabrous, not strongly dilated below; anthers usually dark, 0.6—-0.8 mm. long; follicles 3, rarely 5, inflated, divergent, usually rather abruptly narrowed at apex, 8-14 mm. long, 3-4.5 mm. wide, the style slender, 2-3 mm. long; seeds brownish, almost 2 mm. long, with a spiral band of 5-6 turns of confluent scales. Spur often longer than sepals, 10-22 mm. long; sepals 9-11 mm. long. ...... a ae Var, anthriscifolium. Spur often equal to sepals, shorter than above. 260 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 The spur 7-10 mm. long; sepals mostly 6-8 mm. long. ...... Var. callerii. The spur 5 mm. long; sepals ca. 5 mm. long. .............. Jar. savatieri. Delphinium anthriscifolium Hance var. anthriscifolium. Fic. 2, B. Delphinium anthriscifolium Hance, Jour. Bot. 6: 207. 1868. ? Delphinium anthriscifolium var. majus Pamp. and var. ramosum Pamp. Nouv. Giorn.-Bot. Ital. (n.s.) 22: 288. 1915 re naan cavaleriense Lévl. & Vaniot, Bull. Acad. Geogr. Bot. 11: 49. 2 and 15: 711. 1905. Type, Cavalerie 2344 from Pin-fa do Lofou (E, K) seen. Delphinium cerefolium Lévl. & Vaniot, Bull. Acad. Geogr. Bot. 11: 49. 1902. Type, Chaffanjon, June 2, 1898, Mont du College, Kouy-Tcheou (£), seen. Cf. Lauener & Green, Notes Bot. Gard. Edinb. 23: 584. 1961. Flowers usually several; sepals 9-11 mm. long; spur 10-22 mm. long. LECTOTYPE: “prope rupem calcaream Kai-kun-shek,” Kwantung prov., China, June, 1867, Sampson, Hance no. 10125 (holotype K; isotypes BM, GH); chosen since original description mentions this collection as having “afforded excellent material for the above diagnosis.’ DIsTRIBUTION: Widely distributed in China. REPRESENTATIVE COLLECTIONS. Hupeh: Henry 5862, 1348, 1966; Wilson 93; H-Ch’ang Chow 89; Watters in 1879. Szechwan: Fang 758; Farges 1146, 1441, 1440; Faber 725; Yong 58311. Anhwei: Fan & Li 08. Yunnan: Tsai 50809; Covdlenc 2957. dias Handel-Mazzetti 11802. Fukien: Metcalf & Chang 797; Tang Siu Ging 13795. Kwangtung: Chun 6276; Lamont 1065. Delphinium anthriscifolium Hance var. callerii (Franchet) Finet & Gagnep. Bull. Soc. Bot. France 51: 471. 1904. , Delphinium callerii foe Bull. Soc. Linn. Paris 1: 329. 1882; Bull. Soc. Philom. Paris 5: Delphinium exiguum E. patel ex Diels, Bot. Jahrb. 29: 327. 1900. Type from Szechwan: Nan ch’uan, von Rosthorn 2022; not seen a eg robertianum Lévl. & Vaniot, Bull. Acad. Geogr. Bot. 11: 2. Type, near Kauy Yang, Kwangtung, Sept. 12, 1897, Bodinier ? (s); seen. Delphinium minutum Lévl. & Vaniot, Bull. Herb. Boiss. II. 6: 505. 1906. Based on Esquirol 24, Kouy-tchéou (printed as 23 in original descrip) oe , uc); seen. Cf. Lauener & Green, Notes Bot. Gard. Edinb. 23: 584 Baie kweichowense W. T. Wang, Acta Bot. Sinica 10: 283. 1962. Based on Y. Tsiang 8571 from Tenfan, Kweichow. Seen ee us). Some- what intermediate between vars. anthriscifolium and callerii Flowers one to few; sepals mostly 6-8 mm. long; spur 7-10 mm. long. Type: Near Macao, China, Callery 6, in 1841 (Pp), photo. (E), seen; and Callery 51, in 1844, photo. (£), seen. RANGE. Widely distributed i in China. 1967 } MUNZ, ASIAN SPECIES OF DELPHINIUM 261 ‘9 ‘at ae ‘he ict peteiirelt o staphisagria; — petal 10 mm. long, - = wn fr 3 (Mo Group III. Upper petal with a wing poe the base of the spur. B. D. akties var. anthriscifolium; upper lamina 11 mm. long, lower lamina 5 mm. long; drawn from Henry 5862 (cH). C. D. a. var. callerii; lamina of upper petal 9 mm. lo pe of lower petal 4 mm. long; drawn from Cavaler rie 2822 (uc). D. D. a. var. savatieri; i of ae petal 7.5 mm. long, of lower petal 4.5 mm. long; drawn from Beach 20 (v ROUP D. syncarpum; lamina of upper petal 7 mm. long, cla Bt of lower iE 3 mm. long and wide; drawn from Pichler, Monchtuee Uw) REPRESENTATIVE SPECIMENS. Szechwan: Henry 7263; Wang 20450; Sun 1765; Fang 12258; Chien 5877. Kiangsu: Feng 112; Schindler 275 C; Stewart 2071; Merrill 11429; Keng 1311; Pei 2646. Fukien: Dunn 2321; Chung 2602. Kwangtung: Cavalerie 2169, 2822. Chekiang: Barchet 11; Ching 1363; Bullock 7. Anhwei: Ching 8442. Aig anthriscifolium Hance var. savatieri ( oad me comb. noy —— Delphinium savatieri shbyse: Bull. Soc. Linn. Paris 1: 330. 1882; Bull. Soc. Philom. Paris 5: 162. 18 Sepals ca. 5 mm. long; spur ca. 5 mm. long, subconic. Type: China: Chekiang: “in siccis ad pedem montium Shao-Shin, prope Ning-po,” May, 1863, Lud. Savatier (p?); not seen. REPRESENTATIVE SPECIMENS. Less common than the other two forms. Kiang- su: Teng 153; Chang 386. Chekiang: Kay Beach; Keng 580. Hupeh: Henry 200, 760. Anhwei: Bullock 10125 (sm), cited by Hance as D. anthriscifolium, A spurless form seen from Hupeh: Morse 553 (K, NY). 262 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 GROUP IV Plants perennial, with a cluster of short, tuberous, mostly vertical roots. Leaves palmatifid, at least the upper dissected into linear or oblong laciniae. Inflorescence usually a rather dense raceme with pedicels short (to 1 cm. long) and flowers small (sepals to 1 cm. long). At least the lower petals bearded. Plants of rather arid regions, largely from the Mediterranean to Turkestan and Afghanistan. This group is in large part the Section DiEpRoPETALA, Tribus Macro- centra of Huth, Bot. Jahrb. 20: 341. 1895; or, in the text of his monograph on later pages of the same reference, it would fall into his Tribus Gzb- berula, p. 423, Tribus Lasiocarpa, p. 428, and Tribus Leiocarpa, p. 439 All of these categories proposed by him, however, include other species not in my Group IV. Since the fifty species which I recognize in this group do not fall into readily separable minor groups, I am treating them in alphabetical order, hoping that in this way their study will be facilitated. KEY To SPECIES 1. Bracts broad, ovate-lanceolate. 2. Bracts longer than the pedice a ete terete below; sepals piles to heii or so blue; spur 8-10 m. long. Northern Anatolia to Bulgaria. ...... D. albiflorum. a Stem angled and sulcate below; iene bluish: su 10 mm. long. ger Himieatee to Fe PO Eee 20. D. dasystachyum. 2. Brects equaling pedicels; sepals bluish white to pale rose; spur 15 mm LOne Serre ta SOCOM: TARE sa pe ase 8 by & D. ithaburense. 1. Bracts narrower, mostly lin 4. Spur definitely longer roe its sepal Stop 5. Sepals bluish to purple, occasionally 6. Stems strigulose with — sacaian hairs. 7. Flowers purplish blue 8. Sepals strigulose; ‘spur 10-15 mm. long; blade of lower petal longer than wide. Iraqui Kurdistan. ............-- Be ihae eg cee, Am As Big AD, packarnei 8. Sepals with scattered spreading hairs; spur 20 mm. | blade of lower petal about as wide as long. Afg Rainn: pe Oe oias arenas NETS eke EE 8 52. D. tetanoplectrum. 7. Flowers pale blue to smoky gray. 9. Stem subglabrous below, strigulose above; lower petioles only subvaginate at base: spur usually curved; carpels hirsute-pubescent. Turkish Rurdistan.. 35. oo: eS Te Mien EGG Boy ee 21. D. dolichostachyum. 9. Stem strigulose throughout; lower — strongly vagi- nate; spur straight; carpels strigulos 10. Bracteoles 3-6 mm. long, near the middle of the dicel. 11. Upper petals glabrous at ame spur 16-20 mm. long, the upper sepal ca. m. long. Tauria. 20 . D. pallasii. 1967] MUNZ, ASIAN SPECIES OF DELPHINIUM 263 11. Upper petals often with some hairs at tip; spur 2-15 mm. long, the upper ier 8-10 mm. long. m Hewae Rr ns eee . D. tuberosum. 10. oe 1.5-3 mm. long, near the eee — petal notched at apex; spur 10-15 mm. ong. Afghanistan, western Himalaya. ......... 12. Upper petal entire; spur 17 mm. ped “Turkestan. Pe AO eS ek Be ees . D. batalinii. 6. Stems glabrous or with spreading hairs (at least is e). 13. Sepals serena) glabrous, sometimes with minute puberu- lence near the tip. 14. pews leaves with broad primary lobes (1 cm. or wider), some usually still evident at anthesis; sepals violet blue, the spur 11-12 mm. long. 15. Spur of upper petals ca. 3.5 mm. wide at base, 12 mm. long; lobes of lower petals lanceolate. Northern Trane one eS 12. D. aquilegifohum. 15. Spur of upper petals 2 mm. wide, 9 mm. pike lobes of lower petals oblong-ovate. pai iby ish... RRA eG ee ay D. pence ren 14. Lower leaves with mostly narrower lobes and largely withered by anthesis. 16. Sepals pale ine ca. 8 mm. long; spur 15 mm. long; racts 5-8 mm. long. Turkey. ..... 22. D. fissum. 16. Sed a ly 9-10 mm. ee s 3-5 mm. ae te 11-13 mm. lon Hees Nh Sas 4. D. longipedunculatum. 17. Bracts 8-12 mm. ra spur mm. long. Tean Shan, middle Asia. .............-----: 11. D. albomarginatum. 13. Sepals with hairs on the outside. 18. Upper parts of o not glandular; era 8-10 mm. long Persian Riirtistan. .. 6 as ics p, quercetorim. 18. ve parts of een with some aie hairs . Spur steeply ascending; carpels glabrous; racemes elongate. Caucasus and beyond. ............-.. is os aeRO rig oe dy agatha ae a 47. D. schmalhausenii. 19. Spur more or less oe 20. Sepals pale 21. Carpels sk eee lower petals not ex- ceeding sepals. 22. Spur curved at apex, ieee seed as long as sepals. — ie aR risa 5 1a Dp. nine 22. Spur straight, ae: one and one-half times as long as sepals. Mascate, Iran 42, 2; penicillatum. . Carpels pubescent; lower petals exceeding sepals. Persian Kurdistan. ...........-- 44. D. quercetorum, ho — 264

un petals completely covered with long hairs. Middle bucharicum. Asia ? Upper petals am lower petals with few short hairs in distal part 46. Spur of upper petals ca. 3.5 mm. wide, mm. long; lobes of lower petals aie Northern Iran. ..... 12. D. aquilegifolium. . Spur of upper petals 2 mm. wide, 9 mm. long; lobes of lower aie oblong-ovate. Southern OO bes ae gpa 46. D. saniculifolium. de ia) JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 44. 7 sepal 10-16 mm. lon Flowers pale rose; ‘ain — yas EC). SCE MES OR ee en Maree 2 ret ame: 8. D. chodatii. . Flowers blue or violet. 48. Racemes dense, elongate, ms flowered; sepals deep violet. Kurdistan. ............ 6 D. a ee > ~I > oo . Racemes 49. Sepals oo blue; lower petals hairy. jes | eee Pere 57. D. vanense. 49. Sepals sordid-lilac; lower petals almost glabrous. Asia M eda, Se Pe MS re etn 39. Sepals with some pubesce 50. Bracts inflated at a cam, with long narrow terminal lobes. NN ce ke es ae ee ae, 20. D. dasystachyum. Bracts not as above. St. yar = some spreading hairs. Spur curved into a terminal hook; sepals bluish ie ane villous. Western Himalaya. ve eas 55. D. uncinatum. 50. Oo 52. Spur not hooked at end. 53. Sepals aes blue to dark blue-violet, 7-13 mm. 54. Plant. 26-60 cm. tall; ans Sua pubescent. Turkish Kuciist ia yee a pent ae I. 2. see ere 54. Plant 16-20 cm. tall; sepals viscid- pubescent. Kastamonu, Turkey. ...... 25. D. ilgazense. . Sepals paler, the upper sepal 12 mm. long; carpels glabrous. aia eee eer faperta vee on WwW Pedicels nied closely appressed o. 53. inflorescence glabrous; sepals 10-11 mm. aes spur ialiineniicdianiae 2 mm. long. cis 7 1 Baers elem enamel rer raat . D. karategini. 55. Axis of inflorescence more or ious arson’ 56. pag sepal 6-9 mm. lon . Racemes to ca. 10- dowered. rather open; r 5-7 mm. long; upper petal-blade with some hairs. Himalaya. .........-- Ey ese as ee 29. D. kohatense. . Racemes 20- or more-flowered; spur 8-10 mm. long; upper petal-blade glabrous or nearly so. : 58. Raceme dense; spur 8 mm. long. Tran 6. D. ursinum. 58. Raceme open; spur 10 mm. long. East anna _. 24. D. griseum. . Upper sepal 10-13 59. Flowers blackish " violet- purple; petals wn — on ~ un oO 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 267 purple. Georgia to the Volga River. ... 43. D. puniceum. 59. Sepals bluish; petals lilac. 60. Lamina of lower petals longer than wide. 61. Ultimate leaf segments 1-2 wide; lamina of lower petals oblong, with 2 lanceolate divergent lobes. Russian Ar- menia, Transcaucasia. ........ D. laxiusculum. fo.) js S ct 5 ro] pa Pas) — ee SAT n q oO =} —- wm ve w . wide; lamina of lower petals elliptic, with 2 broader parallel lobes. Southern Iran. aspire 19 ; 60. Lamina of lower petals roundish, lobed ca. 3 mm. Western Himalaya. 45. D. roylei. 10. Delphinium albiflorum DC. Syst. Nat. 1: 353. 1817. Fic. 3, A. Delphinium hybridum Steph. ex Willd. var. albiflorum (DC.) Koch, Linnaea 15: 247. 1841. Delphinium ochroleucum auct. non Stev. ex DC. Delphinium ponticum Hausskn. & Bornm. Osterr. Bot. Zeitschr. 40: 211. 1890. Based on Bornmiiller 972 from Amasia, Mt. Sana Dagh (sM, Kk, P). ie; . Seen, Delphinium albiflorum var. candolleanum Huth, Bot. Jahrb. 20: 439. 1895. Not typified. Delphinium albiflorum var. eginense Huth, Bot. Jahrb, 20: 439. 1895. Based on Sintenis 1890 from Egin, Armenia. Not seen. Stem terete, 5—6 dm. tall, 2-5 mm. thick, simple, glabrous or glandular- pubescent above, glabrous and somewhat angled below; leaves petioled, multipartite into acute long linear lobes 1-3 mm. wide, the petioles strongly dilated at their base; lower leaf blades 4-7 cm. broad, strigose; racemes crowded, elongate, many flowered, pubescent or glabrous; bracts mem- brane margined, white to pale blue, broadly lanceolate, acuminate, longer than the pedicels; bracteoles membranaceous, 3-4 mm. long and just below the flowers; median pedicels shorter than the sepals, mostly ca. 5 mm. long; flowers, including spur, 19-22 mm. long; sepals white, or nearly so, to yellowish or pale blue, glabrous or pubescent; upper sepal oblong-ovate, 7-8 mm. long, the spur straight, horizontal, obtuse, 8—10 mm. long, 2-2.5 mm. wide at base; lateral sepals oblong, 8-10 mm. long, 34 mm. wide, rounded at apex, with greenish midrib; lower sepals ovate, obtuse, ca. 8 by 5 mm.; upper petals glabrous, bluish, bilobed, the lamina somewhat oblique, ca. 9 mm. high and with 2 short rounded apical lobes; the spur ca. 10 mm. long; lower petals bifid, the lobes subacute, unequal, bearded, the lamina 6 mm. long, elliptic rounded in outline, the claw 5 mm. long; stamens 5—6.5 mm. long, the upper part of the filaments bluish or 268 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 whitish; anthers yellowish or blue, roundish, 0.6 mm. in diameter; follicles 3, pubescent, erect, ca. 8 mm. by 2.5 mm., the beak an additional 2-2.5 mm Type: Armenia, Tournefort (P). Disrianorion. Northern ee ey “T se at 1800 to 2100 m., growing on limeston Paine gee DELESSERT, Icon. Pl. 1: ¢. 58. 1820. REPRESENTATIVE COLLECTIONS. Paphlagonia: Sintenis 4883; Kars: Davis 30593; Tunceli: Davis 31305; Agri: Furse 3654; Erzerum: Furse 3785; Armenia: Calvert & Zohrab 1347. A species characterized by its broad bracts. 11. Delphinium albomarginatum Simonova, Key Pl. Environ. Tash- kent, ed. Popov, 124. 1924. Delphinium karataviense Paviov, Bull. Soc Mosc. sect. Biol. [ Mosk. Obschch. Isp. Pri. a II. 47: 80. 1938. poner as having D. albomar- ginatum as a synonym Stem 8—15 dm. tall, branched and glabrous above, with short white un- equal hairs below; lowest leaves long petioled, strongly widened at the base, still present at time of flowering, when the laminae fall away; laminae trifid, the lobes unequally palmate; upper leaves less divided; flowers in a dense long raceme; lower bracts longer than the flower, the upper narrow- linear and equal to it; pedicels not more than half as long as the calyx; bracteoles near the middle of the pedicel or lower; sepals dark violet-blue, 9-10 mm. long, glabrous; spur almost one and one-half times as long as the sepals, straight; petals white, the upper notched; lower ciliate, but glabrous on the surface. Type: Middle Asia, from Humsana in Umganski Mts. Type at Tash- kent, but I was unable to see it or any other material. Apparently this plant is endemic in the western part of the Tean Shan region. 12. Delphinium aquilegifolium (Boiss.) Bornm. Bull. Herb. Boiss. II. 4: 1083. 1904, reprinted as Fl. Elbursgeb. N.-Persiens 11. 1904. Fic. 3, B. peli acrtesang saniculifolium Boiss. var. aquilegifolium Boiss. Fl. Orient. 1: 91. Delpininm aquilegifolium var. hebecarpum Bornm. Bull. Herb. Boiss. I. 4: 1904, reprinted as Fl. Elbursgeb. Aaeign 13. 1904. Based on pein i 6062, May 29, 1902, N. Persia, mts. Schimran. n (BM, P, ? Delphinium aguileife Nee vars. eu alintin. oLibalacdston and violaceum Bornm. Stem more or less zigzag, 3—6 dm. tall, openly few branched, subglabrous except for small areas on one side of the pedicels or all the way around with dense spreading gland-tipped hairs inflated at the base; basal leaves 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 269 palmately 3—5-lobed into broad obovate parts 1-2 cm. wide or shallowly lobed and toothed at the apex, the blades 3-4 cm. wide, subreniform in outline, subglabrous, on petioles 3-10 cm. long, vaginate at base; cauline leaves few, more deeply dissected into parts with lance-linear lobes; in- florescence an open panicle of ca. 5-20 flowers in rather lax racemes; bracts lance-linear, 3-5 mm. long; pedicels elongating in fruit, ascending and becoming 1-3 cm. long, with 2-3 linear bracteoles ca. 2 mm. long and near the middle; flowers violet-blue; sepals glabrous or pubescent, the upper sepal ovate, 8 mm. by 3 mm., subacute, the spur 11-12 mm. long, 3 mm. wide at the base, somewhat attenuate at the apex; lateral sepals 7 mm. by 3 mm., oblong-elliptic, rounded at the tip; lower sepals 8 mm. by 4 mm., elliptic, obtuse; blade of upper petals oblique, ca. 8 mm. long, short-lobed at the apex, the spur ca. 12 mm. long; lower petals ca. 8 mm. long, the claw 4 mm., the broadly elliptic-oblong blade 4 mm. long, lobed to ca. 2.5 mm., ciliate in the cleft and more or less bearded on the surface; stamens 4-5 mm. long, subglabrous; anthers bluish, 0.7 mm. long; fol- licles 3, divergent, pubescent, to ca. 10 mm. by 4 mm., the styles an additional 2 mm. Lectotype: northern Iran, Mount Elbrus near Derbend, Kotschy 239 (holotype undoubtedly at G; isotypes seen at A, BM, K, P). DistriBuTION. At 1900-2500 m., northern Iran and western Pakistan. REPRESENTATIVE COLLECTIONS. Iran: Rechinger 3838; Bornmiiller 6060; Furse 2627: Trott 887; Furse & Synge 329 and 330; Koelz 18257, W. O. Doug- las 236. W. Pakistan: Chitral, Bowes Lyon 25; Swat, Stewart & Rahman 25229. The species is distinguished by the broad basal lobes and is near p. saniculifolium, but with broader, somewhat longer petal-spur and more sharply pointed lobes on the lower petals. D. saniculifolium is a species of southern Iran. 13. Delphinium batalinii Huth, Bot. Jahrb. 20: 432. pl. 8. 1895. Fie: 3, C. Stem erect, slender, subsimple, 2.5—7 dm. tall, short-strigulose through- out, scattered-leafy throughout; petioles dilated at base, 2-5 cm. long, ascending; leaves palmately parted into ultimate broad oblong lobes, the lower blades 2-3 cm. wide, the upper gradually widened from the petiole and with several lanceolate to sublinear ultimate lobes ca. 1 cm. long and 2-4 mm. wide; inflorescence racemose, branched at base or simple, elongate, with rather many, not crowded flowers; bracts all entire, the lower oblong-lanceolate, the upper linear, 5-8 mm. long; pedicels erect, adpressed to the axis, 10-15 mm. long, strigulose; bracteoles 2, linear, near the flower or more often 1/3-1/4 way below it, ca. 2 mm. long; flowers dilute blue; sepals strigulose, the upper sepal round-ovate, 9 mm. by 7 mm., apiculate, the spur straight, 17 mm. long, 3 mm. wide at the base; lateral sepals round-ovate, ca. 8 mm. long; lower sepals ca. 8 mm. by 5 mm.; upper petals yellowish or tinged blue, the lamina ca. 7 mm. JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 271 long, slightly oblique, entire at the roundish apex, the spur 17 mm. long; lower petals short-bifid near the apex, sparsely pilose, blue or whitish, the blade strongly oblique, oblong, 6 mm. long, the claw 5 mm.; stamens ca. 6 mm. long, somewhat dilated at the base, glabrous, the round- oblong anthers bluish, 1 mm. long; carpels 3, short- hirsute, 9-10 mm. long. Type: Huth cited four specimens, all in the Turkestan Herbarium in Leningrad. The one sent me on loan from Leningrad was “Trajectus montium Gasi Mailik, supra Choschbalak,” 6000 feet, Regel in 1883 (LE). I therefore choose dis as lectotype. DIstrIBUTION. At about 2000-3000 m., Turkestan or Asia Media. REPRESENTATIVE SPECIMENS. Lipsky 1685; Lipschitz 685; Regel at Chosch Rawat 14. Delphinium biternatum Huth, Bot. Jahrb. 20: 422. 1895. Fic. 3, D. etc floribundum Freyn & Sint. ex Freyn, Bull. Herb. Boiss. II. 3: 561. 1903. Based on Sintenis 850, Aschabad, Suluklii (seen at BM, E, MO, P). Stems glabrous below, more or less strigulose above, branched at middle or above, 4-12 dm. tall; basal leaves gone by anthesis; cauline leaves subglabrous, short petioled, the petioles dilated at the base, the blades ternate with petiolulate segments cuneate at the base and divided into 3 subsessile, lanceolate, entire or lobed parts, mucronate, 2~5 mm. wide, the leaf laminae 4-10 cm. broad; terminal raceme to 30 cm. long, the lateral shorter, ascending; pene bracts leafy, trifid, to ca. 1.5 cm. long, the upper entire, to 12 mm. long; pedicels densely glandular-pubescent, shorter than the flower, the lance-linear bractlets near the base of the flower, 2-3 long, claw 6.5 mm.; drawn from Rechinger ly (us). E.D. ee: upper amina 9 mm. long, spur 10 mm.; lower lamina 6 mm. long, claw from type (LE). F. D. carduchorum: cock aed 7 mm. long, cic 11 mm.; wer lamin . long, claw 4 mm.; drawn from Davis rola srk = D. centeteroides: upper lamina 9.5 m m. long, spur 20 mm.; lowe mm. ri claw 4m Apa from Grifith 45 — H. DP. chodat “pe ina m. long ca - lower lamina 6 mm. long, claw m drawings in git description. ct cyphoplectrum: upper . 9 mm. penn spur 10 mm.; low n . long, claw 5 mm.; drawn from Kotschy 153 (Mo). eye dasystachyum; upper lam toi spur 9 mm.; lower | long, cla .; drawn from Balansa 859 (cH). K. D. dolichostachyum; upper lamina 7 mm. long, spur 11 mm.; lower lamina : lon, r 22 mm.: lower lamina 4.5 mm. lon : = Hk Sintewis 589 (us). M. D. iivesip aii upper lamina 9 mm. long, spur 10 6m m.; lower lamina 4.5 mm. long, claw ; drawn from Rechinger 6191 (us). ND iseum; upper ina 10 mm 1onig, ‘spur 10 mm.; er lamina 5 mm long, claw 5 mm.: drawn from type (w). O. D. inopinatum; upper lamina 8 mm long, spur 12 wer lamina 5 mm. long, cla m.; drawn from TY ) D. ithaburense; upper lamina 10 mm. long, spu ; lower lamina 6 mm. long, claw 6 mm ouglas 39 (us). Q. D. phe, ak upper lamina 8 mm. ees, spur 10 mm.; lower lamina 7 mm. long, claw 6 mm.; drawn from TYPE (LE 272 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 mm. long; sepals ochroleucous, strigulose, the upper 11-12 mm. long, ovate, subacute, the spur pointed, 13-15 mm. long; lateral sepals more or less oblong, strigulose on midrib and finely puberulent at apex; lower sepals elliptic-ovate, strigulose especially in upper parts; upper petals with lamina ca. 11 mm. long and with 2 lobes 2.5 mm. long, the spur 11 mm. long; lower petals ca. 12 mm. long, including a very short spur (1.5 mm. long), bearded, bilobed for ca. 4 mm.; stamens 4-6 mm. long, glabrous, dilated at base; anthers yellow, 0.7 mm. long; follicles 3, hirsute, 10 mm. by 2.5 mm., the style an additional 2 mm.; seeds obpyramidal, 1.2 mm. long, densely clothed with transverse rows of scales. Type: (lectotype) in montibus Kuh-i-Frusch, Turkestan, 7000 ft., June, 1883, Regel (LE); seen. DistrisuTion. At about 1700-2500 m., Turkestan, Transcaspia, Kho- rasan (Iran). ILLUSTRATION. Nevski in Komaroy, Fl. U.S.S.R. 7: pl. 11, fig. 2a, 20. 1937. REPRESENTATIVE COLLECTIONS. Transcaspia: Sintenis 500. Turkestan: Socalski 292; Popov 259; Mikelson 2279; Chaffanjon 804; Komarov, June 26, 1893. Iran: Khorasan, Rechinger 1416. 15. Delphinium bucharicum M. Popow in Korowin, Kultiasow & Popow, Descr. Pl. Nov. Turkestan. 49. 1916. Fic. 3, E Stem 3-4 dm. high, straight, very slender, sulcate, short-pubescent below, naked above; leaves 2-3 at base of plant, the petioles 4-5 cm. long, vaginate; lamina 3-lobed, the lobes sessile, many times dissected into long narrow, somewhat pubescent laciniae less than 1 mm. wide; racemes simple, with about 8 crowded flowers; bracts membranaceous on margins, linear-lanceolate, ca. 4 mm. long; pedicels 4-7 mm. long, glabrous; brac- teoles small, below the middle of the pedicel; flowers pale, whitish-blue; sepals glabrous, with dark blue veins, elliptic-oblong, 7-8 mm. long, the spur 8-9 mm. long, curved upward and with a small hump near the tip; petals scarcely lobed at apex, bluish white, the upper laminae perpendicu- lar to the spur, ca. 9 mm. long, with a few bristles near the tip, the spurs 10 mm. long; lower petals bearded throughout, the laminae 6 mm. long, oblong, bilobed 1/3 their length, the claw 6 mm. long; follicles 3, glabrous. Type: In the Tashkent Herbarium. A specimen loaned me from Len- ingrad as type material was Kultiasow 626, April 29, 1914 (LE). DistriBUTION. Endemic on mountain slopes, middle Asia; described from mountains west of Bishkent. opov in a corrected supplement to his work, reduced his species to D. leiocarpum Huth, but Nevski in Komarov, Fl. U.S.S.R. 7: 170. 1937, maintains it. It has the hairiest lower petals that I have seen. 16. Delphinium carduchorum Chowdhuri & Davis, Notes Bot. Gard. Edinb. 22: 407. 1958. Fic. 3, F. Stem 2.5-6 dm. tall, erect, acutely angled and subsulcate below, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 273 branched, leafy above, strigulose to subglabrous below, spreading glandu- lar-pubescent above; lower petioles 4-8 cm. long, more or less vaginate, the blades orbicular, 3—6 cm. in diameter, palmately trisect into glabrous laciniate segments 2-4 (—5) mm. wide, linear-lanceolate to linear or ob- long-lanceolate; upper cauline leaves short petioled, trisect, the linear segments entire or divided almost to the base; raceme terminal, 7- to 20- flowered, lax or somewhat dense, 8-15 cm. long, glandular-pubescent; lower bracts 2.5—-5 cm. long, mostly deeply tripartite, strigose, equaling or surpassing pedicels; upper bracts entire, lance-linear, ciliate, shorter than flowers; pedicels 8-15, or the lower to 30 mm., strict; bracteoles in upper part of pedicel, 5-8 mm. long; sepals intense blue or somewhat violet, oblong-ovate or elliptic-obovate, acute or obtuse, 9-13 mm. by 5—7.5 mm., somewhat strigose; spur 11—16.5 mm. long, attenuate, hori- zontal or somewhat curved; petals included, the upper falcate, the lamina 7 mm., short-bifid, the spur 11 mm. long; lower 7.5-10 mm. long, the ovate limb 4-6 mm. by 3-3.5 mm., acutely lobed to about the middle, bearded; filaments glabrous; follicles 3, oblong, pubescent, 7 mm. by 3 mm.; seeds triquetrous, squamate. Type: Turkey, prov. Hakkari, Cilo Dag above Diz deresi, 2400 m., Davis 23954 holotype (£), isotypes (BM, K) ; seen. DistrIBuTIon. Rocky places, 2250 to 2850 m., Turkish Kurdistan and adjacent Iran. REPRESENTATIVE COLLECTIONS. Turkey: Davis 24370, 23349, 2532 cena, 22801, 23061, 22734. Related to D. micranthum of Iraq, which has a more panicled strigose inflorescence. 17. Delphinium centeteroides (Briihl) Munz, comb. nov. Fic. 3, G. Delphinium saniculifolium Boiss. subsp. centeteroides Briihl in Briihl & King, Ann, Bot. Gard. Calc. 5: 95. 1896. Stem ca. 3-9 dm. high, openly and rather freely branched, sparsely pubescent above, the hairs short, thickened at base, apparently glandular; leaves mostly withered by anthesis, the lower long petioled (to 12 cm. and with long dilated base), the upper shorter, remote, lower laminae ap- parently with few broad cuneate lobes, the uppermost trisect, 2-5 cm. long, the ultimate divisions linear; inflorescence an open panicle of short, rather few flowered, compact racemes; bracts simple, lance-linear, 4-5 mm. long; pedicels ascending, pubescent, mostly 1—2 cm. long; bracteoles lanceolate, ca. 3 mm. long, near the middle of the pedicel; sepals blue, strigulose; upper sepal ovate, ca. 10 mm. by 4 mm., pointed; spur 16-20 mm. long, slender, subcylindric, more or less curved; lateral sepals broadly oblong-elliptic, 10 mm. by 4 mm., rounded at tip, strigose on mid-line; lower sepals asymmetrically oblong-obovate, 10 mm. by 4.5 mm.; upper petals with oblique glabrous lamina 9-10 mm. long, unequally shallowly bilobed: spur to 20 mm. long; lower petals with a claw ca. 4.5 mm. long, 274 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 the oblong-ovate, bearded blade 7 mm. long, shallowly 2-lobed; stamens 6-7 mm. long, glabrous; anthers 1 mm. long; follicles hairy; styles 2.5 mm. long; seeds with short appressed scales. Type: Afghanistan, Griffith no. 45, Herb. East India Co.; isotypes seen (GH, K). DIsTRIBUTION. At about 2200 to 3300 m., Afghanistan and West Paki- stan (Baluchistan). ILLUSTRATION of parts of flower, BRUHL & Kine, Ann. Bot. Gard. Calc. 5: pl. 117, fig. 4g, 5h, 6e. REPRESENTATIVE SPECIMENS: Afghanistan: Hay 373. W. Pakistan: Helen Crookshank 324, 366; Lace 3862; Jafri & Akbar 2197; Nasir 28459. This proposed species is of slender, strigulose, and pale-flowered plants. 18. Delphinium chodatii Oppenheimer, Bull. Soc. Bot. Genéve II. 26: 94. 1936. 3, aa Slender-stemmed, ca. 2.5 dm. tall, the stem sulcate, glabrous through- out; lower leaves 3—4, sparsely hirsute, the others glabrous, palmatifid, the segments divided into narrowly linear lobes, the blades ca. 2 cm. long; petiole somewhat dilated at the base, the lower longer than the upper; raceme leafless, rather lax, ca. 10-flowered; bracts lance-ovate, shorter than the pedicels; pedicels terete, recurved in flower, erect in fruit, thickened at summit; bracteoles basal, linear; sepals 9-10 mm. long, glabrous, pale rose; spur subattenuate, 9-12 mm. long; upper petals falcate, strongly oblique, the lamina ca. 8 mm. long, 2-toothed at apex, the spur 8 mm. long; lower petals obovate, the lamina densely bearded, divided 2/3 its length and ca. 6 mm. long, the claw 6-7 mm.; stamens glabrous; follicles 3, hairy; seeds obovoid-triquetrous, brown, winged. Type: Ouadi Kelt, Palestine, “entre le Moulin et l’embouchure de la vallée de Jéricho,” April 21, 1934, apparently at (G). Not seen. ILLusTRATION. Bull. Soc. Bot. Genéve II. 26: 94. fig. 1, 2. 1936. No specimens seen. Drawings copied from those in original description. The glabrousness of the plants and the pale rose flowers seem distinctive. 19. Delphinium cyphoplectrum Boiss. Diagn. I. 6: 7. 1845. Fic. 3, 3: Delphinium cyphoplectrum var. micranthum Boiss. Fl. Orient. 1: 91. 1867. Delphinium schroedingerianum Hand.-Mazz. Ann. K. K. Nat. Hofmus. Wien 27: 44. 1913. Based on Kotschy 393a and 400, near Persepolis. The latter seen (kK). Stem terete, branched above, 3-10 dm. tall, subglabrous to somewhat strigulose below, more loosely appressed-pubescent above; cauline leaves few, remote, with petioles to ca. 1 dm. long and somewhat vaginate at the base, the lower leaves palmately parted, then divided into oblong lobes 3-5 mm. or more wide, the laminae 5-8 cm. wide; racemes several, lax, 1967] MUNZ, ASIAN SPECIES OF DELPHINIUM 275 several flowered; bracts mostly entire, the lower to 2 cm. long, 1.5-3 mm. wide, the upper much reduced; pedicels densely appressed-puberulent, 5— 10 mm. long, the 2 bracteoles near the base of the pedicel, 2-3 mm. long; sepals blue to smoky gray, the upper 12 mm. by 7 mm., obtuse, strigose, the spur straight, obtuse, subcylindric, 10 mm. long, ca. 3 mm. wide at the base; lateral sepals subelliptic, ca. 12 mm. by 4.5 mm., rounded at apex, somewhat strigose near end of midrib; lower sepals elliptic-obovate, broad- ly obtuse, ca. 13 mm. by 5 mm.; upper petals blue, the lamina obliquely truncate, 9 mm. long, bilobulate; lower 10-11 mm. long, the claw 4.5 mm., the limb ca. 6 mm., elliptic, bearded, the 2 parallel lobes 2.5 mm. long; stamens 8-9 mm. long, glabrous, slightly dilated in lower half; anthers 0.8 mm. long; follicles 3, pubescent Type: Mt. “Kuh-Malo prope Dalechi Persiae australis,” March, 1842, Kotschy 153, presumably (G); isotypes seen (A, BM, K, MO, P DIsTRIBUTION. Between 2000 and 2700 m., southern Iran. ExampLes. Haussknecht, April, 1868; Furse 3061; Loftus, May 3, 1852; Horner 153; Kotschy 400 20. ai aciiia dasystachyum Boiss. & Bal. in Boiss. ee 135.83 7. 1859. G. 3, J. eet dasystachyum var. ochroleucum Boiss. Fl. Orient. 1: 88. 1867. Based on Mt. Berit Da, Hausskn., cited by Davis, Fl. Turkey 1: 113. 1965, as example of D. dasystachyum. Plant cinerascent-strigulose throughout, the stem simple, more or less angled, 2—4.5 dm. tall, leafy to the inflorescence; lowest leaves 2-3 cm. broad, with petioles twice as long and vaginate at base, the laminae multi- fid into ultimate linear lobes ca. 1 mm. wide; upper cauline leaves more crowded, with short broad membranous petioles; inflorescence dense, many flowered, to ca. 1 dm. long; bracts 3—5-fid to the broad short petiole, Strigulose, 1-2 cm. long; pedicels arcuate, 5-10 mm. long, the lanceolate acuminate bracteoles ca. 5 mm. long, 2 mm. wide, just beneath the flower; sepals broad based, bluish to dirty white, densely and loosely strigulose; upper sepal 8 mm. by 3 mm., obtuse, lance-ovate, the spur 10 mm. by 3 mm., acute, descending, slightly curved; lateral sepals 8 mm. by 3 mm., oblong, obtuse; lower sepals ca. 9 mm. by 3 mm., obtuse; upper petals glabrous, bluish to yellowish, oblique, ca. 8 mm. long, bidentate, the spur the same length; lower petals with claw 4-5 mm. long, lamina oblong- ovate, bearded, 4-5 mm. long, with terminal lobes ca. 2 mm. long; stamens yellow, 4-5 mm. long, glabrous, anthers 0.7 mm. long; follicles 3, ca. 6 mm. long, hairy, rugulose toward apex; seeds truncate, with transverse undulating rows of white lamellae. Type: Summit of Ali-Dagh, at 1700 m., 7 km. se. of Caesarea, Cappa- docia, Turkey, Balansa 859 (probably G); isotypes seen (BM, GH, K, P). DistriputTion. Rocky slopes, 1500-2500 m., eastern Turkey. Examptes. Davis 31669, 31999; Balls & Gourlay 1992, 1977; Purse 3732. 276 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 21. Delphinium dolichostachyum Chowdhuri & Davis, Notes Bot. Gard. Edinb. 22: 408. 1958. Fm: 3,i ta Plant 5—8.5 dm. tall, the stem erect, acutely angled, subsulcate, branched above the middle, strigulose above, subglabrous below; lower and median cauline leaves palmately 3—5-sect, 5-8 cm. broad, subglabrous, long petioled, the segments 3—5-parted into lance-linear or lanceolate ultimate parts 2-5 mm. wide; petioles subvaginate at the base; upper leaves few, trisect; terminal raceme 13-23 cm. long, dense, many flowered, longer than the lateral ones; bracts prominent, the lower 1-3 bracts foliaceous, plurisect, longer than the pedicels, the middle tripartite into linear seg- ments, the upper entire, linear-subulate; pedicels 9-25 mm. long, ascend- ing-erect; bracteoles linear-subulate, on upper part of pedicel; sepals pale blue, 6-9 mm. 3-4 mm., oblong-obovate or elliptic-ovate, obtuse or rounded at tip, densely strigulose, the spur 9-12 mm. long, slender, at- tenuate, acute, usually curved; petals included, the upper glabrous, with a short-bifid almost straight lamina 7 mm. long, spur 10 mm. long; lower petals with broad-ovate limb 4~5 mm. long and broad, bearded, bifid above the middle, the claw puberulent; stamens 5—6 mm. long; anthers yellowish, rounded, almost 1 mm. long; follicles 3, mostly 7-9 mm. by 3—4 mm., hirsute-pubescent; seeds 1.5 mm. long, scaly, triquetrous. Type: Turkish Kurdistan, prov. Bitlis, Karz Dag above Kamer, 2200 m., Davis 24566 (Davis & Polunin 8-24-1954); holotype (£); isotypes (BM, K); seen. A specimen collected by Are Rowi 12213 and one by Gillett 11782 from Iraq (K) seem to belong here. The species resembles D. dasystachyum in its small pale flowers, but differs in its tall branching, strigosity, less vaginate petioles, and tripartite (not multi-) lower bracts. 22. Delphinium fissum Waldst. & Kit. subsp. anatolicum pei & Davis, Notes Bot. Gard. Edinb. 22: 404. 1958. c. 3, L. gee ane leiocarpum Huth, Bull. Herb. Boiss. 1: 334. 1893, as to type: Sintenis 589, Kaz Da, near Kareikos, Turkey; isotype seen ( Kj). Dame-as alien Delphinium amani Post ex Huth, Bull. Herb. Boiss. 3: 152. 1895. Based on Shepard (Barbey 295), in 1892, Gaiour Da (Amanus); isotype seen (K). Stem 5—9 dm. tall, 3-6 mm. thick, sometimes with vertical lines of short spreading hairs above; leaves few, the lower glabrous, with petioles very slightly dilated at the base, to 1 dm. long, the blades 7-18 cm. wide, palmately multifid into ultimate lance-linear to oblong segments 2~8 mm. wide; upper leaves reduced; raceme dense, 5-10 or more cm. long, many flowered, subglabrous; bracts simple, linear, 5-8 mm. long; pedicels re- curved and subclavate at summit, mostly less than 1 cm. long; bracteoles near middle of pedicel, linear, 2-3 mm. long; sepals pale blue, glabrous: upper sepal ovate, 8 mm. long, 4 mm. wide, obtuse, the spur often sub- 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 277 vertical, straight, ca. 15 mm. long; lateral sepals oblong, obtuse, ca. 6 mm. m.; the lower pair oblong-ovate, 6 mm. by 3 mm.; upper petals glabrous, pale, 7.5 mm. long, bilobulate, the spurs ca. 15 mm. long; lower petals with lamina ca. 4.5 mm. long, white bearded, the lobes 3.5 mm. long; stamens 5—6 mm. long, filaments dilated below, glabrous; anthers yellow, 0.8 mm. long; follicles 3, glabrous. Type: Mt. Ida, near Kareikos, prov. Cannakale, Turkey, Sintenis 589, holotype (kK), isotypes (BM, P, US) ; all seen. DIsTRIBUTION. Coniferous forest, 1000-1650 m., Anatolia. Subsp. fis- sum is European. EXAMPLES. Anatolia: Manissadjian 675; Davis 13140, 14703; Haradjian 571, 552. 2IS4. 23. Delphinium gorganicum Rech. f. Ang. Osterr, Akad. Wiss. Math.- Nat. 88: 226. 1951. Fic. 3, M Stem erect, 1-1.3 m. high, subsimple or with few branches above, some- what fistulose, scattered-leafy, pale green, striate, glabrous or sparsely strigulose, especially abcve; lower leaves gone by anthesis, the petioles apparently long, the upper much shorter, subglabrous, slightly dilated at the base; leaf blades angled-cordate in outline, to ca. 1 dm. broad, 2—3- palmatisect into petiolulate segments that are long-cuneate at the base, obovate above and divided into 5—7 lance-linear lobes 5-8 mm. wide, 2—4 cm. long, subglabrous, paler beneath than above, acute, submucro- nulate; uppermost leaves simpler, passing into leafy divided bracts 3-5 cm. long, upper bracts strigose, trifid to lance-linear, 5-10 mm. long; terminal raceme lax, many flowered, to 3 dm. long; pedicels suberect, 1-10 mm. long, with yellow glandular pubescence; bracteoles 2, subulate, near the base of the pedicels; sepals yellow, puberulent, the upper sepal broadly ovate, 10 mm. by 8 mm., acutish, the spur 9-12 mm. long, almost straight, ca. 3 mm. wide at base, narrowed rapidly to the narrow apex; lateral sepals elliptic-obovate, 12 mm. by 5 mm., obtuse; lower sepals 12 mm. 7 mm., obovate, obtuse; petals yellowish, the lamina of the upper slightly oblique, entire at the obtuse tip, ca. 9 mm. long, the spur 10 mm. long; lower laminae bearded, ciliate, oblong, 4.5 mm. by 3 mm., deeply cleft, the lobes lance-oblong, claw 6 mm. long, serrulate on upper edge; stamens almost 8 mm. long, somewhat ciliate, moderately widened; anthers light in color, round-oblong, 0.8 mm. long; follicles 3, pubescent. Type: Iran: prov. Gorgan (Asterabad), Mt. Shawar near Hadijilang, 2400-2600 m., K. H. & F. Rechinger 6191 (w) ; isotypes seen ( K, UC, us). Very near D. biternatum; doubtfully distinct. No other collections seen. 24. Delphinium griseum Gilli, Repert. Sp. Nov. 57: 96. 1955. Fic. 3, N. Stem clothed at base with brownish black sheaths formed by old petiole bases, erect, angular-striate, subflexuous to bases of petioles, 6—7 dm. tall, 278 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 with few short branches in upper part and with short retrorsely appressed pubescence; lower petioles to 13 cm. long, dilated into a sheathing base, short-puberulent; upper shorter; blades roundish in outline, the lower 5-8 cm. broad, 3-parted almost to base, the parts cuneate, 2-3 cm. wide, 3-4 cm. long, their upper part divided into lobes 1-1.5 cm. long, 3-15 mm. wide; bracts linear, strigulose, ca. 3 mm. long; terminal raceme lax, 2-3 dm. long, ca. 20-flowered, the lateral racemes 12-15 cm. long, 10- flowered; pedicels ca. 1 cm. long, elongating in fruit, bibracteolate, closely and finely strigulose; sepals gray-violet, oblong, attenuate at base, obtuse, 9 mm. long, 2 mm. wide, appressed-puberulent on whole outer face, the upper sepal with a slightly curved obtuse spur 1 cm. long, puberulent, somewhat gibbous above the base; petals whitish, the upper glabrous or nearly so, the blade and spur each 10 mm. long, the former with 2 lanceo- late lobules; lower petals lanceolate-spatulate, long stipitate, deeply lobed into acutish lobes with long hair on both sides, blade and claw each about 5 mm. long; filaments gray violet, dilated at base, sparsely short-hairy; anthers dark brown; follicles 3, slightly pilose toward the tips to almost glabrous; immature seeds trigonous, squamate or rugulose. Type: In Kabul Gorge, southeast of Surobi, east Afghanistan, at 1050 m., May 28, 1951, Gilli 839 (w); seen. Another collection from “Sarobi,” May 12, 1951, O. H. Volk 1873 (w). This species is very near Delphinium kohatense of the northwestern Himalaya, but apparently has more flowers in a raceme and they are larger. I am greatly indebted to Professor Rechinger and Dr. Riedl of the Natural History Museum of Vienna for the loan of the two specimens cited above. 25. Delphinium ilgazense Davis, Notes Bot. Gard. Edinb. 26: 169. 1965. Stems 1.5-2 dm. tall, simple, terete, striate, sparsely strigulose below, with spreading glandular hairs in inflorescence; lower leaves long petioled, suborbicular, 2-3 cm. wide, palmatisect, strigose, the segments divided into ultimate linear laciniae ca. 0.5 mm. wide; inflorescence dense, 2.5—5 cm. long, 4- to 7-flowered; bracts viscid-pilose, the lower trisect or all entire, linear; pedicels erect, 5-7 mm. long; bracteoles in upper part of pedicel, lance-linear, 2.5-4.5 mm. long: sepals dark violet-blue, divergent, ovate-oblong, spreading-pilose, 7-10.5 mm. long; spur 13-14 mm. long, gradually attenuate, acute, horizontal: upper petals slightly exserted, the lamina glabrous, short-bifid; lower petals with a broadly ovate, deflexed, deeply bifid lamina bearded above and equal to the claw; follicles strigose. Type: Kastamonu, Turkey, north side of Ilgaz Dag, 2200 m., on rocky limestone slopes, Aug. 28, 1962, Davis 38373 () which I have not seen. Known from the type collection only; illustrated as plate 17, with the original description. Endemic; Davis considers it near D. cyphoplectrum. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 279 26. Delphinium inopinatum Nevski in Komarov, Fl. U.S.S.R. 7: 73, 721, 19345 Fic. 3, 0: Stem 7.5—8.5 dm. tall, leafy, almost wholly yellow glandular-pubescent; petioles 5-10 cm. long, widely dilated at base, densely pubescent; laminae of leaves 4-8 cm. broad, ternately parted, the first segments short petiolu- late, divided bi- to triternately into terminal narrowly linear (1.25—2.5 mm. wide), densely pubescent laciniae; inflorescence mostly branched, the central raceme rather dense, 2.5—3.5 dm. long, the axis sparsely glandular- pilose; bracts minute, 2.5 mm. long, linear, glabrous or ciliolate; pedicels mostly less than 1 cm. long, glabrous; bracteoles linear-subulate, inserted above the middle of the pedicels; sepals pale, sordid bluish-lilac, 8-9 mm. long, 3.5-4 mm. wide, oblong-obovate, obtuse, glabrous, white-membra- nous on margins; spur nearly or quite straight, spreading to erect, glabrous, 10-11 mm. long, 2—2.5 mm. wide; petals whitish, the upper laminae perpendicular to the spur, glabrous, ca. 8 mm. long, bidentate, the spur 12 mm. long; lower laminae slightly oblique, long-hairy and ciliate, more or less laciniate, ca. 4 mm. long, the claw 4 mm. long; follicles 12-15 mm. long, 4 mm. wide, erect, glabrous; seeds pale, 2.5—3 mm. long, with trans- verse rows of scales. Type: Asia Media, “in montibus Ak-Tau” (Kazakhskiy), May 19, 1932, Russanov 119 (LE); seen. The type is the only specimen that I have seen. This species differs from D. bucharicum in its indument of yellow glan- dular hairs. 27. Delphinium ithaburense Boiss. Diagn. I. 8: 9. 1849. Fis. 3, P. Delphinium sulcatum Reichb. ex Hoffmannsegg, Verz. Pfl. Nachtr. 1: 132. 4, Nomen Delphinium salmoneum Mouterde, Fl. Djebel Druse 98. pl. III, no. 2. 1953. No type cited, but in the illustration, the legend says ‘‘Tell Ahmar, 15 May, 1942.” I have seen no specimen, and cannot separate it as a species from the description alone. Stem 3-10 dm. tall, angled, simple, subglabrous to somewhat glandular- pubescent; leaves palmately 5—7-parted into cuneate segments, In turn coarsely toothed or lobed, the upper leaves laciniate into oblong-linear lobes ca. 2 mm. wide; petioles vaginate at base; raceme elongate, simple, rather many flowered, often rather dense at anthesis, lax in fruit; bracts lance-ovate, 5-10 mm. long, ciliate or glabrous; pedicels mostly 5-10 mm. long, the lance-ovate bractlets 2-3 mm. long and near the base of the pedicel; flowers bluish white to pink or yellowish green; sepals mem- branous margined; upper sepal lance-oblong, 9-10 mm. long, subglabrous except for the pubescent patch below the acutish tip, the spur straight, glabrous, horizontal, ca. 15 mm. long, 2.5 mm. wide at base; lateral sepals oblong, ca. 9 mm. by 3.5 mm., rounded at the apex, with broad base, slightly pubescent toward the tip of the midrib; lower sepals elliptic- 280 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 oblong, 9 mm. long, 4 mm. wide, slightly asymmetrical; upper petals gla- brous, obliquely truncate, the lamina 10 mm. long, 2-lobulate for 1 mm., the spur 15-17 mm. long; blade of lower petals 6-7 mm. long, bearded, with 2 lance-oblong lobes 3-4 mm. long, claw 3-4 mm. long; stamens 5—7 mm. long, lower half white, oblong, upper half yellow, narrow; anthers yellow, scarcely 1 mm. long; follicles 3, erect, hirsute, 13-16 mm. long, 4-5 mm. wide, the beak 3-5 mm. long; seeds ca. 2 mm. long, with several rows of scales. Type: Mt. Thabor, Galilee, Palestine, Boissier in 1846 (G) not seen; isotype (P) seen. DIstTRIBUTION. Wooded places at 100-2000 m.. mountains of Syria, Lebanon, Israel and Jordan. ILLUSTRATION. Mouterpe, Fl. Djebel Druse PI. IFT; id a T9538: Examples. Lebanon: W. O. Douglas 39. Jordan: Dinsmore 1864. Israel: Davis 4608a, 4894A, 4896, 4608; F. Meyers B4864; Egger 6-29-1906. 28. Delphinium karategini Korsh. Bull. Acad. St. Pétersb. sér. 5. 9: 402. 1898. Fic. 3, Q. Stem 8-15 dm. tall, freely branched above, rather equably leafy, with scattered more or less retrorse short hairs; petioles vaginate at base, 5—12 cm. long, the blades to 15 cm. broad, parted into long-petiolulate segments, these in turn divided into petiolulate parts which are dissected into linear laciniae to 3 cm. long and 1.5—2 mm. wide, more or less pubescent; upper- most leaves reduced; inflorescence of several divergent strict branches 1.5-5 dm. long, each rather openly many flowered, the axis glabrous; lowest bracts trifid into linear segments, others linear-subulate, glabrous, 6-9 mm. long; pedicels erect-divaricate, to ca. 10 mm. long, glabrous, the bracteoles near or below the middle of the pedicel, lance-linear, ciliate, ca. 3 mm. long; sepals pale blue, somewhat strigose, membranous margined; upper sepal 10-11 mm. by 4-5 mm., acute, the spur spreading upcurved, 12 mm. long, 3 mm. wide at base, blunt; lateral sepals oblong-obovate, rounded at apex, ca. 9 mm. by 4 mm.; lower sepals ca. 10 mm. by 4 mm., acute; petals whitish, the upper laminae strongly oblique, 7 mm. by 3 mm., glabrous, entire, the spur 10-11 mm. long; lower laminae almost straight, oblong-ovate, 7 mm. by 3 mm., bifid over halfway into lance- oblong lobes, long-ciliate and bearded, the claw 6 mm. long; stamens 7 mm. long, glabrous, the anthers pale, oblong, 1 mm. long; follicles 3, glabrous. Type: Turkestan: Karategia, “ad ostium fl. Muksi,” June 23, 1897, Korshinsky 80 (LE); seen. 29. Delphinium kohatense (P. Brithl) Munz, comb. nov. Fic. 4, A. Delphinium saniculifolium Boiss. var. kohatense P. Brihl in Briihl & King, Ann. Bot. Gard. Calc. 5: 95. 1896. Delphinium saniculifolium Boiss. var. gilgitense P. Brihl in Briihl & King, /oc. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 281 cit. Based on J. L. Stewart, Sept. 23, 1885, at 13,000’; Gilgit Expedition s. of Hindu Kush (x); seen. Stem slender, terete, mostly few branched above, 2—5 dm. high, sparsely strigulose below, densely so in the inflorescence; lower leaves 2—4 cm. broad, subglabrous to somewhat strigulose, trifid to base into cuneate segments, of which the two lateral are deeply divided so as to make 5 sub- equal, pinnately few lobed or toothed parts, the ultimate divisions 1-3 mm. wide, mucronate; petioles 2-7 cm. long, slender, somewhat dilated at base; cauline leaves few, remote, reduced; racemes rather open, to Ca. 10-flowered, slender, strigulose; bracts linear, 2-3 mm. long; pedicels ascending, 5-25 mm. long; bracteoles lanceolate, near the base of the pedicel; sepals pale blue, blunt, the upper 6-8 by 3 mm., strigulose, the spur cylindric, obtuse, 5-7 mm. long; lateral sepals 7.5-8 mm. by 2.5 mm.; lower sepals ca. 8 mm. by 3 mm., oblong-oblanceolate; upper petals ca. 7.5 mm, long, pale, oblique, the bearded lamina with terminal lobes 1 mm. long; spur 5 mm. long; lower petals ca. 8 mm. long, lamina and spur subequal in length, the former bearded, divided for 2.5 mm.; stamens 4.5— 5.5 mm. long, the filaments bluish, somewhat ciliate on expanded lower part; anther 0.8 mm. long; follicles 3, pubescent at tip. Lectotype: Pakistan: Kohat & Wagiristan, Drummond (cat), seen; but specimen old and almost without flowers. I am greatly indebted to Mr. K. S. Srinivasan for loan of the above and other specimens from the Calcutta Herbarium. What is apparently the same thing, collected by J. L. Stewart, at dry places up to 3500’, (K) was seen. ILLUstRATION. Briwy & Kino, Ann. Bot. Gard. Calc. 5: pl. 117, fig. 1 & 2. 1896 DistripuTion. Dry places, mostly below 3000 m., Afghanistan and West Pakistan to northwest India. REPRESENTATIVE SPECIMENS. Afghanistan: Herb. Griffith 43. Punjab: Aitchi- son 1004; Drummond 14811; R. R. Stewart 10936. West Pakistan: Chitral, Stainton 2529; Siddiqui & Arahman 26865; Toppin 370; Dr. Giles, June 24, 1886, 30. Delphinium kurdicum Boiss. & Hohen. in Boiss. Diagn. I. 1: 67. 1842. Fic. 4, B. Delphinium kurdicum var. elongatum Freyn & Sint. Osterr. Bot. Zeitschr. 41: 363. 1891. Based on Sintenis 2822 from Mt. Sokardidagh, Armenia; not seen, but cited by Davis as example of D. kurdicum in F\. Turkey 1: 116. 1965. Stem somewhat zigzag, few branched, 3—5 dm. tall, more or less strigulose with very short hairs, many of which are enlarged at base; lower leaves withered by anthesis, upper multifid into few linear or lanceolate lobes 8-20 mm. wide, 3-8 cm. long; racemes in an open panicle, rather few to many flowered, more or less peduncled; bracts lance-linear, 5-10 mm., rarely to 30 mm. long; pedicels somewhat divaricate, 5-15 mm, long; 282 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 48 [phi the upper spurred, the lower clawed. A. D. kohatense; upper lamina 7.5 mm. long, s ; lower lamina long, claw 4.5 mm.; drawn from R. R. Stewart in 1936 ( B. D. kurdicum; upper lamina g, spur 10 mm.; lower ina 4.5 mm. long m.; drawn from Kotschy 427 (mo). C atesquamatum; upper in long, spu mm.; lower lamina 6 mm. long, cl m.; drawn from Gilli 840 (w). D. D laxiusculum ; upper lamina m. long, spur 10 mm.; ower lamina 5 mm. long, c 5 ; drawn from Sjovits (GH) 2 mm. long, spur 10 mm.; lower lamina 8 mm. long, claw, 5 mm.; drawn from Balansa 1353 (us). F. D. longipedunculatum ; upper lamina 8 mm. long, spur 11 mm.; lower lamina 7 mm. long, claw 5 mm.; drawn from Mo. Bot. Gard. 1,720,217. GD, macropogon; upper lamina 7 mm. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 283 bracteoles 3-4 mm. long, situated above the middle of the pedicel; sepals yellowish to dirty white, strigulose to glabrous, the upper sepal ovate, 7.5-8 mm. long, 3.5 mm. wide, acute, the spur horizontal, 10-11 mm long, 3.5 mm. wide at base; lateral sepals 8 mm. by 3 mm., oblong, broadly unguiculate, pubescent on median line or more broadly; lower sepals 8 mm. by 3 mm., rather pointed, broadly unguiculate; upper petals glabrous, obliquely truncate, yellowish, the limb 7 mm. long, bidentate, the spur 10 mm. long; lower petals with claw 4 mm., blade 4.5 mm. long, the latter broadly elliptical, bearded, lobed ca. 2 mm. deep; stamens 4-5 mm. long, glabrous; anthers 0.8 mm. long; follicles 3, glabrous, 8 mm. by 3.5 mm., the styles an additional 2 mm. Type: “Ad fontes montis Gara antlers) . on 427 (originally published as 627) (G) not seen; isotypes (BM, se DISTRIBUTION. mie slopes and plains, 1600-1400 m., northern Iraq and eastern Turkey EXAMPLES. Turkey: Kotschy 503; Sintenis aa Davis 31496. Iraq: Ali al- Rawi & Serhang 23807; R. Wheeler Haines 4 31. Delphinium latesquamatum Gilli, Repert. Sp. Nov. 57: 94. 1955. Fic. 4, C. Stem 2.5—6 dm. high, clothed at base with whitish sheaths of old leaves, terete, striate, erect, subflexuous, with divaricate branches 2—4 dm. long, with short retrorse pubescence throughout, mixed with glandular hairs in the upper parts; petioles 1.5—3 cm. long in lower leaves; blades cuneate, 3-parted to the lower third, 2-8 mm. wide, the parts incised into ultimate laciniae 0.5—2 mm. wide; cauline leaves with petioles 0.3—-3 cm. long, the ultimate divisions of the laminae 0.5 mm. wide, acute; bracts linear, puberulent, 3-10 mm. long; pedicels 1-2 cm. long, bibracteolate mostly long, spur 10 mm.; lower lamina 4.5 mm. long, claw 4 mm.; drawn from TYPE (LE). H. D. macrostachyum:; _— lamina 6 mm. lon | spur 9 mm.; lower lamina 5 mm. long, ae ~ st .; drawn from Sintenis 1266 (put). I. micranthum; upper lam a On , spur 12 mm.; lower lamina 6 mm. ong, aw 4m mm.: drawn rary *Cillett 9501 (us). i dey muristanicum: upper lamina 5 . long, claw 6 m - drawn from W. E. Brown 804 (Kk). K. D. ochroleucum; upper lamina 7 mm "lee spur 8 mm . 4 J fr (GH). D. pallasii; upper lamina 10 mm. long, — 18 mm.; lower lamina 6.5 mm. long, claw 6 mm.; drawn from TYPE (LE). D. pallidiftorum ; upper lamina 9 mm. lon pur 10 mm.; lower lamina 5 mm. gine claw 5 mm.; from Heideman in 193 . N. D. penicillatum; upper lam spur wer lamina 5 mm. long, claw wn from Aucher 4034 Net O. D D. punice um; upper lamina 10 mm. long, spur 9 mm.; lower lamina m. long, claw 6 mm.; drawn from Hohenacker i in 1834 (cas). P. D. querce- paces upper lamina 9 mm. oe spur 1 ower lamina 5 mm. long, claw 5.5 m.; drawn from Haussknecht in . (pm). Q. D. roylei; upper bone 8 mm. long, spur 14 mm.; amina 5 mm. long, claw 5 mm.; T —. Ab 9261 (cH). R. D. sancufoium: upper lamina 7 mm. long, spur 9 mm.; lower lamina 5 mm. long, claw 4 mm.; drawn from Rechinyer 3364 (US). 284 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 above the middle, the bracteoles 2-3 mm. long; sepals blue, obovate, obtuse or acutish, 6-9 mm. long, 4-6 mm. wide, puberulent and glandular without, the spur slightly curved, not gibbous, 13-17 mm. long; petals somewhat yellowish below, bluish toward the tips, the upper lamina 10-11 mm. long, glabrous, almost round near the tip, where 5 mm. in diameter, crenulate, the spur ca. 12-13 mm. long; lower laminae 6 mm. by 5 mm., glabrous save for the patch of hairs near the base, with 2 lobes 1-2 mm. long, claw 3-4 mm. long; filaments glabrous; anthers dark brown; fol- licles 3, pilose, to ca. 8 mm. long; seeds with many tiers of broad scales. Type: Afghanistan, near Bamian, 2950 m., Aug. 29, 1951, Gilli 840 (w); seen. REPRESENTATIVE COLLECTIONS. All from Afghanistan: prov. Orozgan, Rechinger 18812; Lindberg 875, 964; central Nuristan, Kerstan 1399, at 2600 m.; Edelberg 1208 at 2000 m.; prov. Ghazni, Rechinger 17552, 17629, at 3300 m.; prov. Bamian, Rechinger 18698, 18449, 18250. The species seems near D. batalinii of Turkestan, but has much wider petals. I am very grateful to Dr. K. H. Rechinger and Dr. H. Ried] of the Natural History Museum at Vienna for the loan of the material cited above. 32. Delphinium laxiusculum (Boiss.) Rouy, Fl. France 1: 143. 1893. Nevski in Komarov, FI. U.S.S.R. 7: 175. 1937. Fic. 4, D. Delphinium hybridum Willd. var. laxiusculum Boiss. Fl. Orient. 1: 89. 1867. Delphinium hybridum subsp. laxiusculum N. Busch in Fl. Cauc. Crit. III. 3: 56. 1903 Delphinium hybridum var. gracillimum Ledeb. Fl. Rossica 1: 61. 1841. Placed ere by Nevski in Komarov, Fl. U.S.S.R. 7: 175. 1937. Delphinium hybridum var. hirtula Trautv. Acta Horti Petrop. I. 1: 21. 1878. Placed here by Nevski in Komarov, Fl. U.S.S.R. 7: 175. 1937. Delphinium cyphoplectrum Boiss. var. stenophyllum Boiss. Fl. Orient. 1: 91. 67. Isolectotype (Davis), Schambulagh, Iran, Szowits (K), seen. Delphinium talyschense Tzvelev, Not. Syst. ex Hort. Inst. Bot. Komarovii 19: 6. 1959. Type, Azerbaidschan, Talysch, Matveeva 555 (LE), seen. Stem 2-5.5 dm. tall, somewhat angled and sulcate, strigulose, quite densely so in the inflorescence, simple; lower leaves with petioles to 1 dm. long and vaginate at the base, blades 3-10 cm. wide; cauline leaves re- mote, the uppermost with shorter petioles, blades palmately 5-fid, the segments openly laciniate into few lobes 1-2 mm. wide; raceme 12-18- flowered, lax, strigose; bracts simple, lanceolate, 3-20 mm. long; pedicels mostly 5-15 mm. long; bracteoles 2, near base of pedicel, 2-3 mm. long; sepals blue, the upper strigulose, 12 mm. by 4 mm., lance-ovate, acute, the spur cylindrical, blunt, 12 mm. long, 2.5 mm. wide at the base; lateral sepals elliptical, 10 mm. by 4 mm., obtuse, strigose near the tip of midrib; lower sepals 12 mm. by 4 mm., acute, more generally strigose; upper petals bluish, oblique, ca. 8.5 mm. long, shallowly bilobulate, the spur 8.5—9 mm. long; lower petals bluish, claw 4.5 mm. long, blade 7.5 mm. long, bearded, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 285 narrow, with 2 lanceolate lobes 3-4 mm. long; stamens 6-7 mm. long, glabrous, dilated in lower half; anthers bluish, 0.7 mm. long; follicles 3, strigose, 11-20 mm. long, the beak an additional 2-3 mm. Type: “Hab. inter segetes prope Elisabethpol,’ Hohenacker in 1834 (G) not seen, isotypes (GH, K, LE, MO) seen. DISTRIBUTION. Transcaucasia, northern Turkey, and northern Iran. EXAMPLES. Transcaucasia: Gurvitsch in 1937. Iran: Furse & Synge 784; Pichler, in 1882. This plant is treated in Davis, Fl. Turkey 1: 115. 1965, as D. cyphoplectrum var. stenophyllum. 33. Delphinium longebracteatum (Boiss.) Munz, comb. nov. Fic. 4, E. Delphinium dasystachyum Boiss. & Bal. var. longebracteatum Boiss. FI. Orient. 1867. oe szovitsianum var. longebracteatum (Boiss.) Huth, Bot. Jahrb. 20: 4 895. Stem simple, subterete, 2-3 dm. tall, substrigulose below, densely pubescent above with soft yellowish spreading hairs; lower leaves with petioles 4-7 cm. long, much dilated at base, middle cauline and upper petioles shorter, the broad base membranous, 1-1.5 cm. long, almost 1 cm. wide; blades 2-3 cm. wide, multifid into linear lobes to ca. 1 mm. wide; inflorescence dense, 4-8 cm. long, soft-hairy, many flowered; lower bracts to 3 cm. long, with long linear segments and dilated membranous base, the upper ca. 1 cm. long, simple, with dilated hairy base and long attenuate tip; pedicels to 1.5 cm. long, more or less curved; bracteoles to ca. 9 mm. long, membranous, with subulate tip and placed about the middle of the pedicel; flowers yellowish, the sepals with long yellow or white hairs, upper sepal 11 mm. by 5.5 mm., obtuse, the spur 12 mm. long, descending, acute; lateral sepals oblong-obovate, 10 mm. by 4.5 mm., rounded at apex; lower sepals elliptic-oblong, 10 mm. by 4.5 mm.; upper petals oblique, the blade 12 mm. long, somewhat bearded on inner face of lamina and with terminal lobules ca. 1 mm. long; lower petals unguiculate, the claw ca. 5 mm. long, the lamina bearded, oblong, 5 mm. long, the 2 parallel lobes ca. 3 mm. long; stamens glabrous, 5—6 mm. long; anthers 0.8 mm. long; follicles 3, hairy. Type: Northeastern Turkey: o monte Bingoldagh Armeniae alt. 6000’,” Kotschy 754 (c). Not see Dassprinéceoiy. At 1500 to 2200 1 m., northeastern Turkey. Examp tes. Balansa 1353; Balls & Gourlay 1977a. So far as I can see these collections are referred to D. dasystachyum by Davis, Fl. Turkey 1: 113. 1965, but the more spreading pubescence of the upper stems, more yellowish sepals, longer upper petals, etc., seem to me to separate Delphinium longebracteatum from that species. 286 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 48 34. Delphinium longipedunculatum Regel & Schmalh. Acta Horti Petrop. 5: 226. 1877. Fic. 4, F. Stem slender, 3-6 dm. high, zigzag, slightly angled, sparsely pilose below, glabrous above, subdivaricately branched above; leaves remote, largely withered by anthesis, the petioles dilated at the base, 2—5 cm. long, the blades rounded-cordate, palmately lobed, the lobes incised into broadly oblong, obtuse lobules; inflorescence open, subpaniculate, the flowers on each branch rather few; bracts entire, lance-linear, 3-5 mm. long; pedicels stiff, slender, 2—6 cm. long; bracteoles opposite, ca. 3 mm. long, placed well below the flower; sepals deep blue, glabrous; upper sepals lance-ovate, 9 mm. long, 4 mm. wide, acute, spur straight, subcylindric, blue, horizontal, acute, 11-13 mm. long; lateral sepals oblong-ovate, 9 mm. by 5 mm., abruptly narrowed into an acute tip; lower sepals elliptic-obovate, 9 mm. by 5.5 mm., rounded apically; upper petals white with bluish tinge, 8—9 mm. long, the terminal lobules slightly more than 1 mm. long; lower petals bluish, white bearded, the lamina 7-8 mm. long, with divaricate lobes 3 mm. long and the claw 4.5 mm. long; stamens 5—6 mm. long, bluish, dilated below, glabrous; anthers 1 mm. long; follicles 3, glabrous, 10-12 mm. long, 2 mm. thick. Type: “In montibus Karatau inter Turkestan et Tschemkent,” A. Regel, May, 1876 (LE) seen; possible isotypes seen (BM, K, P). DISTRIBUTION. Turkestan. EXAMPLE. Vuorrili 43. 35. Delphinium macropogon Prokhanov, Not. Syst. ex Herb. Ko- marovii 21: 417-437. 1961. Fic. 4, G. Stem 2-10 (—14) dm. tall, 2-5 mm. thick, ridged, glabrous below, some- what strigulose in inflorescence, more or less zigzag, remotely leafy, simple or few branched in inflorescence; lower leaves long petioled (14-25 cm.), cauline petioles to ca. 1 dm., somewhat vaginate; blades suborbicular, 7-15 cm. broad, pubescent on both surfaces, paler beneath, palmately 5—7-fid to base into obovate segments narrow-cuneate below, more or less trifid above into lance-linear lobes 13-30 mm. long, 2~7 mm. wide; racemes 5—12 cm. long, dense, many flowered; bracts linear, 5-15 mm. long, white margined, somewhat pubescent on margins; pedicels 5-8 mm. long, strigose; brac- teoles linear, 3-6 mm. long, situated near middle of pedicel; sepals och- roleucous to whitish, pubescent; upper sepal 9 mm. by 5 mm., spur conic- cylindric, 10-12 mm. long; lateral sepals 7 mm. by 4 mm.., elliptic, rounded at apex; lower sepals ca. 8 mm. by 4 mm.; petals white, the upper laminae strongly oblique, 7 mm. long, glabrous, bidentate, the spur 10 mm. long; lower laminae round-ovate, 4.5 mm. in diameter, bearded and ciliate in upper parts, bifid almost to base into approximate broad segments, claw af mm. long; stamens 6-7 mm, long, glabrous; anthers pale, elongate; fol- licles 8-10 mm. long, glabrous to strigose; seeds obovoid, to 2 mm. long. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 287 Type: Daghestan: Kukurt-Tau, 2 km. from Mukku, Prokhanov 43, (LE) seen. No other material available. 36. Delphinium macrostachyum Boiss, ex Huth, Bot. Jahrb. 20: 441. 1895. Fic. 4, H. Stem erect, flexuous, 6-15 dm. high, 3-4 mm. thick, glabrous below, sometimes puberulous above; leaves remote, long petioled, the petioles broadly vaginate at the base; lower leaf blades to 3 dm. across, sub- glabrous, multifid, the lobes broadly lanceolate; upper leaves 3—5-parted, the parts subentire; inflorescence simple and racemose or openly few branched, 3—5 dm. long, densely and many flowered; bracts entire, lance- subulate, 4-6 mm. long; pedicels 6-10 mm. long; bracteoles linear, 2—3 mm. long, near the middle of the pedicel; sepals deep violet to bronze- black, glabrous, the upper sepal ovate, 9 mm. by 4 mm., obtuse, the spur obtuse, straight, ca. 10 mm. long, 3 mm. wide at the base; lateral sepals 10-11 mm. long, 2.5 mm. wide, subobtuse; lower sepals 9 mm. by 4 mm., rounded at apex; upper petals violet, glabrous, ca. 7 mm. long, the 2 lobes 1 mm. long, the spur at right angles to the lamina and ca. 9 mm. long; lower petals violet, the claw 3 mm. long, the lamina roundish, white bearded, the lobes 3 mm. deep; stamens 5-6 mm. long, glabrous; anthers dark, 0.8 mm. long; follicles 3, erect, glabrous, 10 mm. long, the styles another 3 mm. Lectotype: Turkey: Bakakri near Mardin, July 2, 1888, Sintenis 1266 (K); isotypes (P, PHIL); all seen. DistripuTion. Kurdistan from Turkey to Iran; on rocky limestone slopes at 1200-2000 m. SEEN. Iran: Furse 3404. Iraq: Wheeler Haines 513; Rami, Hoshen & Nuri 29481. 37. Delphinium micranthum Boiss. & Hohen. in Boiss. Diagn. I. 1: 68. 1842 Fre. 4, I. Stem angled, subglabrous to strigulose, slender, 3-5 dm. tall; petioles dilated-vaginate at base, the lower to 8 cm. long, the upper shorter; leaves 3—5-parted, then multifid into linear lobes ca. 1 mm. wide; racemes in loose panicles, but only 5—7-flowered; bracts linear (the lower trifid), 5-15 mm. long; pedicels 5-20 mm. long, bibracteolate just below the flower or lower, the bracteoles linear, 4-5 mm. long; sepals purple-blue, 7-10 mm. long, strigulose; spur straight, acute, 10-15 mm. long; upper petals glabrous, bluish, the blade about 7 mm. long, the spur 10-12 mm.; lower petals with the limb ovate, bifid to the middle and almost 6 mm. long; stamens somewhat bluish, 5-6 mm. long; anthers 0.6 mm. long; follicles 3, glabrous or glandular-pubescent, 6-8 mm. long, the styles an additional 1.5—2 mm, Type: Iraq: Mt. Gara, Kurdistan, Kotschy 610, in 1843 (G), not seen. I have seen Kotschy 410, July 1841 (P) from same region. 288 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 DIsTRIBUTION. At elevations below 2800 m., Iraqui Kurdistan. Examp.es. J. B. Gillett 9501, 11206, 12473; Gillett & Rawi 7404, 7223; Guest & Ludlow-Hewitt 2691; Lazar 118; Al-Rawi 24731; Wheeler Haines 1320; Guest 2916. 38. Delphinium nuristanicum Tamura, Acta Phytotax. ee 16: 141. 1956. Fic. 4, J. Stem 4-8 dm. tall, 3 or more mm. thick, simple or branched in inflores- cence, sparsely white-hispidulous below, glabrous above the very base; radical and lower cauline leaves round-reniform in outline, 3—5 cm. long, 4—6 cm. wide, 3-parted almost to base, the lateral segments bipartite, all 5 parts incised-laciniate into lance-linear, obtuse lobes white-hispid on margins and lower surface; petioles 6-10 cm. long; upper cauline leaves glabrous, tripartite almost to base into linear-oblong segments 2—4 cm. long, 2.5-5 mm. wide, short petioled; raceme simple or branched below, 1-2 dm. long, few to many flowered; lower bracts to 18 mm. by 1.8 mm., petioled, the upper linear-filiform, 3-8 mm. long; pedicels 5-12 mm. long, more or less divaricate, elongating after anthesis, with some spreading, yellowish, glandular hairs near the summit; bracteoles near middle of pedicel, ca. 3 mm. long; sepals blue with pale median band, glabrous; upper sepals 12-13 mm. long, 4 mm. wide, obtuse, the spur horizontal with decurved tip, 12 mm. long, 2.5—3 mm. wide at base; lateral sepals 12 mm. by 4 mm., rounded at apex; lower sepals 10-12 mm. by 4-5 mm., oblong with rounded apex; upper petals whitish, the lamina strongly oblique, ca. 10 mm. long, glabrous, deeply bidentate, the spur ca. 12 mm. long; lower petals bluish, the oblong-obovate lamina ca. 6 mm. by 4 mm., bearded, cleft into 2 lanceolate divaricate lobes ca. 3 mm. long, the claw 6 mm. long; stamens 7-8 mm. long, glabrous, the roundish anthers 1 mm. long; carpels 3, said to be pilose in the type, glabrous in specimen cited below. Type: Afghanistan: Nuristan: Chatras at 2100 m., Kitamura, Aug. 2, 1955 (KYO), not seen. ILLUSTRATION. Kitamura, Fl. Afghan. 126. fig. 50. 1960. SPECIMEN SEEN. West Pakistan: Chitral, W. E. Brown 804 (kK). Resembling D. penicillatum of southern Iran, but that species is viscid- pubescent, has shorter bracts and pedicels, smaller sepals. 39. Delphinium ochroleucum Stev. ex. DC. Syst. 1: 353, 546. 1817. Fic. 4, K. soya hybridum var. ochroleucum Boiss. Fl. Orient. 1: 89. 1867, as to Dehn hybridum ssp. ochroleucum Busch in Fl. Cauc. Crit. III. 3: 54. Deiphinium ochroleucum var. sulphureum Regel ex Huth, Bot. Jahrb. 20: 2. 1895. Not typified. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM — 289 Stem slender, mostly 3—7 dm. tall, sparingly strigulose below, almost glabrous in inflorescence; lower leaves mostly withered by anthesis, long petioled, the petioles broad and vaginate at base, cauline leaves crowded, short-petioled, largely 2—4 cm. in diameter, 3—5-parted, then again divided into many narrowly linear lobes 0.5-1.5 mm. broad, strigulose; raceme generally simple, dense, many flowered, mostly 1-2 dm. long; lower bracts trifid, upper entire, linear, 8-15 mm. long; pedicels shorter than flowers, bibracteolate in middle; sepals ochroleucous, glabrous, the upper one ca. 7 mm. long, ovate, acute, its spur erect, 10-11 mm. long, 2 mm. wide at base; lateral sepals 5.6 mm. by 2 mm., more or less oblong; lower sepals 7 mm. by 3 mm.; upper petals with bluish tinge, glabrous, 7 mm. long, bidentate, the spur 10 mm. long; lower petals with blade ca. 5 mm. long, the 2 bearded lobes ca. 3.5 mm., the claw serrulate; stamens 4-5 mm. long, glabrous; anthers yellow, 0.8 mm. long; follicles 3, erect, 12-15 mm. long, 3 mm. thick, glabrous, the style an additional 2.5—3 mm. Type: “Hab. in Iberia,” now East Georgia (G) not seen. I have, how- ever, seen a photograph of the Leningrad specimen “Iberia, comm. Steven”’ (Kk). DistrIBUTION. Apparently largely below 1000 m., mountains of Cau- casus and Transcaucasia. EXAMPLES SEEN. Kolenati 1590; Hohenacker near Schuscha; Grossheim 37; Woronow 990; Davis 33907, 33853; Capus 50, 51, 53. 40. Delphinium pallasii Nevski in Komarov, Fl. U.S.S.R. 7: 17d. 424, 1937. Fic. 4, L. Delphinium tauricum Pallas, Ind. taur. in Neueste Nordische Beitrage 3: 432. 1796; nomen. Delphinium hybridum Willd. var. debilius Ledeb. Fl. Rossica 1: 61. 1841. Based on Bieber, Tauria. Not seen. Stem 3.5-11.5 dm. tall, simple, densely strigulose throughout, more or less leafy; petioles strongly dilated at base, the lower 1—2 dm. long, the upper 1-3 cm.; blades of leaves 3-parted almost to base into cuneate- rhombic segments, these divided deeply into few divaricate, more or less falcate, linear-lanceolate lobes 1-3 cm. long, 1.5—4 mm. wide, somewhat pubescent especially beneath; raceme erect, dense, many flowered, 1-3 dm. long; bracts linear, the lower rarely 2-3-parted, 7-15 mm. long; pedicels suberect, 8-12 mm. long, strigulose; bracteoles at about the middle of the pedicel, linear-subulate, 3-6 mm. long; sepals lilac, obtuse, somewhat strigose, the upper oblong-ovate, ca. 12 mm. by 6 mm., the spur spreading, slightly curved, 16-20 mm. long; lateral sepals oblong- elliptic, 12 mm. by 6 mm., rounded-truncate; the lower 11 mm. by 5 mm., rounded-truncate; petals pale, the upper lamina almost straight, 10 mm. by 2 mm., glabrous, bidentate, the spur ca. 18 mm. long; lower laminae 6.5 mm. by 5 mm., bifid to near the base into acute lanceolate lobes, long- ciliate, the claw ca. 6 mm. long; stamens 5—6 mm. long, glabrous; anthers round-oblong, 0.8 mm. long; follicles 3, pubescent, 290 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Type: Tauria (in Azerbaydzhan?), Herb. Pallas (LE); seen. No other material seen. 41. Delphinium pallidiflorum Freyn, Bull. Herb. Boiss. II. 1: 254. 1901. Fic. 4, M. Delphinium cyphoplectrum Boiss. var. pallidiforum (Freyn) Davis, Notes o . Edinb. 26: 171. 1965. Delphinium nachiczevanicum Tzvelev, Not. Syst. Leningrad 19: 3. 1959. Based on Egorova, Tzevelev & Czerepanov 418 (LE) from prov. Nachie- zevan, Azerbaidshan. Seen. Stem terete, zigzag, 3-8 dm. tall, simple or sparsely branched, retrorse- strigulose below, with longer spreading glanduliferous hairs below the in- florescence; leaves few, short strigulose, the long petioles vaginate at the base, the blades round in outline, 5—6 cm. in diameter, palmately many parted, the lower into lanceolate, the upper into ultimate lance-linear laciniae 1-3 mm. wide; raceme short to long, lax, with spreading more or less glandular hairs; bracts entire, linear to linear-lanceolate, 3-7 mm. long; pedicels glabrous or pubescent, 6-20 mm. long; bracteoles below middle of pedicel, 2-4 mm. long; sepals violet-blue, bluish or light lilac, subglabrous, 11-13 mm. long, the spur 10-13 mm. long, blunt; petals pale, the upper glabrous, with narrowly lanceolate, oblique, bilobed blade, the spur ca. 10 mm. long; lower blades long-pilose, 5—6 mm. long, 3 mm wide, deeply cleft into parallel lobes, the claw 5-6 mm. long; filaments pale lilac, glabrous; anthers yellow, minute; follicles 3, glabrous. Type: Turkey, Armenia, Wan, Warack-Dagh, 2700 m., June 27, 1899, A. Kronenburg (G?), not seen. DistRIBUTION. Steppes and rocky places, 1900-2400 m., Turkish and Russian Armenia, northern Iraq and northwestern Iran. AMPLES. Transcaucasia: Heideman, May 27, 1932. Russian Armenia: Besser. Iraq: Gillett 11782; Rawi 12213. Iran: Koelz 18015. a Ps hinium penicillatum Boiss. Ann. Sci, Nat. II. 16: 369-370. 41. Plant with short, viscid pubescence throughout; stem angled-striate, 2-3 dm. tall; leaves petioled, tripartite, the segments bipinnatifid into short linear setaceous lobes scarcely 1 mm. wide; lower blades roundish, to 5 cm. broad, upper smaller; racemes spiciform, strict, open, to 15- flowered, 5-12 cm. long; bracts deltoid-acute, 2 mm. long; pedicels 3.5—4 mm. long, bibracteolate below the middle; bracteoles linear, pubescent, ca. 2 mm. long; sepals hairy, oblong, pale blue, the upper sepal ovate, 6 mm. by 3 mm., subacute, spur straight, 9-9.5 mm. long, 2 mm. wide at base, gibbous at tip; lateral sepals oblance-spatulate, 6 mm. by 3 mm., pubescent along the middle line; lower sepals slightly wider, more pointed and pubescent; upper petals glabrous, the blades obliquely truncate, slightly divided, the lamina ca. 8 mm., the spur 10 mm. long; lower laminae ovate, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 291 bearded, penicillate, ca. 5 mm. by 3 mm., deeply divided into 2 divaricate pointed lobes, claw ca. 4 mm. long; stamens 4-5 mm. long, ciliate; anthers yellow, 0.6 mm. long; follicles 2-3, subglabrous. Type: Aucher 4034 from Mascate (Muscat?), Arabia?, (Pp); isotype (K); both seen. Rance. Apparently in arid places, Oman and southern Iran. EXAMPLE SEEN. Iran: Kotschy 393a. 43. Delphinium puniceum Pallas, Reise 3: 327. 1776. Fic. 4, O. Delphinium hybridum Willd. var. puniceum Claus, Ind. des in Gobel It. 2: : i Delphinium hybridum subsp. puniceum N. Busch in Pe Cauc Crit. Te. 3: 55. 1903. Plant somewhat strigulose throughout; stem 3-8 dm. tall, slender, terete, simple or few branched above; leaves few, the lower long petioled, the upper with shorter petioles; petioles dilated at base; blades 3-15 cm. in diameter, 3—5-parted to base, the parts then divided into linear strigu- lose lobes mostly 1-3(-6) mm. wide and with 1-2 divergent lobules; racemes dense, many flowered, 5-30 cm. long, densely, sometimes loosely strigulose; bracts entire, lance-linear, 4-6 mm. long; pedicels ascending, 7-12 mm. long; bracteoles 3-4 mm. long, mostly near the base of the pedicel; sepals deep bluish-purple, strigulose; upper sepal ovate-lanceolate, ca. 12 mm. by 3 mm., acute, the spur straight, obtuse, 11-12 mm. long, 3 mm. wide at the base; lateral sepals lanceolate, 12 mm. by 2 mm., subob- tuse, slightly narrowed at base; lower sepals lance-oblong, 13 mm. by 4 mm., acutish; upper petal lamina purple, ca. 9 mm. long, the terminal lobules 1.5 mm. long, spur 10-11 mm. long; lower petals purple, the blade 6 mm. long, white bearded, the lobes lance-oblong, 4 mm. long; claw 6 mm.; stamens 6-7 mm. long, glabrous; anthers dark, 0.8 mm. long; follicles 3, erect, 10 mm. by 3 mm., more or less reticulate and strigulose, the styles 2-2.5 mm. long; seeds obscurely angled, covered with scales. Type: near Salinas Tschaptschatschi near “Volgam australem” in Sibiria, Pallas misit 1781, (LE), seen. Photographs (£, K), seen. : Rance. Region of Volga River to Georgia and Asia Media (Pawlowski, Fragm. Fl. & Geobot. 9: 437. (1963). ExampLes. Georgia: Hohenacker in 1834; Radde in 1870. oiss. & Hausskn. in Boiss. Fl. Orient. Fic. 4, P. pallidiflora Busch in Fl. Cauc. 44. Delphinium quercetorum B Suppl. 20. 1888 Delphinium hybridum ssp. laxiusculum vat. Crit. III. 3: .57.., 1903. Stem to 3 mm. thick, glandular-pubescent in upper half, 3-4 dm. tall, mostly simple; lower leaves suborbicular, 1.5-3 cm. in diameter, strig- ulose, palmatisect into short divergent linear or oblong lobes; petioles 292 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 dilate-vaginate at base; raceme 7-12 cm. long, lax, 7—12-flowered, glan- dular-pubescent; bracts lanceolate, membranous, 4-7 mm. long, entire; long; sepals pale bluish, somewhat pubescent; upper sepal lance-ovate, ca. by 3 mm., obtuse, the spur somewhat inflated at apex, 12-15 mm long, 2.5 mm. wide at base, straight, horizontal; lateral sepals oblong- 2 short apical lobes, hirsute, the thick spur 11-12 mm. long; lower laminae 5.5 mm. long, heavily bearded, the linear-lanceolate lobes 5 mm. long, claw 5.5 mm. long; stamens glabrous, 5-6 mm. long; anthers bluish yellow, rounded, 0.8 mm. long; carpels 3, pilose their entire length. Type: Mt. Pir Omar Gudrun, Persian Kurdistan, at 4000-5000’, June 1867, Haussknecht (Gc) not seen; isotypes (BM, K) seen. DiIsTRIBUTION. At 600-2000 m., Iraq, Iran. EXAMPLES. Iraq: Rawi 23431, 22706, 22200; Kass & Nuri 27607. Iran: Merton 3935; Wheeler Haines 1712. There is some question as to the distinctness of this species, since there is considerable variation in the specimens. It is near D. cyphoplectrum. 45. Delphinium roylei Munz, nomen novum. Fic. 4, Q. Delphinium incanum Royle, Ilustr. Bot. Himal. 55. 1834, not E. D. Clarke, 1812. Stem 5-10 dm. tall, simple or usually few branched above, terete, 3-5 mm. thick, closely and finely strigulose, so as to be more or less ashy, espe- cially above; lower leaves withered by anthesis, 5-8 cm. in diameter, palmately multipartite into cuneate, deeply and sharply incised segments, the ultimate lobes sharp pointed, 1.5-3 mm, wide, greenish above, paler and closely strigulose beneath, the petioles to 1 dm. long, vaginate at base; middle cauline leaves remote, short petioled to subsessile; principal raceme commonly 1-2 dm. long, rather densely many flowered, the lateral racemes fewer and more loosely flowered; bracts linear, acuminate, 5—10 mm. long; pedicels commonly 1—2.5 cm. long, recurved at apex; bracteoles 2.5-3 mm. long, located near base of pedicel; sepals deep blue, densely strigulose; upper sepal 13-14 mm. by 8-9 mm., pointed, the spur hori- zontal, cylindric, straight, somewhat pointed, 15-16 mm. long, 3—3.5 mm. wide at base; lateral sepals broadly elliptic-ovate, obtuse, 15 mm. by 10 mm., strigulose on midline; lower sepals 15-16 mm. by 8-9 mm., subobo- vate, pointed; upper petals pale, the oblique glabrous lamina 8 mm. long, shallowly 2-toothed, spur ca. 14 mm, long; lower laminae deep blue, 6 mm. long, bearded, roundish, lobed for ca. 3 mm., the claw 5 mm. long; stamens 5-7 mm. long, subglabrous; anthers blue, 1 mm. long; follicles 3, strigose, 10-15 mm. long, 3-4 mm. wide; seeds ca. 1 mm. long, 3-angled, with ca. 5 transverse rows of scales, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 293 Type: Kashmir: Purbunee, R. Jnglis (L1v?); not seen. Rance. Grassy meadows and hillsides, 2000-5000 m., Kashmir. ILLUSTRATIONS. BLATTER, Beautiful Fl. Kashmir 1: pl. 4. 1927; Cov- ENTRY, Wild FI. Kashmir 3: pl. 6. 1930. Examp.es. A few of the many specimens, apparently all from Kashmir or environs are: R. R. Stewart 6530, 5694, 3156, 10546, 19586, 19693; Duthie 13440, 14091; Falconer 61; Dickason 365, 372; Koelz 9261; Lace 467; Ludlow 868, 8213; Ludlow & Sherriff 7560, 8238, 9376; Jacquemont 673/791; Schlagint- weit 3066, 5171. The earlier use of the name Delphinium incanum by E. D. Clarke, Travels 2: 451. 1812, for a species from Cana, Israel, now referred to the genus Consolida, makes it necessary to give up Royle’s later Delphinium incanum. 46. Delphinium saniculifolium Boiss. Diagn. I. 6: 6. 1845. Fic. 4, R. ? Delphinium saniculifolium Boiss. var. refracta Brihl & King, Ann. Bot. Gard. Calc. 5: 95. 1896. Not typified; no specimens cited. Stem erect, terete, flexuous, mostly openly branched above, 3-8 dm. tall, glabrous or sparsely hairy (especially below); lower leaves with petioles 2-8 cm. long and more or less vaginate at the base, the blades palmately 5-lobed into broad cuneate divisions with few coarse terminal teeth, the laminae 2-5 cm. broad, glabrous or somewhat strigulose; cauline leaves few, smaller, palmately parted to the base; racemes lax, several flowered, in an open panicle, peduncled; bracts lance-linear, mostly 3-5 mm. long; pedicels divaricate, very slender, 5-25 mm. long; bracteoles 2 (3), ca. 2 mm. long, inserted near the base or the middle of the pedicel; sepals blue, subglabrous, the upper ovate, 7-8 mm. long, obtuse, the spur 7-9 mm. long, subcylindric, obtuse; lateral sepals elliptic-oblong, rounded at apex, 7 mm. long; lower elliptic, rounded at apex, 7 mm. long; upper petals bluish, the laminae oblique, glabrous, 7 mm. long, more or less bidentate, the spur ca. 9 mm. long; lower petals blue, the blades 4-5 mm. long, sub- ovate, bearded, lobed for 2-3 mm., the claw 4 mm. long; stamens 4-5 mm, long, somewhat bluish, glabrous; anthers 0.6 mm. long; follicles 3, largely glabrous, 7-10 mm. by 3 mm., the styles ca. 2 mm. in addition; seeds scarcely 1 mm. long, obpyramidal, with several rows of scales. Type: “In alpe Kuh-Delu Persiae australis,” Kotschy 536, (c?), not seen; isotypes (A, BM, K, MO, UC) seen, but all in rather poor condition. DistRIBUTION. Southern Iran at 900-2400 m. Examp.es, Rechinger 3364; Haussknecht, July, 1868; Stapf 2300. 47. Delphinium schmalhausenii Alboff, Trav. Soc. Hort. Odessa (Descr. nouv. esp. pl.) 1891; Huth, Bot. Jahrb. 20: 441. 1895. : Fic. 5, A. Delphinium hybridum Willd. subsp. schmalhausenii (Alb.) Busch in Fl. Cauc. Grits: 3:53: 294 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Delphinium hybridum Steph. ex Willd. Sp. Pl. 2: 1229. 1799, not L., 1756. Delphinium freynii Conrath ex Freyn, Bull. Herb. Boiss. 3: 36. 1895. Based on Conrath, 1889, Somchetia, Lok. 2300 m. (G); not seen. Delphinium somcheticum Conrath & Freyn, Bull. Herb. Boiss. 3: 37. 1895. Based on Conrath, 1889, Somchetia, Tschatach (G); not seen. Delphinium halophilum Huth, Bot. Jahrb. 20: 487. 1895. New name for D. freynii Conrath. Stem single, 2-10 dm. tall, more or less zigzag, striate and usually gla- brous below, glandular-pubescent above; lower leaves gone at time of anthesis; cauline rather long petioled (1-2 dm.) and petioles with broad base, upper petioles almost lacking; blades to 2 dm. broad, palmately multifid, then pinnately dissected into numerous very narrow laciniae; raceme solitary, long-cylindrical, rather dense; lower bracts foliaceous, laciniate, the upper trifid or entire; pedicels 5-10(—18) mm. long, glan- dular pubescent; bracteoles linear in upper half of pedicels; sepals dark blue, strigulose, the upper sepal linear-elliptic, obtuse, 9-10 mm. long, 3.5 mm. broad, spur ascending, straight, 15-20 mm. long, 3 mm. wide at base, acute; upper petals blue, glabrous, ca. 8-10 mm. high, short lobed, the spur 15—16 mm. long; lower petals with roundish hairy lamina 5 mm. long, rather deeply bilobed, claw 5 mm. long; filaments yellowish; anthers blue; follicles mostly glabrous, 10-12 mm. long; seeds with numerous mem- branous scales, Type: Transcaucasia, Alboff 190, Aug., 1893. Not seen. DIsTRIBUTION. At 2000 m. or above, northern Turkey, Transcaucasia. Examples. Turkey: Prov. Kars, Davis 32512. Transcaucasia: Radde 428; Radde, June 16, 1867; Szowitz 391; Hohenacker in 1834. Davis, Fl. Turkey 1: 112. 1965, unites the above three proposed species (Delphinium schmalhausenii, D. freynii, and D. somcheticum) into one. The variation in size of flowers, character of pubescence (whether glan- dular or not), and in division of bracts does not correspond with any dis- tinction in geographical distribution. 48, ea semibarbatum Bienert ex Boiss., Fl. Orient. 1: 92. 1867. 5, &. ? ? Delphinium semibarbatum var. hoeltzeri Huth, Bot. Jahrb. 20: 424. 1895. Based on a Hoeltzer collection from Persia (LE); not seen. se ce hybridum var. sulphureum Regel, Acta Horti Petrop. 5: 226. has Delphinium zalil Aitch. & Hemsl. Trans. Linn. Soc. II. 3: 30. 1888-1894. Based on Aitchison 671, Khorasan, Iran; isotypes seen (BM, GH). Delphinium biternatum Huth var. leiocarpum Freyn & Sint. ex Freyn, Bull. Herb. Boiss. II. 3: 561. 1903. Aschabad, Sintenis 383; isotypes seen (BH, BM, E, K, MO, US). Stem erect, 2-8 dm. tall, subglabrous to retrorse-strigulose, simple or few branched above; lower petioles to 1 dm. or longer, scarcely dilated at base, the cauline leaves several, gradually shorter petioled upward, the 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 295 blades 3-parted, then again ternate, ultimately divided into long linear lobes mostly 1-1.5(—2) mm. wide, the blades 2-10 cm. in diameter, sub- glabrous to somewhat strigulose; inflorescence with rather open, rather many flowered racemes 1—3 dm. long, often with smaller supplementary racemes as side branches; bracts mostly lance-subulate, 3-8 mm. long; pedicels 2-10 mm. long, stout, more or less glandular pubescent; brac- teoles lance-subulate, at about the middle of the pedicel, 2-4 mm. long; flowers yellow or with a greenish tinge; sepals glabrous, broadly ovate, obtuse, 9-10 mm. long, the spur subcylindric, yellow, 7-10 mm. long; upper petals with the laminae ca. 10 mm. long, 3 mm. wide, attenuate into a bidentate minutely puberulent tip; spurs ca. 8 mm. long; lower laminae almost as long, bifid, yellow bearded, subserrate on the claw; stamens yellow, 4-6 mm. long, glabrous; anthers 1 mm. long; follicles 3, glabrous, suberect, 10 mm. by 3.5 mm., the styles 2 mm.; seeds quadrate, trans- versely fimbriate-lamellate. Type: “In montosis prov. Khorassan Persicae orient. ad orientem urbis Mesched” (Bunge) (c), not seen. Isotype seen (P). DISTRIBUTION. Mountain meadows and grassy plains to 2000 m., Transcaspia to Turkestan and Iran. ILLustRATIONS. Trans. Linn. Soc. II. 3: pl. 3. 1888-1894; Bot. Mag. pl. 7049. 1889. EXAMPLES. Transcaspia: Litvinov 443; F. N. Meyer 626. Turkestan: Regel, May 17, 1880; Lipsky 1706; Sovetkina & Czausova 4267. Iran: Rechinger 1698, 4934; Merton 3888. 49. Delphinium semiclavatum Nevski in Komarov, FL: US oa. 7 168, 726. 1937. |) (a a Or Stem 5.5-7.5 dm. high, few branched with long virgate branches, densely retrorse-pubescent below, leafy at the base, the petioles 5—10 cm. long, dilated at their base, subglabrous; leaf blades suborbicular, almost glabrous, 3-parted to base, the segments sessile, the middle one obovate, cuneate at base, shallowly incised into 3-5 obtuse, oblong, rounded lobes 2-3 mm. broad, the lateral segments unequally incised into further divi- sions; raceme lax, few flowered, long and with a glabrous axis; pedicels glabrous, to 1.7 cm. long; bracts narrowly lance-linear, glabrous, 2-3 mm. long; bracteoles 1.5-2 mm. long, linear-subulate, inserted above the middle of the pedicels; sepals sordid lilac, glabrous, obtuse, 9-10 mm. long, 4.5 mm. wide; spur straight or nearly so, obtuse, gibbous below the tip, 10-12 mm. long, to 2.7 mm. wide at base; upper petals whitish, the laminae glabrous, oblique, bidentate, ca. 8 mm. long, the spur 9 mm. long; lower laminae round-ovate, 3 mm. long, bifid half its length, sparsely bearded, the claw 4 mm. long; follicles glabrous, 8 mm. long. Type: Asia Media: Pamiro-Alaj occidentalis, Sangardak, June 20, Lipsky 1887 (LE); seen. ae Near Delphinium saniculifolium, but flowers sordid lilac, not blue, spur longer; pedicels shorter, petals subglabrous. 1896, 296 JOURNAL OF THE ARNOLD ARBORETUM [voL, 48 FicureE 5. Delphinium, Group IV. Petals, the upper spurred, the lower ; lower Sitahcnatiin D. schmathausenii; upper lamina 8 mm. long. spur 15 mm.: lowe pes - long, claw 5 mm.; wn from Hohenacker in 1834 (BM). D semibarbatum ; upper lamina 10 mm. long, spur 8 mm.; er lamina 5.5 mm long, claw 9.5 mm.; drawn from von Knorring at Samarkand (cH). Dp; semiclavatum ; upper lamina 8 mm. long, spur 9 mm.; lower lamina 3 mm. long, wna heen sk Si Lipsky type (LE). D. D. suave; upper lamin €; u a 8 mm. an sate beDesed desis) lamina 6 mm. long, claw 4 mm.; drawn from Aitchison ] : og hho (sm). E. D. Szowitsianum ; upper lamina 7 mm. long, spur 1 } Ower lamina 4 mm. long, claw 4 mm.; drawn from Szowitz (GH). F. D. tetan- ng, spur 11.5 mm.; lower la As, Pichler June 20, 1882 (type material of D. caerulescens, wu). I. D. turk- ener ee lamina 11 im. long, spur 21 mm.; lower lamina 5 mm. long, claw 6.5 mm.; drawn from Sintenis 1672 (mo). J. D. uncinatum; upper lamina 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 297 50. Delphinium suave Huth, Bot. Jahrb. 20: 470. 1895. Fic. 5, D. Delphinium penicillatum Boiss. var. macroplectrum Buser in Boiss. Fl. Orient. Suppl. 19. 1888. Based on Aitchison 862. Delphinium saniculaefolium Boiss. var. kurramensis P. Brihl in Brihl & King, Ann. Bot. Gard. Calc. 5: 95. 1896. New name for D. penicillatum var. macroplectrum. Stem or stems slender, erect, simple or few branched, terete, more or less strigulose throughout or more loosely so in the inflorescence, 1.5—4.5 dm. tall; lower leaves apparently trifid into broad cuneate segments that are somewhat toothed or lobed, the laminae 1.5—2.5 cm. broad; petioles slender, 2-5 cm. long, vaginate at base; cauline leaves remote, short petioled, the blades 2—3 cm. long, with few lobes 1-3 or more mm. wide; inflorescence simple, racemose, several flowered, bracts entire, lance-linear, 3-8 mm. long; pedicels ascending, 5-15 mm. long; bracteoles lance- linear, 1.5—2 mm. long, above the middle of the pedicel; sepals pale blue, slightly strigulose, the upper about 7 mm. by 3.5 mm., the spur 10-15 mm. long, spreading, straight, pointed; lateral sepals round-obovate, pubescent on midrib, 7 mm. long, 5 mm. wide; lower sepals 8 mm. by 6 mm.; upper petals glabrous, the blade oblique, dilute yellow or bluish, 9-11 mm. long, with terminal lobes ca. 1 mm. long, spur 12 mm. long; lower petals with an oblong or roundish-oblong, sparsely bearded, blade divided to 2.5-3.5 mm., the claw 4 mm. long; stamens 4.5—5 mm. long, glabrous; anthers dark, 1 mm. long; follicles 3, strigulose. Type: Afghanistan, 2600-3600 m., Kurrum Valley or “in pratis vallis Alikel,” Aitchison 862; holotype (LE) and isotypes (BM, GH, P, PH), seen. DistrrBuTion. Rocky slopes, meadows, and the like, 2600-3800 m., Afghanistan, West Pakistan. ILLUSTRATION. Brix & Kine, Ann. Bot. Gard. Calc. 5: pl. 11 7, he. J: 1896. REPRESENTATIVE COLLECTIONS. Afghanistan: mt. above Sari Casma, Bacon 73. W. Pakistan: Chitral, Harriss 15837. Kohat, Burtt & Nairn 586. Punjab, Ram Bahsli 14811. Beluchistan, R. R. Stewart 579; Appleton 3862. 51. Delphinium szowitsianum Boiss. Fl, Orient. 1: 89. 1867. Fic. 5, E. Delphinium hybridum Steph. ex Willd. var. szowitsiana Trautv. Acta Horti Petrop. II. 2: 493. 1873. Delphinium hybridum var. dasyanthum Avé Lallem. Ind. Horti Petrop. 1842, Huth. Delphinium dasystachyum Boiss. & Hohen. ssp. szowitzianum N. Busch in FI. Cauc. Crit. III. 3: 58. 1903. Delphinium tricolor Bernh, ex Steud. Nomencl. ed. 2. 1: 489. 8 mm. long, spur 5.5 mm.; lower lamina 5 mm. long, claw 4 mm.; drawn from Stewart 12688 (Ny). K. D. ursinum; upper lamina 7 mm. long, spur 9 mm.; lower lamina 4 mm. long, claw 4 mm.; drawn from Rechinger 6083 ISOTYPE (us). . D. vanense; upper lamina 11 mm. long, spur 11 mm.; lower lamina 8 mm. long, claw 4 mm.; drawn from Davis 22735 (BM). 1840: nomen. 298 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Stem simple, 3-9 dm. tall, closely strigulose or subglabrous below, densely strigulose and pubescent above, more or less angled and sulcate; leaves largely dead by anthesis, the cauline scattered, with petioles dilated at the base, short in upper leaves; blades largely 2-4 cm. in diameter, ternate, then laciniate into many strigose or pubescent linear lobes, 0.5- 1(—2) mm. wide; inflorescence very dense, many flowered, to ca. 1 dm. long; lower bracts trifid, from a broad membranous base, upper mostly less than 5 mm, long, densely pubescent or strigose; pedicels generally less than 5 mm. long, the 2 lance-linear bracteoles 5-6 mm. long; sepals densely short-hairy, yellowish with greenish tips, the upper sepal ovate, 8 mm. long, spur short-hairy, suberect, 12-13 mm. long, 3-4 mm. wide at base; other sepals about 8 mm. by 4 mm.; upper petals with bluish tinge, ca. 7 mm. long, emarginate, very short-puberulent near base, the spurs ca. 11 mm. long; lower petals 7.5-8 mm. long, bearded, with rounded lobes 2 mm. long; stamens glabrous, 5—6 mm. long; anthers yellow, 1 mm. long; follicles pubescent. Type: Russian Armenia near Nachitchewan, Szowitz, holotype (G?) not seen; isotype (K) seen, has “Szovits 439.” DistriBuTion. At 1800 m. or thereabouts, Transcaucasia, Russian Armenia. ILLustRATION. Belg. Hort. 22: p. 48, pl. 3. 1872. SPECIMENS SEEN. Russian Armenia: Szovits (GH, Pp, us); Yaroshenka & Ubanove, July 23, 1945. Transcaucasia: Hohenacker, July, 1838; Heideman & Aliev, July 27, 1932. 52. Delphinium tetanoplectrum Rech. f. Anz. Osterr, Akad. Wiss. Math.-Nat. 91: 72. 1954. Fic. 5, F. Stem 2-5 dm. tall, strict, erect, with short, more or less deflexed-ap- pressed, non-glandular pubescence; lower and middle leaves with petioles 3—6 cm. long, membranous-dilated at base, the blades ashy-pubescent, round-cordate in outline, palmately lobed or incised, 1-2 cm. long, the divisions relatively simple; inflorescence very lax, few flowered, the axis with short crisped, mostly eglandular hairs; bracts narrowly linear-lance- olate, 4-6 mm. long; pedicels 1-2 cm. long; bracteoles inserted in upper third of pedicel, narrowly linear, acute, 3-4 mm. long; sepals pale purplish blue, ca. 8 mm. long, to 5 mm. wide, with short scattered hairs, the upper sepal ovate, the others elliptic-obovate: spur 15-20 mm. long, ca. 2 mm. wide when pressed, straight, acute, more densely pubescent than sepals; petals palish, the upper almost straight, the laminae oblong, glabrous, emarginate, ca. 12 mm. long, the spurs slender, 15-20 mm. long; lower petals with blade 5.5 mm. long and wide, round-obovate, sparsely bearded, shallowly bilobed, the claw 5 mm. long; stamens ca. 5 mm. long, glabrous; anthers pale, oblong, 1 mm. long; carpels 3, coarsely pubescent. Type: Afghanistan, Lorinj, 2460 m., Aug. 27, 1939, W. Koelz 13725 (us). The type and the other three collections were on loan to Dr. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 299 Rechinger, in Vienna, and I am grateful to him and Dr. Riedl for kindly sending them on to me. CoLLECTIONS SEEN, Afghanistan: Gurzan, 2600 m., Koelz 14100; Baligali Pass, 2800 m., Koelz 13853; Sufedsang, 3070 m., Koelz 13986. The species keys out with D. micranthum, of Kurdistan, but has the lower petal blade much wider and less hairy. 53. Delphinium tuberosum Auch. ex Boiss. Ann. Sci. Nat. IE. 16: 370. iG. 5, Fi. Delphinium tuberosum var. leiocalycinum Bornm. Beih. Bot. Centr. 19: 201. 1905. Type, Strauss & Bornmiiller, Mt. Latétar; not seen. Delphinium caerulescens Freyn ex Stapf, Denkschr. Akad. Wien Math. Na 25; not Paxton, 1849. Based on a type from Jalpan (Djal- pan), northern Iran (wu); seen. At the University of Vienna there is a series of four sheets from Mt. Elwend, each with “Legit Th. Pichler, 1882” and also “Iter Persicum, Dris. J. E. Polak.” Of these, two have an addi- tional label “Jalpan, 6-20-1882.” Some of these specimens have the upper petals glabrous, others with a few or several stiff hairs. All have the same general flower color and form and the spur gibbous at the slightly curved apex. There seems to be the same general variation as I have found in other D. caerulescens. Stem erect, flexuous or almost zigzag, 3-5 dm. tall, terete, slender, few branched, closely strigulose, sometimes apparently with a zone of gland- tipped hairs near the middle; leaves often largely gone by flowering time, the lower long petioled, the petioles long-vaginate at the base, middle cauline leaves remote, short petioled; blades 2-4 cm. wide, palmately parted into numerous mostly linear lobes, 1-2.5 mm. wide, strigulose ; terminal racemes to 2 dm. long, usually with not more than 15 flowers, rather lax, the lateral racemes smaller, if present; bracts linear, 3-25 mm. long; pedicels 5-15 mm. long, ascending; bracteoles 1.5-5 mm. long, placed near or above the middle of the pedicel; sepals bluish or smoke- gray; upper sepal 8-10 mm. long, 4.5 mm, wide, pointed, mostly strigulose, the spur cylindric, horizontal, almost straight, bluish, gibbous at the curved apex, 12-15 mm. long; lateral sepals oblong-elliptical, 8-10 mm. long, 4-4.5 mm. wide, obtuse, strigose on midline; lower sepals 8.510 mm. long by 4 mm., oblong-elliptic, rounded at apex; upper petals oblique, the blade 8.5-10 mm. long, from glabrous to having a few stiff hairs on the surface or near the shallowly 2-lobed tip; spur 12-14 mm. long; lower petals heavily long bearded, the claw 4.5-5 mm. long, serrulate on one edge, the blade 5—6 mm. long, with 2 rather divergent pointed lobes ca. 4 mm. long; stamens 5-6 mm. long, glabrous; anthers whitish to bluish, 0.8-1 mm. long; follicles 3, ashy-strigulose to subglabrous, veiny, 8-12 mm. long, the style an additional 2.5 mm. long; seeds umbilicate, 1.2—1.5 300 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Typr: Mt. Elwind, near Hamadan, northern Iran, Aucher 79, holotype (p), isotype (K); both seen. Upper petals glabrous to few-haired. DisTRIBUTION, At about 2000 m., Iran. REPRESENTATIVE SPECIMENS. H. F. Macmillan 7, 8, 9; Koelz 18110a, 18144; Miss Lindsay 1288; Pichler in 1882; Knapp in 1884. 54. Delphinium turkmemum Lipsky, Acta Horti Petrop. 18: 2. 1901. FIG. 5, 1. Delphinium eas Steph. ex Willd. var. pilosulum Trautv. Acta Horti Petrop. 9: 437 Delphinium pilosulum B. Fedtsch. Fl. Tian-schan Occid. 1: 98. 1904. Stem or stems terete, 2.5—7 dm. tall, slender, pilose at base, glabrous in middle, more or less glandular-puberulent on pedicels and in inflorescence; lower leaves 3—6 cm. broad, subreniform in outline, 5-fid into cuneate seg- ments with pinnate lobing into oblong parts 2-4 mm. wide, glabrous or with a few long slender hairs; lower petioles 4-10 cm. long, basally dilated; cauline leaves remote, with narrower lobes; petioles shorter; inflorescence pedicels ascending, 7-10 mm. long; bracteoles 2, lanceolate, 1.5-2 mm. long, at about the middle of the pedicel; sepals pale violet-blue, the upper sepal ovate, glandular-puberulent, 8-10 mm. long, 4-5 mm. wide, the spur stout, horizontal, 16-20 mm. long, 3-4 mm. wide at the base, obtuse; Jateral sepals oblong, 12 mm. by 6 mm., obtuse, glandular-puberulent on midrib; lower sepals oblong-elliptic, 12 mm. by 6 mm., rounded at apex, glandular-puberulent; upper petals oblique, yellowish, the lamina ca. 1 cm. long, the spur ca. 2 cm. long; lower petals 12-13 mm. long, the claw 6 mm., the blade bearded, lobed for about 3 mm.; stamens 5—7 mm. long; anthers yellow, 1 mm. long; follicles 3, glandular-pubescent. Type: Kisil-Arwat, Turcomania, AJ. Becker 29, Dec. 18, 1896 (LE); seen DIsTRIBUTION. Turcomania, Transcaspia. REPRESENTATIVE COLLECTIONS. Turcomania: Litwinow 447, 448. Transcaspia: Lipsky, May 8, 1912; F. N. Meyer 552; Sintenis 1672; Lipsky 3070. 55. Delphinium uncinatum Hook.f. & Thoms. in Hook.f. Fl. Brit. Ind. 1: 24. 1872. Fic. 5, J. — penicillatum Hook.f. & Thoms. Fl. Ind. 48. 1855, not Boiss., pastor vestitum ses Diagn. II. 1: 13. 1853; Fl. Orient. 1: 92. 1867, not Wall. ex Royle, 183 Delphinium uncinatum var. Pe Briihl in Brith] & King, Ann. Bot. Gard. Calc. 5: 94. 1896. Based on Stocks 994, from Beluchistan; seen. Delphinium uncinatum var. aitchisonii Boiss. in Brihl & King, Ann. Gard. Calc. 5: 94. 1896. Based on Aitchison, Dec., 1879 from Kurrum Valley; (CAL), seen. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 301 Stem simple, terete, subflexuous, 1.5-9 dm. tall, subglabrous below, densely pubescent above with spreading hairs; radical leaves crowded, the petioles dilated at the base, 2-12 cm. long; laminae 3—5-sect to base, 1.5—6 cm. long, somewhat broader, subglabrous to villous, the segments narrow- cuneate at base, pinnately dissected into narrow toothed lobes commonly 1.5-8 mm. wide; cauline leaves few, somewhat reduced, the lobes nar- rower; inflorescence racemose or subpaniculate, 1—-2.5(—4) dm. long, lax, spreading-hairy; bracts lanceolate, 3-5 mm. long; pedicels commonly 4-6 mm. long; bracteoles lance-linear, basal, ca. 2 mm. long; sepals bluish purple to pale, sparsely villous, the upper oblong-ovate, 7-8 mm. long, pointed, the spur hooked, cylindric, 7-10 mm. long; lateral sepals hairy on midline, rounded at apex, 7-8 mm. long; lower sepals oblong, rounded at tip, ca. 8 mm. long; upper petals 7-10 mm. long, short bilobed at apex, the spur 5-6 mm. long; lower petals 8-9 mm. long, the laminae oblong, 5 mm. long, bilobed for ca. 3 mm., bearded, the claw 4 mm. long; stamens ca. 5 mm. long, glabrous or nearly so; anthers 0.6 mm. long; follicles 3, glabrous or hairy. Type: Stocks 994, among wormwood on plains, Beluchistan (K) seen. Cited by Boissier as D. vestitum, by Hook. f. & Thoms. for their penicil- latum, and by Briihl and King for their var. pseudovestitum. DistrisuTion, At 1000-2000 m., West Pakistan (Punjab, Beluchistan), Afghanistan, Kashmir. ILLUSTRATION. Britut & Kine, Ann. Bot. Gard. Calc. 5: pl. 116, fig. 2,3. 1896. REPRESENTATIVE MATERIAL. Afghanistan: Kurrum Valley, Aitchison, Dec. 1879. Beluchistan: Stocks 994; J. H. Lace 3351, 3326. Punjab: R. R. Stewart 7002; G. Watt 603; Jacquemont 672. Kashmir: Stewart 12688, 27322; T. Thom- son, P. Briihl, in Briihl & King, Ann, Bot. Gard. Calc. 5: 94. 1896, proposed D. uncinatum var. vicaryi from the Punjab, based on Vicary from near Rawal Pindi. I have not seen the specimen, which seems to have been characterized largely by being glabrous. It probably is an unusually glabrous plant of D. uncinatum. 56. Delphinium ursinum Rech.f. Anz. Osterr. Akad. Wiss. Math.-Nat. 88: 225. 1951. Fic. 5, K. Stem grayish, 5-10 dm. tall, slender, erect, 1—2-branched above, strig- ulose, terete, few leaved; petioles somewhat dilated at base, the lower to 15 cm. long, the upper shorter, strigulose; leaf blades rounded-cordate in out- line, 6-10 cm. broad, palmatifid almost to the base into 7 cuneate-obovate divisions, these 3-laciniate to about the middle, then again 3-lobed into oblong parts 3-6 mm. wide, mucronulate and obtuse or acute at apex; uppermost leaves trifid into lanceolate segments; inflorescence of 1 or more racemes, the terminal one many flowered and to ca. 3 dm. long, rather dense above, strigulose; bracts lanceolate, entire, 5-15 mm. long; 302 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 pedicels 2-10 mm. long, appressed to axis; bracteoles lance-linear, usually below the middle of the pedicel, 2—4 mm. long; sepals white, suffused with blue, nerves greenish, somewhat strigulose; upper sepal elliptic-ovate, 9 mm. by 5 mm., obtuse, the spur 8 mm. long, straight, subcylindric, ca. 2.5 mm. wide at base, 2 mm. at the blunt tip; lateral sepals elliptic, 8 mm. by 5 mm., obtuse; lower sepals like the lateral, but more pointed; upper petals with a bluish, almost straight lamina, bidentate at the apex, 7 mm. long, the spur 9 mm.; lower laminae bearded, 4 mm. by 3 mm., deeply cleft, the claw 4 mm. long, serrulate on back edge; stamens 4.5—5 mm. long, the upper filaments and anthers bluish; anther suborbicular, 0.6 mm. long; follicles 3, hairy. Type: Iran, prov. Gorgan (Asterabad), at 2400 m., on Mt. Shahvar near Hadjilang, K. H. & F. Rechinger 6083, holotype (w), not seen; iso- types (K, UC, US), seen. 57. Delphinium vanense Rech. f. Symb. Bot. Upsal. 11(5): 8. 1952. BiG. 5; & Delphinium cyphoplectrum Boiss. var. vanense (Rech.f.) Davis, Notes Bot. Gard. Edinb. 26: 171. 1965. Stem zigzag below, leafy, 3-5 dm. high, striate, strigulose, subsimple or short branched in lower half, the internodes 4-6 cm. long; petioles of lower leaves broadly vaginate, to 7 cm. long, those of lower cauline leaves 3-4 cm., the blades 3.5—6 cm. long and wide, round-cordate in outline, narrowly palmate-tripinnatisect, the ultimate lobes 10-15 mm. long, 1-1.5 mm. wide, divaricate, oblong-linear, acute, minutely strigulose, while in the upper cauline leaves the ultimate lobes are linear, 1 mm. broad; racemes glabrous to hirsute and viscid, lax, strict, erect; pedicels ca. 10 mm. long, divergent, the 2 bracteoles near the base, linear, ca. 3 mm. long; sepals deep blue to violet blue, almost glabrous, the upper sepal 10-12 mm. long, 5-6 mm. wide, lance-ovate, subapiculate, spur almost straight to slightly curved, obtuse, subglabrous, 10-12 mm. long, 2.5 mm. wide; middle sepals oblong-oblanceolate, ca. 10 mm. by 3 mm., the lower oblanceolate with a broad base, 10-12 mm. long, 5-6 mm, wide, obtuse; upper petals about as long as sepals, the laminae glabrous, vertical to the spur, ca. 11 mm. long, narrow, bidentate at apex, the spur 11 mm. long; lower laminae long- hairy, 8 mm. long, lance-ovate, deeply divided into 2 lance-oblong lobes, the claw 4-5 mm. long; stamens 6-7 mm. long, glabrous; anthers yellow- ish-blue, round-oblong; follicles 3, subglabrous to hairy. Tyre: Kurdistan (prov. Van, Turkey), 20 km. from Tatvan, at 1900 m., Frédin 328 (ups); not seen. DistRIBUTION. At 1900-2900 m., Turkish Kurdistan. SPECIMENS SEEN. Turkey: Bitlis, Davis 22542; Artos Dag, Davis 22753. The specimens cited have the narrow leaf segments of D. laxiusculum, but that species has strigose racemes. |To be continued | 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 303 THE GENERA OF EUPHORBIACEAE IN THE SOUTHEASTERN UNITED STATES ? Grapy L. WEBSTER EUPHORBIACEAE A. L. de Jussieu, Gen. Pl. 384. 1789, “Euphorbiae”’ (SpuRGE FAMILy) Trees, shrubs, herbs, or vines; stems in some taxa succulent and often with colored or milky latex. Leaves alternate or opposite (or, rather u ‘Prepared for a generic flora of the southeastern United States, a joint project of the Arnold Arboretum and the Gray Herbarium of Harvard University made pos- established in the first paper in the series (Jour. Arnold Arb. 296-346. 1958) and continued to the present paper. The area covered includes Nerth and South Carolina, Georgia, Florida, Tennessee, Alabama, Mississipp, Arkansas, and Louisiana The descriptions are based primarily on the plants of this formation in brackets. References which the author has not seen are marked by an asterisk. Part of the data reccrded in this presentation have accrued from a long-term study of the reprcductive morphology of the Euphorbiaceae which has been sup- ported by several grants from the National Science Foundation (G-—7148, G—-18768, taking review of the manuscript and his are due my wife, Barb D. Webster, for her critical proofreading of th u- scr 0 Gordon W. Dillon has helped greatly in checking the abbreviations Delzie Demaree, and Dr. John Thieret. The chromosome count furnished through the courtesy of Dr. Jack Ellis. y Clapman from dissections by Dr. Wood. These materials were collected for the Generic Flora of the Southeastern United States by Dr. Wood, Dr. F. C. Craighead, Dr. R. C. Rollins, Mr. Alan Strahler, and Dr. J. L. Thomas. Eprror’s Notre. This treatment of the Euphorbiaceae, adhering to the plan of ive than the other papers published in this series because of the unusual interest a relevance of this work to the taxonomic framework of the entire family. — C. oop, JR 304 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 3—6-lobed or -parted in most taxa, but sometimes lobes more numerous, or reduced or absent. Corolla of mostly 3—6 separate or connivent petals (rarely more numerous or united), or often rudimentary or absent. Disc often present in flowers of one or both sexes, usually extrastaminal, of separate segments or cuplike. Stamens (1—)3—20[—400], inserted on a flat or convex receptacle; filaments free or united; anthers mostly 2- locular, dehiscing longitudinally; pollen tectate or intectate, inaperturate to polytreme (mostly tricolporate). Gynoecium of 3 or 4 united carpels [in the majority of taxa; reduced to 1 or multiplied up to 25 in others]; stigmas entire to bifid or lacerate; styles distinct to connate; ovary su- perior, 3- or 4-locular; placentation axile; ovules 1 or 2 per locule, anat- ropous or hemitropous, inserted beneath an obturator, crassinucellate, with 2 integuments, nucellus often prolonged as a beak in contact with the obturator, Fruit typically a capsular schizocarp, of 3 elastically de- hiscent mericarps falling away from a persistent columella, but in some taxa drupaceous [baccate, or samaroid]. Seeds 1 or 2 per locule, or (by reduction) solitary in the fruit; seed coats thin to bony or fleshy; endo- sperm usually present; embryo straight or bent; cotyledons usually broader than the radicle, plane or (rarely) folded. Embryo sac normal, mono- sporic (Polygonum type) or in some disporic or tetrasporic. (Including Acalyphaceae KI. & Garcke, Antidesmaceae Walp., Peraceae KI. & Garcke, Phyllanthaceae KI]. & Garcke.) Type GeENus: Euphorbia L. At once one of the largest and most diverse families of angiosperms, the Euphorbiaceae comprise some 7000 species in about 300 genera. Al- though the family is predominantly tropical, several hundred species of the tribe Euphorbieae, together with scattered species of primarily tropical genera such as Croton, are found in temperate regions. In the United States, 22 indigenous genera are known, 18 occurring in the southeastern United States. The 24 native and naturalized genera in the Southeast are represented by about 115 species. The amplitude of morphological variation is so great that it is difficult to characterize the family, and, for this reason, many authors have suggested that it may be polyphyletic in origin. For the majority of genera, the family may be recognized by the unisexual flowers, the presence of a floral disc, and the trimerous gynoecium which typically ripens into a septicidally dehiscent schizocarp with three or six anatropous seeds. Vegetatively, the family is marked by having mostly alternate, stipulate leaves, often with glandular spots or appendages on the blades. Latex is present in many genera, but is far from being as ubiquitous as is implied by many textbooks. The cymose inflorescences of many Euphorbiaceae are distinctly pro- terogynous, the female flowers at the lower dichotomies of the inflorescence maturing before the males; this is especially clearly seen in Jatropha, Cnidoscolus, and many cyathia of Euphorbia. Pollination relationships have not been studied in the vast majority of genera, but the observations of Warmke on Hevea and the analysis of various floral characters suggest 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 305 that pollination of many tropical Euphorbiaceae is carried out mainly by small Diptera. Despite the frequent citation of the Euphorbiaceae as an example of a wind-pollinated family, only a few genera, such as Acalypha and Ricinus, seem to be truly anemophilous. Pollination by humming- birds has been observed in Pedilanthus and may occur in other red-brac- teate taxa of tribe Euphorbieae. Observations on weedy species of Euphorbia and Phyllanthus suggest that self-compatibility is widespread in many herbaceous Euphorbiaceae. Embryo sac development is of the normal, monosporic type in most genera, but tetrasporic types are prevalent in Acalypha, and, in Euphorbia, both bisporic and tetrasporic types have been recorded. Apomixis was long ago demonstrated by Strasburger in the Australian Caelebogyne ilicifolia J. E. Sm. which produces nucellar embryos (often with poly- embryony). Adventitious embryony has also been found in Euphorbia dulcis L., but gametophytic apomixis has not been observed in the Euphorbiaceae. Anatomical studies have shown that there is great variation in wood anatomy (particularly in fiber thickness, parenchyma distribution, and ray configuration), but no sufficiently comprehensive systematic survey of euphorbiaceous woods has yet been made. There is tremendous di- versity in trichome shapes in some tribes (particularly the Crotoneae), but very little in others (e.g., the Phyllantheae). Petiolar anatomy was studied by Dehay, whose conclusions as to family affinities seem exces- sively speculative; but within certain groups it may provide an excellent indication of affinity (e.g., in evaluating the supposed relationship of Cnidoscolus with Jatropha). The laticiferous organs of Euphorbiaceae have been studied intensively in such economically important genera as Hevea and Manihot, but still have not been surveyed in the manner necessary to obtain taxonomically significant data (cf. summation by Gaucher), Latex is not reported in any genus of the subfamily Phyllan- thoideae, although Bischofia and Uapaca possess tanniniferous vessels which seem homologous with the “laticifers” of many Crotonoideae. The distinction between laticiferous and tanniniferous vessels is to some ex- tent arbitrary, since the so-called laticifers in the Crotonoideae produce a great variety of compounds, including tannins, saponins, starch, resins, terpenoids, and rubber. Clearly, a study of the origin and evolution of the laticiferous systems in the Euphorbiaceae should throw a great deal of light on phylogenetic relationships within the family. The Euphorbiaceae contain an extraordinary diversity of organic com- pounds, possibly more than in any other plant family. In addition to the compounds enumerated above, a variety of alkaloids has been found in both the Phyllanthoideae (especially subtribe Phyllanthinae) and Croton- oideae (mainly in Croton). The oily seeds of many Crotonoideae contain unusual fatty acids and also some peculiar, highly poisonous proteinaceous compounds (phytotoxins), the structures of which have not yet been elucidated. It seems clear that there is an exceedingly bright future for chemotaxonomic studies here. y 306 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Although there have been many studies on floral morphology in the Euphorbiaceae, few general surveys of the entire family have been made, other than those of Baillon and Michaelis. A disproportionate amount of effort has been expended on investigations of the cyathium of the Euphor- bieae, but the results have, in many instances (e.g., Haber, Bodmann), been vitiated by the failure of authors to correlate meaningfully anatomical structure with the systematic position of the taxa studied. The use of floral characters as indicators of relationship has been complicated by the circumstance that trends of modification of floral parts have been both fluctuating and reversible. For example, the primitive euphorbiaceous flower was probably pentamerous with two whorls of stamens, to judge from morphological relationships among the primitive genera of Phyl- lanthoideae such as Astrocasia, Heywoodia, and Wielandia. During the evolution of other taxa, there has been a general overall trend towards reduction in number of parts, so that the gynoecium in the great majority of taxa is three-carpellate, and stamens are often reduced to five, or even to only one, per flower in the Euphorbieae and some Hippomaneae. On the other hand, a secondary increase in number has occurred in calyx- lobes (e.g., up to 10 or 12 in Dalechampia), stamens (over 100 in several species of Croton), and carpels (10-20 in Hura and Hippomane). One of the most fruitful subareas of floral. morphology has been the investigation of pollen structure, which was started by Erdtman and further elaborated by Kohler, Punt, and Webster. Pollen characters seem, in many instances, to be extremely valuable indicators of affinity, as Erdtman originally suggested for the taxa with Croton-type pollen; Punt has shown that examination of the pollen makes possible an immediate and confident assignment of many genera to the proper tribe. The greatest diversity of pollen types is found within Phyllanthus, where the micro- spores may be tectate or intectate, prolate to oblate, inaperturate to tri- colporate, polycolporate, pancolporate, or porate. The causes for this variation are still obscure, but it is notable that the gamut of pollen types in Phyllanthus rather strikingly parallels that in Polygonum, which has flowers of much the same size and configuration that probably are pollinated by similar agents (largely Diptera). Ovule and seed characters are also important in the classification of the Euphorbiaceae but have not yet received the attention they deserve. Although the ovules are prevailingly anatropous, in the tribe Phyllantheae hemitropous ovules are common and seem to characterize natural group- ings of genera. Plants such as Tetracoccus, with anatropous ovules which develop into shiny black seeds, clearly must be removed from the tribe. The seeds furnish important systematic characters at various levels of affinity, from characterizing species to separating tribes. Carunculate seeds are rare in the Phyllanthoideae but are quite common in the Crotonoideae, so much so that a number of uncritical writers have de- scribed the seeds of Euphorbiaceae as generally carunculate. In fact, however, the development of the caruncle is a very fluctuating character, 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 307 and within single genera (e.g., Euphorbia) rather closely related species may be separated by the presence or absence of the caruncle. The schizocarpous fruit of Euphorbiaceae is generally described as a tricoccous capsule, and the mericarps as cocci. There seems to be no reason to use these special names, since the fruit is essentially similar to that found in, for example, the Malvaceae. The elastic dehiscence of the fruit is characteristic, and seeds may be hurled to a considerable distance; those of Hura travel several yards, with a noise like a gun shot. Drupaceous fruits are found in a considerable number of Phyllanthoideae (many of these cauliflorous), and those of some tropical species of PA»l- lanthus are baccate. The carunculate seeds of various taxa are said to be dispersed by ants, but there is still insufficient field evidence to cor- roborate this. Cytologically, the Euphorbiaceae are as diversified as in most other respects, but such a small proportion of the species (less than 5 per cent) has been studied that chromosome data are at present only of limited systematic usefulness. Except in some succulent species of Euphorbia, the chromosomes are relatively small, so cytological comparisons have to depend largely on number alone. Perry, who completed the first chromo- somal survey of the family, concluded that the basic number is 8 and that annual species are more primitive than perennials. Subsequent studies, however, suggest that neither conclusion is justified. In the sub- family Phyllanthoideae it seems most likely that x=13 in the more primitive taxa, while in the Crotonoideae the most widespread basic num- bers are x=9, 10, and 11. Even in Euphorbia itself 10 seems more ) taxa and five tribes of uniovulate taxa were taken over by Pax, with some modification, as the basis of the system currently in general use. 308 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Klotzsch and Garcke, writing contemporaneously with Baillon, initiated a policy of splitting the Euphorbiaceae which has been followed by a minority of workers; they recognized six families which were roughly the equivalents of Mueller’s tribes. The most recent adherent of this philosophy is Hurusawa, who has promoted the four subfamilies of Pax (two of Stenolobeae and two of Platylobeae) to separate families. This procedure has little to recommend it, since the palynological evidence suggests that the Stenolobeae are probably an artificial group which should be returned to positions within the Platylobeae. Furthermore, while there is a rather sharp gulf between the uniovulate and biovulate Euphorbiaceae, they seem unquestionably to be related, and separation into different families overrates differences at the expense of similarities. The proposal of Erdtman to realign the tribes of Crotonoideae on the basis of pollen morphology seems reasonable and is, on the whole, borne out by the detailed studies of Punt. As noted by Erdtman, this results in an arrangement much closer to that of Bentham than to that of any other of the 19th century systematists. Although no formal, revised system of the Crotonoideae has yet been published, it appears that many of the data necessary are already at hand. Kohler’s thorough analysis of pollen morphology in the biovulate Eu- phorbiaceae has, in general, confirmed and extended the observations of n proper places among the platylobian genera represents a valuable con- tribution to the classification of the family. A few of his innovations (e.g., placing Poranthera in the Antidesmeae, rather than in the Andrachneae) do not accord with evidence from other data, but, on the whole, his new arrangement is undoubtedly much closer to the lines of natural affinity than any heretofore proposed. The most interesting proposal of Kohler is to separate the mainly Old World biovulate genera with spinulose pollen grains as a separate sub- family. Because of the strong correlation with other morphological characters, this suggestion appears warranted, and the biovulate Euphor- biaceae are herewith assigned to two subfamilies: subfam. Phyllanthoi- deae Pax, with alternate leaves, nonspinulose pollen, ovules often hemitropous, and seeds not carunculate; and subfam. Oldfieldioideae Kohler & Webster,” with leaves often opposite or whorled, mostly spinulose pollen, ovules strictly anatropous, and seeds mostly carunculate. * Kohler described several new suprageneric taxa which are invalidly published because he failed to provide Latin diagnoses. His roposed subfamily merits cae and is here validated: Subfam. Oldfieldioideae Kohler & Webster, subfam. Arbores fruticesve saepe dioicae; foliis alternis, oppositis, verticellatisve, simplices vel palmatipartitis; floribus apetalis, disco plerumque nullo; staminibus liberis, granis eo hay paar ovulis in quoque 08 collateralibus. Typus: Oldfieldia Benth. e Kohler’s tribe Oldfieldieae was also invalidly published, but an earlier tribal name, Paivaeuseae, is validly published and available. As a co nsequence of Article 19 of the International Code of Botanical Nomenclature (1966), this virtually unpronounce- 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 309 By analogy with Kdéhler’s proposal for the biovulate Euphorbiaceae, it might be possible to divide the uniovulate Euphorbiaceae (Crotonoideae and Ricinocarpoideae of Pax) into two subfamilies based on pollen morphology. Punt has shown convincingly that the genera with crotonoid pollen should be placed together in a single taxonomic grouping. However, the morphological and anatomical evidence does not show as good a cor- relation with palynological characters as is true for the biovulate Euphor- biaceae. For example, Hevea has articulated laticifers and in this and other respects seems related to Manihot; yet it has colporate pollen uniovulate Euphorbiaceae are less diversified than the biovulate taxa, and there is no consistent difference in seed morphology. Tentatively, therefore, it seems preferable to retain all of the uniovulate Euphorbiaceae (including the stenolobian Ricinocarpoideae of Pax) in a single sub- family Crotonoideae. The replacement of the subfamilial quartet of Pax by the present trinity does not make much difference in the classification of North American Euphorbiaceae, since the only genus of Oldfieldioideae repre- sented is Tetracoccus. That small genus of xerophytic shrubs is confined to the western United States and Mexico and does not enter our area. The extent of phylogenetic relationship between the Euphorbiaceae and other families has long been controversial. Baillon, in 1858, imagined a tetrahedral relationship, with the Euphorbiaceae on one face and the Malvales, Geraniales-Rutales, and Rhamnales on the other faces. Pax, summing up his studies in 1924, regarded the Euphorbiaceae as of poly- phyletic (or at least diphyletic) origin, with part of their ancesters in the Geraniales-Sapindales complex, and part in the Malvales. Hutchinson, whose view is perhaps the most extreme, suggested an origin from at least four orders: Bixales, Tiliales, Malvales, Celastrales (and possibly also Sapindales). Croizat, in his principal consideration of the problem (1940), emphasized the affinity to the Malvales, especially the Sterculiaceae. The existence of so many multifarious and disparate hypotheses sug- gests that little understanding of the problem of relationship at the family level has been achieved. Part of the confusion surrounding the affinities of the Euphorbiaceae resides in the unsatisfactory delimitation in the Hutchinsonian sense. able name is automatically replaced by Oldfieldieae when both subfamilies and tribes are used in the subfamilial classification. Although the result is the possibility of two same taxon with precisely the same taxonomic circumscription e€ nk (an unfortunately unforeseen consequence of Article 19), it seems far preferable to describe a new subfamily based on Oldfieldia than to per- petrate a subfamilial name such as Paivaeusoideae ! 310 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 While within the Euphorbiaceae the pollen grains furnish the most critical characters for delimitation of infrafamilial taxa, at the family level the ovules may be more significant. All bona fide Euphorbiaceae have pendent epitropous ovules, with the micropyle abaxial to the funicle often treated as a distinct family Irvingiaceae, gynoecia with epitropous ovules rather similar to those of the Euphorbiaceae occur. Recently, Capuron has described from Madagascar a new genus of Irvingiaceae, Cleistanthopsis, which, as its name indicates, has an extraordinary general similarity to Cleistanthus, of Euphorbiaceae subfam. Phyllanthoideae. Desp‘te the solitary style and uniovulate locules, which preclude its being directly ancestral to the Ur-Euphorbiaceae, Cleistanthopsis obviously de- serves consideration as at least a possible euphorbiaceous “great-uncle.” Further botanical exploration in such regions as Madagascar may there- fore eventually provide the solutions to some of the problems over which several generations of morphologists have quibbled in vain. Part of the difficulty involved in tracing phylogeny within the Euphor- biaceae and in relating it to other families comes about because of the relatively poor fossil record. Kirchheimer (1957) disposed of most of the reports of fossil Euphorbiaceae from Central Europe; and in general, it would appear that records based on leaf impressions are even more un- reliable than is usual in paleobotany, due to the large amplitude of foliar variation in Euphorbiaceae.* Recently, however, the picture has improved somewhat due to the discovery by Chandler of a considerable variety of fossil euphorbiaceous fruits from the Eocene of England, and the descrip- tion of several fossil woods from India. The preservation of the English fruits and seeds is unexpectedly good, considering the dehiscent nature of the fruits and the thin seed-coats of most genera. The striking genus Paleo- wetherellia Chandl. has fruits which, in some respects, suggest the neo- tropical genus Hura, although the orientation of the ovules precludes identification with any known euphorbiaceous genus; it may possibly *For a particularly egregious example of the ill-founded identifications perpetrated by some paleobotanists, see the paper by Potbury, who described a new species of = a Oo rt) ° o i) & ° = ry ° oO ist) oO Ko] o wm — co _ ° Ss ° po 3 ° n % re) a _ co © fs) = r) san] ry re) sag oO = & oO [= etc. Her as a wild guess, and the fossils she referred to Acalypha and Drypetes are scarcely more convincing. Even the fossil leaves classified as Aleurites are only possibly con- generic. All of these remains could at best be ascribed to form-genera such as Euphor- ; arius), but, to the writer’s knowledge, there is as yet no unequivocal record of a living gerontogean euphorbiaceous taxon as a fossil in the New World, and vice versa. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 311 represent some extinct branch of the family. Other carpic fossils are much closer to living taxa. For example, Chandler’s Euphorbiotheca lakensis, which she compared to Andrachne and ‘Securinega’ (i.e., Flueg- gea), is a very close match for Flueggea suffruticosa (Pall.) Baill., a species widespread in Siberia and China and with a close relative in Japan. Madel, in a careful review of fossil euphorbiaceous woods, has shown that at least seven form-genera from the Cretaceous and Tertiary can be recognized. Unfortunately, the woods of many taxa of subfam. Phyllan- thoideae are not very distinctive and can easily be confused with woods of other families, such as the Flacourtiaceae. Madel rather convincingly compares her Upper Cretaceous Paraphyllanthoxylon capense with Mar- garitaria discoidea (Baill.) Webster,* a tree widespread in tropical Africa today. Bailey’s Paraphyllanthoxylon arizonense, a roughly con- temporaneous New World fossil, is less similar to the neotropical Mar- garitaria nobilis L. f., but may prove to be more similar to Margaritaria or Phyllanthus than to any other genera. There are now rather good records of Tertiary woods similar to those of living species of Antidesma, Bridelia, Aleurites, and Hevea. If these fossil woods can be matched with fossils of fruits and seeds (as found in the English deposits), there would appear to be a good opportunity to make much more nearly definitive identifications and thus to produce — at last — some historical documenta- tion to prop up the rather shaky phylogenetic classification of the family. A number of botanists over the past century have sought to resolve some of the problems of the phylogenetic position of Euphorbiaceae by ex- cluding various extraneous elements. Baillon proposed the removal of the Buxaceae and was followed in this by Mueller, though not by Bentham. Today it appears that the resemblance of Buxaceae to Euphorbiaceae is superficial, as the boxwoods differ in a number of fundamental charac- ters. The perianth of the Buxaceae does not appear to be strictly homol- ogous with that of Euphorbiaceae, as it consists of four decussate tepals in the male flower and about six to twelve imbricate bractlike tepals in the female flower. The gynoecium of Buxaceae is very different in hav- ing the styles emerging laterally, not apically, on the ovary, as well as in the apotropous ovules; and at least in Buxus and Pachysandra, endo- sperm development is cellular, rather than nuclear. Despite the tantalizing similarity between the pollen grains of Buxus and Pachysandra with those of various Euphorbiaceae (subfam. Crotonoideae), it appears most probable that this resemblance is due to convergence and is not indicative of affinity. Somewhat more similar to the Euphorbiaceae is the genus Daphniphyl- lum, which resembles some taxa of subfam. Phyllanthoideae in both habit and its gynoecium with paired, anatropous, epitropous ovules. However, Daphniphyllum diverges markedly in its exstipulate leaves, imperfectly septate ovary, ovules without an obturator, and especially in its seeds itaria discoi aill.) Webster, comb. nov. Cicca discoidea Baill. eeacue ne ge an is one oe Peaccvwent (Baill.) Muell. Arg. Linnaea 32: 51. 1863 312 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 48 with a minute embryo. The latter feature suggests Ranalian affinities for wood of Daphniphylium, according to Janssonius, is very different from any Euphorbiaceae and resembles that of Hammamelidaceae, thus con- firming the suggestion of affinity made by Hallier. In its gynoecial con- formation, Daphniphyllum shows some resemblance to the Buxaceae de- spite its differently oriented ovules; and it also resembles Buxus in its exstipulate leaves and cellular endosperm. It would appear, therefore, that the Daphniphyllaceae are at least much more closely related to the Buxaceae than either family is to the Euphorbiaceae; and both probably belong in the Hammamelidales-Trochodendrales complex, but with their exact positions yet to be determined. Still other genera which have been viewed with suspicion as to their euphorbiaceous parentage include Aextoxicon, Bischofia, and Picrodendron. Aextoxicon, which somewhat resembles the euphorbiaceous genus Pera in habit (partly because of its lepidote indumentum), has been excluded from the Euphorbiaceae by Pax on the basis of its aberrant perianth, apotropous ovules, and ruminate endosperm. As the only representative of the family Aextoxicaceae, it would according to Pax take a position adjacent to the Icacinaceae. Picrodendron, a West Indian genus with compound leaves and fruits superficially like walnuts, has been variously placed in the Simaroubaceae, Euphorbiaceae, or in its own family, Picro- dendraceae. The female flower of Picrodendron (illustrated by Fawcett and Rendle) is quite typical for the Euphorbiaceae, as the anatropous ovules are inserted below an obturator. The genus therefore may be truly euphorbiaceous, and its leaves and spinulose pollen suggest a possible affinity with genera of the subfamily Oldfieldioideae. Bischofia presents a more difficult problem. Although the genus re- sembles some of the subfam. Oldfieldioideae in habit, the pollen is dif- ferent, and the wood structure is not readily distinguishable from that of some taxa of subfam. Phyllanthoideae. Airy Shaw has proposed to create a separate family Bischofiaceae, to be placed near Staphyleaceae because of the resemblance between Bischofia and the Chinese genus Tapiscia. Bischofia differs from Tapiscia in so many ways (e.g., unisexual flowers, nearly free calyx segments, completely different gynoecium, and larger embryo) that a very close relationship between the two genera can scarcely be maintained. Nevertheless, Bischofia is so isolated in the Euphorbiaceae that its exclusion is quite possibly warranted, and Airy Shaw's suggestion of a staphyleaceous kinship deserves further investi- gation. Airy Shaw has also created new unigeneric families for Androstachys, Hymenocardia, and Uapaca; but, although each of these taxa is certainly aberrant, they have the fundamental gynoecial characters of Euphorbia- ceae, and there does not seem to be any compelling reason why they should be removed. A rather surprising realignment of the boundaries of the Euphorbiaceae 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 313 has recently been proposed by Forman, who has removed Microdesmis and Galearia to the family Pandaceae, which heretofore included only the anomalous African genus Panda. Forman’s arguments for an affinity between the three genera seem convincing, and it must be noted that Microdesmis and Galearia have always occupied a rather anomalous position in the Euphorbiaceae; Bentham created a special tribe Galearieae, apparently because he could not place them with any other group o genera. Microdesmis is aberrant from most other Euphorbiaceae in its basic chromosome number of «=15, in the compression of the anthers against the vestigial gynoecium or petals in the male flower bud, and in the absence of an obturator. Galearia shares at least some of these characters, and both genera have a wood structure unlike that of other uniovulate Euphorbiaceae. Although it is difficult to write a diagnosis for the revised family Pandaceae which will unequivocally exclude all Euphorbiaceae, the inclusion of Panda — with its orthotropous ovules and unusual fruit — would make the Euphorbiaceae more difficult to circumscribe. Conse- quently, it seems expedient to recognize the Pandaceae as an offshoot of the Euphorbiaceae and the only other family which is at all closely allied. When the Buxaceae, Daphniphyllaceae, and Pandaceae are excluded, the question remains as to where the closest affinities of the Euphorbiaceae are to be found. Despite all the additional information which has accrued, it must be admitted that no single taxon ancestral to the Euphorbiaceae can be designated at the present time, and it is not certain that any living taxon will ever be identified as the progenitor. The relationships of the polypetalous woody dicotyledons are notoriously reticulate, and the best that can be done at present is to specify the most plausible phylogenetic groupings. The inclusion of the Euphorbiaceae within the Geraniales of Engler appears to be justified, since in wood anatomy, floral morphology, and ovule and seed structure they resemble such families as Linaceae, Oxalidaceae, and Irvingiaceae. On the other hand, with respect to certain characters there are undeniable similarities to other groups, especially the Sterculiaceae and Thymelaeaceae, which show some striking similarities in pollen and seed structure. Croizat, among others, has emphasized the sterculiaceous affinity, but it should be noted that most of the malvalian- appearing Euphorbiaceae belong to the Crotonoideae, and the presumably more primitive Phyllanthoideae show little resemblance. Perhaps the similarities of Euphorbiaceae with the malvalean families and with the Flacourtiaceae may be explained as parallelism, in the sense that all of these families represent divergent branches of a plexus that may ultimately be of rosalian origin. It is not possible to go any further than that on the basis of current botanical knowledge. Because of their chemical diversity, the Euphorbiaceae include a rather large number of plants of economic importance. Although rubber from Hevea is perhaps the most valuable commodity, commercially important products are also obtained from euphorbiaceous seed oils (Aleurites, Ricinus, and, to a lesser extent, Croton and Jatropha), starchy tubers (Manihot), and waxes (Aleurites, Sapium, and Euphorbia). Still other 314 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 compounds, such as the alkaloids and saponins in the Phyllanthoideae, may yet prove to be of medicinal value. Finally, a number of Euphorbiaceae are grown as ornamentals. The most important species is doubtless the poinsettia (Euphorbia pulcherrima Willd. ex Grah.), but especially in tropical and subtropical gardens one may encounter species of PAyllanthus, Acalypha, Croton, Codiaeum (the widely cultivated “Croton” of horticul- turalists), Euphorbia, and Pedilanthus. In addition to the 24 genera treated in detail below, some of these orna- mental or economically important plants may be found cultivated in our area, especially in southern Florida. The Para rubber tree, Hevea brasil- iensis (Willd. ex A. Juss.) Muell. Arg., may be found in Key West and perhaps extreme southern Florida; it can be recognized by its long- petiolate trifoliolate leaves, paniculate inflorescences, apetalous flowers, and large fruits. Another euphorbiaceous tree planted in the Miami area, Bischofia javanica Blume, also has trifoliolate leaves but the leaflets are toothed (instead of entire as in Hevea) and the twigs produce no latex. Various species of Antidesma are cultivated, especially A. Bunius (L.) Spreng.; these are easily distinguished from other Euphorbiaceae by the combination of entire laurel-like leaves and the spicate female flowers which mature into edible drupaceous fruits. One of the most characteristic trees which may be seen in the Miami area is the sandbox tree, Hura crepitans L., immediately recognizable by virtue of its spiny stem, laticifer- ous twigs, heart-shaped leaves, bizarre flowers (the female with a dilated umbrella-like stigma), and large multicarpellate fruits which dehisce in a violent explosion. The latex and the seeds of this plant, as in man other local Euphorbiaceae of subfam. Crotonoideae, are dangerously poisonous, and plantings close to houses should be made with care. REFERENCES: Ary SHAw, H. K. Notes on Malaysian Euphorbiaceae. I. Kew Bull. 3: 484. 1949; II-XV. Ibid. 14: 353-397. 1960; XVI-XIX. Ibid. 469-479. ———. Notes on Malaysian and other Asiatic Euphorbiaceae. XX—XLVIII. Ibid. 16: 341-372. 1963; XLIX-LV. Ibid. 19: 299-328. 1965; LVI- LXVI. Jbid. 20: 25-49. 1966; LXVII-LXXXII. Ibid. 379-415. [Many new spp. and several new genera described in this series. | Diagnoses of new families, new names, etc., for the seventh edition of Willis’s “Dictionary”. Ibid. 18: 249-273. 1965. [Describes 4 new families: ARNOLDI, W. Zur Embryologie einiger Euphorbiaceen. Trav. Mus. Bot. Acad. Sci. St.-Pétersb. 9: 136-154. 1912. [Gives some details of megasporo- genesis in 2 spp. of Phyllanthus and individual spp. of Acalypha, Codiaeum, Glochidion, Jatropha, Pedilanthus, and Trigonostemon. | ASSAILLY, A. Contribution 4 la détermination des Euphorbiacées par la méthode anatomique. Bull. Soc. Hist. Nat. Toulouse 89: 157-194. 1954. [Deals largely with cultivated or medicinal spp. of France. | BAILEY, I. W. The problem of identifying the wood of Cretaceous and later age oe, Paraphyllanthoxylon arizonense. Ann. Bot. 38: 439-451. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE ats BaILiLon, H. ipneCeE i sive Euphorbiacearum monandrarum descrip- tionem. Ann. Sci. Nat. Bot. IV. 9: 192-204. 1858. . Etude générale os groupe des Euphorbiacées. 684 pp. 27 pls. Paris. 1858. . Species Euphorbiacearum. A. Euphorbiacées Africaines. Adansonia 1: 58-87. 1860; 139-173, 251-286. pl. 5. 1861; 2: 27-55. 1861; 3: 133-166. 1863 (?). —. Enumération des Euphorbiacées cultivées dans les Jardins Botaniques de Paris. /bid. 1: 104-117. 1860; 340-352. 1861 . Species Euphorbiacearum. Euphorbiscese Nev- Caledonicae. Jbid. 2: 211-242. 1862. Species Euphorbiacearum. Euphorbiacées Americaines. /bid. 4: 257- 377. 1864; 5: 221-240, 305-360. 1865. ‘ ‘Spectes Euphorbiacearum. Euphorbiacées Australiennes. /bid. 6: 282- 345. 1866. f a la parthénogénése et la suppression du genre Caelebogyne. Ibid. 368-379. . Nouvelles observations sur les Euphorbiacées. /bid. 11: 72-138. pl. 9. 1873. [Interesting discussion of Baillon’s philosophy of classification in the Euphorbiaceae, including a defense of his extremely broad circumscrip- -tions of such genera as Amtidesma, Phyllanthus, Tournesolia, and Ex coecaria. | pay arg Hist. Pl. 5: 105-256. 1874. BANERJI, I., & M. K. Durr. The development of the female gametophyte in some members of the Euphorbiaceae. Proc. Indian ve Sci. B. 20: 51-60. 1944. [Putranjiva, Trewia, Phyllanthus, Chamaesyc BentHAm, G. Notes on Euphorbiaceae. Jour. Linn. Soc. Bot. 17: 185-267. 1878 . HooKEr. eee Gen. Pl. 3: 239-340. 1880. [Treat- ment prepared by Bentham. | Bioum, H. Poisonous ates of Venezuela. xvi + 136 pp. Cambridge, Mass. 1962. pashowbincae 47-61.] Braun, A. Uber Polyembryonie und Keimung von Caelebogyne. Abh. Akad. Wiss. Berlin 1859: 109-263. pls. 1-6. 1860. [Discussion of Caelebogyne mainly on pp. 109-131.] Brown, N. E., Hutcuinson, & D. Prartn. Euphorbiaceae. Jn: W. T. THISELTON- Dyer, ed., Fl. Trop. Afr. 6(1): 441-1020, 1034-1059. 1911- 1913. [Pp. 441- 576. ‘1911; pp. 577-960. 1912; pp. 961-1059. 1913.] . Euphorbiaceae. Jn: W. T. THISELTON-Dyer, ed., Fl. Capensis m2): 216-516, 585, 586. 1915-1925. [Pp. 216-384. 1915; pp. 385-516. 1920; pp. 585, 586. Brown, W. H. The bearing of a on the phylogeny of flowering plants. Proc. Am. Philos. Soc. 79: 549 pls. 1-12. 1938. Capuron, R. Une Irvingiacée as Adansonia II. 5: 213-216. 1965. [Cleistanthopsis multicaulis Cap., gen. & sp. nov.; similar in habit and some rupee characters to Euphorbiaceae subfam. Feistcsemwr aie ae CHANDLER, M. J. Some Upper Spor - Eocene fruits from Egypt. Bull. Brit. Mus. Geol. 2: 147-187. pls. 10- 1954. Sates Euphorbiaceae, Lagenoidea and new genus sence iba 166-178. | CocKErELL, T. D. A. Fossil Euphorbiaceae, with a note on Saururaceae. Tor- reya 9: 117-119. 1909. 316 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Croizat, L. Glands of Euphorbiaceae and of Euphorbia. Chron. Bot. 4: 512-514. 1938. —. On the phylogeny of the Euphorbiaceae and some of their presumed 1940. allies. Revista Univ. Chile 25: 205-220. Notes on the Euphorbiaceae. II. Bull. Bot. Gard. Buitenzorg III. 17: 204-208. 1941, Peculiarities of the inflorescence in the Euphorbiaceae. Bot. Gaz. 103: 771-779, 1942 a. . New and critical Euphorbiaceae chiefly from the southeastern United States. Bull. Torrey Bot. Club 69: 445-460. ne [Notes on Croton, Manihot, ip sp ret including new spp. & combs. Notes n Americ n Euphorbiaceae, with description of eleven new species. some Wash. Road Sci. 33: 11-20. 1 . Novelties in American Euphorbiaceae. Tout? ‘Arnold Arb. 34: 165-189. 43b. Bibliographical notes on the Euphorbiaceae. Revista Acad. Colomb. Ci. Exact. Fis. Nat. 5: 541-547. 1944. [A valuable but not readily available review of exact dates of publications by Baillon, Mueller, and Klotzsch. ] . Novelties in American Euphorbiaceae. Jour. Arnold Arb. 27: 289-291. 1946. De Witpeman, E. Les latex des Euphorbiacées. I. en ta générales. Mém. fast: Colon. Belge Sci. Nat. Med. Collect. 8° 12(4): 1-6 DAnIKER, A. U. Uber die Euphorbiaceen und die eeatiea ey der Mono- chlantydene: Arch. Julius Klaus-Stiftung 21: 465-469. 1945. [An effort to find relationships between Euphorbiaceae, Balanopaceae, and Juglandaceae. | Denay, C. L’appareil libero-ligneux foliaire des Euphorbiacées, Ann. Sci. Nat Bot. X. 17: 147-290. pls. 1-4. 1935. DELPINO, F. Applicazione di nuovi criterii per la classificazione delle piante. Terza memoria. Mem. Accad. Sci. Ist. Bologna IV. 10: 565-599. 1 pl. 1889, [Pseudanthia in Euphorbiaceae, 572-580 EIcHLer, A. W. Bliithendiagramme construirt und erlaiitert. vol. 2, xx + 575 Leipzig. 1878. [Euphorbiaceae, 385-398; Euphorbieae treated in con- siderable detail. | RDTMAN, G. Pollen morphology and plant taxonomy. Angiosperms. xii + 539 Hook. f. in the Pandaceae. Kew Bull, 20: 309-321. pl. 5. 1966. [Includes appendices on leaf and stem anatomy hs e R. METCALFE and on wood structure by N. PARAMESWARAN & MET FROEMBLING, W. Anatomisch-systematische Catersuhung von Blatt und Axe der Crotoneen und Euphyllantheen. 76 pp. 2 pls. Inaug.-diss. Cassel. GAGNEPAIN, F., & L. Bete, Euphorbiacées. Fl, Gén. Indo-Chine 5(4): 229- 372. 1925; 5(5): 373-516. 1926; 5(6): 517-673. 1927. [Phyllantheae by BEILLE; remainder by GAGNEPAIN. GAUCHER, ‘fe Recherches anatomiques sur les Euphorbiacées. Ann. Sci. Nat. Bot _ VIII. 15: 161-309. 1902. GIBBS R. D., . T. Epwarp, & J. M. Fertanp. A novel colour reaction of some Euphorbia and Oxyanthus stiecied, Phytochemistry 6: 253-257. 1967. [Orange color reaction to 2.5 per cent HC] in methanol shown by 3 spp. of Euphorbia, and by certain other Euphorbiaceae, especially Phyllanthoideae.] GriseBacH, A. Erlaiiterungen augeswahlter Pflanzen des tropischen Amerikas. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 317 Abh. Ges. Wiss. Gottingen 9: 3-58. 1861. [Euphorbiaceae, including a sub- fandiital classification, 11-21. Grininc, G. Euphorbiaceae-Porantheroideae et Ricinocarpoideae. Pflanzen- reich IV. 147(Heft 58): 1-97. 1913. Hanscirc, A. Ueber die phyllobiologischen Typen einiger Fagaceen, Monimia- Melastomaceen, Euphorbiaceen, Piperaceen und Chloranthaceen. Beih. Bot. Centralbl. 10: 458-480. 1901. [Euphorbiaceae, 472-479. ] HEGENAUER, R. Chemotaxonomie der Pflanzen. Band 4. Dicotyledoneae: Daph- niphyllaceae-Lythraceae. 551 pp. Basel & Stuttgart. 1966. [ Euphorbiaceae, 103-140, 490-492; a valuable compendium of information with extensive references, examined too late for much of the appropriate information to be incorporated into this paper. | HERBERT, H. Anatomische Untersuchung von Blatt und Axe der Hippomaneen. Inaug.-diss. 62 pp. Munich. Hurvusawa, I. Eine nochmalige Durchsicht des herkémmlichen Systems der Euphorbiaceen i im weiteren Sinne. Jour. Fac. Sci. Univ. Tokyo Bot. 6: 209- 342. pls. 1-4. 1954. Incram, J. Notes on the cultivated Euphorbiaceae. 1. The flowers of the Euphorbiaceae 2. Chi idoscolus and Jatropha. Baileya 5: 107-117. 1957. JABLONSK Bridelieae. Pflanzenreich IV. 147- Gea 65): 1-98. 1915. _ Notes on neotropical Euphorbiaceae. 1. Synopsis of South American Sapium. Phytologia 14: pe ein 1967. [Includes key to 58 spp. recog- nized, Lyre: map. | New species and transfers. bid. 450-456 pls. 2-4. [All in the Seana piven 2 se 1890. Nachtrage zu Teil III. Abt. 5: 210-213. 1897; Ibid. 37, 38. ; Ibid. (Ergainzungshefte II): 191-195. 1906-1907; Ibid. Senet iit): 166-185. 1914. Einige neue Euphorbiaceen aus Amerika. Repert. Sp. Nov. 8: 161, 162. . Die Phylogenie der Euphorbiaceae. Bot. Jahrb. 59: 129-182. 1924. . Euphorbiaceae americanae novae II. Repert. Sp. Nov, 41: 224-226. 1937. . HorrMAnn. Euphorbiaceae. Pflanzenreich IV. 147-[{I], M-VII, IX-XVII. 1910-1924. [I.] Jatropheae (Heft 42): 1-148. 1910; II. Ad- rianeae, Additamentum I (Heft 44): 1-111. 1910; III. Cluytieae, Addi- tamentum II (Heft 47): 1-124. 1911; IV. Gelonieae (Heft 52): 1-41. 1912; V. Hippomaneae, Additamentum III (Heft 52): 1-319. 1912; VI. Acalypheae-Chrozophorinae, Additamentum IV (Heft 57): 1-142. 1912; VII. Acalypheae-Mercurialinae, Additamentum V (Heft 63): 1-473. 1914; IX. Acalypheae-Plukenetiinae (Heft 68): 1-108. 1919; X. Acalypheae- Epiprinae (Heft 68): 109-111. 1919; XI. Acalypheae-Ricininae (Heft 68): 112-134. 1919; XII. Dalechampieae (Heft 68): 1-59. 1919; XIII. Pereae (Heft 68): 1-14. 1919; XIV. Additamentum VI (Heft 68): 1-81. 1919; XV. Phyllantheae (Heft 81): 1-349. 1922; XVI. Acalypheae-Acalyphinae, 320 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Additamentum VII (Heft 85): 1-231. 1924. [Incomplete; no treatments published of Phyllantheae subtribes Phyllanthinae and Glochidiinae, Cro- toneae, and Euphorbieae. Parts I and II by Pax alone. | Systematische Stellung der Gattung Aextoxicon. Jahresber. Schles. Ges. ‘Vaterl. Cult. 1916(Bd.I, Abt.II.b.): 17-21. 1917. [Refers Aextoxicon to a separate family. | & . Euphorbiaceae. Jn: O. von KircHner, E,. Loew, & C ScHROTER, Lebensgeschichte der Bliitenpflanzen Mitteleuropas Bp 241- 308. 1930. ——_. Euphorbiaceae. Nat. Pflanzenfam. ed. 2. 19c: 11-233. 1931. Perry, B. A. Cytological icocipeaag in the Euphorbiaceae. Virginia Jour. Sci. 3: 140-144. 1943 (“1942” . Chromosome number an = phylosenstic relationships in the Euphorbi- aceae. Am. Jour. Bot. 30: 527-543. 1943. Potuamus, L. G. Rubber. Botany, production, and utilization. xvii + 448 pls. 1-64. London. 1962. [Includes discussions of Hevea, Manihot, and other laticiferous Euphorbiaceae Potsury, S. S. The La Porte iia of Plumas County, California. Carnegie Inst. ote 465: 29-81. pls. 1-18. [Four dubious taxa of Euphorbiaceae, 74-76. Punt, W. Pollen morphology ag a Euphorbiaceae with special reference to taxonomy. Wentia 7: 1-116. Raju, M. V.S., & A. N. Rao. Te achat of the male and female gameto- et kiad - Malotus albus Mull. Half-yearly Jour. Mysore Univ. II. B. 13: 5-8. . [Reports Drusa-type embryo sac as found by Ventura in M. Ramanvujam, C. G. K. Fossil woods of Euphorbiaceae from the Tertiary rocks of South Arcot District, Madras. Jour. Indian Bot. Soc. 35: 284-307. 1956. lapse 3 new genera: Bridelioxylon, Glochidioxylon, and Putranjivoxy- ia ae Ae ire American woods of the family Euphorbiaceae. Trop. Woods 938. pee F. sy A ac Untersuchung von Blatt und Axe der Acalyphen Inaug.-diss. 123 pp. 1 pl. Munich. 1892. ROTHDAUSCHER, H. Ueber die anatomischen Verhaltnisse von Blatt und Axe der Phyllantheen (mit Auschluss der Euphyllantheen). Bot. Centralbl. 68: 65-79, 97-108, 129-136, 161-169, 193-203, 248-253, 280-285, 305-315, ere 385-393. 1896, [Also published as repaged diss., 89 pp. Cassel. Santos, J. K. The laticiferous vessels and other anatomical structures of Ex- coecaria Agallocha. Philip. Jour. Sci. 47: 295-304. 1932. i R. E. Studies in be genus Hevea. I. Bot. Mus. Leafl. Harvard Univ, 13: 1-11. 1947; II. Zbid. 97-132. pls. 8, 9. 1948; III. Ibid. 14: 79-86, pls. 18, 19. 1950; V. Ibid. 15: 247-254. 1952; VI. Ibid. 255-272. ei S in the genus Micrandra I. The relationship of the genus Cunuria to saa: Ibid. 201-221. pls. 65-74. [Reduces Cunuria to Micrandra.] An n the genus Joannesia. Ibid. 17: 25, 26. 1955. [J. insolita Pitt, reduced. ie a tonne of J. princeps. | A new generic concept in the Euphorbiaceae. bid. 27-36. pis. 12-14. [ Vawbesis peeaaal i Schult.; related to Joannesia.| 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 321 ScHWEIGER, J. Beitrage zur Kenntnis der Samenentwicklung der Euphorbiaceen. Flora 94: 339-379. 1905. Scott, D. H. On the occurrence of articulated laticiferous vessels in Hevea. Jour. Linn. Soc. Bot. 21: 566-573. 1885. [Points out differences between laticiferous organs of Hevea and Manihot vs. those of Jatropha SHERFF, E. E. Additional studies of the Hawaiian Fuphorbiackae,. Publ. Field Mus. Bot. 17: 547-576. 1939. [Includes a revision of Hawaiian Claoxylon, reduction of Neowawrea to Drypetes. StncH, R. P. Forms of ovules in Euphorbiaceae. Pp. 124-128 im Plant Embry- ology, a symposium. vi + 273 pp. New Delhi. 1962. . Structure and development of seeds in Codiaeum variegatum Blume. Jour. Indian Bot. Soc. 44: 205-210. 1965. SmirH, J. J. Euphorbiaceae. Jn: Koorpers & VALeTon, Addim. Cogn. FI. Arb. Java. 12: 9-637. 1910. [A critical regional monograph; in Dutch and Latin. ] SOLEREDER, H. Systematic Bigs ge of oe Si Vol. II. Monochlamy- deae. (Transl. L. A. DLE & F FritscH.) vi + pp. 645-1182. Ox- ford. 1908. ncaa rsonness 1047-1055. ] THatacuar, T. Morphological studies in the Euphorbiaceae. Half-yearly Jour. Mysore Univ. II. B. 13: 43-68. 1953. [Mainly embryological studies o spp. of Breynia, Croton, Euphorbia, Putranjiva, and Sebastiania. | TuorNeE, R. F. Vascular plants previously unreported from Georgia. Castanea 16: 29-48. 1951. [Includes Aleurites, Manihot. Upuor, J. C. T. Certain minor rubber seem eee in the Western Hemi- sphere during times of emergency. Pp. 201, 202 in F, VERDOORN, ed., Plants and plant science in Latin America. wadice Mass. 1945. [Discusses spp. of Sapium, Cnidoscolus, and Euphorbia. | Ventura, M. Sulla poliembrionia di Mallotus japonicus Muell. Arg. Ann. Bot. Roma 20: 568-578. pls. 15, 16. 1934. [Embryo sac 16-nucleate ios type. | . Nuovo contre alla ee delle ee aera ge 22: 42-52. sion et systématique. Trav. Inst. Sci. Chérifien 6: i-xx, 1-217. p Hen map. 1953. Deuxiéme Partie. Anatomie. Ibid. 19: i-xxix, 219-533. ue. H. E. Studies on pollination of Hevea brasiliensis in Puerto Rico. Science 113: 646-648. 1951. Studies on natural pollination of Hevea brasiliensis in Brazil. Ibid. 116: 474, 475. 1952. [Reports that major pollinating agents are Heleid midges. Wexsster, G. L. The status of Agyneia and Glochidion. Taxon 9: 25, 26. 1960. revision of the genus Meineckia. Acta Bot. Neerl. 14: 323-365. tis. Cytotaxonomic studies in the Euphorbiaceae, subtribe Phyllanthinae. Am. Jour. Bot. 49: 14-1 “i “196 &K LLER. The genus Reverchonia (Euphorbiaceae). Rhodora 65: 193-207. 1963. [Discusses systematic importance of ovular morphology in subfam. Phyllanthoideae. | Wueeter, L. C. A miscellany of New World a [I.] Contr. Gray Herb, 124: 35-42; II. Ibid. 127: 48-77. pls. 3, 4. 1939 322 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 . Dichapetalaceae et Euphorbiaceae novae. Proc. Biol. Soc. Wash. 53: 7-11. 1940. Wiuiams, L. Woods of northeastern Peru. Publ. Field Mus. Bot. 15: 1-587. 1936. [{Euphorbiaceae, 264-283. | ZIMMERMANN, W. G., G. Hect, & H. Becer. Euphorbiaceae. /n: HEct, I Fl. Mittel-Europa 5(1): 113-193. pls. 177, 178. 1923. ie cake Euphorbia; includes original morphological observations by Zimmermann. | Kev TO THE GENERA OF EUPHORBIACEAE A. Ovules paired in each locule of the ovary; sap never milky; leaves not glandular; flowers axillary, solitary or in glomerules, not in spikes; pollen (in local taxa) mostly tricolporate (Subfam. Phyllanthoideae). Disc in male flower extrastaminal; res rs with or without petals; fruit dehiscent, or at least with more than ls C. Plants dioecious; petals present; btcoilaih gynoecium evident in male D. Male flowers subsessile; pollen a short colpi; trees or arbo- rescent shrubs with leathery We cs wa pen es as 1. Savia. nee flowers pedicellate; pollen with elongated colpi; Ee erbs, leaves with thinner texture. ............ Andrachne. : Plants monoecious (in need Aiea petals absent; cab flower without a rudimentary gynoe E. Disc usually present in ek sexes; male calyx open at anthesis, lobes not inflexed; seeds with dry coat, cna a ventral i es der) RapeeMen i aermRReGT. Seat NOG 1a alee I Seis SMR R See Phyllanthus. E. Disc absent; male calyx turbinate, lobes iene a rib seeds with somewhat fleshy coat, ventrally i invagina pies [ Breynia. | B. Disc in male flower intrastaminal, at least in 0 flowers apetalous; ae bidbhistegt seeds one per lociile: dioecious trees or shrubs with QO WO Pe ee Re . Drypetes. A. Oviiles solitary in each ae of the ovary; sap often milky or colored; leaves sometimes with petiolar or laminar glands; sulted ticslperae or e). inaperturate (Subfam. Pes Ph F. Flowers solitary, spicate, or in cymes, not in bisexual See a. Inflorescence dichasial or else stamens inflexed in the bud; petals present or else calyx petaloid; pollen spheroidal, senna porate or inaperturate [tricolporate in Hevea]. H. Stamens scarcely or not at all inflexed in the bud; trichomes simple and uniseriate, rarely stellate or lepidote; inflorescence dichasial. I. Perianth biseriate; pollen inaperturate; stem exudate colored, not a milky latex J. Calyx lobes free, imbricate; fruit eS gy caruncu- late; stipules more or less persistent. 5. Jatropha. fe Calyx lobes valvate, fused into a spathe: « feuit ied seeds ecarunculate; stipules caducous. ....... 6. Aleurites. I. ont uniseriate ; pollen aperturate; stems exuding milky atex w. K. Pollen tricolporate: seeds ecarunculate; leaves —— lence a aie i Lae ie BA: vea. | 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 323 K. Pollen porate; seeds carunculate; leaves often deeply lobed but usually not compound. L. Male disc intrastaminal; stamens free; male perianth usually yellowish, greenish, or purplish: stinging hairs OER Sette Orica, Nee, Prodi State 7. Mani - L. Male disc extrastaminal; stamens connate; male p anth white; stinging hairs present. .... 8. Cn Pbcitad H. Stamens more or less inflexed in the bud; trichomes often stellate or lepidote; inflorescence racemiform or spiciform; pollen inaper- turate. M. Fruit a dehiscent, 3-locular schizocarp; seeds carunculate. _. . 9. Croton. TUN ere Osetra hp ete ae Meh, ee oe aes . Crotonopsis. G. Inflorescence not clearly dichasial — racemiform or spiciform) ; petals absent or, if present, then male calyx valvate; dictinc various, mostly tricolporate, never verrucose. N. Petals eae. at least in male flower; seeds ecarunculate; calyx lobes v O. syria gynoecium present in male flower; trichomes simple or glandular (not malpighiaceous) ; leaves finely serrate, lateral veins distinctly parallel. .............. 1 aperonia. O. Rudimentary gynoecium absent; trichomes malpighiaceous (at least in part); leaves entire or coarsely serrate, veins not distinctly parallel. fied eee Soe eee Soa! 12. Argythamnia. N. Petals absent in the male flow P. Male calyx valvate; sap ae styles conspicuously papillate to laciniate (or, if not, then plants with stinging hairs Stamens free or connate, not branched or fasciculate: flower; veined, at most slightly lobed; stipules discrete, not united; inflorescence spiciform, racemiform, or capitulate; female owers proximal to the male R. Styles laciniate or pinnatifid; herbs or shrubs, never twining; stinging hairs absent; seeds usually caruncu- late. S. Leaves opposite; plants usually dioecious; bracts small; anthers not much elongated; heer flower with 2 elongated staminodia; carpels usually 13, “ul ercurialis, S. Leaves alternate; plants monoecious (in local taxa) ; female bracts enlarged: anthers elongated, vermi- ee female flower lacking staminodia; carpels usu- ally 14. Acalypha. R. sone Seu at most papillate; herbs, often twining, rmed with stinging hairs; seeds not carunculate. _. oir fee ee Ne ee ee £5; Tragia. Q. Stamens branched and fasciculate, anthers many (up to 1000) per flower; leaves palmately lobed; stipules fused into a circular sheath; inflorescence paniculate, normally with female flowers distal to the male; seeds carunculate. 16. Ricinus. 324 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 P. Male calyx imbricate or reduced; styles undivided, not laciniate or pinnatifid; female bracts not accre nay ye sap often milky; inflorescences spiciform or racem T. Styles not connate into a esially, ‘enlarged column; male pitesiaes not cig and conelike; carpels 2 or 3 (ex- cept in Hippoma U. Seeds ae siamese dry; fruit capsular. V. Floral bracts conspicuously biglandular at base, not displaced from rachis of inflorescence; ovary not conspicuously stipitate. W. Columella not 3-horned at base; male calyx lobes and stamens 3; female flowers distinctly ror CUE Es adie rscece Rare) a7 ge eae oP 17. Sebastiania. ’,Columella with a 3-horned gynobase persistent after dehiscence of fruit; male calyx lobes and renga usually 2; female flower sessile or nearly OT sees een ale Ma TS Reade eS Ge Nbr 19. Stillingia. Nn: Floral bracts not glandular at base, adnate to lateral axes of inflorescence (and thus displaced from rachis); ovary conspicuously cect fae rears eae ere eet Vit errs Es, aces 8. Gymnanthes. U. Seeds ecarunculate, either seed coat or oo wall fleshy. X. Fruit thin walled, capsular; seed coat fleshy (‘aril- late’); ovary 3- locular; styles connate to about the WM ee, cia Nel Faget 20. Sapium. X. Fruit drupaceous; seed coat dry; ovary 6—10-locular; styles connate only near the base. .. 21. Hippomane. T. Female flower with umbraculiform stigmatic disc terminat- ing a long stylar column; male flowers aggregated into a fleshy pedunculate conelike structure; carpels more than 5. = ura. F. Flowers aggregated into a usually bisexual cyathium (pseudanthium), usually with one central female flower surrounded by 4 or 5 male mono- chasia; glands Ponies of cyathium usually conspicuous, often with ages petaloid append Y. Cyathi a more or less actinomorphic, at least not conspicuously spurred; styles mostly united only below the middle and stems not NB N markedly succulent (in native taxa). . Leaves alternate or opposite, if opposite, then mot inequilateral at base and with chlorenchyma-sheathed v veins; main axis not aborting, —- monopodial at least below; stipules often reduced or abse 22. Euphorbia. . Leaves entirely pene usually distinctly oe at base, stipulate, with chlore enchyma-sheathed veins; main axis aborting just above the cotyledons, branching sympodial throughout vo -G ml OQ) ee 28 & N _ ag ic} 3 3 E: Re. = aos ‘Oo yi ~ 3 om wm fe ae a a @ 5 Z. ae =7 ry 5 bo) Q ° Bo Z. a c 3S .c nm w 1967 | WEBSTER, GENERA OF EUPHORBIACEAE $25. Subfam. PHYLLANTHOIDEAE Pax Tribe PHYLLANTHEAE [Dumort. |] Subtribe Andrachninae Muell. Arg., ““Andrachneae” 1. Savia Willdenow, Linn. Sp. Pl. 4(2): 771. 1806. Dioecious shrubs or small trees. Leaves alternate, entire, chartaceous to coriaceous, short-petiolate, stipulate. Flowers axillary, the male in dense glomerules, the female few or solitary. Calyx ordinarily 5-parted, lobes imbricate; petals 5, much smaller than sepals [large and conspicuous in some Malagasian taxa]. Male flower: stamens 5, free; anthers introrse, opening longitudinally; pollen subglobose, reticulate, colpi short, endocol- pus with diffuse or rounded ends; rudimentary gynoecium 3-lobed. Female flower: disc annular; carpels 3; styles more or less free, bifid, branches slender; ovary glabrous or pubescent; ovules 2 in each locule, anatropous. Fruit capsular; columella persistent; seeds 1 (2) per locule, smooth, chalaza ventral; endosperm copious; embryo straight; cotyledons broad, plane, much longer than the radicle. Type species: S. sessiliflora (Sw.) Willd. (Named in honor of Gaetano Savi, 1769-1844, professor at Pisa.) A genus of 20 to 25 species with a remarkable tricentric distribution: West Indies, southern Brazil, and Madagascar. The circumscription adopted here is approximately that of Bentham and is also near that of Pax and Hoffmann, except that their Savia phyllanthoides is returned to Andrachne. Our single representative, Savia bahamensis Britton, is a common shrub of coastal and lowland thickets from the Florida Keys south and east to the Bahamas, Caicos Islands, Cuba, Cayman Islands, and Jamaica. It is very similar in appearance to the common Cuban species S. erythroxyloides Griseb., from which it differs in its glabrous (rather than sericeous ) ovary and its more strongly reticulate leaves (the tertiary veinlets prominent beneath, whereas they are obscure in S. erythroxyloides). In these charac- ters it agrees more closely with the Cuban S. clusiifolia Griseb., but that species differs in its more cuneate-truncate leaves and larger seeds. In the Keys, Savia bahamensis has been little collected except on Big Pine Key, although records are available from Little Torch Key, No Name Key, and Key Largo. Possibly the species is often overlooked, as the flowers (which seem to appear mainly in April and May) are relatively inconspicuous. Chromosome counts of the plant would be most desirable since not a single species of Savia has ever been reported on cytologically. According to the classification proposed by Urban, Savia bahamensis would fit into sect. HETEROSAVIA Urb., characterized by fruits with 2-seeded locules and an embryo with the radicle about half as long as the cotyledons. In this circumscription, sect. SAVIA is monotypic, containing only S. ses- siliflora, which has capsules with 1-seeded locules (the second ovule abor- tive in development). This difference is not absolute, however, since the 326 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 seeds in various species of sect. HETEROSAVIA are often of different sizes and in some instances are probably not viable. However, other characters (e.g., leaf texture, petal size) are correlated with this difference, and Urban’s proposed distinction may well be valid. The biseriate perianth, unspecialized pollen, and generalized vegetative characteristics of Savia mark it as one of the more primitive taxa of Euphorbiaceae. Although Rothdauscher reported vessels with simple per- forations in S. sessiliflora, Solereder found scalariform perforations in other species. Our own species, S. bahamensis, appears to be rather highly specialized, since it has simple vessel perforations and mostly uniseriate rays. Comparisons with the Madagascar species are much needed. Kohler has noted that Savia is palynologically heterogeneous even after some ex- traneous taxa brought in by Pax are excluded. Most of the primitive taxa in subtribe Andrachninae have oblate spheroidal grains, as found in the West Indian S. sessiliflora and S. andringitrana Leandri from Madagascar. However, in S. Danguyana Leandri of sect. PETALopIScUs and in species of the West Indian HETEROSAVIA (e.g., S. pre the grains are prolate and have a much finer ornamentatio Savia cannot be delimited satisfactorily ns the Madagascar ai are carefully studied. It is notable that these plants differ from the Wes Indian ones in being monoecious and having a more conspicuous ae further study may show that they should be segregated into a separate genus, Petalodiscus, as was done by Pax in 1890. This would make Savia an entirely American group but would not affect the obvious affinity be- tween the West Indian and Malagasian taxa. The evidence from pollen characters, in fact, seems contradictory, since it suggests two groups of affinity, each represented in both the West Indies and in Madagascar. Examination of seeds in the Madagascar taxa might possibly help to resolve the impasse. Leandri has reported exalbuminous seeds in the Madagascar species §. Bojeriana Baill., whereas in all West Indian plants examined copious endosperm is present. If a correlation between seed and pollen classes can be established, the circumscription of the subgeneric taxa in Savia can at last be fixed. Whatever the circumscription of Savia may prove to be, the genus cer- tainly belongs in the taxonomic group which includes woody Old World genera such as Blotia and Wielandia. Its only near relative in the New World (and a rather distant one at that) would appear to be Astrocasia, which differs strikingly in appearance due to its long-petiolate leaves and more conspicuously petaliferous flowers; the extrorsely dehiscent stamens adnate to the vestigial gynoecium in Astrocasia furnish an additional tech- nical distinction. The genus Andrachne is difficult to distinguish from Savia on a world-wide basis because of the variability of both taxa, al- though Kohler has shown that the pollen is quite different, and at least the typical species of Andrachne have hemitropous ovules. The local repre- sentatives of the two genera may be easily distinguished by the larger and thicker leaves, subsessile male flowers, and solitary seeds of Savia. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 327 REFERENCES: Under family references see BAILLON (1858), GAUCHER, KOHLER, LEANDRI, Punt, REcorp, ROTHDAUSCHER, and SOLEREDER. Pax, F., & K. HorrmMann. Eup Phyllanthoid Phyllantheae—Wie landiinae. Pflanzenreich IV. 147-IX(Heft 81): 180-189. 1922. Ursan, I. Nova genera et species II. Symb. Antill. 3: 280-420. 1920. [ De- scription of sect. Heterosavia, 284. | ———. Sertum antillarum XXX. Repert. Sp. Nov. 28: 209-236. 1930. [De- scribes several spp. of Savia and discusses relationships of S. bahamensis, 209-212.] 2. Andrachne Linnaeus, Sp. Pl. 2: 1014. 1753; Gen. Pl. ed. 5. 444. 1754 Herbs or subshrubs. Leaves alternate, stipulate, petioles abbreviated to as long as the blade; blades thin, entire, often small. Plants dioecious [in our representative, most other species monoecious]; flowers in axillary clusters, the female often solitary. Flowers petaliferous, but petals often reduced or (in the female flower) rudimentary; disc usually segmented (except in our representative). Male flower: stamens 5, opposite the calyx lobes; filaments free [sometimes united]; anthers introrse or laterally dehiscent; pollen more or less prolate, 3-colporate, colpi elongate, endocol- pus pointed at ends; vestigial gynoecium usually lobed or divided. Female flower: carpels 3; stigmas capitate; styles more or less free, spreading, bifid; ovary glabrous or pubescent; ovules 2 in each locule, anatropous [or transitional to hemitropous]. Fruit capsular; columella persistent. Seeds usually 2 in each mericarp, smooth or roughened; endosperm copi- ous; embryo straight, cotyledons not folded, broader than the radicle. 2n = 24, 26. (Including Lepidanthus Nutt.) LecToTyPe spEciEs: A. Telephioides L.; see Small in Britton & Brown, Illus. Fl. No. U. S. ed. 2. 2: 453. 1913. (Name from Greek, andrachne, supposedly applied to Portulaca oleracea L. by ancient authors. ) As here delimited in the sense of Mueller, Andrachne is a small but distinctly heterogeneous genus of approximately 15 species widely scat- tered in both temperate and tropical regions of the Old World and New World. Only four species are known in North America; a single one, 4. phyllanthoides (Nutt.) Coulter, 2m = 26, enters our area. ‘ope ee Until recently, this species was unknown east of the Mississippi, the outlying localities being in southern Missouri (Texas and Shannon coun- ties) and central Arkansas (Garland, Saline, and Hot Springs counties). In 1963, Mrs. Blanche Dean discovered a colony along the banks of Scarum Creek in Blount County, Alabama; according to Mrs. Dean the plant here grows in deep sand among boulders in association with Rhus radicans, Salix nigra, and Amorpha fruticosa. In the western parts of its range (Texas and Oklahoma), A. pAyllanthoides appears to be an obligate calciphile, usually growing in cracks in limestone; at the same time, how- ever, it does seem to prefer creekbed localities which are periodically flooded. In central Arkansas, on the other hand, Demaree (on labels ) reports finding it on steep shale or novaculite slopes. 328 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 The discovery of the Blount County station, over 300 miles east of the nearest population, adds another interesting relict to the ones already known from central and northern Alabama and recalls the extreme restric- tion of Croton alabamensis in the Warrior and Cahaba River basins. The Alabama plants seem taxonomically indistinguishable from those farther west, although they may prove to have certain minor differences (viz., possibly smaller petals) when the population is better sampled. No other species of this family in the Southeast shows such an outstanding dis- junction, although the bicentric population (Texas-Florida) of Phyllan- thus abnormis and the vicariant pair P. polygonoides-platylepis (Texas- Louisiana and northwestern Florida) offer somewhat analogous distribu- tion patterns. The relatively slight morphological differentiation which has occurred in all three instances suggests that the disjunctions may date only to the Pleistocene. Within the genus Andrachne, A. phyllanthoides appears to be quite isolated, except for its western vicariant A. arida (Warnock & Johnston) Webs ter.5 This latter plant, which is known Sal from a few desert local- ities in trans-Pecos Texas and Coahuila, obviously resembles A. phyllan- thoides but differs in its shrubbier habit, smaller more rigid leaves, and shorter pedicels. As Warnock and Johnston suggest, the two species appear to represent the vicarious offspring of an originally widespread and enim population the range of which has been dissected by climatic change. This species-pair (A. phyllanthoides-arida) occupies a taxonomically isolated position within Andrachne. The smooth, anatropous seeds and dioecious inflorescences suggest Savia, where, in fact, both species have been placed by most recent workers. However, Punt and Kohler have shown that the pollen grains of A. phyllanthoides, with elongate colpi and large, sharply defined ora, resemble those of typical species of Andrachne much more than they do any species of Savia. Furthermore, in habit and in gross flower structure both species show a much greater resemblance to the Asiatic species of Andrachne sect. ARACHNE Endl. than to any species of Savia. Some contemporary workers (e.g., Hurusawa, Pojarkova) accept Arachne as a distinct genus from Andrachne (sensu stricto) and would pre- sumably place A. phyllanthoides in Arachne. However, in most charac- teristics the latter taxon is so close to typical representatives of Andrachne that it would be difficult and inconvenient to recognize two genera. In tackling this problem in the future, notice should be taken of the rela- tively neglected feature of ovule configuration: at least some taxa of sect. ARACHNE have anatropous ovules, whereas in A. Telephioides (§ AN- “ _ achne arida (Warnock & — Webster, comb. nov. Savia arida nock & Johnston, Southw. Nat. 5: 3. 1960. This new tecnienedane is rather em- barca ssing to the author, since it was He least partly on his advice that Warnock an nd Johnston described their plant as a Savia rather than as an Andrachne. Since then, ee accumulating evidence shows that their plant must be excluded from Savia, e and the most Preaserpcah course is to place it in Andrachne, rather than in som segregate — 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 329 DRACHNE) they are hemitropous. However, it is not yet certain whether this difference will hold for other species assigned to Andrachne. For the time being, therefore, it appears that the best disposition of the two species of the United States is still that of Mueller, who placed A. phyl- lanthoides in Andrachne as the only representative of section PHYLLAN- THOPSIs (Scheele) Muell. Arg., characterized by the undissected male disc, somewhat woody habit, and deeply divided vestigial gynoecium. According to both Kohler and Punt, Andrachne is palynologically sim- ilar to the Old World genus Actephila; an affinity between these two genera was pointed out long ago by Baillon. However, Actephila differs markedly from all species of Andrachne in its seeds, which are solitary (by abortion) in each locule and nearly or quite without endosperm, the cotyle- dons more or less folded. Through sects. ARACHNE and PHYLLANTHOPSIS, Andrachne seems more closely related to Savia, of which it may be re- garded as the herbaceous derivative (cf. Savia for generic distinctions). In another direction, Andrachne stands in approximately an sey position to various apetalous genera of subtribe Phyllanthinae. For ex- ample, A. ovalis of South Africa has many of the attributes of the ete. thetical ancestor of both Meineckia and Chascotheca, while A. Tele- phioides and its relatives show a possible significant resemblance to some nearctic herbaceous species of Phyllanthus subg. Isoctapus. From all of these genera Andrachne may be distinguished by its diminutive habit, delicate petaliferous flowers, characteristic pollen grains, and paired seeds with copious endosperm. REFERENCES: Under yas seach see ASSAILLY, CroizaT (1943), GAUCHER, HurUSAWA, KOHLER, MICHAELIS, PUNT, RoTHDAUSCHER, and VINDT Crark, RK. C. aw phyllanthoides (Nuttall) Muell. on the Cumberland Plateau of Alabama. Castanea 32: 73, 74. April 1967. [Reports a visit in Aug. 1966 to the Blount County station. | Jounston, M. C. Savia arida in Coahuila, Mexico. Southwest. Nat. 7: 80. liad A Ph llantheae An PAX Echt FMA Phyl An “dachomnae: ka “Wy. eee abe 169-179. 1922. Poyarkova, A. I. Contribution a la systématique des représentants du genre Andrachne s. 1. habitant le Caucase et la partie de la région méditerranéene. (In Russian.) Bot. Zhur. 25: 341-348. Reese, G. Uber die Polyploidiespecktren in der nordsaharischen Wiistenflora. Flora 144: 598-634. 1957. [Reports chromosome number of A. Tele- 1 Se eteg H., & M. C. Jounston. The genus Savia (Euphorbiaceae) in extreme western Texas. Southwest. Nat. 5: 1-6. 1960. Subtribe Drypetinae Griseb., “Drypeteae” 3. Drypetes Vahl, Eclog. Am. 3: 49. 1810. Dioecious trees or shrubs, usually with dense wood. Leaves alternate, short-petiolate, stipulate, blades often leathery, entire or sharply toothed, 330 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Flowers in dense axillary clusters, sometimes almost sessile. Flowers apetalous; calyx in both sexes of 4 or 5 (rarely 6 or 7) more or less deciduous imbricate sepals. Male flower: disc intrastaminal, marginal lobes sometimes projecting between stamens; stamens 3-12 (rarely up to 50), filaments free, anthers basifixed and extrorse to introrse; pollen pro- late, tricolporate, tectate, with large endocolpus; rudimentary gynoecium present or absent. Female flower: disc cupuliform; ovary of 1 or 2 [very rarely 3 or 4] carpels; styles nearly obsolete, the more or less dilated stig- mas nearly sessile atop the ovary; ovules 2 in each locule, anatropous. Fruit indehiscent, becoming more or less drupaceous, exocarp fleshy or leathery, endocarp crustaceous or bony; seeds usually solitary in each locule, ecarunculate, testa smooth; endosperm copious; embryo straight, cotyledons broad. (Including Cyclostemon and Hemicyclia.) TYPE sPecIES: D. glauca Vahl. (Name from Greek, dryppa, overripe olive, in allusion to the fruit of the type species.) — GUIANA PLUM, WHITE WOOD. Perhaps 150 species of circumtropical distribution, the vast majority in World species were assigned to Cyclostemon and Hemicyclia. However, Pax and Hoffmann (1922) appear to have combined these taxa correctly into a single genus characterized by an indehiscent fruit with reduced seed number, abbreviated styles, and an intrastaminal male disc. Hurusawa (1954) has even combined Putranjiva and Drypetes, and it must be admitted that the former scarcely differs in any essential character, except its lack of an intrastaminal disc Approximately eight or nine species of Drypetes are known from the West Indies and Central America. Two, each belonging to a different section, have entered the United States. Representing sect. OLIGANDRAE Pax and Hoffm. is D. lateriflora (Sw.) Urb., which normally has only four stamens isomerous with a four-lobed calyx, and a two-locular ovary which develops into a thin-walled, subglobose, scarcely fleshy fruit. It is fairly common in hammocks throughout the Florida Keys and extends northward on the Atlantic coast of Florida as far as Brevard County. Since all the other species ® of sect. OLIGANDRAE are restricted to the Old World (mainly Africa), the relationships of D. lateriflora provide an interesting problem. Further study may show that it should be relegated to a different section. Section Drypetes (§ Hemicyclia of Pax & Hoffmann), with 30-40 species, includes all the remaining American species of Drypetes, among them our other local species, D. diversifolia Krug & Urb. (including D. keyensis Krug & Urb.). This plant differs greatly from D. lateriflora in having a five-lobed calyx, eight to ten stamens, and a one-locular ovary which develops into an oblong drupe with thick exocarp and endocarp. “The only other American species placed in sect. OLIGANDRAE by Pax & Hoffmann is the se Hs intel Cuban Drypetes triplinervia Muell. Arg. (DC. Prodr. 15(2): 456. 1866), which must be excluded from Dr rypetes because of its dehiscent fruit. It ete properly dassitod as Chascotheca Aplnlinéevia (Muell. Arg.) Webster, comb. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 331 The species can ordinarily be easily distinguished in the vegetative state since D. lateriflora has thinner leaves which are abruptly pointed at the tip, whereas the leaves of D. diversifolia are thicker, more rigid, and mostly rounded to obtuse (or at least not abruptly narrowed) at the tip. Further- more, D. diversifolia shows an interesting leaf polymorphism. Partic- ularly on seedlings and sprout-shoots, the leaves may be conspicuously spinulose-serrate, contrasting greatly with the “normal” entire leaves; some individual branches of a mature tree with entire leaves may show the toothed kind. No such variation occurs in D. lateriflora, which always has entire leaves. Longwood reports that D. Brownii Standl., bullhoof, of British Hon- duras produces strong, hard timber with some of the properties of English oak; it is used for railway ties, rafters, beams, and other heavy construc- tion, No studies of the timber quality of the two Florida species have been reported, but it seems doubtful that they can ever be of any com- mercial value because of their small size in our area (trunks not over one foot in diameter). The dense wood of Drypetes, which makes it potentially valuable for lumber, is correlated with characteristics of considerable systematic interest. The vessels have scalariform perforations, xylem parenchyma is abundant, and fibers are thick-walled. Janssonius grouped Drypetes with the Old World genera Aporosa and Baccaurea on the basis of these characteristics, and Metcalfe and Chalk associated it with the “4 porosa type,” from which it is divergent mainly in having rather narrow rays. Anatomically, Drypetes appears to be the most primitive of our taxa of Euphorbiaceae, and does not show any close relationship to other local genera of Phyllanthoideae except Savia. Another indication of the taxonomic isolation of Drypetes within the Phyllanthoideae is the report by Mangenot & Mangenot of 2m = 40 in an African species, D. mottikoro Leandri. Chromosome counts are unreported for any American species, but would be of unusual interest, since the one known count suggests a base number (x = 10) different from that in most other Phyllanthoideae, which have x = 13. : Palynologically, Drypetes appears to be heterogeneous, for Kohler has recognized two different pollen types within sect. DRYPETES. The closely related genus Putranjiva, which Hurusawa has combined with Drypetes, has very similar pollen, and Putranjiva Roxburghii also has an = 40 in agreement with the single count in Drypetes. Possibly connecting Dry- petes with taxa in subtribe Andrachninae is Lingelsheimia, which Léonard has shown to differ from Drypetes by its capsular fruit. Perhaps the most their monoecious inflorescences, an sy because these are porate and echinulate, Ko 0 Petalostigma to the subfamily Oldfieldioideae. Although Kohler regards their pollen as derived within the subfamily, the suggestive resemblance to Drypetes warrants further analysis, and it seems possible that Drypetes 332 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 may be the closest surviving taxon to the ancestral stem of the Oldfield- ioideae. REFERENCES: Under family references see DeHay, GAUCHER, HurUSAWA, KOHLER, LEANDRI, Loncwoop, MicwaeELis, Punt, Recorp, ROTHDAUSCHER, and SMITH. Keno, H. New or critical Euphorbiaceae from eastern Asia. Jour. Wash. Acad. Sci. 41: 200-205. 1951. [Describes Liodendron, related to Drypetes and Putranjiva. | LEoNARD, J. Notulae systematicae XXXIII. Sur les limites entre les genres Drypetes Vahl et Lingelsheimia Pax (Euphorbiacées.) Bull, Jard. Bot. Bruxelles 32: 513-516. 1962. Pax, F., & K. HorrMann. Euphorbi J petinae. Pflanzenreich IV. 147-XV(Heft 81): 227-280. 1922. SARGENT, C. S. Drypetes. Silva N. Am. 7: 23-28. pls. 307, 308. 1895. Urpan, I. Additamenta ad cognitionem florae Indiae occidentalis. Bot. Jahrb. 15: 286-361. 1892. [Revision of West Indian spp. of Drypetes, 351-357. ] Dill Pe egies | DL. +} ip ee i’ cee Subtribe Phyllanthinae [Muell. Arg., “Phyllantheae”’ | 4. Phyllanthus Linnaeus, Sp. Pl. 2: 981. 1753; Gen. Pl. ed. 5. 422. 1754. Trees, shrubs, or herbs; stems not succulent, often entirely glabrous; lateral axes in some taxa deciduous, subtended by scale-like leaves. Leaves alternate (spiral or distichous), stipulate, pinnately veined, entire and unlobed; petiole short. Plants monoecious or, less commonly, dioe- cious; inflorescences axillary, cymose, the cymes usually highly condensed and sometimes reduced to solitary flowers. Flowers apetalous; calyx synsepalous, of 4-6 imbricate or decussate lobes; disc usually evident, segmented or cupular. Male flower: disc extrastaminal, usually segmented; stamens (2)3-5[- i ing longitudinally or horizontally; pollen prolate to globose, 3- or 4-col- porate [pancolporate or porate in some exotic taxa]; vestigial gynoecium absent [very rarely present|. Female flower: disc segmented or more often patelliform to cupular [rarely absent]; staminodia absent [very rarely present]; carpels usually 3 [rarely 2 or 4-12]; styles free or basally connate, more or less bifid [rarely entire], sometimes multifid; ovules 2 in each locule, hemitropous, contact with obturator; embryo sac normal (Polygonum) type. Fruit usually capsular, explosively dehiscent [less commonly baccate or drupa- ceous|; mericarps separating from a more or less persistent columella; seeds usually 2 in each locule [rarely only one maturing]; testa dry and usually thin [rarely thickened and bony], not ventrally invaginated; endo- sperm copious; embryo straight or slightly curved, cotyledons broader than and about as long as the radicle. (Including Cicca L., Emblica Gaertn.., Xylophylla L.) Lectotype spEctrs: P. Niruri L.; see Small in Britton ? & Brown, Illus. Fl. No. U. S. ed. 2. 2: 453. 1913. (Name from 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 333 Greek, phyllon, leaf, and anthos, flower, in allusion to the production of flowers on specialized leaf-like lateral branches.) Approximately 750 species may be assigned to this highly diversified genus which is predominantly represented in the Old World tropics. Over 200 species have been reported from America, mostly from Brazil and the West Indies. Less than a dozen attain temperate latitudes, and the genus is unknown in Europe and temperate Pacific America. In the southeastern United States, Phyllanthus is represented by eight native and two natural- ized species belonging to five sections in three subgenera. Subgenus Isoctapus Webster, which includes about 70 herbaceous species with alternate phyllotaxy and unspecialized ramification patterns, has three Southeastern species. Belonging to section PARAPHYLLANTHUS Muell. Arg. because of their spiral phyllotaxy and dissected floral disc are two closely related species, Phyllanthus polygonoides Nutt. ex Spreng., 2n = 16, and P. platylepis Small. A mainly calciphilous plant of lime- stone regions in Texas and Oklahoma, P. polygonoides barely enters our area in prairies of western Louisiana. The very similar P. plat ylepis, which differs mainly in its more or less rhizomatous habit and larger fruits and seeds, is confined to a small region in the hammocks of northwestern Florida (Dixie and Taylor counties). It shows an even more striking resemblance to P. Liebmannianus Muell. Arg., of coastal swamps in Vera- cruz, Mexico, These three species appear to be the vicariant relicts of an ancestral population which probably had a much more nearly continuous distribution along the perimeter of the Gulf of Mexico during the late Tertiary. cae The third species of subg. Isociapus, Phyllanthus caroliniensis Walt., belongs to sect. Loxopopium Webster by virtue of its distichous phyllotaxy and very short, geniculate fruiting pedicels. This, the most widespread American Phyllanthus, is remarkable in being one of the few Euphorbia- ceae which is distributed from temperate holarctic regions, south through the tropics into temperate southern latitudes. In the Southeast it is repre- sented by ssp. caroliniensis, with smooth stems and an entire or angled fe- male disc, occurring throughout our area except in southern Florida, where it is replaced by ssp. saxicola (Small) Webster. The latter taxon, mainly West Indian, has established a beach-head on the mainland only in the Florida Keys and the Dade County pinelands. An extralimital taxon, ssp. guianensis (Kl.) Webster, has been shown to have 2m = 36, so that at least two different basic chromosome numbers (x = 8, 9) appear to be present in subg. ISOCLADUS. ; Subgenus K1rGANELIA (Juss. ) Webster, a primarily Old World group, is represented by a single naturalized species, Phyllanthus tenellus Roxb., 2n to sect. PENTANDRA Webster,’ which includes about ten herbaceous Africa suffruticosae; ramifica _Sectionis é calycis lobis 5, disci segmentis distinctis ; bae monoicae annuae vel 7 Phyllanthus sect. Pentandra Webster, sect. nov. Herbae r ae é tione more sectionis Phyllanthi; pedicellis capillaribus ; flore staminibus 5, liberis; granis pollinis 334 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 species with thin-walled capsular fruits, undilated style-branches, and three- or four-colporate pollen grains. The section is significant phyloge- netically because most of its taxa have precisely the habit and appearance of species of subg. PHYLLANTHUS, from which they scarcely differ in anything more than the five-merous rather than three-merous androecium. Since P. tenellus is the only herbaceous diploid species with phyllanthoid branching, it and closely related taxa such as P. capillaris Schum. may be regarded as the nearest living equivalents of the taxa ancestral to subg. PHYLLANTHUS. Apparently the earliest collection of Phyllanthus tenellus in the United States was made by Hunnewell in Orange County, Florida, in 1924. Since then, however, its spread appears to have been rapid, for it is now known from a considerable number of localities between Sumter County, South Carolina, and Key West. The plant is an aggressive greenhouse weed due to its rapid flowering and effective ballistic seed dispersal; it seems to show a decided preference for sandy soil, and hence is to be found almost ubiquitously in Citrus groves in parts of central Florida. Although it has often been confused with native species of subg. PHYLLANTHUS, it is easily distinguished by its five free stamens and long capillary fruiting pedicels. Subgenus PHYLLANTHUs, which includes well over 100 species in both hemispheres, comprises herbaceous or suffruticose plants with only two or three stamens and specialized “phyllanthoid” branching; the leaves on the main stem are reduced to scales which subtend deciduous leafy floriferous branchlets. In the Southeast there are five species of this subgenus, two or three of them native. The only representative of sect. URINARIA Webster is Phyllanthus Urinaria L., 2n = 52, a weed of Old World origin which has been introduced into a number of localities in the United States during the past 25 years. Although it spreads much less contagiously than P. tenellus, it has been found in Texas, Louisiana, and Alabama, and doubt- less will be encountered in other Gulf states. While P. Urinaria is occa- sionally confused with taxa of the next section, it differs markedly in its verrucose ovary, sessile female flowers, hispidulous leaf margins, and trans- versely (rather than longitudinally) ribbed seeds. Section PHYLLANTHUs includes our remaining native species of subg. PHYLLANTHUS, characterized by pedicellate flowers, smooth ovaries, and verruculose or longitudinally ribbed or striate seeds. Subsection NIRURI Webster, with stamens partially or entirely free, slender scale-leaves, and verruculose seeds, is represented in the United States only by a relict population of Phyllanthus Niruri L. in east-central Texas, and does not enter our area. Subsection Swarrzianr Webster, comprising mostly annual plants with connate stamens, broader scale-leaves, and striate or ribbed seeds, includes four local species. Phyllanthus amarus Schum., 2n = 52, the most abundant and widespread weed in the genus, is common in southern Florida (Dade and Monroe counties) and has been collected subglobosis 3- vel 4-colporatis; flore Q disco patelliforme, ovario loculis 3, stylis bifidis; fructo capsulare, seminibus asperulis. Typus: Phyllanthus pentandrus Schum. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 335 as far north as Gadsden County. Because of its nearly ubiquitous distri- bution, it is impossible to be certain that this species is native to Florida. The closely related P. abnormis Baill., which is definitely native, stands extremely close to P. amarus morphologically, but can be distinguished by its larger seeds, four-parted male flowers, more deeply dissected female disc, and thicker, more reddish stems. Unlike P. amarus, it shows a dis- tinct preference for sandy soils and has a strikingly disjunct range: the western population extends from northeastern Mexico and western Texas to near Houston, while the eastern population is confined to peninsular Florida, from Cedar Key to Dade County. At least one additional species of the subsection, P. fraternus Webster, has recently been discovered in Louisiana. It is also very similar in appearance to P. amarus, but differs in having the male and female flowers segregated at different axils (in P. amarus and P, abnormis a male flower is paired with a female at all distal axils on the flowering branchlet). Subsection PENTAPHYLLI Webster, a West Indian taxon of about 13 or 14 species, characterized by the perennial rootstock and blackened indurate scale-leaves on the main stems, has only a single species in the Southeast: Phyllanthus pentaphyllus Wright ex Griseb., 2n = 52. This is the most widespread species in the subsection, with populations extending from Florida and the Bahamas south to Guadeloupe and Curacao. The Florida population belongs to ssp. pentaphyllus by virtue of its deeply lobed female disc, and it has been distinguished as var. floridanus Webster be- cause of its primarily dioecious inflorescences (in contrast to the pre- dominantly monoecious condition in var. pentaphyllus of the Bahamas and Cuba). The recent report of P. pentaphyllus from South Carolina is erroneous, as it was based on a mislabelled specimen; in the United States the species is confined to limestone areas (mainly pinelands) in Dade and Monroe counties, Florida. Several other species of Phyllanthus have been reported from Florida, but these reports lack confirmation by specimens. Small cited the com- monly cultivated P. acidus (L.) Skeels, of subg. Crcca (L.) Webster, from southern Florida, but there is thus far no evidence that it has become naturalized. Small also reported from Key West P. epiphyllanthus Lot subg. XYLOPHYLLA (L.) Pers. sect. XYLOPHYLLA, a taxon outstanding be- cause of the branchlets transformed to phylloclades and with concomitant reduction of the leaves to scales. His identification was erroneous, for the specimen from Key West collected long ago by Blodgett, as well as Small’s own collection made in 1921, represent P. angustifolius (Sw.) Sw., a species otherwise known only from Jamaica and the Cayman Islands. Since the regions closest to Key West (Cuba and the Bahamas) have only P. epiphyllanthus, it seems highly unlikely that P. angustifolius could have occurred on Key West as a native plant. Recent attempts to locate the species on Key West have been unsuccessful, and it seems most reasonable to conclude that P. angustifolius, which was widely planted in Caribbean gardens a century and more ago, became temporarily natural- ized and then died out again. 336 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 The remarkable amplitude of variation in both vegetative and floral characteristics evident in Phyllanthus has led various investigators to propose over 50 segregate genera. However, it has not been demonstrated that the genus is polyphyletic as presently constituted, and there are few large groups of species within it which could be defined by clear-cut diagnostic features. Chromosomal, palynological, and gross morphological characters, although displaying a protean diversity, fail to show the sort of correlation on which segregate genera could easily be justified. For example, the periporate pollen grains of subg. XYLOPHYLLA are correlated with the specialized branching pattern, but in the American species placed in subg. BotryANTHUS Webster this type of pollen is associated with definitely nonphyllanthoid branching. Until additional compelling evi- dence is unearthed, therefore, it seems expedient to maintain the circum- scription of Phyllanthus as established by Mueller and modified by Hooker (when he recognized Glochidion as a separate genus). The subgeneric classification of Phyllanthus is based primarily on char- acters of branching pattern and pollen, in addition to the usual floral characteristics. In the vast majority of species, an unusual pattern of vegetative development occurs which has been designated as “phyllan- thoid branching.” In plants of this kind, the main axis produces ordinary leaves only at the first few nodes above the cotyledons; distal nodes have the leaves reduced to scale-like cataphylls subtending deciduous leafy floriferous axes. There is thus a striking division of labor among the por- tions of the axial system of the plant: permanent axes produce only scale- like leaves and branchlets (never flowers) and are capable of indefinite growth, whereas the lateral axes produce both the leafy and reproductive portions of the plant and are of more or less limited growth (although the terminal bud of the branchlet does not necessarily abort). In a relatively few species, including most of those of temperate regions, such as P. poly- gonoides, there is no such vegetative specialization; flowers and leaves may be produced indiscriminately on both main and lateral axes. “These microspores, designated as “synrugoidorate” by Erdtman_ because the pores are connected by marginate streaks (presumably homologous with the colpl 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 337 and probably significant, that the three-colporate syncolpate grains of subg. KrrGANELIA and Cicca occur in diploid species (2n = 26), whereas all known species with periporate grains are at the tetraploid level (2n = 52) or higher. Although polyploidy is not necessarily an indication of phylogenetic advancement, the correlation of chromosome number with pollen type and other morphological characters suggests that it does indi- cate advancement in Phyllanthus. The cytological picture is not a simple and straightforward one, however, since more than one basic number occurs in the genus, and the correlation of karyotypes with morphologically definable taxa is at best imperfect. Until recently, it appeared that the taxa in subg. IsocLapUS were very divergent from other subgenera in having basic numbers of x = 8, 9 instead of x = 13. Since these plants also had unspecialized branching patterns, in contrast to the phyllanthoid branching found in most of the other subgenera, it might seem that a rather good argument could be presented for the generic segregation of subg. Isocrapus, including our commonest local species, P. caroliniensis. This was not attempted, how- ever, because the flowers in subg. Isoctapus are so typical for the genus that a segregate genus would be very difficult to characterize morphologi- cally. Furthermore, Miller and Webster (1967) have reported x = 9 in tal Niruri, a species with typical phyllanthoid branching (and, in fact, the type species of the taxon including plants with this branching pattern). Although cytological data are still more scanty than one might wish, it now appears that in at least two unrelated lines the basic chromosome numbers of x = 8, 9 have been derived by reduction from the prevailing chromosome complement of x = 13. This bimodality of chromosome number in Phyllanthus appears to be partly explicable on the hypothesis that evolution of the herbaceous habit proceeds in association with a re- duction in chromosome number. However, this appears to hold only for species extending into or restricted to temperate regions; tropical herba- ceous weeds such as P. amarus and P. Urinaria show a precisely contrary tendency towards (euploid) polyploidy. Cytological examination of a considerably larger number of herbaceous species will be necessary (and even then possibly not sufficient) to resolve this curious and puzzling ssibly evolved as a result of increase in number of colpi grain; mechanical considerations apparently preclude if more than three colpi are in subg. KIRGANELIA), po in a syncolpate type of pollen the colpi in syncolpate grains from meeting at the poles present. 338 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 tion seems neither convenient nor necessary. The closest genus to Phyl- lanthus is the Old World Glochidion, which technically differs in its un- divided styles and ventrally invaginate seeds. Among the numerous Australian taxa of Phyllanthus subg. GompHIpIuM ® are some of uncertain generic status, but it seems probable that additional study will make it possible to define the boundary between the two genera more sharply. It seems fairly clear that Glochidion and the smaller Old World genera Breynia, Sauropus, and Synostemon have arisen from Phyllanthus subg. KirGANELIA by loss of floral disc and specialization of seeds, styles, and male perianth. All of these genera together make up the subtribe Phyllan- thinae, sensu stricto. In comparison with many other genera of Euphorbiaceae, Phyllanthus includes relatively few species of economic importance; no latex or seed oils, such as are known in the Crotonoideae, occur. Phyllanthus Emblica L., a tree with branchlets imitating those of Leguminosae with small leaflets, is an important source of tannin and vitamin C; it is cultivated in the Miami area as an ornamental. Of some potential importance, chemotaxonomically, if not economically, are the alkaloids reported from various species of Flueggea and Margaritaria, two genera rather closely related to Phyllanthus. At least 10 different alkaloids are known at present (Hegnauer), and investigation of the less specialized taxa of Phyllanthus may produce data of interest in assessing both intra- and intergeneric relationships. Saponins are known in a number of tropical species of Phyllanthus, and may be the active principle in the use of foliage of various species as fish-poisons; South American species of subg. CONAMI appear to be outstanding in this regard. A few species of Phyllanthus have some promise as ornamentals in tropical and subtropical gardens. The most popular are probably the xylophyllas, various phylloclade-bearing species of sect. XyLopHyi1a which have some interest as botanical curiosities. The most common greenhouse xylophylla, P. elongatus (Jacq.) Steud., is apparently a hybrid between P. angustifolius and P. epiphyllanthus. In outdoor gardens in southern Florida, on the other hand, it is the parental species which are commonly grown, and the hybrid appears to be rare REFERENCES: Under family references see ARNOLDI, BAILEY, BAILLON, BANERJI & DvuTT, SmirH, WesstTeR & ELLIS, WEBSTER & MILLER, WHEELER (1939), and WILLIAMS. Auts, H. E., C. R. Bett, & A. E. Raprorp. Species new to the flora of North and South Carolan “Whodows 60: 10-32. 1958. [P. Niruri and P. penta- phyllus erroneously reported from S. Carolina 4 & A. E. Raprorp. Species new to the flora of North Carolina. Jour. * Phyllanthus subg. Gomphidium (Baill.) Webster, stat. nov. Phyllanthus sect. Gomphidium Baill. Adansonia 2: 234. 1862. LECTOTYPE para Phyllanthus Chamae- cerasus Baill. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 339 Elisha Mitchell Sci. Soc. 75: 140-147. 1959. [Reports P. Niruri, erroneously, from New Hanover Co. | Bartton, H. E. Monographie des Phyllanthus. Premiére Partie. Adansonia 1: 23-43. 1860; 2: 13-20. 1861. [Never completed; treats only certain American spp. | BANCILHON, L. Sur la mise en évidence d’un réle “organisateur” du méristeme apical de l’axe orthotrope de Phyllanthus. Compt. Rend. Acad. Sci. Paris 260: 5327-5329. 1965. _ Mise en évidence de gradients de différenciation lors du développement végétatif de Phyllanthus amarus Schum. et Thonn. Ibid. 262: 1228-1231. 1966 _R. Nozeran, & J. Roux. Observations sur la morphogénese de l’ap- pareil végétatif de Phyllanthus herbacés. Nat. Monspel. Bot. 15: 5-12. Croizat, L. Three new Amazonian species of Phyllanthus L. Trop. Woods 78: 5-9. 1944. [Includes speculations on inflorescence homologies. DESHPANDE, P. K. Contributions to the embryology of Kirganelia reticulata Baill. Jour. Biol. Sci. 2: 76-83. 1963 KaussMANN, B. Beitrage zur Morphologie von Phyllanthus Niruri L. Planta 38: 586-590. 1950. [Includes references to earliest reports on phyllanthoid branching. | . Histogenetische Untersuchungen zum Flachsprossproblem. Bot. Stud. Jena 3: 1-136. 1955. [Development of phylloclades in Phyllanthus, 97— 105.] Lupwic, F. Uber das Bliihen eines brasilianischen Phyllanthus (Ph. Niruri?). Kosmos Stuttgart 18: 35-37. 1886. [Pollination by Diptera. ManesHwakrl, P., & O. R. CHowpry. A note on the development of the embryo- sac in Phyllanthus Niruri Linn. Curr. Sci. Bangalore 5: 535, 536. 1937. Mayumopar, G. P., & M. ArsHap ALI. Developmental studies of Phyllanthus Niruri Linn. and P. reticulatus Poir. (Euphorbiaceae) with special reference to the origin and nature of axillary vegetative buds. Proc. Indian Acad. Sci. B. 43: 149-160. 1956. Rogrnson, C. B. Philippine Phyllanthinae. Philip. Jour. Sci. Bot. 4: 71-105. 1909. [Includes critical review of generic delimitations in the subtribe. | Sencupta, P., & J. MUKHOPADHYAY. Terpenoids and related compounds — VII Triterpenoids of Phyllanthus acidus Skeels. Phytochemistry 5: 531-534. 1966. [Reports isolation of phyllanthol, a pentacyclic triterpenoid. | StncH, R. P. Development of endosperm and embryo in Phyllanthus Niruri L. Agra Univ. Jour. Res. Sci. 5: 163-167. 1956. [Sp. probably misdetermined. | Wesster, G. L. Studies of the Euphorbiaceae, Phyllanthoideae. I. Taxonomic notes on the West Indian species of Phyllanthus. Contr. Gray Herb. 176: 45-63. 1955; II. The American species of Phyllanthus described by Lin- naeus. Jour. Arnold Arb. 37: 1-14. 1956; III. A monographic study of the West Indian species of Phyllanthus. Ibid. 37: 91-122, 217-268, 340-359. pls. 1-14. 1956; 38: 51-80, 170-198, 295-373. pls. 15-23. 1957; 39: 49— 100, 111-212. pls. 24-32. 1958. [Includes several spp. of southeast. Use extensive bibliography, 37: 249-253. _ The origin of the cultivated Xylophylla hybrid, Phyllanthus X elongatus. Brittonia 11: 177-182. 1959. _ The species of Phyllanthus ( Euphorbiaceae) collected by Sessé and Mocifio. Sida 2: 377-380. 1966. [Includes notes on P. pentaphyllus. | 340 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Subfam. EUPHORBIOIDEAE (Crotonoideae Pax) Tribe CROTONEAE Dumort., “‘Crotonieae” Subtribe Jatrophinae Griseb., “Jatropheae” 5. Jatropha Linnaeus, Sp. Pl. 2: 1006. 1753; Gen. Pl. ed. 5. 437. 1754. Shrubs or small trees, or a few species herbaceous. Leaves alternate, stipulate (stipules often glandular), petioles short to longer than the blade, blades entire, toothed, or palmately lobed. Plants monoecious [rarely dioecious]; flowers in terminal, often long-stalked, dichasia, the lower flowers female, the distal ones male. Flowers with biseriate perianth; calyx synsepalous, 5-lobed, sepals imbricate or open in aestivation; petals 5, imbricate to contorted, free or coherent (but not truly connate). Male flower: disc [entire or] dissected; stamens 8-10, connate, anthers biver- ticellate; pollen globose, inaperturate, clavate; vestigial gynoecium absent. Female flower: disc cupular or pulviniform; carpels [2] 3 [rarely 4]; styles more or less connate at the base, bifid, the style-branches entire [bifid] ; ovary glabrous or pubescent; ovules 1 in each locule, anatropous, nucellar beak conspicuous. Fruit capsular; mericarps crustaceous to woody; columella persistent. Seeds 1 in each locule, more or less ellip- soidal, testa crustaceous, carunculate; endosperm copious; embryo straight, cotyledons thin, broader and longer than the radicle. (Including Aden- oropium Pohl, Curcas Adans., Mozinna Ortega.) Type species: J. gossypiifolia L.; see McVaugh, Bull. Torr. Bot. Club 71: 457. 1944. (Name from Greek, iatros, physician, and phago, to eat, in allusion to medicinal and edible properties of the seeds and roots.) In the circumscription of Pax & Hoffmann and McVaugh, Jatropha includes 125-150 tropical species of very diverse habit; a majority (70- 80) of these are American, but 40-50 occur in Africa. Because of the diversity within the genus and the confused concept held by Linnaeus and many subsequent workers, the literature is beset with both nomenclatural and taxonomic problems. Although Cnidoscolus has been retained within Jatropha as recently as the revision of 1910 by Pax, it is actually very distinct, as pointed out by McVaugh. Various segregate genera have been recognized by writers of local floras in the southeastern United States and elsewhere, but Miller & Webster have shown that there is no good evidence for doing so. According to the treatment of McVaugh, the two Southeastern taxa of Jatropha belong to different sections. Representing sect. MACRANTHAE Pax, which is composed of species having a nonimbricate male calyx and glabrous reddish petals, is J. multifida L. This shrubby or arborescent species, easily recognized by its deeply divided leaves with 10 or more sharply toothed lobes, occurs from Mexico and the West Indies south to Brazil, and has been introduced into the Old World. Although no speci- mens have been seen, Small reported it from the Everglade Keys in Dade County, Florida, and it seems likely that the species may become exten- 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 341 sively naturalized in extreme southern Florida, as it is very commonly cultivated there. Our other species, Jatropha gossypiifolia L., has thus far been reported only from the island of Key West; but it seems probable that it, too, will become more widespread, at least in the Florida Keys, as it is extremely weedy in other areas of the Caribbean. It belongs (as type species) to sect. JATROPHA (§ Adenorhopium of McVaugh), a group characterized by having distinctly imbricate male calyx lobes, pubescent petals connivent into a short “tube,” and seeds with prominently fimbriate caruncle. Jatropha gossypiifolia is easily distinguished from its local congener, J. multifida, by means of its broadly and more shallowly lobed leaves beset with many stalked glands. The single specimen seen of J. gossypiifolia from Key West would fit into var. elegans (K1.) Muell. Arg. because of its glabrous leaves. Miller and Webster have shown that five species of Jatropha investigated cytologically have 2n = 22; the species are then presumably diploid, with x = 11. More recently, Miller & Webster have reported two additional counts, one of which (for J. dioica Sessé) represents a tetraploid chromo- some complement (27 = 44). This cytological evidence strongly supports the taxonomic association of Jatropha with Aleurites, as proposed by Bentham, since the latter genus also has a base number of 11. A number of other tropical species of Jatropha are cultivated in southern Florida and may be found as escapes. These include the very ornamental J. integerrima Jacq. (J. hastata Jacq., J. pandurifolia Andr.) with brilliant scarlet flowers; J. Curcas L., with yellowish-green flowers and somewhat fleshy fruits; and J. podagrica Hook., with succulent stems, peltate leaves, and small, red flowers. ‘ Species of Jatropha are of some economic importance in the tropics, although not the basis for any major industry. The seeds of J. Curcas are valued for their purgative properties and for the oil which can be used in the manufacture of soap (Lanjouw, Wilbur). The seeds are dangerously poisonous when fresh, due to the presence of the alkaloid curcine; but when roasted they are edible and distinctly palatable. Other species, such as J. multifida, have seeds with somewhat similar properties. REFERENCES: Under family references see ASSAILLY, BENTHAM, DeHAy, DEWILDEMAN, FROEMBLING, GAUCHER, INGRAM, MICHAELIS, MILter & WEBSTER, Punt, and Scott; under Cnidoscolus see McVaucH (1943, 1944). Hoim, T. The seedlings of Jatropha multifida L. and Persea gratissima Gartn. ‘ot. Gaz. 28: 60-64. 1899. [Describes the unusual seedling of J. multifida in which cotyledons remain inside seed coat. Mackenzie, K. K. Type of the genus Jatropha. Bull. Torrey Bot. Club 56: 213-215. 1929. i McVavucu, R. The jatrophas of Cervantes and of the Sessé and Mociho Herbarium. Bull. Torrey Bot. Club 72: 31-41. 1945 . The genus Jatropha in America; principal intrageneric groups. Torrey Bot. Club 72: 271-294. 1945. Bull. 342 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Miiter, K. I, & G. L. WesstTErR. Systematic position of Cnidoscolus and Jatropha. Brittonia 14: 174-180. 1962. Pout, J. E. Plantarum Brasiliae. Icones et Descriptae. vol. 1. xvi + 136 pp. pls. 1-100. Vindobonae. 1826-1828. Rao, A. R., & M. Mataviya. On the latex-cells and latex of Jatropha. Proc. Indian Acad. Sci. B. 60: 95-106. pls. 4, 5. 1964. Wizzur, R. L. A synopsis of Jatropha, subsection Eucurcas, with the description of two new species from Mexico. Jour. Elisha Mitchell Sci. Soc. 70: 92— 101. 1954. 6. Aleurites G. Forster, Char. Gen. Pl. 111. pl. 56. 1776. Deciduous [evergreen] trees; branches and leaves with indumentum of stellate and simple hairs. Leaves alternate, stipulate (but stipules cadu- cous and mature leaves thus appearing exstipulate), long-petioled; blades entire or lobed, palmately veined, glandular at the base. Plants monoe- cious; flowers in terminal panicle-like cymes. Perianth biseriate: calyx calyptrate in bud, rupturing valvately into 2 or 3 lobes at anthesis: petals 5, large and showy, imbricate or contorted. Male flower: disc of 5 seg- ments [entire]; stamens 8-10[—20], connate; anthers introrse [extrorse |, dehiscing longitudinally; pollen globose, inaperturate, clavate; vestigial gynoecium absent. Female flower: disc dissected into 5 segments; carpels [2] 3-5; styles free, bifid; ovules 1 in each locule, anatropous. Fruit drupaceous, more or less tardily dehiscent. Seeds 1 in each locule, ecarun- culate, testa thick and woody; endosperm copious, oily; embryo straight, cotyledons broad and palmately veined, much longer than the radicle. Type species: A. triloba G. Forst. = A. moluccana (L.) Willd. (Name from Greek, aleurites, floury, in allusion to the pale coating of pubescence on leaves and stems.) — CANDLE NUT, TUNG TREE. A small but economically important genus of five or six species native to the Old World. The genus was widely separated from Jatropha in the treatment of Mueller but fairly closely associated with it by Bentham and by Pax. Evidence from cytological and palynological studies definitely favors the latter view (Erdtman, Miller & Webster). In the southeastern United States Aleurites Fordii Hemsl., the tung oil tree, is extensively cultivated, and has become naturalized to a limited extent. Herbarium specimens have been seen from Florida and Louisiana, and Thorne has reported it from Georgia. The species belongs to sect. DryANnpra (Thunb.) Muell. Arg., which comprises four eastern Asiatic species. The section is characterized by having an indumentum of bifur- cate hairs, showy flowers with petals 1.5 cm. long or more, 3—5 carpels, and a more-or-less dehiscent fruit. Aleurites Fordii, the only deciduous species in the genus, is adapted to a more temperate climate than its congeners. It is native to forested regions in central and western China, extending as far north as Hupeh and Szechuan provinces. As grown in the United States, it is an ornamental tree in aspect somewhat resembling Firmiana simplex (L.) F. W. Wight (Sterculiaceae). The attractive 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 343 flowers, cea net in early spring before the leaves are expanded, are white with pink v: Despite its a ae economic importance, the tung tree was an confused with the Japanese wood-oil cs a cordata (Thunb.) R. Br. e Steud.) and the Mu-yu tree (A. montana (Lour.) Wilson), of aie China and Viet Nam. Both these Rec: however, are evergreen trees adapted to tropical climates and have wrinkled fruits, rather than the smooth ones of A. Fordii According to Newell et al., tung oil trees were first planted in the south- eastern United States in 1906, and since then they have become extensively cultivated in Mississippi, Alabama, Georgia, and northern Florida. The oil, which is contained in the endosperm of the seed, is a valuable drying oil which is essential for the manufacture of the best waterproof varnishes. The seeds have drastic purgative properties similar to species of Jatropha and can cause severe poisoning if eaten raw; the plant may therefore have (from Palm Beach County southward), but they are not seriously culti- vated on a commercial scale. Aleurites moluccana (L.) Willd., candle nut or kukui nut, is cultivated all around the tropics; it can be grown as far north as Brevard County and may quite possibly become naturalized to some extent. The type (and perhaps only) species in sect. ALEURITES, it differs from the species previously discussed in having an indumentum of stellate hairs, smaller flowers (petals usually less than 1 cm. long), and indehiscent fruits of only two carpels.1° Cytological evidence provided by Stockar shows that A. moluccana is a tetraploid, with 2” = 44, whereas the four species of sect. DRYANDRA are all diploids, with 2n = 22 Some interesting experimental taxonomy of Aleurites has been accom- plished (more or less inadvertently) by tung-oil breeders who have crossed A. Fordii with other species in an effort to obtain cultivars with improved disease resistance or a delayed flowering period (and thus less chance of frost damage). Merrill and Kilby report that crosses can be made readily between A. Fordii and A. montana, but that the F, hybrids are partially sterile. Unfortunately, most of their data involves the results of back- crosses, and neither they nor other workers have presented any sort of a © Airy Shaw has recently proposed (Kew Bull. 20: 393-395. 1966) a narrow generic Se ins 5 in which Aleurites would include only A. moluccana (and the probably synonymous A. Remyi Sherff), while A. Fordii and its relatives would be ici tri. ea to some extent bridges the gap between the extremes within the genus; ‘and all these species have in common similar leaves, calyptrate calyx, and large, ecarunculate, oily seeds. In my opinion, the propriety of these segregations must remain in doubt until the generic limits in this entire complex (including Deutzianthus Gagnep. et al.) are revised monographically. 344 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 summary of crossing results which would indicate the degree of reproduc- tive incompatibility between the species. REFERENCES: Under family references see Arry SHAW (1966), BAILLON, DEHAY, GAGNEPAIN, GaucueEr, MApDEL, MicHaetis, Pax (1910, gid PERRY, PoTBURY, Punt, Rir- TERSHAUSEN, SMITH, THORNE, and WHEELE ANGELO, E., R. T. Brown, & H. J. AMMEN. ga studies with tung trees. Proc. Am. Soc. Hort. Sci. 41: 176-180. 1942. Biswas, K. Cultivation and systematic study of the tung oil yielding trees (Aleurites) in India. Jour. Sci. Industr. Res, 4: 260-272. 1945. Brackmon, G. H. Tung oil—a gift of China. Econ. Bot. 1: 161-175. 1947. [General review of the in Bray, G. T. The cultivation and production of tung oil. I. World Crops 3: 247-250. 1951. Brown, R. T., & E. FisHer. Period of stigma receptivity in flowers of the tung tree. Proc. Am. Soc. Hort. Sci. 39: 164-166. 1941. Dickey, R. D., G. Grpert, & C. M. Gropp. The genus Aleurites in Florida: I. Botanical characteristics; II. Chemical and physical properties of the oils. Florida Agr. Exp. Sta. Bull. 503: 1-40. EmMeL, M. W. The toxic principle of as Fordii Hemsl. Am. Veterin. Med. Assoc. Jour. 103: 162. A ag The toxic principle of the tung tree. Florida Agr. Exp. Sta. Bull. 431: 1-35. 1947. FERNHOLZ, D. L. Cold resistance of buds, flowers and young fruits of tung. Proc. ‘Am, Soc. Hort. Sci. 41: 124-126. 1942. Gorz, O. C. Cromosomos en Aleurites moluccana Willd. Arg. Jard. Bot. Rio Janeiro 7: 5-10. 1947. GRANER, E. A. Notes on the chromosome number and morphology in root tips of tung (Aleurites Fordii Hemsl.). Arch, Inst. Biol. Veg. Rio Janeiro 2: 81-82. pl. 1. 1935. Hems.ey, W. B. Aleurites Fordii. Hooker’s Ic. Pl. 29: pls. 2801, 2802. 1906. Revision of the synonymy of the species of Aleurites. Bull. Misc. Inf. Kew 1906: 119-121. 1906. The wood-oil trees of China and Japan. Jbid. 1914: 1-4. 1914. HiInKut, S. G. Tungovoe derevo. (In Russian.) Bull. Appl. Bot. 2: 137-153. 1935, [Discusses and illustrates distinctions between A. cordata, A. Fordii, and A. montana. Hon, H. C. Genus Aleurites in Kwantung and Kwangsi. Lingnan Sci. Jour. 18: 303-327, 513-524. pls. 17-21. 1939. [Provides a good key to 3 spp., original illustrations, and discusses distributions in S. China.] Kurmono, K. Hybridization of the tung tree. Soviet. Subtrop. ‘5; 87-90. 1937.* LANGERON, M. Le genre Aleurites for “Shere i systématique— eRe ee 160 pp. 4 pls. Pari Laycock, D. H., & L. J. Foster. Rainfall and pea bearing in tung (Aleurites montana). 176: 654. 1955. McCann, L. P. evelopment of the pistillate flower and structure of the fruit of tung (Aleurites Fordii). Jour. Agr. Res. 65: 361-378. 4 pls. 1941. . Embryology of the tung tree. Jbid. 71: 215-229. 1945. , W.S, Coox, & C. R. Camppett. Factors affecting time of initiation and 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 345 rate of oe - pistillate flowers of the tung tree. Proc. Am. Soc. Hort. Sci. 39: 157-1 941. MERRILL, S. Heterosis in ast Proc. Am. Soc. Hort. Sci. 71: 231-236. 1958. . Kitpy. Progress in hybridizing Aleurites Fordii with A. mon- tana. Proc. ‘Am. Soc. Hort. Sci. 78: 225-229. 1961. Mowry, H. ee in the tung-oil tree. Florida Agr. Exp. Sta. Tech. Bull. 247: 1-32. 1932. NEWELL, W., es Mowry, & R. M. Barnette. The tung-oil tree. Florida Agr. Exp, Sia. Bull. 280: 1-67. 1935. SELL, H. M., & F. A. Jounston, Jr. Biochemical changes in terminal tung buds during their expansion prior to blossoming. Pl. Physiol. 24: 744-752. 1949. SHERFF, E. E. Additional pet of the Hawaiian Euphorbiaceae. Field Mus. Pu bl. Bot. 17: 547-576. 1939. [Description of A. Remyi Sherff. ——. Some additions to our knowledge of the flora of the Hawaiian Islands. Am. Jour. Bot. 31: 151-161. 1944. [Additional notes on A. Remyi. | StocKar, A. Complemento cromosémico diploide de algunas especies Aleurites. Revista Argent. Agron. 13: 253-255. pl. 6. 1946. [Includes all spp. | ommunicacion ee ian sobre hibridaciones entre varias especies de Aleurites. Ibid, 14: 33-38. 1947. Warp, F. K. A hybrid p Rae Gard. Chron. III. 122: 128-129. 1947. [A. montana X A. Fordii; no quantitative data. Wesster, C. C. A note on pollination in budded plantations of tung trees (Aleurites montana). Nyasaland Agr. Quart. Jour. 3: 17-19. 1943.* Observations and experiments on flowering and pollination of the tung tree. E. Afr. ecih Jour. 9: 136-143. 1944.* WHEELER, L. C. Notes on the genus Aleurites. Harvard Univ. Bot. Mus. Leafl. 7: 119-122. 1939, [Largely details of nomenclature. | Witson, E. H. The “wood-oil” trees of China and Japan. Bull. Imp. Inst. London 11: 441-461. 1913. [The first clear discrimination of A. montana. | Wirt, F. Het botanisch onderzoek van Aleurites, Landbouw Nederl. Indié 15: 9-27. 1939 Subtribe Manihotinae Muell. Arg., ‘““Manihoteae” 7. Manihot Miller, Gard. Dict. Abr. ed. 4. 1754. abi or shrubs [rarely trees], often with tuberous roots, stems laticif- ous. Leaves alternate, stipulate (stipules often very small), long- Bete blades mostly palmately lobed, stipellate at the base but without distinct foliar glands. Plants monoecious; flowers in terminal, sometimes pseudo-axillary panicles, the female flowers on several lateral basal axes, the male flowers more or less racemose along the central axis. Perianth uniseriate (presumably petals missing), often petaloid; calyx lobes im- ricate. Male flower: calyx synsepalous, more or less cyathiform; disc central, intrastaminal, 5-lobed, lobes more or less bifid; stamens 10, free, biseriate, the outer opposite the calyx-lobes; anthers introrse, dehiscing longitudinally; pollen globose, periporate, exine clavate; rudimentary gynoecium present or absent. Female flower: calyx synsepalous but more deeply lobed than the male; disc pulviniform, staminodia sometimes pres- ent as well: carpels 3; styles 3, connate below; stigmas dilated and lacer- 346 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ate or multifid; ovules 1 in each locule, anatropous. Fruit capsular, de- hiscent, sometimes winged; columella dilated above, often persistent. Seeds smooth, carunculate; seed coats thin and crustaceous; endosperm starchy; embryo with broad palmately veined cotyledons. LrcToTYPE species: Jatropha Manihot L. = Manihot esculenta Crantz; see Adanson, m. Pl. 2: 356. 1763. (Name from manioc, aboriginal Indian word for the cassava plant.) — CASSAVA, TAPIOCA PLANT. Over 150 species of Manihot are known, according to Pax and Hoff- mann; the number may, however, prove to be considerably smaller than this when the genus is monographically revised. The majority of the species is native to Brazil, but a respectable concentration is present in Mexico, and two or three cross the border into the United States Although originally confounded with species of Jatropha by Linnaeus, Manihot is very distinct by virtue of its apetalous flowers, central male disc, and periporate pollen grains. The basic chromosome number of n = 18 (x probably = 9) rather than x = 11 in Jatropha also supports a rather wide taxonomic separation. Mueller’s placement of the two genera in adjacent subtribes presents a reasonable picture of their affinity. As pointed out by Miller & Webster, the closest genus to Manihot appears to be Cnidoscolus (q.v.). The tapioca plant, Manihot esculenta Crantz (M. utilissima Pohl), is cultivated in southern Florida, as elsewhere in the Caribbean area, and has been reported by Small (as Jatropha Manihot L.) to be naturalized in the Florida and Everglade Keys. Lakela & Craighead report it from Collier, Dade, and Monroe counties. One specimen (Small, Nov. 1904) from Dade County does seem to represent this species, so that the above records may well be correct. However, collections of naturalized Manihot north of extreme southern Florida probably do not represent the cassava plant, which is not sufficiently frost-resistant to overwinter in northern Florida. The record of M. esculenta cited by Thorne from Georgia is erroneous. His collection (Thorne 5866), as well as naturalized plants from northern Florida (e.g., Godfrey 54844), may tentatively be identified as M. Grahamii Hook. (M. Tweedieana Muell. Arg.), a species native to Paraguay and western Brazil.1! The somewhat notched leaflets of these collections suggest M. carthaginensis (Jacq.) Muell. Arg., and at least one sheet has been so identified. However, that species has much smaller stipules, and the leaflets are distinctly lyrately lobed. According to the classification of Pax, both of these naturalized species belong to sect. PARvIBRACTEATAE Pax, which includes about 35 to 40 mainly Brazilian species characterized by deeply lobed leaves and small, ee bracts. Manihot Grahamii, assigned to subsect. ELATAE e recognized by its slender, attenuate-acuminate, laterally fiat stipules which become 1 cm. or more long, its large glabrous male _ —— (1943) has reduced M. Grahamii to a synonym of M. Roletiiene: Be: and he may be correct in doing so. However, the long, glabrous, toothed stipul chisipdelely aalasigs male calyx of he oe and Florida plants do not cued * with Mueller’s description of Pohl’s speci 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 347 calyces (mostly 12-15 mm. long), thin leaves with (5—)7-10 abruptly acuminate lobes, and wingless capsules with relatively large seeds (7 mm. or more across), The true cassava plant, M. esculenta, belongs to subsect.. UTILIssIMae Pax because of its internally pubescent male calyx and deeply lobed leaves. It may be distinguished from M. Grahamii by its shorter stipules, smaller male calyces (less than 10 mm. long), thicker leaves with fewer (3—7) lobes which are gradually acuminate and paler beneath, and its distinctly wing-angled capsules with somewhat smaller seeds (mostly 6.5 mm. or less across). Cytological studies on eight species of Manihot have demonstrated a uniform chromosome number of 2” = 36, except that one accession of M. esculenta was tetraploid (2n = 72). Actually, the genus may be basically tetraploid, since x = 9, in all probability. As pointed out by Miller & Webster, the cytological evidence alone suggests that Manihot is more closely related to Cnidoscolus, which also has x = 9, than to Jatropha, in which x = 11. Except for Cnidoscolus, Manihot does not appear to have any close relatives, and these two genera would appear to be the sole members of subtribe Manihotinae. Manihot esculenta is a plant of great economic importance as a carbo- hydrate source in lowland tropical areas. Although cassava is most im- portant as a starchy food in Latin America, it is also cultivated fairly intensively in Africa and to some extent in Asia as well. Despite the importance of the crop, the taxonomy of the cassava plant and closely related species is surprisingly imperfect. Rogers has shown that there is no morphological correlation between the two major “convariants’’ of manioc defined on root and stem characters with the “sweet” or “sour” nature of the roots. In sweet cultivars, the hydrocyanic acid is confined to the phelloderm of the root, but in the sour cultivars it occurs in the cortex as well, and the root is consequently dangerously poisonous unless boiled and mashed or strained. Rogers notes that sweet cultivars are found mainly west of the Andes and bitter to the east, and suggests that possibly cultivation began on the eastern slopes of the Andes. Although of little present importance, arborescent species of Manthot were at one time an important source of rubber (Zimmermann, Polhamus). The most valuable species appears to have been the Ceara rubber tree, M. Glaziovii Muell. Arg., of sect. GLAZIOVIANAE Pax, which was extensively planted in East Africa prior to World War I. Manihot dichotoma Ule was also planted in Africa to a considerable extent. Although plants of Mani- hot could be cultivated in drier regions than Hevea, greater difficulty in tapping and obtaining high yields led to its drop from favor. ee Because of the low protein content in cassava starch, it is an inferior carbohydrate source and a potential cause of malnutrition in areas where it is used heavily; consequently, efforts have been made to develop culti- vars with enriched roots by crossing with other species (Bolhuis). These crossing programs, although purely utilitarian, are beginning to yield interesting data with regard to species relationships within Manihot, as noted by Rogers. Such species as the Guianian M. saxicola Lanj. and M. 348 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 melanobasis Muell. Arg. are easily crossed with M. esculenta; according to Bolhuis and Jennings, they may not even be distinct species. This evidence suggests that the taxonomic arrangement of Mueller, which was followed in general by Pax, in 1910, may be quite artificial, and that the infrageneric taxa of Manihot may have to be drastically revised. REFERENCES: Under family references see DEHAY, GAUCHER, HERBERT, INGRAM, LAKELA & CRAIGHEAD, LANJouw, McVaucu, MIcHAELIs, MUELLER (1874), Pax & Horr- MANN (1910, II), Perry, PoLHAMUS, PUNT, SCHWEIGER, Scott, THORNE, VEN- TURA (1940), De WILDEMAN, and WiLL1AMs. Under Jatropha see MACKENZIE. ABRAHAM, A. Natural and artificial a in tapioca (Manihot utilissima). Proc. 31st Indian Sci. Congr. 1944(3): 1945? BraNcuHint, M. La geografia della manioca. “i ll. Soc. Geogr. Ital. VIII. 4: 26-53. 1951. ee economic, but with extensive bibliography. | Botuuts, G. G. A survey of some attempts to breed cassava-varieties with a high content of ite in the roots. i. 2: 107-112. 1953. [De- scribes crosses between M. esculenta and M. saxic rhea J. M.,& V. C. Bruce. Floral biology and raoas of Manihot utilis- ma. Philip. Agr. 39: 306-316. 1955. Ceara M. F., & K. D. S. S. NANAYAKKARA. Studies in cassava. classification of races occurring in Ceylon. Trop. Agr. Ceylon 101: 3- 12, 214-222. 1945; II. Production of hybrids. Jbid. 104: 59-74. CrFErRRI, R. Saggio di classificazione delle razze di Manioca as anihot esculenta C rantz), Relaz. Monogr. Agrar.-Colon. 44: 1-58. . Fondamenti per una classificazione cae tients ie Manihot esculenta Crantz. Arch. Bot. Forli 18: 27-33. 1942. Cours, G. Le manioc 4 Madagascar. Mém. Inst. Sci. Madagascar B. 3: 203- 400. pls. 1-15. 1951. [Includes morphological notes, 207-273; results of rears crosses, 363-365. | Croizat, L. A study of Manihot in North America. Jour. Arnold Arb. 23: 216- 225. 1942. [Discussion and description of new spp., but not a full revision. | . Preliminari per uno studio del genere “Manihot” nell’America merid- ionale. Revista Argent. Agron. 10: 213-226. 1943. [Listing, without keys; combines M. Grahamii and M. Tweedieana with M. flabellifolia.| Manthot Tweedieana Mueller is unacceptable. Ibid. 11: 173, 174. 1944. [Correctly points out priority of M. Grahamii. Graner, E. A. Genetica de manihot. I. Hereditariedad da forma da folha e da coloracio da pelicula externa das raizes em Manihot utilissima Pohl. Bragantia 2: 13-22. 1942. Hooker, W. J. Manihot Grahami. Hooker’s Ic. Pl. 6: pl. 530. 1843. Jennincs, D, L. Further studies in breeding cassava for virus resistance. E. Afr. Agr. Jour. 22: 213-219. 1957, . Manihot melanobasis ee Arg. —a useful parent for cassava breed- ing. Euphytica 8: 157-16 9. Lanjouw, J. Two interesting yee of Tcdrowrat L. from Suriname. Rec. Pay Bot. Néerl. 36: 543-549. pl. 9. 1940. [M. saxicola Lanj., new sp. NicuHots, R. F. W. Breeding cassava ne virus resistance. E. Afr. Agr. Jour. 12: 184-194. 1947. Pax, Ie Euphorbiaceae-Crotonoideae—Adrianeae. Pflanzenreich IV. 147-II( Heft : 1-111. 1910. [Manihot, 21-99.] 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 349 Rocers, D. J. Intraspecific categories of Manihot esculenta. Science 126: 1234, 1235. 1957. _ Studies on Manihot esculenta Crantz and related species. Bull. Torrey Bot. Club 90: 43-54. 1963. _ Some botanical and ethnological considerations of Manihot esculenta. Econ. Bot. 19: 369-377. 1965. [Includes speculations as to interspecific relationships, as well as place of origin of cassava. Scuery, R. W. Manioc—a tropical staff of life. Econ. Bot. 1: 20-25. 1947. Scott, D. H. On the laticiferous tissue of Manihot Glaziovii (the Ceara rubber). Quart. Jour. Microscop. Sci. 24: 194-204. 1884. SENARATNA, J. E. Bisexual flowers in the manioc, Manihot esculenta Crantz (M. utilissima Pohl). Ceylon Jour. Sci. A. 12: 169. 1945. Totepo, A. P. Anatomia e desenvolvimento ontogenético da flor de mandioca. Bragantia 22: 465-476. 1963. _ Anatomia e desenvolvimento ontogenético do fruto e da semente de mandioca. (English summary.) bid. 22: Ixxi-lxxvi. : Tracy, S. M. Cassava. U. S. Dep. Agr. Farmer’s Bull. 167: 1-31. 1903. Viécas, A. P. Anatomia da parte vegetativa da mandioca. Inst. Agron. Sao Paulo Bol. Téc. 74: 1-32. 1940. ZIMMERMANN, A. Der Manihot-Kautschuk. xli + 342 pp. Jena. 1913. [De- scription of rubber industry based on M. Glaziovi and related spp. | 8. Cnidoscolus Pohl, Pl. Brasil. Ic. Descr. 1: 56. 1827, nom. cons. prop. Herbs, shrubs, or trees; stems and leaves containing milky latex and armed with stinging hairs; pith septate. Leaves palmately [or pinnately] veined or lobed [sometimes parted into segments and almost compound] ; petioles usually long, glandular at junction with blade; stipules entire to laciniate [rarely obsolete]. Plants monoecious; flowers in terminal (some- times pseudo-axillary) dichasia, the female flowers at the lower (proximal) nodes, male at the upper (distal) nodes. Flowers apetalous; calyx more or less petaloid, usually white. Male flower: calyx synsepalous; calyx lobes imbricate: disc annular; stamens 8—10[—25], the outer free and inner connate [or all connate]; pollen globose, periporate, clavate; stam- inal column without terminal filiform staminodia [these present in some exotic taxa]. Female flower: calyx aposepalous; sepals deciduous [syn- sepalous in some species]; disc annular, sometimes associated with stam- inodia; carpels 3 [rarely 5]; styles more or less free, several times bifid or laciniate [rarely once bifid], sometimes apically dilated; ovule 1 in each locule, anatropous. Fruit capsular [rarely drupaceous | ; columella slender: Seeds 1 in each locule, carunculate; endosperm starchy; embryo straight, with broad cotyledons longer than the radicle. (Bivonea Raf.) LECTOTYPE species: C. hamosus Pohl; see Small in Britton and Brown, Illus. FI. No. U. S. ed. 2. 2: 462. 1913. (Name most appropriately derived from Greek, cnide, nettle, and skolos, thorn, in allusion to the stinging properties of the plants.) — STINGING NETTLE, BULL-NETTLE. This distinctive genus of about 50 American species has long been con- fused with Jatropha, due to the influence of Mueller, who so treated it. However, McVaugh (1944) pointed out its distinguishing characteristics 350 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Chi doscolus. a-l, C. stimulosus: a, port on of flowering: and fruiting flowers, a ; te en from unopened bud to show attachment of anther, ; d, androecium after anthesis, showing disc, ten stamens, and staminode, < 6; e, female flower, with adjacent male buds removed — note quincuncial aestivation, x , gynoeciu i i nd staminodia, < 6; &, gy wi nearly mature fruit, Gee h, columella after dehiscence of schizocarp, < 2; i, one valve of schizocarp after dehiscence, showing bony endocarp, a 2; j, k, lateral and adaxial surfaces of seed with caruncle, xX 2; 1, embryo, X 2 and correctly insisted that it is very distinct from Jatropha. Miller & ebster subsequently showed that evidence from petiolar anatomy, pollen morphology and chromosome counts agrees and furthermore suggests that a stinging hairs, petiolar glands, laciniate style tips, septate pith, and other 1967] WEBSTER, GENERA OF EUPHORBIACEAE 351 characters (McVaugh). Baillon noted that, although Cnidoscolus was much closer to Manihot, it differed in its free stamens and extrastaminal disc. The interesting South American C. urnigerus (Pax) Pax appears to be partly transitional between the two genera, since it has free stamens as in Manihot but the extrastaminal disc and stinging hairs of Cnidoscolus. There can now be little doubt that Cnidoscolus is more closely related to Manihot than to any other genus in the family, as is indicated in this treat- ment by the association of the two genera in the subtribe Manihotinae. In our area Cnidoscolus is represented by two species belonging to sect. Jussteuia (Houst.) Pax, which, in the circumscription proposed by McVaugh, is characterized by distinct outer filaments (but see C. texanus) and small multiple petiolar glands. Our representatives belong to the subsect. URENTES (Pax) Pax & Hoffm., a small group of five more or less herbaceous species which have the inner stamens united, and styles 2—3 times bifid. The only other representative of sect. JussteutA, the aberrant C. urnigerus, differs strongly in its completely distinct filaments and simply bifid styles, and perhaps should be referred to a different section. Cnid- oscolus stimulosus (Michx.) Engelm. & Gray is widespread in dry, often sandy areas along the Coastal Plain from southeastern Virginia to the Florida Keys and west to Mississippi and eastern Louisiana; it is not recorded west of the Mississippi River in Louisiana. It may be recog- nized by its small male flowers (tube under 1 cm. long) with glabrous tube and small seeds (under 9 mm. long). Our other species, Cnidoscolus texanus (Muell. Arg.) Small, differs in being a much coarser and more viciously stinging plant with larger stipules, male flowers (tube 1-2 cm. or more, spinose), and seeds (mostly 12-15 mm. long). The androecium of C. texanus also differs in having the outer whorl of stamens connate below, so that the species does not agree with the sectional character. It is a characteristically Texan species of limestone areas, but it enters our range in southwestern Arkansas and western Louisiana. A gap of about 150 miles separates its easternmost station in Rapides Parish, Louisiana, from the westernmost known locality of C. stimulosus in Washington Parish. The two species are “classical” vicariants, therefore, and may represent the collateral descendants of a widespread ancestral population on the pre-Pleistocene coastal plain. Cytological observations (Miller & Webster) indicate a chromosome number of 2 = 36 in both local species, suggesting that they are tetra- ploid if x — 9. The same chromosome number has been reported for four Mexican species (Miller & Webster), so the genus is cytologically homo- geneous so far as is known, and its karyotype appears close to that of Manihot. The relationships between our two species and the other taxa of sect. Jussteura remain to be worked out. Mueller included all of the taxa of subsect. URENTES except C. texanus in his inclusive Jatropha urens.’* As 72 4 complicating factor in the taxonomy of the C. urens group is the typification of C. urens (L.) Arthur. Mueller (1866) correctly pointed out that the specimen in the Linnaean Herbarium is not the common weedy plant but rather a very different 352 JOURNAL OF THE ARNOLD ARBORETUM | VOL. 48 noted by McVaugh, however, both C. stimulosus and C. texanus seem to be sufficiently distinct species, whatever the disposition of the Brazilian taxa of the complex. It is perhaps not too surprising, in view of its relationship with Hevea and Manihot, that Cnidoscolus also includes species with a high rubber content in the latex. At least two arborescent Mexican species, C. elasti- cus Lundell and C. tepiquensis (Cost. & Gall.) Lundell, the ‘“chilte rojo” and “chilte blanco”’ of Mexicans, yield rubber somewhat resembling that of Hevea (McVaugh, Lundell). During World War II these were investi- gated as possible emergency sources of rubber, but the latex proved to have too high a resin content, and no commercial production of any impor- tance was sustained. Since the war, some commercial exploitation has continued, for the latex provides a useful additive to other gutta-like gums (Williams). The characteristic stinging hairs of Cnidoscolus have never been inten- sively studied, although Solereder and Lutz mentioned their resemblance to those of Urtica. The mechanism may be similar, as in both genera the unicellular hairs end in a minute spherical tip which breaks off to permit injection of the poison. In Cnidoscolus the size of the hairs and their distribution on parts of the plant, especially the flowers, varies consider- ably, although the fruit is usually quite spiny, even if the perianth is not. A few tropical species, especially C. aconitifolius (Mill.) Johnston and C. chayamansa McVaugh, have nearly unarmed leaves which are used as a vegetable and may prove to be a good source of vitamin C. REFERENCES: Under family references see BAILLON, DEHAY, GAUCHER, INGRAM, HERBERT, Lourteic & O’DoneLL (1943), MILLER & WepsteR, Pax & HorrMANN (1910, I), PotuHamus, Punt, SOLEREDER, and Upnor. Bonpar, G. Penao, Cnidoscolus Marcgravii Pohl, novo recurso oleifero da Bahia. Inst. Centr. Fomento Econ. Bahia Bol. 12: 1-16. pls. 1-4. 1942. Croizat, L. New and critical Euphorbiaceae of Brazil. Trop. Woods 76: 11-14. 1943. [Mainly on Cnidoscolus; discusses nature of the perianth. | Jounson, P. R. Texas bullnettle and its control. Bull. Tex. Agr. Exp. Sta. 052: 1-11. 1966. [Includes descriptions and illustrations of underground parts of C. texanus. | LEON, HNo. Cnidoscolus y Victorinia gen. nov. en Cuba y Espafiola. Mem. Soc. Cuba. Hist. Nat. 15: 235-244. pls. 23, 24. 1941. [Describes new genus on basis of 2 Greater Antillean spp. with 5 carpels and drupaceous fruit. ] LuNDELL, C. L. Cnidoscolus elasticus, the source of highland chilte, a remarkable new rubber yielding tree from the states of Durango and Sinaloa, Mexico. Field Lab. 12: 33-38. 1944. species belonging to sect. CALYPTROSOLEN (ie., C. Kunthianus (Muell. Arg.) Pax & Hoffm.). This specimen is so different, however, from the illustrations of Commelin, Marcgrav, and Plukenet cited by Linnaeus that it is hardly desirable to designate it as the type. The Commelin illustration accords rather well with the plant later de- scribed by Pohl as C. M arcgravii, so that it may perhaps serve to typify the species. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 353 . The genus Cnidoscolus in Mexico: new species and critical notes. Bull. Torrey Bot. Club 72: 319-334. 1945. sible sources of “chilte” rubber). 23 pp. Rubber Development Corp., Wash., D. C. 1943. [Primarily on spp. of Cnidoscolus, including a key and illustrations. | . The genus Cnidoscolus: generic limits and intrageneric groups. Bull. Torrey Bot. Club 71: 457-474. 1944. Menavt, P. A chemical analysis of Jatropha stimulosa. Jour. Agr. Res. 26: 259, 260. 1923. [Seeds contain ca. 51% fat, yielding oil similar to castor oil. | Mutter, K. I., & G. L. WesstTer. Systematic position of Cnidoscolus and Ja- tropha. Brittonia 14: 174-180. 1962. Pax, F. Euphorbiaceae—Jatropheae. Pflanzenreich IV. 147(Heft 42): 1-148. 1910 SEIGLER, D. S., & J. J. BLoomrretp. The chemistry of the genus Cnidoscolus — I. The fatty acid components of the seed oil. Phytochemistry 6: 451. 1967. [Contains 71% linoleic acid. ] Wueeter, L. C. Pedilanthus and Cnidoscolus proposed for conservation. Contr. Gray Herb. 124: 47-52. 1939. Wiurams, L. Laticiferous plants of economic importance. II. Mexican chilte (Cnidoscolus): a source of gutta-like material. Econ. Bot. 16: 53-70. 1962. Subtribe Crotoninae Benth., “Eucrotoneae’”’ 9. Croton Linnaeus, Sp. Pl. 2: 1004. 1753; Gen. PI. ed. 5. 436. 1754. Herbs, shrubs [or trees]; stems often with colored or resinous sap but not with milky latex; indumentum at least in part of branched or lepidote trichomes. Leaves alternate, sometimes crowded and subopposite or thyrses, the female flowers solitary at lower (proximal) nodes, the male female flowers. Male flower: calyx 5-lobed (lobes rarely 4 or 6), lobes imbricate to valvate; disc e the petals; petals isomerous with calyx lobes, mostly about as long (rarely absent): stamens [3—]8-20[-400], free, filaments usually inflexed in bud; anthers extrorse in bud; pollen globose, inaperturate, clavate; vestigial gynoecium absent; receptacle often densely lanate. Female flower: calyx lobes (4) 5-7(-10), sometimes accrescent, disc annular or rarely dissected, each locule, anatropous, nucellus elongated into a slender beak. Fruit capsular [rarely somewhat fleshy]; columella persistent, usually slender; 354 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 seeds 1 in each locule, terete to compressed, carunculate; testa smooth and dry, usually thin; endosperm copious; embryo straight, cotyledons broader than and as long as or longer than the radicle. LecroTyPE SPECIES: C. aromaticus L.; chosen here.!* (Name from Greek, roton, tick, in allusion to the resemblance of the seed of Ricinus to the arachnid; arbitrarily applied to the present genus by Linnaeus to replace Tournefort’s Rici- noides ). A very large genus comprising more than 600 species, according to Pax Hoffmann, or around 1000 species, in the opinion of Croizat. It seems probable that the latter estimate will prove to be somewhat inflated when the genus is monographically revised. More than two-thirds of the species are American, and, of these, the majority are South American and West Indian; perhaps 75 or 80 species occur in mainland North America. In the circumscription established by Mueller and followed by later workers, Croton is a very natural and rather isolated genus, without any close relatives other than such small satellite genera as Crotonopsis and Eremocarpus.‘* The characteristic oe pollen grains with large, hexago- nally hen excrescences are much like those in other subtribes of the tribe Crotoneae. However, Croton is well characterized by the distinctive the bud. Furthermore, the stellate or lepidote pubescence and the non- milky sap of Croton permit its easy distinction from many other taxa of Crotonoideae. The circumscription of the infrageneric taxa of Croton is as contro- versial as that of the genus is not. Bentham and various later workers sharply criticized the system of Mueller as highly artificial and unnatural; but it is most difficult to find clear-cut assemblages of species, and pro- tracted acquaintance with the genus tends to induce greater sympathy for Mueller’s adoption of an artificial arrangement. Johnston has sensibly placed the Texas species of Croton in informal groups, and Ferguson did not use any subgeneric designations at all. In the present treatment an attempt has been made to incorporate the results of recent anatomical and cytological investigations without introducing any more taxonomic changes than are absolutely necessary. Small (in Britton and Brown, Illus. Fl. No. U. S. . 2: 454. 1913) mee Croton Tiglium L. as the type species, and he has been yee by recent work so it would appear that one of these must be designated the gidag Since Geiseler combined: the two species under the latter name, that one is chose r “In agreement with the opinion of Macbride (Publ. Field vi Bot. 13(3A): is ; : ov. this section have as yet been reported from our area, although C. argenteus L. enters southern Texas and may possibly become established in peninsular Florida. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 355 The 13 or 14 species of Croton in our area belong to seven sections. Section ANDRICHNIA Baill. (§ Eluteria Griseb.), the only section in which female flowers with well-developed petals occur, is represented by the interesting and rare C. alabamensis E. A. Sm. ex Chapm., which is known only from two populations occupying less than 100 acres in central Alabama (Bibb and Tuscaloosa counties). Farmer carefully studied the life history of the species, the reproductive biology of which is now the best known of any species of Croton. He found the Bibb County plants to be diploids (2n = 32), whereas the larger-leaved plants from the War- rior River area, in Tuscaloosa County, are tetraploids (2m = 64). The species does not appear to have any close relatives in North America, un- less Croizat was correct in suggesting that C. argyranthemus is a depau- perate member of this section. Section Croton, although by far the largest subgeneric group in the Muellerian sense (with over two-thirds of the species), is represented in our area by only three species. Plants of this section, which includes the “typical” crotons with smooth seeds, pubescent male receptacle, and pentamerous regular calyces, may be assigned to several subsections. Plants with stellate, rather than lepidote, indumentum and stamens rela- tively few in number (20 or less) belong in subsect. MEDEA (KI.) Pax; our representatives are C. linearis Jacq. and C. humilis L., both tropical species confined to southern Florida. Subsection ARGyRocROoTON Muell. Arg., including plants with lepidote indumentum, is represented by C. argyranthemus Michx., which occurs along the Coastal Plain from Georgia and Florida west into eastern and southern Texas. Section Astrara (KI.) Baill. is in many ways the most distinctive group within Croton and has the best claims to generic status.’ It is an entirely American group of less than 10 species, all of which except the local one, C. lobatus L., are confined to South America. Croton lobatus appears to be a newcomer to our flora, since it was first reported in 1965 by Lakela and Craighead from Dade County, Florida. It may be easily distinguished from other local species of Croton by its three-to-five-lobed leaves, glabrous male petals and receptacle, and cylindrical seeds. Despite the undoubted distinctiveness of Astraea, its species resemble some of those in the fol- lowing section, and, except for the absence of oil cells mentioned by Froembling, there really seems to be no basis for excluding the group from Croton. Section GEISELERIA (KI.) Baill. (§ Decarinium of Mueller) is an en- tirely American group of about 30 species brought together by Mueller on the basis of the irregular female calyx which has two of the lobes greatly reduced. As Mueller himself admitted, the section may be unnatural; the species with lepidote leaves placed in series LAMPROCROTON Muell. Arg. should perhaps be referred to one or more other sections. Series GEISELERIA ** Mueller’s circumscription of sect. ASTRAEA was arbitrarily based on the single character of the le receptacle, and is consequently unnatural. However, if the tw orth American species (C. humilis and C. panduriformis) are excluded, the remainder of the section represents a very natural and distinctive group. 356 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ———s ———_-s 2. Croton. a-b, C. linearis: a, inet of male plant, 2; b, female fiowee two sepals removed to show d staminodia, X 6. c—m, C. al bam turator above ovule, 10; h, branchlet pons nearly mature fruit, X 1; i, ma- ture schizocarp before dehiscence, X 2; j, same in p Rema ae section to show columella at center, seeds with EF embedded in endosperm (stippled), and bony endocarps Sama he X 2; k, persistent sepals, petals, and columella after dehiscence of fruit, x 2; 1, seed, adaxial face, X 3; m, embryo, X 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 357 (Kl.) Muell. Arg., comprising plants with stellate pubescence, is repre- sented in our area by four taxa of controversial relationship. The com- modest and best known, C. glandulosus L., occurs from Maryland to Iowa, south to western Texas and Florida, and beyond our limits into South America. It differs from the species so far mentioned in its sharply dentate leaves with large patelliform glands at the base of the blade, male flowers with only seven to nine stamens, and much sparser coat of stellate tri- chomes. In peninsular Florida occur a number of similar populations which have been recognized as separate species. Perhaps the most distinctive is C. arenicola Small, which includes plants with shorter and much more densely pubescent leaves; it was mistaken for the West Indian C. betulinus Vahl by Ferguson. Croton floridanus Ferg., on the other hand, is a name applied to plants with much less pubescence than usual. Further studies are needed to determine whether either of these taxa should be ranked as distinct species or whether they should take their place among the plethora of varieties of C. glandulosus already recognized by Mueller, Ferguson, and Croizat. A fourth species which seems to be quite distinct, C. trinitatis Millsp., is known in our area only from a single collection at Pensacola, but it may be expected in waste places farther south in Florida. It is easily distinguished from C. glandulosus by its broader, more coarsely toothed leaves. A small, entirely North American group of four taxa, sect. HEPTALLON (Raf.) Muell. Arg., well represented in our area, differs little from sect. Croton except for the calyx of usually six to ten lobes; otherwise, the stellate pubescence and twice bifid styles would relate it to herbaceous taxa of subsect. Mepea. Croton capitatus Michx., as treated by Johnston, is a wide-ranging species with three varieties, two of which, vars. capitatus and Lindheimeri (Engelm. & Gray) Muell. Arg., occur scattered through most of our area, except southern Florida. Further study may show that these “varieties” should indeed be recognized as distinct species. Croton Elliot- tii Chapm., a seldom-collected plant of Georgia and northern Florida, is very similar but has narrower leaves and less deeply divided styles. Section ANGELANDRA Muell. Arg. includes two annual North American species, both of which occur in our area. Mueller’s “key” character of the male calyx “inaequaliter (3—) 5-partitus’” was poorly chosen, since the male flowers are in fact not particularly zygomorphic. A more distinctive feature is provided by the rather highly reduced inflorescences often with a single female flower which becomes reflexed (except In an extralimital Texas population). Croton monanthogynus Michx., common and wide- spread from Maryland and Iowa, south to Georgia, Texas, and northern Mexico, was made the basis of the genus Gynamblosis, primarily because of its female flowers with only two carpels, one of which usually becomes abortive in fruit. However, C. Lindheimerianus Scheele, a closely related but coarser and more heavily pubescent species which enters our area In n, 0, C. glandulosus: n, hair from leaf, « 50; 0, seed, lateral view, * 5. P; 4 C. punctatus: hairs from leaf, X 50. r, C. argyranthemus: seed, lateral view, > ae F 358 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 western Arkansas, has three carpels which develop into a pers aie fruit; it thus bridges the gap between C. monanthogynus and more “typ- ical” taxa of the genus. Johnston has provided detailed Sen and distribution maps of these taxa as they occur in Texas Our last group, sect. DREPADENIUM (Raf.) Muell. Arg., comprises eight or nine American species, mostly of temperate North America (except for the aberrant C. Eichleri Muell. Arg. of Brazil). It is set apart from all the other sections by apetalous male flowers and a distinctive pale indu- mentum of trichomes often more or less intermediate between typical stellate hairs and scales. Croton punctatus Jacq. (C. maritimus Walt.), the most widespread species, common on sandy coastal beaches from Cape Hatteras, North Carolina, south throughout our area and beyond, differs from related taxa in its long-petioled leaves and monoecious inflorescences. Croton texensis (Kl.) Muell. Arg., a dioecious species with narrower, more sparsely pubescent leaves, is adventive in Alabama, according to Johnston. The chromosome number 2m = 14, at variance with all other counts for Croton, has been reported from an extralimital species of this section, C. californicus ; cytological oe on C. punctatus would therefore be of especial taxonomic inter cent cytological sara on Croton suggest that the genus may prove to have a diversity of karyotype exceeded only by Euphorbia (Miller & Webster). Chromosome numbers are now reported from seven sections, and, although scarcely more than ten species have been sampled, it would appear that there are four base numbers represented: x = 7 in sect. DREPADENIUM; x = 8 in sects. ELEUTERIA, GEISELERIA, ANGELANDRA, and Croton (subsect. MEDEA); x = 9 in sect. ASTRAEA (C. lobatus); and = 10 in sects. Croton (subsect. CycLosticMA), ASTRAEA, and HEPTAL- LON. These results suggest that additional chromosome studies, when cor- related with anatomical data, will give a considerably better insight into relationships and provide the basis for a much more natural classification. Froembling, in a rather comprehensive anatomical investigation, showed that there is a tremendous diversity within Croton, particularly in trichome conformation and in the distribution of secretory (oil-bearing and laticif- erous) elements. Further detailed investigations of trichome types would appear warranted, since it may prove possible to correlate anatomical characters with systematic divisions better than Froembling was able to do. Especially interesting was Froembling’s demonstration that many species of Croton lack the characteristic “latex”’-bearing elements (“Un- gegliederte Milchsaftréhren”’) and that the typical species of sect. ASTRAEA lack oil cells. Compared with other genera of Crotoneae such as Jatropha, Hevea, and Manihot, there are few economically important species of Croton. The best known, doubtless, is the Asiatic species C. Tiglium L., the seeds of which furnish the well-known croton-oil, which possesses properties similar to castor oil or the seeds of Jatropha Curcas. At least one American species furnishes a similar substance in the stems which has been utilized under the name of Cascarilla Bark. Unfortunately, the botanical identity of the 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 359 plant is uncertain due to long-standing confusion, but it appears probable that it is the West Indian Croton Eluteria (L.) Sw. It would not be sur- prising, however, if further investigation were to show that many other species of Croton are also valuable sources of similar substances. REFERENCES: Under family references see ASSAILLY, BAILLON (1858), BENTHAM, Brown, Crorzat (1942; 1945, pp. 188, 189), DEHAY, ERDTMAN, FROEMBLING, GAUCHER, Laketa & CRAIGHEAD, Lanjouw, LEonarp, McVaucH, MARTICORENA, Mr- CHAELIS, MILLER & WEBSTER, MODILEWSKI, MUELLER (1874), Narr & ABRAHAM, Pax (1884), Perry, Punt, Recorp, SmitH, and THATACHAR. BACHMAN, O. Untersuchungen iiber die systematische Bedeutung der Schildhaare. Flora 69: 387-400, 403-415, 428-448. pls. 7-10. 1886. Bracuo, R., & K. J. Crowtey. The essential oils of some Venezuelan Croton species. Phytochemistry 5: 921-926. : Carasta, J. P. El género Croton en Cuba. Carib. Forester 3: 114-135. 1942. [Includes key. ] . The question of Croton Eluteria and Croton Cascarilla, Ibid. 110-113. 1942. [Reviews problem of botanical identity of the Cascarilla plant; bibliography includes earlier references not included here. | Crorzat, L. New and critical Euphorbiaceae chiefly from the southeastern Us: Bull. Torrey Bot. Club 69: 445-460. 1942. [Discusses spp. delimitation in sect. Heptallon. | Preliminari per uno studio del genere Julocroten Martius. Revista Argent. Agron. 10; 117-145. 1943. [An enumeration, without keys; in- cludes a discussion of reasons for maintaining Julocroton as a distinct genus. | Farmer, J. An ecological life history of Croton alabamensis E. A. Smith ex Chapm. 91 pp. Ph. D. diss. (unpubl.). Univ. Alabama. 1962. Fercuson, A. M. Crotons of the United States. Missouri Bot. Gard. Rep. 12: 33-73. pls. 4-31. 1901. FROEMBLING, W. Anatomisch-systematische Untersuchung von Blatt und Axe der Crotoneen und Euphyllantheen. Bot. Centralbl. 65: 129-139, 177-192, 241-249, 289-297, 321-329, 369-378, 403-411, 433-442. pls. 1, 2. 1896. GEISELER, E. F. Crotonis monographiam. x + 83 pp. Halle. 1807. Hooker, J. D. Croton Eluteria, Bot. Mag. 123: pl. 7515. 1897. [Construes the Cascarilla bark species in the same sense as Carabia. | Jounston, M. C. The Texas species of Croton. Southwest. Nat. 3: 175-203. 1959 & B. H. Warnock. The ten species of Croton (Euphorbiaceae) occurring in far western Texas. Southwest. Nat. 7: 1-22. 1962. [Includes descriptions of some spp. found in the southeastern U. S.] Léonarp, J. Observations sur les plantules de quelques Croton du Congo. Bull. Soc. Bot. Belg. 94: 23-28. 1962. [Notes on some systematic implications of seedling morphology. | Marcuanp. L. Du Croton Tiglium. Recherches botaniques et thérapeutiques. 94 p _ 2 pls. Paris. 1861. [Includes excellent illustrations of habit and floral details; botanical part also printed in Adansonia 1: 232-245. 1861.] Mone. C. The last addition to the shrubs of eastern North America (Croton alabamensis). Garden Forest 2: 592, 594. 1889. [Illustration inaccurate in some respects. 360 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ae satay ae Translocation heterozygosity in Croton californicus. Madrono 18: 31. TAILLANDIER, zie ‘Etude al du latex de Croton gossypifolius. 120 pp. ons-le-Saunier. 1939 10. Crotonopsis Michaux, Fl. Bor.-Am. 2: 185. pl. 46. 1803. Delicate annual herbs, stems and foliage with lepidote trichomes; latex monoecious; flowers in abbreviated spiciform racemes, bracts 1-flowered. Flowers with pentamerous calyces; calyx lobes imbricate. Male flowers: petals present, imbricate; disc dissected into 5 small segments; stamens 5, opposite the petals, free, filaments incurved in bud; pollen globose, inaper- turate, clavate; gynoecium vestigial or absent. Female flower: calyx lobes 3—5, more or less unequal and some lobes often obsolete; disc incon- spicuous, deeply lobed or divided; petals absent; carpel solitary; style twice or more bifid; ovary lepidote; ovule solitary, anatropous. Fruit dry, achene-like, indehiscent; seed solitary, not carunculate; seed coat smooth and thicker than carpel wall; endosperm copious; embryo straight, coty- ledons broader than and about as long as the radicle. Type sprectes: C. linearis Michx. (Name alluding to resemblance of the plant to herbaceous species of Croton.) A genus of two species confined to eastern temperate North America. In branching pattern, leaves, and the characteristically reduced inflores- cences, these plants somewhat resemble herbaceous species of Croton, espe- cially C. monanthogynus Michx., of sect. ANGELANDRA. However, the indehiscent unicarpellate fruit with an ecarunculate seed is quite different from the dehiscent fruit with carunculate seeds of C. monanthogynus. While it would scarcely do any violence to nature to combine Crotonopsis with Croton, the group is sufficiently distinctive and well characterized to make it seem best to let it stand as an independent genus Although Gray and Chapman, as well as many other floristic writers, followed Pursh in construing the genus as containing but a single species, Pennell showed that there are two and gave a clear summary of their distinctions. Crotonopsis linearis Michx., recognizable by its narrow leaves with equal-rayed stellate hairs above, elongated 3—6-fruited spikes, and fruits with small pale scales, occurs from South Carolina, south to Florida, and west to Missouri and Texas, with a disjunct population in eastern Iowa and northwestern Illinois. Crotonopsis elliptica Willd., often confused with C, linearis, may ordinarily be distinguished by its relatively broad leaves which are more densely and coarsely hispid-stellate above (due to unequal development of one of the radii of the hairs); more ab- breviated spikes with only one or two female flowers; and fruits with darker and larger scales. Furthermore, in C. elliptica, branching occurs closer to the base so that the plants often are lower and more spreading than the erect, more sparsely branching plants of C. linearis. Crotonopsis 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 361 elliptica has a much broader range than its sister species, as it is known to extend from New Jersey south to northern Florida, and west to Texas, southeastern Kansas, and southern Illinois. According to reports in the literature (e.g., McVaugh; and on specimen labels) it is a characteristic species of bare sandstone or granitic outcrops, although it may occur in disturbed weedy habitats as well. In contrast, C. linearis is most often found in sandy soil in clearings or open woods. REFERENCES: Under family references see FROEMBLING, MICHAELIS. Under Croton see BACHMAN. McVaucH, R. The vegetation of the granitic flat-rocks of the southeastern United States. Ecol. Monogr. 13: 120-166. 1943. [Includes distribution map of C. elliptica. | PENNELL, F. Notes on plants of the southern United States —IV. The genus Crotonopsis. Bull. Torrey Bot. Club 45: 477-480. 1918 (To be concluded) VoLuME 48 NuMBER 4 ~ JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY B. G. SCHUBERT EDITOR T. G. HARTLEY C. E. WOOD, JR. LAZELLA SCHWARTEN CIRCULATION OCTOBER, 1967 Mizrour! RorTen (cat NOV 9 1967 Xo GARDEN LIBRAR PUBLISHED THE ARNOLD A SSACHU SETTS ; fee : THE JOURNAL OF THE ARNOLD ARBORETUM Published quarterly by the Arnold Arboretum of Harvard University. Subscription price $10.00 per year. Volumes I-XX, reprinted, and back issues of volumes XXI-XLYV are available from the Kraus Reprint Corporation, 16 East 46TH Srreet, New York, N.Y. 10017. Subscriptions and remittances should be addressed to Miss Dutere A. Powerit, ArNotp ARBoRETUM, 22 Divinity Avenvz, CamBripce, Massa- cHUSETTS 021, CONTENTS OF NO. 4 Tue GENERA OF EUPHORBIACEAE IN THE SOUTHEASTERN UNITED Srates (Concluded). Grady L. Webster 363 ” FREEMAN AND Custis’ Account of THE Rep River EXPEDITION or 1806, an OveRLooKeD Pusiication or Botanica INTER- est. C. V. Morton . 431 REVISION OF THE GeNus LuNasia (RuTAcEAE). Thomas G. A Synopsis or THE Astan Species or De.pHinrum, SENSU Srricto (Continued). Philip A. Munz . 476 Tue Drrector’s Report .. 546 | Invex to Votums 48 570 Volume No.8 , incl ae 203-84 JOURNAL OF THE ARNOLD ARBORETUM VoL. 48 OcTOBER 1967 NUMBER 4 THE GENERA OF EUPHORBIACEAE IN THE SOUTHEASTERN UNITED STATES * Grapy L. WEBSTER Tribe ACALYPHEAE Dumort. Subtribe Ditaxinae Griseb., “Ditaxideae” 11. Caperonia St. Hilaire, Hist. Pl. Remarq. Brésil. 244. 1825. Annual or perennial herbs, sometimes rhizomatous, often with glandular pubescence. Leaves alternate, undivided, petiolate, serrate, stipulate. Plants monoecious [rarely dioecious]; inflorescences spiciform, axillary, usually bisexual, with 1-5 basal female flowers and several distal male flowers: bracts uniflorous. Male flower: pedicel articulate; calyx usually 5-lobed, lobes valvate in the bud; petals 5, often unequal, basally adnate to the staminal column; disc absent; stamens 10, connate, anthers in 2 superposed whorls; pollen subglobose, reticulate, 6-colporate, germ pores large: vestigial gynoecium terminating the staminal column. Female flower: subsessile [rarely pedicellate]; calyx deeply 5-lobed, with 0-5 smaller supernumerary lobes; disc obsolete; carpels 3; styles 3, more or less deeply 3—7-lobed; ovary usually densely muricate; ovules solitary in each locule, anatropous, nucellus not elongated. Fruit capsular, more or less verrucose; columella slender, apically dilated. Seeds spheroidal, minutely foveolate, with narrow raphe; caruncle absent; endosperm co- pious; cotyledons broader than radicle. LECTOTYPE SPECIES: C. castanei- folia (L.) St. Hil.; see Britton and Wilson, Sci. Surv. Porto Rico 6: 486. 1924. (Named in honor of Noél Capperon, 16th Century pharmacist of Orléans, France, the discoverer of Fritillaria M eleagris L.) A genus of 35 to 40 species, about three-fourths of these native to tropical America and the remainder to Africa. Although placed near Argythamnia in most treatments, the plants look very different from that genus because of their serrate leaves with conspicuously parallel lateral veins and simple, often glandular, trichomes (as opposed to the bifurcate trichomes of Argythamnia). A single introduced species, Caperonia palustris (L.) St. Hil., is now * Continued from volume 48, p. 361. 364 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 found in ditches and swampy areas in Texas, Louisiana, and Florida. It belongs to sect. CAPERONIA because of its annual habit, prickleless stems, and female calyx usually with one or more small supernumerary lobes. conspicuously glandular-pubescent stems, narrow lanceolate stipules, and petals about equalling the calyx in the male flower; whereas C. castanei- folia lacks glandular pubescence, has ovate stipules, and the male petals are exserted beyond the calyx. There are no authentic records of C. castaneifolia in the continental United States. REFERENCES: Under family references see BAILLON (1873, 1874), LANJouw, LEONARD, MIcHAELIS, MILLER & WEBSTER, and O’DoNELL & LouRTEIG. Lronarp, J. Notulae systematicae XX. Contribution a l’étude des Caperonia Africains. Bull. Jard. Bot. Bruxelles 26: 313-320. 1956. [Discusses typifi- cation of C. palustris. ] Pax, F. Euphorbiaceae—Acalypheae—Chrozophorinae. Pflanzenreich IV. 147- VI (Heft 57): 1-142. 1912. [Caperonia, 27-49.] 12. Argythamnia P. Browne, Civ. Nat. Hist. Jamaica 338. 1756. Herbs or shrubs; stems and foliage with indumentum of bifurcate (mal- pighiaceous) hairs, tissues often stained with purplish pigment. Leaves alternate, petioles short, stipules small; blades entire or dentate, not glandular. Plants monoecious [rarely dioecious]; inflorescences short, racemiform, bisexual, with one to few basal female flowers; bracts small, uniflorous. Male flowers: calyx [3—]5-lobed, lobes valvate in the bud; petals 5, narrowed to a claw at base, more or less adnate to base of staminal column; disc dissected, segments opposite the calyx lobes; sta- mens 5-15, monadelphous; anthers 1—2(-3)-seriate; 1-3 filiform stam- inodia sometimes present atop staminal column; pollen oblate, tectate, tricolpate, bilaterally symmetrical; vestigial gynoecium absent. Female flower: calyx 5-parted, lobes imbricate; petals 5, imbricate, entire, some- times reduced; disc more or less dissected into sometimes elongated seg- ments; carpels 3; styles free or basally connate, bifid, style-branches more or less dilated at the tips; ovules 1 in each locule, anatropous, nucellus not elongated. Fruit capsular; columella persistent. Seeds subglobose, foveolate to reticulate, caruncle absent: endosperm present; embryo straight, cotyledons broader than the radicle. (Including Ditaxis Vahl.) Type species: A. candicans Sw. (Name from Greek, argios, white, and thamnos, shrub; the type species noted by Browne as having white bark.'®) “The generic name has been variously spelled Argithamnia, Argothamnia, and Argyrothamnia; but Bentham (Gen. Pl. 3: 304) pointed out that the last name was based on Mueller’s erroneous derivation from argyros, silver. As affirmed by Wheeler (Contr. Gray Herb. 127: 55, 56. 1939), there seems to be no reason why the original spelling should not be followed. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 365 An American genus of about 50 species, mostly tropical but some attain- ing temperate latitudes in the southern United States and in Argentina. The circumscription adopted here is approximately that of Bentham, who followed Mueller in including Ditaxis within Argythamnia but excluding Speranskia. Pax in 1912 adopted a radical view in which most of Mueller’s sections (e.g., Philyra, Chiropetalum) were treated as genera. The pollen studies of Punt provide some support for Pax’s view, since most of the taxa in question have easily distinguishable pollen types. Argythamnia is rather closely related to our other genus of Ditaxinae, Caperonia. How- ever, it not only differs strikingly in its entire, triplinerved leaves, but is technically separable on the basis of its malpighiaceous hairs, male flowers without a rudimentary gynoecium, three- or four-colporate tectate pollen, and ornamented seeds. malpighia- 5 Our representatives of Argythamnia all belong to subg. D1TAxIs (Vahl) sually biseriate androecium of Croizat, which includes the species with a u 366 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 approximately 10 stamens and bifid styles (the branches sometimes bi- lobed). Section SEROPHYTON (Benth.) Webster ' is represented by two species which have been reported from Arkansas: A. mercurialina (Nutt.) Muell. Arg., with elongated inflorescences (over 5 cm. long), petals ob- solete or wanting in the female flower, and large (4-5 mm.), wrinkled seeds; and A. humilis (Engelm. and Gray) Muell. Arg., with abbreviated inflorescences scarcely 1 cm. long, petals present in the female flower, and small (ca. 2 mm.), reticulate seeds. In Florida, sect. Diraxis [ (Vahl) Muell. Arg.| is represented by A. Blodgettii (Torr.) Chapm.; this species differs from members of the preceding section in having the petals adnate to the staminal column. Pax followed Mueller in assigning the Florida plants to A. Fendleri Muell. Arg. [= A. argothamnoides (Bert.) Ingram], a species otherwise known from South America. The plants involved do look much alike, but differ in various floral details (e. g., petals broader and staminodia ciliate-tipped in A. Blodgettii) and in the sculpturing of the seed-coat. Apparently, therefore, A. Blodgettii is a distinct species endemic to pine woods and hammocks on limestone in extreme southern peninsular Florida and the Florida Keys. A single chromosome count of m = 13 has been recorded in A. Brandegei Millsp., a species of Mexico. Our local species have not yet been studied cytologically, but if they prove to have the same number, this would provide interesting evidence toward defining relationships in the subtribe Ditaxinae. So far counts in Chrozophora and Caperonia show x = 11, indicating a more distinct separation of Argythamnia than might otherwise be suspected. REFERENCES: Under family references see BAILLON (1873, 1874), GAUCHER, MICHAELIS, Punt, WHEELER (1939). IncRAM, J. New species and new combinations in the genus Argythamnia. Bull. Torrey Bot. Club 80: 420-423. 1953 ———. New names in Argythamnia subgenus Ditaxis. Ibid. 84: 421-423. 1957. . A revisional study of Argythamnia subgenus Argythamnia (Euphor- biaceae). Gentes Herb. 10: 1-38. 1967. Jounston, M. C., & B. H. Warnock. “The four kinds of Argythamnia (Euphor- biaceae) in far western Texas. Southwest. Nat. 7: 154-162. 1962. [Includes description of A. mercurialina. | Subtribe Mercurialinae Pax 13. Mercurialis Linnaeus, Sp. Pl. 2: 1035. 1753; Gen. Pl. ed. 5. 457. 1754. Annual |or perennial] herbs [sometimes rhizomatous|; leaves opposite, stipulate, petiolate, more or less pellucid-punctate, with crenate margins. Plants dioecious (rarely a few male flowers developed on female plants) “ Argythamnia sect. Serophyton (Benth.) Webster, comb. nov. Serophyton Benth. Bot. Voy. Sulphur 52. 1844. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 367 [or monoecious in some polyploid taxa]; flowers in modified thyrses, the disc absent; stamens 8-15 (—20), free, filaments variously oriented in bud, anthers extrorse, anther-sacs subglobose, dehiscing transversely; pollen ellipsoidal, tectate, tricolporate, colpi narrowly operculate; vestigial gy- noecium absent. Female flower: calyx Icbes 3, imbricate; disc absent; 2 elongate staminodia usually present; carpels 2 (very rarely 3); styles free except at base, unlobed, conspicuously papillate; ovule 1 in each locule, anatropous, nucellus not elongated. Fruit capsular; valves echinate; columella slender. Seeds not compressed, testa rugose, caruncle present but small and delicate; endosperm present; cotyledons broader than and about equalling the radicle. LecroTypE species: M. perennis L.; see Small in Britton and Brown, Illus. Fl. No. U. S. ed. 2. 2: 460. 1913. (Classical name used by Pliny to honor the god Mercury, the mythical discoverer of the therapeutic properties of the plant.) — MERcurY. A small genus of seven or eight Old World species, mostly in the Med- iterranean region but with one or two taxa in China, Japan, and Formosa. The genus is similar to Acalypha but lacks the highly specialized male flowers of that genus, and it is easily distinguished by its opposite leaves and small female bracts. In fact (cf. Pax, 1914), the closest relationships are probably with three small genera of Pax’s series Mercurialiformes: Seidelia and Leidesia, from South Africa, and Dysopsis, from the southern Andes. The only species of Mercurialts naturalized in the United States is M. annua L., which has been reported from widely scattered localities mented on by Camerarius (1691) in his demonstration of sexuality in plants; he simply isolated female plants and proved that seeds did not occur in the absence of fertilization. Actually, he may have been lucky in getting positive results, for exactly three centuries later Kerner made the claim that isolated female plants did set seed and that M. annua was therefore parthenogenetic. Further investigations showed that Kerner’s observations were correct but his explanation wrong. Isolated female eds, but Bitter and Malte showed that megasporogenesis and seed development is normal and that the apparent parthenogenetic fruits are actually produced as the result of fertilization by scattered reduced male flowers on otherwise strictly female plants; these small male flowers develop rapidly and are difficult to detect. The remarkable variability in sexual expression in mercuries has at- tracted the interest of many investigators (e.g., Yampolsky), but for a long time no very convincing explanations could be offered. The key to the problem was found only recently, when Durand and Thomas discov- 368 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 ered that different kinds of sexual expression were correlated with a remarkable polyploid series within M. annua (sensu lato). Durand has shown that within the Linnaean M. annua there are no less than seven levels of ploidy: from 2x (2m = 16) to 14x (2n = 112). Strictly dioecious plants are found only in the weedy diploid species, M. annua (sensu stricto), which is the form immigrant in the United States and elsewhere. In south- that the opposite is more likely true. As in Castanea, the flowers of Mercurialis seem to have acquired many anemophilous characteristics, 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 369 without however entirely losing the potentiality of a certain lesser amount of insect pollination. Cytological studies have yielded much information of taxonomic and evolutionary interest, as already partly discussed for the M. annua com- plex. The base number in all species investigated of Mercurialis is defi- nitely « = 8, which contrasts with the base numbers of 11 and 13 in sub- tribe Ditaxinae, and with x = 10 in Acalypha, which has been placed close to Mercurialis in many taxonomic treatments. So far no other morpholog- ically similar genus with x = 8 has been detected. Gadella & Kliphuis have made a remarkable report that in M. perennis 2n = 64 in the female and 66 in the male. If confirmed, this would apparently mean that the mechanism of sex inheritance is different from that found in M. annua, where there is no difference in number nor any morphologically detectable sex chromosomes. REFERENCES: Under family lied see DeHay, EICHLER, GAUCHER, MICHAELIS, Noz- ERAN, PAx & HoFFMANN (1914, 1930), PunT, RITTERSHAUSEN, SCHWEIGER, Vint, and ZIMMERMANN BERGFELD- GAERTNER, ae PicickledsipeathichuieSe und cytologische Unter- suchungen der “von Hanstein’schen Zellen” bei Mercurialis annua und perennis L. Zeitschr. Bot. 52: moe 302. 1964. Bitter, G. Zur Frage der Geschlechtsbestimmung von Mercurialis annua durch Isolation weiblicher Pflanzen. Ber. Deutsch. Bot. Ges. 27: 120-126. 1909. BLARINGHEM, L. Etudes sur le polymorphisme florale III. Variations de sexual- ité en rapport avec la multiplication des carpelles chez le Mercurialis annua L. Bull. Soc. Bot. France 69: 84-89. 1922 Danc-VAN Liem. Embryogénie des Euphorbiacées. Développement de ]’embryon chez le Mercurialis perennis L. Compt. Rend. Acad. Sci. Paris 250: 3217- 3219. 1960. [Considers embryonic development similar to that in Buxus.] Duranp, B. L’organisation morphologique de la fleur des Mercuriales annuelles. Nat. Monspel. Bot. 8: 105-124. 1956. . Les groupements de fleurs chez les Mercuriales annuelles. bid. 9: 21-43. 1957. Polymorphisme, polyploidie et répartition des sexes chez les Mercuriales annuelles. Compt. Rend. Acad. Sci. Paris 244: 1249-1251. . Le complexe Mercurialis annua L, s. 1. — une étude biosystematique. Ann. Sci. Nat. Bot, XII. 4: 579-736. 1963. Gaze, D. R. Inheritance of sex in Mercurialis annua L. in relation to cytoplasmic theory of inheritance. Dokl. Akad. Nauk. SSSR II. 23: 478-481. GabELLA, T. W. J., & E. Kiipnvuts. Chromosome numbers of flowering plants in the Netherlands. Acta Bot. Neerl. 12: 195-230. 1963. [Reports different Gittot, P. Observations sur le polymorphisme floral de Mercurialis annua L. Bull. Soc. Bot. France 71: 684-692. 1924. Remarques sur le déterminisme du sexe chez Mercurialis annua L. mpt. Rend. Acad. Sci. Paris 178: 1994, 1995. 1924 Hebie-Toeenen , J., & Y. Hestop-Harrison. The effect of carbon monoxide on sexuality in Mercurialis ambigua L. fils. New Phytol. 56: 352-355. 1957. KERNER VON Maritaun, A. The natural history of plants. Half-vol. III. 370 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 (Transl. & ed., F. W. OLtver.) iv + 496 pp. pls. 9-14. London. 1895. [Remarks on experiments to demonstrate parthenogenesis in M. annua, 465, 466; German ed. of this volume publ. 1891. | Mukerji, S. K. Contributions to the autecology of Mercurialis perennis L. Jour. Ecol. 24: 38-81, 317-339. 1924. [Includes distribution map of most taxa, discussion of interspecific relationships. | Porteres, R. Taux sexuel chez Mercurialis annua L. Jour. Agr. Trop. Bot. Appl. 3: 443-445. 1956. SAUNDERS, J. Monoecious and hermaphrodite esbaboertente perennis. Jour. Bot. 21: 181, 182. 1883. [Nectaries said to represent staminodia. Souvitte, M. G. Observations sur le uae de Mercurialis annua L. Revue Gén. Bot. 37: 49-62. Tuomas, R. G. Effects of temperature and length of day on the sex expression of monoecious and dioecious angiosperms. Nature 178: 552, 553. ; | Reports experiments on M. ambigua . Sexuality in Picci and hexaploid races of Mercurialis annua L. Ann. Bot. II. 22: 55- 958 Troupin, G. Etude de la croissance et = la sexualité de Mercurialis annua L. Bull. Soc. Bot. Belg. 79: 96-99. Weiss, F. E. Die Bliitenbiologie von tous Ber. Deutsch. Bot. Ges. 24: 6 501-505 1 2 WETTSTEIN, R. von. Das he igor rp ee Bliiten bei Mercurialis. Ber. Deutsch. Bot. Ges. 34: 829-8 pl. 24. 1917. YAMPOLSKY, C. Sex intergradation in wie i. of Mercurialis annua, Am. Jour. Bot. 7: 95-100. pl. 5. 1920. The cytology of the intersexual flowers of Mercurialis annua —a morphogenetic study. Jbid. 21: 651-672. 1934. [Includes references to the author’s numerous earlier papers. | Subtribe Acalyphinae Griseb., ‘““Acalypheae” 14. Acalypha Linnaeus, Sp. Pl. 2: 1003. 1753; Gen. Pl. ed. 5. 436. 1754. Annual or perennial herbs or shrubs [rarely trees]; indumentum of simple hairs or glands. Leaves alternate, stipulate, petiolate, unlobed, often punctate. Plants monoecious or rarely dioecious; inflorescences spiciform [rarely paniculate|], unisexual or bisexual, terminal or axillary; female flowers 1-3 per bract, bracts large and foliaceous, often lobed or parted [rarely small and entire]; male flowers several per bract, bracts very small. Flowers apetalous, disc absent. Male flowers subsessile; calyx valvate, parted into 4 segments; stamens 4-8, filaments free or connate at base; anther- -sacs pendent, unilocular, more or less elongated and vermi- form: pollen oblate-spheroidal, 3-5-pseudoporate, tectate, psilate; vestigial gynoecium absent. Female flowers sessile; calyx lobes 3{—5], imbricate; carpels 3 (rarely only 1 or 2); styles free or basally connate, laciniate or lacerate into filiform segments [rarely subentire]; ovary smooth, pubes- 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 371 smooth to pitted or tuberculate; endosperm whitish; embryo straight, cotyledons reniform, broader than the radicle. LECTOTYPE sPECIES: A. virginica L.; see Small in Britton and Brown, Illus. Fl. No. U. S. ed. 2. 2: 457. 1913. (Name from Greek, akalos, unattractive, and aphe, touch; Linnaeus’s derivation obscure, as the plants he placed in the genus are not stinging.) A very natural genus of about 400 species, the majority American, and the largest concentration (about 100 species) in Mexico and Central America. The genus is isolated in the Euphorbiaceae, as the only other taxon in the subtribe Acalyphinae is the monotypic Acalyphopsis, based on an aberrant species from Celebes. Mueller and others have placed Acalypha close to Mercurialis, perhaps because of the similarity in habit and wind-pollinated flowers. However, the basic chromosome number and pollen grains are different in the two genera, and the resemblances ma prove to be superficial. As suggested by Pax & Hoffmann, the relationship of Acalypha probably lies with some group within the subtribe Mercurial- inae, but a more exact affinity cannot be pointed out at this time. The infrageneric classification of Acalypha presented by Pax & Hoff- mann was modelled closely on that of Mueller, except that the ranks of the subgeneric taxa were inflated so that no less than 40 sections were recog- nized in place of the two proposed by Mueller. Because of the narrow amplitude of divergence among the species of Acalypha, these “sections” of Pax and Hoffmann are really comparable to subsections or series in other genera of Euphorbiaceae, and they are so treated here.’ “Pax & Hoffmann Sanne 147-XVI(Heft 85): 12. 1924) derive the generic name from akalyphes, not covering, in apparent allusion to the female bracts; but this is contaien by en ea explicit derivation of the name (Philos. Bot. te is difficult to reconcile the systems of infrageneric a erected by students and w of saiateae with biological reality, on the o d, h the requirements of the present f N re on the er. Hurusawa pe ac. Sci. Univ yo t. 6: 295-301. 1954) has inflated the subgeneric taxa to an e eate ann. Mueller’s division of (and more unacceptable) extent than did Pax and Hoftm the genus into two sections with many infrasectional taxa app relationship better, but on the basis of his own explanation (Linn 4 his ultimate taxa designated by the sign “§” were not near “ag simply “Gruppe.” antedly treated these groups a ugh Mueller had pub- Ss are ae all illegitimate under ank here, and pre 8 series Rees in ine United States or mentioned in this discussion are then as follows: (1) Series Sclerolobae Muell. arn » Ser. NOV.; os on § Sclerolobae Muell. Arg. Linnaea 34: 23. 1865. Lectotype: A. setosa (2) Series Plumosae Muell. Arg., ser. nov.; cenit on § Plumosae Muell. Arg. loc. cit. 49. Lectotype: A. alopecuroidea Jac (3) Series Dhlevidens Muell. Arg., = nov.; based on § Phleoideae Muell. Arg. loc. cit. 46. Lectotype: A. phleoides gd (4) Series Brachystachyae Muel Arg r. nov.; based on § Brachystachyae Muell. Arg. loc. cit. 42. LECTOTYPE: - oe eke Hornem. 372 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Of the 17 species of Acalypha recorded from the United States, nine species in four series occur in our area. All of these plants belong to sect. AcALYpHwa because of the sessile female flowers with large bracts; the few taxa in sect. Lrnostacuys (KI.) Muell. Arg., characterized by pedicellate female flowers subtended by small bracts, occur mostly in tropical America and do not reach our limits. The basic division of the genus was made by Mueller largely on the basis of the relative position of the female and male flowers, which may be either proximal or distal on terminal or axillary unisexual or bisexual spikes. The group Pantogyne-Acrogyne of Mueller, distinguished by uni- sexual spikes with the female spike terminal, is represented in our area by ser. SCLEROLOBAE Muell. Arg. ex Webster, which is characterized by nar- row female spikes and deeply lobed female bracts. The two local repre- sentatives are A. ostryifolia Ridd., with 13—17-lobed female bracts, occur- ring throughout our area; and A. setosa A. Rich., with mostly 7—8-lobed female bracts, restricted to the Coastal Plain at scattered localities from South Carolina to Florida and Louisiana. The latter, native to the West Indies and Central America, is strictly a weed in the southeastern United States and appears to have been introduced. Mueller’s group Polygynae-Acrogyne, in which the spikes are usually bisexual and terminal (often with axillary ones as well), is represented by two series, each with a single species. Acalypha alopecuroides Jacq., belonging to ser. PLuMosAE Muell. Arg., and recognizable by its dense, thickly pubescent spikes, has been sparingly introduced into southern Louisiana from tropical America. Acalypha chamaedrifolia (Lam.) Muell. Arg., of ser. PHLEOIDEAE Muell. Arg., is a native species restricted in our area to extreme southern Florida and the Keys; its principal distribution is in the West Indies. It is easily distinguished from our other species by the perennial habit and small seeds (usually only about 1 mm. long). Most of our local plants of Acalypha belong to Mueller’s group Polygy- nae-Pleurogynae, in which the spikes are bisexual and strictly axillary. Our four or five species, the weedy members of the notorious Acalypha virginica complex, are all annuals referable to ser. BRACHYSTACHYAE Muell. Arg. because of their abbreviated spikes with few female bracts. Weatherby recognized 3 species in this group, one with two varieties; (5) Series Boehmerioideae Muell. Arg., ser. nov.; based on § Bochmerioidene Muell. Arg. DC. Prodr. 15(2): 871. 1866. panei A, hashed Mig. [ = 4. lanceolata Willd. (6) Series Repandae Muell. Arg., ser. nov.; eee on § Repandae Muell. Arg. oe 34:4 65. LECTOTYPE: A ‘mollis Series magrerte Sean Muell. Arg., ser. nov.; based on § Monostachyae Muell. Arg. by cit. 53. Lectotype: A. monostachya Cav (9) Series Palnines vine Muell. Arg., ser. nov.; based on § Palminervia Muell. Arg. loc. cit. 9. TOTYPE: A. macrostachya HBK. (10) Series Caturoideae (Pax & Hofim.) Webster, stat. nov.; based sect. Caturoideae Pax & Hoffm. Pflanzenreich 147- XVI(Heft 85): 139. 1924. ema: A. Caturus Blume. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 373 whereas L. Miller more recently has distinguished five species, one with two varieties. The group is taxonomically difficult, and Steyermark has described putative hybrids between three species in Missouri; but hybrid- ization has still not been conclusively demonstrated. Weatherby in a critical study of typification established that the name A. virginica L. is to be associated with the relatively narrow-leaved plant having female bracts cut mostly into 10-14 lanceolate lobes with long, spreading hairs. This species is common in Arkansas, Tennessee, and North Carolina but almost completely avoids the Coastal Plain. Far more abundant, widespread, and weedy is A. rhomboidea Raf., which has broader leaves and mostly 7—9-lobed female bracts that lack long hairs. It is common throughout the eastern United States and is the only species which crosses into Canada (in Ontario) ; in our area it is found both in the highlands and the Coastal Plain but does not enter peninsular Florida. Very similar to A. rhomboidea in appearance is A. Deamii (Weatherby) Ahles, a sibling species only recently discriminated. It has leaves and bracts much like those of A. rhomboidea, but differs in its two- rather than three-seeded fruit and in its larger seeds (2.2—3.2 mm. long vs. 1.2—-2.0 mm, in A. rhomboidea). Acalypha Deamii seems to be restricted to more meso- phytic sites and is known only from scattered localities in the Mississippi- Ohio River drainage, entering our area in Arkansas and Tennessee. The most controversial populations in ser. BRACHYSTACHYAE include those to which the name A. gracilens Gray has been applied. These are plants with narrow leaves, the margins nearly or quite entire, and with much shallower lobing on the female bracts than is found in A. rhomboidea and related taxa. Sj leis A still unpublished analysis of this group made by L. Miller indicates that there are at least two major taxa involved. Throughout the Coastal Plain in our area, and in some inland localities as well, occurs ssp. gracilens, with leaves crenate-lanceolate and staminate spikes very short. To the west, this population is more or less replaced by an Ozarkian population which extends from central Texas into Oklahoma, Arkansas, and Missouri. Plants of this taxon, which differ in their narrower leaves, long male spikes, and, most strikingly, in having one-seeded instead of three-seeded capsules, may be designated as A. gracilens ssp. monococca (Engelm. ex Gray) Webster.2° Lillian Miller regards this taxon as meriting specific rank, and it is possible that further investigation may substantiate her viewpoint. Field studies are needed in eastern Texas and Louisiana, where there are plants with the narrow leaves and long male spikes of ssp. monococca combined with the 3-seeded capsules of ssp. gracilens. These plants have been interpreted by Miller as an undescribed variety of A. gracilens, but the alternative explanation (followed here) is that this more or less intermediate population may represent a transition between the two subspecies. Additional studies of the natural populations may eventually provide a generally acceptable answer. Engelm. ex Gray) Webster, stat. nov. A. — i : nococca ( acd seth te tedlons Mn US. ed. 2. 408. 1856. gracilens var. monococca Engelm. ex Gray, Man. Bot. No. 374 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. 48 Cytological studies in Acalypha have so far led to conflicting results, since earlier workers such as Perry and Landes reported x = 7, whereas more recent reports (Kapil; Miller; Miller & Webster) suggest a base number of x = 10. Miller found 2n = 40 in A. virginica, A. Deamii, and A. monococca, which suggests that at least the taxa of ser. BRACHYSTACH- YAE are tetraploids. Miller & Webster found both tetraploid and octo- ploid levels among three tropical American species. Perhaps the most interesting aspect of reproductive biology in Acalypha is its embryology. In contrast to most genera of Euphorbiaceae, only tet- rasporic embryo sacs have been reported in Acalypha, and these are of several different kinds; reviews of embryo-sac diversity have been pro- vided by Mukherjee and by Kapil, among others. According to Mukherjee the basic type of female gametophyte in Acalypha is the Penaea type, in which the mature embryo sac typically contains four egg-apparatuses (each of three cells) and four polar nuclei; this type has been reported in A. rhomboidea by Landes, as well as in several Asiatic species. In A. indica L. there are eight polar nuclei which fuse, and the four lateral groups are two-celled (Maheshwari & Johri). Finally, there is a report in A. lanceolata Willd. of the Peperomia hispidula type, in which there is one two-celled egg-apparatus and a 14n polar fusion-nucleus. The taxonomic implications of this diversity are obscure, partly because of problems in identification. For example, Banerji reported the Acalypha indica type in A. fallax Muell. Arg.; that name, however, is a synonym of A. lanceolata Willd., of ser. BoEHMERIOIDEAE, in which Thatachar reported the Peperomia hispidula type. Furthermore, the results so far do not cor- relate well with the system of Pax & Hoffmann. In A. malabarica the em- bryo-sac is of the least specialized Penaea type, whereas in the closely related A. lanceolata the presumably most specialized Peperomia hispidula type has been reported. In a similar, although somewhat less disconcerting manner, A. australis L. and A. indica L. are reported to have the Penaea type and A. indica type respectively, even though they are closely related species of ser. REPANDAE. It may therefore be questioned whether the embryological results thus far throw any light on evolutionary relation- ships within the genus. However, further studies on a wider range of care- fully identified taxa, especially on the less specialized representatives of sect. LINostAcHys, are needed before we can dismiss the embryological work as completely irrelevant taxonomically. In marked contrast with most other sizeable genera of Euphorbiaceae, Acalypha is almost devoid of species with economic utility, although various species are cultivated as tropical ornamentals. Acalypha Wilkes- tana Muell. Arg., a rather unattractive species of ser. PALMINERVIAE Muell. Arg., probably of Fijian origin, is grown presumably because of its variegated foliage. Much more striking is A. hispida Burm f., of ser. CaTUROIDEAE (Pax & Hoffm.) Webster, possibly of Papuasian origin, which has attractive reddish female spikes. Both are grown in southern Florida, but there is no evidence that either has become naturalized. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE Ags REFERENCES: Under family references see ARNOLDI, DEHAY, GAUCHER, HurusAwa, Lovur- TEIG & O’DoNELL (1942), McVaucH, MIcHAELIs, MILLER & WEBSTER, MUELLER (1865), Murrey, Punt, Recorp, RITTERSHAUSEN, and WILLIAMS. BANERJI, I. A contribution to the life-history of Acalypha fallax Muell. Arg. Bull. Bot. Soc. Bengal 3: 29-32 Jounston, M. C., & B. A. Waite: The four species of Acalypha (Euphor- biaceae) in far western Texas. Southwest. Nat. 7: 182-190. 1962. [In- cludes description of A. ostrytfolia. | Jour, B. M., & R. N. Kapri. Contribution to the enSEEUPIEY and life history of Acalypha indica L. Phytomorphology 3: 137-151. Kajsace, L. B., & K. S. N. MurtHy. The embryo-sac of ede ciliata Forsk. Jour. ele Bot. Soc. 33: 417-422. 1954. ee os N. Embryology of Acalypha Linn. Phytomorphology 10: 174-184. Pres L. Dimorphic carpellate flowers of Acalypha indica L. Jour. Indian Bot. Soc. 1; 1-5. 1919. Lanpes, M. Seed development in Acalypha rhomboidea and some other Euphorbiaceae. Am. ibs! Bot. 33: 562-568. 1946. Mauesuwakrt, P., & B. M. Jourt. The embryo sac of Acalypha indica L. Beih. Bot. Centralbl. 61A: 125- 136. 1941. Mitter, L. W. A taxonomic study of the species of ay re in the United States. 198 pp. Ph.D. diss. (unpubl.). Purdue Univ. & G. L. Wesster. Population studies on the Indiana nee of Acalypha. Am. Jour. Bot. 48: 548. 1961. Muxuerjeg, P. K. The female gametophyte of Acalypha malabarica Muell. with a brief discussion on the Penaea type of embryo sac. Jour. Indian Bot. Soc. 37: 504-508. 1958. NitscHkE, R. Die Sash Verbreitung der Gattung oe Bot. Arch. 4: 277-317. 1923. [A tedious and uninteresting compilati PAX PS UR: pee MeAnn Pee Hct tse CoscansiteaNcaly AP RADY! phinae. Pflanzenreich IV. 147-XVI(Heft 85): 1-231. Pratn, D., & J. HutcHrnson. Notes on some species of Acalypha. Bull. Misc. Inf. Kew 1913: 1-28. 1913. [Discusses A. a and related taxa. | SreYERMARK, J. A. Flora of Missouri. Ixxxiii + 1725 pp. Ames, ei 1963. [ Acalypha, 978-982; Ageaee some putative Spaeth hybri Swamy, B. G. L., & B. P. BALAKRISHNAN. Female gametophyte a Acalypha tricolor. Jour. eh Bot. Soc. 25: 67-69. 1946. Taretsut, S. On the development of the embryo-sac and fertilization of Acaly- pha australis L. (In Japanese; English summary.) Bot. Mag. Tokyo 41: 477-485, pls. 15-17. 192 THatTacHarR, T. Mo esladical studies in the ir obmenaas I. Acalypha lanceolata Willd. Phytomorphology 2: 197-201. 19 Weatuersy, C. A. The group of Acalypha virginica in see North America. Rhodora 29: 193-204. 1927. [Revision, with key. . The typification of Acalypha virginica L. Ibid. 39: 14-16. 1937. ae AEE. enero! types in the group of Acalypha virginica. Ibid. 42: 96. 940 376 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Subtribe Plukenetiinae Pax 15. Tragia Linnaeus, Sp. Pl. 2: 980. 1753; Gen. Pl. ed. 5. 421. 1754. Perennial herbs, sometimes suffrutescent, decumbent to erect or twining; indumentum of uniseriate hairs, stalked glands, and stinging hairs. Leaves alternate, stipulate, petiolate or sessile, entire to more commonly toothed or lobed [rarely divided]. Plants monoecious; inflorescences racemiform, bisexual, opposite the leaves or terminal on lateral branches; female flowers at 1 to several proximal nodes, remaining nodes male; bracts small, sub- tending solitary flowers. Flowers apetalous, calyx 3—6-lobed, disc absent. Male flower: pedicel articulate; calyx lobes valvate; stamens 2—5 (-8) [-50]; filaments connate at least at base, anthers dehiscing longitudinally; pollen spheroidal, finely reticulate, tricolpate with colpi operculate, or inaperturate; rudimentary gynoecium small or absent. Female flower pedicellate; staminodia absent; carpels usually 3: styles 3, united at least at base, unlobed, often papillate; ovary hispid with stinging hairs; ovules 1 in each locule, anatropous, nucellus not beaked. Fruit capsular: columella persistent, with 3 apical interlocular points. Seeds subglobose, smooth or slightly roughened, not carunculate: endosperm whitish; cotyle- dons foliaceous, considerably broader than the terete radicle. LECTOTYPE SPECIES: T. volubilis L.; see Small in Britton and Brown, Illus. Fl. No. U.S. ed. 2. 2: 458. 1913. (Name in honor of the German herbalist Hier- onymus Bock, 1498-1554, whose name was latinized as Tragius.) A rather protean genus of more than 100 species found mostly in the tropical regions of Africa and Latin America, but with a few in warm- temperate areas. The intrageneric diversity of Tragia is indicated by Mueller’s recognition of 12 sections, nine of which were retained by Pax & Hoffmann. Various segregates, such as Ctenomeria, have been recognized in the past, and the generic limits still present an interesting problem for study. Punt pointed out that the taxa in sect. BIA (K1.) Muell. Arg. have nonaperturate pollen and suggested that this group possibly might warrant generic status. The discovery of a somewhat transitional kind of pollen in some species of the United States (Miller; Miller & Webster) favors the retention of Bia within Tragia, but the issue can still not be regarded as settled. : To some extent, the taxa of Tragia in the United States represent the outliers of mainly tropical groups, but there are several species which are primarily of eastern temperate North American origin. Both sections found in the United States are represented in the southeastern United States. Section Tracra, which includes those species with mainly 3-5 stamens and tricolpate pollen grains, is represented by 11 species in the United States. These include the weedy “nose-burns” characteristic of disturbed areas in Texas and adjacent states, which Johnston interpreted as all belonging to a single species, T. nepetifolia Cav. In the judgment of Miller & Webster, however, that species is restricted to Mexico and Arizona, while some of the taxa reduced to synonymy by Johnston are accepted as 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 377 distinct species. Five species of sect. TRAGIA enter our area; the common- est is T. urticifolia Michx., found from North Carolina to Texas and Arkansas and barely entering northern Florida. It occurs mostly in the Piedmont and foothill regions and enters the Coastal Plain at only a few points. It is easily separated from related species by the long, persistent base of the staminate pedicel, which equals the bract (in other species it is much shorter); and it often has conspicuously hispidulous stems. Two species which superficially resemble 7. urtici folia enter our area from the west in Arkansas. Tragia ramosa Torr., which includes the majority of populations placed in T. nepetifolia by Johnston, is an extremely variable species which may, however, be generally recognized by its small female calyx lobes (shorter than the gynoecium) and its typically narrow (sometimes even linear) upper leaves. Often confused with it is T. betoni- cifolia Nutt., which has broader leaves like those of 7. urticifolia and a larger female calyx with lobes which extend above the tips of the styles. More isolated both geographically and morphologically is T. saxicola Small, which has very slender, wiry stems and broad, sharply toothed leaves; the female flowers and seeds are smaller than those of T. betonici- folia, which it somewhat resembles. It is restricted to pine woods growing on limestone in the Florida Keys and the adjacent Florida mainland. Occupying a rather divergent position among our species of Tragia is T. cordata Michx., a twining vine with large deeply cordate leaves and large seeds (more than 4 mm. long). Although widespread, it has a spotty distribution and seems nowhere very abundant; in our area it is known from a few localities in Arkansas and Louisiana, east to Tennessee, Georgia, and extreme northwestern Florida. It appears to be distinctly less weedy than other species and is probably restricted to relatively undis- turbed tracts of deciduous forest. Our two other species of Tragia constitute the section LEPTOBOTRYS (Baill.) Pax, which is nearly confined to our area. Tragia urens L., ex- tending from southeastern Virginia to southern Florida and west to eastern Texas, is the most abundant representative of the genus on the Coastal Plain. Uniquely, it has terminal inflorescences (on lateral leafy twigs) and very characteristic leaves whi sharply toothed as in many con recognized as a distinct species seems unjustified. Tragia Sma cium and pollen of T. urens, toothed leaves and in its inflorescences, wh leave: in all other temperate North American taxa of Tragia. It is primarily a species of the Gulf Coastal Plain, central Florida west to extreme ea of section Tracta both of these species have two stam colpate pollen. ; Miller (1963) reported the first chromosome counts in the genus, and known for six species. In both sect. TRAcra chromosome numbers are now ; c number is apparently 11, as five species and sect. LepTrosotrys the basi 378 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 are tetraploids with 2 m = 44; an extralimital species, Tragia amblyo- donta (Muell. Arg.) Pax & Hoffm., is apparently a decaploid, with 2” = ca. 110. Unfortunately, the significance of these counts cannot be eval- uated at the generic level, since not a single other genus of this subtribe has been studied cytologically. Pollinating mechanisms are still scarcely known in Tragia, but one curious feature of the reproductive biology has now been detected in several species. Ule long ago pointed out that some South American twining plants of Tragia (T. volubilis L. vel aff.) produce dimorphic fruits: in addition to the usual three-locular capsules there are some inde- hiscent one-seeded fruits which have two or three long horns. Miller found similar fruit dimorphism in the North American 7. brevispica Engelm. & Gray; here the production of two kinds of fruit is an important taxonomic character, since it provides one of the main distinctions between that species and the very similar T. ramosa, which never seems to produce aberrant capsules. In T. brevispica the one-seeded, winged fruits develop from a three-locular ovary with normal ovules, but two of these abort and the outer part of the ovary wall develops unequally to produce the wings. Although Ule rather naively suggested that the winged fruits functioned to raise the plants’ competitive ability through animal dispersal, the role the fruits actually play during the life cycle has not yet been elucidated. In a number of South American species horned or crested three-locular fruits occur, as illustrated for T. Paxii Lourt. & O’Don., so that a tendency to produce “aberrant” fruits appears to be widespread in the tropical American taxa of the genus.2! The distinctive stinging hairs of Tragia were studied by Knoll and found to be of subepidermal origin. Their structure is unique and unlike the nettle-hairs of Cnidoscolus and various Urticaceae, because in Tragia the stinging element is a large, sharp-pointed crystal which represents one arm of a highly modified druse. Such a crystal type is found in no other family, and the other genera of Euphor- biaceae possessing it (e.g., Cnesmosa, Acidoton, and Dalechampia) are probably all related to Tragia. REFERENCES: Under rene references see BROWN ef al., GAUCHER, MICHAELIS, PUNT, and RITTERSHAUSE Haumay, L. Sobre una supuesta “‘heterocarpia” de Tragia volubilis L. Physis : 304-306. 1922. JouNston, M. C. The noseburn (Tragia, Euphorbiaceae) of western Texas. Rhodora 64: 137-142. 1962. Knott, F. Die Brennhaare der Euphorbiaceen-Gattungen Dalechampia und Tragia a ber. Math.-Nat. Akad. Wiss. Wien 114(Abt. I): 29-48. pls. 1, 190 cecidomyid larva). As far as the North Ame stat Laer are concerned, however, this explanation is highly improbable, since microtome sections of a considerable number of fruits failed to disclose any evidence of Poa ake and winged fruits regularly developed on plants of Tragia brevispica grown in the greenhouse. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 379 Lourteic, A., & C. A. O’DONELL. Tragiae Argentinae. Lilloa 6: 347-380. 1941. Miter, K. I. Preliminary studies on the genus Tragia (Euphorbiaceae). Proc. Indiana Acad. Sci. 72: 257. 1963. . A taxonomic study of the species of Tragia in the United States. 161 pp. Ph.D. diss. (unpubl.). Purdue Univ. 1964. . L. Wesster. A preliminary revision of Tragia (Euphorbiaceae) in the United States. Rhodora (in press). Pax, F., & K. HorrMann. Euphorbi Acalypheae-Pl reich IV. 147-IX(Heft 68): 1-108. 1919. [Tragia, 32-101.] Rao, J. S., & D. D. SuNpARARAJ. Stinging hairs in Tragia cannabina L. f. Jour. Indian Bot. Soc. 30: 88-91. 1951. Wiccrns, I. L. Notes on certain Euphorbiaceae in the Sonoran desert. Contr. Dudley Herb. 4: 343-353. pls. 13-17. 1955. [Comments on ¢ flowers of Tragia spp., 346, 347.] 1 Lee Pflanzen- Subtribe Ricininae Griseb., “Ricineae”’ 16. Ricinus Linnaeus, Sp. Pl. 2: 1007. 1753; Gen. PI. ed. 5. 437. 1754. Shrub or tree (annual in temperate regions); twigs smooth, usually glaucous; sap watery. Leaves alternate; stipules fused into a caducous sheath, leaving a circumaxial scar; petiole elongated, with patelliform glands at least at base and at junction with blade; lamina peltate, palm- ately 7-11-lobed, lobes serrate. Plants monoecious; inflorescences panicu- late, terminal (appearing opposite the leaves and axillary due to sympodial growth), proximal nodes with several-flowered male cymules, distal bi- sexual or female (the bisexual with female flowers central and males lateral); bracts papery, glandular at base. Flowers apetalous, disc absent. Male flower: calyx completely synsepalous in bud, valvately 3—5-partite at anthesis; anthers extremely numerous, up to 1000; filaments partially connate at base, irregularly branched; anthers introrse in the bud; pollen spheroidal, tectate (psilate), tricolporate, colpi narrow; rudimentary A genus of a single highly variable species native to Africa and perhaps to India. Mueller distinguished 15 “varieties” and a number of forms which are really cultivars; it is not clear (and probably never will be) whether the species ever showed geographic subspeciation. Because of the economic importance of the seeds, Ricinus has received considerable at- 380 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 48 tention from botanists, and it has become a classical experimental plant during recent years. The seeds are valuable for their high content (50 per cent or more) of castor oil, which is a mixture of ricinoleic acid and smaller amounts of oleic, linoleic, and other fatty acids. The unique properties of ricinoleic acid, which thus far has not been isolated from any other plant, have led to the use of castor oil, not only as the familiar purgative, but in many industrial applications (e.g., in the manufacture of paints, inks, plastics, soaps, and linoleum). Almost all of the castor oil used in the United States today is imported, although the plant was grown as a crop and pedicel length, X 1; i, carpel wall after dehiscence of schizocarp, < 1%; j, columella, X 1%; k, 1, seed before and after drying of caruncle, X 114; ™, embryo oriented as in seed, & 114. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 381 in the Midwest during the 19th century, and the industry has revived to some extent at the present time. In our area, climatic conditions permit successful cultivation of Ricinus only in Arkansas and western Tennessee, as a serious fungal disease prevents it from being cultivated on the Gulf Coastal Plain. The plant is commonly grown as an ornamental, however, and escaped individuals may be expected throughout most of our range. In the tropical area of southern Florida the plant attains its normal wild expression and becomes a sizable tree very different in appearance from the usual garden herb. It is well known that the seeds (specifically, the seed-coats) of Ricinus are extremely poisonous. This is due not to the alkaloid, ricinine, but rather to the phytotoxin, ricin, which is one of the most toxic compounds known: the minimum lethal dose by injection is only 0.001 mg. per kg. It is interesting that similar phytotoxins have been reported in the seeds of Jatropha Curcas and Aleurites Fordii, but our present inexact knowledge of the structure of these complex proteins does not permit placing much taxonomic significance on the distribution of the substances. Various aspects of the morphology of Ricinus have been investigated, especially the remarkable androecium. Delpino called attention to the explosive dehiscence of the anthers, which is similar to that in some Urti- caceae, and Steinbrinck and others have seen a resemblance between the Ricinus microsporophyll and that of pteridophytes; suggestions have even appeared that the Ricinus androecium is a mass 0 dichotomizing telomes! Such woolly speculations have been effectively discredited by Van der Pijl, but the fascicle-formation of the stamens still has not received a satisfactory explanation. si Various assertions of myrmecophily have been made for Ricinus, but the plant has not been sufficiently studied in the wild state, and the sig- nificance of the various observations is not clear. The extrafloral nectaries found on the leaves and bracts secrete noticeable quantities of nectar, and Reed has reported visits by ants. Gates has speculated that the caruncle on the seed is attractive to ants, but no observations of large tropical gr e for slow-dehiscing fruits. 1as though not to the same extent as in Mercurialis. Plants which are female in early ontogeny have been obtained, but these eventually produce male flowers, and truly dioecious strains of Ricinus are extremely rare, if they exist at all. ot ve Jakob has found evidence of secondary association between two of the ten pairs of chromosomes. This might suggest that the chromosome com- 382 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 plement was originally x = 9, but the cytological evidence is equivocal, and secondary association was not observed among the 10 chromosomes in haploid plants (Poole & Hadley). Ricinus seems to be related to the Asiatic genus Homonoia, which was associated with it in the subtribe Ricininae by Pax & Hoffmann; Z. javensis (Blume) Muell. Arg., in particular, has similar male and female flowers, but its habit is utterly different. The evidently wind-pollinated flowers and chromosome complement of « = 10 suggest a possible distant relationship to Acalypha. The lobed, glanduliferous leaves and large, foliaceous cotyledons might appear to suggest an affinity with Jatropha, but that seems negated by the differences in androecium, pollen, chromo- some number, and many other characters. REFERENCES: Under family references see ASSAILLy, aaa (1858), BLonm, Denay, DELPINO, EICHLER, GAUCHER, KINGSBURY, MANDEL, MIcHAELIs, Pax & Horr- MANN (1930), Mop1tewskt, Punt, Daten aie and SCHWEIGER. ALexanprov, W. G., & O. G. ALEXANDROvA. Uber konzentrische Gefassbiindel im Stengel von Ricinus communis. Bot. Arch. 14: 455-461. 1926. [Reports these bundles in pith of inflorescence region. BALASHEY, L. N. M. Sannrkova. Effect of root nutrition on sex charac- teristics of the castor oil plant. (In Russian.) Dokl. Akad. Nauk SSSR 60: 1061-1063. 1948. [Ammonium nitrate and potassium sulphate BEILMANN, A. P. The castor bean — an important crop for the future. Aes aue Bot. Gard. Bull. 35: 171-175. 1947. BLARINGHEM, L., L. Cuevet, & L. RoHLFs. Une nouvelle race de ricin (Ricinus communis L. var. pseudoandrogyna). Compt. Rend. Acad. Sci. Paris 233: 770-774. 1951. & L. Routrs. Sur la fixation rapide du Ricinus communis L. vat. pseudoandrogyna f. " ali Blar. et Rohlfs. Compt. Rend. Acad. Sci. Paris 237: 291-293. Biocu, R. The a a ay of the secretory cells of Ricinus and the problem of cellular differentiation. Growth 12: 271-284. 1948. BocbASHEvsKalA, O. V. Physiological conditions of ricinin biosynthesis. (In Russian.) Dokl. Akad. Nauk SSSR 99: 853, 854. 1954. [Biochemistry of castor beans. | BorATYNSKA, W. Mutagenic effects of ionizing radiations in Ricinus communis Genet. Polon. 3(2): 137-153. 1962.* BricHa, R. D., & B. R. Spears. Castorbeans in Texas. Bull. Tex. Agr. Exp. Sta. 954: 1-11. 1960. | Mostly economic. CHANDRASEKARAN, S. N., & D. D. Sunpararaj. A note on the inflorescence of Ricinus communis Lin Jour. Indian Bot. Soc. 25: 103. 1946 Cocconr, G. Anatomia dei nettarii estranuziali del Ricinus communis L. Mem. Accad. Sci. Ist. Bologna V. 5: 423-431. 1 pl. 6. Creecu, R. G., & H. H. Kramer. Gene interaction in Ricinus communis, L. Agron. Jour. 51: 642-644. 1959 DANDENO, J. B. The pomp of seed- dispersion in Ricinus communis. Bull. Torrey Bot. Club 31: 1904 Datta, R. M. On the Hingelera of “che inflorescence of Ricinus communis Linn. Sci. Cult. 10: 451, 452. 1945. 1967] WEBSTER, GENERA OF EUPHORBIACEAE 383 DeELPINno, F. Anemofilia e scatto delle antere presso il Ricinus communis. Malpighia 3: 337, 338. ABSD, Dominco, W. E. Amount of pene outcrossing in the castor oil plant. Jour. Am. Soc. Agron. 36: 360, 3 4 Dusarp, M., & P. EBERHARDT. . ricin. 2™¢ ed. 120 pp. Paris. 1917. Dumont, D. La culture du ricin aux Etats-Unis. isa 12; 273-278. 1957. Ferry, J. : The morphology and anatomy of the floral organs of Ricinus com- munis [Tourn.] L. Abstr. Doct. Diss. Ohio State Univ. 19: 41-50. 1936. [Presents an unconvincing interpretation of the ¢ and @ flowers as con- densed inflorescen Furtuata, R. Studies on the agglutinating and hemolysing factors contained in Ricinus communis. I. Studies on the agglutinating factor. Jour. Bio- chem. 38: 361-369. 1951. Gates, B. N. Carunculate seed dissemination by ants. Rhodora 45: 438-445. 1943. [Includes notes on Ricinus, 439-440. GoNncALVEs pA CunnaA, A. L’anatomie des nectaires du pétiole . la feuille de Ricinus communis. Compt. Rend. Soc. Biol. Paris 107: 90, 91. 1931. . Etudes cytophysiologiques sur les wipers a pétiole es la feuille de Ricinus communis L. Bol. Soc. Brot. IT. pls. 1, 1939. [In- cludes historical review; regards nectaries as Hiei ert Hartanp, S. C. The genetics of Ricinus communis. Bibliogr. Genet. 4: 171- Ae eration in gene frequency in Ricinus communis L. due to climatic coniione: Heredity 1: 121-125. 1947. [Based on field work in Peru. Harte, O., & I. THALER. Mikrochemische und fluoreszenzoptische untersuchun- gen an ieee von Ricinus communis L. Osterr. Bot. Zeitschr. 103: 44-52. 1956. . The castor bean plant and laboratory air. Bot. Gaz. 46: 439— 442. 1913. [Highly sensitive to ethylene. Hivpert, F, Ueber die Trennungszonen an den Bliiten und den Bau der Fruchtstiele von Ricinus communis L. Jahrb. Wiss. Bot. 88: 862-892. 1939. [Interprets the portion of the male flower pedicel above the articulation as “pericladium . ee Morphologie der Bliitenstande von Ricinus. Biol. Zentralbl. 61: 182-208. 1941. [An exhaustively thorough study of inflorescence structure and growth form. | Istam, A. S A. A. Kwan. Studies on colchicine-induced tetraploids in Ricinus communis L. Jour. Sci. Industr. Res. 3: 83-86. IwANowa-ParaorsKaja, M. Zur Anatomie des Ricinus as Textilpflanze. on Russian; German summary.) Bull. Univ. Asie Centr. 15: 97-115. pls. 4, 5 1927. [ Abstracted in Bot. Centralbl. 153: 451. 1928.] Jaxos, K. M. The pachytene chromosomes of the castor oil plant. Cytologia 21: 76-80. 1956. Secondary association in the castor oil plant. Ibid, 22: 380-392. 1957. asynaptic castor-oil plant (Ricinus communis L.). (In . Bull. Kyushu Univ. Fac. Agr. 14: 1-6. KATAYAMA, TT. Ans Japanese; English summary.) Sci 953. ape P. C. Observations of hermaphroditism in the flowers of Ricinus communis Linn. (castor plant). Sci. Cult. 26: 83. 1960. Mernpers. H. C., & M. D. Jones. Pollen shedding and dispersal in the castor plant Ricinus communis L. Agron. Jour. 42: 206-209. 1950. 384 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Nakamoto, M., & H. Yokoyama. Studies on the blooming habits of castorbean (Ricinus communis. L.) (In Japanese; English summary.) Gifu Univ. Fac. Liberal Arts Educ. Sci. Rep. Nat. Sci. 2: 175-180. Nararn, A. Mutants in castor oil plant. Sei, Cult. 16: 484, 485. 1951. . Artificial production of tetraploids in Ricinus communis Linn. Curr. Sci. Bangalore 22: 268, 269. 1953 [Use of colchicine. ] NarayAN, A. Inheritance of spininess of capsule in Ricinus communis. Indian Jour. Genet. Pl. Breed. 20(3): 223-226. 1960. Otsson-SeFFER, R. J. The castor oil soy (Ricinus communis, Linné). Am. Rev. Trop. Agr. 1: 102-1 1910 ParKEY, W. Cytoplasmic inhgence in ‘the Spree a = pistillate sex ex- pression in castorbeans, Agron. Jour. 49: 427, . G. SCHOENLEBER. A new device achat nomograph, for measuring the sex expression in castor beans, Ricinus communis. Agron. Jour. 46: 288, sie 1954. AB, ey ees FMANN. Euphorbiaceae—Acalypheae-Ricininae. Pflanzen- reich IV. i47-XI(Heft 68): 112-127. 1919. [Ricinus, 119-127. Penrounp, W. T. The anatomy of the castor bean as conditioned by light in- tensity and soil moisture. Am. Jour. Bot. 19: 538-546. Z, Pryt, L. vAN DER. The stamens of Ricinus. Phytomorphology 2: 130-132. 1952. Poot, D. D., & H. H. Haptey. Haploidy in castorbeans. Jour. Hered. 45: 285- 288. 1954. Popowa, G. M. The castor bean in central Asia. Bull. Appl. Bot. Pl. Breed. 16(4): 145-240. 1926. [Includes extensive morphological observations; in Russian, but with detailed English summary, 227-240. Rao, M. B. S., & K. THANDAVARAYAN. Occurrence of multi-locular fruit in Ricieas communis, L. (Castor). Madras Agr. Jour. 41: 329, 330. Reep, E. L. Extra- floral nectar glands of Ricinus communis. Bot. Gaz. 76: 102-106. 1923. ReynoLps, M. E. Development of the vascular network in the node of Ricinus communis. Am. Jour. Bot. 28: 728. 1941 . Development of the node in Riciaus communis. Bot. Gaz. 104: 167- 170. 1942, SArKAny, S. Uber Entwicklung und Funktion des interfaszikularen Kambiums bei Ricinus communis. Jahrb. Wiss. spe 82: 625-656. Scott, F. M., & H. M. SHarsmitH. The transition region in the beste of Ricinus communis: a physiological feast on. Am. Jour. Bot. 20: 176-187. 1933. [Not very illuminatin , V. SJanoim, & E. Bower. Light ‘md electron microscope studies of the primary xylem of Ricinus communis. Am. Jour. Bot. 47: 162-173. 960. Ss N. J., & W. E. Domino. Effect of duration and intensity of light upon flowering in several varieties and hybrids of castor bean. Bot. Gaz. 108: 556-570, 1947. SHirFRiss, O. Sex instability in Ricinus. Genetics 41: 265-280. 1956 HRADER, J. H. The castor-oil rae ene U.S. Dep. Agr. Bull. 867: ‘1-40. 1920. Sunpact, S. S. Castor hybrids. I. Manifestation ie hybrid vigour in F,s during early seedling stages. Indian. Oilseeds Jour. 8(2): 159-161. 1964.* StncH, P. Pharmacognostic study of the root of Ricinus communis Linn. Jour. Sci. Industr. Res. 15C(12): 259-262. 1956. [Includes morphology. | 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 385 Sincu. R. P. Structure and development of seeds in Euphorbiaceae: Ricinus communis L. Phytomorphology 4: 118-123. 1954. STEINBRINCK, I. C. Ueber die physikalische Verwandtschaft der pollenschleu- dernden Ricinus-Anthere mit den sporen-schleudernden Farn- und Selaginel- la-Kapseln. Ber. Deutsch. Bot. Ges. 28: 2-7. 1910. TAKAHASHI, T., G. Funatsu, & M. Funatsu. Biochemical studies on castor bean hemagglutinin. Jour. Biochem. 52(1): 50-53. 1962. Tuomas, C. A., & R. G. OreELLANA. Biochemical tests indicative of reaction of castor bean to Botrytis. Science 139: 334, 335. 1963 Tremazr. S. A.. & N. Munammap. Formation of oil in castor (Ricinus com- munis). Pakistan Jour. Sci. Res. 13(3): 137, 138. 1961.* Water, G. R., & L. M. Henperson. Biosynthesis of the pyridine ring of ricinine. Jour. Biol. Chem. 236(4): 1186-1191. 1961. WEIBEL, R. O. The castor-oil plant in the United States. Econ. Bot. 2: 273-283. 1948 & W. L. Burtson. A review of the experimental work on castor beans in the United States. Chemurgic Dig. 5: 167-172. 1946.* Wuire, O. E. Breeding new castor beans. Jour. Hered. 9: 195-200. frontisp. 1918 - Inheritance studies on castor beans. Brooklyn Bot. Gard. Mem. 1: 513-521. pls. 23-28. ZIMMERMAN, L. H. Castorbeans: a new oil crop for mechanized production. Advances Agron. 10: 257-288. 1958. Tribe HippoMANEAE Bartl. ex Endl. Subtribe Hippomaninae Muell. Arg. 17. Sebastiania Sprengel, Neue Entdeck. Pflanzenk. 2: 118. pl. 3. 1821. Shrubs [or trees, rarely herbs] ; twigs glabrous or with simple hairs [spinescent in some species]. Leaves alternate [very rarely opposite], stipulate; petiole short; blade undivided. Plants monoecious [rarely dioecious]; flowers in usually bisexual spikes which are terminal or (due to sympodial growth) opposite the leaves [rarely axillary]; female flowers solitary in axils of lower bracts, male flowers usually in clusters of 2 or 3 subtended by upper bracts; bracts biglandular at base. Flowers apetalous, disc absent. Male flower: calyx synsepalous, lobes usually 3 (2), subequal [sometimes distinctly unequal |, imbricate; stamens [2 or] 3; filaments free or nearly so, anthers extrorse; pollen subglobose, tectate (psilate), tri- colporate, 3-lobed in polar view; rudimentary gynoecium absent. Female flower subsessile; calyx 3-lobed, lobes imbricate, with minute adaxial processes at base; carpels 3 [rarely 2]; styles free or connate at base, en- tire; ovary smooth [sometimes muricate]; ovules 1 in each locule, ana- tropous. Fruit capsular; columella dilated above, usually persistent. Seeds subglobose [cylindric in some species], smooth [or roughened], caruncu- late; endosperm copious; embryo straight, cotyledons broader and longer than the radicle. Type species: S. brasiliensis Spreng. (Named in honor of Antonio Sebastiani, 1792-1831, Italian professor of botany.) As circumscribed by Pax, a diversified genus of some 75 species, at least 386 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 three-quarters of which are native to southern Brazil. Although Pax enumerated three species from the Old World, one of these (S. inopinata, from Africa) has been made the type of a segregate genus Duvigneaudia by Léonard, primarily on the basis of its drupaceous fruit and ecarunculate seeds. Three species from Indonesia (two described by Van Steenis) are of dubious affinity, and their possible relationship to Duvigneaudia needs to be examined. Sebastiania Chamaelea (L.) Muell. Arg., an herb found across the tropical belt from Africa to northern Australia, appears to be fairly closely related to the weedy American herb, S. corniculata (Vahl) Muell. Arg., and there is no reason to question its position in the genus. Sebastiania consequently has at least one gerontogean representative, but its remarkable paucity of development there remains unexplained. Our only species, Sebastiania fruticosa (Bartr.) Fern. (S. ligustrina (Michx.) Muell. Arg.), is a shrub 3-4 m. high which is found in hard- wood forests (often along streams) from North Carolina to Florida, west to eastern Texas. The species is the only representative of sect. STILLING- 1opsis (Muell. Arg.) Webster,?? which differs from sect. CNEMIDOSTACHYS (Mart.) Webster 3 in its ovoid seeds, smooth ovary, connate male sepals, and woodier habit. It is much closer to sect. SEBASTIANIA, which differs only in its reduced male calyx lobes and larger bracteal glands; the type species, S. brasiliensis, strongly resembles S. fruticosa in many respects. A peculiarity which is shared by some other Hippomaneae is that Sebastiania fruticosa does not produce noticeably milky juice, even though laticifers are present (according to Herbert). The single chromosome count, 2n = 56 (Perry), suggests that the species may be an octoploid Pringlei Wats. and S. Palmeri Rose; Pax (under the latter species) has reviewed the literature, which is more voluminous than the apparent triviality of the phenomenon would seem to warrant. Within a few years after Hooker provided the first botanical description in 1854, the cause of the erratic behavior of Sebastiania seeds was shown to be due to the activity of the larva of a microlepidopteran moth (Laspeyresia saltitans, Tortricidae). Similar “springenden Bohnen” have also been reported from other taxa of Hippomaneae (e.g., Colliguaya odorifera Mol., Sapium biloculare (Wats.) Pax), and in all instances (where known) the causa- tive agent has proved to be a tortricid larva. Although little has appeared in the literature since the plethora of papers between 1890 and 1926, a number of questions remain to be resolved. The biological relationship ™ Sebastiania sect. Stillingiopsis (Muell. Arg.) Webster, comb. nov. Gymnanthes sect. Stillingiopsis Muell. Arg. Linnaea 32: 96. : tiania sect. Cnemidostachys (Mart.) Webster, comb. nov. Stillingia sect. Cnemidostachys (Mart.) Baill. Etud. Gén. Euphorb. 515. 1858, Cnemidostachys art. Nov. Gen. Sp. 1: 66. 1824, 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 387 between lepidopteran and plant has not been studied in sufficient detail (i.e., through entire life-cycles) and the classification of the taxa of both host. and parasite is still imperfect. Mueller originally determined the infected Mexican plant as Sebastiania Pavoniana (Muell. Arg.) Muell. Arg., and it is possible that most of the various Mexican host plants may prove to be varieties of that species. Generic limits in the tribe Hippomaneae have long been subject to controversy, and it is not certain that taxa such as Sebastiania can be maintained in their present circumscription when more data become avail- able. Baillon, taking a broad view of generic limits, included Sebastiania, Stillingia, Maprounea, and Sapium within a very inclusive genus Excoe- caria, Mueller separated Excoecaria, Maprounea, and Sapium, but kept Sebastiania united with Gymnanthes. Bentham, followed by Pax, sep- arated Gymnanthes on the basis of its reduced perianth. There is little doubt that Gymnanthes and Sebastiania are closely related, but they can be fairly readily distinguished, so there seems to be no urgent necessity to combine them. The Old World genus Excoecaria is also closely related to Sebastiania but is sufficiently distinguishable by virtue of its axillary dioecious inflorescences and ecarunculate seeds. REFERENCES: Under family references see BAILLON (1873, pp. 120-123), HERBERT, NozeEr- AN, Pax & HorrMANN (1912, V: 88-153), PERRY, and THATACHAR. BucHENav, F. Die springenden Samen aus Mexiko. Abh. Naturw. Ver. Bremen 3: 373-377. 1873. Die “springenden Bohnen” aus Mexiko. /bid. 12: 47-52, 277-290. 1891. [Probably the most scholarly review of the jumping-bean problem. | Dampr, A. Las semillas brincadoras de México. Mex. Secretaria Agr. Bol. Mens. Ofic. Def. Agr. 2: 440-451. 1 fold. pl. 1928. [The most recent treatment ; includes most earlier references. | Léonarp, J. Notulae systematicae XXV. Duvigneaudia J. Léonard: genre nouveau guinéo-congolais. Bull. Jard. Bot. Bruxelles 29: 15-21. 1959. [Discusses relationships of this segregate genus with Sebastiania. | Narr, N. C., & M. Marrreyr. Morphology and embryology of Sebastiania Chamaelea. Bot. Gaz. 124: 58-68. 1962. Ritey, C. V. Mexican jumping beans and the plant upon which they are pro- duced. Am. Garden 12: 552-554. 1891. [Illustrated; the best available general account in English. ] STEENIS, C. G. G. J. vAN. Provisional note on the genus Sebastiania in Malaysia. Bull. Bot. Gard. Buitenzorg III. 17: 409, 410. 1948. 18. Gymnanthes Swartz, Prodr. Veg. Ind. Occ. 95. 1788. Glabrous trees or shrubs; sap not appreciably milky. Leaves alternate, stipulate, unlobed, entire or dentate; petiole short, not glandular. Plants monoecious [rarely dioecious]; inflorescences spiciform, axillary, pro- tected by a conspicuous bud, bisexual; female flower usually only 1 per inflorescence; long-pedicellate from the lowermost node; male flowers 388 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 [1-]3 per cymule, occupying remaining nodes; bracts more or less glandular at base. Flowers apetalous; calyx small to absent; disc absent. Male flower: calyx rudimentary, of 1 or 2 small sepals often confounded with bractlets, or absent; stamens 2 or 3 in lateral flowers of the cymule, 3-5 in the central flower; filaments free or basally connate; pollen sub- globose, tectate, tricolporate, circular or 3-lobed in polar view; rudimen- tary gynoecium absent. Female flower: calyx minute, of 2 or 3 reduced sepals [or absent]; carpels 3; styles free or connate at base, unlobed, slender and recurved; ovary smooth, stipitate [or sessile]; ovules 1 in each locule, anatropous. Fruit capsular; columella 3-winged, persistent. Seeds subglobose, smooth, dry, carunculate; cotyledons much broader than radicle. Lectotype species: G. lucida Sw.; typification effected by Grisebach, Fl. Brit. W. Indian Is. 50. 1859. (Name from Greek, gymnos, naked, and anthos, referring to the highly reduced perianth of the flowers.) — CRAB-woop. As circumscribed by Pax, a small genus of 12 species confined to the Caribbean region, mostly in the West Indies. Gymnanthes is extremely difficult to separate from Actinostemon, on the one hand, and from Sebastiania, on the other, and Baillon’s suggestion to merge these and some other genera under Excoecaria is not entirely without merit. How- ever, until additional data, such as chromosome counts (unrecorded for any species of Gymnanthes), are forthcoming, the traditional circumscrip- tion is diffidently maintained here Our only representative, Gymnanthes lucida Sw., the most widespread species in the genus, has been recorded from southern Florida throughout the Bahamas and Greater Antilles to Guadeloupe, in the Lesser Antilles. In Florida it is mainly confined to hammock vegetation on limestone in the Miami area and Florida Keys, extending to Key West; the northern- most collection known is from Palm Beach County. The species is easily distinguished from all its congeners by the remarkable stipitate ovary, the gynophore attaining a centimeter in length and appearing as a pro- longation of the peduncle. The solitary, long-pedicellate female flower and stiff, crenate leaves also make it readily separable from Sebastiania fruticosa, even though the genera on the whole may be difficult to dis- tinguish., REFERENCES: Under pd bese see HERBERT, NozerAN, Pax & HorrMANnN (1912, V: 81-88), and P SARGENT, C. S. oe Silva N. Am. 7: 29-32. pl. 309. 1895. 19. Stillingia Garden in Linnaeus, Mant. Pl. 19. 1767. Glabrous herbs, shrubs, or small trees; sap not evidently milky. Leaves alternate [opposite, or verticillate]; stipules small and often glandular; blade unlobed, pinnately veined, [with or] without basal glands, serrate. Plants monoecious; inflorescences terminal, spiciform; female flower 1 per 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 389 bract, proximal; male flowers [1—]3 or more per bract, distal; bracts biglandular. Flowers apetalous, disc absent. Male flower: calyx 2-lobed, lobes imbricate; stamens 2, filaments connate below, anthers extrorse; pollen ellipsoid to spheroid, with [1] 3 colpi, tectate but sometimes pseudoreticulate, slightly to distinctly 3-lobed in polar view; rudimentary gynoecium absent. Female flower: sepals 3, imbricate [sometimes absent | ; carpels 3 [2]; styles more or less connate, unlobed, slender; ovary smooth (not muricate); ovules 1 in each locule, anatropous. Fruit capsular; columella dilated distally, often deciduous, surrounded by 3-horned struc- ture (“gynobase”) left after dehiscence of the carpels. Seeds not or scarcely compressed, [smooth or] rugulose, the caruncle sunk in a ventral depression [absent in some taxa]; endosperm present; embryo straight, the cotyledons broader than the radicle, cordate at base. TYPE SPECIES: S. sylvatica Garden ex L. (Named in honor of Benjamin Stillingfleet, 1702-1771, English botanist and disciple of Linnaeus). A well-defined genus of 26 or 27 known species with a strange distribu- tion somewhat paralleling that of Sebastiania. The vast majority (23) of the species are American, but two or three occur in the Mascarene Islands and one is known only from Fiji. There seems to be no reason to question the assignment of these Old World species, which belong to the single sect. PACHYCLADAE Pax, to Stillingia. The American species have been revised by Rogers, who has reaffirmed the distinctness of the genus. Rogers regards Sapium as the genus most closely related but differing in its united sepals, fleshy seed coat, usually elongated foliar glands, and especially in its lack of the three-horned “eynobase” which characterizes the dehiscent fruit of Stillingia. The infrageneric taxa of Stillingia have been differently treated by the last two monographers, Pax dividing the genus into six sections, Rogers recognizing two subgenera with a total of five series. As construed here, Stillingia comprises four sections, two of which are represented in the United States. However, none of the southwestern representatives of sect. LEPTOSTACHYAE Pax (ser. Treculianae Rogers), which are characterized by reduced female calyx and ecarunculate seeds, enter our area. Baldwin County, Alabama. Closely related, although placed in a different series, SYLVATICAE Rogers, on account of its herbaceous nonlenticellate stems from an enlarged rootstock, is S. sylvatica, queen’s delight. This species, which is found from southeastern Virginia to Florida, westward into Kansas and New Mexico, occurs in our area mainly in dry habitats (e.g., sandy flats or pine woods) on the Coastal Plain. It is thus ecologi- cally separated from S. aquatica, but Rogers has suggested that it hybridizes both with that species and, to the west, with S. texana I. M. Johnston. } 390 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 J Fic. 5. Stillingia. a-i, S. aquatica: a, branchlet with male flowers and nearly mature fruit, X 14; b, male cymule, showing bract, two cup-shaped glands, pedi- cel of central male flower, two male flowers at anthesis, two buds, X 8; c, male two e : 2 lower part of brittle columella, x 4; h, segment of schizocarp wall, x 4; i, seed, X 4. j, k, S. sylvatica: j, segment of schizocarp with seed, X 2; k, em- bryo, X 2. 3 that S. aquatica and S. sylvatica should probably be placed in the same series, As in the other Hippomaneae, almost nothing is known of the reproduc- tive biology of the species of Stillingia, although the pollen structure and 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 391 conspicuous glandular bracts suggest that they are at least partly ento- mophilous. The chromosome number of S. sylvatica was reported to be 2n = 36 by Perry, but Rogers was unable to confirm this and thought that Perry’s count was probably too low. The only firm count for any species of Stillingia appears to be 2m = 60 in S. sanguinolenta Muell. Arg., of ser. OpposiTIFOLIAE. Until more taxa in this and related genera are studied, it will not be clear whether the basic number in that species is 10 or 15. REFERENCES: Under family references see BAILLON (1873, pp. 120-123), HERBERT, MILLER & WepsTER, O’DonELL & Lourteic, Pax & HorrMann (1912, V: 180-199), Perry, and Punt. Jounston, M. C., & B. H. Warnock. The three species of Stillingia (Euphor- biaceae) in far western Texas. Southwest. Nat. 8: 100-106. 1963. | Includes description of S. sylvatica. | Rocers, D. J. A revision of Stillingia in the New World. Ann. Missouri Bot. Gard. 38: 207-259. 1951. 20. Sapium P. Browne, Civ. Nat. Hist. Jamaica 338. 1756. Glabrous trees or shrubs; stems usually with copious milky latex. Leaves alternate, stipulate; petiole usually biglandular at apex; blade denticulate, pinnately veined. Plants monoecious [rarely dioecious | ; flowers in terminal bisexual spikes; female flowers several per spike, soli- tary in the axil of each proximal bract; male flowers in several-flowered clusters in the axils of distal bracts; bracts conspicuously biglandular at flower: calyx lobes [2] 3-5, small; carpels [2] 3; styles [free or] con- nate, unlobed, slender; ovary smooth; ovules 1 in each locule, anatropous. Fruit capsular [sometimes subdrupaceous] ; valves often somewhat woody; columella slender and usually persistent. Seeds rounded (not com- pressed), outer seed coat fleshy and pseudoarillate, ecarunculate; cotyle- dons broader than radicle. (Including Triadica Lour.) TYPE SPECIES: S. aucuparium Jacq.24 (Classical name used by Pliny for a resin-exuding species of Pinus; ultimately derived from Celtic, sap, fat, perhaps alluding to the sticky latex.) — MILK-TREE. A pantropical genus of over 100 species, the majority American, only one native to the United States. In terms of discrimination of species, lant on which Browne ba ium was later designated as S. auc u Jacquin. Although Jacquin subsequently applied the name to a South American pecies, his original disposition must b we Consequently, S. aucuparium Jacq. is the correct name the species treated in most floristic manuals as S. jamaicense Sw.; the plant generally called S. aucuparium probably should take the name S. salicifolium HBK. 392 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Sapium is one of the most difficult genera in the Euphorbiaceae. Our comprehensive ignorance of the reproductive biology and ecology of the species makes it impossible at present to pinpoint where the difficulty lies. Hemsley’s work, as far as it went, appears to be sound, and his illustra- tions at least provide a point of departure for further studies; but Pax’s monograph is extremely difficult to use, and his supraspecific taxa are probably highly unnatural. In the southeastern United States, Sapium is represented only by the naturalized S. sebiferum (L.) Roxb. (Triadica sebifera (L.) Small), Chi- nese tallow tree, which is commonly cultivated as an ornamental. Wild trees have been found in low, often swampy, areas in South Carolina, Georgia, and Louisiana. The species belongs to sect. Tr1apica (Lour.) Muell. Arg., a group of six Asiatic species characterized by having a some- what fleshy fruit and by the seed coat being white and adherent to the columella, rather than reddish and free from the columella as in most tropical American species. In the United States, Sapium sebiferum is grown mainly as an orna- mental, especially for the foliage, which turns bright red in autumn. In central and southern China, Japan, and northern India, however, the tree is cultivated for the seeds. Both waxy seed coat and endosperm are used, the former providing the “tallow” used for candle-making in China, and the latter yielding an oil with some of the properties of tung oil. Unfor- tunately the labor costs involved in harvesting the fruits preclude com- mercial exploitation of the tree in countries such as the United States. Small recorded a second species, Sapium glandulosum (L.) Morong, evidently on the basis of a Curtiss collection from near Pensacola, but the plant has not been re-collected and cannot be regarded as naturalized. The only other species of Sapium in the United States is S. biloculare (Wats.) Pax, one of the Mexican jumping-bean plants, which enters a Perry has reported 2m = 36 in Sapium sebiferum, which is the only published chromosome number for the species. Ventura found that em- bryo-sac development was of the normal type in trees studied in Italy, but reported that only one seed developed per capsule; this is quite at Morro with our plants, in which three seeds regularly develop in each ruit. Although merged by Mueller with Excoecaria, Sapium has been recog- nized by most botanists on the basis of its monoecious inflorescence pro- duction, fleshy-coated seeds, and biglandular petioles. As pointed out by Pax, the relationship to Stillingia is probably closer, but Rogers has shown that Sapium can usually be distinguished from that genus by its fleshy seeds, elongated foliar glands, and lack of a persistent gynobase upon dehiscence of the fruit. According to Horn and Polhamus, the rubber contained in latex of Sapium species is comparable in quality to the best Hevea rubber and, in fact, surpasses the latter in strength and elasticity. However, the trees are difficult to tap, and significant commercial production has never been 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 393 established. Because of the abundance of the plants in Latin America, it would appear that further investigation of the potentialities of Sapium latex might be warranted. REFERENCES: Under family references see BAILLON (1873, pp. 120-123), GAUCHER, HER- BERT, HURUSAWA, JABLONSKY, LANJouw, LEonarD, Pax & HorrMann (1912, V: 199-245), PoLHAmMus, REcorpD, and WILLIAMS. BrapLey, C. E. Yerba de la fleche — arrow and fish poison of the American Southwest. Econ. Bot. 10: 362-366. 1956. [Latex of S. biloculare. | BRITTEN, f be in the collection of Ruiz and Pavon. Jour. Bot. 47: 422- 424. HEMSLEY, sg 'B. Sapium. Hooker’s Ic. Pl. 27: pls. 2647-2650, 2677-2684. 1901-1902; ibid. 29: pls. 2878-2900. 1909. [Discussions of several spp. ] Horn, E. F. Another rubber-producing Euphorbiaceae (Sapium). Trop. Woods 86: 13, 14. 1946, Howes, F. N. The Chinese tallow tree (Sapium sebiferum Roxb.), a source of drying oil. Kew Bull. [4]: 573-580. 1949. Huser, J. Revue critique des espéces du genre Sapium Jacq. Bull. Herb. Boiss. II. 6: 345-364, 433-452. 1906. [Illustrated; with keys. ] LEONARD, J. Notes sur les espéces africaines continentales des genres Sapium P. Br. et Excoecaria L. (Euphorbiacées). Bull. Jard. Bot. Bruxelles 29: 133-146. 1959. [Includes a review of problems in delimiting Sapium from allied genera. | Lin, W. C., e¢ al. An investigation and study of Chinese tallow tree in Taiwan (Sapium sebiferum Roxb.). (In Chinese; English summary.) Bull. Taiwan Forestry Res. Inst. 57: 1-37. 1958.* Pax, F. Die Verbreitung der Gattung Sapium. Pflanzenareale 1: 21. map 13. 1 PITTIER, H. The Mexican and Central American species of Sapium. Contr. U. . Natl. Herb. 12: 159-169. pls. 10-17. 1908. Ue, E. Die Kautschukpflanzen der Amazonas-Expedition und ihre Bedeutung fiir die Pflanzengeographie. Bot. Jahrb. 35: 663-678. 1905. [Includes notes on Hevea and Sapium VenturA, M. Sopra alcuni dati embriologici di Sapium sib hat ee Roxb. Atti ‘Accad. end Rend. Sci. Fis. Mat. Nat. VI. 23: 91, 193 21. Hippomane Linnaeus, Sp. Pl. 2: 1191. 1753; Gen. Pl. ed. 5. 499. 1754. Glabrous trees or shrubs; milky latex copious, poisonous. Leaves alter- nate, stipulate (stipules caducous); petioles elongated, with a single gland at the apex; blade pinnately veined, more or less cordate. Plants monoe- cious [dioecious]; inflorescence spiciform, terminal, bisexual, the rachis more or less thickened; female flowers few, solitary at proximal nodes; male flowers several to numerous in dense glomerules at each of the distal nodes; bracts conspicuously biglandular at base. Flowers apetalous, disc absent. Male flower: calyx 2- or 3-lobed, lobes imbricate; stamens 2, fila- ments connate; anthers extrorse; pollen subglobose, tectate, 3-colporate, 3-angled in polar view; rudimentary gynoecium absent. Female flower 394 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 sessile: calyx 3-parted; carpels (5) 6-9; styles connate at base, unlobed, spreading and recurved, densely papdlate: ovary smooth; ovules 1 in each locule, anatropous. Fruit drupaceous, mimicking an apple, with yellowish or reddish fleshy exocarp, bony endocarp. Seeds not compressed, smooth, ecarunculate; endosperm present; cotyledons flat, much longer and broader than the radicle. Lectotype species: H. Mancinella L.; see Baillon, Etude Gén. Euphorb. 540. 1858. (Name from Greek, Aippos, horse, and mania, fury, referring to the effect of the poisonous latex.) — Mancut INEEL. A small tropical American genus of two or three species, well character- ized by the multilocular drupaceous fruit which is so apple-like in appear- ance that it has been the cause of serious poisonings. Two very closely related species with smaller drupes and spinose leaves, H. spinosa L. and H. horrida Urb. & Ekm., are endemic to dry scrub on limestone in south- western Hispaniola. Our representative is the well-known manchineel tree, Hippomane Mancinella L., n = 11,2° which is widespread along seashores in the Caribbean area from Florida to Mexico, Colombia, and Venezuela, with outlying populations in the Galapagos and Revillagigedo Islands. In Florida, the manchineel is restricted to the Keys and the immediately adjacent mainland, where it is found in hammock vegetation behind the mangrove zone. From the earliest days of European colonization the manchineel has had a sinister reputation, and its toxicity has been exaggerated by some writers; still, there is no doubt that the latex is extremely dangerous if taken in- ternally, and it can produce severe irritation of the skin. The wood has, in the past, been used here and there in the West Indies for making furniture, but this has now almost ceased, as the trees have been ruth- lessly extirpated from many parts of their previous range Despite its widespread occurrence and familiarity, the reproductive biology of the manchineel remains almost completely unknown; but in our area it flowers in April. The pollination process and embryological development have not been followed, nor has germination and seedling establishment been carefully studied. The affinities of Hippomane have seldom been discussed, but there seems to be no reason to dissent from Baillon’s opinion that it is very close to Sapium. The overall habit of the manchineel certainly recalls species of that genus, and it can scarcely be pa gest generically except on the basis of its indehiscent multilocular ruit. REFERENCES: Under family references see BAILLON esi oe Krincspury, Pax & HoFFMANN (1912, V: 261-263), Punt, and Rec LauTER, W. M., L. E. Fox, & W. T. ArratL. rican of the Ac principles of Hippomane Mancinella L. 1. Historical review. Jour. Am. Pharm. ~ Previously unpublished chromosome count by D ack El ased on examina- lis, b tion of meiosis in male flower buds of shrubs growing a Auchenskeoch Beach, Tobago (Webster et al. 9825) 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 395 Assoc. Sci. Ed. 41: 199-201. 1952. [Mostly anecdotal; for parts II and III, with various authors, see ibid. 44: 361-363. 1955; 46: 93-97. 1957.] Oakes, A. J., & J. O. ButcHeR. Poisonous and injurious plants of the U.S. Virgin Islands. U. S. Dep. Agr. Misc. Publ. Agr. Res. Serv. 882: 1-97. 1962. [Hippomane, 50, 51; also includes references to spp. of Croton, Jatropha, Euphorbia, and Pedilanthus. | SARGENT, C. S. Hippomane. Silva N. Am. 7: 33-38. pl. 310. 1895. Urpan. I. Plantae Haitienses novae vel rariores III. Ark. Bot. 20A(15): 1-94. pl. 1. 1926. [Discussion of H. spinosa, 66, 67.| _ Plantae Haitienses et Domingenses novae vel rariores V. a cl. E. L. Ekman 1924-27 lectae. Ibid. 22A(8): 1-98 pls. 1-4. 1929. [Description of H. horrida, 64.| Tribe EupHorsiEAE [Dumort. | Subtribe Euphorbiinae Hurusawa 22. Euphorbia Linnaeus, Sp. Pl. 1: 450. 1753; Gen. Pl. ed. 5. 208. Herbs, shrubs, or trees of diverse habit, aérial parts sometimes succulent, latex whitish, Leaves alternate, opposite, or whorled, persistent or decid- uous (or sometimes caducous and then plants appearing leafless); stipules present or absent, sometimes represented by glands; blade undivided (sometimes lobed), usually pinnately veined, eglandular. Plants monoe- cious or rarely dioecious; inflorescence a bisexual (rarely unisexual) pseudanthium (cyathium); cyathia borne in terminal or axillary dichasia or pleiochasia (or, by reduction, apparently solitary). Cyathium with a solitary terminal female flower subtended by (4) 5 male cymes, the bracts opposite the male cymes fused into a radially or isobilaterally symmetrical, campanulate to hemispheric involucre; tips of bracts alternating with 4 or 5 (rarely fewer) glands (nectaries); glands entire, 2-lipped, or variously lobed, sometimes with petaloid appendages; male cymes monochasial or possibly dichasial at base, each developing 1-10 or more flowers; brac- teoles of male monochasia conspicuous to reduced or absent, often fringed or lacerate, sometimes partly fused into the involucral partitions. Male flowers monandrous, simulating a single stamen; perianth usually absent, very rarely developed as a small cup (more commonly represented by an articulation): anthers dehiscing longitudinally and “introrsely” (i.e., centripetally): pollen subglobose, tricolporate, reticulate to tectate (psil- ate), sometimes operculate, with conspicuous intine thickenings parallel to colpi. Female flower pedicellate, pedicel often reflexed; perianth of 3-6 more or less united scalelike sepals or more commonly rudimentary or absent; carpels 3 (very rarely 2 or 4); styles free or basally connate, bifid or rarely entire; ovary glabrous or pubescent, often angled or carinate; ovules 1 in each locule, anatropous, with a more or less evident nucellar beak; embryo sac normal (Polygonum type) or rarely of various aberrant tetrasporic types. Fruit capsular [rarely drupaceous]; columella more or less persistent. Seeds smooth, ridged, or variously pitted or tuberculate, with or without a caruncle; endosperm copious; cotyledons fleshy, broader 396 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Ns ae va and Chamaesyce. a-i, Euphorbia corollata: a, upper part of oe plant, x 4; b, cyathium, 8; c, cyathium with walls removed to show cymules of male pare and hairy siatteolak < 8; d, male flower with bracteole —note articulation with pedi x 82:¢, nearly mature schizoc ocarp, G4, segment of Apa abit after ehiossnee e, X 6; g. columella of schizocarp, X 6; h, seed, X 6; i, embryo oriented as in seed, < 6. j, E. inundata: cyathium and female Pes with three-lobed perianth, x 4. . commutata: athium from above, < 8; 1, embryo, oriented as in seed (compare vie ij ”, 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 397 than the radicle. (Including Agaloma Raf., Dichrophyllum Kl\. & Garcke, Elaeophorbia Stapf, Galarhoeus Haw., Poinsettia Graham, Tithymalopsis Kl. & Garcke, Tithymalus J. Gaertn. [nom. cons.], and Zygophyllidium (Boiss.) Small.) Lectotype spectes: E. antiquorum L.; see Millspaugh, Field Mus. Publ. Bot. 2: 306. 1909. (Named for Euphorbus, physician to King Juba of Numidia, a contemporary of Pliny, Euphorbus having drawn attention to the properties of the succulent E. officinarum vel aff.) — SPURGE. In the broad circumscription of Pax & Hoffmann, Euphorbia, with over 1500 species, is the largest genus in the family and one of the largest among the angiosperms. This inclusive genus is characterized by its more or less regular bisexual pseudanthium, to which the term ‘“‘cyathium” has been applied by most workers. Boissier, in the first general monograph, placed the cyathium-bearing Euphorbiaceae into only four genera: An- thostema, Euphorbia, Synadenium, and Pedilanthus. Among these, Eu- phorbia included all the species in which the male flower lacks a perianth and the cyathial involucre bears four or five glands. Pax & Hoffmann admitted a number of additional genera, such as Stenadenium and Monadenium, which differ from Euphorbia only in relatively slight modi- fications of the cyathium. The least justifiable segregation would appear to be the recognition of the genus Elaeophorbia Stapf for certain African species with drupaceous fruits; these species should be returned to Euphorbia subg. EupHorstA, where they find a place adjacent to sect. TEKEANAE Croizat.*6 Croizat and Dressler, in a number of articles, have severely criticized the traditional circumscription of Euphorbia (without, however, using most of the segregate taxa in their own publications). Some of these criticisms have considerable merit, as the current concepts of generic limits in the Euphorbieae are indeed illogical and unnatural. The African genera Monadenium and Synadenium, although generally accepted by most 20th century workers, are evidently allied to Euphorbia sect. Trrv- CALLI and are scarcely more distinctive overall than are many of Boissier’s sections which are usually retained within Euphorbia. If one accepts the position that the cyathium is a tribal character of the Euphorbieae rather than diagnostic for the genus Euphorbia, other criteria must be sought for generic delimitation in the tribe. The tre- mendous diversity in habit, leaf morphology, chromosome number, and pollen configuration might at first appear to favor the dissolution of oe into a number of smaller genera. However, a review of the Euphorbia sect. ee baron Webstir. stat. nov. Elaeophorbia Stapf, Bi i Ic. Pl, IX. 4: pl. 2823. 190 x 6. . dentata: cyathium, X 8. n-q, Chamaesyce maculata (C. supina): n, cau of rostrate plant to show hari th habit, arrangement of leaves, and interpetiolar stipules, X 14; 0, cyathium, X 16; p, q, seed, adaxial surface and from above, < 16. 398 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 taxonomic history of the Euphorbieae is sufficient to indicate that the problem does not have an obvious solution. Variations in such features as the cyathial glands (appendaged or not), seeds (carunculate or not), and female flowers (naked or “calyculate”’) do not seem to show strong correlation with major phyletic groups. There is the additional practical difficulty that the numerous exceptions to such characteristics (e.g., carunculate seeds) make it extremely difficult to find good “key” charac- ters to diagnose segregate taxa. If the various microgenera of Euphorbieae cannot be easily distinguished, there seems little reason to adopt them simply because they represent evolutionary units; it is quite as easy to discuss the evolution of these taxa if they are referred to as infrageneric components of Euphorbia. The present author finds himself no more able to resolve this taxonomic impasse than have the many others who studied the situation earlier. A major difficulty appears to reside in the lack of basic morphological and cytological data. Although many papers have been published on Euphor- bia, most have dealt either with species in subg. Esuta or with the African succulents. Chromosome numbers are available for over 80 spe- cies, but many of these require confirmation, and no counts have been reported for any species in 18 of the 27 sections recognized by Boissier. Although much (perhaps too much!) has been published on the morph- ology of the cyathium, anatomical data — even for the cyathia — are in- adequate for comparisons between sections, and the nature of the spines in the succulent taxa is still controversial. In view of this present scarcity of systematically meaningful facts, and in the face of the highly contra- dictory opinions of experts living and dead, a compromise view on generic limits is adopted in the present treatment: Chamaesyce is accepted dif- fidently as a distinct genus, whereas all other segregate taxa, including Poinsettia, are relegated to subgeneric status within Euphorbia. This disposition is frankly one of expediency and does not claim either logical or phylogenetic justification. Chamaesyce happens to be a large and doubtless natural group of several hundred species which is rather easily definable. It seems more convenient, in practice, to recognize Chamaesyce as an independent genus than to retain it within Euphorbia; whereas, in my opinion, this is not true, in our present state of knowledge, for the other taxa of Euphorbieae in question. As thus delimited, Euphorbia, sensu stricto, is still a large, polymorphic genus consisting of more than 1000 species distributed into about seven subgenera ; in the southeastern United States it is represented by about 30 native or naturalized species belonging to four subgenera. All species of three other subgenera recognized by Wheeler (EREMoPHYTON, LyYCIOP- sis, and RHIZANTHIUM) are confined to the Old World and in our area will be found only in cultivation, Subgenus Esura Pers. (§ Tithymalus, of Boissier), with over 500 species, is probably the largest and most diversified group within Euphor- bia; it includes some of the most primitive as well as most highly special- ized taxa in the genus. Despite this remarkable amplitude of diversifica- Ag 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 399 tion, the great majority of taxa of subg. Esuta have in common alternate phyllotaxy (at least below), exstipulate leaves, and bibracteate cyathia with exappendiculate glands. Boissier’s treatment of this group in De Candolle’s Prodromus is unsatisfactory as regards supraspecific categories, since he crowded over 300 species into a single one of his 27 sections. Prokhanov, in Flora URSS, has tended toward the other extreme in recog- nizing ten sections and a large number of subsections and series for the relatively limited diversity of taxa found in the Soviet Union. In the present circumscription, the 17 Southeastern species of subg. EsuLa are referred to three sections. Section Latuyris Godron (§ Epurga Prokh.) includes only Euphorbia Lathyris L., a European species occurring as a weed (perhaps partially escaped from herb gardens). The species differs sharply from other repre- sentatives of subg. Esura in its decussate phyllotaxy; thick-walled, tardily dehiscent capsules; and large (4-5 mm. long), roughened, carunculate seeds. Although it was placed at the opposite end of subg. Esuxa in Boissier’s arrangement, the group closest to sect. LarHyris would appear to be sect. MyrsINITEAE (Boiss.) Pojero, a group of Mediterranean peren- nial species with similar cyathia (horns of glands dilated, male bracteoles _ reduced). Our remaining species, which have the leaves alternate (at least below) and less massive capsules with smaller seeds, are about equally divided between two sections (very perceptively distinguished by Roeper in 1824, although without distinguishing names). Section Esura [ (Roeper) Koch] includes those species of subg. Esura with truncate to crenulate or bicornate nectaries, linear to oblong cotyle- dons, smooth ovaries, and prevailingly entire cauline leaves. Over 200 species have been described, primarily from temperate regions of the Northern Hemisphere, but extending into tropical latitudes along high mountain ranges. The section is close to sect. TiIrHYMALUS and not always easily distinguishable from it, but the placement of most individual species is usually not in doubt. Perhaps the local species of most dubious relationship is Euphorbia trichotoma HBK., a West Indian coastal plant which occurs in southern peninsular Florida as far north as Tampa Bay. Norton appears to have been correct in transferring it from sect. TiTHy- MALUS, and it seems best regarded as our only representative of subsect. PARALIOIDEAE Prokh., diverging from other taxa in the section by its crenulate leaves and peculiar branching pattern. There is a rather striking resemblance between E. trichotoma and E. Paralias L., a Mediterranean plant of similar maritime habitats, and the two may perhaps be vicariants. Subsection Esutar Boiss., a primarily Old World group of perennials with smooth, carunculate seeds, is represented in our area only by the pernicious introduced weed Euphorbia Cyparissias L., 2n = 20, 40; other related species (e.g., E. Esula L.) have been introduced into the northern United States and Canada and may eventually be found in mountainous areas of the Southeast. Subsection FoveosPpERMAE Hurusawa, comprising Holarctic annuals or short-lived perennials with pitted, usually carunculate seeds, includes the 400 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 remaining five or six species in our area. Both exotic weeds, such as the ubiquitous Euphorbia Peplus L., and native species, such as E. commutata Engelm., are represented. Section TrrHyMatus Roeper, characterized by rounded, entire cyathial glands; suborbicular cotyledons; and usually smooth seeds, includes at least 150 species, mainly of north temperate regions but with perhaps a majority in Eurasia. Our species fall into three subsections. The type species of the section, Euphorbia Helioscopia L., is an introduced European weed belonging to subsect. GALARRHAEI Boiss., a group of about ten species including annuals or biennials with obovate serrulate leaves and carunculate reticulately ornamented seeds. Our only native species, the rather variable E. spathulata Lam. (E. dictyosperma Fisch. & Mey.) ete from E. Helioscopia in having verrucose, rather than smooth, cap- Bessie PurpPurATAE Prokh., a group of perennial ee species with generally entire leaves oy counibaaa ovaries and seeds, primarily Eurasian. Its only typical representative in North nndaitéa’s is the taxonomically rather isolated Euphorbia purpurea (Raf.) Fern. (E. Darlingtonii Gray). Two additional species in our area, E. platyphylla L. and E. obtusata Pursh, are anomalous in their annual habit, and E. obtusata is somewhat transitional to subsect. GALARRHAEI in its serrulate leaves and slightly wrinkled seeds. Subsection INUNDATAE Webster,27 an entirely New World group, is represented in the Southeast by three related Coastal Plain species: EF floridana Chapm., E. inundata Torr., and E. telephioides Chapm. These plants resemble E. Jpecacuanhae in their habit and female flowers with perianth, but they differ in lacking stipules and appendages on the cyathial glands. It would appear that subsect. INUNDATAE is a group transitional in certain respects between subg. Esura and subg. AcALomA. Boissier included a dozen South American species in the same subsection, but further study is needed to determine whether these should be closely asso- ciated with the species of the United States Subgenus AcAtoma (Raf.) House, in the circumscription of Wheeler, is an American group of ten sections with approximately 100 species. Although plants of very diverse habit are included, the subgenus can be characterized by the appendiculate glands of the cyathium; the majority of species also have stipulate leaves and ecarunculate seeds. Boissier’s treatnient of the taxa now gathered into subg. AGALOMA is not very satisfactory, as his ten sections (II-XI) are not very well defined and are really equivalent to his subsections in sect. TrrHyMaAtus; doubtless they “ Euphorbia sect. Tithymalus subsect. Inundatae Webster, subsect. nov., fo calvin — e stipulis nullis, capsulis calyculatis, seminibus ecarunculatis. TyPUS: Norton (Missouri Bot. Gard. Rep. 11: 2. 1899) pointed out that Euphorbia ae’ cuanhae L. t be transferred from subg. ULA to su Acatoma. Since lcci is 4, leaves alternate below to rpeeed opposite; 4— 5, reduction from five glands to four; 5 — 6, perennial to annual ha 426 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 there is no record of the intentional widespread cultivation of any species. Any possible economic importance would seem to be negative, since many species are undesirable weeds and few are both sufficiently large and abundant to serve as sources of latex. The arborescent Hawaiian species of sect. SCLEROPHYLLAE have been investigated as a possible source of rubber (McGeorge & Anderson), but the trees are limited in distribution, and the latex, with over 55 per cent resin, does not appear to be of a commercially promising quality. In common with species of Euphorbia (mainly subgenera Esuta and EupHorsiA), various weedy species (e.g., C. hypericifolia, C. prostrata, C. thymifolia) have been shown to harbor trypanosomes of the genus Leptomonas in their laticifers (see Mesnil, and references under Euphorbia). REFERENCES: Under family references see BANERJI & Dutt, Croizat ake pp. 182-189), GAUCHER, Hurusawa, KiotzscH, McVaucH, MUuRLEY, & FFMANN (1930), Punt, and SCHWEIGER: under Euphorbia see oo CroIzaT (1940) D’AMATO, DENIS, GAUCHER, HABER, Harvey & Lee, HuruSAwA, KrocHMAL, Mesnit, Moyer, PAMMEL, ROEPER, SHERFF, WENIGER, and WHEELER Burcu, D. Two species of Chamaesyce (Euphorbiaceae) new to the United States. Rhodora 67: 185, 186. 1965. The application of the Linnaean names of some New World species of ‘Euphorbia subgenus Chamaesyce. Ibid. 68: 155-166. 1966. [Includes a complete bibliography of the earlier nomenclatural polemics of Crorzart, FosBerc, and WHEELER. | . Two new species of Chamaesyce (Euphorbiaceae), new combinations and a key to the hae members of the genus. Ann. Missouri Bot. Gard. 53: 90-99. 1966. Croizat, L. thar ee in the United States. Bull. Torrey Bot. Club 72: 312-318. & O. DEGENER. mec In: O. DEGENER, Fl. Hawaii. 1936-1946. [Chamaesyce,, 1936, revised 1938; Chamaesyceés,s, 1937; Chamaesyce,, 1936, revised 1946; in cludes discussion of generic relationships and illustrations of vegetative habit. ] Hucuet, P. Euphorbia maculata L. Observations sur la morphologie des stipules. Bull. Mus. Hist. Nat. Paris IT. 36: 407-411. pls. 1-3. 1964. KajaLe, L. B. Fertilization and the development of embryo and seed in Eu- phorbia hirta Linn, Proc. Natl. Inst. Sci. India 20: 253-260. 1954. KroMBEIN, K. V. Some insect visitors of mat euphorbia in southeastern Ari- zona. (Hymenoptera, Diptera.) Entomol. News 72: 80-83. 1961. [Cham- aesyce albomarginata; visits mainly by Sphecidae, Andrenidae, and Halic- tidae. McGeorce, W., & W. A. ANDERSON. a aietiges lorifolia, a possible source of rubber and chicle. Press Bull. Haw i Agr. Exp. Sta. 37: 1-16. 1912. [Physical and chemical properties of ie of Euphorbia, by McGerorce, 1-13; occurrence of Euphorbia lorifolia and tapping, by ANDERSON, 14—16.] MuKERJEE, P. K. Study of mature embryo, seed coat and fruit wall in Exphor- bia cristata Heyne. Jour. Biol. Sci. 4: 1-5. 1961. Orcutt, C. R. The golondrina plant. W. Am. Sci. 7: 190-195. 1890. 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 427 RAMAKRISHNAN, P. S. Edaphic ecotypes of Euphorbia thymifolia. Jour. Ecol. >. 157-162. 1965. SHINNERS, L. New names of Texas Chamaesyces. Field Lab. 17: 69, 70. 1949, denda on Texas Chamaesyce. Ibid. 20: 24~26, 1952. SIQUEIRA- -Jaccoun, R. J. pe. Contribuicgao para o estudo da Euphorbia brasil- tensis Lam. Mem. Inst. Oswaldo Cruz 54: 103-113. 1956. [Observations n anatomy, especially on laticifers, and occurrence of trypanosomes. | TAKeMoTo, T., & M. Inacakr. Constituents of Euphorbia maculata. Constit- uents of Euphorbia pilulifera. (In Japanese; pe summary.) Jour. Pharm. Soc. Japan 78: 292~294; 294, 295. 1958 THELLUNG, A. Die in Europa bis bet secepeaerer Explor Arten der Sek- tion Anisophyllum. Bull. Herb. ad 7s TANAT72. Ven, R. von. Beitrag zur Kenntnis a ronan lbplinuitese und einige vergleichende und entwicklungsgeschichtliche Untersuchungen iiber die vrei des Cyathiums. Ann. Jard. Bot. Buitenzorg 38: 131-162. pls. 13~22. 928. Warminc, E. Disposition des feuilles de Euphorbia buxifolia Lam. Danske Vid. Selsk. Forh. 1896: 326-334. 1896. WarteRFALL, U. T. A new species of Euphorbia from pegnviny Rhodora 50: 63, 64. 1948. [E. carunculata Waterfall, a sand-dune p WHEELER, L. C. Euphorbia subgenus Chamaesyce in Caisdh and the United States exclusive of southern Florida. Rhodora 43: 97~154, 168-205, 223~ 286. pls. 654-668. 1941, [Detailed monographic revision; includes exten- sive bibliography and references to Wheeler’s earlier paper Wrecanp, K. M. A neglected North American Euphorbia. Bot. Gaz. 24: 49-52. I. 3. 1897. [E. hirsuta Wieg. critically arama and _ illustrated; synonymized with £&. apa ain he by Wheele 24. Pedilanthus Poiteau, Ann. Mus. Hist. Nat. Paris 19: 388. 1812, nom. cons : Trees or shrubs, twigs and foliage more or less succulent, with whitish [rarely yellow] latex. Leaves alternate, distichous, deciduous or persis- tent, stipulate; blade entire, pinnately veined, eglandular. Plants mon- oecious; inflorescences bisexual pseudanthia, these borne in terminal or (by reduction) lateral dichasia. Pseudanthium (cyathium) bilaterally symmetrical, the involucral tube produced into a posterior spur containing 4 lor, by modification, 2 or 6| glands; female flower solitary in the cyathium, central; male flowers in 5 more or less bracteolate monochasia. Flowers lacking a gaol disc undeveloped. Male flower monandrous, simulating a stamen; filament separated from pedicel by an articulation; anther dehiscing longitudinally and abaxially; pollen tricolporate, per- forate-tectate, with paired intine thickenings parallel to the colpi. Female flower pedicellate; perianth perhaps represented by a disciform structure at base of ovary; carpels 3; styles connate along most of their length into a slender column much longer than the ovary; style tips bifid; ovary angled or carinate; ovules 1 in each locule, anatropous; nucellus short- beaked; embryo-sac development normal. Fruit capsular [rarely inde- hiscent]; columella slender, expanded distally. Seeds subterete, testa smooth or minutely tuberculate, ecarunculate; endosperm copious; em- 428 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 bryo straight, cotyledons somewhat fleshy, broader than the radicle. Tithymalus sensu Small.) Lectotype species: P. Tithymaloides (L.) Poit.; see Millspaugh, Field Mus. Publ. Bot. 2: 353. 1913. (Name from Greek, pedilon, sandal, and anthos, flower.) — SLIPPER-SPURGE. A very natural and well-characterized genus of 14 neotropical species, mostly concentrated in Mexico. Dressler has given the genus an exhaustive revision and provided an excellent model and stimulus for further work in the tribe Euphorbieae; the present account leans heavily on his work. Our single native species is Pedilanthus Tithymaloides (L.) Poit., 2n = 34, 36(?), which is represented in Florida by ssp. Smallii (Millsp.) Dressler, a shrub up to 2 m. high with succulent, conspicuously zig-zag stems. It has been collected in pinelands and hammocks from the eastern shore of Lake Okeechobee, south to Meigs Key (Monroe County), as well as in northern Cuba, near Matanzas. Dressler regards this Floridian sub- species as closely related to ssp. parasiticus (Kl. & Garcke) Dressler, of Central America and the Greater Antilles, and suggests that both taxa may have invaded the West Indies from northern Central America. If his reconstruction of the migrational history is correct, this species has made a remarkable double invasion of the West Indies: once from Central America to Cuba and Florida, and once from the northern coast of South America via the Lesser Antilles to the Greater Antilles. Unfortunately, the popularity of the species as a hedge-plant has led to its cultivation and escape in many areas, so that the provenance of many older collections may be doubtful. In the Miami area ssp. Tithymaloides (and perhaps other subspecies) is cultivated and may possibly be found escaped to straight, rather than zig-zag, stems, but identification of the variegated- leaved cultivars is often difficult. Pollination of the slipper-like cyathia by hummingbirds has been re- ported for two species by Dressler, doubtless to the relief of the ethologists who had dubbed Pedilanthus a “humming-bird flower” on the basis of circumstantial (i.e., purely morphological) evidence. Additional observa- tions, including studies in Florida, are much needed. Dressler has not unreasonably speculated that the unusual bilaterally symmetrical con- formation of the Pedilanthus cyathium represents an “adaptive peak” attained as a result of pollinator selection from an ancestral type of ior Jayolucre. be zat has ae (1962) launched a violent attack on Dressler’s “ornitho- genetic” ne of the origin of the cyathium of Pedilanthus. It is true that this hypothesis is Speedin cet eat! circumstantial, particularly with regard to estimations of geological time. I hope, however, that Croizat does not expect that eschewing attempts to synthesize a unified picture of the phenomena they are in- vestigating. Croizat’s own “explanation” of the cyathium of Pedilanthus is not espe- cially illuminating, perhaps because he has been overly influenced by Bodmann’s mis- leading work on Euphorbia. To the extent that morphologists have neglected develop- mental considerations and attempted to draw mechanical one-to-one homologies be- ween organs such as the parts of the cyathium, some of Croziat’s comments are 1967 | WEBSTER, GENERA OF EUPHORBIACEAE 429 According to Dressler, Pedilanthus is related to and has been derived from Euphorbia subg. AGALOMA, with which it corresponds in many re- spects, differing mainly in its zygomorphic cyathia. As clearly shown in Dressler’s diagrams, the projecting anterior “beak” of the Pedilanthus cyathium is formed mainly from a pair of involucral bracts, whereas the posterior spur has been derived mainly from the remaining three bracts and the four adjacent partially connate (and highly modified) gland ap- pendages. The closest surviving species to the putative ‘“‘ancestral type” is perhaps the familiar Mexican species Euphorbia fulgens Karw., of sect. TRICHEROSTIGMA (KI. & Garcke) Boiss., which has alternate leaves and brilliant-red petaloid appendages; a plant of this kind might be con- sidered “pre-adapted,” in the Dresslerian sense, to enter on the road of number and appendages of cyathial glands). The West Indian genus Cubanthus (Boiss.) Millsp., which was included in Pedilanthus as sect. Cusantuus by Boissier and treated as a genus by Dressler and Mill- spaugh, differs strongly from Pedilanthus in its basically actinomorphic cyathia with exappendiculate glands. As pointed out by Dressler, it stands quite apart from both Pedilanthus and subg. PormNsETTIA, and ap- ears to have been derived from sect. ADENORIMA of subg. Esuta. Com- parative studies of pollinating agents in these three groups, although necessitating field work in some difficult areas, would be of great evolu- tionary interest and might help elucidate certain aspects of cyathial evolution in the more primitive neotropical Euphorbieae. REFERENCES: Under family references see BAILLON (1858), BLou, Crorzat (1940, 1942a, < ATTEI, MIcH- 1943), E1cHLER, GAUCHER, HurusAwa, INGRAM, KLOoTZSCcH, M AELIS, Perry, and Punt; under Euphorbia see Borsster, Croizat (1936, 1937), Porscu, Ripota, and WHEELER. Banerjt, I. Pollen and embryo-sac of two Euphorbiaceae. Proc. Indian Acad. Sci. 34B: 172-181. 1951. [Normal embryo-sac development reported in P. Tithymaloides. | However, in my opinion, he goes too far in adopting a counsel of despair , etc.” of the Pedilanthus cyathium are morpho- h much time or effort to quibble over of the glands of primitive gosh further study of the cyathia of that remarkable plant is indicated. 430 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Buttock, A. A. Pedilanthus versus Tithymalus. Bull. Misc. Inf. Kew 1938: 468-470. 1938. Croizat, L. Tithymalus or Pedilanthus? Nomenclatural considerations, notes, new names and combinations. Am. Jour. Bot. 24: 702-704. 1937. . Peculiarities of the inflorescence in the Euphorbiaceae. Bot. Gaz. 103: | Animadversions on Dressler’s monograph, 103-137. . The genus Pedilanthus (Euphorbiaceae). Contr. Gray Herb. 182; 1-188. pls. 1-21. 1957. [Definitive monograph, includes extensive bibliography of references not itemized here. MarkowskI, A. Beitrage zur Kenntnis der Gattung Pedilanthus. 51 pp. Diss. Halle. 1912.* MituspaucH, C. F. The genera Pedilanthus and Cubanthus, and other American Euphorbiaceae. Field Mus. Publ. Bot. 2: 353-373. 1913. Riwora, F. Interpretazione morfologica del ciazio di Pedilanthus. Bull. Orto Bot. Univ. Napoli 1: 415-418. 1903. Wueeter, L. C. Typification of the generic synonyms of Pedilanthus. Contr. Gray Herb. 124: 43-46. 1939a. . Pedilanthus and Cnidoscolus proposed for conservation. Ibid. 47-52. 1939b. , L. Curax, & A. Wuite. Confusion among the slipper flowers. Cact. Succul. Jour.116: 81-83. 1944. [Discussion of two West Indian taxa, '?. angustifolius and P. Grisebachii. | ADDENDA (1) Beguiled by long usage of and excessive familiarity with the subfamily name Crotonoideae Pax, the author [and editor] failed to notice that since it includes the type genus of the family (Euphorbia), the subfamily must under the current rules of nomenclature take the name Euphorbioideae. Consequently, the reader should substitute Euphorbioideae for Crotonoideae wherever it occurs in the earlier part of the text. The following reference was omitted from the general bibliography (Jour. Arnold Arb. 48: 315): CHANDLER, M. E. J. The Lower Tertiary floras of southern England. I. Palaeocene floras. London Clay flora (supplement). xi + 354 pp. Atlas, 34 pls. (separately bound). Brit. Mus. (Nat. Hist.), London. 1961 [Euphor- biaceae, 203-211]; II. Flora of the pipe-clay series of Dorset (Lower Bag- shot). xi + 176 pp. 29 pls. Brit. Mus. (Nat. Hist.), London. 1962 [ Euphorbiaceae, 80-90]. In the generic key (Jour. Arnold Arb. 48: 322), the final part of the second lead A should read: pollen colporate, porate, or inaperturate (Subfam. Euphorbioideae). The generic descriptions of Manihot and Cnidoscolus are excessively reticent with regard to the calyx; it is pentamerous in both genera. oo bh dl a w — “-_ _ — DEPARTMENT OF BOTANY UNIVERSITY OF CALIFORNIA Davis, CALIFORNIA 1967 | MORTON, RED RIVER EXPEDITION OF 1806 431 FREEMAN AND CUSTIS’ ACCOUNT OF THE RED RIVER EXPEDITION OF 1806, AN OVERLOOKED PUBLICATION OF BOTANICAL INTEREST C. V. Morton My ATTENTION WAS DRAWN some twenty-five years ago by the late Dr. John Swanton, of the United States Bureau of American Ethnology, to a small book in the library of the U. S. Geological Survey that was of interest to him for its information on certain Indian tribes. This anonymous book is entitled “An Account of the Red River in Louisiana, drawn up from the returns of Messrs. Freeman and Custis, to the War Office of the United States, who explored the same in the year 1806.” * Swanton mentioned this book and quoted extensively from it, in con- nection with the Caddo Indians, in his book “Source Materials for the History and Ethnology of the Caddo Indians” (U. S. Bur. Am. Ethnol. Bull. 132. 1942). The work is mentioned briefly and with scant regard by William H. Goetzmann in his fascinating and wonderfully documented book “Army Exploration in the American West 1803-1863” on pages 34-36 (Yale University Press). This Freeman and Custis work is of historical and ethnological interest, but in addition it contains lists of plants (including some novelties), animals, birds, insects, amphibians, and so forth, and the plants, at least, have been completely overlooked by subsequent botanists up until the present day. There is, for instance, no mention of this publication at all in Mrs. Susan Delano McKelvey’s monumental book “Botanical Exploration of the Trans-Mississippi West, 1790-1850.” The botanical part of this narrative is quite interesting, being the first published information concerning the native plants of Loui- siana after the Louisiana Purchase, and it includes perhaps the first mention of some important plants, such as the Osage Orange. HISTORICAL ACCOUNT OF THE EXPEDITION AND THE PUBLICATION The historical background of the expedition to the Red River is briefly as follows: While the English were colonizing the eastern seaboard of the United States, the French were in control of Canada and Louisiana, which included all the territory on the western side of the Mississippi River, in- cluding the present states of Louisiana, Arkansas, Kansas, Missouri, Ne- braska, Iowa, North and South Dakota, Minnesota, and Montana. At 2 Octavo, 63 pp., 2 folded tables. Without place of publication, without date. The Library of Congress card and Swanton give the date as 1806, but Goetzmann gives it as 1807, and that the latter is probably correct is indicated below (see p. 43 432 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 the end of the French and Indian War, in which the French were disas- trously defeated, all of Louisiana was ceded to Spain by the Treaty of Paris, February 10, 1783. However, for reasons best known to Spain and Napoleon, Lenisiand was ceded back to France by the secret Treaty of San Ildefonso in 1800. Alarmed by this evidently not so secret treaty, President Jefferson sent James Monroe to Paris to negotiate for the sale of New Orleans, which was needed as a seaport by the settlers in Tennessee, Kentucky, and elsewhere in the Middle West. Surprisingly, Napoleon, at the time in firm control of France as First Consul of the Republic, agreed not only to sell New Orleans but all of Louisiana for $15,000,000, surely the greatest bargain in history since the sale of Manhattan Island to the Dutch by the Indians. The sale was agreed to with alacrity by Jefferson and ratified by the United States Senate in 1803. Jefferson’s first concern was to ascertain the boundaries of the new territory. An expedition to explore the upper part of the Purchase was immediately organized, the Lewis and Clark Expedition of 1803-1806, which was brilliantly successful and justly acclaimed. The Red River, a tributary of the Mississippi River running northwest from about the middle of the present state of Louisiana, was supposed to be the south- western boundary between Louisiana and the Spanish possessions in Texas and the west. Two expeditions were authorized. One, authorized by Henry Dearborn, Secretary of War, March 30, 1804, was for an expedi- tion under the command of William Dunbar and George Hunter.” This expedition got under way October 16, 1804, from St. Catherine’s Land- ing, on the Mississippi River below Natchez. It proceeded up the Red River only about 26 miles, and then turned north on the tributary Black River (now the Ouachita River) and went up into present-day Arkansas. Hunter returned to St. Catherine’s Landing on January 31, 1805. The journal of this expedition was in family hands until very recently, but it has now been published as “The Western Journals of Dr. George Hunter, 1796-1805” edited by John Francis McDermott, in the Transactions of the American Philosophical Society (new series, vol. 53, part 4, 1963). At almost the same time as the Dunbar and Hunter expedition, another expedition to explore the Red River was authorized April 14, 1804, under the command of Thomas Freeman, an expedition doomed to failure and consequently known today only to historians. The expedition was delayed for almost two years by the opposition of the Spanish and also by the Osage Indians, but it finally started April 19, 1806, perhaps using some of the supplies left behind for its use by Hunter after his return the pre- vious year. The party consisted of “Mr. Thomas Freeman, Surveyor, unbar’s account of this expedition has been published from a manuscript in the library of the American Philosophical ragny “Journal of a Voyage Commencing at St. Catherines landing, on the East bank of the Mississippi, proceeding downwards to the mouth of the Red river, and from thence ascending that river, the Black river and the Washita river as high as the Hot-Springs in the proximity of the last men- tioned river,” in “Documents relating to the Purchase and Exploration of Louisiana.” Houghton, Mifflin & Company, 1-189. 1904. 1967 | MORTON, RED RIVER EXPEDITION OF 1806 433 who was furnished with the requisite instruments, for determining geo- graphical positions by astronomical observations; Dr. Peter Custis, whose attention was directed to botany, and natural history; Captain Sparks, and Lieutenant Humphreys, two non-commissioned officers, seventeen pri- vate soldiers, and a black servant.’”’ The connection, if any, between the Dunbar-Hunter Expedition and the Freeman-Custis Expedition is not at all clear. In Jefferson’s letters to Dunbar, sai by McDermott, there is no mention of Thomas Freeman or Peter Custi The party entered the Red River from the Mississippi May 3, expecting to be able to ascend in their boats to the country of the Pawnee Indians. Their orders were * to purchase horses from these Indians and to ascend to the top of the mountains, the distance being supposed to be about 300 miles. It is evident that the distance was vastly underestimated. As a matter of fact, the real source of the Red River was unknown and un- explored until the famous Marcy Expedition of 1852, almost 50 years later.* The expedition proceeded upstream through country inhabited by friendly Indians until July 28, when a point was reached that was held by a force of perhaps 300 Spanish cavalrymen under the command of Captain Don Francisco Viana. It is said that by his strategy and diplo- macy Mr. Freeman saved the party from destruction. However, from the published account it seems that the Spanish commander was civil and not unfriendly but firm in his purpose not to allow the expedition to proceed. Since the Americans were far outnumbered, there was no alter- native but to turn back, which the party did on July 30, arriving back in Natchitoches on August 23. Since the expedition did not reach its des- tination it was evidently considered a failure, even though it did provide some information on the course of the river, the countryside, the climate, the Indian inhabitants, and the flora and fauna. The farthest point reached by the expedition, that where it was stopped by the Spanish, has been a matter of conjecture. Sabin ® summarizes the information as follows: “Map makers and later writers are far from agreement as to where on the Red River Freeman was halted by Captain Viana. Thwaites, editor of the James- Long expedition (Early Western Travels, Vol. XVII) is clearly wrong in stating that the halt occurred where the Little River joins the Red. This is fifteen miles or so east of the Texas-Louisiana north-south boundary and thirty miles or so east of the western boundary of Arkansas.6 The long note in Wagner-Camp to the 1810 Pike is delightfully indefinite on this point, as is Thomas M. Marshall ® The original orders are in the manuscript division of the Library of Congress. *Marcy, Randolph B. Exploration of the Red River of Louisiana in the year 1852, pp. i-xv, 1-286. 1854. Botany of the Marcy Expedition by John Torrey, op. cit. pp. 245-272, pl. I-XX. The botanist-collector of the Marcy Expedition was Dr. G, C. Shumard. ° Sabin, Joseph. Bibliotheca Reticana, A dictionary of books relating to America 1875. from its discovery to the present time. * Sabin is right; it is obvious from the bllshed account that the expedition pro- ceeded much farther than this point. 434 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 | (JS 1e)s Py < ~~ 2 : : ) OTe a5 SS aS ; “Si. P ee 4s CC a FIGURE I. Photograph of left half of the original “Map of the Red River in Louisiana” of Freeman and Custis. 1967 | MORTON, RED RIVER EXPEDITION OF 1806 435 ‘ a - | b> HW : | ‘ | f _ r’ i fe ; 4 if | | ; 4 : é : * N ¥ * L i ale a : i i j ! : ; a * : i ba i] ]x 3 * “i we iy lac a ee. > Oe A ee ; (| OY 4 3) is Oe. | 7, “a ¥ or ' ‘ SRiSSs Se ae tke * Mi. : lal s, Fos 5 he Ficure 2. Photograph of ke half of the original “Map of the Red River in Louisiana” of Freeman and Cus 436 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 in his History of the Western Boundary of the Louisiana Purchase. My guess is that the Freeman party was halted near the little Texas village of River View on the Red River about thirty miles or so west of the north-south Texas-Louisiana boundary. This may be the ‘handsome bluff’ mentioned in the Freeman account. This would fairly well agree with the location for the halt given in the famous Robinson Map of Mexico, Louisiana. . . Philadelphia, 1819 (No. 1073),* which places it twenty-five miles or so west of the present Texas-Louisiana line. The map in the Philadelphia, 1810, Pike (No. 1047), ‘The First Part of Captn Pike’s Chart of the Internal Part of Louisiana’ is invaluable for following the route of the Freeman expedition up the Red River, but its legend for ‘the explor- ing Party stopped here’ seems to be too far west by at least a degree of longitude; and Map No. 55, showing Arkansas Territory, in P. M. G. Van der Maelen’s Atlas Universel, Bruxelles, 1827 (No. 1095), is almost certainly in error in the legend it has at the confluence in Oklahoma of the Red River and Boggy Creek (about 80 miles west of the Texas-Louisiana line), ‘La R. rouge a été reconnue jusqu’ici par Th. Freeman.’ Sabin and the writers mentioned by Sabin seem to have been unaware that the original map of Freeman and Custis was in existence. It is still preserved in the National Archives (Map M33-1); it is reproduced in this paper as Figures 1 and 2. This map shows that Sabin was right, that the stop was at or near the present River View, Red River County, Texas, a point very near the Texas-Oklahoma-Arkansas meeting point. The expedition was, therefore, mostly confined to present-day Louisiana, but did traverse the southwestern part of Arkansas, and a short distance along the boundary between Texas and Oklahoma. The published account of the Red River expedition is a small book of 63 pages, probably always very rare. Copies are in the Library of Congress; U. S. Geological Survey; Bancroft Library, University of Cali- fornia, Berkeley; Howard-Tilton Memorial Library, Tulane University; Boston Athenaeum; Clements Library, University of Michigan; Harvard University; Streeter’s Collection, Morristown, New Jersey; and the Maryland Historical Society, Baltimore. The book does not have any place or date of publication. The place was undoubtedly Washington, D.C., as given on the Library of Congress card, which gives the date as “18062”, and Swanton (U. S. Bur. Am. Ethnol. Bull. 132. 1942) gives the date as 1806 without a question mark. However, Custis’ final report is dated at Fort Adams, October 1, 1806. For this report to reach Wash- ington (probably requiring several weeks overland by stage coach), and for it to be combined with the previous report and with Freeman’s re- ports, would seem to require too much time to have the account published before the end of 1806. I am indebted to my friend, Mr. A. DeWeese, chief of the Information Desk, New York Public Library, for looking into the matter. He has found a reference establishing the probable date as 1807, which is the date to be expected. Item no. 1040 in Thomas W. Streeter’s Bibliography of Texas is devoted to this book of Freeman and Custis, and the third paragraph discusses the date of publication as follows: “These numbers refer to the appropriate entries in Sabin’s Bibliotheca. 1967 } MORTON, RED RIVER EXPEDITION OF 1806 437 “The 1807 date given here to the Freeman pamphlet was suggested by the late Wilberforce Eames, on what was for me a memorable occasion when, not long before his death, Mr. Eames, accompanied by Lathrop Harper, spent the day with me in Morristown. He pointed out that as the copies of the Freeman pamphlet and the undated Account of a Voyage up the Mississippi River . in the Years 1805 and 1806. Compiled from Mr. Pike’s Journal had come to the Library of Congress and to the Bancroft Library sewn together and as in his copy of the Account there was laid in a letter of presentation signed by General Dearborn, the Secretary of War, and dated, ‘War Department, March 9, 1807,’ there was a strong presumption that the Freeman pamphlet was published early in 1807. Edwin James in his Account of An Expedition from Pittsburgh to the Rocky Mountains, Philadelphia, 1823, says he had access to the Freeman journal, now lost, and at pages 306-314 of Volume III gives an account of the expedition which, in places, supplements the Account entered here. There is an interesting article on Freeman in the Dictionary of American Biography. That article gives the place where Freeman was halted as ‘near where the present boundaries of Arkansas, Oklahoma and Texas meet.’ ” Thomas Freeman, who led the expedition, was born in Ireland * and emigrated to the United States in 1784. On March 25, 1794, he was ap- pointed to survey a part of the newly created District of Columbia, a survey completed June 25, 1795. After planting the stones for the bound- ary, he commenced the first topographical survey of the District, but resigned July 7, 1796, to accept an appointment as U. S. surveyor to determine the boundary between the United States and Spain, which at that time held Louisiana in control. After the Red River Expedition, he mapped a part of the boundary between Alabama and Tennessee in 1807. In 1811 he was appointed U. S. Surveyor of Public Lands south of Tennessee, a position he held until his sudden death at Huntsville, Alabama, November 8, 1821. I have been unable to find out anything definitely about the ancestry of Peter Custis, the botanist who accompanied the expedition, perhaps as a substitute for George Hunter, who, it appears, did not wish to return to the Red River for personal reasons. In spite of its prominence in con- nection with George Washington and with Robert E. Lee, both of whom had connections through marriage, there is no published genealogy of the Custis family. Peter Custis was the son of one Robinson Custis, of Accomac County, Virginia, in whose will of 1797, he was left a part of the estate for “a Latin education and to be brought up in one of the learned professions.” ® Peter Custis is also mentioned in R. T. White- law’s “Virginia Eastern Shore,” 1° where he is said to be a doctor. He did receive the degree M.D. from the University of Pennsylvania in 1807, that is, after he returned from the Red River Expedition. Since he is mentioned in the published account of the expedition as “Dr. Peter Custis” this may indicate that he obtained the degree early in 1807 before * Dictionary of American Bicgraphy. 7: 13. 1931. ; : ® This information was kindly supplied by a member of the staff of the New York Public Library who has Custis family connections. 10 2, 438 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 the account was published or that he was given the title “Dr.” in an- ticipation of his receiving the degree shortly. The other two leaders of the expedition are not of especial interest in this connection. Richard Sparks was appointed First Lieutenant of the U.S. Infantry in 1791, a Captain in 1792, a Major in 1806 (presumably after his return from the Red River Expedition), a Lt. Colonel in 1807, and a Colonel in 1812. He died in 1815. Enoch Humphreys was ap- pointed Lieutenant, First Artillery and Engineers in 1801, a Captain in 1809, and later a Major. He died in 1825. Although the Account of the Red River is anonymous, it must be attributed to Freeman and Custis, as it is by the Library of Congress, by Streeter in the work mentioned above, and by Sabin.11 The work was compiled from one or more reports by Freeman to the Secretary of War and from two reports by Custis, obviously by someone fairly incompetent because of the numerous misspelled words, inconsistencies in punctuation, and especially by the many misspelled scientific names. The original manuscript reports by Custis are still in existence, in the National Archives in Washington, available on microfilm.12 Custis’ handwriting is excep- tionally clear but even so the published work has numerous errors of transcription. All the data regarding the natural history of the region and the meteorological data come from Custis, the general remarks re- garding the progress of the expedition and the encounter with the Spanish evidently come from a report by Freeman, the manuscript of which is apparently lost. Dr. Custis’ second letter, enclosing the final lists, follows. “Fort Adams ugie Oct. 1, 1806 “Inclosed I have the honor to forward my last communications, consisting of Lists of such vegetables & animals as you have not already had an account of, together with one of all the trees of Red River & my meteorological observations. — I have to lament that a more complete opportunity had not occurred, that I might have been enabled the better to perform the duties alloted me, but it is a thing well known that a person subject to the movements of another has little, or I might add, no time to make botanical excursions. — With respect to the country it will be unnecessary to say anything, as it is similar to that of which you have before had accounts. I could however report what I mentioned in my last letter that in point of beauty, fertility and salubrity there is not its equal in America, nay in the world. — The water of Red River above Coashatta Village is so strongly impregnated with salt as to render it unfit to drink. When this water is low we find the sand beaches rendered white with salt. “T have to return my most sincere thanks for the appointment with which you have honored me, & to regret the failure of the expedition. — Permit me, Sir, to tender you my best wishes. Yrs “The Honble. Peter Custis Henry Dearborn Secretary of War.” abin, op. cit. 16. ™ Letters Received by the Secretary of War 1806, M221, Roll 4, C140, C207. 1967 | MORTON, RED RIVER EXPEDITION OF 1806 439 Perhaps the very first list of the trees of the lower Mississippi River region was that given in The Journal of Andrew Ellicott (first edition, Philadelphia, 1803, republished by Quadrangle Books, Inc., Chicago, Illi- nois, 1962). LEllicott’s Journal is concerned mostly with political and military matters, but on pages 285 to 289 of the republished edition he gives a list of some of the common trees and other plants. Although he professes himself an indifferent botanist it appears that he did know a good deal, and many of his names are more correct than those supplied later by Custis. Although Ellicott was never on the Red River, many of the same plants occur along the banks of the Mississippi also, and, therefore, it is not surprising that Ellicott should mention many of the same species as those commented on by Custis. There is no indication that Custis ever saw this list of Ellicott; if he had, he might very well have made some changes in his nomenclature. A list entitled “Common Names of some of the Trees, Plants, Shrubs etc. growing in the Country adjacent to the Ouachita” was prepared by the Dunbar and Hunter Expedition of 1804, probably by William Dun- bar. This list is in the Hunter Papers in the library of the American Philosophical Society but was never published until] by McDermott in The Western Journals of George Hunter, 1796-1805, in 1963 (pp. 120, 121). It consists of common names only; the plants listed are in part the same as those listed by Custis, but they can not be at all definitely determined. Another shorter “List of Vegetables (perhaps non-descripts) from the River Washita” was also prepared, and this also was published by McDermott (pp. 121, 122). This list contains extensive notes on a few plants considered to be of especial interest, among them being the “Bois d’arc” or Osage Orange. There is no reason to believe that Custis saw these lists either. Peter Custis wrote a letter from Natchitoches, dated June 1, 1806, to his teacher, Benjamin Smith Barton, and signed it ‘“‘your sincere and ever-devoted Friend and Pupil.” This letter contained a brief account of the trip up to that date, and included, of course, a number of the same observations that were published later in the full account. Barton lost no time in publishing this letter, in his “The Philadelphia Medical and Physical Journal, collected and arranged by Benjamin Smith Barton, M. D., Professor of Materia Medica, Natural History, and Botany, in the University of Pennsylvania,” vol. 2, part 2, pp. 43-48, Sept. 26, 1806. This is the first published record of the natural history of the expedition, but it also has been completely overlooked. A good many of the trees noted in the later account are also recorded here, including the first men- tion of Juglans petiolata Custis. Custis presumably did not take along equipment for preparing her- barium specimens which was probably not considered necessary, since this was not primarily a scientific expedition for studying the natural history of the region. Moreover, there would probably not have been room for much equipment in his small boats. His botanical observations 440 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 were of course based on living material, but he did manage to collect 26 numbers of herbarium specimens, which he enumerated (see below, p. 455) but did not name. These probably represented species that he could not name from the books at hand, and it may be conjectured that most of them were undescribed at the time. These herbarium specimens were never formally identified and reported on, apparently. They were given to Custis’ teacher, Benjamin Smith Barton who, in 1791, had been appointed the first Professor of Natural History and Botany in the United States, according to Francis W. Pen- nell (The Elder Barton — His Plant-Collection and the Mystery of his Floras, Bartonia 9: 17-34. 1926). Among the contents of Barton’s her- barium Pennell mentions (op. cit. p. 21) “Peter Custis specimens from Virginia and ‘450 miles up the Red River.’” Frederick Pursh was em- ployed in 1806 and 1807 as a collector by Barton and apparently also as a herbarium assistant, since most of the specimens in the Barton her- barium bear determinations in Pursh’s hand. It is likely that he saw and may have identified the Custis Red River collections, and it may be that some of these formed the basis of some records in Pursh’s later Flora Americae Septentrionalis (1814), which was prepared after Pursh went to England. That the Custis specimens have remained unknown is the result of the general neglect of the Barton Herbarium, which, after Bar- ton’s death in 1815, was presented to the American Philosophical Society. It was kept but remained essentially unknown and unused until it was turned over to the Academy of Natural Sciences some 80 years or more later. Of course, the Barton Herbarium is historically and botanically very important because it contains more than 800 of the collections of Frederick Pursh. Since the Custis collections do not even bear generic names, except for three indicated as Mimosa, Lonicera, and Polypodium, and they have not been cited in the botanical literature, they can not be otherwise identified. As may be seen from a perusal of the lists and my commentaries, which follow, Custis’ identifications of the plants observed leave much to be desired, but in extenuation one must consider the circumstances. It is hard for us now to realize the paucity of literature existing on the plants of the southern United States in 1806 and particularly on those of the western part, in Louisiana. Michaux’s Flora Boreali-Americana had ap- peared in 1803, but it is evident that Custis had not yet seen it, for he mentions none of Michaux’s species. He was familiar with Marshall’s Arbustum Americanum (1785) and Walter’s Flora Caroliniana (1788), for Marshall and Walter are mentioned, but it is doubtful that he had these works along on the expedition. Professor Ewan suggests that he was probably also familiar with Bartram’s Travels and with Thomas Jefferson’s Notes on Virginia, which is certainly likely, but he would not have had these works along with him either, and they would not have been very helpful in any case. What is certain is that he did have and use Linnaeus’ Systema Vegetabilium, ed. Murray (either ed. 14, 1784, or ed. 15, 1797, 1967 | MORTON, RED RIVER EXPEDITION OF 1806 441 which are very similar in format and content), because all the names used are found in this work and many of them are not in any American works available at that time. The identifications came from Linnaeus’ Systema, a work in which all of the known plants of the world were crammed into one volume. The treatment is the ultimate in condensation, consisting only of diagnoses of from six to ten words, and sometimes a citation or two. All the de- scriptions of the Species Plantarum (ed. 3) are omitted, and also all men- tion of geographic distribution. I can picture Custis sitting in the open flat boat while it was proceeding up the river at a snail’s pace trying to match up his specimens with these brief Linnaean diagnoses, throwing up his hands at the impossibility, and finally tossing a coin as to which name to enter in the list. This explains why he reports from Louisiana SO many species known only from Europe, Africa, Asia, Australia, and tropical America. I had much the same experience when I was a student at the University of California. Dr. Harvey M. Hall asked Tom Howell and me to identify the plants for a local wildflower show in Berkeley. We were unexpectedly deluged with specimens, many not at all local but from the Sierras and distant Coast Ranges, and many belonging to diffi- cult genera like Lupinus, Gilia, Phacelia, and Arctostaphylos. As the time for the opening of the show grew close we, in desperation, just put down species names almost at random, in order for the plants to have some designation for the public, confident that no one was likely to ques- tion the names, except perhaps our teacher, Professor Jepson, who was unlikely to make an appearance. Similarly, Custis must have felt that in order to uphold his position as a botanist and to earn his pay he must supply names for the plants. He could be fairly confident that no one in the office of the Secretary of War was going to question them, and he was right. It is only now after 150 years that the names are coming to light. It is well for botanical nomenclature that he was chary of assign- ing new Latin names, because many of the plants that he saw really were undescribed species, which is why he could not find their diagnoses in Linnaeus’ work. He actually created only three new botanical names, which are discussed below, and these all prove to be synonyms at the present time. The next list of plants of the Red River Valley was not published until Torrey’s list of 1854, in the Marcy report mentioned previously. A com- parison with Torrey’s list gives some clues to the identity of the plants recorded by Custis. ACKNOWLEDGMENTS This account, first written some twenty years ago and recently revised, has been read by my colleagues Drs. Mason E. Hale, David B. Lellinger, Harold Robinson, Velva Rudd, Lyman B. Smith, and William Stern, to whom I am grateful for many valuable suggestions. I am especially grateful to Professor Joseph Ewan of Tulane University, who, from his 442 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 vast knowledge of botanical history, was able to draw my attention to several references I would otherwise have overlooked; and to Dr. Clair A. Brown, of Louisiana State University, who was able, from, his wide knowledge of the native flora of Louisiana, to suggest the identification of several of the plants mentioned that were doubtful to me. EXTRACTS FROM THE NARRATIVE Some extracts from the narrative in which the vegetation is particularly mentioned are given below, with my commentaries in footnotes, and fol- lowing them are the complete plant lists, again with my commentaries. In the quoted parts the old spellings (and misspellings) of the original are reproduced. “Left Fort Adams, on the Mississippi, April 19, 1806... (p. 5) For two or three miles [from the confluence with the Mississippi] this river [the Red River] is beautifully bordered with willow trees,’* which extend back from its margin nearly half a mile to the second rise in the land or bank, about six feet higher than that which borders the river. The trees of this second bank or plane are various. The Pecan is the most abundant; this is a species of Juglans, which bears the much admired Illinois nut. In its habits the Pecan is so very different from the Juglans alba,'* as not to be mistaken for it, by the most (p. 6) careless, yet the only difference which can be pointed out, is in the leaf: both have pinnate leaves; the Alba has an odd leaflet sessile, the Pecan has odd leaflet petiolate. Juglans petiolata'® would be an appropriate name for this species. The cotton tree, which of late has excited so much attention, is found here; although considered a populus, and by some termed Lom- bardy poplar, there is no other resemblance than the leaf.'® “The other trees are oaks, (Quercus rubra, alba and phellos.) * Per- simmon (Diosphyros Virginiana),8 Hagberry (Prunus padus),1® Syca- * Salix nigra Marsh. and S. interior Rowlee. “Presumably the Mockernut Hickory, Carya tomentosa Nutt., which should per- haps be known as C. alba (L.) Nutt. It is doubtful if Juglans alia L. should be re- jected as a nomen ambiguum, as it was by Rehder, who has been followed more re- cently by Little. *° Juglans pet iolata Custis, a name overlooked in the botanical literature, was P 2(2): 45 —as Juglans pecan Marsh. (1785), J. illinoensis Wangenh. (1787), and J. oliviformis Michx. eae The name currently accepted as correct is Carya illinoensis (Wan- genh.) K. Koch. gal oe Eastern Cottonwood, Populus deltoides Bartr., common and of large size on the river-banks in the area traversed. Quercus phellos L., the distinctive Willow Oak, was surely identified correctly, but Q. alba L. and Q. bubra L. were probably applied to any of the White and Red Oaks soaeay. pote of which had not been distinguished in the time of Custis “Properly Diospyr ” Prunus padus Le is the European Bird Cherry. Presumably the plant referred 1967 | MORTON, RED RIVER EXPEDITION OF 1806 443 more, or the American Plane tree (Plantanus Occidentalis) °° and the Crataegus aria.” After passing three miles up the river, no trees but the willow and Pecan are seen from it, until approaching the Black River, where the Cyprus (Cupressus disticha) ** is noticed. The Pecan is cov- ered with Misleto. “The bank of Black river on the right side is high, the soil very rich and light, and occasionally flooded, and besides all the trees above enu- merated, except the Sycamore and Cotton trees, which disappear shortly after entering the river, it produces the Red Gum (Liquid amber styraci- flua) 2° of a very large size, and the Ironwood (Lyder axylon mite.) ** . (p. 7) The banks of the river are from 18 to 25 feet higher than the eve at this season, and the land generally declines from them, for about one fourth of a mile to Cypress swamsp “At the Avoyell’s settlement, about 35 sites higher than Black river, the Plantanus Occidentalis, and Cotton trees begin to make their appear- ance, with the Cornus Sericea *° and Cypress; the Pecan and Persimmon are most abundant, the first of which generally grows to the height of 100 feet. Nine miles above the settlement is a beautiful bluff, nearly a mile in length, and fifty feet in height, formed of a reddish yellow sandy clay; here is first seen the Quercus Nigra or Black Oak, the Myrica Cerifera or Candle berry bush, and Maple (Acer Pennsylvanicum.) *° “Six miles above this bluff, a stratum of large trees and leaves, thirty feet below the surface of the ground, and covered by 10 or 12 feet of hard marle or clay, was exposed to view in the bank of the river. The Pine (Pinus Sylvestris,) 27 Dogwood (Cornus Florida,) Sassafras (Laurus Sassafras,) 28 Chesnut Oak (Quercus Esculus,) *® Holly (llex aqui- to is the Black Cherry, Prunus serotina Ehrh., which was formerly called hagberry or hack malate with some glandular hairs. Trans- caucasu 28. Sepals blue, variegated with white; bracteoles ca. 2 mm. long, near the flo ed POU EA tas 202. D. elisabethae. Sepals blue, not variegated with white; bracte- oles 10-12 mm. long, just above the middle of the pedicel. (Treated in oe ey 4. D. pyramidatum. to 2 27. Inflorescence strigose, not duce on 2 nflorescence an open panicle, the central axis densely many flowered. Transcaucasus to Tur- key 91. D. crispulum. Tapriensesticl racemose or few ae each branch few flowered. Nepal. .. 97. D. incisum. 25. Upper sepals 16-25 mm. long. 30. The upper sepal ca. 25 mm. long. 31. Lower petal lamina bifid almost to base into lanceolate ines. Tie So aS eee a 08. D. nortonit. 31. Lower petal aan bifid halfway aN ge lobes. Kansu, Szechwan. ......-------- D. pylzowii. The upper sepal 15-20 mm. lon Be 32. Sepals slaty to aie blue 8-20 mm. long, the spur 15-25 mm. long; lower pine with aes lobes. Kansu. . albocoeruleum. 29. \o w o 32. Sepals darker ons Bracteole well below the flower. wer petal lamina oblong, ca. 8 mm Sikang. CG ees 116. BD. smithianum. 34. Lower petal lamina lipicovate Kansu to Tibet. D. pylzowii. 33. Bracteoles very near the Fb lower petal lamina broadly elliptic. JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 35. Flowers 1 to few in a subumbellate group; bail petal emarginate at tip. Western CRA ei ER BREE... 96. D. henryi. : — more numerous. 36. Inflorescence open-paniculate; upper petal entire. Eastern Tibet........... POR RIDR # Aare enler Wine gs LS. e rockit. 36. Inflorescence racemose; upper petal emarginate. Honan. 1 22. D. sd 20. SEP er petals bluish to whitish, not darker than the sepals everal in an inflorescence; sepals largely 3- 15 mm. long. 38. Stem sean even up into the inflorescence. 39. Upper petal notched; lower petal lamina ovate. 40. Sepals intense blue to violet, glabrous, the upper one ee mm. long, the spur 15-18 mm. long. Northern BURA cry bine RL ES cls na 8S LO. 22, ses adel 40. Sep violet-blue, pubescent, the upper sepal 11 m ong, the spur 13-18 mm. long. Western China. ..... rie REEL Be alg Deen ea eA FT ge 118. D. sutchuense. 39. Upper petal entire; lower petal lamina more or less obo- vate. Western China. 41. Bracts mostly 3-parted; upper sepal 12-15 mm. long; lamina of upper petal 9 mm. long. ..............-- eons ets ee Pea cg ie es 103. D. maximowiczit. 41. Bracts entire; upper sepal 20 mm. long; lamina of upper petals 12- 13 mm. long. .... 82. D. autumnale. 38. Stem with some hairs, at least in the inflorescence. 42, Inflorescence with some yellow glandular hairs. 43. Bra - KE ce mm. long; bracteoles 2 mm. long. Af- CU ne a ee 121. D. tenuipes. 43. he . 45 mm. long; bracteoles 6-10 mm. long. 44. ca ee dm. tall; sepals 8-11 mm. long. Son 80. D. aemulans. 44. Stems | to 1.5 dm. tall; sepals 13- ae mm. long. 114. D. pylzowitt. 42. Inflorescence with neither yellow nor glandular hairs. 45. sn petals blue, the blade 12-16 mm. long; bracts w On 46. Blade of lower petal mostly entire, glabrous ex- cept for the basal patch of md carpels densely pubescent. Siberia. ........ . D. cheilanthum. . Blade of lower petal a, more or less cleft, gh almost throughout; carpels subglabrous. eh ee ee haoe 27. D. yuanum. 45. Upper le pale, the blade 8-10 mm. long; bracts mostly not leaflike, but linear to setaceous. 47. Blade of lower petal almost round, 8-10 mm. in diameter, entire to emarginate. Eastern Siberia, through Mongolia to western China. ......---: pee Eee cee aa 95. D. grandiflora. 47. Blade of lower petal deeply cleft, 4-5 mm 48. Pedicels mostly 1-2 cm. long, with a oa a nN 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 501 what recurved tip; bracteoles 2-3 mm. long, placed below the middle of the pedicel. 49, Inflorescence spicate; sepals 13-15 mm, long; spur 15-16 mm. long. Kashmir. (Treated in Group IV). 45. D. roylet. . Inflorescence corymbose. 50 0 mm. long; spur 13 mm long. Punjab. ... 100. D. koelzii. . Sepals 13-17 mm. long; spur 15- 17 mm. long. ie 2 ERS SS abies Lape sd area ae " bhutanicum. . Pedicels 1-4 cm. long; Se 2-4 mm. long, placed near the middle of the pedicel; sepals 12-15 mm. long. Poop be patie Sues Cuba waco ue ow: denudatum. 37. Flowers mostly 1 to a stem; sepals 15~25 mm. stile 51. Plant quite glabrous on stems and axes. 52. Spur decurved or almost hamate, 20-25 mm. long; eae teoles placed below the middle of the pedicel, 4-10 m long. Szechwan. (Treated in Group X)............... Fa Bn eet Ls Rl pea apnea See pa ea, Tena D. potaninit. 52. Spur straight or slightly curved; bracteoles ‘above the middle of the pedicel. 53. Spur 15 mm. long; flowers deep blue. Kumaun. ..... 98. D. kamaonense. 53. Spur 18-21 mm. long; flowers me white. Nepal. . 125. williamsii. > \o unr oS f oo . Plant strigose to hirsute on stems and/or axes of inflorescence. 54. Blade of upper petal emarginate at apex. 55. Spur 10-15 mm. long; plant 7-10 dm. high, short- imite, Tibet. ss cate awe 108. D. nortonit. 55. Spur 20-25 mm. lon 56. Upper netal yellowish at apex; flowers more or 8. D. on on less racemose. Shantung. .... 88. chefoense. 56. Upper petal violet or blue at apex; flowers mostly 1-2 on a stem 57. Sepals sordid- brown or purple-brown, pale on margins; upper seri setose. Tibet. . .D. — 57. Sepals blue; upper anh glabrou 58. Follicles 5; bracteoles at middle of pedicel. Lichiang a unnan. ca RARE ae z. beesianum, 58. Follicles 3; scale below the middle of the pedicel. cee POT he a 54. ape of upper petal entire at apex. Upper sepal 13-18 mm. long. 60. Spur about 20 mm. long; limb of upper petal 8 mm. long. Yunnan. ...... 106. D. mosoynense. Spur 15-19 mm. long; limb of a petal 6-7 mm. long. Tibet, Kansu. ..... 4. D. tsoongit. an 9 502 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 59. Upper sepal 20-25 mm. long. 61. Spur 14 mm. long; lamina of upper petal 8 mm. long; blade of lower petal entire, ovate, pointed. 10 Re RG Ste ee et ; muscosum. 61. Spur 20-30 mm. long; lamina of upper petal 10- 12 mm. lon 62. Lamina of lower petal oblong, rounded at tip, much longer than wide; stem 1.5—5 dm. tall. On Be — 63. The lamina of lower petal not cleft; petiole bases not conspicuously inflated. Lichiang Range, Yunnan. ........... Neen DINAS Bete 84. D. beesianum. 63. Lamina of lower petal bifid; petiole bases very conspicuously inflated. Yun- Bi els Es aro: 90. D. coleopodum. 62. Lamina of lower petal almost round, more or less cleft. 64. Stem to 1.5 dm. tall, subglabrous; a nN pp n L g i iy le ¢) E au pet) 2, ee a fa") i= | wm oO ©, e Lee J oO cr Lon} o Ls | wm oO 1 strigulose; flowers laxly racemose. Bre ee 88. D. chefoense. 80. Delphinium aemulans Nevski in Komaroy, Fl. U.S.S.R. 7: 161, 725... 1937. Fic. 8, A. Stems 8-10 dm. tall, to 8 mm. thick at the base, lightly ribbed, with retrorse white hairs below and violet-tinged, densely glandular pubescent under and in the inflorescence, leafy; leaves glabrous, round-reniform, to 1 dm. in diameter, truncate or slightly cuneate at the base, palmatisect almost to base, the segments pinnatifid into lance-linear lobes 2-3 mm. wide, 5-25 mm. long; petioles scarcely enlarged at base, 10-12 cm. long, or the lowermost longer; raceme branched below, dense, many flowered, the axis densely glandular pubescent; bracts linear, 5-15 mm. long; pedi- cels erect-spreading, 1-2 cm. long, glandular; bracteoles linear-filiform, glandular-pubescent, 6-10 mm. long, 0.5 mm. wide, placed in the upper part of the pedicel; sepals deep blue, ovate, obtuse, glandular pilose; upper sepal ovate, 10 mm. by 6 mm., prolonged into a narrow point; spur 12-13 mm. long, 1.5-1.8 mm. thick, obtusely acuminate and incurved at apex; lateral sepals 11 mm. by 8 mm., ovate, obtuse; lower sepals obovate, 9 mm. by 6 mm., obtuse; petals bluish, the upper laminae somewhat oblique, pale, ca. 8 mm. long, entire, glabrous, the spur 13 mm. long; lower petals strongly oblique, the blade oblong-ovate, 4.5 by 3 mm., bearded and ciliate, bifid more than halfway; claw 4 mm. long: stamens 5.5—6 8; mm. long, the anthers dark; follicles 3, densely glandular pubescent. Type: Dzungaria, in Monrak Mts. near Zaissan, July 22, 1930, Gont- scharov & Borissova (Le); seen. Nevski speaks of this species as near 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 503 D. dyctiocarpum, but differing in its glandular pubescence. I have, how- ever, placed that species in Group IX. 81. Delphinium albocoeruleum Maxim. Bull. Acad. Pétersb. 23: 307. 1877; Fl. Tangut. 1: 23. pl. 4. 1889. Fic. 8, B. Delphinium albocoeruleum var. pumilum Huth, Bot. Jahrb. 20: 409. 1895. Described as 10-15 cm. high and from Tibet. The only Tibet specimen cited is Przewalski in 1884, which I have not seen. But I have seen several other collections from there. Stem erect, 1-12 dm, tall, more or less angulate, often purplish, sparsely retrorse-strigulose, leafy, simple or subcorymbosely branched above; basal petioles to 10 cm. long, somewhat dilated at base, upper much shorter; leaf blades cordate at base, 5-8 cm. broad, strigulose above, more loosely pubescent or even silky beneath, 3—5-parted almost to base, somewhat less so in basal leaves, the subrhombic divisions deeply 2~—3-fid and incised into lance-oblong, subacuminate teeth or lobules; inflorescence branched, the branches few flowered; bracts trifid or multifid, ca. 2 cm. long; pedi- cels to ca. 5 cm. long, gray-strigose; bracteoles near summit of pedicel, trifid to simple and lanceolate, 4-8 mm. long; sepals slate purple to pale blue, sometimes pubescent; upper sepal broadly oblong-ovate, 18-25 mm. by 13-14 mm., obtuse to truncate, the spur straight or somewhat decurved, 15-25 mm. long, 33.5 mm. wide at base; lateral sepals oblong-obovate, 18-25 mm. by ca. 18 mm., rounded-truncate at apex; lower sepals rhombic- ovate, 18 mm. by 13 mm., with an obtuse apiculation; petals very dark, almost glabrous, the upper laminae slightly oblique, 10-13 mm. long, bidentate at apex, the spur 16-19 mm. long; lower petals strongly oblique, the blade slightly bearded, oblong, 8-20 mm. long, cleft into 2 rather lance- ovate lobes, claw 7 mm. long; stamens 6—7 mm. long, ciliate on filaments; anthers dark, linear, 1.5 mm. long; follicles 3, hairy, 13-14 mm. long, the styles an additional 2 mm.; seeds 1.5 mm. long, triquetrous, with hori- zontal rows of fused scales. Kansu, “in regione Tangut prope l’Amdo, Przewalski 1872” TYPE: (LE) seen; isotype (K) seen. DistRIBuTION. At 3000 to 5000 m., eastern Tibet and southwestern ansu. ILLUSTRATION, Maximowicz, Fl. Tangut. 1: pl. 4. 1889. REPRESENTATIVE COLLECTIONS. Tibet: Ludlow & Sherriff 523, 6899; Ludlow, Sherriff & Taylor 6921; Rock 14467, 14459, 14301, 14505; Kingdon Ward 6123, 12244, 12224, Kansu: Farrer & Purdom 242, 243, 686 in part; Rock 13194. 82. Delphinium autumnale Hand.-Mazz. Symb. Sinicae 7: 276. pl. 5, fig. 5. 1931. Fic. 8 Stem slender, 2.5—4.5 dm. tall, rigid or flexuose, striate, glabrous or with a few spreading hairs and short retrorse-strigulose ones below, equally few- leaved above, often branched; leaves orbicular, 3-9 cm. in diameter, trifid 504 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 X35 Ficure 8. Delphinium, Group VI. Petals, the upper spurred, the lower clawed. A. D. aemulans; blade of upper petal 8 mm. long, spur 13 mm., blade of lower petal 4.5 mm. long, claw 4 mm.; drawn from TYPE (LE). B. D. albo- 9 blade of upper petal 12 mm. long, spur 25 mm.; bla long, claw 8 mm.; drawn from Handel-Mazzetti 5495 (ze). D. D. batangense; blade of upper petal 14 mm. long, spur 21 mm.; blade of lower petal 7 mm long, claw 8 mm.; drawn from Soulié 3042 bis (Pp). E. D. beesianum; blade of upper petal 12 mm. long, spur 24 mm.; blade of lower petal 9 mm. long, claw 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 505 to base, the lateral segments bifid or trifid, all parts irregularly and dis- tantly divided subpinnately, the ultimate divisions elongate-linear, spread- ing, 1.5—2.5 mm. wide, acute, thickish, strigulose above and on the revo- lute margins, with few longer hairs beneath; petioles of lower leaves ca. half as long as the stem, the upper shorter, short-sheathing at base, bearded-ciliate, strigulose above; racemes terminal on stems and branches, to 5-flowered, short, subcorymbose; bracts subulate-linear, to 15 mm long, ciliate; pedicels equaling spurs or shorter, 1-2.5 cm. long, sub- glabrous except for some more or less retrorse hairs above; bracteoles near the middle of the pedicels, but often above or below, subulate, ciliate, 4-10 mm. long; flowers blue, horizontal; sepals broadly elliptic, ca. 20 mm. by 10-12 mm., the upper and 2 lower mucronate-thickened at apex, the 2 lateral broadly rounded, more or less erose; spur 25 mm. long, obtuse, 4.5 mm. thick at base, somewhat decurved to the slender tip; upper petals ca. two-thirds the length of the sepals, the blade slightly oblique, pale, 13 mm. by 4 mm., rounded and entire at the apex, glabrous, the spur slender, ca. 25 mm. long; lower petals with an exappendiculate claw 7 mm. long and an oblique limb, roundish or short obcordate, 11 mm. by 8 mm., erose, the surface and margin white-hairy, the base yellow-hairy; stamens 5 mm. long, ciliate; anthers blue, oblong, 1.5 mm. long; follicles 3, hirsute-villous. TYPE: Sikang province, China; on the Hwang-liangdse, between Yen- yuen and Kwapi; growing on limestone at 3600-3900 m., Oct. 5, 1914, Handel-Mazzetti 5495, holotype (probably w), not seen; isotype (e), seen, 83. Delphinium batangense Finet & Gagnep. Bull. Soc. Bot. France 51: 478. 1905 Fic. 8, D. Stem terete, erect or flexuous, 1.5—-5 dm. tall, strigulose, divaricately corymbose at summit, few flowered; leaves several, puberulent, sessile or the lower with petioles to 6 cm. long, crowded on lower stem or scattered; blades rounded in outline, 3—4-parted almost to base into obcuneate parts, these deeply cut or dentate, with lance-oblong laciniae 2-5 mm. wide and obtuse at apex; branches of inflorescence to 12 cm. long; bracts sessile, 6.5 mm.; drawn from Forrest 6694 (us). F. D. beesianum; blade _ upper petal 13 mm. long, spur 18 mm.; blade of lower petal 10 mm. long, claw 7 mm.; drawn calcicola) (us). G. D. lle) blade BM). I. D. upper petal 13 mm. long, spur 16 mm.; blade of lower petal 7 mm. long, claw 7 J mm. lo r 18 mm.; blade of lower petal 8.5 mm. long, claw 5 mm.; drawn from Zimmerman 192 (us). K. D area, blade of ain petal 16 mm long, spur 17 mm.; blade of lower petal 13 m ong, claw 9 mm.; drawn from Thonnikov-Galitaky 3428.(uc). L. D. ert ‘ctaluien: blade of space ‘petal 10 mm. long, spur 12 mm.; blade of lower petal 4 mm. long, claw 6 mm.; drawn from Grossheim s Schiskin 305 (uc). 506 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ca. 2 cm. long, 3-parted, the parts 3-lobed, obtuse; pedicels 2—5 cm. long, strigulose; bracteoles 3-lobed, obtuse, ca. 1 cm. long, situated below the middle of the pedicels; flowers 2-4 on a stem, deep blue, ca. half the length of the pedicels; upper sepal round-ovate, 18-20 mm. long, spur arched-recurved, 20-24 mm. long; other sepals ovate, obtuse, short- strigulose, 20-22 mm. long; lamina of upper petals straight, emarginate, glabrous, 13-14 mm. long, the spur 19-20 mm. long; limb of lower petals obovate-oblong, 2-lobed, bearded in the middle, not ciliate, ca. 7 mm. by 5 mm., the claw 8 mm. long, glabrous, appendaged at base; stamens 6—7 mm. long; follicles 3, densely puberulent, 6 mm. long. Type: Tibet, prov. Batang, Zambala & Yargong, in 1903, Soulié 3042 bis (Pp); seen. It is this number at Paris that was dissected and drawn, but Soulié 3040, 3041, 3042, and 3043 were filed with the type. From the same location, also deposited at Paris, Soulié 3898 and 3899. ILLUSTRATION. Bull. Soc. Bot. France 51: plate 5, B. fig. 7-11. 1905. 84. pees a beesianum W. W. Smith, Notes Bot. Gard. Edinb. 8: 150, TSE S: Fic. 8, E, F. Dain beesianum var. malacotrichum Hand.-Mazz. Acta Horti Gothob. : 60. 1939. Based on Harry Smith 2985, Tsepula, Szechwan. Isotypes sort seen. Delphinium beesianum var. malacotrichum f. radiatifolium Hand.-Mazz. loc. cit. Type number H. Smith 11939 from Taofu, Sikang. Isotypes (A, BM, Delphinium. beesianum var. radiatifolium (Hand.-Mazz.) W. T. Wang, Acta Bot. Sinica 10: 264. 1962. ms beesianum var. latisectum W.T. Wang, loc. cit. 265. Based on . Wang 64923 and 64995 from Yunnan; not seen Deiphinam calcicola W. W. Smith, Notes Bot. Gard. Edinb. 8; 130, 1913. . Forrest 6364 from Lichiang Range, Yunnan (£), seen; isotype (US) Dele wen beesianum f. calcicola (W. W. Smith) W. T. Wang, Acta Bot. Sinica 10: 264. 1962. Rootstock slender, covered with fibrous bases of old leaves; stem 1.5—5 dm. tall, flexuous, branched from below, white- and retrorse-strigulose or more loosely pubescent, the hairs sometimes yellow; lower leaves with purplish slender petioles 5-15 cm. long, strigulose, thin and vaginate at the base; leaf blades 5-fid, roundish in outline, 2-6 cm. in diameter, the segments divided to the middle or below into 2-3 lobes 1-3 cm. long, mostly 1-5 mm. wide, acute, strigulose above, longer hairy beneath; flowers 1—few at the ends of the ranches in corymbose fashion; bracts mostly foliose, the petiole short, somewhat inflated, the blade trifid into lanceolate lobes, 1 cm. or longer, pubescent: pedicels pubescent, 2-8 cm. long, pilose; bracteoles above the middle of the pedicel, trifid to entire, 8-12 mm. long; sepals blue with deeper veining, densely pilose, the upper sepal ovate, 22— 26 mm. long, 10-15 mm. wide, obtuse, the spur 20-25 mm. long, straight or somewhat curved to the blunt tip, 4-5 mm. wide at the base; lateral 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 507 sepals oblong with rounded tip, 23-30 mm. long, 10 mm. wide; lower sepals ovate, about the same size; petals blue, the blade of the upper straight, 12-13 mm, long, entire to emarginate, the spur 18-24 mm. long; blade of lower petals oblong-obovate, 9-10 mm. long, 5 mm. wide, erose to deeply emarginate, claw 6.5—7 mm. long; stamens ca. 8 mm. long, the fila- ments ciliate; anthers dark, 1 mm. long; follicles 5, white-hairy, ca. 14 mm. by 3.5 mm., the style an additional 3 mm.; seeds brown, winged, 1.5 mm. long. Type: Yunnan, limestone cliffs on east flank of Lichiang Range, 4500 m., Sept. 1910, G. Forrest 6694 (E), seen; isotypes (K, P, UC, US) seen. DistTRIBUTION. At about 4000 to 5000 m., in eastern Tibet and western hina. EXAMPLES. Tibet: Rock 14518, 14520, 14523, 14522, 14460, 22267, 13398; Ludlow, Sherriff & Taylor 6379. Yunnan: Kingdon Ward 1025; Forrest 22577, 21007, 20204; Rock 24738, 10817, 7763, 5421, 11692, 10788, 10776, 17248, 22852; Wang 64995, 69525, 70146. Kansu: Rock 13092, 14640. Szechwan: Kingdon Ward 4856; Forrest 22967, 20572; H. Smith 4243, 4326, Sikang: H. Smith 12619, 12515, 11712, 12421. As here conceived this is rather a variable species both as to pubescence and whether the upper petals are entire or not at the apex. 85. Delphinium bhutanicum Munz, sp. nov. Fic. 8, G. Perennial, to ca. 1 m. high, the stem slender, openly branched above, more or less retrorse-strigulose throughout, scattered leafy; lower leaves withered by anthesis, upper blades 4-10 cm. wide, divided almost to their base into cuneate-rhombic segments finely strigulose above, more or less pubescent beneath especially along the veins, trilobed about halfway, then lobulate into linear-lanceolate to lanceolate lobules 5~15 mm. long, 2-5 mm. wide; petioles slightly dilated at the base; uppermost leaves much reduced and serving as bracts; inflorescence open, 2- to few-flowered, subcorymbose, strigulose, the pedicels suberect, 2—5 cm. long; upper bracts lanceolate, ca. 1 cm. long; bracteoles broadly linear, 4-6 mm. long, slightly below the flower; sepals blue, strigulose, the upper broadly obovate, 15-16 mm, by 11-12 mm., subtruncate at apex, spur 15-16 mm. long, ca. 3 mm. wide at base, spreading-decurved; lateral sepals oblong-obovate, 16 mm. by 8 mm., rounded at apex; lower sepals slightly shorter and wider; petals blue, lamina of upper somewhat oblique, 9 mm. by 3 mm., glabrous, entire at the somewhat rounded tip, spur ca. 15 mm. long; lamina of lower petal obovate, 8-9 mm. by 5-7 mm., bifid ca. 3 mm. into 2 rounded-oblong lobes, white-bearded throughout, the claw 5 mm. long; stamens 5—6 mm. long, glabrous; carpels 3, pubescent, 11-12 mm. by 3.5-4 mm., the style an additional 2-2.5 mm.; seeds winged on the angles. Planta perennis, ad ca. 1 m. alta, tenuis, supra ramosa, strigulosa ; laminae foliorum superiorum 4-10 cm. latae, alte divisae, segmentibus cuneato-rhombicis, supra strigulosae, infra pubescentes, laciniis ultimis 5-15 mm, longis, 2-5 mm. latis; inflorescentia laxa, floribus 2—paucis, 508 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 subcorymbosa; pedicelli suberecti, 2-5 cm. longi; bracteae foliosae, su- periores lanceolatae, ca. 1 cm. longae; bracteolae 4-6 mm. longae; sepala azurea, strigulosa; sepalum superiore late obovatum, 15-16 mm. longum; sepala lateralia oblongo-obovata, 16 mm. longa, apice rotundata; sepala inferiora breviora latioraque; petala azurea, laminae superiores glabrae, integrae, 9 mm. longae, calcaria 15 mm. longa; laminae inferiores obovatae, —9 mm. longae, bilobatae; ungues 5 mm. longi; folliculi 3, pubescentes, 11-12 mm. longa; semina alata. Type: Northeastern Bhutan, Me La, Aug. 25, 1934, Ludlow & Sherriff 887; holotype (BM); isotype (£). The proposed species has been collected between 3500 and 4000 m., as follows: open grassy meadows, Me La, Bhutan, Ludlow & Sherriff 1016 (pM); Kurmatrang, Pumthang, Cooper & Bulley 2239 (z); Cho La, S. E. Tibet to Bhutan, Ludlow & Sherriff 873 (pm); Pangotang, Tsampa, Ludlow & Sherriff 19686 (pm). Its inflorescence is corymbose, not spicate as in some species of the region, e.g., D. denudatum and D. roylei. 86. Delphinium caeruleum Jacquem. ex Camb. in Jacquem. Voy. Bot. 4: 7. Ht. 6.. 1843. Fic. 8, H. Delphinium grandiflorum L. var. kunawarensis Brihl in Brith] & King, Ann. Bot. Gard. Calc. 5: 98. 1896. New name for D. caeruleum. ? Delphinium caeruleum var. obtusilobum Brihl ex Huth, Bot. Jahrb. 20: 463. 1895 and D. grandiflora var. obtusiloba Briihl in Brithl & King, Ann. Bot. Gard. Calc. 5: 99. 1896, Delphinium caeruleum vars. corymbosum Brihl ex Huth and tenuicaule Brihl ex Huth, Joc. cit. 464. 1895 and D. grandiflorum var. tenuicaulis Brihl in Brithl & King, loc. cit. 98. Unfortunately I have been unable to borrow authentic material from the Calcutta Herbarium. Mukerjee (Bull. Bot. Surv. India 2: 293-296. 1960) refers vars. corymbosum and tenuicaule to Delphinium candelabrum Ostenfeld. ? Delphinium caeruleum var. magnificum Huth, loc. cit. ba Stem branched from base, 7-40 cm. tall, branches spreading, densely strigulose below, more loosely pubescent above; leaves suborbicular in outline, 1-4 cm. in diameter, 5—7-parted into cuneate-oblong segments, these incised into linear strigulose lobes tending to be revolute on the margin or less divided in the basal leaves; lower petioles to ca. 1 dm. long, not widened at the base, upper cauline leaves remote, shorter petioled; flowers few, in a loose subcorymbose raceme or panicle; bracts divided into linear segments or entire, about 1 cm. long; pedicels 1-5 cm. long; bracteoles fairly remote from flower, pubescent, linear, 5-8 mm. long; sepals pale to intense blue, pubescent to hairy, the upper sepal 12-16 mm. long, ovate, acutish, ca. 9 mm. wide; the spur almost straight, subhori- zontal, 18-25 mm. long, 2.5—3 mm. wide at the base, narrowed to an acute apex; lateral sepals oblong, pubescent along the midrib, 13-16 mm. long, mm. wide; lower sepals about as long, more pointed and more pubes- cent; upper petals bluish, slightly oblique, the blade ca. 7 mm. long, en- 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 509 tire, narrow to rounded at apex, spur 14-15 mm. long; lower petals blue, the blade oblong-obovate, 7 by 5.5 mm., with a round hairy patch near the base, otherwise sparsely bearded and ciliate, entire or cleft for about 2 mm., the claw 4.5 mm. long; stamens 5-6 mm. long, the anthers dark, 1 mm. long; carpels 3—5, pubescent, 8 mm. or more long, styles an additional 2.5-3 mm. Type: Cambessedes says “Ad jugum montis Houkio (vulgo Houkio- Ghauti) ad fines Tartariae sinensis et provinciae Kanaor (alt. ca. 5000 m.).”’ At Paris is a specimen “in herboris altissimis inter Hookis Ghat et Doubling, Kunaweer,” Jacquemont Cat. No. 1896 (Pp) (720); (K) isotype seen. The Paris specimen has three pieces, two are low, a few cm. high, the third is taller, spindly, 3-4 dm. high; all have lobules 1-2 mm. wide. DistRIBUTION. Apparently at 3000 to 5800 m., Nepal, Sikkim, Tibet. I am unable to divide the subacaulescent from the caulescent forms, both growing in the same area. As here recognized, then, a rather variable species in height, pubescence, and outline of lamina of lower petal, which may be entire and rounded to emarginate and rounded. EXAMPLES. Nepal: Duthie 5299; Lall Dhwoj 556. Sikkim: Cave 6; Smith & Cave 2131, 1935: Cooper 760. Tibet: Strachey & Winterbottom 11, 13; Ludlow & Sherriff 10051, 9952, 9751, 779, 11067, 2400, 1977, 1968; Spencer Chapman 35: Kingdon Ward 9997. Bhutan: Ludlow, Sherriff & Hicks 17476. 87. Delphinium candelabrum Ostenfeld in Hedin, Southern Tibet 6(3): 80. 1922. (A list of flowering plants from inner Asia collected by Dr. Sven Hedin [compiled by Ostenfeld & Paulsen].) — Fic. 8, I, Low perennial from a slender rootstock arising from a thickened branched elongate root; stem dark, slender, branched or simple, subglabrous or with some spreading hairs, 5-15 cm. high; leaves subrosulate at the base, the lower petioles 2—7 cm. long, very slender, strigulose, dilated at the base; the blades 1-2 cm. in diameter, rounded, palmatisect to the base, then again into oblong-linear segments 1-2 mm. wide, mucronate at the sub- truncate apices, rather long-pubescent especially beneath; cauline leaves 1 to few, remote, the upper reduced to trifoliate bracts; flowers about 3 to 6, in an open subcorymbose inflorescence; bracts 5-10 cm. long, simply or compound-laciniate; pedicels arcuate-ascending, very slender, some- what spreading-villous, 5-10 cm. long; bracteoles entire and linear, to ca. 10 mm. long, 1 mm. wide, or sometimes trilobed, near the middle of the pedicel; sepals blue-purple, sometimes drying brownish, white- or yellow- villous, the upper sepal round-ovate, 20 mm. by 18 mm., the spur often slightly decurved, ca. 20 mm. long, 4 mm. wide at the base; lateral sepals asymmetrically oblong-elliptic, 20-25 mm. long, 13 mm. wide; lower sepals ovate-elliptic, 22 mm. by 15 mm., rounded-apiculate at the tip; upper petals dark violet, quite straight, the lamina ca. 14 mm. long, with a few stiff hairs, rounded and erose at apex, the spur 17 mm. long; lower petals with blade at right angles to the claw, rounded-obovate, 6 mm. by 4.5 510 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 mm., unequally 2-lobed, short-bearded on lower half, the claw 6.5 mm. long; stamens glabrous, 7 mm. long; anthers dark, 1 mm. long; carpels 3, airy Type: Eastern or Inner Tibet, at 5127 m., Aug. 9, 1901, Hedin (s?). Not seen. ILLUSTRATION. OSTENFELD & PAULSEN, List Fl. Pls. Inner Asia, pl. 2. 1922 SPECIMENS SEEN. Tibet: Ludlow 332 (rE); Pomo Tso, Ludlow & Sherriff 323 E). 88. Delphinium chefoense Franchet, Bull. Soc. Philom. Paris VIII. 5: 170. 1893. Fic. 8, J. Delphinium gilgianum Pilger ex Gilg & Loesener, Bot. Jahrb. 34. Beibl. 75: 33. d on Zimmermann 192, May, 1900, from Tsingtau, Kiau-tschou, Schangtung (Shantung), China; isotypes seen (GH, US). Delphinium grandiflorum L. var. gilgianum (Pilger) Finet & Gagnep. Bull. Soc. Bot. France 51: 483. 1905. Stem 5—8 dm. tall, erect, angled, few leaved, retrorse-strigulose and short pubescent, with some glandular hairs; basal leaves largely withered by anthesis, the petioles 5-15 cm. long, but slightly dilated at the base, the laminae parted to the base with short, linear, obtuse segments; upper cauline leaves short-petioled, the blades rounded in outline, 3-8 cm. broad, minutely pubescent mostly with appressed hairs, trifid, then the 2 lateral lobes deeply cut making the blade essentially 5-fid, the parts cu- neate-obovate, incised 1/2—2/3 their length into lanceolate or linear lobes mostly 2-5 mm. wide; inflorescence laxly racemose, rather few flowered, somtimes with 1 or 2 basal branches, the axis densely spreading-pubescent, some of the hairs glandular; bracts mostly entire, lance-linear, pubescent, 5-7 mm. long; pedicels ascending, slightly divaricate, 2-5 cm. long; brac- teoles 2, linear, remote from the flower, 3-4 mm. long; sepals deep blue, with some white, pubescent, the upper sepal broadly ovate, 20-21 mm. by 15-16 mm., subacute, the spur straight, ca. 20 mm. long, 5 mm. wide at the base, obtuse; lateral sepals elliptic, 20 mm, by 12 mm., rounded-ob- tuse, generally pubescent; lower sepals elliptic-obovate, 20 mm. by 13 mm., obtuse; upper petals pale, somewhat yellowish at apex, the laminae very oblique, ca. 8 mm. by 4 mm., rounded and glabrous at apex or slightly emarginate, the spur ca. 18 mm. long; lower petals bluish, oblique, the lamina rounded-obovate, ca. 9 mm. in diameter, shallowly notched, with yellow beard near the base, otherwise sparsely hairy and ciliate, the claw 5 mm. long; stamens 7-8 mm. long, dilated below, glabrous; anthers dark, oblong, 1.2 mm. long; follicles 3, pubescent; seeds winged on angles. Type: Chefoo, northern China, prov. Shantung, Fauvel (P); seen. ADDITIONAL MATERIAL. Chefoo, Carmichael 823 (pm); E. Chungtien, K. M. Feng 2714 (a). 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 511 89. Delphinium cheilanthum Fisch. ex DC. Syst. 1: 352. 1817; Prodr. 1: 53. 1824. Fic. 8, K. Delphinium pauciflorum Reichb. ex Schlecht. Linnaea 6: 582. 1831, nomen; ex Ledeb. Fl. Rossica 1: 61. 1841. Based on “In terra Tschuktschorum, ad sinum St. Laurentii,” Eschscholtz (LE). Seen. Delphinium davuricum Bess. Cat. pl. h. bot. Cremenec. Suppl. 3: 9. 1814, not Georgi, 1797-1802. Delphinium dahuricum Stev. ex DC. Prodr. 1: 53. 1824; pro synon. Delphinium cheilanthum var. dahuricum (Steven in DC.) Huth, Bot. Jahrb. 20: 466. 1895. Delphinium sylvaticum Turcz. Bull. Soc. Nat. Moscou 15: 74. 1842. Type from Dahuria, Turczaninow, Herb. Fischer (LE); seen. Delphinium parviflorum Turcz. loc. cit. 75. Type, “Ad fl. Barguzin, 1834, Turce, : : ‘ Delphinium cheilanthum var. parviflorum (Turcz.) Huth, Bot. Jahrb. 20: 467. 1895 Delphinium middendorfii Trautv. Florula taimyrensis phaenogama, in Mid- dend., Reise Sibiriens 1(2)[Bot.] 63. 1847. Type “Ad. fl. Taimyr, Aug. 1843, Middendorf,” (LE); seen. Delphinium cheilanthum ssp. middendorfii (Trautv.) P. Brihl in Brihl & King, Ann. Bot. Gard. Calc. 5: 100. 1896. ? Delphinium cheilanthum ssp. schizophyllum Brihl, Joc. cit. Based on a collection by Davidson on China Hill near Naini Tal. Not seen. Delphinium cheilanthum var. subglabrum Ledeb. ex Huth, Bot. Jahrb. 20: 4 Delphinium cheilanthum vars. laxiflorum and chinense Trautv. ex Huth, Joc. S78 Herb. Fischer. One certainly belongs here, the other doubtfully so. Perennial from a woody crown; stems 1 to several, 2-9 dm. tall, simple or branched, subglabrous below to more or less strigulose above, somewhat ridged, 4-8 mm. thick, rather leafy throughout, the lower petioles to about 1.5 dm. long, slightly dilated at the base, the upper much shorter to almost sessile; blades rounded-pentagonal in outline, 3-12 cm. broad, from sub- glabrous to strigulose on upper surface, especially the latter along the veins, somewhat paler and pilose beneath, 5-parted, or the upper 3-parted, with cuneate-oblanceolate or -lanceolate segments, these entire in upper leaves or trifid or 5-fid in lower into lanceolate acuminate lobes 2—5 mm. wide; inflorescence racemose, few- to several-flowered, the bracts leaflike, entire to multifid; pedicels erect or ascending, 1.5-3 cm. long; bracteoles lanceolate, to ca. 6 mm. long, 2 just beneath the flower, often a third farther down; sepals rich deep blue, sometimes white, pubescent, the upper one 12-16 mm. long, ovate, the spur straight or somewhat decurved, 15-20 mm. long, 3-4 mm. wide at base, narrowed to an acute tip; lateral sepals rhombic-ovate, 12-16 mm. long, 8-9 mm. wide, obtuse, pubescent along the mid-vein; lower sepals oblong-ovate, 16-20 mm. long, 8-10 mm. wide, acutish, more generally pubescent; upper petals mostly blue, the blade slightly oblique, glabrous, entire, 13-16 mm. long, the spur slightly 512 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 longer; lower petals mostly blue, the oblong-obovate blade mostly not lobed, ca. 13 mm. by 9 mm., with a small patch of yellow hair near the base; claw broad at summit, 6-7 mm. long; stamens 6~7 mm. long, gla- brous; anthers dark, 1.2 mm. long; follicles 3, densely pubescent, 13—15 (~20) mm. long, the styles 2-3 mm. long; seeds winged on the angles. Type: “Hab. in Daouria, circa Doroninsk. Viassof.” Photograph seen (BH); specimens seen (LE). Rance. Eastern Siberia and northern Mongolia. ILLUSTRATIONS. GMELIN, FI. Sibirica 4: 187. pl. 76. 1799; Bot. Reg. 6: pl. 473. 1820. REPRESENTATIVE SPECIMENS. Augustinowicz in 1875; [konnikov-Galitzky 365, 3335, 3428; Karo 188; Turczaninow in 1831; Bunge in 1883; Maack 673. Huth, Bot. Jahrb. 20: 466-467. 1895, lists var. formosum Hort., and var. chauvieri Lem. (D. hendersonii Henfr., Gard. Mag. Bot. 2: 57. pl. 6. 1850), both of which seem to be entirely horticultural forms. 90. Delphinium coleopodum Hand.-Mazz. Symb. Sinicae 7: 275. 1931. Fic. 9, A, B. Stem simple, 2—4 dm. tall, mostly glabrous, leafy in lower half; petioles 5-15 cm. long, dilated at base for 2-3 cm., blades pentagonal, 2-8 cm. broad, somewhat shorter, strigulose or with longer hairs especially be- neath, deeply pentafid into cuneate-obovate divisions that are incised to about the middle into several lanceolate acuminate lobes 5—15 mm. long, 2-5 mm. wide; inflorescence mostly a simple few-flowered secund raceme 5-20 cm. tall; bracts setaceous and simple to leafy and 3-lobed, the lower with conspicuous inflated petiole base, the upper often 1 cm. long, purplish; pedicels erect, spreading-hairy near the apex, 3-8 cm. long; bracteoles setose, purplish, 7-12 mm. long, placed near the middle of the pedicel; sepals dark blue, rather long-pubescent, the upper sepal ovate, 20-25 mm. long, 12-14 mm. wide, pointed, the spur horizontal or slightly decurved, 18-25 mm. long, ca. 3 mm. wide at the base, rather slowly narrowed to an often blunt tip; lateral sepals 25-27 mm. by 12-13 mm., obtuse, pubescent along the midrib; lower sepals about as large, but more generally pubes- cent; upper petals bluish, the blade 11-12 mm. long, slightly oblique, entire, the spur 20-21 mm. long; lower petals blue, the oblong blade ca. 12 mm. by 8 mm., cleft 2.5 mm., with very long hairs near the base, scat- tered shorter ones elsewhere, claw 6 mm. long; stamens 5—6 mm. long, the filaments somewhat ciliate; anthers dark, 1 mm. long; follicles 3, erect, with long yellow hairs, the body 15-18 mm. by 4—5 mm., the style an additional 3 mm.; seeds ca. 1 mm. long, with a whitish membrane folded into longitudinal wings at the angles. Type: Northwest Yunnan at Likiang, no. 4105, presumably collected by Handel-Mazzetti and presumably at (w). Not seen, but it is well illustrated in his Symbolae Sinicae 7: pl. 5. fig. 6. 1931. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 513 Rance. At about 4000 to 5000 m., in the Likiang or Lichiang Snow Range of Yunnan. REPRESENTATIVE SPECIMENS. Rock 10777, 11676, 7761, 6110, 10633, 11688, 10764, 24945, 11436, 7765, 10669, 6092, 24738; Forrest 6690, 22471, 22483, 22362, 3035, 11363, 16942, 15198, 15173; C. Schneider 3029. This is one of the most easily differentiated species from western China with its remarkably inflated petiole bases. 91. Delphinium crispulum Rupr. Fl. Cauc. 34. 1869. Fic. 8, L. Delphinium speciosum var. linearilobum Trautv. Acta Horti Petrop. 4: 102. 1886. Based on “Transcaucasia rossica districtu Achalzich, ad lacum Cho- sapin, 6000’, G. Radde 212, Aug. 19, 1874” (LE); seen. Delphinium linearilobum (Trautv.) N. Busch in Fl. Cauc. Crit. LT. :.3568. 1902 Delphinium triste var. linearilobum Huth, Bot. Jahrb. 20: 413. 1895. Delphinium linearilobum var. crispulum N. Busch in Fl. Cauc. Crit. IT, 3:'68. 902. Delphinium tomentellum Busch, loc. cit. 70. Type, Baku, district Kuba, Mt. Bazar, Alexsenko 2620 (LE); seen, also photo. (A, K). Delphinium tomentellum vars. macranthum and araraticum N. Busch, loc. cit. 70, 7 Delphinium araraticum (Busch) Busch in Davis, Fl. Turkey 1: 111. 1965. Delphinium tomentellum var. angustibracteatum Busch, Fl. Cauc. Crit. III. 3: 70. 1902. Perennial from a woody rootstock; stem 4-7 dm. tall, angled, ciner- eously finely strigulose throughout, leafy to inflorescence, branched above; leaves rather crowded, petioles dilated at base, the lower 1-2 dm. long, the upper shorter; leaf blades strigose above and beneath, 5-10 cm. long, 5- parted almost to base into narrow segments which are in turn deeply in- cised into lance-linear lobes 5-25 mm. long, prominently veined; upper leaves tripartite; inflorescence an open panicle, the central raceme densely many flowered, the lateral more lax and few flowered; bracts strigose, linear to lanceolate, 1-2 cm. long; pedicels 1—4 cm. long, ascending; bracteoles strigose, 5-7 mm. long, linear, subtending the flower ; sepals light blue, strigose, upper sepal ovate, 13-15 mm. by 5 mm., pointed, the spur almost straight, 14 mm. long, 3 mm. wide at the base, rather narrow at apex; lateral sepals oblong-obovate, 15 mm. by 8 mm., rounded-obtuse ; lower sepals 15 mm. by 5 mm., pointed; petals very dark, the upper with a blunt, almost straight, somewhat bidentate, glabrous lamina ca. 10 mm. long, the spur ca. 12 mm. long; lower petals with the blade oblong, bearded, 4 mm. by 2.5 mm., cleft 1/3 its length, the claw 6 mm. long, the blade at right angles to the claw; stamens 6-7 mm. long, with glabrous filaments; anthers dark, oblong, 1 mm. long; follicles 3, pubescent; seeds squamellate. Type: Caucasus orientalis, Gumbet, Owerin, July 26, 1861 (LE), seen; photograph (A, K). 514 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 FicurE 9. Delphinium, Group VI. Petals, the upper spurred, the lower clawed. A. D. coleopodum; blade of upper petal 12 mm. long, spur 21 mm., blade of lower petal 12 mm. long, claw 6 mm.: n upper petal 10 mm. long, spur 18 mm.; blade of lower petal 10 mm. long, claw 4 mm.; drawn from Nikiforow (micuH). F. D. henryi; blade of uppet petal 1967] MUNZ, ASIAN SPECIES OF DELPHINIUM 515 DistRIBuTION. The plants seem to grow in rocky places between 1000 and 3000 m., from eastern Anatolia to the Caucasus and northwestern Tran. EXAMPLEs. Transcaucasia: prov. Tiflis, Grossheim 305. Turkey: ete Kars, Davis 32562 in part; Turkish Armenia: Sintenis 7107 and 7108; prov. Van, Davis 23327, 23612; McNeill 731; prov. Erzincan, Davis 31665; prov. Conk Davis 32443. Iran: above Zinjanab, Gilliat-Smith 2514. Nevski (Komarov, Fl. U.S.S.R. 7: 143. 1937) cites the synonymy much as I have given it, but Davis (Fl. Turkey 1: 110-111. 1965) main- tains D. linearilobum and D. araraticum as species, as well as D. tomentel- lum and D. crispulum, making 4 distinct species. The differences seem very slight; D. araraticum, for instance, having bractlets 2-3 mm. wide, while in the others they are 1-2 mm. Without more material than was available to me I cannot be sure of the status of the taxa and follow Nevski. 92. oa eps denudatum Wall. Cat. no. 4719. 1830, momen; ex Hook. f. & Thoms. Fl. Indica 49. 1855. BiG. 9, C. ion pauciflorum oe Tilustr. Bot. Himal. 55. 1834, not D. Don, r. Fl. Nepal. 196. 182 Stem usually freely and openly branched throughout, 4-8 dm. tall, sub- glabrous to rather coarsely strigose, especially above, 2-4 mm. thick, terete, sometimes spreading-pubescent above; basal leaves 5—15 cm. broad, roundish in outline, 3—5-parted into rather broad, obovate divisions that are pinnately and divaricately coarsely laciniate into suboblong lobes or teeth 2-3 mm. wide, the petioles to 1.5 dm. long, slender, somewhat dilated at the base; cauline leaves remote, gradually reduced upward and more deeply laciniate into narrower lobes; inflorescence an open panicle of rather few-flowered open racemes; bracts linear, commonly 5-15 mm. long; pedicels ascending, 1-4 cm. long: bracteoles 2, commonly near the middle of the pedicel, 2-4 mm. long; flowers fragrant: sepals blue to violet, the upper sepal ovate, 12-13 mm. by 6-7 mm., acute, pubescent, the spur 14-15 mm. long, 3.5 mm. wide at the base, subacute; lateral sepals oblong-ovate, 13 mm. by 7-8 mm., rounded at apex, pubescent on 10 mm. long, spur 25 mm.; blade of lower petal 11 mm. long, claw 5 mm.; drawn we Henry 6952 (GH oo D. incisum; blade of upper Sie 9 mm _ long, spur awn from Dhwoj 49 (e). H. D. koelzii; blade of upper petal 9 mm. long, al 5 mm. long, claw 5 : Rkamaonense; ag of upper petal 8 mm. long, spur 14 mm.; blade of lower petal 7 mm. lon w 5 mm. i drawn from lade 781 (z). ae, oneal mm.; blade of lower petal 7 ‘ieee ya ei mm ie spur 18 anes reg of lower petal 6 mm. long, claw 6 mm.; drawn from Kotschy 430 (mo). L. D. m petal 11 mm. long, spur 28 mm.; blade of lower petal 8 mm. long, claw 7 mm drawn from pees Smith 3385 (pm). 516 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 midline; lower sepals oblong-ovate, 14-15 mm. by 7 mm., rounded at apex, pubescent; upper petals white, with blue tinge at apex, the blade 8—9 mm. long, glabrous, bidentate, oblique, the spur 13-15 mm. long; lower petals blue or violet, the claw 5 mm., the broadly elliptical blade bearded, ca. 6 mm. long, cleft almost halfway; stamens somewhat violet, glabrous, 5-6 mm. long; anthers bluish, 0.8 mm. long; follicles 3, sparsely strigose to subglabrous, 10-16 mm. by 3-3.5 mm., the styles an additional 2-3 mm.; seed dark, obpyramidal, 1 mm. long, the scales rather long and not reg- ularly arranged in even rows. Type: Baramula, Kashmir, Winterbottom 348, isotype (K); seen. Rance. About 1600 m. to 3300 m., Himalaya of northern India, Kash- mir and West Pakistan. ILLUSTRATIONS. COLLETT, Fl. Simlensis 12. 1902; Basu, Indian Medic- inal Plants, pl. 7. 1918; BLatrer, Beautiful Flowers Kashmir 1: pl. 4. 1927; Coventry, Wild Flowers Kashmir 3: pl. 7. 1930. REPRESENTATIVE COLLECTIONS. Punjab: Koelz 3163, 4963, 1944, 4638, 4841, 8344; R. Stewart 6042, 2026 1/2; Sherriff 7310; Lace 164; Duthie 8-21-1897, 5-25-1897. Pradish: Koelz 21113, 8659; Jacquemont 535, 721; Fleming 445, 394: Sahni 24608; Dudgeon & Kenoyer 107; Strachey & Winterbottom 3; Watt 612; Hooker & Thomson, June 11, 1848. Kashmir: Pinfold 164; Ludlow & Sherriff 7671, 8106, 9106; Dickason 366; Stewart 17468, 9268, 17534, 12147, 93. Delphinium dissectum Huth, Bot. Jahrb. 20: 403. 1895. Fic. 9, D. Stem erect, simple, 2.5-5 dm. tall, strigulose throughout, scattered-leafy to the middle, the lower petioles 5-12 cm. long, dilated at base, strigulose, the upper gradually shorter; leaf blades rounded-pentagonal in outline, the lower 5-10 cm. broad, strigulose on both surfaces, parted almost to the base into narrow-cuneate segments, these much broadened above, deeply pinnately incised and laciniate into linear segments mostly 1.5-2.5 mm. wide; raceme elongate, peduncled, 1-3 dm. long, soft-pubescent, rather laxly flowered; lower bracts laciniate, upper entire, linear-lanceolate, ca. 1 cm. long; pedicels divaricate, mostly 1-3 cm. long; bracteoles lanceolate, 4-6 mm. long, subtending the flower; sepals dilute blue, with slender hairs 1 mm. long, both externally and internally; upper sepal ovate, 16-18 mm. by 10 mm., acutish, the spur stout, decurved, 16-17 mm. long, 4 mm. wide at base, blunt at tip; lateral sepals oblong-elliptic, 18-20 mm. by 8-10 mm.; lower lance-ovate, 20-22 mm. long, 9 mm. wide, subobtuse; petals black, the upper setose at entire tip, the spur 13 mm. long; lower petals oblique, the blade 6 mm. by 3 mm., the 2 lobes gaping, ciliate, 2 mm. long, bearded, the claw 6.5 mm. long; stamens 6—7 mm. long, glab- rous; anthers dark, 1 mm. long; follicles 3, more or less pubescent. Type: Mongolia between Tschargachantu and Geltgentei, Kirilov, 1841 (LE); seen. COLLECTIONS SEEN. All from Mongolia: /konnikov-Galitzky 785; Pavlov 336; Krascheninnikov 87. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM S17 94. Delphinium dolichocentroides W. T. Wang, Acta Bot. Sinica 10: 159-160. 1962. Stems 3—5 dm. high, 2.5 mm. thick at base, white-hirsute below, gla- brous above, remotely 1—2-foliate; basal leaves 2—3, long petioled; blades broadly pentagonal, 2.5-5.5 cm. long, 4.5—8 cm. wide, cordate, 3-parted to 1.5—3.5 cm. above the base, the middle part broadly rhombic, cuneate at base, 3-parted to below the middle, the middle lobe broadly or nar- rowly rhombic, trifid to middle, the lobules linear to lanceolate, 3-5 mm. wide, rarely 1-toothed, the lateral lobes broadly obliquely rhombic, smaller, unequally 2—3-fid; lateral parts broadly obliquely flabellate, un- equally 2-parted to base, the upper parts like the middle, the lower broadly rhombic, unequally bifid, densely strigulose beneath, loosely hirsute above; petioles 14-17 cm. long, loosely hirsute, narrowly vaginate at base; upper leaves with few long laciniate entire segments 3-4 cm. long, 5 m. wide; racemes 5-24 cm. long, glabrous, very open, 3—6-flowered; pedicels 2-8 cm. long, bibracteolate near or above the middle; bracteoles linear, 6-14 mm. long, 0.4-1 mm. wide, loosely puberulous; sepals blue- purple, elliptic-obovate, 14-16 mm. long, appressed puberulent, short- corniculate, these tips 1-1.5 mm. long; spur subulate, 27-28 mm. long, 4.2 mm. wide at base, slightly falcate decurved; upper petals flavescent, glabrous, emarginate; lower petals blue, the limb broad-elliptic, bilobu- late, long-ciliate, yellow bearded below the middle, about as long as its exappendiculate claw; stamens glabrous; carpels 3, densely pilose. Type: Szechwan, Muli, Aug. 19, 1937, T. T. Yi 14008 (Px), not seen; photograph (£), seen. I have had no material for study, Delphinium dolichocentroides was said by the author to differ from D. dolichocentrum in its simple hirsute stems, few-flowered racemes, emarginate upper petals; to have the habit of D. yuanum, but with the spurs longer than the sepals and decurved; to be near D. taliense, but with leaves 3-parted to the base and the parts more deeply divided. 95. Delphinium grandiflorum L. Sp. Pl. 1: 531. 1753. Fic. 9, E. Delphinium grandiflorum var. linnaei Brithl & King, Ann. Bot. Gard. Calc. 5: Delphinium chinense Fisch. Cat. Hort. Gorenk. 1808; Lodd. Bot. Cab. 1: pi. V1 tBh: : Delphinium grandiflorum var. chinense Fisch. ex DC. Prodr. 1: 53. 1824. Delphinium grandiflorum vars. bhotanica, paradoxa, setchwanensis and ¢san- gensis Brihl in Briihl & King, Ann. Bot. Gard. Calc. 5: 98, 99. 1896. I have been unable to borrow authentic material. Delphinium sinense Fisch. ex Link, Enum. Hort. Berol. 2: 80. 1822. Delphinium pubiflorum Turcz. FI. Baic.-Dahur. 73. 1842, nomen. j iforum Turcz. loc. cit. I have had on loan the following material from Leningrad: “In arenosis ad Baicalem prope Turxam, 1829, Turcz.” as “pubiflorum mihi” 2 sheets; and “In arenosis ad 518 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 48 ripam Baicalis, prope Turxam, 1823, Turcz. mihi.” I take these to be authentic material. Delphinium grandiflorum var. acaule Turcz. ex Huth, Bot. Jahrb. 20: 461. 1895. A sheet “Ad acidulas Pogromezenses 1829, Turcz.” as var. acaule (LE) seen. The stems were 1-5 dm. tall, very fine-leaved. Delphinium grandiflorum var. gmelinii Rchb. Ill. Aconit. atque Delph. pl. 12. 1823-27. A sheet from (LE) seen, from N. T. Siberia. Delphinium grandiflorum L. var. pallidum Huth, Bot. Jahrb. 20: 461. 1895, described as having white flowers; may belong here. Delphinium grandiflorum var. pumilum Huth, loc. cit. as being about 2 dm. igh; may belong here. Delphinium grandiflorum var. tigridium Kitagawa, Rep. 1st Sci. Exp. Manch., Sect. 4, IV: 17, 83. pl. 2. 1936. Reduced to synonymy by W. T. Wang, Acta Bot. Sinica 10: 271. 1962. Delphinium grandiflorum var. majus W. T. Wang, Joc. cit. 273. Said to have unusually large leaves. Based on Feng 9286 from Yunnan, which I have not seen, Delphinium grandiflorum var. leiocarpum W.T. Wang, loc. cit. 274. Based on C. Fu 7900 (which I have not seen), with glabrous carpels. Delphinium grandiflorum var. glandulosum W.T. Wang, loc. cit. 272. 1962. The type by S. T. Wang, May 2, 1928, from Shensi is not available. Delphinium virgatum Jacq. ex Sprengel, Syst. 2: 617. 1825; not D. virgatum Poir. Encycl. Suppl. 2: 458. 1811. Delphinium fangshanense W. T. Wang, Acta Bot. Sinica 10: 269, 270. 1962. Based on T. N. Liou 8258 which has not been available to me. Stem or stems mostly slender, 2-8 dm. tall (although some forms are semiprostrate), simple or branched above, mostly closely strigulose, more densely so in the inflorescence, leafy, the cauline leaves not crowded, grad- ually reduced upward; petioles scarcely dilated at the base, slender, the lower to 1.5 dm. long, the uppermost nearly or quite lacking; leaf blades palmately multipartite, 3-10 cm. wide, the segments divided into parts that are in turn laciniate into linear lobes mostly 1-2 mm. wide and up to 2 cm. or more long; inflorescence simple or few branched, the flowers from 1-2 on a branch to several in rather a dense raceme; lower bracts laciniate, 1-3 cm. long, the upper entire, linear, to 1 cm. long; pedicels erect or ascending, strigulose, 2-4 cm. long; bracteoles linear, opposite; sepals ca. 18 mm. long, 3 mm. wide at the base, gradually tapered; upper petals pale with slight bluish tinge, the limb almost straight, 8-10 mm. long, subentire or emarginate at the rounded apex, glabrous, the spur 16-18 mm. long; lower petals violet blue, the blade yellow bearded at the center, ciliate or not on margins, almost round in shape, 8-10 mm. in diameter, entire, erose or emarginate, the claw 4—4.5 mm. long; stamens 5—6 mm. long, glabrous or nearly so; anthers dark, 1 mm. long; follicles 3, straight or recurved, hairy or puberulent for most part, 10-16 mm. long, 3 mm. wide; styles an additional 2.5 mm.; seeds 1 mm. long, with a loose membranous envelope folded into wings. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 519 Type: “Habitat in Sibiria”; photograph seen of specimen (LINN). RANcE. Central to western Siberia, Inner and Outer Mongolia, northern China (mostly Hopei, Shansi, Kansu). ILLustRATIONS. GMELIN, FI. Sibirica 4: pl. 78. 1769; Bot. Mag. pl. 1686. 1815. EXAMPLES SEEN. Siberia: Sokolow 97; Maximowicz in 1859; Stukow 1358; Tugarinowa, July 13, 1908; Ehnberg, July 19, 1900; Digby, 1914; Nikiforow, July 10, 1917; Herb. Ledebour 809; Herb. Pallas, Aug. 1877. Mongolia: Karo 207; Prescott in 1830; Ohwi 10662; Hsia 3159; Kozlov 240; Ikonnikov-Galitzky 210; 314a; Litvinov 490, 989; Dorsett & Dorsett 3367. Tibet: Ludlow & Sher- riff 9814, 9073, 11114; Rock 13361, Shansi: Licent 1305; K. Ling 9143; Han Ngan 92: H. Smith 6526, 8150, 7139. Hopei: Dorsett & Morse 7051; Liou 1405; Read 746; Liu 1433, 1968; Breitschneider 1817; H. Smith 173, 468, 1163; T. F. King 12689; C. F. Li 10707; F. N. Meyer 1057. There is great variation in flower size and color, and pubescence, as well as in foliage, the ultimate laciniae of the leaves tending to be wider in Chinese plants than in Siberian and Mongolian. However, a careful study of the matter did not make recognition of the var. chinense seem worth while. The species has been widely cultivated and I have seen in herbaria specimens that I would consider D. grandiflorum bearing the labels tat- sienense, sinense album, chinense, chinense azureum, grandiflorum, sibiri- cum, grandiflorum ‘Blue Mirror,’ ‘Blue Butterfly’ and ‘Farquahar’s "i ? Victory. 96. Delphinium henryi Franchet, Bull. Soc. Philom. Paris VIII. 5: 177. 1893. Fic. 9, F Delphinium henryi {. concolor W. T. Wang, Acta Sinica Bot. 10: 79. 1962. Perennial from a very slender rhizome; stem 1-3.5 dm. high, retrorse- strigulose and with long hairs near the nodes, leafy throughout, the leaves only gradually reduced up the stem; petioles somewhat dilated at their base, more or less pubescent, the lower to 10 cm. long, the upper almost lacking; laminae broader than long, somewhat pentagonal in outline, more or less pubescent with rather long appressed hairs, 2-5 cm. broad, 3- to 5-parted almost to base, the segments cuneate at base, deeply cut into few lance-linear lobes 2-3 mm. wide and acute; flowers solitary to few in a subumbellate cluster; bracts largely foliaceous, divided; pedicels retrorse- pubescent, 2-3 cm. long; bracteoles narrow-elliptic, 5-10 mm. long, close to the flower; sepals blue, pubescent; upper sepal 16-20 mm. long, 10-12 mm. wide, ovate, pointed, the spur 25 or more mm. long; lateral sepals oblong-ovate, obtuse, 15-20 mm. long, 8-9 mm. wide, pubescent along the middle: lower sepals 20-22 mm, long, 8-10 mm. wide, acutish-cornic- ulate; petals brownish, with dark dots, the upper petals with a lamina 10 mm. long, very slightly oblique, bidentate, the spur ca. 25 mm. long; lower petals with claw 5 mm. long, lamina oblong-obovate, 11 mm. by 6.5 mm., cleft 4 mm., bearded somewhat at base and ciliate; stamens ca. 7 mm. 520 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 long, ciliate; anthers dark, oblong, 1.2 mm. long; follicles 3, silky, suberect, divergent at tips. ee Central China, Hupeh, Dr. Henry 6952 (P); seen. Isotypes (BM, NY), seen. The collection number was published as 6932, but an ect cited above are all 6952. OTHER COLLECTIONS. Szechwan: Henry 6952A; K. H. Yang 65383. Chensi: Licent 2760; Giraldi in 1897. Two sheets from Szechwan labeled sutchuense fide Wang, do not fit the description. These are nearer to D. henryi but with a more open habit: Chu 3807 (£) and 3127 (£). 97. Delphinium incisum Wall. Cat. 4717. 1828, nomen. Fic. 9, G. Delphinium ranunculifolium Wall. var. incisum Hook. f. & Thoms. Fl. Indica £:51,..3855: Delphinium elatum L. var. incisum Hook. f. & Thoms. FI. Brit. India 1: 26. 1872. Delphinium englerianum Huth var. incisum Huth, Bot. Jahrb. 20: 418. 1895. To about 1 m. tall, the stem loosely retrorse-pilose, slender, more or less angled, remotely leafy, becoming more retrorse-strigose above, not gland- ular; cauline leaves pentagonal, the lamina 5-10 cm. wide, somewhat strigose on both surfaces and paler beneath, 5-fid to near the base, the segments somewhat rhombic, 3-lobed to below the middle, then incised into rather few lance-oblong lobules to ca. 1 cm. long and 1.5—3 mm. wide, mucronate; petioles not strongly vaginate, to ca. 1 dm. long, the upper- most leaves almost sessile; inflorescence racemose or few branched, each branch rather few flowered, fairly compact; bracts leafy and dissected or the upper entire, linear-oblong, to ca. 8 mm. long; pedicels erect-ascending, strigose, mostly 1-2 cm. long; bracteoles linear-oblong, strigose, 4-6 mm. long; sepals bluish, strigose, the upper sepal ovate, 14-15 mm. long, 11-12 mm. wide, rounded at apex, the spur spreading decurved, ca. 17 mm. long, 2.5-3 mm. wide at base, blunt; lateral sepals rounded-obovate, 13 mm. long, 12 mm. wide; lower sepals 13 mm. by 7 mm., subacute; petals black, the upper laminae slightly oblique, 9 mm. by 2 mm., entire, gla- brous, the spurs 16 mm. long; lower petals with oblong laminae; 6 mm. by 3 mm., with 2 short unequal oblong ciliate lobes and some bearding from the base out, the claw 6 mm. long; stamens 6—7 mm. long, glabrous; anthers dark, 1 mm. long, oblong; carpels 3, densely pubescent. Type: Kamaon (Kumaun), Dr. Wallich, List No. 4717, isotypes (BM, E; K, NY, P); seen. reo SEEN. Nepal: Nilo Bi Rh no. 21 (£); Lall Dhwoj 49 (E), 267 E). Near D. himalayae, but inflorescence more branched; flowers smaller; upper petals glabrous and entire. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 521 98. Delphinium kamaonense Huth, Bull. Herb. Boiss. 1: 333. 1893. Spelled camaonense in Huth, Bot. Jahrb. 20: 460. 1895. Fic. 9, I, J. Delphinium grandifiorum L. var. kumaonense E. Huth ex Briihl & King, Ann. Bot. Gard. Calc. 5: 98. 1896. Perennial from a thick subvertical root; stem branched, 1—5 dm. tall, subglabrous or with some spreading hairs; petioles dilated at the base, the lower 3-8 cm. long; lower leaves pubescent especially beneath, strig- ulose above, the blades 2—4 cm. across, 5-parted, then 5-fid, the segments cuneate or narrower, laciniate into oblong lobes 2-3 mm. wide, or shal- lowly lobed with broader divisions; upper leaves with linear lobes 1-2 cm. long, 1-1.5 mm. wide; inflorescence open, divaricately branched; lower bracts tripartite with linear lobes; upper bracts entire, linear, to 1 cm. long; pedicels much longer than flowers, bibracteolate above the middle; bracteoles linear, 3-5 mm. long; flowers deep blue; sepals slightly pilose, to 20 mm. long, with a dark brown subapical spot; upper sepal round-ovate, obtuse, 17 mm. by 12 mm., the spur straight or slightly curved, 12-15 mm. long, 3.5 mm. wide at the base; lateral sepals elliptic- oblong, rounded at apex, 13-14 mm. long, 8~9 mm. wide, lower sepals 14 mm. by 7 mm., oblong-ovate; upper petals dilute brown, the laminae somewhat oblique, 8 mm. long, 3 mm. wide, entire, glabrous, the spurs 14 mm. long; lower petals blue, bifid, yellow bearded, obovate rounded, 7 mm. long, 8 mm. wide, the claw 5 mm. long; stamens 5~6 mm. long, glabrous; anthers dark, 1 mm. long; follicles densely pilose. Type: N. W. India, Kamaon (Kumaun), Nipshany Valley in Darma, 4300-4600 m., Duthie in 1884 (probably c), not seen; isotypes (E, K), seen; they are numbered 2675, Aug. 31, 1884. Rance. Growing at perhaps 3000 to 4500 m., Kumaun, Nepal. REPRESENTATIVE COLLECTIONS. Kumaun: Duthie 5299; Ludlow 781, 775. Nepal: F. M. Bailey in 1936; Arnold 12; Stainton, Sykes & Williams 1871, 7216, 2026, 2233, 2200, 7351, 1930, 8035, 7742, 8142; Tyson 93, Polunin 1576; Lowndes 1324, 1229. 99. Delphinium kantzeense W. T. Wang, Acta Bot. Sinica 10: 161. 1962. Stem ca. 1 m. tall, 5 mm. thick near base, loosely spreading-hirtellous, branched, equally foliate; median leaves with broadly pentagonal blade, ca. 5.5 cm. long, 7.5 cm. wide, cordate at base, 3-parted to 3-4 mm. above the base, the primary middle parts broadly rhombic, narrow-cuneate at base, 3-parted to below the middle, secondary middle parts narrow-rhombic, trifid to middle into deltoid or lanceolate lobes 2.5-3.5 mm. wide, the secondary laterals lance-ovate, one-lobulate or entire, the primary lateral parts broadly obliquely obovate, 2-parted to 8-9 mm. above the base, upper secondary parts obliquely rhombic, 3-fid, the lower secondary smaller, unequally bifid, loosely strigulose beneath, loosely hirtellous on veins above; petioles slightly longer than blades, not vaginate; racemes 522 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 terminal, the central ca. 14-flowered, lateral 3-—5-flowered, lax; lower bracts trifid, others linear, 5-12 mm. long; pedicels 2—5 cm. long, retrorse- strigulose and yellow-glandular-puberulent, divergent, bracteolate above; bracteoles to 2—3 mm. from base of flowers, linear, 4-9 mm. long, 0.3 mm. wide; sepals ashy-blue, oblong-ovate or narrow-elliptic, 10-11 mm. long, puberulent; spur 16-18 mm. long, subulate, 2—2.5 mm. wide at base, de- curved to a right angle; upper petals ashy-blue, glabrous, entire; lower petals ashy-blue, the laminae ovate or elliptic-ovate, rounded at apex, entire to emarginate, ciliate, yellow bearded at base, claw slightly longer, appendiculate; stamens glabrous; carpels 3, glabrous. Type: Kantze, July 8, 1951, Y. W. Tsui 4303 (presumably PE); not seen. I have seen no material definitely referable here. The species was de- scribed as near D. yunnanense, but the pedicels spreading puberulent; spur strongly decurved; carpels glabrous. 100. Delphinium koelzii Munz, sp. nov. Fic. 9, H. Stem slender, 1.5—3 mm. thick, 8—10 dm. tall, strigulose, simple or with a few short branches above; lower leaves mostly withered by anthesis, the blades commonly 2-8 cm. long, less broad, multisect into divisions that are pinnately and divaricately laciniate into lobes 1-2 mm. wide, sub- glabrous above, somewhat strigulose beneath, but not strongly bicolored, the petioles to 5 cm. long, with dilated base; middle cauline leaves several, short-petioled to subsessile; racemes lax, the terminal to 8-flowered, the lateral with fewer flowers; axis, pedicels, etc., densely and finely strigulose; bracts linear, subsetaceous, to 5 mm. long; pedicels divergent with more or less recurved tip, mostly 1-2 cm. long; bracteoles 2-3 mm. long, below the middle of the pedicel; sepals wine-colored to blue, somewhat strigulose, the upper sepal ovate, 10 mm. by 5 mm., obtuse, the spur 13 mm. by 3 mm., straight, rather blunt; lateral sepals 10 mm. by 6 mm., rounded- oblong; lower sepals rounded, 10-11 mm. by 6 mm.; upper petals pale, oblique, the laminae 8-9 mm. long, shallowly toothed at apex, the spur ca. 13 mm. long; lower petals with laminae 5 mm. long, bearded, subovate, divided for ca. 3 mm., the claw 5 mm. long; stamens 5-6 mm. long, gla- brous; anthers dark, 0.8 mm. long; follicles 3, densely strigulose. Caulis tenuis, 8-10 dm. altus, strigulosus; folia infima marcescentia, laminis 2-8 cm. longis, multifidis, cum laciniis ultimis 1-2 mm. latis; inflorescentia subcorymbosa, racemis laxis; bracteae lineares, 3-5 mm. longae; pedicelli divergentes, 1-2 cm. longi; bracteolae 3-5 mm. longae; sepala vinacea aut azurea, strigulosa, 10-13 mm. longa, calcare 13 mm. longo; laminae petalorum superiorum obliquae, 9 mm. longae, bidentatae; laminae petalorum inferiorum 5 mm. longae, lobatae; stamina glabra; carpella 3, dense strigulosa. Type: Punjab, at Kulu, Kangra, at 1600 m., June, 1933, W. Koelz 4922 (GH); isotypes (F, MICH, NY, US) seen. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 523 OTHER Punjab COLLECTIONS are Kulu, Kangra, Koelz 4921 (micu, us) and 4757 (MICH). The species seems near D. roylei, but has a more corymbose inflorescence and smaller flowers. 101. Delphinium lanigerum Boiss. & Hohen. in Boiss. Diagn. 8: 10. 1849, Fic. 9, K. Perennial from a thickened root, the stem 1-2 m. tall, terete, glabrous, sparsely branched above; lower leaves to 2 dm. across, palmately parted almost to base, the parts divided into broadly lanceolate, acute, elongate lobes 8-10 mm. wide, the petioles to ca. 2 dm. long, dilate-vaginate; upper cauline leaves remote, divided into much narrower lobes; racemes short, lax, few flowered, glabrous; bracts linear, 5-10 mm. long; pedicels slender, spreading, 2—3.5 cm. long; bracteoles near the middle of the pedicels, setaceous, 4-5 mm. long; flowers intense blue; sepals glabrous except for the ciliate apex; upper sepal 12-15 mm. long, 5 mm. wide, ovate, acute, the spur 15-18 mm. long, 4 mm. wide at base, horizontal, almost straight, gradually narrowed to the tip; lateral sepals 12 mm. by 4 mm., oblong, somewhat rounded at apex; lower sepals 13 mm. by 5 mm., elliptic-ovate, acute; upper petals obliquely truncate, the blade ca. 11 mm. long, divided at apex for ca. 1.5 mm., glabrous, the spur 22 mm. long; lower petals 12 mm. long, the claw 6 mm., the blade 6-7 mm. long, divided ca. 4 mm., heavily bearded; stamens 6~7 mm. long, glabrous; anthers dark, almost 1 mm. long; follicles 3, pubescent. Type: Mt. Elbrus, northern Iran, Kotschy 430, July 1, 1843; isotypes (BM, MO, P), seen. Said to be frequent in Dudera valley, Mt. Elbrus. A collection from a wet meadow, Dorud, Luristan, Iran, June 9, 1941, Koelz 18164 belongs here. 102. Delphinium malacophyllum Hand.-Mazz. Acta Horti Gothob. 1a) S24 Be. Oi 1s oe- Fic. 9, L. Stem 1-4 dm. tall, simple or corymbosely branched, the entire plant covered with soft hyaline spreading hairs or strigulose, with longer hairs at the nodes, equally and scattered leafy, 1-7-flowered; leaves pentagonal, 3-6 cm. in diameter, trifid to the base, the outer segments often unequally 2-3-parted, all segments trilobed to the middle or less, the ultimate divi- sions lanceolate, entire, 1-3 mm. wide, acutish, the middle elongate, stri- gose above, scattered white-hairy beneath; petioles slender, the lower to 10 cm. long, the main cauline evaginate, ca. as long as the blades, the upper- most almost lacking; lower bracts foliose, trifid, ca. 1 cm. long, the upper linear, entire; pedicels 2.5-7 cm. long, ascending; bracteoles small, fili- form, in the upper third or fifth of the pedicel; flowers nodding, dark violet; sepals puberulent to hairy; the upper sepal round-ovate, 13-20 mm. long, 10-12 mm. wide, apiculate, the spur slender, straight, 25-35 524 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 mm. long, 3 mm. wide at the base; lateral sepals elliptic, 16-20 mm. long, 10-12 mm. wide, rounded at apex, strigulose on midrib and margin; lower sepals ca. the same, more apiculate; upper petals with a broadly linear lamina 9-10 mm. long, glabrous, emarginate; lower petals dark, the nar- row claw 6 mm. long, the deflexed lamina 7 mm. long, bifid to the middle or more, the disc yellow bearded; stamens 5—6 mm. long, glabrous or ciliate; follicles densely white-silky. Type: Szechwan, Harry Smith 3385 from Hsioeh-shan, 4300 m., Aug. 11, 1922 (w), not seen; isotypes (BM, E), seen Differs from D. pylzowii in softer pubescence and darker petals. 103. Delphinium maximowiczii Franchet, Bull. Soc. ople Paris VIII. 5: 164. 1893. . 10, A. Sa artigo grandiflorum L. var. latisecta Maxim. Plant. Chin. Potanin. in cta Horti Petrop. 11: 30. 1890. Stem subglabrous below to minutely strigulose, more densely so above, slender, much branched, 3-8 dm. tall, scattered leafy; leaves tripartite to near the base, the segments again parted, so that the leaf outline is pen- tagonal, commonly 2—8 cm. in diameter, the 5 parts deeply incised into lance-oblong sharply pointed teeth or lobes 2—3.5 mm. wide, green and somewhat strigulose above, paler and strigulose beneath especially along the veins; petioles slender, somewhat strigose, somewhat dilated at the base, the lower to 20 cm. long, the upper 1-3 cm.; flowers few, in a corym- bose cluster at end of each branch, sky-blue; bracts mostly 3-parted; pedicels slender, divaricate, 2-4 cm. long, strigulose, the bracteoles lance- linear, 3-6 mm. long, near the middle of the pedicel; sepals strigulose, somewhat lighter in color along the middle line; upper sepal ovate, 12-15 mm. long, obtuse, the spur 16-25 mm. long, 2.5—3 mm. wide at base, nar- rowed gradually to a pointed apex; lateral sepals oblong, 17-19 mm., by 6-7 mm., rounded at apex, the pubescence along the midvein; lower sepals more rhombic, 15-18 mm. long, 7-8 mm. wide, obtuse; upper petals yellowish with blue tinge, slightly oblique, the lamina 9 mm. long, gla- brous, somewhat rounded at the apex; lower petals with an oblong-obovate blade 7 mm. by 5 mm., yellow bearded in a patch near the base, entire to lobed, the claw 6 mm. long; stamens 5—6 mm. long, glabrous; anthers 1 mm. long; follicles 3, puberulent, recurved-divergent; seeds squamate. Type: E. Kansu, Fyn-Shan-Ling Mts., Potanin in 1885 (some labels, such as on the sheet at Paris have 1881), (Pp), seen; isotype (K), seen. RANGE. Apparent at about 2500 to 3500 m., western China and adja- cent Tibet SEEN. Tibet: ina Rock 23243, Szechwan: Rock 18327. Yunnan: Rock 24272; Yu 10504 There is considerable variation in the lamina of the lower petal in the above cited specimens, although it has the same general shape throughout. Yii 10504 has the lamina entire, with some hair outside the basal patch. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 525 Rock 18327 and 24272 have entire lamina without outside hair. Rock 23243 and Potanin have the lamina partly cleft and with short ciliation as well as the basal patch. 104, Delphinium mollipilum W. T. Wang, Acta Bot. Sinica 10: 268. 1962. Stem about 3.5 dm. high, 1-2 mm. thick below, loosely spreading- pubescent, simple, equally foliose; middle cauline leaves with pentagonal lamina, 3.3 cm. long, 6 cm. wide, cordate, trisect, the middle segment broadly rhombic, trifid to below the middle, the middle lobes linear, entire, 3.5 mm. wide, the lateral at an obtuse angle, linear, 1—2-lobulate, the lobules obliquely ovate, subglabrous below, loosely pilose above; petioles equaling the laminae, subterete, indistinctly vaginate, spreading-pilose; upper leaves gradually reduced; inflorescence about 2-flowered; lower bracts trifid into linear lobes, the upper entire, linear; pedicels 3-3.6 cm. long, divergent-erect, with spreading white hairs; bracteoles near the middle of the pedicels, 4.5—6 mm. long, loosely puberulent; sepals purplish blue, oblong to oblong-obovate, 16-17 mm. long, white-puberulent; spur ca. 25 mm. long, slender-subulate, ca. 2 mm. thick at base, straight; upper petals glabrous, yellowish, emarginate; lower petals blue, the limb round- ovate, emarginate, yellowish bearded at base, longer than the appendiculate claw; stamens glabrous; carpels 3, loosely puberulent. Type: Kansu, Sept. 24, 1956, Exped. Hunangho 8928 (PE), not seen; photo. (£), seen. A specimen from Kansu at ca. 3000 m., Aug., 1914, Farrer & Purdom 225 (£), in part, seems to belong here. The species is supposed to be close to D. caeruleum, but with pedicels spreading pubescent and carpels 3. My interpretation of D. caeruleum might almost include this, but that species is restricted to an area farther west. 105. Delphinium monanthum Hand.-Mazz. Acta Horti Gothob. 13: 50. 1939. Fic. 10, B. Delphinium candelabrum (Ostenfeld in Hedin) var. monanthum (Hand.- Mazz.) W. T. Wang, Acta Bot. Sinica 10: 78. 1962. The stem low, to ca. 1.5 dm. tall, subglabrous, with several leaves, purplish, flexuous; leaves pentagonal, the blades mostly 1-3.5 cm. in diameter, with fine pubescence, deeply cordate, trisect to the base, the 2 lateral segments deeply divided to make 5 more or less cuneate segments, each of these deeply 3- or more-lobed, the final divisions 1-5 mm. long, 2-4 mm. wide, rounded-oblong, mucronate; petioles 1-5 cm. long, purplish, slightly dilated at base; flowers solitary, large, deep purplish blue ; pedicels erect to horizontal, 5-15 cm. long, retrorse-pubescent, the hairs longer toward the summit; bracteoles usually 2, like reduced leaves, 5-15 mm. long, often near the middle of the pedicel; sepals with rather long fine hairs, the upper 20-25 mm. long, 12-15 mm. wide, broadly ovate, minutely 526 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Ficure 10. Delphinium, Group VI. Petals, the upper spurred, the lower clawed. A. D. maximowiczii; blade of upper petal 9 mm. long, spur 20 mm.; blade of lower petal 7 mm. long, claw 5 mm.; drawn from Rock 24272 (uc). . D. monanthum; blade of upper petal 12 mm. long, sp mm.; blade of lower petal 11 mm. long, claw 6 mm.; drawn from Rock 13185 (GH). Cc... mosoynense; blade of upper petal 8 mm. long, spur 18 mm.; blade of lower petal 6 mm. ‘ .5 mm.; drawn from Delavay 8 (Ny). D. D. muscosum, blade of upper petal 8 mm. long, spur 16 mm.; blade of lower petal 11 mm. long, claw 6 mm.; drawn from the published illustration of the type. E. D. norton, blade of upper petal 15 mm. long, spur 12 mm.; blade of lower petal 7 mm. long, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 527 apiculate, the spur 25-30 mm. long, ca. 4 mm. thick at base, subcylindric; lateral and lower sepals 25 mm. by 15 mm., somewhat erose on the more or less rounded tips; upper petals blue, slightly oblique, the blade entire, 12 mm. long, the spur 24 mm. long; lower petals blue, the lamina almost round, somewhat bearded and crenulate, 11 mm. by 10 mm., slightly bilobed, claw 6 mm. long; stamens 7-8 mm. long, more or less ciliate; anthers dark, 1.5 mm. long; follicles 3, pubescent to glabrous. Type: Szechwan, Dongrergo, 4800 m., Aug. 9, 1922, Harry Smith 3131; isotypes (A, MO), seen Other material — eh ae is from ee ean E. H. Wilson 4691 (GH, us). N. W. Yunnan: G. Forrest 20780 (us). . Kansu: Rock 13024 (GH, K, NY, mace 13185 = GH, K, NY, ae Hitinind Bonet 253 (E). 106. Delphinium mosoynense Franchet, Bull. Soc. Philom. Paris VII. 5: 168. 1893 B.C. Delphinium grandiflorum L. var. mosoynense (Franchet) Huth, Bot. Jahrb. 20: 461. 1895. Delphinium caeruleum var. majus Wang, Acta Bot. Sinica 10: 266. 1962. Wang refers Rock 13730, which I had placed in D. mosoynense, to this species. Stem slender, flexuous, branched above, retrorse-strigulose below, usually densely so above, 1—5 dm. tall; lower leaves with long (5—15 cm.) petioles not dilated at base; upper petioles shorter; leaf blades subglabrous to strigulose, divided to base into 5 parts, each of these further incised into divaricate lance-linear lobes mostly 5—20 mm. long, 1-2.5 mm. wide, sharp pointed, more or less inrolled at margins and pale beneath; cauline leaves 1-3, remote; flowers mostly in an open panicle, deep blue, the bracts linear, 5-20 mm. lone. or the lowest trifid; pedicels divaricate-ascending, 2-5 cm. long; bracteoles near the middle of the pedicel, linear, 5—8 mm. long; sepals strigulose, the upper 15-18 mm. long, ovate, the spur straight, sub- ulate, ca. 20 mm. long, 3 mm. wide at the base; lateral sepals 12-15 mm. long, 8 mm. ede obtuse; lower sepals like the lateral; upper petals pale, claw 6 mm.; drawn from _ BOG (Ek). es Bape ecstet blade of upper 1 ng, claw 8 mm.; blade of lower petal 9 mm. long mm.; ee ateg as D. ded blade of upper petal 71 5 othe ._ long, spur 15 mm.; blade of lower peta . long, claw 6 mm.; drawn ane Rock aay (GH ay Hertogpapanl Bade: of upper petal 13 mm. lon ng, s long, claw 7 mm.; drawn from Harry Smith 11273 (mo). 528 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 somewhat oblique, entire at apex, the limb 8 mm., the spur 18 mm. long; lower petals blue, ca. 6.5 mm. by 5 mm., yellow bearded in center of base of oblong limb, not ciliate; claw 4.5 mm. long; stamens glabrous, 5 mm. long; anthers dark, 1 mm. long; follicles 3, finely pubescent. Type: Yunnan, on Pee-cha-ko, near Moo-so-yn, north of Tali, Abdé Delavay 717, Aug. 31, 1884 (p); seen. Photograph (uc). DistriBuTION. Apparently ranging between 2500 and 4500 m., Yunnan, Szechwan, southwestern Kansu. Examptes. Yunnan: Rock 17338; Forrest 6533, 6643, 15233; K. M. Feng 2514; Maire 7437; Delavey 8; Ducloux 3875. Szechwan: Soulié 2389, 2387; Handel-Mazzetti 5103. Tibet: Kingdon Ward 10621. 107. Delphinium muscosum Exell & Hillcoat, Jour. Roy. Hort. Soc. Lond. 78: 183. 1953. Fic. 10, D. Perennial from a slender rootstock arising from a thickened elongate root; stems 10-15 cm. tall, many branched from base, hirsute-pilose throughout; basal leaves with petioles to 7 cm. long, flattened at base, pubescent-pilose; blade suborbicular in outline, 2-3 cm. in diameter, trifid to base, the segments divided 3-4 times into ultimate linear pubescent- pilose lobes scarcely 1 mm. wide; cauline leaves short-petioled to subses- sile; inflorescence subcorymbose, cymosely branched; flowers solitary at tips of branches, dark blue to dark violet, 3-3.5 cm. long; sepals yellow- hirsute-pilose without, almost glabrous within, the upper elliptic-ovate, 25 mm. by 15-18 mm., the spur 14 mm. by 4.5-6 mm., slightly decurved; lateral sepals elliptic, spatulate, 22 mm. by 11-13 mm.; lower sepals ovate or oblong-ovate, slightly obtusely acuminate, 23 mm. by 14-16 mm., gla- brous within; 2 upper petals narrowly elliptic, almost straight, 26 mm. by 4 mm., pubescent within, almost glabrous without, the lamina 8 mm. long, rounded and entire at apex; 2 lower petals with the lamina irregularly ovate, entire, 10-12 mm. by 6-8 mm., yellowish- or whitish-bearded to- ward claw, the latter 6-7 mm. long; stamens bluish, 7-8 mm. long, glabrous; anthers dark, 1.2 mm. long; follicles 3, tomentellous. Type: Bhutan, Rinchen Chu, at 5000 m., Aug. 7, 1937, Ludlow & Sher- riff 3537 (BM); isotype (£); seen. Rance. Growing on open stony wet scree, 4500 to 5500 m., Bhutan. EXAMPLES SEEN. Ludlow, Sherriff & Hicks 19725 (BM), 19328 (BM), 19375 (BM), 19399 (BM), 16825 (Bm), 17102 (Bm); Cooper 4791 (Bm). A species with leaves as finely dissected as any I have seen. 108. Delphinium nortonii Dunn, Kew Bull. 1927: 247. 1927. Fic. 10, E. Perennial, 7-10 cm. tall, with dense short spreading pubescence espe- cially above; stems several, 1-flowered; leaves long petioled (2-4 cm.), short-vaginate, the blades orbicular; radical leaves 2-3 cm. in diameter, palmately 3-parted, the segments broadly cuneate, short-petiolulate, im- 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 529 bricate, palmately incised into obtuse lobes 3-5 mm. long, 2-3 mm. wide; cauline leaves few, similar, smaller; flowers large, violet, inflated-globose, long-pubescent; upper sepal 25 mm. long, 18 mm. wide, broadly ovate, obtuse, the spur 10-15 mm. long, 5—7 mm. wide at the base, subcylindric, obtuse; lateral sepals asymmetrically elliptic-ovate, 18 mm. by 15 mm., rounded-truncate; lower sepals slightly longer, more acute; bracts folia- ceous, 1 cm. long; pedicels 4-5 cm. long; bracteoles 2, near middle of the pedicel, tripartite to oblong, to ca. 10 mm. long; upper petals almost straight, the lamina 15 mm. long, 3 mm. wide, somewhat emarginate at the blunt apex, spur 10-14 mm. long; lower petals unequally bifid, yellow bearded, the lamina 7—8 mm. long, 5~6 mm. wide, the bearded lobes 5 mm. by 2 mm.; filaments glabrous, 7-8 mm. long; carpels 3, velutinous. Type: Tibet, sandy soil above Kampa Dzong, 12 miles north of Sikkim, 5000 m., June 17, 1922, E, F. Norton 369 (x); seen. A specimen from Gyong, E. Himalaya, at 4300 m., by Rohmoo Lepcha, Sept. 12, 1912 (£), seems to belong here. The species is in the large- flowered group with D. muscosum, and seems to key out independently. 109. Delphinium przewalskii Huth, Bot. Jahrb. 20: 407. 1895. Fic. 10, F. Stem sparsely pilose, increasingly so in the inflorescence, erect, branched at the base, 1.5-2.5 dm. tall; leaves glabrous above, somewhat pilose be- neath, deeply 3—5-parted, the parts deeply laciniate into 2-3 lobules 2~6 mm. wide; bracts passing into cauline leaves; bracteoles broadly lance- olate and near the flower or spatulate and remote from it, 1-2 cm. long, 5~7 mm. wide; flowers usually terminal, deep blue, 1 to few in a corymbose raceme; sepals 2-2.5 cm. long, 1.2-1.5 cm. wide, loosely pubescent, sub- corniculate at tips; upper sepal ovate, 25 mm. by 15 mm., apiculate; lateral sepals elliptic-ovate, 25 mm. by 15 mm.; lower sepals 28 mm. by 18 mm., acute to acuminate; spur ca. 20 mm. long, straight or arcuate, 3.5 mm. wide at base; upper petals more or less ciliate, dark brown, the lamina almost straight, 15 mm. long, 3 mm. wide, glabrous, emarginate, the spur ca. 15 mm. long; lower petals semibifid, white-ciliate, dark brown at base, yellow bearded, less deeply colored in upper part, the lamina to 10 mm. by 4 mm.; claw ca. 8 mm. long, somewhat appendiculate; stamens 7-8 mm. long, glabrous; anthers purple, almost 2 mm. long; carpels 3, densely Type: W. Mongolia, “Montes Alaschan,” Przewalski “1874” in Huth, “1871” on herbarium sheet, “1873” (LE), and “1873” (K). The 2 Lenin- grad specimens (1871 and 1873) are on one sheet, with the label “Synty- 110. Delphinium pseudo-caeruleum W. T. Wang, Acta Bot. Sinica 10: 269. 1962 Stem ca. 9-10 dm. tall, 3 mm. thick at base, indistinctly 5-angled, re- 530 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 trorsely strigulose, many branched, equally foliose; lower cauline leaves long petioled, the upper short or subsessile; blades as in D. caeruleum and D. grandiflorum, the ultimate lobes linear, 1.5~-1.8 mm. wide, more or less revolute, ciliate, ventrally subglabrous, villous along nerves dorsally; petioles subglabrous or retrorse-villous below; inflorescence corymbose, 2—5-flowered; lower bracts trifid into linear lobes, others entire, linear; pedicels densely retrorsely strigulose, 3.2-7.5 cm. long, suberect; brac- teoles near middle of pedicel, linear, 4-6 mm. long; sepals deep blue, oblong-obovate or narrow-elliptic, minutely puberulent; spur 20-22 mm. long, subulate, straight or slightly decurved; upper petals dilute blue, glabrous, emarginate; lower petals blue, the limb and claw subequal, the limb yellow bearded at base, subelliptic, bifid to middle into triangular, obtusish, ciliate lobes; stamens glabrous; carpels 3, densely brownish- strigulose. Type: Kansu, Si-Ku, 2500 m., July 29, 1951, T. P. Wang 14658 (PE), not seen. Isotype and photograph (£), seen This species suggests D. mollipilum, but has strigose stems and pedicels and bifid lower petals. 3 111. Delphinium pseudograndiflorum W. T. Wang, Acta Bot. Sinica 10: 275. 1962. Fic. 10, G. Stem about 3-7 dm. tall, almost glabrous and glaucous, with occasional spreading hairs especially about the nodes and on the pedicels just below the flowers; basal leaves few, the blades to ca. 1 dm. in diameter, sub- glabrous or with a few spreading and some minute appressed hairs, round- pentagonal in outline, divided almost to the base into obovate segments, cuneate and 1 cm. or more wide with little lobing, or more frequently laciniate into linear parts 1-2-3 mm. wide and 1-6 cm. long, acuminate; cauline leaves largely of the latter type, scattered and gradually reduced up the stem; lower petioles to 10 or more cm. long, very slightly vaginate at the base, the upper much shorter; inflorescence open, branched, the racemes lax, few to several flowered; lower bracts trilobed, others entire, linear-subulate, 5-15 mm. long; pedicels arched or erect, 2-10 cm. long; bracteoles linear, 3-7 mm. long, near the flower or lower; sepals deep blue, often tinged purple, strigulose and with longer hairs without, more or less crinkled on margin; upper sepal round-ovate, 13-20 mm. long, rounded- obtuse, the spur 22-26 mm. long, 2.5 mm. wide at base, straight or some- what decurved; lateral sepals elliptic-oblong, 16-20 mm. by 6-10 mm., rounded at apex; lower sepals elliptic-oblong, similar to but slightly wider the lateral; upper petals light colored, the blade slightly ae m. by 2.5 mm., glabrous, subentire, the spur ca. 17 mm. long; low ate blue, the lamina strongly oblique, round-obovate, 8 mm. by 65 mm., deeply emarginate, yellow bearded at base, ciliate, the claw 5—6 mm. long; stamens 5—6 mm. long, ciliate; anthers dark, 1 mm. long; carpels 3, hairy, with dark veining, ca. 12 mm. long by 3 mm. wide, the styles an additional 3 mm.; seeds angled, winged, 1.2 mm. long. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 531 Type: Szechwan, Hei-shin, 2400-3300 m., July—Aug., 1957, H. Li 74016 (PE), not seen; photo. of type and isotype seen (E). Rance. Growing in fields and scrub up to 5000 m., Tibet, western China. EXAMPLES SEEN. Tibet: Hanbury-Tracy 195; Kingdon Ward 10881, 10621, 12077; Ludlow & Sherriff 687, 2595, 499; Ludlow, Sherrif & Taylor 6093, 5162, 13920: Rock 14538. Szechwan: H. Stevens 458 in part; Harry Smith 3985. Yunnan: G. Forrest 16951, 15112, 6538, 23185; Schneider 3250; Yu 13481; H. Wang 41742. Sikang: H. Smith 13764. Kansu: R. C. Ching 922, 1070; Rock 13178, 14594. Near to D. grandiflorum and referred here is much of the more western material that formerly was identified as that species. It has longer spurs, a more erose lower petal and sometimes looser stem-pubescence. Wang (p. 275) proposed also var. glabrescens with glabrous pedicels and var. lobatum (p. 276) with more deeply lobed lower petal-laminae. I have had no material of these, and from descriptions only, must refer them to synonymy. : 112. Delphinium pulcherrimum W. T. Wang, Acta Phytotax. Sinica 65.370. 1957, Fic. 10, H. Stem 1.5-3 dm. tall, equally scattered-leafy, striate, densely retrorse- pubescent or subglabrous, few branched; basal leaves lacking; lower cauline leaves long petioled, upper shorter, 3.5-7.5 cm.; leaf blades pen- tagonal-orbicular, cordate, 3-4 cm. in diameter, 3-parted almost to the base, middle segments elliptic-rhombic, long acuminate, laciniate, the lobules oblong-linear, 2 mm. wide, acute or rounded at apex, strigose; laterals unequally 2—3-parted, the secondary parts long acuminate, lacin- iate; petioles 1.5—7 cm. long, vaginate; flowers blue purple; pedicels to 6 cm. long; bracteoles near the flowers, lanceolate to elliptic, 8-11 mm. long, ciliate; sepals round or broadly elliptic, short-acuminate, loosely puberulent without, 17-22 mm. long, 12-17 mm. wide, spur 23-25 mm. long, 3 mm. wide at the base, straight or decurved, obtuse; upper petals 13 mm. long, 2 mm. wide, obtuse, unequally 2-lobulate, glabrous; lower petals 14 mm. long, the limb elliptical, bifid to the middle into 2 ovate, loosely ciliate lobes rounded at apex, yellow bearded below the middle, the claw equaling the limb, appendiculate; stamens 6-7 mm. long, glabrous; carpels 3, densely silky. Type: Yunnan, A-tun-tze, Pai-ma-shan, Shu-ling, 4240 m., Aug. 3, 1937, T. T. Vii 9330 (PE), not seen; photo. (E), seen. OTHER MATERIAL referable here. Southern Tibet: Ludlow & Sherriff 1997, 1958: Rock 14460. Kansu: E. Licent 4686; R. C. Ching 1123. T. Wang, Acta Bot. Sinica 10: 268. Fie. 10, I. spreading-pilose, 1-foliate 113. Delphinium pumilum W. 1962. Stems 10-16 cm. tall, to 1 mm. thick below, 532 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 above, simple or rarely 1-branched; basal leaves 3—4, long petioled, the blades pentagonal-reniform, 7-14 mm. long, 13-24 mm. wide, cordate, trisect, the middle segment rhombic-obovate, cuneate at base, 3-parted to below the middle, the parts lance- or oblong-linear, 1-2 mm. wide, acutish, entire or with 1 lobule; lateral segments flabellate, unequally bipartite to the base, the upper parts like the middle segment but slightly oblique, the lower subequal to the preceding, unequally twice parted into ultimate linear lobules, slightly revolute, strigose beneath, loosely silky pilose above, the petioles 2-6 cm. long, spreading-pilose, short vaginate; cauline leaf similar; inflorescence corymbiform, mostly 2-flowered; bracts leaflike or small, linear, 6 mm. long; pedicels 2-9 cm. long, subequal, the hairs white, spreading or retrorse and yellow, glandular; bracteoles near or above the middle of the pedicel, linear-subulate, 3-4.5 mm. long, with loose spreading puberulence; sepals blue, elliptic or elliptic-obovate, 12-14 mm. long, loosely puberulent, the spur subulate, 33-38 mm. long, 3-3.5 mm. wide at the base, slightly curved; upper petals glabrous, emarginate; lower petals blue, the limb broad-obovate, yellow bearded at base, ca. 6 mm long and wide, subtruncate at tip, emarginate, glabrous or loosely puberu- lent, claw as long as the limb; stamens glabrous; carpels 3, pilose. Type: Szechwan, Muli, Sept. 14, 1959, S. K. Wu 2662 (PE), not seen; photo. (£), see The author compares Delphinium pumilum with D. caeruleum and D. tatsienense, all three being small plants. I refer here, Szechwan, Muli King- dom, Rock 18316; Mt. Konka, Rock 16886. Yunnan, Vii 14413 114. Delphinium pylzowii Maxim. ex Regel, Gartenflora 25: 289-290. pl. 879. 1876; Maxim. Bull. Acad. Pétersb. 23: 307. 1877. Fie. 10; J; & Delphinium pylzowii var. trigynum W. T. Wang, Acta Bot. Sinica 10: 78. 1962, in part, as to Rock 13194 and 14501 cited by Wang. mipiyetise caeruleum var. macranthum Finet, Jour. Bot. Morot 21: 98. 1908. Based on Soulié 3903, 3911, 3914 from Yargong, Batang. Seen (P). er on caucasicum var. tanguticum Maxim. in schedis Delphinium tanguticum (Maxim.) Huth, Bull. Herb. Boiss. 33t. (pon. 1893. Type. Szechwan, G. N. Potanin, 1885 (LE), seen; isotype (K). seen. Delphinium labrangense Ulbrich ex Rehder & Kobuski, Jour, Arnold Arb. 14: 11, 1933, nomen. Proposed for Rock 14482, 14483, 14504, 14513. Perennial from a more or less woody subterranean base, the stems 1 to several, slender, 1-3 dm. tall, simple and 1-flowered or few branched and few flowered, pilose to retrorsely strigulose, the hairs in the upper part less appressed, very dense, short to rather long and fine; lower leaves 3- or 5- parted, 1.5-2.5 cm. across, the parts more or less cuneate, broad, coarsely 3-toothed at tip, the petioles very slender, 2—5 cm. long, somewhat dilated at base; main cauline leaves few, more deeply laciniate into oblong lobes 2-3.5 mm. wide, 4-10 mm. long; bracts entire to 3-lobed, the lobes entire, to 7 mm. wide or laciniate; pedicels erect, 2-10 cm. long; bracteoles linear, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 533 near the middle of the pedicel, linear to lanceolate, 5-15 mm. long, seta- ceous, to 5 mm. wide; calyx deep blue; upper sepal ovate, 15-22 mm. long, 10-15 mm. wide, obtuse, with fine rather long hairs, spur somewhat decurved, 24-28 mm. long, ca. 3.5 mm. wide at base, only gradually nar- rowed into a somewhat obtuse tip; lateral sepals broadly ovate, 20-22 mm. long, 16—18 mm. wide, obtuse; lower sepals 24 mm. by 15 mm., ovate, obtuse; upper petals from yellowish with a blue margin to almost black- purple, the laminae straight, not toothed to erose or emarginate, about 14 mm. long, the curved spur 25 mm. long; lower petals bearded, the elliptic- ovate lamina ca. 8 mm. long, divided to ca. 5 mm., the claw ca. 5 mm. long; stamens 7—8 mm. long, glabrous; anthers dark, narrow, 1.5 mm. long; carpels 5 or 3, pubescent; seeds with low lamellate papillae. Type: Kansu, Terra Tangutorum, Przewalski, 1872 (LE), seen; isotype (P), seen. Rance. About 3500 to 4500 m., Kansu, Szechwan, eastern Tibet. ILLUSTRATIONS. Maximowicz, Fl. Tangut. 1: p/. 3. 1889; Gartenflora 25: pl. 879. 1876; Bot. Mag. 145: pl. 8813. 1910. REPRESENTATIVE MATERIAL. Kansu: Przewalski in 1880; Rock 13093, 13041, 14644, 14625, 13086, 13088, 13023, 14641; Farrer & Purdom F 710. Szechwan: Wilson 3094, 4691; H. Smith 3135, 2961, 3384, 4283. Tibet: Ludlow & Sherriff 8854, 9902. 115. Delphinium rockii Munz, sp. nov. Fic. 10, L. Stem ridged, dark, to ca. 1 m. tall, 4-5 mm. thick, more or less strig- ulose, branched above, scattered-leafy; petioles scarcely dilated at the base, the cauline to 1 dm. long, somewhat retrorse-pilose; cauline leaf blades generally reduced up the stem, 3—7 cm. in diameter, rounded-pen- tagonal, strigose above, more loosely so beneath, cleft almost to the base, the divisions rather narrow, 3-8 mm. wide, laciniate or lobed into ultimate parts 1-3 mm. wide, acute; inflorescence open paniculate, the ultimate branches leafy-bracted and bearing umbellate clusters of 3 or more flowers: bracts ca. 1 cm. across, deeply laciniate; pedicels 2-3 cm. long, divergent, strigulose, not glandular; bracteoles subtending the flower, strigulose, 2-3 mm. long, 1 mm. wide; sepals bluish purple, strigulose, the upper sepal oblong-ovate, 18-20 mm. long, 10-11 mm. wide, obtuse, the spur decurved, strigulose, 14-16 mm. long, 3 mm. wide at the base, the other sepals elliptic-ovate, 15 mm. long; petals dark, the upper laminae suberect, glabrous, entire, obtuse, 11-12 mm. long, the spur 15 mm. long; lower laminae oblong, 7 mm. long, unequally shallowly bilobed, the upper surface white-hairy at the base, ciliate; the claw 6 mm. long, appendaged; stamens glabrous; carpels 3, glabrous. Caulis ad 1 m. altus, 4-5 mm. crassus, strigulosus, Supra Tamosus; petioli vix dilatati, ad 1 dm. longi, retrorso-pilosi; laminae foliorum caul- ium 3-7 cm. latae, rotundato-pentagonales, supra strigosae, infra laxe pubescentes, alte lobatae, laciniis ultimis 1-3 mm. latis; inflorescentia paniculata; rami cum bracteis foliosis et umbellis 3—5 florium; pedicelli 534 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 2-3 cm. longi, divergentes, strigulosi; bracteolae strigulosae, 2-3 mm. longae; sepala azureo-purpurea, strigulosa; sepalum superiore oblongo- ovatum, 18-20 mm. longum, 10-11 mm. latum; calcar 15 mm. longum; sepala latiora et inferiora elliptico-ovata, 15 mm. longa; petala atropur- purea; laminae superiores glabrae, integrae, obtusae, 11-12 mm. longae; laminae inferiores oblongae, 7 mm. longae; inaequaliter bilobatae, base hirsutae, ciliatae; unguis 6 mm. longus, appendiculatus; stamina glabra; carpella 3, glabra. Type: Eastern Tibet, grasslands between Labrang and Yellow River; valley of Kachiassu, 3000 m., Aug. 5, 1926, J. F. Rock 14543 (cH). The proposed new species is near D. henryi and D. trisectum in the shape of the lamina of the lower petal and in having the bracteoles very near the base of the flower; but differs in its more paniculate inflorescence and its upper petal entire at the apex. 116. Delphinium smithianum Hand.-Mazz. Acta Horti Gothob. 13: 49. 1939. Fic. 10, M. Perennial from an elongate root, with 1 or more stems 3-12 cm. high, simple or few-branched, suberect, more or less purplish, leafy and with short spreading hairs; leaves reniform-pentagonal in outline, the blades 2-5 cm. wide, bifid or trifid to near the base into cuneate obovate divisions, the 2 lateral ones deeply divided to make 5 subequal parts, obovate, 2—3- lobed, these lobes each with 2 or 3 broad teeth rounded at the apex, 1-3 mm. long and ca. 2 mm. wide, the parts somewhat more puberulent be- neath than above; petioles slender, 2—6 cm. long, dilated at base, somewhat pubescent; flowers 1-4 on a stem; pedicels 3-8 cm. long, suberect, arising in upper axils; bracteoles 2, leafy, ca. 3-lobed, 6-8 mm. long, placed above the middle of the pedicel; sepals broadly ovate, obscurely violet, fine- hairy; upper sepal 18-20 mm. long, 14 mm. wide; spur 20-22 mm. long, 5 mm. wide at base, subcylindrical, horizontal, blunt: lateral sepals round- ovate, obtuse, pubescent mostly along the purplish midvein; lower sepals by mm., acute, pubescent; upper petals dark, the laminae straightish, 13 mm. long, slightly and unequally bidentate, the spur 20 m. long; lower petals dark, the oblong blade enlarged somewhat distally, bearded, ca. 8 mm. long, bifid ca. 1/3 its length, the narrow claw 7 mm. long; stamens 7-8 mm. long, glabrous; anthers dark, 1.2 mm. long; carpels silky villous; seeds wingless. yPE: Sikang, Kangting (Tachienlu), 4600 m., Aug. 20, 1934; Harry Smith 11274, type (w), not seen; isotypes (A, Mo), seen 117. Delphinium sordidecaerulescens Ulbrich, Notizbl. Bot. Gart. Berlin 12: 356. 1935. Fic. 13, A. Erect perennial, sparsely branched, 6-8 dm. tall; stem slender, yellowish green with purplish base, very sparsely hairy; basal leaves gone by an- thesis; cauline leaves few, remote, gradually reduced up the stem; lower 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 535 petioles to 18 cm. long, slightly dilated at the base; blades suborbicular, 4—5 cm. in diameter, about septemfid, the segments 2—3-partite into linear lobes 1-3 mm. wide, acutish, sparsely pilose; upper leaves trifid into entire linear lobes; the inflorescence open paniculate, glabrous to somewhat pubescent on the pedicels with white hairs; bracts linear, 5-20 mm. long; pedicels divergent, 3-9 cm. long, suberect, slightly curved; bracteoles linear, 4-6 mm. long, well below the flower; sepals sordid-bluish, somewhat crisped-pubescent, with a greenish tubercle below the apex; upper sepal ovate, 11-12 mm. long, 8 mm. wide, obtuse, the spur straight, spreading, 22-23 mm. long, 3.5 mm. wide at the base, very slender toward the tip; lateral sepals oblong, 9 mm. long, 7 mm. wide, round-truncate at the apex; lower sepals 14 mm. long, 7 mm. wide, obtuse; petals pale, whitish or yellowish, the upper laminae entire, oblique, ca. 7 mm. long, glabrous, somewhat pointed, the spur about 20 mm. long; lower laminae oblong- obovate, somewhat erose, about 6 mm. long, 4.5 mm. wide, yellow-bearded at the base, short-ciliate on the margin, the claw appendiculate, 5-6 mm. long; stamens 6-7 mm. long, glabrous; anthers dark; carpels 3, pilose, 8-9 mm. long, the filiform styles another 2-3 mm.; seeds glabrous, ca. 3 mm. long. 2? Type: “Szechwan, reg. bor. Huang-chen-kuan, in prato, 3400 m., Harry Smith 3983 (ups; seen through the courtesy of Dr. J. A. Nannfeldt). This species is very close to Delphinium pseudograndiflorum Wang, over which its name has priority. Its flowers are apparently much paler in color and smaller, and without more material for comparison I am main- taining the two species. 118. Delphinium sutchuense Franchet, Bull. Soc. Philom. Paris VIII. Ed oF pM | Fie, LI, A. Stem erect, sparsely pubescent to glabrous, scattered-leafy to the top, 4-8 dm. tall; lower petioles slender, not or scarcely dilated; blades 6-9 cm. wide, fine-pubescent, 3—S-sect, the segments oblong, entire at base, narrow-cuneate, ca. 3 cm. wide above, subpetiolulate, distally trifid, then incised into few, long-acuminate, very pointed lobes, 2-4 mm. wide; in- florescence branched, the branches erect, almost appressed to the central stem, bare, with a 2—6-flowered raceme at the summit; upper bracts linear- subulate, the bracteoles subulate, a little below the sepals, 10-12 mm. long, very narrow; pedicels twice as long as flowers, closely strigulose; flowers violet blue, strigulose; upper sepal ovate-oblong, obtuse, 11 mm. by 7 mm., spur 13-18 mm. long, horizontal or slightly ascending, obtuse; other sepals ca. 15 mm. long; petals blue, the upper oblique, the lamina 9 mm. long, emarginate, possibly few-haired at the apex; lower laminae oblong-ovate, 5.5 mm. by 4 mm., bearded at base, ciliate, divided almost halfway into rather broad lobes, claw appendiculate, 5 mm. long; follicles 3, silky. Type: W. China, in north of province of Sutchuen (Szechwan), Potanin, 1885 (Pp); seen. 536 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 (cH). H. D. ts . long, spur 20 mm. long; blade of lower petal 7 mm. long, claw 5 mm.; drawn from Rock 14547 (GH). . williamsii; gist of upper petal 10 mm. long, s spur 17 mm.; blade of lower petal 6 mm. long claw 6 mm.; drawn from Polunin, Sykes & Williams 1907 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 537 119. Delphinium tatsienense Franchet, Bull. Soc. Philom. Paris VIII. 5: 169. 1893. Fic. 11, B Delphinium bonatii Lévl. Repert. Sp. Nov. 7: 99. 1909. Type, Yunnan, Ducloux 596, isotypes (, uc), seen, Lauener & oe Notes Bot. Gard. Edinb. 23: 585. 1961, referred this to D. grandifloru Delphinium soonmingense Chen, Bull. Fan Mem. Inst. Biol. Peiping (n.s.) 1: 175, 1948. Based on Y. P. Chang 450 from Soonming (PE?); not seen. This referred to oo under D. tatsienense by W. T. Wang, Acta Bot. Sinica 10: 267. Delphinium papas var. latisectum W.T. Wang, Acta Bot. Sinica 10: 164. 1962. Type C. W. Wang 66226 (PE), not seen; isotype seen (A). Stem 2—6 dm. tall, slender, simple or with slender branches, subglabrous to strigulose, the hairs tending to be more spreading in upper parts; leaves remote, few, the lower petioles to 1 dm. or longer, slender, strigulose, not vaginate at base; the upper shorter; blades 3-8 cm. or wider, divided to base into 3 or 5 parts that are narrow-cuneiform at their base, deeply in- cised into obtuse linear lobes commonly 2-4 mm. wide, green and scantily pubescent above, paler and more pubescent beneath; flowers few, in an open corymb or short raceme; lower bracts foliaceous, trifid into linear lobes 1-1.5 cm. long or simple; pedicels arched-ascending to straight, commonly 2-several cm. long, strigulose; bracteoles linear, often more than 2, commonly lance-linear, 4-8 mm. long, near the middle of the pedicel or above; sepals deep blue, pubescent; upper sepal broadly ovate, obtuse, 20 mm, by 12 mm., the spur horizontal, quite straight, 30 mm. long, 3 mm. wide at base, slender, narrowed to rather an acute point; lateral sepals oblong, 17 mm. by 9 mm., rounded- ae at tip, pubescent along midvein; lower sepals 20 mm. by 10 mm., narrow-obovate, obtuse; upper petals bluish, oblique, 10 mm. long, glabrous, the spur very slender, almost 30 mm. long; lower petals blue, oblong-obovate, the lamina 9 mm. by 6 mm., subentire to cleft at tip for 2 mm., ciliate, densely bearded near base of the blade, the claw 5 mm. long, narrow; stamens 6~7 mm. long, the filaments slightly ciliate; anthers dark, 1.2 mm. long; follicles 3, short- pubescent, 12-15 mm. long, 3-4 mm. wide, the styles an additional 3 mm.; seeds winged on angles. Type: Szechwan, near Tatsienlou, Henri d’Orléans & Bonvalot, July 29, 1890 (Pp); seen. Photographs of type (BH, E, UC); seen. DistripuTion. Largely at 2500 to 3000 m., western China. Examples. Szechwan: Soulié 2027, 2381, 2382, 2489, 2380; H. Stevens 89, 114; H. Smith 4660, 10798, 13764; W. P. Fang 4276; Cunningham 193, 355, 490; Wilson 3096. an: Vii 12941, 14527; sa 6494, 11449, 22378; Maire 380, 720; Rock 18485, 6292, 7770, 11485, 10686, 11545; Schneider 2739, 2182. Tibet: Rock 23243; Soulié 3529. er petal 9 mm. long, spur 19 mm.; blade of ie mae oe _— 2 ha Se n from E. it Wilson 2396 ‘(xy). K. . yuanum; blade of rue este e mm. “ee spur 12 mm.; blade of lower petal 5 mm. long, claw 4 mm.; 5 aie from Vu 12490 (A dD. yunnanense ; oe of uppet petal 7.5 mm lon ur 22 mm.; blade of lower petal 5.5 mm. long, w 4 mm.; drawn from Dales 967 (xy). 538 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 120. Delphinium tenii Lévl. Repert. Sp. Nov. 7: 98. 1909. Pigl,<., pp. Delphinium dolichocentrum W. T. Wang, Acta Bot. Sinica 10: 162. 1962. Based on Yunnan, Atuntze, T. T. Vii 9885 (PE), not seen; photo. (£), seen. Stems divaricate, slender, 8-15 (—45) cm. tall, white-villous, simple; leaves basal and a few cauline, the lower petioles very slender, 2—4 cm. long, spreading-villous, somewhat dilated at the base, the cauline not dilated, the upper shorter; leaf blades 2—4 cm. wide, more or less triangu- lar, divided to base to form 3 cuneate-obovate segments, green and finely strigulose above, paler and with longer white hairs beneath, these segments trifid and somewhat additionally lobed into oblong-linear, obtuse parts 2-4 mm. wide; flowers few in an open corymb; bracts ordinary leaves; pedicels strigose or spreading-pubescent, to 6 cm. long; bracteoles leafy, ternate or entire, and elliptical, at about the middle of the pedicel; sepals deep blue, strigose; upper sepals 12-15 mm. long, 8—10 mm. wide, obtuse, the spur decurved toward the apex, 20-22 mm. long, 3 mm. wide near the base, gradually narrowed toward the slender tip; lateral sepals elliptic- oblong, 13-15 mm. long, 6-8 mm. wide, rounded at the apex; lower sepals 12-15 mm. long, 5-8 mm. wide, obtuse; petals bluish, the upper laminae almost straight, entire, glabrous, 10 mm. long, the spur 17 mm. long; lower laminae oblong, emarginate, 6 mm. long, 2.5 mm. wide, bearded near the base and sparsely elsewhere, claw 5 mm. long; stamens 6-7 mm. long, filaments moderately widened; anthers blue, 1 mm. long; follicles 3, hairy, 12-14 mm, long; seeds 2 mm. long, narrowly winged on angles. Type: Yunnan, Lou-Fou near Tong-Tchouan, June, 1906, Simedn Ten 581 (£), isotype (UC); seen. Examptes. All from Yunnan: T. T. Vii 9885 (a), 13336 (a); C. W. Wang 64707, 69117, 64733, 69174, 69397, all (A); H. T. Tsai 56240 (a), 51974 (A); K. M. Feng 2841 (a); Rock 18205 (a, us); C. Schneider 2510 (GH). 1, E. Root woody, 5 mm. thick; stem 1 m. tall, 3 mm. thick, simple, terete, erect, glabrous above the white-pilose base; radical and lower stem leaves broadly reniform, 7-9 cm. wide, 3-partite to near the base, the lateral segments bipartite, all cuneate-obovate, pinnatifid into oblong or ovate- oblong lobes, acute or obtuse, glabrous above, sparsely white-hairy be- neath; petioles 8-10 cm. long; upper leaves glabrous, 3-parted, the lateral segments 2-parted, all oblong, 1—3-toothed or entire, the blades 4—6 cm. long, 6-7 cm. wide, petioles 3-4 cm. long; racemes laxly branched below, the branches arcuate-divaricate, with about 20-22 flowers; bracts linear, 2.5-5 mm. long; pedicels 15-45 mm. long, slender, arcuate-ascending, bibracteolate above the middle, viscid pilose apically with yellow hairs; bracteoles linear, 2 mm. long; flowers blue; spur cylindric, gibberulate below the apex, 10-11 mm. long, 2 mm. wide; sepals glabrous, 9-11 mm. 121. Delphinium tenuipes Tamura, Acta Phytotax. Geobot. 16: 141. 1956. Fie. 11, 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 539 long, 3-5 mm. wide; upper petals glabrous, the lamina ca. 8 mm. long, entire, somewhat oblique; lower petals bifid, white bearded; carpels 3, viscid-pilose. Type: Afghanistan, Nuristan, Chatrass, 2100 m., Kitamura, Aug. 2, 1955 (Kyo, not available A specimen from West Pakistan, Swat, Kalam to Utrot, 2500 m., July 20, 1953, R. R. Stewart & A. Rahman 25125 (BM) seems to me to match the illustration by Kitamura, Fl. Afghan. 125. fig. 51. 1960. 122. Delphinium trisectum W. T. Wang, Acta Bot. Sinica 10: 80. Stems 4.5—5 dm. tall, 3-6 mm. thick at base, densely retrorse-strigulose, simple or with one branch above, subequally foliose; cauline leaves long petioled, the blades broadly reniform (3—)4.5—6.5 cm. long, (5.2—)7.5- 12 cm. wide, basally cordate, trisect, the segments broadly obovate or rhombic-obovate, 3-lobed to the middle, the lobes 3-lobulate to the middle, these lobules lance-ovate, entire or 1-toothed, lateral lobes also bifid and much divided; petioles 15-17 cm. long, indistinctly vaginate; racemes 10— 14-flowered, lax; lower bracts short petioled and trisect, the segments lanceolate, the upper bracts linear; pedicels 2—7 cm. long, densely retrorse- pubescent; bracteoles near the flowers, lance-linear, 7-9 mm. long, 1 mm. wide, puberulent; sepals purple, elliptic to elliptic-obovate, 14-17 mm. long; spur slightly longer than sepals, cylindric-subulate, straight or slightly decurved; petals dark brown, emarginate, glabrous; lower petals with elliptical laminae, bilobed, long-ciliate, yellow bearded near the base; stamens glabrous; follicles 3, spreading, puberulent. Type: Honan, Shang-Cheng, 400 m., April 19, 1959, Honan Exped. 149 (PE), not seen; photo. seen (E). Near Delphinium henryi and D. rockii. 123. Delphinium triste Fisch. ex DC. Syst. 1: 362. 1817. Fic. 11, F. Delphinium obscurum Stev. ex DC. Syst. 1: 362. 1817, nomen pro synon. Stem 3-7 dm. tall, slender, simple or few branched above, leafy, retrorse- strigulose below, the hairs less appressed in the inflorescence; lower petioles slender, somewhat dilated at the base, 5-12 cm. long, the upper shorter; laminae broader than long, subglabrous to strigulose (especially beneath), mostly 4-12 cm. wide, divided quite to the base into 3 or parts, these narrow-cuneate below, obovate, deeply laciniate into linear lobes 1.5—4 mm, wide, acute; racemes lax, 3—15-flowered, the lower bracts trifid, pubescent, to 2 cm. long, others linear-lanceolate, to ca. 1 cm. long; pedi- cels ascending, pubescent, 2—5 cm. long; bracteoles lance-linear to linear, mostly somewhat below the sepals, 3-6 mm. long; sepals purplish black, cinereous pubescent, the upper sepal ovate, 22 mm. long, 12 mm. wide, rounded-obtuse, the spur almost straight, 12-14 mm. long, 5 mm. wide at the base, obtuse; lateral and lower sepals rhombic-ovate, 14-15 mm. by 540 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 10-11 mm., rounded at apex; petals black, the upper laminae slightly oblique, entire but slightly irregular and crisped on the rounded setose tip, ca. 12 mm. long, the spur 13 mm. long; lower laminae bearded on both sides, ovate, 7 mm. long, 5 mm. wide, cleft for about 2 mm. into rather broad lobes, claw 6 mm. long; stamens 7-8 mm. long, the filaments much broadened at base, glabrous; anthers dark, oblong, 1.3 mm. long; follicles 3, somewhat divaricate, densely pubescent, 14-17 mm. long, ca. 4 mm wide, the styles an additional 3 mm.; seeds brown, 2.5 mm. long, densely clothed with overlapping transverse rows of scales. Type: “Hab. in Daouria circa Doroninsk (Vlassof ex Fisch.).” I have seen a sheet from Dahuria, Herb. Fischer (LE, Pp) which may be type material. Two other specimens loaned from Leningrad both from the Herb. Ledebour are Turcz., 1829 and Turcz. no. 929. The Turczaninow material agrees well with the Herb. Fischer sheets. ILLUSTRATIONS. KNow es & Westcott, FI. Cabinet 2: pl. 54. 1838; J. W. Loupon, Ladies’ Flower Garden of Ornamental Perennials 1: p/. 10. 1843 This very distinctive species with dark flowers is represented by the following Mongolian collections: “In alpibus sajanensibus, Radde” (K, LE, Pp); “In pratis ad fl. Dschida”’ ee (xk, Pp); Ikonnikov- -Galiteky 82 (uc); Lake Kosogol, Newburg, in 1924 (a, 124. Delphinium tsoongii W. T. Wang, Acta Phytotax. slit 6: 368. 1957, HG. Stem 6—9(—50) cm. long, simple or branched, with short white, usually retrorse hairs, rather equally leafy throughout; lower petioles up to 7 cm. long, slender, with white scattered hairs, somewhat vaginate, ee shorter up the stem; leaf blades broader than long, cordate, 1- : diameter, 5-parted to the base, the petiolulate parts repeatedly divided into ultimate oblong-linear lobes mostly 1-2 (—3) mm. wide; inflorescence glandular-villous with white and yellowish hairs, each branch ending in a 2—7-flowered corymb; lower bracts dissected, foliose, the upper entire, lance-linear; pedicels ascending, 2-5 cm. long; bracteoles linear, slightly above the middle of the pedicel, 2-4 mm. long; sepals dark purplish blue to bright blue or paler, villous; upper sepal elliptic to broadly ovate, 13-16 mm. long, 6—8.5 mm. wide, the spur slender, slightly decurved toward the narrow tip, 20-25 mm. long, 2.5-3 mm. wide at the base; lateral and lower sepals elliptic-oblong, rounded to obtuse at tip, 14-18 mm. long; upper petals dark, with a slightly oblique lamina 10 mm. long, the tip rounded and entire; lower lamina oblong, oblique, 7 mm. by 4 mm., cleft for 2 mm., yellow bearded near the base, subglabrous to white-hairy on remainder, claw 5 mm. long; stamens 6-7 mm. long, glabrous; anthers dark, 1.5 mm. long; follicles 3, hairy, divergent at tips, 15-17 mm. long, the styles an additional 2.5 mm. long; seeds black, obpyramidal, 1 mm. long, covered with minute scales. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 541 Type: Tibet, between Cha-chia and Chiao-tuao, Aug. 9, 1953, P. C. Tsoong 6073 (presumably PE, not seen). If I understand this species at all, I would refer here Rock 14547 from Tibet (GH, Ny); Rock 13154 (GH, us) and 13193 (cH, NY), both from Kansu; and R. C. Ching 534 (cH, us) from Kansu. Delphinium tsoongii is near D. mosoynense but has somewhat smaller floral parts and a different range. 125. Delphinium williamsii Munz, sp. nov. FiG.Lks Perennial from long branching woody roots; stems rather slender, 4—10 dm. tall, freely and openly branched in upper half, scattered leafy, but the leaves more abundant below, largely glabrous except for the strigulose tips of the inflorescence; lower petioles 5-10 cm. long, very slender, sub- glabrous, more or less dilated at the base, the upper petioles shorter; leaf blades to 1 dm. broad, 5-fid to near the base, sparingly strigulose to sub- glabrous, ciliate, variegated, being rather pale near the center, paler on under surface, the segments narrow at the base, obovate-rhombic, lobed then dissected into ultimate parts largely 1-2 mm. wide, acute-mucronate; upper cauline leaves with almost filiform ultimate divisions; inflorescence open, the racemes lax, axes subglabrous, several flowered, to ca. 2 dm. long; bracts linear, 5-12 mm. long, or the lower divided; pedicels diver- gent, 1.5-5 (—10) mm. long, somewhat strigulose toward the tips; brac- teoles near the middle of the pedicels or above, linear to sublanceolate, 2-5 mm. long; sepals dull white with greenish tips, subglabrous except for minute appressed hairs on basal parts; upper sepal oblong-ovate, 14— 18 mm. long, 8 mm. wide, apiculate, the spur straight, spreading, sub- cylindric, 18-21 mm. long, 3 mm. wide at base, rather blunt; lateral sepals elliptic-oblong, 16-18 mm. long, 7-8 mm. wide; lower sepals slightly narrower; petals whitish, the upper lamina strongly oblique, 10 mm. by 4 mm., glabrous, subemarginate, the spur ca. 13 mm. long; lower laminae round-ovate, 6 mm. long, somewhat bearded, short-ciliate, bifid about 1/3 its length into rather broad lobes; claw 6 mm, long; stamens 5-6 mm. long; anthers rather dark, roundish, less than 1 mm. long; carpels 3, strigose. Caules tenues, 4-10 dm. alti, supra laxe ramosi, disperso-foliosi, praeter strigulosos apices inflorescentiae glabri; petiola basalia, 5-10 cm. longa; laminae ad 1 dm. longae, 5-fidae, subglabrae aut parce strigulosae, ciliatae, segmentis dissectis, laciniis 1-2 mm. latis; inflorescentia laxa, racemi laxi, ongi; bracteae lineares (aut inferiores divisae), 5-12 mm. longae; pedicelli divergentes, 1.5-5 (—10) cm. longi, plus minusve strig- ulosi; bracteolae lineares, 2-5 mm. longae; sepala albida, subglabra; sepalum superiore oblongo-ovatum, 14-18 mm. longum, 8 mm. latum, apiculatum, calcar extensum, 18-21 mm. longum; sepala lateralia et inferioria elliptico-oblonga, 16-18 mm. longa; petala albida, lamina su- perior obliqua, 10 mm. longa, glabra, subemarginata; lamina inferior late 542 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ovata, hirsuta, ciliata, bifida, unguis 6 mm. longus; stamina 6 mm. longa; carpella 3, strigosa. Type: Nepal, near Gilam, Tila Valley, 1900 m., April 19, 1952, mainly near cultivated terraces, O. Polunin, W. R. Sykes & L. H. J. Williams 1907 (BM). Another collection is from Nepal, Dillihot, 2600 m., banks of cultivated fields, Polunin, Sykes & Williams 3916 (pm). The proposed species, Delphinium williamsii, is near D. kamaonense, but with white, not blue, flowers and a longer spur. It is a pleasure to dedicate it to Mr. L. H. J. Williams of the Department of Botany, British Museum (Natural His- tory). 126. Delphinium wilsonii Munz, sp. nov. Fic. 11, J. Root woody; stem terete, slender, 6-8 dm. tall, few leaved in the lower part, glabrous except for a few long hairs near the base; basal petioles 5-12 cm. long, spreading-villous, the cauline shorter, not vaginate; leaf blades rounded pentagonal, 3-7 cm. in diameter, with a few slender hairs along the veins of the under surface, 5-fid to near the base into rhombic segments, each rather deeply lobed into few lance-oblong laciniae 1—2.5 mm. wide; cauline leaves 2-3, reduced; inflorescence rather simple, race- mose, few flowered, to ca. 1.5 dm. long, glabrous except for the loosely strigose upper parts of the pedicels; bracts linear, 4-6 mm. long; pedicels suberect, 1-2 cm. long; bracteoles subulate, 2-3 mm. long, near middle of pedicel; flowers blue, the sepals strigulose; upper sepal elliptic-ovate, 13-14 mm. long, obtuse, the spur almost straight, 20 mm. long, 3.5 mm wide at base; other sepals round-elliptic, 13-14 mm. long; petals bluish, the upper laminae oblique, subacute, 9 mm. long, glabrous, the spur 18— 19 mm. long; lower petal laminae oblong-obovate, 7 mm. long, with 2 oblong lobes, hairy near the base, ciliate; the claw 6 mm. long; stamens glabrous; carpels 3, slightly hairy at base of styles. Caulis teres, tenuis, 6-8 dm. altus, infra paucifolius, subglaber; petiola basalia, 5—12 cm. longa, villosa; laminae foliorum orbiculare-pen- tagonales, 3-7 cm. longae, infra in venis sparsim villosae, 5-fissae, laciniis 1-2.5 mm. latis; folia caulium 2-3; inflorescentia subsimplex, racemosa, paucis floribus; bracteae lineares, 4-6 mm. longae; pedicelli suberecti, 1-2 cm. longi; bracteolae subulatae, 2-3 mm. longae, in medio pedicelli; sepala azurea, strigulosa; sepalum superiore elliptico-ovatum, 13-14 mm longum, obtusum, calcar horizontale, 20 mm. longum, base 3.5 mm. latum; sepala lateralia et inferioria orbiculari-elliptica, 13-14 mm. longa; petala subazurea, laminae superiores obliquae, subacutae, 9 mm. longae, glabrae; laminae inferiores oblongo-obovatae 7 mm. longae, bilobatae, base hirsutae, ciliosae, unguis 6 mm. longus; stamina glabra; folliculi 3, apice subhirsuti. Type: W. Hupeh, central China, July, 1901, E. H. Wilson 2596 (pos- sibly 2396) (Ny). Delphinium wilsonii keys to near D. tenii and D. 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 543 dolichocentroides; but it has strigulose rather than glabrous pedicels and comes from Hupeh not Szechwan or Yunnan. 127. Delphinium yuanum Chen, Bull. Fan Mem. Inst. Biol. Peiping (n.s.) 1: 176. 1948. Fic. 11, K. Stem 10-13 dm. tall, 4-5 mm. thick, glabrous and somewhat glaucous branched above; lower petiole to 2 or 3 dm. long, somewhat dilated at base, glabrous or with a few spreading hairs; blades pentagonal-orbicular in outline, 5-15 cm. in diameter, 5-fid to base, subglabrous or with scat- tered hairs and some very short appressed ones, the 2 lower divisions often again deeply parted to make 5 or 7 segments, these cuneate-obovate, the lower leaves with 3 or more blunt oblong lobes 3—5 mm. wide, the upper and middle 3-incised, then again divided into oblong-linear lobes 2-3 mm. wide, subobtuse, mucronulate; inflorescence simple and lax or few- branched, the flowers ca. 15-30; bracts foliaceous or mostly entire, linear, 8-15 mm. long, somewhat hairy; pedicels ascending, more or less arcuate, 1-5 cm. long; bracteoles linear, hairy, 3-6 mm. long, situated somewhat below the flower; sepals blue, white-pubescent with short appressed and somewhat longer, spreading, white and yellow hairs; upper sepal ovate, 12 mm. by 8 mm., obtuse, the spur spreading subcylindric, 14 mm. long, 3 mm. wide at base, obtuse; lateral sepals elliptic, 11 mm. by 6 mm., rounded at apex, the pubescence mostly on the midrib; lower sepals ob- long-obovate, 12 mm. by 5 mm., rounded at apex; petals blue, the upper laminae quite oblique, broadened upward, glabrous, 6 mm. long, the spur 12 mm. long; lower laminae oblique, yellow bearded near base, sparsely white-hairy and ciliate beyond, 5 mm. by 4 mm., oblong-obovate, scarcely to shallowly bilobed, dilated below, glabrous, gradually narrowed into a broad claw 4 mm. long; stamens 6 mm. long, glabrous; anthers dark, oblong, 1 mm. long; follicles 3, subglabrous. Type: N. W. Yunnan, Chungtien, 3000 m., T. T. Yu 12490, July 27, 1937 (PE), not seen; isotype (A) and photograph of type (E), seen. I have seen three collections from Yunnan at the Arnold Arboretum, which seem to me to belong here: Atuntze, 3400 m., T. T. Yu 10547; Muli, Wachin, 4000 m., T. T. Vii 14613; Chungtien, 3300 m., T. Vu 13852. 128. Delphinium yunnanense (F ranchet) Franchet, Bull. Soc. Philom. Paris VIII. 5: 173. 1895. Fic. 11, L. Delphinium denudatum var. yunnanense Franchet, Plant. Delav. 25. 1886. ? Delphinium yunnanense vars. laxum and strictum, Huth, Bot. Jahrb. 20: 465. 1895. No specimens cited. Delphinium esquirolii Lévl. & Vaniot, Bull. Herb. Boiss. Il. 6: 509. 1906. J have seen Esquirol 744 (£) from Lévl. Herb. as D. esquirolii; Esquirol 794 or 793? (gE) as D. esquirolii sp. nov.; and Esquirol 190 (£). Stem simple or branched, 2.5—7.5 dm. tall, more or less retrorse-strig- 544 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ulose below, subglabrous above, terete; leaves light green, few and remote, finely and thinly pubescent, the petioles slender, mostly 4-8 cm. long, strigulose, dilated at the base, blades more or less round-ovate in outline, 4-12 cm. wide, pentagonal with usually 5 segments to near the base, these cuneate and divaricately and coarsely sharp toothed in the lower leaves or lobed and then toothed with narrow divisions 1-3 mm. wide in the cauline leaves; upper leaves few lobed; flowers few in lax racemes or these in an open panicle, the lowest bracts sometimes foliose, dissected, the upper setaceous, 5-15 mm. long; pedicels mostly glabrous, ascending or erect, 2-5 cm. long; bracteoles opposite, setaceous, 2-5 mm. long, a little above the middle of the pedicel; sepals deep blue, finely strigulose, the upper sepal ovate, 12 mm. long, subacute, the spur horizontal, 20-22 mm. long, ca. 2 mm. wide at the base, slender, narrowed and straight to the fine tip; lateral sepals 10-12 mm. long, strigulose on midrib, rounded at apex, broadly ovate; lower sepals more generally strigulose, about as large; upper petals almost straight, the limb 7.5 mm. long, notched at apex, the slender spur 22 mm. long; lower petals bluish, the blade sparsely bearded except at base, long-ciliate, oblong-obovate, 5.5 mm. long, divided for about 2 mm., the claw 4 mm. long; stamens 5-6 mm. long, the fila- ments somewhat ciliate; anthers dark, 1 mm. long; follicles 3, subglabrous to strigose; seeds obtusely obpyramidal, transversely wrinkled, with straight wings on the angles. TYPE: Yunnan, in the plain of Mosoyn, Delavay 967 bis, Aug. 31, 1886 (p), seen; isotypes (K, NY), seen; photographs of type (BH, FE, UC), seen. DISTRIBUTION. The species is found at about 1800 to 3800 m., Yunnan. ExampLes. Forrest 15536, 2977, 7216; Henry 13440, 9272; Maire 389, 1196; Ducloux 7329, 370, 5866; Delavay 9, 6515, 6670, 5866; Simeon Ten pis Bae a ie Yu 17375; Martin & Hodder 1789; Rock 6551, 6902; Ching 21657; Feng 89. Huth (Bot. Jahrb. 20: 465. 1890) proposes vars. laxum and strictum without typification, so one does not know what he had in mind. NOMEN INCERTUM 129. Delphinium lepidum Fisch. & Avé-Lall. Ind. Sem. Horti Petrop. Judging from the description which speaks of narrow segments in the leaves and which compares the flowers with those of D. grandiflorum L.., this plant would fall in my Group VI. It has been compared to D. ériste, which does not reach nearly so far west as the Caucasus, whence D. /epi- dum was described. I have seen no type specimen and cannot place it at all. Herewith is a translation of the original description Leaves 3-sect, the segments divided into entire linear lobes; spur longer than the calyx; lamina of upper petals suboval, subentire, 2- toothed above the middle, long-pilose on both sides and on margin; seeds not lamellate.— Plant showy, 4.5 dm. Stem short-pubescent. Lower 1967 | MUNZ, ASIAN SPECIES OF DELPHINIUM 545 pedicels much longer than bract and flower. Calyx deep blue, subglabrous. Lower petals entire at tip, anterior shallowly bifid, with broad cleft. Flowers larger than in D. grandiflorum and with suborbicular entire lam- ina of upper petals, short-ciliate, short-pubescent beneath, bearded to base above. Caucasus at Kobi. Perennial. Another name about which I am quite uncertain is Delphinium cheilan- thum subsp. schizophyllum Brihl in Brihl & King, Ann. Roy. Gard. Calc. 5: 99. 1896. It was based on a collection by Col. Davidson on the China hill near Naini Tal, Kumaun. I have seen no material by Davidson and can only guess by the name “‘schizophyllum” that the plant may belong in Group VI. | To be continued | 546 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 THE DIRECTOR’S REPORT Tue ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JUNE 30, 1967 THE YEAR 1967 MARKs the ninety-fifth anniversary of the establishment of the Arnold Arboretum in 1872 by the bequest of James Arnold. Early in the year official greetings were received from several institutions with similar interests around the world. Rather than take particular note of this ninety-fifth anniversary, permission was requested from the Harvard Corporation to prepare special activities for the Centennial year, 1972. Our plans, therefore, are directed toward that year when we may indicate the many accomplishments of the past and make known the goals for the future. For nearly a decade a legal controversy over the Arnold Arboretum had been conducted in the courts of Massachusetts. This proceeding, which involved the President and Fellows of Harvard College as trustees of the Arnold Arboretum, concerned the legal right and questioned the decision affirming the scientific importance of locating a portion of the herbarium, library, and other research collections and the staff in a building in Cam- bridge, occupied jointly by members of the Gray Herbarium and the Botanical Museum. A “working herbarium and library” was to be re- tained in Jamaica Plain in association with the living collections. During the preceding fiscal year the Supreme Judicial Court of Massachusetts dismissed the case subject only to the assurance in a report to be filed, indicating that an adequate library and herbarium do exist in Jamaica Plain; that the location of all books is indicated in the general index files; and that a plaque has been placed in the Cambridge building housing the Arboretum materials. A report indicating that these conditions have been met was accepted by Justice Spaulding for the Court and the case finally has been dismissed. In association with this action books belonging to the Arnold Arboretum and deposited in the libraries of the Department of Biology and the Harvard Forest since 1933 were recalled and incorporated with the books of the Arnold Arboretum in Cambridge or Jamaica Plain. The President and Fellows also approved altering the name of the herbarium building in Cambridge to the Harvard University Herbaria, the plural form being the equivalent of “Laboratories,” to indicate the multiple occupancy of the building. A bronze plaque was cast and placed in the foyer of the building listing the several independent collections housed in it, and the date of its completion, 1954. In 1965, the Harvard Corporation approved the establishment of an Institute of Plant Sciences at Harvard to represent the botanical staff of 1967 | THE DIRECTOR’S REPORT 547 as HARVARD UNIVERSITY Photograph of the building newly designatec HER 548 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 RDUNIVERSITY ARIA ree 2 AND LIBRARY BARIUM BOTA ED 1S The plaque recently installed in the building of the Harvarp UNIVERSITY HERBARIA, the University, the Institute to consist of and represent the staff members of the several botanical organizations at Harvard as well as botanical scientists associated with biological organizations. During the past year the Arnold Arboretum was invited to join in association as part of the Institute of Plant Sciences. This invitation was accepted and membership has been approved by the Corporation. The Corporation also noted that its vote of January 19, 1953, concerning the identity of the Arnold Arbo- retum is still in effect, and that the procedures for the Director of the 1967 | THE DIRECTOR’S REPORT 549 Arboretum to submit recommendations for appointments and for approval of budgets directly to the President are still in order. Staff: Although changes in the composition of the staff of an organization are to be expected annually it is always with personal feelings of loss that deaths and retirements of employees of long association are recorded. Professor Irving Widmer Bailey, Professor of Plant Anatomy, Emeritus, died suddenly on May 16th. Although stricken some weeks earlier with a coronary attack, his recovery had seemed probable when death came quickly. Professor Bailey received his A.B. degree from Harvard in 1907, and, except for a few years during World War I, was associated with Harvard continuously. He became associated with the Arnold Arboretum in 1933 when many members of the Bussey Institution staff moved to Cambridge, and he retired officially in 1955. He continued his research in the laboratory of the wood collection he developed for the Arnold Arboretum, and he worked and published on an active schedule to the day before his attack. Tributes to him and for his work have come from all parts of the world. A review of his life will be published in the next volume of this Journal. Thomas H. Curry, assistant superintendent of buildings and grounds of the Arnold Arboretum, served its living collections in Jamaica Plain for 18 years. He had reached retirement age and would have been retired at the end of the fiscal year. His death occurred, after surgery, on April 30, 1967. A new appointment during the year was necessary in anticipation of another retirement. Mr. Victor Ferenc Marx was appointed librarian of the Arnold Arboretum and, jointly, librarian of the Gray Herbarium on May 1, 1967. This appointment overlapped briefly that of Mrs. Lazella Schwarten. Mr. Marx came to Harvard from the University of Washing- ton School of Librarianship. Mrs. Schwarten served the library of the Arnold Arboretum for 23 years and, for the last 13 years, held a joint position as head librarian of the Gray Herbarium and the Arnold Arboretum. She not only served the staff and cared for a major library, but also aided taxonomic botanists, botanical historians, and many others in the field from institutions all over the world. She excelled in all these activities, and her aid has been ac- knowledged with spontaneous gratitude. Although her efforts in preparing the Index to American Botanical Literature published by the Torrey Botanical Club are noted only as items in her bibliography published each year in these reports of the Director, this index is a major tool in tax- onomic and general botanical bibliographic research. Her bibliographies of the published writings of many present and former staff members are also most useful. Mrs. Schwarten’s greatest effort in the past decade how- ever. has been the organization of the books of the Gray Herbarium and the Arnold Arboretum in Cambridge, and those of the Arnold Arboretum 550 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 relocated in Jamaica Plain, into working libraries. Only the staff that saw this develop can truly appreciate the tremendous effort involved. Finally, we would note that Lazella Schwarten served the botanical world through the Journal of the Arnold Arboretum. Articles by her are not numerous, but the word “Circulation” on the cover is also inadequate to describe her work in making certain that botanists around the world received the Journal. When the University announced Mrs. Schwarten’s retirement, to take place this year, the associated staffs scheduled an appropriate gathering. A gift from the Arnold Arboretum and Gray Herbarium recognized her service to both institutions. In addition a bound volume of letters solicited from her friends who had used the library recorded gratitude from many botanists of numerous countries. The July number of this Journal was also dedicated to Lazella Schwarten. Regretfully, this report of the Director records her retirement from the staff. Dr. Harrison Leigh Flint whose appointment as Associate Horticulturist was effective on July 1, 1966, for five years, came to our staff from the University of Vermont. Gordon Parker DeWolf, Jr., was appointed Horticultural Plant Taxon- omist, effective March 15, 1967, for a period of five years. Dr. DeWolf received his Ph. D. degree from Cambridge University, having worked under Dr. E. J. H. Corner on the genus Ficus in tropical America. He served as an Associate Professor of Botany at Georgia Southern College before joining the Arboretum staff. His research interests in cultivated plants and native floras are supported by his field work and study in such areas as New England, Malaysia, Europe, and Africa. Two Mercer Fellows completed their studies during the year. Mr. Pablo Legname returned to his position at the University of Tucuman in Argentina and Miss Sandra Shannon accepted a position at the Waltham Field Station of the University of Massachusetts. Dr. Charles C. Tseng completed his studies of the flowers and fruits of the Araliaceae conducted during the summer as a Mercer Research Fellow, and returned to his position at Windham College in Vermont. Several new assignments were undertaken by the staff. Dr. Richard Howard was appointed to the Commission on Horticultural Nomenclature and Registration for the International Society for Horticultural Science. Dr. Lorin Nevling was appointed a member of the Faculty of Arts and Sciences of Harvard University. He was also appointed Secretary of the American Society of Plant Taxonomists and designated as a representative of the Council of that organization to the American Association for the Advancement of Science, to serve also as Secretary of the Systematic Section of the Botanical Society of America. Dr. Bernice Schubert was appointed Secretary of the Standing Committee on Stabilization of the International Association of Plant Taxonomists, Dr. Carroll Wood was appointed Lecturer on Biology during the year 1966-67, while he was teaching a course in plant taxonomy. 1967 | THE DIRECTOR’S REPORT 551 Horticulture: Recent reports have referred to the drought of the last five years. The rainfall for 1966 was 41.14 inches, 5.55 inches below the normal for Boston. During this period the three conspicuous ponds in the Arboretum were at their lowest since 1937. Watering with metered water was con- tinued through the summer of 1966. The plastic piping which had been run to the top of Bussey Hill made it possible to water and probably to save the large collection of leguminous and ericaceous shrubs there. The dry spell was broken during the winter, however, and between Jan- uary 1 and May 31, 1967, approximately 25.6 inches of rain were recorded at the Arboretum. This was an excess of 5.02 inches over the normal. A total of 7.32 inches in May made it the second wettest May in 97 years of official U. S. Weather Bureau records. The wetness was accompanied by unusual cold so that the common lilacs did not come into full bloom until June 3, the latest flowering time in the 36 years we have kept such records. The conditions of fall drought, and excess water and low temperatures in the spring again altered our planting schedule. Plants could not be dug in the nurseries in Weston nor cared for in many areas on the grounds in Jamaica Plain. Eventually, however, 570 plants representing 480 taxa were added to the collections. Most significant was the planting of 50 Rhododendron fortunei selections which originated on the Dexter Estate in East Sandwich, Massachusetts. Dr. Wyman has served on a committee to collect and evaluate these widely distributed “Dexter Hybrids” of which the most desirable are now planted on Hemlock Hill. Regular embossed labels identify them as completely as possible and include the location of the original plant. A severe ice storm again occurred in New England in February causing considerable branch damage. In the Thuja and Chamaecyparis collection much pruning and repair was needed and duplicate, old, and broken plants were removed to give the healthy specimens additional space. Work continues on the terracing of the slope below the Larz Anderson bonsai house near the Dana Greenhouses. Field stone walls, graded paths and steps made of railroad ties have enhanced the appearance of the area where more than 125 taxa of dwarf conifers have been planted and mulched with gray stone. The old Centre Street roadbed in the northwest corner of the Arboretum was excavated to an average depth of two feet. Large quantities of mulch and any available soil were used to fill the road bed. Grading was hardly completed when snow stopped further work: seeding and some planting were delayed until spring. The cost of reclaiming this old road area amounted to $7,000 for contract labor with heavy machinery. During the winter it was possible to begin a clearing and cleaning operation on the low land adjacent to the Hebrew Rehabilitation Center for the Aged. This hospital was built on land formerly called Joyce Kilmer Park, acquired from the City of Boston. It is adjacent to land owned by Harvard University for the Arnold Arboretum and not under lease from un V1 bo JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 views of the terrace for genetic dwarf conifers below the Larz Anderson head house. The wall is of native field stone. Gray stone is used for mulc ch. 1967 | THE DIRECTOR’S REPORT 553 the City. The unkempt low land served as a barrier during the construc- tion of the hospital but the present operations anticipate development of the area as a formal part of the Arboretum. A major expenditure will be required for fencing, development of roads and paths, and redevelopment of top soil before the land can be fully utilized. The threat to the Peters’ Hill plantings from the bill filed in the Legis- lature to create a ski tow in the Arboretum required action this year to control access in the winter. The area is normally not ploughed by the City and our own equipment is inadequate for the purpose. A chain link fence was installed for half a mile along Bussey Street giving immediate protection to the half of the area most readily approached by automobile. The police were requested to enforce the regulations of the Park Depart- ment regarding skiing. Eventually, additional expenditures for fencing may be required to prevent skiing on the hill and to protect the younger plantings. During the past year air pollution legislation prohibiting open fires was adopted in Boston. Withholding of permission to burn debris and trash from the grounds has proven to be extremely inconvenient and the added expense of commercial waste-disposal is inevitable. Previous reports have referred to changes proposed for the buildings in Jamaica Plain if legal directives did not require the return of materials from Cambridge. During the year the first of these changes was com- pleted by the addition of a glass unit consisting of three controlled- environment sections to the Dana Greenhouses. In addition, air condition- ing was installed in the laboratory for cytology and morphology, new lighting was installed in the library of the Administration Building and safety flood lights added on to the front and rear of the building for auto- matic operation during the winter. New lunch room and additional rest room facilities have been completed. Plans have been developed for the construction of a new garage and vehicle repair facility which will stand free yet close to the Administration Building, permitting the removal of motor vehicles and grounds equipment from the basement. Architects’ drawings are being prepared prior to a request for bids on construction. Also under construction is a laboratory for plant hardiness studies, includ- ing those in cryobiology, in the basement of the Dana Greenhouses. A “Gro-Mor” greenhouse of polyethylene plastic over aluminum frames 11 < 96 feet has been erected in the greenhouse area. A rearrangement of the nursery area has been completed and includes the installation of new ground drainage, watering facilities, and gravel surfaces for the handling of plants grown in plastic containers. Slightly over 1000 square feet are devoted to this purpose. Through the efforts of Dr. Flint a new codperative hardiness testing program has been initiated to obtain more reliable information on the characteristics of some of the more important woody ornamental plants in parts of New England colder than the Boston area. This program is to be carried out with the assistance of about 25 codperators selected as uniformly as possible in hardiness zones 3 and 4. Pot-grown plants of 554 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 ee. Tg ABo unit on ee fiche was newly constructed i BELow: The “Gro-Mor’’ greenhouse of ale ae plastic on aluminum frames and the compacted fabs area for the pot grown plants used experi- mentally for hardiness tests at the Arnold Arboretum. 1967 | THE DIRECTOR’S REPORT 555 about 20 taxa will be distributed annually for trial. At the same time ob- servations will be made on plants already under test. To date 19 cooper- ators have accepted invitations to participate. A limited number of plants was distributed in the Spring of 1967 and the program is expected to reach full scale in 1968. More than 40 taxa have been propagated for distribu- tion in 1967, 1968, and 1969. Mr. Fordham has continued his long-range program of testing seed germination and propagation procedures for the plants within the Arbore- tum collections. Ninety-six additional taxa were handled during the year for such information. The search for abnormal plants in natural habitats also continues; eleven seedling populations and 16 asexual propagations from witches’-brooms are presently under study. Seedling plants developed from seeds of abnormal conifers have been followed for several years and the accumulating records now begin to shed light on the nature and cause of the unusual growth forms. : Three members of the horticultural staff, Dr. Wyman, Mr. Fordham, and Mr. Gensel, spent some time in Puerto Rico conducting studies and observations on the growth and reproductive characteristics of plants in the elfin forest project-area. Plant materials were propagated on the site and in nearby areas while comparable materials have been returned to Jamaica Plain for similar tests under more scientifically controlled condi- tions. The past winter again produced unusual flowering patterns in many shrubs because of snow protection near the ground and varying levels of cold air above. Mr. Gensel, working with Dr. Flint, has undertaken con- trolled studies of cold resistance through flower bud development in several taxa grown in the Arboretum. The variations which occurred in taxa of Forsythia and Rhododendron under natural conditions during the past winter require much further study before an explanation is possible. At the Maryland meetings of the American Association of Botanical Gardens and Arboreta the Arnold Arboretum was again appointed National Registration Center for woody plants not otherwise represented by So- cieties. The International Society for Horticultural Sciences approved the designation of the Arnold Arboretum as International Registration Author- ity for cultivars in Weigela and accepted the list published in Arnoldia as the International Registration List. At a special meeting of the IABG Dr. Howard was appointed funding codrdinator for a project to test the methods of electronic data processing in relation to the record systems of botanical gardens. The Arboretum will work in codperation with the University of Tennessee Arboretum where this program will be developed, using the computer systems at Oak Ridge. Much of the time of the propagation staff is devoted to filling requests for plant parts for propagation or as the basis for special research studies. During the year 66 taxa were propagated for such special requests as part of 184 shipments of material to other gardens and to scientists in 17 coun- tries. The propagation staff received 231 shipments of materials from 32 countries comprising 845 taxa. We supplied healthy materials for other 556 JOURNAL OF THE ARNOLD ARBORETUM [VoL. 48 gardens to grow, aberrant material for special developmental studies, diseased material for pathological investigation, foliage for animal food in special research projects, and dried materials for chemical extractions. Case Estates: With the increasingly large number of people who visit the Case Estates we have had a growing number of requests for literature, information, or guided tours of the land and the collections. During the past year two regularly scheduled field courses were conducted on the grounds, and two special programs for the Friends of the Arnold Arboretum, consisting of a series of evening lectures in the “Red Schoolhouse” and a series of Sunday afternoon lecture-tours on the grounds, were held. Special encouragement is given to students and to school classes to visit the grounds and to learn the purpose of the work being done there. Awards for study projects in the elementary school system, initiated this year, should be continued. The students involved have brought their parents to the grounds and appreciation increases for the efforts of the Arboretum within the town of Weston. During the year a trial bed of 70 taxa of Allium was started to contrast with those of Narcissus, Lilium, and Hemerocallis. Most of the Allium material was grown from seed and much was a gift of Mr. Nathaniel Whittier. The Jris collection previously established for trial was supple- mented by a gift of representatives of the Dykes medal winners which show, for teaching purposes, some of the outstanding variation in the bearded irises. A collection of 79 kinds of native asters of known origin and genetic structure was established on the Case Estates; a section of land was pre- pared for the studies of a new staff member on the species of Dahlia. A collection of crab apples established several years ago in the big field on Newton Street across from the Junior High School flowered spectacu- larly this spring. The cultivar of Malus ‘Henrietta Crosby’ developed by Dr. Sax and planted along Wellesley Street was also unusually beautiful. The isolated location of the Case Estates permits use of the property as a holding area for plants in post-entry quarantine. Almost 500 of the plants introduced by Dr. Wyman from his recent trip to Europe were released during the year and are growing on in the nursery areas. Others remain under continued restriction for future release. Herbarium: During the year 24,264 specimens were mounted and added to the herbarium, bringing the total number of specimens to 849,269. Of these, 5,861 were added to the herbarium of cultivated plants which is main- tained in Jamaica Plain. That collection now contains 130,531 specimens. In order to correct some of the unbalance of content in the herbarium of cultivated plants a program of special collecting and special exchanges has been emphasized recently. It is hoped that the result will be an in- 1967 | THE DIRECTOR’S REPORT 557 creased representation of cultivated herbaceous plants and a_ better indication of the range of distribution of many cultivated woody plants. Such development will increase the significance of this particular collec- tion and broaden its function of providing definitive information. The collection in Jamaica Plain is also being checked to insure that all the taxa cited in Rehder’s Manual of cultivated trees and shrubs and in Bailey’s Manual of cultivated plants are, in fact, represented in the herbarium of cultivated plants. A total of 18,863 specimens was received for the herbarium during the year. Of these, 8,791 were in exchange, 3,627 were gifts, 5,388 obtained through subsidy or support of field expeditions, 411 for identification, and 116 by special exchange, with 350 representing staff collections. This material came from herbaria or collectors in the United States, Papuasia, Europe, Western Malaysia, and Mexico and the West Indies in diminish- ing order, with specimens from seven other geographic areas also. We sent 10,675 specimens in exchange to other institutions; 899 as gifts; and 70 in exchange for identification. For studies by staff and students a total of 76 loans were requested and received from 26 institutions, 14 in the New World and 12 in the Old World. The total of 7,152 specimens represented an average loan of 94 specimens. We received requests for material from our collections and made 143 loans to 78 institutions which averaged 115 specimens and com- prised a total of 16,539 specimens. These were made to 28 foreign institu- tions and to 45 within the United States. These series of figures do not indicate the great amount of organization and labor involved in maintaining high standards of quality in the her- barium and in keeping the collections in good order so that the many routine requests for information, photographs, identifications and so forth may be filled competently. Although the growth of the herbarium and the increase in its use is desirable and commendable, such growth continues to present problems of space in light of which plans for future additions to the collections and the buildings must be considered seriously. The use of the herbarium comprises the basis for diverse studies in the field of botany and may result in monographic work of wide extent, floris- tic studies, manuals of wild or cultivated plants of extensive or limited geographic areas, and, as a complement to field studies, may produce most valuable data for work on breeding programs, chemical analyses, and various other avenues of research. In short, the projects of the staff often develop into studies of broader aspect than is immediately apparent. Brief statements concerning some of the current work follow. Mr. Baranov completed his work on the variations in Sorbus and con- tinues studies on plants of temperate Asia. United States. His studies, however, often involve broader geographic areas or plant relationships. An example is his valuable contribution on the nomenclature of cotton, Gossypium, which forms part of his continu- ing study of the Malvaceae. 558 JOURNAL OF THE ARNOLD ARBORETUM [VvoL. 48 Dr. DeWolf, who has just joined the staff, has had his work interrupted by the change in location. His research interests are in the genera Ficus and Dorstenia of the Moraceae with special emphasis on the Afro-Amer- ican relationships involved. Dr. Hartley completed his study of Zanthoxylum in the Malesian area and turned to the genus Lunasia, also of the family Rutaceae. He has now named approximately 1,600 of his collections from New Guinea and in the process was able to annotate many tropical Asian specimens in the herbarium. Dr. Howard spent a portion of the summer continuing his field studies on the elfin thickets or mossy forest on the summit of Pico del Oeste in the Luquillo Mountains of Puerto Rico. This work, which comprises a study of what is in the forest, the conditions under which the plants live, and a day by day account of what the forest is doing, represents one of the comprehensive ecological studies being conducted in the tropics. Other staff members have contributed their special talents to the study area including Miss Powell, Messrs. Fordham and Gensel, Grime and Canoso, and Drs. Wyman, Nevling, and Wood. The duration of the study has been extended to August, 1968, under the same grant from the Na- tional Science Foundation. Publication of the program and the special studies contributing to it should begin in the coming winter. In addition, Dr. Howard continues his studies of the vegetation of the Lesser Antilles with the identification of special collections from the area and the compila- tion of family treatments. Work has also started on a bibliography of the plants under cultivation in the American tropics. Lack of readily available references to the appropriate valid names and synonyms of tropical plants has handicapped progress in areas of science well beyond plant taxonomy. A bibliography will partially solve the problem of correct name and the authority. This is expected to be a long term project which may attract both workers and independent financial support. Dr. Hu has continued her studies on plants from Asia and has prepared special contributions based on specimens of the Araceae borrowed from European herbaria at Paris and Copenhagen. She has made progress in the study of cultivars of Hemerocallis and in the preparation of special chapters in a forthcoming Hemerocallis handbook. A collection of plants from Mt. Omei, southwest China, made in 1938-40 by collectors for the Arnold Arboretum, has now been mounted and Dr. Hu is beginning the task of naming and inserting these specimens. Mrs. Honor Moore of Canberra, Australia, has been appointed a Re- search Associate during the period her husband is in residence in the area. She is annotating our herbarium material from Australia and indicating the type specimens. The activities of the herbarium have claimed a large proportion of the time and effort of Dr. Nevling during the year. Nevertheless, his research on Schoenobiblus and Schizophragma, and on the Thymeleaceae advances. Dr. Nevling is also codperating with botanists from the Instituto de Biolo- gia. in Mexico, on studies of the flora of Veracruz. Approximately 1,400 1967] THE DIRECTOR’S REPORT 559 specimens including duplicates were received during the year representing the initial effort of collectors in the area. Dr. Nevling will spend some time in this region during the summer. The Arnold Arboretum has agreed to subsidize a portion of the collecting in collaboration with the Jardin Botanico of the University of Mexico, and to identify the cultivated plants of the area. Specimens from this project will be available for distribution to other institutions and specialists willing to identify material represent- ing particular families or genera from Veracruz are being sought. Miss Dulcie Powell is revising for publication a study of the plants transported by the expeditions of Captain William Bligh and is continuing her regular studies of cultivated and native plants of tropical America and the West Indies. Dr. Schubert has handled an unusually large amount of manuscript in her capacity as editor of the Journal of the Arnold Arboretum. Her studies on Desmodium are currently focused on the preparation of a treatment for the Flora of Tropical East Africa. Additional studies and identifica- tions are continuing in the genera Dioscorea and Begonia. As Secretary of the Standing Committee on Stabilization of specific names of plants for the International Association of Plant Taxonomists she has the respon- sibility of codrdinating the opinions of committee members, specialists, and other interested botanists in preparation for the report to be prepared for the next International Botanical Congress. Dr. Wood continues to codrdinate the effort of the small dedicated group of people working on the generic flora of the southeastern United States. With the resignation of Arnold Clapman, who worked as artist on the project, Miss Rachel Wheeler has been appointed staff artist. One of Dr. Wood’s tasks is to keep the presentation of visual artistic material and written material as similar as possible to that previously published in the Journal. Dr. Grady Webster, now of the University of California at Davis, spent a portion of the summer of 1966 with Dr. Wood and prepared the treatment of the family Euphorbiaceae in the Southeastern United States. As indicated elsewhere, a new application submitted to the Na- tional Science Foundation for renewed support of this work has been approved. During the spring semester Dr. Wood taught an introductory course in the Taxonomy of Flowering Plants and continued his ever generous aid to students. Library: The end of the fiscal year marked the retirement of Mrs. Lazella Schwarten as Librarian of the Arnold Arboretum, a post she has held for nearly 23 years. During that time the library has grown from the 59,440 books and pamphlets recorded in the Director’s Report of 1945 to a present total of 53,784 bound volumes and 20,879 indexed pamphlets for a total of 74,663. The high quality and good condition of the library, plus the ready availability of the books through the willingness of the librarian and her staff to assist scholars, have made the library of the Arnold 560 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Arboretum outstanding among international botanical libraries. During the past year 488 bound volumes were added to the collections of which 138 were in areas of horticulture and were deposited in Jamaica Plain. Pamphlets numbering 285 were of value and were indexed and added to the collection. Three additional major changes made during the past fiscal year are not included in the figures cited above. Bound volumes numbering about 600 on temporary deposit in the library of the Biological Laboratories since the 1930’s were returned as were over 2,000 books and volumes of periodicals on the subject of forestry from the Harvard Forest. These volumes have always been included in the total count given in annual reports. Their deposit away from the main library of the Arnold Arbo- retum reflects the crowded conditions existing in Jamaica Plain prior to the construction of the Harvard University Herbaria building and the move to Cambridge. The relocation of these books necessitated changes on all file cards and, in the case of the forestry books, some recataloguing which is in progress. All cards in the general and combined library index of the Arnold Arboretum and the Gray Herbarium now are marked to indicate ownership of the book and its location, by code number and by letters on shelves in Cambridge or in Jamaica Plain. Index cards, as issued, were added to the Gray Herbarium Card Index to American Plants, the Torrey Index to American Botanical Literature, and the Index Nominum Genericorum. Typed cards exceeding 1,200 were added to the Rehder Index to horticultural literature. Under the direction of Miss Stephanne Sutton much work was ac- complished on the historical correspondence of the staff of the Arnold Arboretum. Letters of former staff members have been arranged chron- ologically in new files and organized in alphabetical order by correspon- dents. A master card file indicates the number of letters in each folder and their dates. This work continues. We are pleased to acknowledge here the gifts of Sargent letters as well as photocopies of originals which were loaned to us. Miss Sutton continues her work toward a biography of Charles Sprague Sargent. The problem of publication of quotations from letters to and from Sargent was solved in part, during the year, when permission was granted by the responsible heirs of Professor Sargent to quote from his correspondence. Mrs. George Slate has also given permission to quote from the letters of her father, the late E. H. Wilson, whose work was so closely connected with the development of the Arnold Arboretum. Comparative Morphology: For many years the contributions of the Arnold Arboretum in the area of comparative plant morphology have been primarily those of Professor Bailey and his students and assistants. Our loss is great with his death. Hopefully, we shall seek younger contributors to utilize the great collec- tion of wood samples and microscope slides of plant parts built up during 1967 | THE DIRECTOR’S REPORT 561 his many years of study. Professor Bailey had completed his investigations of the leaf-bearing cacti, a group which had long occupied him during summer work in Arizona. The final paper of the series will be published posthumously in this Journal. The Marcgraviaceae is a tropical family Professor Bailey knew from British Guiana. Upon completion of his cactus studies he turned to it and his excitement over the initial prepara- tions indicated that another study series would be developed. Regrettably these observations will be denied to us. Through the interest of Mr. Frank Hankins in the woods of Panama, we have been able to purchase microscope slides prepared from Panama woods previously unsectioned from our own collections and those of the Vale School of Forestry. In addition, approximately 150 slides added to the collection represent material received in exchange for wood samples. Education: Two regularly scheduled classes within the College of Arts and Sciences of the University were offered by Arboretum staff members. During the fall semester Dr. Howard taught Biology 209, “The Phylogeny of Flower- ing Plants,” and in the spring semester Dr. Wood taught Biology 103, “The Taxonomy of Vascular Plants.” Dr. Hartley taught the course in general botany offered by the Harvard University Commission on Exten- sion Courses. In addition, the staff assisted in the supervision of the research or writing of graduate students and participated, as speakers or as commentators, in the seminars open to undergraduates and graduate students. Non-credit courses open to the public were offered throughout the year in Weston and in Jamaica Plain as a contribution of the staff. Regular field classes were conducted by Drs. DeWolf, Wood, and Wyman, while, in addition, staff members Flint, Fordham, Howard, Nevling, Williams, and Wood also contributed to a special series of lectures and field trips held in Weston during the spring. Guided tours of the grounds of the Arnold Arboretum and special tours of the Case Estates were available to groups, on written request. These are most frequent in the spring. Because of the unusual weather this year many trips had to be repeated when flowering conditions became more satisfactory. Open Houses, in- cluding tours, were held for the Harvard Dames, the staff of the Depart- ment of Biology, and several national touring parties of horticulturists. The staff did not prepare an exhibit for the Spring Flower Show of the Massachusetts Horticultural Society but four staff members, Messrs. Flint, Fordham, and Williams, and Dr. Hu gave lectures and/or demon- strations. A special exhibit of flowering shrubs and trees was prepared for the Worcester County Horticultural Society to supplement its annual Tris show. A codperative program in training in ences has been conducted with Keuka C High School. Students from each of these sc horticultural and botanical sci- ollege and the Jamaica Plain hools have been given short 562 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 periods of employment under supervision in order that they may learn curatorial, laboratory, or greenhouse techniques from the Arboretum staff. At Weston Junior High School an Interdisciplinary Team Teaching Pro- gram this spring selected the Arnold Arboretum and the Case Estates as a study area. Miss Marian Case had encouraged such scholastic interest during the days of the Hillcrest School when she offered prizes for the best studies of Weston School pupils. The Interdisciplinary Team project offered a similar opportunity to reward student interest in the work of the Arboretum. Cash prizes were awarded to Gregory Morrow, Catherine Searle, and Bruce Howard for the three papers judged best by their teachers. Honorary mention awards, reproductions of the drawing ‘“Da- vidia,” by Blanche Ames, were made to Jon Conant, Thomas Griffin, Lois Leonhardt, Richard Powers, and Nancy Raffio. Staff members were guest lecturers to many groups during the year. Dr. Flint spoke to the Gardeners’ and Florists’ Club of Boston, the Connecticut Rhododendron Society, and at the University of New Hamp- shire. Mr. Fordham conducted several special demonstrations and tours of the Arboretum greenhouses for classes from colleges in four New Eng- land states and from the Niagara Parks Commission’s School of Horticul- ture. Dr. Howard spoke at Auburn Community College, Auburn, New York, on a visit sponsored by the American Institute of Biological Sci- ences. He was also a program speaker at the annual Fall Symposium of the Missouri Botanical Garden which considered the subject “Systematics and Natural Areas.” The Department of Botany at Miami University, Oxford, Ohio, sponsored his visit to that campus. During the year two programs were prepared for the School of Global Medicine of the U. S. Naval Medical School. These were ‘Tropical Flora of Medical Impor- tance” and “Edible Plants of the Tropics.” Both programs were recorded on closed circuit television in color and prepared for distribution to the Armed Forces as training tapes or 16 mm. film. Dr. Nevling was a guest lecturer to the Department of Botany of the University of Connecticut. Dr. Wyman was the principal speaker at the annual meetings of the American Rhododendron Society in Asheville, North Carolina, the Garden Clubs of Cincinnati, Ohio, and the Northeastern Section of the Interna- tional Shade Tree Conference in Portland, Maine. Again this year the staff assumed the responsibility for local arrange- ments in Cambridge for the midwinter meeting of the Northeast Section of the American Society for Horticultural Sciences. Travel and Exploration: The International Horticultural Congress and the annual meeting of the American Institute of Biological Sciences were held simultaneously on the campus of the University of Maryland in College Park. All staff members who could attend did so and many took part in meetings, served on committees, organized symposia or presented papers. | r. Wyman organized and conducted a section meeting on amateur 1967 | THE DIRECTOR’S REPORT 563 horticulture for the International Society of Horticultural Science. He was in charge of the presentation of awards at the banquet of the Amer- ican Horticultural Society. Dr. Nevling organized a symposium on the Diverse Applications of Plant Taxonomy for the American Society of Plant Taxonomists. Dr. Howard served as a representative at the meeting of the Commission on Nomenclature and Registration and pre- sented an invitational paper in a symposium on the “Nomenclature and Registration of Cultivars.” Dr. Flint attended the annual meetings of the Society for Cryobiology and with Mr. Fordham attended the annual meeting of the International Plant Propagators Society where Flint spoke on testing landscape plants for hardiness in Vermont, and Fordham reported on “Woody plants dif- ficult to root.” Miss Powell represented the Arboretum at the first annual meeting of the Caribbean Conservation Commission, held in Grenada. She explored the area of Morne Quaqua and later, a mountain on St. Kitts, to deter- mine the possibility of using these areas for further studies of elfin thickets. Material from each area was also returned to Boston by air mail to determine the possibilities of such a method of shipment for scientific studies of living specimens. Dr. Schubert attended the A.E.T.F.A.T. meetings held in Uppsala, Sweden. and before and after these meetings visited and worked in her- baria in Paris, Stockholm, Vienna, and at Kew and the British Museum (Nat. Hist.), in London. Later she again represented the Arboretum at the third Mexican Botanical Congress in Mexico City. She made a short pour l’Etude Taxonomique de la The 1966 plenary session of the Association J Flore d’Afrique Tropicale (A-E.T.F.A.T. ) held in Uppsala, Sweden. Dr. Bernice t Cc Schubert represented the Arnold Arboretum at these meetings. 564 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 field trip following the Congress to locate material of Nanarepenta tolu- cana Matuda, purportedly rare, which may well be a species of Dioscorea already described. Dr. Wood attended a special symposium on terrestrial plant ecology held at St. Francis Xavier University in Nova Scotia. Gifts and Grants: The Arnold Arboretum is fortunate to receive small monetary contribu- tions from many people who visit the grounds throughout the year. This immediate expression of appreciation is warmly received. We value equally highly the generous response to an annual appeal made to the contributors of longer standing who are the Friends of the Arnold Arbo- retum. During the past year the gifts for current use have been assigned to support work in horticulture in Jamaica Plain, except for the few spec- ified for other purposes. Additional summer help on the grounds and in the greenhouses was made possible by these gifts. An additional bequest from the estate of Laura Lucretia Case was added to the endowment funds of the Arnold Arboretum. Gifts and grants were also received in support of specific portions of the living collections and for the research of the staff. Most significant was the grant made for the support of a Generic Flora of the Southeastern United States under the direction of Dr. Wood. The project supported by the grant has been approved for a period of five years although the present grant covers only the first two. This continuing study was sup- ported in its initial stages by Mr. George R. Cooley and, more recently, by other grants from the National Science Foundation. Forty-five papers treating 76 families and 249 genera of flowering plants have already been published for this study in the Journal of the Arnold Arboretum. We are pleased to acknowledge the many gifts of plants, books, manu- scripts, and similar material items which increase the value of the Arboretum as a research and reference institution. Mr. William J. Flemer of the Princeton Nurseries has supplied valuable plant material for the street tree collection, Mrs. A. F. Bonsal of Upper Montclair, New Jersey, has been instrumental in obtaining for us Jris rhizomes of the Dykes medal winners to supplement the display and teaching collection at the Case Estates. Mr. Nathaniel Whittier of Medfield, Massachusetts, has donated many bulbs of interesting taxa of Allium. The University of New Hampshire made a gift of the Latimer Seed Collection which is to be used as a teaching and reference collection at the Dana Greenhouses. Additional research material on the Caribbean area from the library of the late George H. Hamor was a gift of his widow. Mrs. Van Wyck Brooks presented the Arboretum with the original letters she had received from Professor Sargent. These gifts and many others are of significance and value to the increasing collections of the Arnold Arboretum. on 1967 | THE DIRECTOR’S REPORT 56 Publications: The twelve numbers of Arnoldia issued annually, but at irregular intervals, are contributions of the Arboretum staff to the field of horticul- ture. During the past year several invitational articles of special interest were published. A paper by Dr. John Creech of the United States De- partment of Agriculture was presented at the annual meeting of the American Horticultural Society in Georgia and published in Arnoldia as “Expeditions for New Horticultural Plants.” At the International Horti- cultural Congress Dr. Peter Lapin of the Main Botanical Gardens in Moscow spoke of the efforts of that organization in plant introduction and on the hardiness of ornamental shrubs. He agreed to revise this pre- sentation and describe his research in a special article which we were pleased to publish as “Seasonal Rhythm of Development of Woody Plants and its Importance in Introduction.” The work of Dr. Wyman and Dr. Flint in revising the plant hardiness map which serves as a basis for numerical indication of plant hardiness comprised another issue of Arnoldia. The accompanying text compared the several systems of hardiness zonation now in use. Copies of this bright and clear new color reproduction have been distributed to botanists and horticulturists throughout the country. Dr. Schubert served as editor of the Journal of the Arnold Arboretum. Four issues published during the year comprised 387 pages and 22 papers. The first of Dr. Philip Munz’s studies of Delphinium and Consolida of Africa and Asia were published in 1967. This series, treating some impor- tant collections of the Arnold Arboretum, will continue through April, 1968. Dr. Howard was guest editor of the July, 1966, number of the News- letter of the American Association of Botanical Gardens and Arboreta which featured a survey of methods of security and the problems of pro- tection of plants and people in the botanic gardens and arboreta of North America. Bibliography of the Published Writings of the Staff and Students July 1, 1966-June 30, 1967 tomy of the leaf-bearing Cactaceae. XVI. The Baitey, I. W. Comparative ana rated leaves of Pereskiopsis. development of water-soluble crystals in dehyd Jour. Arnold Arb. 47: 273-287. 1966. _ The significance of the reduction of vessels in the Cactaceae. Jour. Arnold Arb. 47: 288-292. 1966. Baranov, ANpREY I. Publications of A. I. Baranov 1940-1965. Taxon 15: 265- 269. : _ Taxonomic studies in Sorbus alnifolia (Rosaceae). Baileya 15: 36-42. _—— & Sxvortzov, B. Plantae novae et minus cognitae florae Chinae boreali- orientalis (II). Quart. Jour. Taiwan Mus. 19: 155-164. 1966 566 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 BrizicKy, GEorGE K. The Goodeniaceae in the southeastern United States. Jour. Arnold Arb. 47: 293-300. 1966. . Nomenclatural notes on Gossypium (Malvaceae). Jour. Arnold Arb. 48: 152-158. 1967 yaar Harrison L, How to have a good clipped hedge. Arnoldia 27: 17-27. 1967. : nae tet plants for Vermont. Univ. Vermont Ext. Serv. Spec. Publ. 8: 1-84. ——.. sae ee and dehardening of woody plants. Pl. Gard. 22: 16-18. 1967. . Seasonal hardening in trees and shrubs. Arnoldia 26: 57-60. 1966. . Seasonal hardening in trees and shrubs useful even though not winter- hardy. Arnoldia 26: 61-64. 1966 , Boyce, B. R. & Beattie, D. J. Index of injury — a useful expression of freezing injury to plant tissues as determined by the electrolytic method. Canad. Jour. Pl. Sci. 47: 229, 230. 1967. = ALFRED J. Weather as it concerns ro practice of cutting selection. t. Pl. Prop. Soc. Comb. Proc. 15: 235-237. 1966. ———. Dwarf conifers from witches’-brooms. pet 27: 29-52. 1967. ene Tuomas G. The flora of the ‘“‘Driftless area.”” Univ. Iowa Stud. Nat. t. 21: 1-174. a revision of the Malesian species of Zanthoxylum (Rutaceae). Jour. Amold Arb. 47: 171-221. 1966. Howarp, Ricuarp A. The Director’s Report. The Arnold Arboretum during the fiscal year ended June 30, 1966. Jour. Arnold Arb. 47: 323-345. 1966. . Edible plants of the tropics. Unit 9B, Suppl. U.S. Nav. Med. School Kinescope film HA-PMB 619: 1-16. 1967. ———. Fire protection. Quarterly Newsletter AABGA 67: 21, 22. ———. Notes on the cultivated woody species of Clitoria (Leguminosae). Baileya 15: 14-19. 1967. . Security and protection at Botanic Gardens and Arboreta. Quarterly Newsletter AABGA 67: 7-14. 1966. . The persistence of the double-flowered form of Celandine Poppy. Rhodora 69: 179-184 Tropical flora of medical importance. ue 9A. Suppl. U.S. Nav. Med. School Kinescope film HA-PMB 618: 1-17, 1966. EBER, CLAUDE. The botanical garden of Saint-Pierre 1803-1902. Am. Hort. Mag. 45: 398-403, 1966. Hv, Suiv- Yinc. The Compositae of China (III). Quart. Jour. Taiwan Mus. 19: 238-310; The Compositae of China (IV). Ibid. 311-409. NEVLING, Lorin I., Jr., & Woopvgury, Roy. Rediscovery of Daphnopsis hel- leriana, Jour. Armold Arb. 47: 262-265. 1966. Sax, Kar. Biological problems of the age of science. Wash. State Rev. 10: 5-9. 1966. The Bussey Institute: Harvard errant s Graduate School of Applied Biology. Jour. Hered. 57: 175-178. 196 & Sax, Hatty J. Radiomimetic eae drugs and mutagens. Proc. Nat. Acad. 55: 1431-1435. 1966. ———————. Radiomimetic effects of beverages, nla and _ insecticides. Cranbrook Inst. Sci. News Letter 36: 46-49. 1966 SCHWARTEN, LAZELLA (with Rocerson, Crark T.., RICKETT, H. W. & BECKER, 1967 | THE DIRECTOR’S REPORT 567 Herman). Index to American Botanical Literature. Bull. Torrey Club 93: 360-384, 463-489. 1966; 94: 111-124. 1 Woop, Carro_t E., Jr. Lectotypification of specific names in floras. Taxon 16: 23-28. (with BucHHEIM, GUNTHER). Gymnospermae in Appendix II. Nomina Familiarum Conservanda. International Code of Botanical Nomenclature. Edinburgh, 1964. p. 209. 1966 Wyman, Donatp. A few poisonous plants. Arnoldia 26: 65-75. 1966. New England for autumn color. Am. Hort. Mag. 45: 281-285. 1966. & Firint, Harrison L. Plant hardiness zone maps. Arnoldia 27: 53-56. 1967. Ricuarp A. Howarp, Director 568 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 The Board of Overseers of Harvard College Committee to Visit the Arnold Arboretum AucusTIN H. Parker, Jr., Chairman, Boston, Massachusetts. GeorcE R. CLark, Vice-Chairman, Philadelphia, Pennsylvania. Mrs. GeorcE L. BATCHELDER, Jr., Beverly, Massachusetts. Mrs. RALPH BRADLEY, Canton. scant ALBERT C. BuRRAGE, Ipswich, Massachusetts. WILLIAM J. Devine, Commissioner, City of Boston, Department of Parks and Recreatio WILLIAM EE III, Princeton, New Jersey MItForp R. LAWRENCE, Falmouth, Massachusetts, Harotp E. Moore, Jr., Ithaca, New Yor R. HENRY Norwes, Jr, Mentor, Ohio. Mrs. RicHARD W. Pratt, Chestnut Hill, Massachusetts. FRANCIS W. SARGENT, Boston, Massachusetts. Mrs. Homer N. Sweet, Weston, Massachusetts. GEORGE TAYLOR, Kew, England. Joun J. TIERNEY, Jr., Milton, Massachusett Mrs. G. KENNARD WAKEFIELD, Milton, has ate Mrs. CHARLES D. WEBSTER, New York, New York. RicHArD P. WHITE, Washin gton, D.C. NATHANIEL Wuittier, Medfield, Massachusetts. Mercer Research Fellows during at least part of fiscal year 1966-1967 Witiiam GEnsEL, University of Rhode Island, sy chaeriee Rhode Island. PABLO RauL LEGNAME, Instituto Lillo, Tucuman, Argen SANDRA SHANNON, University of Massachusetts, perce “Massachusetts. Cuartes C. TSENG, Windham College, Putney, Vermont. Francis H, Wotre, University of New Hampshire, Dover, New Hampshire. + Died January 6, 1967. 1967] STAFF OF ARNOLD ARBORETUM 569 Staff of the Arnold Arboretum 1966-1967 RicHARD ALDEN Howarp, Ph.D., Arnold Professor of Botany, Professor of Dendrology, and Director. IrvING WIDMER BaILey, S.D., Professor of Plant Anatomy, Emeritus.t Kar Sax, S.D., Professor of Botany, Emeritus. PRISCILLA JEAN JAMES-ASHBURNER, Business Secretary. Anprey I, BaraAnov, Curatorial Assistan GEORGE KONSTANTINE Brizicky, R.N Dr,, Botanist, Southeastern Flora Project. nt. t.* Tuomas Gorpon Hart ey, Ph.D., Associate Curator of Pacific Botany. HEMAN ARTHUR Howarp, Assistant Horticulturist. Suiu-Yinc Hv, Ph.D., Botani MARGARET CATHERINE eas. Herbarium Secretary. Victor Ferenc Marx, M.Libr., Librarian. Honor CApPeL Moore, M.Sc., Research Associa Lorin Ives NEVLING, JR., Ph.D., Associate See and Supervisor of the Herbaria. Dutcre AuicrA PowEtt, M.A., Research Fello BERNICE GipUzZ SCHUBERT, Ph. a Japan pestis and Editor. LAZELLA SCHWARTEN, Librarian. STEPHANNE Barry SuTTon, A. i ae Fellow. Rosert Gerow WILLIAMS, B.S., ‘Superintendent. CarRoLL Emory Woop, Jr., P D. ., Associate Curator. DonaLp Wyman, Ph.D., Horticulturist, t Died May 1 * Appointed aie ee the Gray Herbarium. bled Retined June 30, 19 570 JOURNAL OF THE ARNOLD ARBORETUM Acacia, 451 Acaly sha. 323, 370-375 — ser. Caturoideae 37 — ser. Dentatae, 3 — ser. Monostach — ser. Palminerviae, 372, 374 Account of the Red River Expedition of 1806, an Overlooked Publication of Botanical Interest, Freeman and Cus tis’ Acer negundo, 447 = pens vancum, 443, 2 a. Acerates, 454 Achillea ‘santoling, 454 Achras, 2 Aconitopss passe 169 Aco monogynum, 170 Adenoropium, 340 Aérial Axis and Rhizome, Vascular of Vegetative oe atomy of the Palm Rhapis excelsa, IV, 122 velopment in Apex Aesculus parviflora, 444 — pavia, 444 Aextoxicaceae, 312 447 rum dr ummondii, 443, 447 , 447 INDEX [voL. 48 African Species of ee and Con- 30 solida, A Synopsis 7 Agaloma, Agathis alba, 209 ave virgi , 457 —odorum, 452 Althaea rosea, 451 American Palms, Two New, 143 Amorpha fruticosa, 449 Ampelopsis arborea, 449 Ampris, Anacardium excelsum, 207 Anatomy of the V. Palm Rhapis seg I ascular Development in Apex bi saa Aérial Axis and Shine ee 2 —inermis, 206 pea ng 327-329 sect maps ripe — sect. Badia — sect. roa ee. 329 — arida, —telephioides, 327 Androcephalium, 461 — pac ay "461, 464 Aneilema vaginatum, 455 n mia, 458 An- hg of the Palm Rhapis excelsa, I Vv, ee — 455 Apple, Balsam, 454 Aquilicia sambucina, 449 Arbutus uva ursi Arctium, 450 — lappa, 452 —minus, 452 1967] Arctostaphylos uva-ursi, 454 Argythamnia, 323, 364-366 — cans, 364 Arisaema pentaphylla, 455 — j m, 5 — triphylla, 450 Aristeyera, 143 —ramosa, 144 —spicata, 143, 144 Aristolochia pistolochia, 455 — ntaria, 451 Asclepiodora, 454 Ascyrum hypericoides, 454 Ash, 445, 447 Ash, sya be Ash, Prickly, Ash, Water fe Asian Species of Consolida Somat aceae), A Synopsis of reas Asian Species of inium, Sens i De Stricto, A Synopsis = the, cues, 476-545 Asimina triloba, 448 Astragalus canadensis, 454 — carolinianu Astronium eraveolens 110 Athyrium Atropa alee 450 Austrobaileya, The Chromosomes of, 241 Austrobaileya, 241-244 —m , 241 — scandens, 241, 242 Austrobaileyaceae, 241 Bald Cypress, 447 B @) p ean 79 Bean, Sona Bay, 453 Bearberry, 454 eech, Beech, American, 448 Beech, European, 448 INDEX Benzoin, 4 Bernoullia flammea, 117 Bignonia capreolata, 457 —catalpa, 4 —radicans, 453 — sempervirens, 457 shri oo 459 — ungui Biology es Sonate of Eucnide (Lo- asaceae), Flora Bird Cherry, European, 442 Bischofiaceae, 312 Borassus ee 458 Botanical Interest, Freeman and Custis’ Account of the Red River Expedition of a an Overlooked Publication of, Box Elder, 447 English Sweet, 451 Briar, Seustey®, 451 I : rorcE K. Nomenclatural es on orn (Malvaceae), 152 e ud, Red, 44 Bull-Nettle, 349 Bumelia 443, 448 — smallii, Bur eae sing Burdock, 452 Burdock, Common, 452 Burdock, Great, 452 206 aR European, 450 utton Wood, oe Seren ae, 309-3 Byrsonima a 117 actus opuntia, es Callitvichaceas; Calophyllum Sapo 206, 207 Candle ects nen ony 449 aE Cane Cin ae 445 572 Caperonia, 323, 363, 364 —sect. Caperonia, 364 — castaneifolia, 363 uus virginianum, 455 cue seit 448 alba = Gilocenaes ta, ce 459 — tomentosa, Cassava, Cassava. Stin 455 Cassia cha ae. 453, 457 —marylandica, 455, 458 — occidentalis, 45 — tora, 45 Castanea Paco 444, 448 Castor bean, 379 Catalpa bignoniokdes, 448, 459 Cedrela mexicana, 206 Celastrales, 309 Celtis occidentalis 447 Cephalanthus occidentalis, 446, 448 Ceratosanthus ajacis, 176 ate! i 9 » 420-4 hamaesyce, 421-423 — sect. ae 421, 423, 426 ubse t. Chamaesyceae, 422, 423 Bare Elegantes, 42 — ser. Peplides, 422, 423 — ser. Prostratae, 422, 423 —cristata, 42 — maritina. 420 — perlignea, 421 J Chascotheca triplinervia, 330 Chenopodium anthelminticum, 458 4 Chestnut Oak, Swamp, 444 Chickasaw Plum, 446 Chokecherry, id 46 Chromosomes of presen set, The, 241 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Chrysalidocarpus lutescens, 127 457 peri 323, 349-353 . Calyptrosolen, aoz Coccoloba cozumelensis, 206 Comfrey, European Commelina va ginata, 455 — virginica, 455 mpar rative tg Stneige in . Wood Anatomy, 1; II. Consolida, ‘i Synopsis of African Species of Delphinium and, Consolida (Ranunculacte) A Synopsis 159 —ambigua, 4 ‘51, 53, 64, 6 Rperieeenll 161, 170, —armeniaca, 164, 198, ried — axilliflora, 162, 174, 200-202 —barbata, 161, 172, 200-202 —camptocarpa, 164, 193, 200-202 2 bk 02 7 48, 535 53, ye 165, 184, 200, bi ical 202 — de ao pa 164, 187 — divaricata, 179 —euphratica, 167 — flava, 165, 183, 201, 202 — glandulosa, 165, ved 200, 201 — gombaultii, 1 162,:°175 — halophila, 163, 189, 200, 20 TCs cma 164, 196, me 19 —leptocarpa, 163, 185, ty oe ap 191, uritanica, 48, 53 ie ii 163, gis 200-202 1967] Consolida oliveriana, 165, 182, 200-202 — olopetala, 163, 199, 200, 202 — orientalis, 48, 52-55, 164, 177, 200-202 — —sub — subsp. regalis, He 179, 200 Eero 81 La eamalose. 161, 163, 188, 200-202 —saccata, 161, 171, 200-202 — schlagintweitii, 163, 164, 191, 202 wm ie) C oO Len J io} Oo 5. > a & _— an bo Ll ht ngor 193 tae 162, er 163, oe 200-202 — glabrescens, 175 —kabuliana, 175 Gani 163, 199, 201 — teheranica, 161, 165, 201, 202 —thirkeana, 162, 167, 200, 202 — tomentosa, 164, 195, 200, 201 — — subsp. oligantha, 194 a te 164, 190, 200 Co nse arvensis, 450 Copaifer ot apealts olia, 20 Coreopsis lanceolata, 454 Cc elsaihens grandis (Gesneriaceae), A Species from the Solomon Islands, 245 ie oronanthera, 245 —aria, 443, ee INDEX 573 Creeper, Virginia, 455 re escentia cujete, 111 — sect. , 354 — subsect. Argyrocroton, eg — subsect. Cyclostigma, — subsect. wens 355; ee 358 —ser. Geiseleria, 35 — ug omer $55 horns 54 — reflexifolius, 110 —tiglium, 354 Crotonopss, 323, 360, 361 ae 454 Cupania guatemalensis, 207 Cupressus disticha, 443, — thyoides, 447 Curatella, 9, 233 Staring 17; 111, 238 Curcas Custis’ Account of the Red River Expedi- dition of 1806, an Overlooked Publica- tion J Botanical Interest, Freeman and, 431 Cybistax ceapre smithii, 206 Cypress, Bald, 447 Cyrtandra, 245 ee. be Daniellia, Daphniphy nie $12,313 3 Dayflower, Common, 455 574 JOURNAL OF THE ARNOLD ARBORETUM Delphinium and Consolida, A Synopsis of African Species o Delphinium, Sensu Stricto, A Synopsis of the Asian Species of, 249-302, 476- 545 Delphinium, 30-47, 53-55, 160, 161, 249- 302, 476-545 — sect. Anthriscifolium, 259 ——eginense, 267 Pisses ae 499, 503 — — pumilu — siboinaapinntia 263, 264, 268 — amani, ambiguum, a 53, 176 —— callerii, 260 60, 261 — aquilegifolium, 263, 265, 268 —— slandulosissmurn, 268 —w— hebecarpum, 268 —— Sonpiseduietdala ae 268 ——violaceum, 268 ——— pallidum, 34, 33 icooaareceal ogpreesgan 172 — — hirs - ola aa "263, — batangense, ves wp 505 esian 5 5 —bhutanicum, 501, 507 — biternatum be ey A m, 294 ——magnificum, 508 —majus, 527 —_—— obtusilobum, 508 —callerii, 2 Mar wuts — — brev Lager 193 ——d iso , 193 iw _ 2 oense, "501, 502, 510 phe thaablni 500, 511 [voL. 48 1967 | INDEX 575 Delphinium subsp. mi anaptions 511 Delphinium englerianum incisum, 520 — subsp. schizophyllum, 5 —eriocarpum, 252 — — albiflorum, 511 — esquirolii, 543 —w—chinense, 511 — exiguum, —w—dahuricum, 511 —exsertum, 18 — — laxiflorum, 511 — fangshanense, 518 —— parviflorum, 511 — fengii, —-— subglabrum, 511 — fissum ch 517 —— pie anatolicum, 276 — chodatii, 266, 274 — flavum, — cinereum, 252, 255 —— deserti, 184 —consolida, 48, 50, 53, 161, 178 — floribundum, 271 —w— subsp. divaricatum, 179 — foetidum, 478, 484 —w— subsp. paniculatum, 179 — forrestii —— micranth 79 — forskolii, 252 — cornutu — freynii, 189, 294 Nee bore _ cossonianum, 33, 42, 53-55 —— laxiflorum, 42, 53 —crispulum, 499, 513 73 — cruciatum, — cyphoplectrum, we 274 —— lasiostemon, 182 ——micranthum, 2 —— leiostemon, 18 — pallidiflorum, 700 brea fay 284 — vanense, : — dahuricum, i ete hn sees pee art ‘& “sis, ee — grandiflora sus, 508 — grandiflorum, 254, 498, 500, 517 ——undulatum, 482 Soca genle 3 — sig nats "32, 35, 53-55 eT ——minor, 35, 53 — —chinense, 517 a Aion 5 202. 200, 275 ae eae ma — — gilgianum, 510 — — sub owitzianum, 297 —— glandulosum, 518 — — longebracteatum, _ — gmelinii, 518 —— ochroleucum, 275 — —kumaonense, 518 — davisii, 252, 256 — —kunawarensis, 508 — davuricu 1 — — latisecta, 5 — densiflorum, 477, 483 ——leiocarpum, 518 —w—platycentrum, 483 Cee Saal ‘ —w— stenocentrum, 483 m 51 —denudatum, 501, 515 = geenynehion, 527 nnanense, 543 — — pallidum, 518 _— deserti, Si; 53, 184 pir ae paradoxa, 51 eserti-syriaci, 187 — — pubiflorum, 517 — dissectum, 499, 5 —-— pumilum, 51 oe Pie — — setchwanensis, 517 — — gillia — — tenuicaulis, 50: — dtiehocentroies Pie 517 — —tigridium, 518 — dolichoce — — tsangensis, 517 _ dcletinnahoia 262, 276 —griseum, 266, 277 — elatu cisum —halophilum, 189, 294 as elisabethae, 499 —_— halteratum, aay 46, 53-55 —emarginatum, 32, 41, 53-55 ——cardiopetalum, 46, 53 — -—nevadense, 42, 53 —— gracile, 45, 53 576 JOURNAL OF THE ARNOLD ARBORETUM Delphinium halteratum macropetalum, 45, 53 —-—obcordatum, 44, 5 ——tribracteolatum, 46, 53 ——venulosum, 255 — hellesponticum, 196 ~—— — subsp. aintabense, 196, — aa 183 ? —incisum, 499, 5 ee 265, 279 —intricatum, 1 atu —ithaburense, 262, 279 — lacostei, 477, 486 Delphinium lanigerum, 500, 523 latesquamatum, 264, 283 — laxiflorum, 18 —Jaxiusculum, 267, 284 — leiocarpum, 276 — lepidum, 344 — leptocarpum, 185 — losccsii brevirostratum, 48, 53 — macedonicum, 196 —macrocentron, 32, 38, 53-55 — — viride, — spe oe 497, 523 — mauritanicum 48 . 28 — obcordatum, 33, 44, 53-55 — obscurum, 53 ae aye oo 288 — orientale, 52, 54, —— subsp. hisianiteiine: 52; 54,471 —— brevicalcaratum Ban — ae el 266, 290 [voL. 48 1967] Delphinium paniculatum, 179 ae = SEREVASO Eee aradoxum, 1 2 aredioen onl — pauciforum, 511,:515 — pellucidum, 492 —peniilatum, a 290, 300 roplect 297 Sees iy 39, 53-55 — — glabratum, 39, 9, — peregrinum, 32, 43-46, oh oy 252 bsp. halteratum, 46, tee: junceum, 252 — subsp. nanum, 44, 45, 254 um, 192 — py enacentiraiies latisectum, 537 — pygmaeum — pylzowii, 476, 497, 499, 500, 532 — — trigynum, 532 — pyramidatum, 499 — quercetorum, "263, 9 o ranunculifolium incisum, 520 — robertianum, 260 INDEX 5 Delphinium rockii, 500, 533 —roylei, 267, 292, 501 —rugulosum, 188 —ruspolianum, 37, 54 —saccatum, 1 —_ salmo , | | Lan ° =z Q oO & 5 $ oO _- schmathausenii, is 264, 293 — schroedingerianum, 27 —sclerocladum, 166 — -anthoroideum, 167 777 pum —senibarbatum, 265, 294 —— hoeltze ees 266, 295 a 3 5 oO 5h in 5° _ soonmin MALS if a stentruetzct eae 498, 499, 534 — speciosum brevicalcaratum, 484 Lennie fs — staphisagria, 32, 33, 54, 55, 258, 455 —stenocarpum, 169 stocksianum, 175 tS asain 175 00, 535 — sylvaticum, i oe 7 he eg aE — syncarpum — szov paca lngebractetum, 285 265, 00, 538 —tetanoplectrum, 262, 298 —thirkeanum, 167 —tomentellum, 513 ba | 578 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. 48 Delphinium tomentellum angustibractea- I. Wood Anatomy, 1; II. The Pollen, um, 5 231 —— araraticum, 513 Didesmandra, 7, 233 — ranthum, 513 —_— phe 239 —tomentosum, 195 Didynamia, 458 — —longecalcaratum, 196 Digitalis, 456 — tribracteolatum, 46, 54 — flava, 45 ichophorum, 477, 492 —!l —— platycentrum, 49 Dillenia, 5, 232 —— subglaberrimum, 492 —a tibeticum, 4 — aurea, 17 — tricolor, 297 —b —- trigonelloides, 190 — excelsa, 17, 239 — trilobat 95 — eximia — triste, 499, 539 — indica, 18, 239 —— linearilobum, 513 — ingen — trisectum, 500, 33 -— Juzoniensis 18, 239 — tsarongense, 476, 493 — megalantha, 239 —— patentipilum, 493 — ovata, — tsoongii, 501, 540 — papuana, 239 — tuberosum, 263, 299 ein oats 4 — — leiocalycinum, 299 — penta —turkmemum, 264, 300 — phpinenss 18, 239 —uncinatum, 266, 300 — pulch — — aitchisonii, 3 — que — — pseudovestitum, 300 — reifferscheidia, 239 —ursinum, 301 — retic vanense, 265, 266, 302 — retusa, 1 — venulosum, 252, 255 — scabrella — verdunense, 33, 46, 53, 54 — solomonensis, 18 — vestitum, 3 494 — suffruticosa, 239 mR ore) 4 = oe 19 F —-— sphenolobum, 49: ae — virgatum, 252, 253, 518 Dilleniaceae, ee OI: : natomy, 1 , — viride, 48: — viscosum, 476, 477, 495 —— chrysotr ichum, 495 The Pollen, 231 Dilleniaceae, 1-29, 447 Director’s Report, The. on: Arnold Ar- waa bie’ < Bi - oretum iba! - Fiscal Year Ended — williamsii, 501, 541 ke o aa — wilsonii, 498, 542 : ogwood, 44. — wrightii, 478, 496 fala Bee — yuanu 00, 543 Doliocar as 7 — yunnanense, 485 543 oe 19 - > — — laxum, -—dentatus, 19, 239 — — strictum, te ~— guianensis, 239 — zali ~—lasiogyne, 19, 239 Depa : — major, 19, 239 ialium guianense, 117 — oliva ‘2 Dichapetalace e, 309 Drosera brevifolia, 454 Dichrophyllum, 397 — cuneifolia, 454 Dicxison, Wi1iam C. Comparative Drummond’s Red Maple, 443 Morphological Studies in Dilleniaceae, Dryopteris, 455 1967] Drypetes, 322, 329-332 — ae ae — taylorii, 90, 93 Echinacea purpurea, 452 , 450 Echinops — sphaerocephalus, 4 Ecology, Som beak of Tropical Plant, Eglantine, 451 Elaeag Elder, Box, 44 — scaber, Elm, 445, 447 Emblica, 332 Empetraceae, 309 _ Enterolobium cyclocarpum, 206 Eperua falcata, 210 RNST, WALLA Henry J. HOMPSON Floral. Biology and a. tematics of Eucnide (Loasaceae), 56 Erythrina herbacea, 451, 453 Eucalyptus, 20 ide, 57, — sect. Mentzeliopsis, a 88, 68, 84 — sect. Sympetaleia, 57, 58, 68, 85 — ser. Eucnide, 57 — ser. as 57, 38, 82 —a ious Th Natenbaand: o a ne 79 — lobata, 57, 58, 78, 81 — nelsonii, 83 — parryi, 84 — pringlei, 83 INDEX 579 Eucnide Sagas ani cee 83 — rupestris, sh 86 — cee — sy nandra — tenella, oe 78, 87 — urens, ee 58, es — sect. / — sect. : — sect. ! — sect. Aphy — sect. Balsamis, — sect. — sect. — sect. — sect. — sect. — sect. — sect. — sect. — sect. — sect. — sect. — sect. Myrsinitea — sect. — sect. — sect. — sect. — sect. — sect. 410 404, 405 Dichilium, 402, eh 429 coeeatln = Epurga, 399 Esula, 1399, 408, 409 Tirucalli, 397, 404, 405 Tithymalopsis, 400-402, 408, 425 Tithymalus, 398-401, 405, 408- : ilmenite ait 429 — subg. Chamaesyce, 421, 425 — subg. Eremophyton, 398 — subg. Esula, 397-402, 405-410, 414, 420, 424-426, 429 42 — subg. Euphorbia, 397, 404, 405, 407, 409, 410, 418, 426 — subg. Lyciopsis, 398 — subg. Poinsettia, 402-404, 406, 407, 410, 417, 429 — subg. Rhizanthium, 398, 405 — subsect. Corollatae, 400, 401 00, 401 . Ipecacuanhae, 400-402, 425 — subsect. Pachycladae, 405 580 Euphorbia subsect. eg coreeaig 399 —subsect. Pleiadeniae, — subsect. Purpuratac. ir 0. Euphorbiaceae in the Southeastern United States, The Genera of, 303-361, 363- 430 Euphorbiaceae, 303-361, 363-430 —subfam. Crotonoideae, onetergg 311, 15231455322, 1338 53 = — subfam. Euphorbioide oy 0, 430 Rca Oldfieldioideae, 308, 309, 312, Sout —su ae Paivaeusoideae, — subfam. sing rire ay a, 310- 314, 322, 325, 331 — tribe Euphorbieae, 304, 306, 395-430 395 — tribe Phyllantheae, cate 306, 325, 336, 337 —subtribe Acalyphinae, 370-375 — subtribe Andrachninae, 325,-326, dal 53 — subtribe Fanta Ae 347, 351 — subtribe rcurialinae, 366-370 Erecous a 305,329, 332, pe Plukenetiinae, 376-379 — subtribe arent S 9-385 n Overlooked Pub- canbe) Interest, Freeman and Custis’ Account of the Red River, 431 Fagus grandifolia caroliniana, 448 — dass _ — sylva shee Foxglove, 456 Ficu — gla Field Gate 453 Fieldia, 245 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Flacourtiaceae, 311, 313 Floral Biology nA Systematics of Eucnide (Loasaceae), 5 Flowers, Die they 450 Flowers, Studies in Swietenia (Melia- ceae): Observations on the Sexuality of Forestiera acuminata, 444 Fox Grape, 446 Foxglove, False, 456 Fragaria vesca, 453, 454 Freeman and ss stis’ pig af ne ie River Expedition 1806, looked Publication a se ie est, 431 Fungi, 455 Galarhoeus, 397 Galearieae, 313 Garlic, Field, 453 Garlic, Wild, 453 elsemium semperv , 457 Genera of Eu eect in the South- eastern United States, The, 303-361, 3-430 Genipa americana, 206 , The, Genus Lunasia Pomel A Revision of he, 460 Geraniales, 309, 313 — tribe Coronanthereae, 245 E, Coronanthera pera A New Species from the Solomon ey 245 oe aquatica, 445, raceojonag —triacanthos, 4 7 Globe Thistle, saat ly GomeEz-Pompa, ARTURO, Some Problems of Ele gor Plant Ecology, 1 Gonolobus, 4 Gossypium cence Nomenclatural Notes on, Gossypium, 152-158 —sect, Anomala, 157 1967 | sist pium sect. Anomopambak, 157 cee Erioxylum, 157, 158 sect. sect Sect. Sect, Sect. Sect. — sect. io) uta, Ingenhousia, Klotzschiana, 157, 158 sect. sect. sect. sect. Sect. — sect. 158 ieee iech Caducibracteolata, 158 —subsect. Ingenhousia, 158 —subsect. Integrifolia, 158 5 Grape, Fox, Grape, June, 446 rape, Wine, m, White, 445 ymnanthes ' 324, 387, 388 — sect. Stillingiopss, 386 — lucida, Hagberry, 442, 447 Halesia carolina, 448 as G. A Revision of the Genus Lunasia (Rutaceae), 460 Haw, Black, 446 INDEX Hazel, Witch, Heders qtingoefola, 455 —s Hemicyclia, 330 Hevea, 322 Hibbertia, 6, 233 ‘© 9 — cuneiformis, 19, 23 hirtus, 457 —lasiocarpos, 457 militaris, 453 Hickory, 444-447 Holly, English, 444 Hollyhock, 451 Honey Locust, 445, 447 453 e, rp, Ricwarp A. The Director’s Re- ate The Arnold Arboretum During the Fiscal Year Ended June 30, 1967, 546 cvopampr tree, 455 Hura eae Ie cane Hydrocotyle fiver 451 — vulgaris, Hymenaea courbaril as a Resin Producer, Preliminary Investigations of, 203 reana, 20 — courbaril, 203-230 582 JOUR Hymenaea gros obtusifolia, 205 — — subsessilis, — davisii, lificum, 452 — Haan 450 Icacinaceae, 312 lex, 44 ae 443, 444, 448 455 aea courbaril as a Resin pee Peclisaag ary, 203 foxylon, 4 moea wucee 456 — solanifolia, 45 Tridaceae, 455 Ironwood, 443 Irvingiaceae, 310, 313 a 322, 340-342 ect. Ad denorhopium, 341 ioe Jatropha, 341 — sect. Macranthae, 340 — gossypiifolia, 340 — manihot, 346 — urens, 455 JAYAWEERA, Don M. A. The Genus Dua- ban, nga, acai Peed 457 Joxylon NAL OF THE ARNOLD ARBORETUM [voL. 48 Juglans, 442 —alba, 442, 444, 446, 447 a PiEnvesain: 44 —nigra, 448 ieeceaeoos 442 os 442 ae. 439, 442, 447, 459 Julocroton, 354 June Grape, 446 Juniperus virginiana, 446, 448 — repens, 452 Karwinskia calderoni, 206 Kibara, 244 — grandiflora, 96 LANGENHEIM, JEAN H. Preliminary In- i of apm courbaril as esin Produce 3 paar montan, 450 ) Dg . Studies i (Meliaceae): Observations on the Sex- h ers, 101 — subfam. Cacsalpinioideae, 204 Leon ae ‘el wang 453 Lepidanthus, 327 a tsgedticn oneaicae 96 Leptospermum citra , 208 Licania arborea, 0 a 1 era benzoin, Liquidambar eae. poh 447 Loasaceae, Pcie Loase coe Pe sn penta 457 Lotus, Old World, 452 1967] Luehea candida, 20 Lunasia iutacen), A Revision of the Genus, 46 Lunasia, page —am oe. 460, 464 ——amara, 464 Pesan i 473 —— costulata — — lanceolata, oy 4 — truncata, 465 Maclura, let — pomifera Magnolia ae 448 Magotty Bay bean, 453 Mallow, Rose, 453 Malva spieltnlitia: 451 Malvaceae, 307 — subsect. bs means 347 — esculenta, 346 Manilkara “tapetile. 117 Maple, 447 Maple, Sree Red, 443 Maple, Red, 447 Maple, Striped, Margaritaria discoidea, 311 arrubi vulgare, 4 Mastichodendron capiri tempisque, 110 Meliaceae, 101-104 Mentha sativa, 453 Mentzelia cordata, 82 — gronoviaefolia, 79 — hirta, — lobata, 81 — urens, 84 Mercurialis, 323, 366-370 — ua, 452 — perennis, 367 INDEX 583 Mercu 367 Mexican Fin 446 — strigillosa, 450 — rap tag 451 Mint, Mirandacelis monoica, Misleto, 4 aca Hickory, 442 Modiola caroliniana, 451 110 a veg Morning Glory, 456 Morphological Studies in Dilleniaceae, Wood Anatomy, 1; IT. Morton, C. V. Freeman and Custis’ Ac- count of the Red River Expedition of 06, an Overlooked Publication of Botanical Interest, 431 Mozinna, 340 Mulberry, 445, 447 Mulberry, Black, 447 Mulberry, Red, 448 Munz, Puirie A. A Synopsis of African Species of Delphinium and Consolida, 30 Munz, Poitier A. A Synopsis of the Asian Species of Consolida (Ranun- Hinip A. A Synopsis of the sian Species of Delphinium, Sensu fae 249-302, 476-545 Myrica cerifera, 443, 444, 449, 453 Myroxylon, 21 Mytilicoccus costulatus, 464 — grandifolius, 464 Mytilococcus, 461 — quercifolius, 461, 464 Staenigrin 206 Negria, Nemasty lis rats 455 Neptunia, — lutea, 480 — plena, 450 584 Nettle, Bull-, 349 Nettle, Horse, 451 Nettle, Spurge, 455 Nettle, Stinging, 349 Nicandra physalodes, 450 Nomenclatural Notes on Gossypium (Mal- vaceae), 152 Notes on Gossypium (Malvaceae), No- menclatural, 152 ata, 452 Nyssa integrifolia, 445, 447 — sylvatica, 445, 447 Oak, Oak, Oak, Oak, Oak, Swamp Chestnut, 444 Oak, White, 4 45 Oak, Willow ie Observations on the Sexuality of the Flower vi Studies in a iteenke (Melia- 447 Black, 443, 445, 446 Osage Orange, 449, 459 3 Osier, red, Oveiitked Publication of aise In- terest, Freeman and Custis’ Account of the Red River Expedition of 1806, an, 431 Oxalidaceae, 313 alis — violacea, 451 Pachira aquatica, 206 Pachynema, 7, 233 —complanatum, 19 Preemie 239 — junc 5239 Palm, Old ' Wo rld, 458 Palm Rhapis excelsa, Anato omy of the. IV. Vascular Development in Apex of JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Vegetative Aérial Axis and Rhizome, 122 Palms, xe "pay American, 143 Palmett Pal ae A 458 Pandaceae, 313 Papaw, 448 Parthenocissus quinquefolia, 455 epee Soin — incarn — lutea prdena 454 — minima, 4 Pecan, 442, 443, 445, 447 Pedilanthus, 324, 427-430 — sect. Cubanthus, 429 — tithymaloides, 428 Peltogyne, 205 Pennywort, European, 451 Peraceae, 3 Periploca secamone, 454 Phyllanthaceae, 304 phidium, 338 ennh, se hanes 334 — sect. Urinaria, — sect. awn oe 338 — subg. Botryanthus, — subg. — subg. — subg. Go — subg. — subg. — subg. — subg. Xylophylla, 335,:336 — subsect. Niruri esr sites Pentaphylli, 335 . Swartziani, 334 ms, 452 Phytolacca americana, 451 — decandra, 451 Picrodendraceae, 312 Pilocarpus amara, 464 Pine, 447 Pine, Late. 447 Pine, Scotch, 443 Pine, Shortleaf, 443 1967] Pinus, 208 — caribaea, 111, 222 —echinata, 443, 447 Plantago major, 450, 453, 4 Platanus Secidentalis, 443, pa 447 Plum, Chickasaw, 446 Plum, Guiana, 330 Plum, Plum, 446 Podophy lum ain 452, 453 Comparative nay clipart Studies in Dilleniaceae, II, pee me —vu Po ibe patio, 453 — pennsylvanicum, 457 — scandens, 454 Poly podium, 440, 455, 456 Polystichum, 455 Pond Apple, in Poplar, edad 442 Populus — delto 447 porn 3 el 327, 450, 451, 457 Preliminary Tnvestigations of Hymenaea ourbaril as a Resin Producer, 203 Prenanthes alba, 458 Prickly Ash, 444 Prinos verticitiatws, 449 paifer se ee xian — lauro-cerasus — virgini 6, 44 Pueudolmedia oncaeid 117 Psidium sartorianum, 206 Publication of Botanical Interest, Free- INDEX 585 man and Custis’ Account of the Red River Expedition of 1806, an Over- looked, 431 Quercus alba, 442, 446, 447 ——C i — rubra, 442, 446, 447 — velutina, 446 Rabelaisia, Lig — parvifolia, 464 — pilipinensis 461, 464 Ranales, 2 escke ae 159-20 Ranunculus bulbosus, 450 — hederaceus, 45 Red River Expedition of 1806, r- ical saat looked Publication of Botan est, Freeman and Custis’ ppc of the, 431 Redbud, American, pe Redbud, European, Resin Producer, Lega Investiga- tions of Hymenaea courbaril as a, 20. Revision of the Genus Lunasia (Ruta- ceae), A, 460 Rha bdothamnus, 245 Rhamnales, 309 Rhapts iat Anatomy of the Palm, scular Development in Apex of Voqstatv Aérial Axis and Rhizome, eee excelsa, 122- Rhizome, Vascular Development in Apex of Vegetative Aérial Axis . Anatomy of the Palm Rhapis oa EV, 122 Rhus copallina, 452 — coriaria, 452 —cotinus, 45 — glabra, 452 — radicans, as siege ng 450 hina, 457 Ricinus, 323, 379-385 — communis, 379 inia, he — pseudacacia, 445, 448 ecvhacen ict 111 586 Rosa eglanteria, 451 45 Rud eke hirta, 451 — purpurea, 452 Ripenserc, Liry. The Chromosomes of Austrobaileya, 241 Rutaceae, 460-475 —subfam. Rutoideae, 460 — tribe Xanthoxyleae, 460 —subtribe Lunasiinae, 460 Rutales, 309 Sabal minor, 454, 458 Sagittaria falcata, 455 —nigra, 442 Sa mbucs canadensis, 451 Secindule, 309, 310 Sapium, 324, 391-393 — sect. Triadi —albidum, 443, 447 Satureia sey hepin 454 Savia,’322. 325-327 sect. envi ee 326 — sect. Petalodiscus 325 — sect. Stillingiopsis, 386 — brasiliensis, 385 Sensitive Briar, 451 Serophyton, 366 Sexuality of the Flowers, Studies in Swie- tenia (Meliaceae): Observations on the, Shortleaf Pine, 44. Sickingia aa 110 Sicyos angulatus, 454 Sida, — occidentalis, 452 JOURNAL OF THE ARNOLD ARBORETUM [voL. 48 Sideroxylon, 443 —mite, 443, 448 Silphium laciniatum, 454 Silverbell, Carolina, 448 Simaroubaceae, 310, 312 — subfam. Irvingioideae, 310 Sison ammi, 455 Slipper-spurge, 428 Smilax repre nox, 452 AS — sarsaparilla, 452 Solanum carolinense, 451, 452 — panicula tum, 451 — virginia num, 452 Solidago altissima, 458 Solomon Islands, ee grandis (Gesneriaceae), A New Species from the, 245 Some Problems of Tropical Plant Ecology, 105 Sorrel, Common Wood, 449 Sorrel, Violet Wood, 451 , 447 ited States, The Genera orbiaceae in the 303-361, 363- Southern White Cedar, 447 scsi from the Solomon Islands, Cor- gee grandis (Guanwsiacesr), A a cits ca Consolida ape ies: ana of the nia elphinium and Consolida, A bids. . Stricto, A sei of the Asian, 249-302, 476- Spi pice ; Wood, 444, 449 Spigelia anthelmia, 453 — marilandica, 451, 453 Spurge, 39 Spurge family, 303-361 Spurge pos 455 Spurge, Slip , 428 Staiisera 3 saseiicue. 33, 54,258 — sect. Stillingia, —ser. Sylvaticae, 389 reculianae, 389 — sylvatica, 389 1967] Stinging Cassava, 455 Studies in Swictenta (Meliaceae): Obser- vations on the Sexuality of the Flowers, 101 Styrax, 210 — argenteus, 206 Sumac, Smooth, 452 Sweet Briar, English, 341 Sweet Gum, 447 Observations on the Sexuality of the Flowers, Studies in, 101 Swietenia, 101-104 — hum 06 = ttscophyl, 101-104 —mahagoni, 101- bes Symphytum officinale, 451 Synopsis of African Species of Delphin- ium and Consolida, A, 3 Synopsis of the Asian Ape of Consolida (Ranunculaceae), A, Synopsis of the Asian Haein of Delphin- ium, Sensu Stricto, A, 249-302, 476- 54 Systematics of Eucnide (Loasaceae), Floral Biology and, 56 Tabe seer “ape 207 Tetracera, 9, 235 — alnifolia, 239 | atiorescens, 239 — asiatica, 239 — portobellensis, 240 — scandens, 19, 240 — volubilis, 19, 240 INDEX 587 Thelypteris, 455 Thistle, European Globe, 450 Thistle, Yellow, 45 THOMPSON, Henry J., and WattaAce R Ernst, Floral Biology and Systematics of Eucnide (Loasaceae), 56 Thurberia thespesioides, 157 Thymelaeaceae, 313 Thymus capitatus, 454 Tiglium, 35 Tilia americana 445, 448 Tiliales ce ace usneoides, 453, 458 Tithymalopsis, 397 Tithymalus, 397, 428 Tomuinson, P. B., and Martin H. Zim- MERMANN. Anatomy of the Palm Rhapis, excelsa, IV. Vascular Develop- ment in a of Vegetative Aérial Axis and Rhizome, 122 Tooth ke. rent 444, 448 wW ~ = Oo ? — sect. Leptobotrys, 377 —sect. Tragia, 376, 377 — volubilis, 376 Triadica, 391 hoses cuneata, 206 ae: anthes aera 454 Trochodenda les ro pete Seat "Plant ro a Some Problems of, 105 ae ese 453 Tung Tree, Two New Cates Palms, 143 Typha angustifolia, 457 Ulmus americana, 445, 447 Umbrella Tree, 448 Vaccinium arboreum, 455 Vantanea, 207 Vascular Development in Apex of Vege- tative me ial Axis and Rhizome. Anat- the Palm Rhapis excelsa, IV, 122 Vatairea lundellii, 117 Verbascum thapsus, 450, 452 Viburnum prunifolium, 446, 449 fi 49 Virginia Creeper, 455 588 JOURNAL OF THE ARNOLD ARBORETUM Virola, 207 Vitis labrusca, 446 —riparia praecox, 446 — vinifera, 446 Vochysia, 207 Walnut, oy Water A 7 Water pel 445, 447 Waterlily, “inti White, 4 WEssTER, Gravy L. The re ra of Eu- horbiaceae in the Southeastern United States, 303-361, 363-430 White Cedar, SC mane 447 White Gum “White Man’ s Footsteps 454 White Oak, 4 White safe ae 452 White Wood, we WwW Oo Willow Oak, 442 [voL. 48 Wine Grape, 4 Wine ae European, 446 Wood Anatomy. Comparative Morpho- logical Studies in Dilleniaceae, I, 1 Wood Sorrel, Common, Wood Sorrel, Violet, 451 Xylophylla, 332 Yaupon, 455 Yellow. Jessamine, 457 Yellow Thistle, 455 Zanthoxylum americanum, 444, 448 —clava herculis, 444, 448 Rhapis excelsa, IV. Vascular Develop- ~ in Apex of Vegetative Aérial Axis Rhizome, 122 From by lite 397