JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY EpItorRtAL BOARD A. C. SMITH, Editor P. C. MANGELSDORF I. W. BAILEY E. D. MERRILL J. H. FAULL H. M. RAUP I. M. JOHNSTON A. REHDER C. E. KOBUSKI K. SAX VOLUME XXVIII JAMAICA PLAIN, MASS. Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION N Ty > 1968 eae OC, DEC 121969 LZ A er AP . f ‘Oo Ml37 DATES OF ISSUE No. 1 (pp. 1-144, 4 pl.) issued January 15, 1947. No, 2 (pp. 145-260, 3 pl.) issued April 15, 1947. No. 3 (pp. 261-374, 1 pl.) issued July 15, 1947. No. 4 (pp. 375-468) issued October 15, 1947, Printed in U.S.A. TABLE OF CONTENTS STUDIES IN THE THEACEAE, XV. A REVIEW OF THE GENUS ADINANDRA. By CAGP ONCE, OUR ion Bis hee loe 6 4,b:8 5 ORG DE aa nee BOR MSs eek 1 THE CYPERACEAE COLLECTED IN NEW GUINEA BY L. J. Brass, I. With foun texistigures: (By So TP . Biake 65 oie ck kd eeeeva Pawnee Meee dawn e oee 99 Notes ON SOME PLANTS OF CuBA. With two plates. By Richard A. Howard 117 RELATIONSHIP AND TAXONOMY OF THE GENUS BRANDISIA. By Hui-Lin Li.. 127 THE CYTOGENETICS OF GENERIC HyBrips OF Sornus. With one plate. By Hally J. Sax and Karl Sax 137 A FiytaAn AcmopyLe. With one plate. By John 7. Buchholz and Netta LF. Oe ce eae oa ae Ae SRE Ge 56-55 6 o 41g 5-0 ic A RR eG 41 A MONOGRAPH OF THE GENUS STREBLOSA KORTHALS (RUBIACEAE). With two text-figures. By C. E. B. Bremekamp.........0...0..0000002... pa) Pao A MONOGRAPH OF THE GENUS POMAZOTA RIDLEY (RUBIACEAE). By C. E. Be ea are ate eae ia ss eh ee ee 186 SIDEROBOMBYX Brem. Nov. GEN. RUBIACEARUM HEDYOTIDEARUM. — By ss aS aE MN AINE od gh sy Ae Sa Nt, Goh hd Md Sw A Le 204 THE CYPERACEAE COLLECTED IN NEW GUINEA BY L. J. Brass, Il. With two plates and one text-figure. By S. 7. Blake... ...00000000.00...00... 207 NeW AND NOTEWORTHY CHINESE FAGACEAE. By Woon-Young Chun....... 230 A New LactTIFEROUS PLANT IN SoutTH CHINA. By Woon-Young Chun and De Pe aa liye + 2 24. ¢ 4 i en one oe eee eea Grad 245 NOTES ON THE DATES OF ISSUE OF DESVAUX’S JOURNAL DE BOTANIQUE. By E. D. Merrill. ....... 24 ON THE IDENTITY OF BRANDISIA SOULIFI BONATI. With one plate. By EF. D. Merrill She, tA MREME beg Sak 4 4c % Sod 9h dae ee nc ge ASN 251 NOTES ON SOME CULTIVATED TREES AND SHRUBS, V. By Alfred Rehder..... 20 Notes ON Two SPECIES OF ARAUCARIA IN NEW GUINEA AND A PROPOSED NEW SECTION OF THE GENUS. By C. T. While... 2.202060 ee 259 A) MONOGRAPH OF THE GENUS ACRANTHERA ARN. EX MEISN. (RUBIACEAE). With two text-figures. By C. EF. B. Bremekamp.........0.0.00.00.000000.... 261 TROPICAL FERN Hosts or Rust Funer. By J. A. Faull... 2.00.0... 309 SPECIES NOVAE FAGACEARUM SINICARUM. By Woon-Young Chun and Ying SOON 5.2 6H AD ie bla bd hg: ea aks a lw hee 320 TABLE OF CONTENTS NOTES ON CHINESE STERCULIACEAE. By Woon-Young Chun and Hsiang-Hao WORE aco 5 kk eae GO ee eS PGE ORS PoE AREAS ones teanadd KAJEWSKIELLA, A NEW RUBIACEOUS GENUS FROM THE SOLOMON ISLANDS. With one plate. By FE. D. Merrill and L. M. Perry... 0.000.000.0000... Two NEW FLOWERING PLANTS FROM St. Lucia. By A. C. Smith and H. A. CS | a ae oe ee ee ASTRAGALUS IN ARGENTINA, BOLIVIA AND CHILE. By Jvan M. Johnston... SECOND SUPPLEMENT TO THE SPONTANEOUS FLORA OF THE ARNOLD ARBORE- TUM. By Ernest J. Palmer. occ ce ea cent sbusnanes Notes ON A NUMBER OF NEW GUINEAN SPECIES. By C.G.G. J. van Steenis. . ” oo 336 419 ENDLICHER’S “(GENERA PLANTARUM,” ‘“ICONOGRAPHIA GENERUM PLANTA- oo... 424 "AND “ATAKTA BOTANIKA.”” By William T. Stearn ON THE IDENTITY OF THE GENERA CUPULISSA RAF. AND PLATOLARIA RAr. By Ek. D. Merrill and N. VY. Sandwith............... ee STUDIES IN THE THEACEAK, XVI. BIBLIOGRAPHICAL NOTES ON THE GENUS LAPLACEA. By Clarence E. Kobuski.. 00.00.0200... re New PLANTS FROM NORTHEASTERN SONORA, MEXICO AND NOTES ON EXTEN- SIONS OF RANGE. By Stephen S. Whit — ound peusiice awa § Norres ON THE Asiatic FroraA. By Hui-Lin Liv... .0 00... NOTES ON SOME CULTIVATED TREES AND SHRUBS, VI. By Alfred Rehder... | THE ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JUNE 30, 1947. BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND STUDENTS, 7 ree Jury 1, 1946-Juner 30, 1947...0000000., STAFE OF THE ARNOLD ARBORETUM, 1946-47....................... EOIN oss asd tg hho oP OE EPL ERE AEG kbd 0d 4s A tal cea en heed TITLE-PAGE AND TABLE OF CONTENTS................ 430 445 447 JOURNAL OF THE ARNOLD ARBORETUM VoL. XXVIII JANUARY, 1947 NuMBER 1 STUDIES IN THE THEACEAE, XV A REVIEW OF THE GENUS ADINANDRA CLARENCE E. KoBuskI THE GENUS Adinandra was described by William Jack in Malay Misc. 2(7): 50. 1822. At that time two Malaysian species, 4. dumosa and A. sylvestris, were described. Today the type-species, A. dumosa, has the widest distribution and is probably the best known of all the species in the genus. Eighteen years later (1840), P. Korthals added five new species to the genus and described at the same time Sarosanthera, a new genus closely related to Adinandra. This later genus was short-lived. It was reduced to a synonym of Adinandra and the single species, S. exce/sa Korthals, was given the name A. Sarosanthera by Miquel (FI. Ind. Bat. 1(2): 477. 1859), who used the specific name Sarosanthera since Korthals (1840) had described earlier a species named A. excelsa. The generic synonymy of Adinandra is very simple. The only described generic svnonym is Sarosanthera, mentioned in the preceding paragraph. However, both Szyszylowicz (1893) and Melchior (1925) in their respec- tive treatments of the Theaceae in Nat. Pflanzenfam. listed Haenkaea Usteri, Ann. Bot. 6: 117. 1793, as a synonym of Adinandra without any explanation. In Index Kewensis (vol. 1), under Adinandra Jack, is cited the synonym Haenkaea Usteri. This is merely a mechanical error which occurred in the compilation of the material for Index Kewensis. It is strange that the earlier date 1793 caused neither Szyszylowicz nor Melchior to investigate the generic status and the possible species described under Haenkaea. Haenkaea Usteri, in the alphabetical arrangement of Index Kewensis, is referred not to Adinandra Jack of the Theaceae, but to Adenandra Willd. of the Rutaceae, where it truly belongs. By 1868 nearly twenty species, all from the Malay Peninsula or Nether- lands East Indies, had been added to the genus by various authors. Asa result the genus was considered to be localized to these areas. In 1868 Oliver (Fl. Trop. Afr. 1: 170) described A. Mannii from the island of St. Thomas in the Bay of Guinea off the west coast of Africa. This species 2 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII is quite different from the majority of species in the genus and several workers have suggested the possibility of generic segregation. A close study of the entire genus shows that the variations in A. Mannii are not sufficient for generic limitation. In 1874 Dyer described A. Griffithii from East Bengal, and in 1878 Cleyera Millettii Hook. & Arn., of China, was transferred to the genus, thus extending the range to its northern limits. In 1905 the genus was first recorded from the Philippines, when Merrill introduced J. baiag It received its southernmost range in 1925 when Baker described . Forbesti from New Guinea. From 1822 until 1893 there had been no attempt on the part of any worker to treat all the species of the genus in a group. Frequently the genus was listed in local floras, but further than that little was offered as a composite study. In 1893 Szyszylowicz, in his treatment of the Theaceae for Nat. Pilanzenfam., although in very abbreviated form, listed the species of Adinandra and offered a key for their identification. He introduced three new sections and separated them as follows: Sect. I. Eleutherandra: Stamens free. Sect. IIT. Euadinandra: Stamens in bundles. Sect. III. Symphiandra: Stamens joined and disposed in a tube, It is difficult to understand, considering the material available at the time, how Szyszylowicz could separate the species into Sections II and Il. He lists 4. Sarosanthera Miquel in Sect. II and A. lamponga Miquel and A. macrantha Teijsm. & Binn. in Sect. II. I consider all three con- specific. I have studied an isotype of A. Griffithii Dyer and know only on the word of the author that the stamens are uniseriate. e lists A. Miquelii King, a combination never made by King, under Sect. ITI. Adinandra eon is a synonym of Ternstroemia bancana Miq. Melchior (1925), in the second edition of the same publication, accepted the sont of Sine and added a fourth section as follows: Sect. I. Eleutherandra Szyszy.: Style entire. Filaments free. Sect. Il. Euadinandra Szyszy.: Style entire. Filaments in bundles. Sect. IIT. Symphiandra Szyszy.: Style entire. Filaments joined and disposed in a tube. Sect. IV. Eleutherostyla Melch.: Styles free. As noted, Melchior added a second character, the entire style, which combined the first three sections and separated them from his new section, Eleutherostyla, which was created to include A. calosericea Diels (1922), described three years earlier. In 1940 this new section was eliminated by the creation of a new genus, Archboldiodendron, to which Adinandra calosericea was transferred. The entire style cannot be attributed to all members of the other three sections created by Szyszylowicz because a few species are known to have styles that are three-, four-, or five-parted. Adinandra Griffithii Dyer (Bengal), A. lasiopetala (Wight) Choisy (Cey- 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 3 lon), A. filipes Merrill and A. nitida Merrill (China), and A. Macgregorii Merrill (Philippine Isl.) are known to have parted styles, whereas in A. quinquepartita Kobuski (Borneo), in which the floral parts are described from a bud, the stigmas are distinctly five in number and the style, though apparently entire, is deeply five-sulcate. However, in his treatment Melchior states that the knowledge of the stamens is too incomplete to group the species by sections, especially in Sections II and III, and so he groups them in geographical regions, namely Ceylon, Indo-Malaya, East Asia, and Papua. This proved to be an excellent method of treatment. After a study of all the species described to date, I find that the majority of species, save perhaps A. dumosa Jack and A. Millettii (Hook. & Arn.) Benth. & Hook., are of rather limited range. These two mentioned species, although of more extensive geographical distribution, remain in their respective areas. Although whenever possible the stamens are described in detail, I find it unfeasible to treat the species under the sections as outlined by Szyszylowicz and Melchior. Eight out of every ten herbarium specimens, and this is a conservative estimate, have been collected in the fruiting stage, and, as a result, many species have been described from the fruit alone. Further- more, it is difficult, in working with herbarium material, to place a species definitely in its respective section even with the stamens present. In this study I have employed the same method of treatment as used by Melchior and feel that in this manner the species can be identified more easily. In preparing the keys, I have avoided the use of the stamen characters when possible. sea c ie Jack in Malay. Misc. 2 (7): 50. 1822 [repr. in Hooker, Comp. Bot. Mag. 1835; in Calcutta Jour. Nat. Hist. 4: 207. 1843, et in Miscel. Papers Indo- a a 2: 271. 1887]. — Korthals, Verh. Nat. Gesch. Bot. ed. Temminck 103. 1840. — Choisy in Mém. Soc. Phys. Hist. Nat. Genéve 1: 111 (Mem. Ternstr. 23). 1855. — Miquel, Fl. Ned. Ind. 1 (2): 477. 1859; Fl. Ned. Ind. Suppl. 1: 478. King in Jour. As. Soc. Bengal 59 (2): 187 (Mater. Fl. Malay Penin. 1: 127). 1890. —Szyszylowicz in Nat. Pflanzenfam. III. 6: 189. 1893.—Koorders & Valeton, Meded. ’S Lands Pl. 16: 219 (Bijdr. Boomsoort. Java 3: 219). 1896, — Ridley, Fl. Malay Penin. 1: 193. 1922. — Melchior in Nat. Pflanzenfam. ed. 2, 21, 143;, 1925. Craib, Fl. Siam. Enum. 1: 124. 1925.— Metcalf in Lingnan Sci. Jour. 11:18. 1932. _- Melchior in Notizbl. Bot. Gart. Mus. Berlin 11: 1076, 1097. 1934. — Gagnepain in Not. Syst. Mus. Hist. Nat. Paris 10: 112. 1941; in Fl. Gén. Indo-Chine, a 1: 282. 1943. Surocantherc Korthals, Verh. Nat. Gesch. Bot. ed. Temminck 103. 1840. Flowers aoe Sepa ls 5, imbricate, persistent, thick, concave, accrescent, unequal, gradually increasing in size from the outer lobes to the inner lobes. Petals 5, imbricate, connate at the base, glabrous or sericeous on the outer surface, the inner surface glabrous. Stamens glabrous; anthers oblong, erect, basifixed, extrorse, hispid, very rarely 4 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII glabrous; connective projected into an apicule. Ovary pubescent or glabrous, 3- or 5-loculate, rarely 2- or 4-loculate, the ovules usually oo (20-100) in each locule, rarely few (4), affixed with prominent placentae. Style 1, persistent, entire, rarely 3—5-fid; stigma 1, rarely 3-5, entire or lobed. Fruit indehiscent. Seeds o to few, usually minute, dark, shining, scrobiculate, the embryo inflexed, the cotyledons semi-terete. Evergreen trees or shrubs, alternately branched. Leaves alternate, coriaceous, occasionally chartaceous. Flowers axillary, solitary, frequently in pairs; peduncles 1-flowered, recurved, seldom erect, 2-bracteolate, the bracteoles at the apex of the pedicel, opposite or alternate, persistent or caducous. Type spectes: Adinandra dumosa Jack. An amplification of the foregoing formal description, with observations on the characters used in specific delimitation throughout the genus, follows: Leaves: Throughout the genus the leaves seldom furnish sound specific characters. Ordinarily the leaf is coriaceous. A very few species are known to possess chartaceous or subcoriaceous leaves. In size and shape one may find considerable variation on a single specimen. In most in- stances the leaves are obtuse, bluntly acuminate at the apex and tapering at the base. Occasionally there are consistent variations from this pattern, as found in the long-acuminate leaves of A. acuminata Korthals and J. Bockiana Pritzel and its variety, the cordate leaf-base of A. cordifolia Ridley and its variety, and the subsessile leaf-base of A. subsessilis Airy- Shaw. In these few instances specific delimitation may be based on the leaves. The terminal leaf-buds and very young unfolding leaves possess charac- teristics helpful in indicating the degree of pubescence on the leaves. In any group of plants with coriaceous and evergreen leaves there is apt to be a question concerning the interpretation of the degree of pubescence of the leaves. The mature leaves may have been glabrous in the juvenile stage or they may have been extremely pubescent, later becoming glabres- cent or even glabrous. In Adinandra, observation of the terminal leaf-bud offers an accurate means of determining the actual degree of pubescence or glabrosity of a given species. There are only a very few glabrous species, and in these cases only are the terminal leaf-buds strictly glabrous, The veining as a whole is not an outstanding character, nor is it con- sistent. The majority of species fall into a group where the veins vary from 8 to 15 pairs. Very often the veins are obscure and of little use in identification. Exceptions may be found in A. luzonica Merrill, where even the veins of lesser degrees are of - same prominence as the primary veins, and in A. myrioneura Kob., polyneura Kob., A. dasyantha Korthals, and A. phlebophylla Elance. rae the veins exceed more than twenty pairs in number, thereby presenting a conspicuous feature. The midrib is usually flattened or plane on the upper surface and semi-terete on the lower surface. In two species, A. verrucosa Stapf and A. collina Kob., the midrib on the lower surface is bisulcate. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 5 Pedicels: The pedicel length is 1-2.5 cm. for the majority of species. Unless the pedicel is less than 0.5 cm. or more than 3 cm. long it can hardly be used as a delimiting factor. The pedicel is seldom erect, usually being recurved. Pubescence on the pedicel plays the same rdle as pubescence elsewhere. Bracteoles: The persistence of the bracteoles is important. The greater number of species have persistent bracteoles, while in a smaller group of species the bracteoles are caducous. In the latter species the bracteoles may fall so early that their presence is recognized only by the scars, or the bracteoles may remain until anthesis before dropping away from the flower. Only in the persistent bracteoles and those remaining until anthesis can one find characters to be used for identification. Their position, shape and texture are of importance. In the majority of species, the bracteoles are opposite, at the apex of the pedicel, immediately below the calyx-lobes. When they are alternate, and alternate here means a variation up to 5 mm. in their positions, it is in the outer bracteole where the variation in position may be found. The inner bracteole arises imme- diately below the calyx-lobes, whereas the outer one is placed lower on the pedicel. In shape the bracteoles usually simulate the calyx-lobes. Often the two bracteoles vary considerably in size. Calyx: Always persistent, the calyx-lobes vary considerably from the period of anthesis to that of mature fruit. Hence size, unless very small or exceedingly large, is of little importance. There is a marked variation in size in the lobes of a single calyx. The outer lobes are smaller, often more pubescent, and of heavier texture. In many species there is a grada- tion in size starting with the outer or smaller lobes and increasing in size to the largest or inner lobe. The margin of the inner lobes often is considerably more scarious than that of the outer lobes. The texture is of diagnostic importance. In the Eastern Asiatic species and those of the Philippine Islands, the texture of the calyx-lobes is of a decidedly thinner quality than can be found in the other groups or more tropical regions where the lobes are definitely thicker, often subligneous. Corolla: The corolla remains as part of the flower for only a very short period. While still present it offers good characters in color, shape, length and pubescence. The lobes are always connate at the base; hence when dropping off the corolla falls in its entirety. White or cream is the color for the majority of species. Occasionally the corolla may be tinged with pink, and in British North Borneo and New Guinea a few species have purple lobes. In length the corolla may vary from a measurement equalling or slightly exceeding that of the calyx to three times the length of the calyx, as found in the two African species, A. Mannii Oliver and A, Schliebenii Melchior. Pubescence is an excellent character in the corolla. The species can be divided almost equally by this character. When present, the pubescence is usually of the appressed type and is concentrated on the median portion of the external surface. Occasionally one may find species where the pubescence extends over the complete 6 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII external surface. Unlike that in Gordonia, the pubescence is never found on the interior surface Stamens: The stamens offer many excellent characteristics, when present, to aid in identification. However, being adnate to the base of the corolla, they usually fall away with the corolla, They are arranged in the flower in series which vary in number from one to five. When in series, one finds that the filaments vary in length according to the number of series. The filaments are rarely free, but are connate for at least part of their length, usually that of the shortest filaments. Extreme difficulty is experienced working with dried material to determine the exact number of series. Stamens in a single series may average as low as fifteen in number (A. Sana Oliver and 4. Schliebenii Melchior), while those in several series will number as high as 55-60, as found in A. Brassii Kob. and A. Forbesii Ba Variation in pubescence is considerable. In some species the filaments are glabrous, in others they are densely pubescent, and in still others one may find them pubescent on the upper portion adjacent to the anthers or on the a surface only. The anthers are oblong, and with the exception of A. Mannii and A. Schliebenii are always extrorse-hispid. In only the two a neater species are the anthers glabrous. In some instances an elongated apicule may be of diagnostic importance. Pistil: Some of the best delimiting characters are found in the pistil. The style is persistent and either the ovary or fruit is present in every fertile specimen. The number of cells of the ovary has been little used even in descriptions. In immature flowers a cross-section is often difficult to obtain. In all species studied for this treatment the number of cells in the ovary and fruit are recorded. In a very few instances there is a variation within the species of number of cells in the ovary, and in such rare instances it is sometimes difficult to ascertain which is the usual number and which is the variation. The cells are for the most part quite clearly defined as to number. In several instances the placentae are of such length and consistency that they may be mistaken for cell-walls and as such they have obviously been interpreted in the past. Also; in a few species ‘the cell-wall opposite the placentae may project so far into the cell that the branches of the placentae extend beyond the resulting indentation of the wall. A cursory study of such a dissection may very easily be misinterpreted. The fact that the ovary and the fruit are so filled with either ovules or seeds may very well lead to erroneous conclusions. Also, in rare instances the seeds may crowd the cells to such an extent that the walls are broken down and the seeds tend to displace the central axis or push into adjoining cells. However, these variations are exceptional. The number of cells in the ovary and fruit of the majority of species is five. In all species where the ovary is five-celled the ovules are small and very many in number. ‘The seeds, correspondingly, in the fruit are minute and usually number over a hundred. On the other hand, when the ovary and fruit are three-celled, the number of ovules and seeds is fewer and the seeds are larger than those in the five-celled species. The seeds here 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV fi measure tip to 8 mm. long and seldom exceed ten in number for the entire fruit. Exceptions to this are rare but are found in A. javanica Choisy and A, lasiopetala (Wight) Choisy, where the fruit is three-celled but the seeds are minute and for all practical purposes countless. A few species have four-celled or two-celled ovaries. Of the two-celled group, A. sudbsessilis Airy-Shaw is interesting in that a single large seed is developed in each cell of the fruit. Pubescence of the ovary plays an important role in identification. When pubescent, the ovary is usually densely so. This pubescence may be evanescent and the resultant fruit may be glabrescent. However, in such instances one may always find a few stray evidences of pubescence at the apex of the fruit just below its juncture with the style. The style may be glabrous or pubescent. If the ovary is glabrous the style is also always glabrous. However, with a pubescent ovary one may find either a pubescent or glabrous style, depending upon the species. In any given species, however, the styles are usually consistent. In the majority of species the style is entire. However, in a few instances the style is divided into three or four parts, in accordance with the number of cells in the ovary. When the style is parted the number of stigmas is the same as the number of parts to the style. Ina few species the style may be entire and the stigma lobed, in accordance with the number of cells in the ovary. In other species the style may be entire and the stigma distinctly five-parted. The following review of species has been divided into five geographical groups, namely: (1) Eastern Asia; (2) Indo-Malaysia; (3) Philippine Islands; (4) New Guinea; and (5) Africa. Material has been borrowed from most of the leading American herbaria where the genus Adinandra has been deposited in abundance. Added material on critical specimens was obtained later from the Royal Botanic Gardens at Kew. The author is deeply grateful to the curators of these institutions for the privilege of examining their material. In the citations of specimens throughout this paper these institutions are abbreviated as follows: (A) Arnold Arboretum of Harvard University. (C) Chicago Museum of Natural History. (G) Gray Herbarium of Harvard University. (K) Royal Botanic Gardens, Kew, England. (M) Missouri Botanical Garden. (NY) New York Botanical Garden. (PBS) Philippine Bureau of Science, Manila, P.I. (US) United States National Herbarium. EASTERN ASIATIC SPECIES In reviewing the species of Eastern Asia and comparing them with those of the other geographical regions, as arranged in this study, one is im- pressed by the close relationship of all the species of this area and the basic similarity of their characters. Specific differentiation is not as clear 8 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII cut in this area as in the Indo-Malaysian area. Many species have been described on the basis of rather feeble variations. The focal species is A, Millettii (Hook. & Arn.) Benth. & Hook. Herbarium material has been rather plentiful for the Chinese species, and further material may show that the number of species listed below, although conservative, is not conservative enough. The species described from Formosa and the Liu Kiu islands are very closely related to A. Millettii and its varieties. However, little could be done with these species, since the descriptions are often too meagre, and sufficient material for study lacking. Almost equally sparse have been herbarium specimens from Indo-China. During the course of this study, there appeared a belated (although pub- lished in 1942) study of the Theaceae of Indo-China by Gagnepain (Not. Syst. Mus. Hist. Nat. Paris 10: 112-131. 1942). In this paper seven new species have been described. Fortunately, types of four of Gagnepain’s species were found in the material gathered for this study and are treated in this section of the review. The remaining three, A. annamensis, A. caudata, and A. donnaiensis, are quoted at the end of this section, The original descriptions and citations are inserted in this paper in order to make these records as complete as possible. Material is essential before an attempt be made to relate them to better known species. Immediately following is a chart of the characters used in separating the majority of the species of Adinandra. Under each element are listed the differences, and the species are recorded, whenever possible, under these differences. This chart has been helpful to me in the present study and it is hoped that it may be of assistance to other workers. Accompanying charts will be found in the two following groups. CHART FOR EASTERN ASIATIC SPECIES OF ADINANDRA Bracteoles Calyx-lobes PERSISTENT CADUCOUS PUBESCENT GLABROUS Bockian Bockiana Bockiana Bockiana tonkinensis Bockiana tonkinensis Bockiana epunctata acutifolia epunctata acutifolia glischroloma filipes filipes Howi irta glischroloma glischroloma laotica Millettii glischroloma glischroloma nitida obtusissima jubata hirta VWillettii nitida glischroloma glischroloma Millettii ryuk yuensis macrose pala jubata formosana Zen-Tasi hainanensis glischroloma Millet tii owit racrose pal obtusissima laotica hainanensis lasiostyla asiostyla megaphylla megaphylla Villettii btusissima Millet tii rvukyuensis formosana yvae yamensis vaevamensis Zen-Tasiroi 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV Corolla-lobes Filaments PUBESCENT GLABROUS PUBESCENT GLABROUS Bockiana glischroloma glischroloma Bockiana Bockiana hirta hirta Bockiana acutifolia laotic acutifolia Bockiana Millettit Bockiana tonkinensts Millettii tonkinensis epunctata formosana epunctata glischroloma Millettii glischroloma glischroloma obtusissima glischroloma jubata nitida ubata glischroloma glischroloma macrosepala macrose pala hainanensis hainanensis lasiostyla laotica megaphylla lasiostyla megaphylla Millettii Millettii formosana Millettii obtusissima itida Ovary Style PUBESCENT GLABROUS PUBESCENT GLABROUS Bockiana Howiti epunctata Bockiana Bockiana laotica ipes Bockiana acutifolia nitida glischroloma acutifolia Bockiana rvukyuensis glischroloma Bockiana tonkinensis Zen-Tasiroi hirta tonkinensis epunctata glischroloma Howii ] jubata laotica glischroloma glischroloma Millet tii glischroloma macrose pala Millettii irta ainanensis formosana glischroloma lasiostyla nitida jubata megaphylla ryukyuensis glischroloma Millettii yaeyamensis macrose pala obtusissima Zen-Tasiroi hainanensis lasiostyla megaphylla Millet ti Millettii formosana Millettii yaeyamensis JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Ovary FIVE-CELLED THREE-CELLED glisclroloma Paring glischroloma Bockiana hirta acutifolia glischroloma Bockiana jubata tonkinensis lischroloma epunctata macrose pala ili hainanensts Howii aotica lasiostyla megaphylla Millet tii Mille ttii formosana Millettii obtusissima nitida obtusissima KEY TO EASTERN ASIATIC SPECIES A. Ovary pubescent; fruit pubescent or obviously glabrescent. B. Style glabrous. C. Calyx-lobes oe long-acuminate, widest near the base; corolla- lobes glabro D. Leaves cial entire; pedicels up to 2 cm. long. = vier Bev caSussice a nasal actin been ale stghoete bia nogeaiaciva dunce-g ist aetdes acace at eats ax | Millettii. DD. Leaves usually serrulate; pedicels up - 3 vi long. (For ae sig natetacs agua ale Gud Gerd 4 alec Sate btu Rade ease eee Millettii var. formosana. CC. Calyx-lobes rounded or broadly obtuse, ee tie occasionally rounded and sharply apiculate, not long-acuminate, widest portion remote from the base; corolla-lobes pubescent D ples al buds, the un nder surface of the leaves, the current season’s wth of branchlets, and the flower parts covered with a loose, ake pubescence in varying degrees of density, not appresse E. Pubescence dense, persistent on the above-mentioned ees (Kwangsi, Kweichow, Szechuan).................. . Bockiana. Ek. Pubescence dense only in the young stages with a distinct tendency toward glabrescence at maturity F. Leaves long-acuminate, ie margins gore (Fukien, Hunan Kwangsi, Kweichow)............ 2a. A. Bockiana var. icattiolid. FF. Leaves obtuse at the apex, the margins rae >, (Liu Kiu SIATIOS Jick Ges 06 ahyetins wis se Ps atc ete Acosenaieeauatn uk 14. A. vaeyamensis. DD. Terminal buds, the under surface of the leaves, the hes season’s sie of branchlets, and the flower parts covered with a short ap- essed eal (Indo-China)...2b. A. Bockiana var. tonkinensis. BB. te pubescent ves large, 20-24 cm. long, 4-7 cm. wide, the aaa on surface ae finely pubescent............... 000.000.0000. 3. 2 A. megaphylla. CC. Leaves seldom 15 cm. . long, the midrib on upper surface soe ee brous. cence on under surface of leaves (especially young leaves) very long and dense, projecting beyond the margin, noticeable from the upper surface. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 11 E. Bracteoles caducous; style densely pubescent two-thirds the entire F. Pubescence on under surface of the older leaves tending toward glabrescence along the midrib and blade, maintaining denseness along the margin. G: se ae mm, long, occasionally up to - mm. long in ngtung, Kwangsi)........... : glischroloma. GG. eae lobes ac 14 mm pe (Chekiang, ee Pe 53%, 8 Ree ee A. glischroloma var. macrose pala. FF. a ia on the under bie: of the older as well as the unger leaves a dense ferrugineous iridescent tomentum up to ae m (Kwangtung)..... 4b. A. glischroloma var. jubata. EE. Bracteoles persistent; style densely pubescent the entire length. Cimdo-Chinia) .i..5 iss a'ak wo se catandan 4c. A. glischroloma var. hirta. DD. Flowers rainy solitary ; pubescence on the under surface of the Teaves, when present, not pe aieetinn beyond the margin so as to be noticeable from upper surface. E. Style 3- mer pedicel ca. 4 cm. long. (Kwangsi).....5. A. filipes. EE. Style entire; pedicel 2.5 cm. long or less. eaves distinctly dark punctate-dotted on the lower surface; ovary and fruit 5-celled. (Hainan, Kwangtung).............. Be Ie ics Sati te Gres a ee sl a ee BA ee 6. A. hainanensis. FF. Leaves without dark punctate-dots on the lower surface; ovary and fruit 3-celled. G. Bracteoles caducous. (Formosa) 2.2.24 40003 7, A. lasiostyla. € dicels short, seldom over 0.7 cm. long; filaments pubescent; fruit small (ca. 5 mm. diam.), the seeds normal in siz e (2-3 mm. long), but few (9-10) in number. GEiagina mm) eee res Saeco ee 8. A. epunctata. HH. Pedicels up to 2.5 cm. long; filaments glabrous; fruit many-seeded (100 or more). (Formosa)............... Re er ee eee ee lb. A. Millettii var. obtusissima. apts eM ite Saletan nate aaa sane ees 15. A. obtusissima. AA. Ovary and fruit glabro . Entire plant etnus exteiek for the terminal buds and the branchlet growth of current year. C: Ovary and: fruit 5-celledS “Gindo-Ghina)) i acu a eyed 9. A. laotica CC. Ovary and fruit 3-celled. D. Style 3-parted; calyx-lobes ca. 15 mm. long, 7—9 mm. bees bracteoles persistent. (Kwangsi, Kwangtung)..................-. A. nitida. 1B) Dis ele entire; calyx-lobes ca. 8 mm. long, 5 mm. ae bracteoles cous. aes Se Ore ee ees Howit. BB. Under oa of th ves, pedicels, and calyx-lobes, as well as the termina) buds and aie. ae Clit Kansans) ae. oe eee ict wa a 12. A. ryukyuensis. 3. A. Zen-Tasiroi. Paar ir er ce Ne ee Se Dee ee aS 1. se tre ras (Hooker & Arnott) Bentham & Hooker f. ex Hance in Jour. Bot — Maximowicz in Bull. Acad. Imp. Sci. St.-Petersb. 31:17. 1886 oe Tuk ins. spec.). —Szyszylowicz in Nat. Pflanzentam. III. 6: 189. 1893, — Melchior in es Pflanzenfam. ed. 2, 21: 144. 1925.— Rehder & ees in toma mold Arb. 8: 177. 1927. ae in Lingnan Sci. Jour. 11: 30. 1932. — Aerrill in Jour. ae Arb, 19: 55, Clevera ae Hooker & Arnott, Bot. ee Voy. 171, t. 33. 1841. Adinandra Drakeana Franchet in Nouv. Arch. Mus — ser. ‘ 208 (PI. David. 2. — Szyszylowicz in Nat. Pflanzenfam. TI. 6: 12 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII — ili Handel-Mazzetti ex Metcalf in Lingnan Sci. Jour. 11: 19. 1932, n obs., nome DisrripuTioN: China (Kwangtung, Fukien, Anhwei, Chekiang, Kiangsi, Kwangsi). KWANcTuNG: Summit of Pakwan Mt., above Canton, 7. Sampson sae (G), May 1870.— White Cloud Hills, Canton, ai summit, 7. Sampson 1239 (C), May 25, 1879. — Canton and vicinity, C. O. Levine 613 (C, G, M, US), April 24, ps 1230 (M), Aug. 30, 1917. Loh Fau Shan, C. O. Levine 2069 (A, M), May 8, 1918. — Poon Yue Distr., C. O. Levine 3199 (A, G, M), Sept. 4, 1918.—San-on Distr. Ng-Tung Shan, mountain forest, alt. 900 m., T. M. Tsui 219 (M, NY, US), April of A. Hemsleyi, A), Sept. 5, 1917. — Tapu Distr., Tai Mo Shan, Wonk Pan Ch’a, dry steep slope in sandy ie - T. Tsang 21076 (A, NY), July 5, 1932 (shrub 2 m. with edible black fruit) .— i (Kaying) Distr., Yam Na Shan (Yit Nga Shan), Wong Pan Ch’a Shue, along ee of dry, clay meadow, W. T. Tsang 21353 (A, NY), Aug. 4-31, 1932 (erect, woody, 3 m. with edible black fruit). — Wung Yuen Distr., Tsing Wan Shan, Wong chuck I and vicinity, clay thickets on steep dry slopes, S. K. 2025 (A), Aug. 16-31, 1931 (woody, 8 m.).— Hongkong, New Terr., Wu-kau- eg in open woods, Y. Tsiang 2965 (A, C, NY, US), Aug. 13, 1929 (low ayes with deep purple fruit). sin-fung Distr., Lo-lo-ha village, Sha Lo Shan, abundant in forest, Y.W. Taam 928 (A), July 1938 G7 m. high; fruit black, edible). — Hongkong, C. Ford s. nm. (A, NY). Fuxkien: Lui Sang, Kuliang and vicinity, rich woods, alt. 800 m., H.C. Chen 10065 (A), july 20, 1925. — Kutien, H. H. Chun ee (A, NY), May 26, 1928. — Ne Lau San, Diongloh and vicinity, wooded hillside, Y. T. Lin 11720 (M), Aug. 6, 1926.— Yenping, Peh Men Wei, near north gate, aa ‘hibade, K. H. Chou 8748 (M), hig 13, 1926, — Pehling, Minhow Hsien, in Spo H. H. Chung 2074 (A), Aug. 3, 1923 Lana 1 m.).— Foochow, I-Su-Shan, H. H. Chung 2814 (A, M), Aug. 13, mo (shrub 2 m.).— Yenping, Cha-ping, on ned alt. 730 m., H. H. Chane 2846 (A), eas 30, 1924 (shrub 1 m.).— Minhow Hsien, Kushan Monastery, in shady ailogs H. H. Chung 2265 (A), Aug. 17, 1923 (tree or shrub 3—4 m.).— Chua nchow, A. Chung 3097 (M), April 23, 1925 lanier — Foochow, Kushan, H. H. Chane ee (A), Aug. 18, 1925 Pola a — Inghok, Fang ven Yen, H. H. Chung 4047 (A), May 10, 1928. — Kuliang, H. H. Chung 6445 (A), July 7, 1926 (shrub with cease Gowers). —~ Kuchan, Seat HH. H. Chung 7826 (A), Aug. 22, 1926. ANHW Li Kan, along banks of aged wooded stream, alt. 100 m., R. C. Ching 3153 (A), Aug. 7, 1925 (shrub 5 m. with smooth, gray bark).—Kimen, N. K. Ip 7671 (US), Aug. 19, 1924. abet Tsing Tsien, on hill near rocks, ¥. L. Keng 79 (A), July 21, 1926 (low shrub with the fruit solitary in the axils).— Tsing Tsien, on mountain near roadside, Y. L. Keng 144 (A), July 26, 1926 (1 m. high). —S. Yentang,’ shady woods, alt. me m., H. H. Hu 227 (A), Aug. 26, 1920 (shrub 2-3 m. with edible black fruit). — 150 li south of Ping Yung, alt. 70 m., R. C. Ching 1984 (A, G, US), July 4, 1924 (shrub 8 m.; nodding flowers with green calyx and corolla, pistil purple). — 15 li northeast of Tai Suan, open bushy ee alt. 500 m., R. C. Ching 2207 (A, US), July 22, 1924 (big shrub 7 m.). Kwancst: Yung Hsien, Ta Tso Shan, on hillside and in forest, alt. 380 & 540 m., A. N. Steward & H.C. Cheo 819, 905 (A, NY), Aug. 1933 (shrub or tree 6-10 m. with whitish gray bark).— Yao Shan, Ping Nan, thin woods near shit ed . Wang yee (A), May 17, 1936 (shrub with white flowers). Kranost: ahaa, A. David s.n. (photo of type of A. Drakeana, A), 1873 (‘petit arbre, a hei blanches, sans odeur”).— Lungnan Distr., Oo Chi Shan, near Lam Uk Tung village, dry steep — in clay forest, S. AK. Lau 4424 (A, ae Sept 1934 (5 m. high). — Kiennan Distr., Sai Hang iacnr. near Tung Lei ie . Lau Sk (A, US), July 1934.— On trail hig Kit-shan and Hong San, ae ie alt. 580 m., i, L. Gressitt 1425 (A, M), June 21, cig (shrub 1 m. high with pink Roweies: mn Feng Hsien, shady aie alt..650 m., H. H. Hu 1012 (A), May 31, 1921 (shrub 1 m. high with white flowers). a Chuan. near Siang-ton-ton, in small woodland, alt. 100 m., Y. Tsiang 9851 (NY), June 17, 1932 (shrub 1-2 m. hint) 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 13 Shrub or small tree. Branchlets glabrous, terete, grayish brown, the very young branchlets and terminal buds pilose in varying degrees. Leaves coriaceous, oblong-elliptic, 4.5—9 cm. long, 2—3 cm. wide, acuminate, bluntly acuminate to nearly obtuse at the apex, acute at the base, the very young unfolding leaves densely pubescent, the mature leaves showing green, glabrous above, lighter green, glabrous or glabrescent, occasionally with light scattered appressed pubescence beneath, the margin generally entire, occasionally (on the same specimen) lightly denticulate along the upper half, the veins obvious on both surfaces but not conspicuous, the petiole short, ca. 3 mm. long, usually glabrous, rarely slightly pubescent. Flowers axillary, solitary; pedicels up to 2 cm. long, slender, glabrous; bracteoles 2, quickly caducous, alternate, near the apex of the pedicel when (rarely) present in the young buds, similar to the calyx-lobes, ca. 2 mm. long; calyx-lobes 5, imbricate, subequal, ovate-deltoid, 7-8.5 mm. long, 4-6 mm. wide, quite sharply pointed at the apex, widest at the base, glabrous or glabrescent, rarely slightly appressed-pubescent on the exterior surface, the margin subscarious, both glandular-denticulate and ciliolate; corolla-lobes 5, connate at the base, glabrous, white, similar to the calyx- mm. long, the filaments ca. 3 mm. long, glabrous, somewhat dilated, measuring nearly 0.5 mm. across, quite free from each other, adnate to the base of the corolla, the anthers 2-3 mm. long, broadly linear, up to 1 mm. across, covered with a dense white pubescence, the apicule 1—1.5 mm. long; ovary nearly globose, tapering slightly at the apex, white-pubescent, 3-celled, multi-ovulate, the style glabrous, entire, ca. 9 mm stigma obtuse, simple. Fruit glabrescent, with evidence of some straggling pubescence, subglobose, ca. 7-8 mm. diameter, many-seeded, the seeds minute, black, shining. Adinandra Millettii is the earliest described species of the genus from China. It has been most often collected, shows the widest distribution, and probably is the oldest Chinese species phylogenetically. In these respects it compares very favorably with the type, 4. dumosa Jack, from Indo- Malaya. The species itself is confined, according to the material available for this study, to the mainland of China; it ranges along the southeastern coast including the provinces of Chekiang, Kiangsi, Fukien and Kwangtung, and extends inland and westward to include Kwangsi and Anhwei. Two varieties are found in Formosa, A. Millettii var. formosana (Hayata) Kob. (the erstwhile A. formosana Hayata) and A. Millettii var. obtusissima Kob. Closely allied to A. Millettii is the species A. Bockiana Pritzel and its two varieties A. Bockiana var. acutifolia (Hand.-Mazz.) Kob. (A. acutifolia Hand.-Mazz.) and A. Bockiana.var. tonkinensis Kob. All these entities have many diagnostic characters in common, i. e. the small flowers, the calyx-lobes thin in texture, the stamens similar in number, size, shape and arrangement, the filaments showing a definite and charac- teristic dilation toward the anthers, and the 3-celled ovary with a distinct spreading grayish pubescence which becomes glabrescent in the fruit. 14. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII The corolla-lobes are similar in size, hardly extending beyond the calyx-lobes. The characters separating the two species are the shape of the calyx- lobes, the pubescence on the corolla-lobes, and the leaf-shape. These characters, none too strong, are consistent, however, in the two species. Whereas, in A. Millettii and its varieties the calyx-lobes are always long- acuminate with the attenuation starting near the base, in A. Bockiana and its varieties the calyx-lobes are more rounded, occasionally abruptly apiculate but not long-tapering. In A. Millettii the corolla-lobes are always glabrous on the external surface whereas in A. Bockiana the corolla- lobes are always pubescent on the median portion of the external surface, In the latter species the leaves are larger and usually long-acuminate. As might be expected, from the extensive range of the species, there is considerable variation in 44. MJillettii, especially in the leaves. One of the earliest collected specimens and perhaps the most cited (Sampson 15598, Pak-wan, Kwangtung) shows, on the same sheet, branchlets in which the leaves vary from obtuse to long-acuminate. This same variation can be found on several other specimens. Pubescence is another variable char- acter. Generally speaking, except for the current year’s growth, 4. Millettti may be considered quite glabrous, quickly losing its early pubescence. However, in some instances this glabrescence is less exten- sive, hence scattered hairs may be found on the undersurface of the leaves, the pedicels, and the calvx-lobes. In describing the stamens, Hooker and Arnott refer to the filaments as hairy and the anthers glabrous. The illustration accompanying the description portrays them thus. In the dissections made for this study, the filaments were always glabrous and the anthers pubescent. The reverse condition, as given ‘by Hooker and Arnott, would be very unusual for this genus. A photograph and fragment of the type of A. Drakeana Franchet, col- lected by A. David in Kwangsi in 1873 and deposited in the Museum d'Histoire Naturelle, Paris, show clearly that 4. Drakeana is synonymous with A. Millettii. Probably Franchet was unaware of the existence of A. Millettii when he proposed his new species, because his suggested rela- tionships are with the Indo-Malaysian species A Palen Jack and A. — T. Anderson rather than with A. Mi other name, found on herbarium ee inee pe never adequately published, is A. Hemsleyi Handel-Mazzetti. Metcalf, evidently with some authority, stated that Handel-Mazzetti had abandoned this name and never published the species because he felt the relationship was too close to A, Millettii. Mell 769 from Kwangtung is a representative of this distribution. la. Adinandra Millettii (Hooker & Arnott) Bentham & Hooker f. ex Hance var. ormosana (Hayata), comb. nov. Adinandra formosana Hayata in Jour. Coll. Sci. Tokyo 22: 45. 1906; in Jour. Coll. Sci okyo 30 (Art. 1): 42. 1911.—Kanehira, Formosan Trees 51, fig. 1917; ‘e) — revised ed. 452, fig. 409. 1936.—- Melchior in Nat. Pflanzenfam. ed. 2, 21: 149 1925. — Sasaki, Cat. Governm. Herb. 346. 1930 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 15 ssh manana Hayata in Jour. Coll. Sci. Tokyo 30 (Art. 1): 43. 1911. — Kan Te ae an Trees 54, fig. 1917. — Melchior in Nat. Pflanzenfam. ed. 2, 2s tae kspocior ee Icon. Pl. Formos. 3: 44. 1911. — Kanehira, Formosan ees 52. 1917. — Melchior in ag Pflanzenfam. ed. 2, 21: 145. 1925. — Sasaki, Cat. phan Herb. eae DIsTRIBUTION: Formosa. ORMOSA: Rengachi, 7. Hayashi 21213 (A), June 1924.— Bankinsing, A. Henry s. n. (A), Aug. 26, 1931. — Vicinity of Kiirun (Kiirung), 7. Tanaka 340 (A, NY, US), Aug. 17, 1929. — Taihoku-shu, Shinten, 7. Tanaka 11178 (C, NY, US), July 24, 1932. — Taihoku, Mt. Sozan, S. Sasaki 21627 (A, NY), Aug. 14, 1927. __'Tameuy, R. Oldham 37 (G), 1864. This variety differs from the species in several characteristics, none of them sufficiently consistent to warrant specific status. The leaves of the variety are usually distinctly serrate along the upper half of the margin. Generally speaking, the leaves of the species proper are quite entire. However, throughout all the material of the species examined one may occasionally find serration along the margin. The specimen of this variety most closely resembling typical A. Millettii from Formosa is Oldham 37. Here the margin of the leaves is more nearly entire with slight evidence of serration. The shape of the leaf is extremely varied. However, this same variation is found in the species. In Sasaki 21627 one may note, on a single branch, leaves which are obtuse, bluntly acuminate, or acuminate at the apex. The pedicels are generally longer, in rare instances attaining 3 cm. (Henry 514) in length. However, 2 cm. is the average length. Both the species and the variety have the same type of thin scattered pubescence on the ovary and fruit. Occasionally there is a slight evidence of pubes- cence at the very base of the style. This variation is not unusual in any species of Adinandra where there is a pubescent ovary and glabrous style. However, seldom is this variation found within the species. Hayata describes the stamens as “0, saepius 15-17” in number. I found the number to be about 25, the same as in the speci Authentic specimens of A. hypochlora vii and A. pedunculata Hayata have not been examined. Kanehira, in his revised edition of Formosan Trees (1936), lists the two as synonyms of A. formosana. The early descriptions, such as they are, seem to bear out this transfer by Kanehira Lib: pecauine Millettii (Hooker & Arnott) Bentham & Hooker f. var. obtusissima, A typo ¢ “dilfert bracteolis persistentibus, oppositis (linearibus, circiter x ; foliis obovatis, basi longo-attenuatis; stylo pubescente. DISTRIBUTION: Formosa. Formosa: South Cape, A. Henry 1985 (NY). This variety differs from the species in (1) the obovate leaves tapering 16 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII from the middle to the base; (2) the linear bracteoles (ca. 2 < 1 mm.), persistent even on the fruiting specimen; (3) the pubescent style. Matsumura and Hayata in 1906 listed an unnamed var. of A. Millettii collected by Henry on South Cape. The more definite locality of ‘“Tamsui” was added in the citation. Probably these specimens are the same. Tamsui is not truly on South Cape but in the southern portion of Formosa, close to Ariko-banti, Ako, and Mt. Daizyurin, Takao, the type-localities of Adinandra obtusissima. As mentioned under A. obtusissima in this paper, I feel that a very close relationship exists between this variety and the species of the same name. The varietal name obtusissima was selected sees with this in mind. . Adinandra Bockiana Pritzel ex Diels in Bot. ey ag 474 (Fl. Centr.-China). 1900. — Melchior in rue ie snes Hines ed. 2, 21: . 1925. eed in Ling- nan Sci. Jour. 1]: 20. 1932.— Chun in sae net ef 189. 19 DistrRIBUTION: China (Kwangsi, Kweichow, Szechuan). Kwancst: San Chiang Hsien, Ling Wang Shan, valley roadside, alt. 2100 m., A. Steward & H. C. Cheo 964 (A, NY), Sept. 17, 1933 (shrub 2.5 m. high with back fruit). — N. Luchen, Chu Feng Shan, 30 li southwest of Shan Fang, in open woods, alt. 1000 m., R. C. Ching 5838 (NY), June 8, 1928. Kweicnow: Kiensi, Lung-kai-ching, open hillside, ¥. Tstang 8778 (A, NY), Aug. 29, 1930 (small tree 3 m. high; leaves green above, light green below). — Yao Shan, in mixed woods, C, Wang 40107 (A), Oct. 14, 1936 (tree 20 m. high; fruit black when ripe). — Tsunyi Hsien, vee: fey Yah, shaded forest slope, A. N. Steward, C. Y. Chiao & H.C. Cheo 149 (A, C, NY, US), Aug. 31, 1931 (shrub 2 m. high with leaves dark glossy green a. floes Nan ch'uen, Tu ma ton, C. Bock & A. von Rosthorn 317 (vype, fragment and photo, A). Shrub or small tree; branchlets terete, brown, glabrous or seinen the very young branchlets densely reddish tawny-sericeous, Leaves ceous , oblong- ovate, 10-14 cm. long, 3—4.5 cm. wide, acuminate at ira apex, ‘the acumen 1-2 cm. long, acute at the base, the terminal buds and very young leaves densely reddish tawny-sericeous, the mature leaves deep dark glossy green, glabrous above, paler beneath covered with a soft tawny pubescence of varying density concentrated on midrib and the margin, the margin entire, lightly revolute, the veins ca. 12 pairs, obvious but not conspicuous, the petiole 5-7 mm. long, densely sericeous. Flowers axillary, solitary; pedicel slender, arching, 1—2 cm. long, densely spreading- pilose; bracteoles 2, opposite, immediately below the calyx-lobes, quickly . lo is densely pilose; calyx-lobes 5, im ricate, thin with little differentiation sei the middle portion and the margin, subequal, ca. 6—6.5 mm. long, 3.5-4 mm. wide, the outer lobes broadly ovate, occasionally apiculate at the apex, the inner lobes more rounded at the apex, the margin glandular- denticulate and ciliate; mature corolla and stamens not seen; ovary globose, sparsely sericeous-pubescent, 3-celled, multi-ovulate, the style entire, glabrous, ca. 9 mm. long (post anthesis). Fruit globose, glabrescent, black, ca. 1 cm, in diameter; seeds many, reddish brown, shinin In all the specimens examined, none had fully developed iawaes for dissection. However, Ching 5838, in bud, upon dissection showed the corolla to be pubescent on the exterior surface. Otherwise, it appeared very similar to that of A. Millettii. The stamens, although minute, were 25 in number, uniseriate, lightly adnate to the base of the corolla; the 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 17 filaments were glabrous, small and free from each other; the anthers broad and densely pilose. In fact, they appeared like those of A. Millettii in miniature. Adinandra Bockiana differs from A. Millettii in: (1) the pubescent exterior surface of the corolla; (2) the calyx-lobes of A. Bockiana are more rotund at the apex with the maximum width considerably above the base; and (3) the dense, tawny pubescence of the branchlets, leaves and floral parts Further collections may prove this species to be worthy of varietal status only. It is confined to the provinces of western China, only one collection of A. Millettii having been made so far from Kwangsi. 2a. Adinandra Bockiana Pritzel ex Diels var. haps (Hand.-Mazz.), comb. Adinandra a hi Handel-Mazzetti in Anz. Akad. Wiss. Wien i. 105 (Pl. Nov. Sin. Forts. 16, 5). 1922. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. — Metcalf in ee Sci. Jour. 11: 20. 1932.— Chun in Seams 1: 297. 1934, ce pe ae (Fukien, Hunan, Kweichow, Kwangsi). Buong Kang, in aoe forest, alt. 1000 m., H. H. Chung 3649 (A, a y, July 2, 1925 (tree 10 m.). Hunan: Near city of Wikany. on Yiin-shan Mt. in deep forest, alt. 1170 m., H. Handel- cand 12350 (1SO-SYNTYPE, A), July 29, 1918 (shrub 2 m. with yellow flowers). — Sinning Hsien, Ma-Ling-Tung, under shade on slope, alt. 640 m., . Fan & Y. Y. Li 516 (A), Sept. 21, 1935 (tree 10 m. high with dark purple fruit). Kweie cHow: In dumetis ad pagum Badschai, alt. 950 m., H. Handel-Mazzetti 10759 (1so-SYNTYPE, A), July 14, 1917 (frutex elatus, fl. albis). — Tan ling, Tuhshan, in light shade, Y. Tsiang 6893 (A, NY), Sept. 5, 1930 (tree 4 m. high; leaves deep lustrous green above, lighter below).— Kweiting, along river, in open, alt. 400-550 m., VY. Tsiang 5464 (A, NY), June 30, 1930 (shrub; leaves deep green above, lighter below). — Yao-ren- shan, San hoa, in light ete alt. 400-700 m., VY. Tsiang 6394 (A, NY), Aug. 10, 1930 (common tree 3-10 leaves deep apa green above, lighter below). Kwancst: Shang-sze Distr., Shap M an Cha Muk, near Hoh Lung village, southeast of Shang-sze (Kwangtung border), Wong Pan Muk, in dry, clay thickets, W. T. Tsang 22486 (A), July 17, 1933 (tree 3 m. tall with edible black fruit). — Shang-sze Distr. (Kwangtung border), southeast of Shang- sze, Shap Man Taai Shan, Tang Lung village, Wong Pan Cha Shue, in swamp thickets, W. T. Tsang 24013 (A, M, NY), Aug. 11, 1934 (common, 2 m. high; fruit black, edible). — Shang-sze Distr., southeast of Shang-sze (Kwangtung border), near Iu Shan village, Wong Pan Cha Shue, in dry silt, sandy soil, W. 7. Tsang 22167 (A), April 29-30, 1933 (scattered shrubs 2 m. tall; flowers yellow). — Kwei-lin Distr., Hsi-chang village and vicinity, Ch’i-fen- ae n, Wong Pan Cha Shue, in thickets on steep slope, W. T. Tsang 28485 (A, US), Oct. 1-11, 1937 (fairly common, 2 m. high; fruit edible, black). ApbDED DESCRIPTION: AES axillary, solitary or rarely in twos; pedicel usually recurved, ca. 1 cm. long, lightly strigillose or glabrescent; ee 2, quickly caducous, he present linear-oblong, ca. 2 mm. long an . wide, slightly pubescent; calyx-lobes 5, imbricate, broadly ovate, ae l, 4-5 mm. long, 3-3.5 mm. wide, glabrescent or ‘glabrous, the outer lobes somewhat acumin nate, the inner r lobes more rounded at the apex, the margin glandular and ciliate- fimbriate; corolla-lobes 5, connate at the base, ca. 7 mm. long, 4.5—5 mm. wide, rounded at the apex, pubescent on the median portion of the exterior surface: stamens ca. 25, uniseriate, ca. 5 mm. long, the filaments lightly adnate to the base of the corolla, free from each other, glabrous, ca. 2 mm. long, somewhat dilated in the middle, 18 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII the anthers ca. 2 mm. long, quite wide at the base, somewhat sagittate, densely hirsute, the apicule ca. 1 mm. long; ovary subglobose, white- sericeous, 3-celled, multi-ovulate, the style ca. 8 mm. long, glabrous, oe Fruit globose, dark purple when ripe, soft, thin-walled, ovoid, ca. 1 c more in diameter, many-seeded; the seeds dark, shining. This variety agrees with the species in most characters. The chief feature of separation is the glabrescent or glabrous character of most of its parts. 2b. Adinandra Bockiana Pritzel ex Diels var. tonkinensis, var. nov. Adinandra Millettii Merrill in Jour. Arnold Arb. 19:55. — Gagnepain in FI. Gén. Indo-Chine, Suppl. 1: 284. 1943. Non Hooker & rer tt) Bentham Hooker f. A typo differt gemmis, juventutibus ramulis, pedicellis, bracteolis, calyci- bus pubescentibus adpressis. DistRiBUuTION: Indo-China (Tonkin). ONKIN: Chapa, on banks of brooks and ravines, alt. 1500 m., A. Pételot 3751 (type, NY, US), July 1930 (tree 10 m. high with very pale salmon-colored corolla). — Station du Tam Dao, alt. 900 m., A. Pételot 3976 (NY, US), May 1931.— Chapa alt. 1500 m., A. Pételot 4308 (NY, US), July 1931 oth Merrill and Gagnepain associated the specimens cited above with A, Millettii. However, the acuminate leaves, the broadly ovate calyx-lobes and the pubescent corolla-lobes show it to be a variation of A. Bockiana and oe! allied to A. Bockiana var. acutifolia. The real difference lies ubescence. Whereas, in A. Bockiana a ockiana var. nai the pubescence is spreading, in this variety the pubescence is short and closely appressed. ' aera me paaeee Hu in Bull. Fan. Mem. Inst. Biol. 6: 172. 1935; in Icon. Pl. Sin. 5: t. Pe senty Peet Cea in Not. Syst. Mus. Hist. Nat. Paris 10: 115. 1942; in Fl. Gén. Indo-Chine Suppl. 1: 289. Adinandra se ae Li in Jour. Arnold Arb. 26: 66. 1945. DistriputTion: China (Yunnan, Kwangsi); Indo-China (Tonk'n). Cinna: YUNNAN: Ping-pien Hsien, in ravine, alt. 1200 m., H. T. Tsai 60288 (1soTyPE, A) «tee 19, 1934 (small tree 7 m. high). Kwancst: Chen Pien Distr. on sa in iirest, S. P. Ko 55926 (A), Oct. 17, 1935 (tree 8 m. high; fruit ae 24 Ling Wan Distr. in light forest, S. K. Lau 28705 (ryvpx of A, serrulata, A), July 1937. Inpo-CHina: ToNKIN: near Chapa, alt. 1800 m., A. Pételot 3788 (asoryre of A. Petelotii, A), Aug. 1930 Small tree (6 m.); branchlets terete, ferrugineous-pilose; terminal buds elongate-conical up to 2 cm. long, densely appressed ferrugineous-sericeous. Leaves coriaceous, oblong-lanceolate, 16-24 cm. long, 4-7 cm. wide, acuminate at the apex, rounded to broadly cuneate at the base, the very young unfolding leaves densely ferrugineous- pubescent, the mature leaves glabrous on the upper surface except within the channelled midrib, scattered-appressed- pilose beneath in varying degrees, occasionally quite glabrescent, the margin serrulate with occasional apiculate glands varying according to the age of the leaf, the. veins ca. 20-24 pairs, slightly curved. arching, more prominent beneath, the reticulations obvious, the petiole 1947 ] KOBUSKI, STUDIES IN THE THEACEAE, XV 19 stout, ca. 12-15 cm. long, appressed-ferrugineous-pilose above and beneath. Flowers axillary, solitary; pedicel up to 3 cm. long, thickened at the apex, recurved, appressed- pilose; bracteoles 2, quickly caducous, alternate, when present oblong, acute at the apex, 5-6 mm. long, ca. 3 mm. wide, pilose; calyx-lobes 5, ee thick, unequal, broadly ovate, outer lobes 11-13 cm. long, 10-11 cm. wide, sericeous on the median portion of the dorsal surface, the inner lobes pubescent, smaller, less thick, the margins more scarious; ae lobes (fide Hu) “5, unequal, broadly ovate-oblong, to m mm. broad, slightly. grey- ferrugineous-sericeous outside, glabrous inside serrulate along the margins;” stamens 40-45, seemingly uniseriate, rather uniform in size, the filaments short, thickened, ca. 2 mm long, glabrous, the anthers ca. 4 mm. long, densely sericeous on the dorsal surface, apiculate; ovary conical, densely sericeous, 5-celled, multi-ovulate, tapering at the apex into the short blunt style which is appressed-pubescent, the pubescence thinning out towards the glabrous apex. Mature fruit not seen, maturing fruit very thick-walled. There are several very obvious characteristics which aid in easily dis- tinguishing this species. First, the large, oblong-lanceolate leaves 20-2 cm. long and definitely serrate; second, the large almost velvety calyx-lobes, similar in size and appearance to those of A. glischroloma var. macrosepala (Metcalf) Kob.; as third, the pubescence along the midrib of the upper surface of the Gagnepain pare the same species under the name A. Petelotit. There is no need for discussion because the two are obviously the same. Li’s species, A. serrulata, was described from specimens with leaves some- what smaller (16-20 * 4-5 cm.), but otherwise matching those of A. megaphylla perfectly. u’s original description is excellent and very detailed. He states, however, that the stamens are glabrous. Although only buds were available for dissection in this study, it is clear from these dissections that the stamens are very hirsute along the dorsal surface of the anther. Also Hu gives the impression from his description and later drawings in Icon. PI. Sin. that the bracteoles are persistent, whereas in the two specimens examined they are obviously quickly caducous. He fails to mention that the ovary and style are pubescent and that the ovary is five-celled. His later illustration, however, depicts pubescence on these parts. 4. seri Sapa Handel-Mazzetti in Anz. Akad. Wiss. Wien 60:96 (PI. ov. Sin. Forts. 16, p. 5). 1923. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 145. ee a in “Lin an Sci. Jour. 7: 315. [1929] 1931. — Metcalf in Lingnan Sci. Jour. Adinandra iba al ex Metcalf, loc. cit., nomen. DistrRiBUTION: China (Kwangtung, Kwangsi). KWANCTUNG: oe Distr., Sei Hsien, C. L. Tso 20328 Ae NY), May 9, 1929, — Canton, near town of Lien pines a 500-900 m., R. Mell 679 (1so-syNTYPE, A), Aug. 11, 1920 (very ees g T’au Mt., wear a of Iu, in ravine on side of mountain, K. P. To & U.K. Ts'o ang 12041 (M, US), May 23, 1924 (tree 2 m. high; flowers red and white). — Sing-fung Distr., Hau T’ong Shan, Fuk Lung Mon- astery, in thickets, Y. W. Taam 758 (A), oe -June 1938 (scattered shrubs with black edible fruit). — Mei Distr. (Kaying), Yam Na Shan (Yit Nga Shan), in dry silt, W. T. Tsang 21387 (A, NY), Aug. 1932 ae shrubs ca. 3 m. tall; fruit black). 20 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII Kwancst: Yuan Tang Shan, 15 li west of Shan Fang, N. Luchen, in forest, alt. 600 m., R. C. Ching 5719 (A, NY), June 6, se ae 2 m. high; leaves dark green pie paler below). — Yao Shan, in mixed w , C. Wang 39479 (A), 40155 (A), 1 (A), June-Oct. 1936 (tree 10-15 m. high eae black). Shrub or small tree; branchlets thick, terete, villous-pubescent, the oung branchlets and terminal buds covered with a brown, more dense villous pubescence. Leaves coriaceous, oblong-elliptic, 8-14 cm. long, 3-5 cm. wide, acuminate at the apex, tapering at the base, usually cuneate, occasionally somewhat rounded, glabrous above, tawny-hirsute beneath, more densely concentrated along the margin and the midrib, the margin revolute, usually entire with occasional apiculate glands, the con- centrated pubescence (extending from beneath) appearing as a halo along the edge of the upper surface, the veins 10-12 pairs, conspicuous on both surfaces, the petiole 8-10 mm. long, densely hirsute. Flowers axillary, in twos or threes, rarely solitary; pedicel 6-15 mm. long, erect or recurved, tawny-brown- hirtellous; bracteoles 2, caducous, opposite or alternate; calyx-lobes 5, imbricate, broadly ovate, 6—7 mm. long, hirtellous; coll and stamens not seen; immature capsule 5-celled, hirtellous, the style ca. 1 cm. long, the lower two-thirds hirtellous, the upper third glabrous. The tawny-hirtellous pubescence of the young branchlets, the under surface of the leaves and the flowering parts make this species and its varieties easily distinguishable. Concentrated along the margin, even when the leaf has become glabrescent, this pubescence projects beyond the edge of the lower surface in such quantity and to such an extent that when viewed from the upper surface it presents the appearance of a tawny halo. The species is unique in this character. Another unusual feature, for Chinese species, is the occurrence of the axillary flowers in twos and threes; only occasionally does one find solitary flowers. All material cited above is either in bud or in the post anthesis stage. Dissections of very young buds do show, even so, that the corolla-lobes are pubescent on the median portion of the external surface and are somewhat ovate in shape. The stamens, although they appear to be in series and over 30 in number, are quite uniform in size. The filaments are glabrous and the anthers hirsute. In Tsang 21387, the only specimen with developing fruit, the calyx-lobes measure close to 10 mm, in length and 5—6 mm. in width. Cited here as a synonym is the herbarium name A. chinensis Merrill. Metcalf (Lingnan Sci. Jour. 11: 20. 1932) mentioned this unpublished species and listed it as a synonym of his own A. macrosepala. Examina- tion of Merrill’s specimens shows that Merrill had designated as the type of his unpublished species Wulsin’s (Canton Christian College 12041) specimen and not Chung 3384, which belongs to the variety macrosepala. 4a. aan glischroloma Handel-Mazzetti var. pani aor cas comb. nov. nandra macrosepala Metcalf in Lingnan Sci. Jour. 11: DISTRIBUTION: China (Fukien, Chekiang). Fukien: Yenping, Buong Kang, in thickets, alt. 800 m. H. H. Chung 3384 (syNTYPE of A. macrosepala, A, Iso-SyNTyPE M), June 14, 1925 (shrub 3 m.).— Chung an Hsien, open hillside, alt. 1100 m., H. H. Hu 1322 (syNtype of A. macrose- pala, A) July 23, 1921 (shrub ca. 1 m.).—No data, H. H. Chung 7369 (A). 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 21 CHEKIANG: 80 li northeast of Tai Suan, in open pata: alt. 800 m., R. C. Ching 2183 (A, NY, US), July 20, 1924 (shrub 5 m. high h brownish Bark. — Taishun Hsien, on mountain slope, Y. L. Keng 312 (A), hee e 1926 (shrub 1-2 m.). The chief difference of this variety from the species is found in the size of the calyx-lobes. Whereas, in the species, in bud, the lobes measure 6-7 mm. in length and in a ee early fruiting specimen the calyx-lobes measure 9-10 mm. X 5-6 mm., in the variety, the calyx-lobes in bud measure 11-14 mm. in length. Their full development is not known. Dissections of flowers in bud, more nearly mature than those in the species, show that, like the species, their corolla-lobes are ovate, acute at the apex, and pubescent on the dorsal surface. The stamens, many in number, appear to be seriate. The anthers are quite uniform and measure ca. 3 mm. in length. The filaments are very short and glabrous, while the apicule varies considerably in length, from 1—2—3 mm ne may assume that, when mature flowers of the species and variety are collected, all the floral parts of the variety will be uniformly larger than those of the species. 4b. Adinandra glischroloma Handel-Mazzetti var. Hist (Li), comb. nov. Adinandra jubata Li in Jour. Arnold Arb. 25: 422. 1944. DistrRIBuTION: China (Kwangtung). Kwanctunc: Hwei-yang Distr., Shek Shing village, Lin Fa Shan, W. T. ng 5001 (TYPE of A. jubata, A), Aug. 25-26, 1935 (scattered shrubs 2 m. high; ae edible, black). This variety is characterized by a beautiful, dense, ferrugineous, iridescent tomentum (sometimes as much as 5 mm. long) on the terminal buds, young branchlets, under surface of the leaves, and flowering parts. Li separated it from A. glischroloma Handel-Mazzetti but it appears to be nearer the variety var. macrosepala and except for degree of pubescence is a perfect match for Chung 7369 collected in Fukien. In reality, except for pubescence, this variety seems to be midway between the species and the variety macrosepala. In the two latter entities the pubescence is very similar. However, they both agree with var. jubata in type of pubescence differing only in densit ity. 4c. Adinandra glischroloma Handel-Mazzetti var. ee (Gagnepain), comb. nov. Adinandra hirta Gagnepain in Not. Syst. Mus. . Nat. Paris 10: 113. 1942; in Fl. Gén. Indo-Chine, Suppl. 1: 286, fig. 26. ae DistTRIBUTION: Indo-China (Tonkin). TonkKIN: Chapa, alt. 1500 m., A. Pételot 3818 (1soTYPE of A. hirta, A, US), Aug. 930 This variety agrees with A. glischroloma Handel-Mazzetti in most characteristics. The flowers are axillary and like the species are in pairs, or in threes, rather than solitary. The buds are small and very hirsute. The mubescetce of the leaves, although of the same type, is less dense. However, in the ee young leaves, the pubescence appears more dense and typical of the s The eee which separate var. hirta from the species, although 22 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII of dubious importance, are: (1) the ovate-acute or deltoid, persistent bracteoles; (2) the style, pubescent its entire length; and (3) the pubescent filaments. 5. Adinandra filipes Merrill in herb., sp. nov. Frutex 4 m. altus, subglaber (orbs ignotis), ramis ramulisque eriaed bus, ramulis ultimis 2 mm. diame , adpresso- -pubescentibus; gemm leviter adpresso-pubescentibus; foliis oabee oblongo-ellipticis, 7-10 cm. longis, 2—2.5 cm. latis, apice graciliter acuminatis, basi cuneatis, glabris, haud punctatis, margine minute apiculato-serrulatis, dentibus inter se 2—3 mm. distantibus, deorsum remotioribus, margine in partibus inferioribus integris vel subintegris, nervis primariis utrinque 12-15, cirgeidae haud perspicuis, arcuato-anastomosantibus; petiolo crasso, circiter 2 mm. longo; fructibus solitariis, axillaribus, longe (4 cm.) Se vedicellatis, ovoideis, circiter 1 cm. longis (immaturis) consperse pubescentibus, stylis gracilibus, circiter 1 cm. longis, breviter ciliato-pubescentibus, apice breve tripartitis; bracteolis 2, caducis; sepalis 5, imbricatis, persistentibus, re- flexis, oblongo-ovatis, circiter 8 mm. longis, acutis, obscurissime adpresso- pubescentibus. DistrRiBUuTION: China (Kwangsi). KWANGSI: o Mor Shan, N. Lin Yen, rare in forest, alt. 1400 m., R. C. Ching 7138 (TYPE, ee een, A), Aug. 25, 1928 (shrub 3 m. high a eae bark; leaves deep shining green above, pale beneath. This species, considered rare by the collector, is characterized by the slender, elongated pedicels which in fruit are 4 cm. in length, by the reflexed sepals, and the three-parted, pubescent style. diss panel a. Havata, Icon. Pl. Formos. 3: 43. 1913.— Melchior in Nat. eee ed. 2, 21: 144. 1925. — Merrill in Lingnan Sci. Jour. 6: 283. 1930. — fiat & Odashima in Elks Soc. Trop. Agric. 10: 375. 1931.— Metcalf in 1: 20 ead Sci. Jour. . Adinandra hainanensis “sat in - Piip: Jour. ioe 23: 259. 1923. Adinandra Macluret Merrill in Lingnan Sci : 129. 1927. Adinandra rubropunctata Merrill & Chun in enema 1: 70. 1930. Poe in Lingnan Sci. Jour. 11: 20. 1932. — Merrill & Chun in Sunyatsenia 2: 41. 193 Chun, Icon. Pl. Sin. 5:3, t. 203. 1937; in Sunyatsenia 4: 188. 1940 DistRIBUTION: China (Hainan, Kwangtung). NAN: Ch’and-kiang Distr., Ka Chik Shan and vicinity, in thickets on dry cliffs, S. K. Lau 1486 (A, NY), 1533 (A, i ), April 8-13, (tree 3-4 m. high, 9 m. iam.).— Lam Ko Distr., Fu Hoi and vicinity, W. Tsang 201 (A, M, NY, US), July 29, 1927 (4 m. high). — Taam Chau Distr., Sha a ay in forest, W. T. Tsang 543 (A, NY, US), May 27, 1928 (flowers white). —Lam K oo Mai _ Districts, Taai Wong Shan, summit of mountain, W. 7. Tsang 711 (A , June 15, 1928 (2 m. high; flowers white). — Dung Ka to Wen Fa Shi, in nen ait. 550 m., NV. K. Chun & C. L. Tso 43778 (A, NY), Sept. 1, 1932 (tree 6 m. high, cm. diam. with y branches and brownish gray branchlets ; leaves . Ages nent above, . Chun & paler Sibeouit beneath). — Dung Ka, in forested ravine, alt. m., NV. Cc Tso 43936 (A, NY), Sept. 25, 1932 (tree 8 m. high one gray bark and branch- lets; leaves eh dark = Seg eae nape ). — Ng Ch i Ling, Fan Yah, in thin forest, alt. 300 m., NV. K. 2 & C, o 44012 (A, C, NY), Oct. 5, 1932 fiteen 8 m. high, 20 cm. diam. ane gray ba rs tola deep green above). — Po-ting, in forest ravine, alt, 300 m., F. C. How 72081 (A), April 23, 1935 (tree 7 m, high with gray bark; leaves {ustebus green above, paler beneath). — Tai Pin, J. L. Gressitt 1076 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 23 (A), June 21, 1935 (tree 10 m. high, 15 cm. diam.).— Five Finger Mt., in wooded ravine, F. A. McClure 9571 (A, M), May 13, 1922 (shrub 3 m. high, 8 cm. diam.). — Ching Mai Distr., Tai Wong Ling and vicinity, Tung Pin Tin village, dry sandy slope in thicket, C. J. Lez 816 (A, NY, US), July 5, 1933.— Hung Mo Shan and vicinity, by stream, Tsang & Fung 292 (A, NY, US), June 13, 1929 (fruit black).— Seven Finger Mt., in thicket on slope, H. ¥. Liang 61760 (A, NY, US), May 6, 1930 (tree 7 m. high with gray bark and branches). — Ngai Distr., Yeung Ling Shan, in thickets on moist clay slope, S. K. Lau 196 (A, M, NY, US), July 2, 1932.— Bak Sa, in thickets, S. K. Lau 26138 (A), April 8, 1936.— Loktung, in dense woods, S. K. Lau 27209 (A), June 20, 1936. — Po-ting, in forested ravine, alt. 400 m., F. C. How 72471, 73514 (A), May re 1935 (tree 9-25 m. high; flowers white). a heat locality lack ing, in mixed forest, C. Wang 33649 (A, NY), Aug. 12, 1933 (tree 10 m. high) Kwanctunc: Ta Mien Shan, Sup Man Ta Shan, in shaded woods, H. Y. Liang 69638 (A), 69640 (A), July 14, 1937 (tree 6 m. high; fruit villous). — Weishang, Sunyi, in light shade, alt. 930 m., Y. Tsiang 2713 (TyPE of A. rubropunctata, NY), June 15, 1929 (tree 7 m. high; leaves lustrous green above, light green beneath). Shrub or small tree up to 10 m., rarely up to 25 m.; branches brown or grayish ‘brown, terete, the very young branchlets and terminal buds appressed- Basing Leaves coriaceous, oblong-elliptic to oblong-obovate, 6-8 X cm. and 10-13 x 5-6 cm., shortly acuminate at the apex, acute at ihe base, shining green above, paler beneath, glabrous on both surfaces, conspicuously dark-punctate beneath, the margin distinctly glandular- denticulate, the veins ca. 12 pairs, distinct on both surfaces, reticulate, the petiole ca. 5 mm. long, on larger leaves up to 10 mm. Flowers axillary, solitary; pedicel ca. 7 mm. long, strongly recurved, generally glabrous or glabrescent, those of the very young flowers appressed-pubescent especially at the very tip of the branchlets; Sin 2, opposite, quickly caducous, when present pubescent, unequal, c oo be m.; calyx-lobes 5, imbricate, densely eres ceous-pubescent, subequal, 6-8 mm. long, ca. 6 mm. wide, ovate, sometimes aetna at the apex, outer lobes thicker, glandular-denticulate along the scarious margin, the inner lobes thinner, the margin usually entire; corolla- lobes 5, connate at the base, white, pubescent on ae median portion of the exterior aaa oblong-subelliptic, 7-8 mm. long, ca. 4 mm. wide, somewhat obtus at t e apex; stamens 30-35, eat uniseriate, rather uniform in size, “S 5-7 mm, long, the filaments 2-3 mm. long, glabrous, quite free from ‘each other, adnate to the base of the corolla, projected at cence, 5-celled, multi-ovulate, tapering into a densely pubescent style 6-7 mm. long. Fruit globose, black, pubescent, 5-celled, many-seeded, Ler; The outstanding characteristics of this species are: (1) the 5-celled ovary; (2) the dense tawny pubescence on the floral parts; (3) the distinct reticulate veining obvious on both surfaces and the dark-punctate dots on the lower surface of the leaves; (4) the short, recurved pedicel; and (5) the glandular-denticulate margin of the leaves. Its closest relative is A. lasiostyla Hayata, which can be separated by the 3-celled ovary, the obscure veining, the usually entire margin, and the scattered pubescence on the lower half of the style Several synonyms are cited above, only one of which, A. rubropunctata 24 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Merrill & Chun, needs comment. In 1940, Chun (Sunyatsenia 4: 188) listed seven specimens from Hainan as belonging to A. rubropunctata. Of these, I have examined S. K. Lau 196, H. Y. Liang 61760, F. C. How 2471 and 73514. Ican find no characters in these specimens that would warrant separation from the above species. Metcalf (1932) separates A, rubropunctata from A, hainanensis on the presence of red-punctate dots on the lower surface of the leaf, and the apiculate denticulations along the margin. The original description stressed these two characters. Examin- ing considerable material, one finds great variation in the color of the punctate dots which range from black through brown to reddish brown, even to a lighter yellowish shade. This color variation is undoubtedly due to the age of the leaf. The apiculate denticulations may be found on nearly every leaf, varying in size and number. Age probably governs this characteristic also, as these apiculations are very fragile and are easily broken off. The type specimen of 4. rubropunctata, Tsiang 2713, offers the only variation worthy of consideration. Whereas in most of the specimens cited above the pubescence is tawny or straw-colored, in the above-men- tioned type the pubescence is more dense and ferrugineous. The pedicels of Tsiang 2713 are densely pubescent. As it happens, all the flowers of Tsiang 2713 are concentrated at or near the apex of the young branchlets, so near as still to possess the pubescence found on the early growing parts. In the other specimens examined the flowers are scattered along the stems, and the pedicels are usually glabrous or only slightly pubescent. In Tstang 2713, the bracteoles are consistently persistent. In the other specimens of A. hainanensis, the bracteoles are quickly caducous, but when present, are of the same size and shape as those found on the above-mentioned Tsiang specimen. 7. Adinandra lasiostyla Hayata in Jour. Coll. Sci. Tokyo 30 (Art. 1): 42, 1911. — Kanehira, Formosan Trees 53, fig. ae revised ed. 453, fig. 410. 1936. — Mel- chior ‘in Nat. Pflanzenfam. ed. 2, 21: 145. 1925. — Sasaki, Cat. Governm. Herb. 347. 1930 DIsTRIBUTION: Formosa. Formosa: Arisan, R. Kanehira ahs (NY, US), Dec. 1, 1933.— Arisan Prov. Kagi, in forest, alt. 1600-2800 m., E. Wilson 9667, 9770, 10819 (A, US), Jan. 29, Feb. 4, Oct. 18, 1918 (tree 10-15 m. na Trees 8-15 m. high, 1-1.5 m. circumference; branches terete, brown to grayish brown, the branchlets and terminal buds tawny-pubescent. Leaves coriaceous, oblong-elliptic, 8-13 cm. long 2—4 cm. wide, acuminate at the apex, acute at the base, shining glabrous above, pubescent beneath, the margin entire or obscurely crenulate, the veins ca. 12 pairs, obscure above, somewhat conspicuous beneath, the petiole short, 3-5 mm. long, pubescent beneath. Flowers axillary, solitary: pedicels ea ‘short, recurved, ca. 5 mm. long; bracteoles 2, quickly caducous; calyx-lobes 5, imbricate, subequal, rounded, ca. 6 mm. long, 5 mm. wide, pubescent, oar denticulate along the margin; corolla-lobes 5, lightly connate at the base, ca. 9 mm. long, 5 mm. wide, broadly acute or obtuse at the apex, ea aah along the median portion of the exterior surface; stamens 20-25, 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 25 seemingly uniseriate, rather uniform in size, 5—5.5 mm. long, the filaments glabrous, 1.5-2 mm. long, quite free from each other but adnate to the base of the corolla, the anthers ca. 2.5 mm. long with dense tawny pubes- ence, the apicule distinct 1-1.5 mm. long; ovary subglobose, tapering slightly at the apex, scattered pubescent, 3-celled multi-ovulate; style entire, up to 10 mm. long, 1-1.5 mm. in diameter near the base, sparingly pubescent along the lower half; stigma obtuse. Fruit immature. This species is closely related to A. Milletti (Hooker & Arnott) Benth. & Hook. f. and perhaps belongs to the same alliance of species. Certain consistently uniform characteristics, however, are present to warrant specific separation. ‘The large pubescent acuminate leaves resemble those of A. Bockiana Pritzel and its variety acutifolia (Hand.-Mazz.) Kob. A dense pubescence covers the young branchlets, lower leaf-surface, pedicels and calyx-lobes. Oddly enough, the style (whence the specific name) is only slightly pubescent with the scattered pubescence extending only along the lower half The pedicel is very brief, not more than 5 mm. long, and always recurved. The calyx-lobes are rounded rather than acute and the exterior surface of the corolla-lobes is pubescent Whereas A. Millettii is considered a shrub and rarely a small tree, this species is definitely a tree of considerable magnitude according to Wilson, measuring ‘‘30—50 ft. x 2-5 ft.” I presume the latter measure is cir- cumference. Wilson states that the species is very common at Arisan. All specimens examined by the author were collected from this locality or nearby Mt. Morrison. All other specimens cited in literature have been collected in the same general limited locality. 8. Adinandra epunctata Merrill & Chun ex Tanaka & Odashima in Jour. Soc. Trop. 931, nomen; Merrill & Chun in Sunyatsenia 5: 132, fig. 14. 1940. DISTRIBUTION: China (Hainan). Harnan: Mo San Leng, in forest, alt. 1300 m., N. K. Chun & C. L. Tso 44289 (A, C, NY, US), Nov. 20, 1932 (tree 12 m. high with sie eae ers and gray branchlets; leaves deep lustrous green; flowers white). — mixed forest, C. Wang In mixed forest, C. Wang 36007 (A, NY, US), Dec. 30, 1933 (tree 12 m. high with white flowers). — In shaded forest on mountain side, H. Y. Liang 64384 (NY), Jan. 10, 1934 (tree 12 m. high, 1 m. diam.; leaves thick, green above, paler beneath). Small tree 12-18 m. high; branches terete, glabrous, gray or brown-gray, the branchlets terete, gray or gray-brown, finely appressed-pale- to ferru- gineous-pubescent, the terminal buds conical, covered with a very dense tawny to ferrugineous velvety pilose pubescence. Leaves coriaceous, elliptic-oblong, 5-11 cm. long, 2-4 cm. wide, acute at both the apex and the base, lustrous, deep green, glabrous above, paler, at first densely appressed- pubescent beneath, later glabrescent or nearly so, entire along prominent above, less so beneath, the petiole 8-15 mm. long, appressed- pubescent. Flowers axillary, solitary, occasionally in twos: pedicels up to 7 mm. long, densely pubescent, usually somewhat recurved; bracteoles a opposite, persistent, sepaloid, long-triangular, ca. 3 mm. long, 1.75- 26 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII mm, wide, acute at the apex, densely pubescent; calyx-lobes 5, imbricate, ovate, acute at the apex, ca. 5 mm. long, 2.5-3 mm. wide, entire, densely appressed- pale- -pubescent; corolla-lobes 5, white, lightly connate at the base, oblong, obtuse at the apex, 6—7 mm. long, 3-4 mm. wide, pubescent on the external surface, glabrous ae stamens 25, unequal, 4—6 mm. long (in the same flower), the filaments 2—4 mm. long, pubescent on the same distance joined to each other, the anthers oe nearly equal, 1.5—2.0 mm. long, pubescent; the apicule short, 0.5 mm. or less; ovary somewhat conic-ovoid, densely gray-pubescent, 3- ind. tapering at the apex into an entire, pubescent style ca. 3-4 mm. long. Fruit small, ovoid, ca. 5 mm. long, 4 mm. diam., rales eda 3-celled, few-seeded, seeds 9-10, brown, scrobicular, ca. 2 mm. long. A few of the outstanding pana tare of this species are the small flowers and fruit, the persistent and comparatively large bracteoles, the very short pedicels, the pubescent filaments of the anthers, the entire leaves, and the small number of seeds developing in the fruit. Although not described until 1940, this species was mentioned in a list of the Hainan plants by Tanaka and Odashima in 1931, the name probably recorded from an herbarium sheet. Merrill and Chun, in their description, do not mention the number of cells in the ovary and fruit but in the accompanying illustration the ovary is pictured as six-celled. Because of the minute size and dense pubescence, dissections of the ovary are very difficult and quite unsatisfactory. On the other hand, the fruit, when mature, is so crowded with the few seeds which are normal in size, even though the fruit is small, that dissections again are rather unsatisfactory. After several dissections, none of which showed clear-cut cell-structure, my conclusions are that the fruit is three- celled rather than six-celled. It may be that in the original drawing the placentae, which often extend nearly across the cell, were mistaken for cell-walls . Adinandra laotica Gagnepain in Not. Syst. Mus. Hist. Nat. Paris 10: 114. 1942; in Fl. Gén. Indo-Chine, Suppl. 1: 283. 1943. DistRIBUTION: Indo-China (Laos). Laos: ea pie Ao, Xieng-khuang, by stream among evergreens, alt. up to 1100 m., A. . Kerr 21172 (1SOTYPE, %). ieee 21, 1932 (tree 15 m. high). 15 m. high; branchlets tawny brown, terete, glabrous; terminal buds large, glabrous. Leaves thick-coriaceous, glabrous, oblong-lanceolate, 15-20 cm. long, 5.5—6.5 cm. wide, shortly acuminate at the apex, tapering at the base, the margin entire, lightly revolute, the primary veins 12-15 pairs, anastomosing and arching upward near the margin, somewhat pro- nounced above, obscure beneath, frequent secondary veins interspersed between the primary veins, the petiole 15—20 mm. long, glabrous. Flowers axillary, in pairs; pedicel 2—2.5 cm. long, glabrous, thickening towards the apex; bracteoles 2, opposite or nearly so, quickly caducous; calyx-lobes 5, imbricate, broadly o vate, += 8 mm. long, 6-7 mm. wide, glabrous, the margin ciliolate; corolla- lobes (in bud) 5, connate at the base, ca. 6 mn long, glabrous; stamens (bud) ca. 25,3? -seriate, 2-6 mm. long, the filaments 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV ra glabrous, the anthers pilose; ovary glabrous, quite flat in bud, 5-celled, multi-ovulate, the style entire, glabrous Adinandra laotica is in many respects a very close relative of A. angulata idley. Both species have thick branches and branchlets, terete in the former and decidedly angled in the latter. The leaves of both are stiff and heavy-coriaceous. The flowers are axillary and in pairs. In 4. angulata the two pedicels remain attached at the base when removed from the stem and it appears that such may be the case in A. laotica. Both species are very glabrous, even to the terminal bud. In A. angulata, the ovary and fruit are 3-celled with few ovules and later with only a few large seeds developing. Gagnepain describes the ovary of A, laotica as 4-celled. My single dissection showed the ovary to be clearly 5-celled. Only buds were available to both Gagnepain and myself, and in the bud, the ovary is very flat, thin and difficult to section. The bracteoles of A. laotica are caducous while in A. angulata they are persistent. Also, the veins of the leaves of A. angulata are less in number and much more prominent. 10. Adinandra nitida Merrill ex Li in Jour. Arnold Arb. 25: 422. 1944. DistRIBUTION: China (Kwangsi, Kwangtung). Kwancsr: Shang-sze Distr., Shap Man ie Shan, near Iu Shan aes southeast of Shang-sze, near the Kwangtung border, in dry sandy thickets, W. Tsang 22322 (type, A), May 18, 1933 (4 m. high; Aen fragrant, ate Leo nae Distr., Shap Man Taai Shan, near Hoh Lung village, southeast of Shang-sze, near Kwangtung border, in thickets on dry steep rocky slope, W. T. Tsang 22571 (A), June 26, 1933 (fruit black). — Shang-sze Distr., Shap Man Taai Shan, Tang Lung village, southeast of Shang-sze, near Kwangtung border, W. T. Tsang aie aa NY), Oct. 1934 (scattered shrubs 2 m. high; fruit yellow, ae —N. Luchen, Chu Feng Shan, 30 li southwest of Shan Feng, in forest, alt. sks Caching pay (NY), June 8, 1928 san Pe of Nanning, R. C. Chine. 8431 (NY), Nox, 5, 1928 (tree 20 m. thi an 6. ngy (small tree with white flowers). KWANGTUNG: “Sigete ang Distr., Ling Fa Shan, Lin Fung Monastery, W. T. Tsang 25656 (A), Aug. 11-31, 1935 rs 7m. high; fruit black, edible). icon or small tree up to 20 m. high; branches terete, glabrous, gray- brown, the branchlets usually purplish brown, terete, glabrous, the terminal a abscurely appressed- deci near the apex, otherwise glabrous. Leaves coriaceous, ovate-oblong, 7-12 cm. long, 2.5—4 cm. wide, acuminate or obtusely acuminate at the apex, acute at the base, glabrous on both surfaces, dark glossy green above, paler beneath, the margin serrulate, varying in degree, the veins 12-16 pairs, somewhat conspicuous on bot surfaces, the petiole glabrous, 1-1.5 cm. lon Flowers axillary, solitary; pedicel 1-2 cm. long, glabrous; Ae is 7 ke persistent, opposite, unequal, subsepaloid, glabrous, 9-11 mm. long, 4-5 mm. wide, ovate, acute at the 28 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII apex, the smaller bracteole usually keeled along the center of the external surface; calyx-lobes 5, imbricate, glabrous, unequal, ovate, ca. 15 mm. long, 7-9.5 mm. wide, sharply acute at the apex, often subapiculate: corolla- heed 5, connate at the base, ora unequal, 17-19 mm. long . wide, obtuse to rounded at the apex; stamens ca. 25, unequal, 6-11 mm. ne aa the emails eae at the base, adnate to the base of the corolla, glabrous, 2-5 mm. (in the same flower), the anthers linear, 4—6 mm. long, hirsute on ces exterior surface; ovary glabrous, ovoid, ca. 4 mm. diam., 3-celled, multi-ovulate, tapering into the glabrous style 10 mm. long, the style 3- -parted at the apex for as much as 3 mm. Fruit globose, ca. 1.5 cm. in diameter Adinandra nitida, as the name signifies, is characterized by dark green leaves, extremely lustrous on the upper surface and drying to a dark brown or near black. The whole plant, except for a slight and almost unnotice- able pubescence on the terminal bud and anthers, is glabrous, an unusual feature in the genus. Large flowers and flowering parts, including the persistent almost sepaloid bracteoles, are features of distinction. The style, glabrous and three-parted up to 3 mm. is still another character of note. Except for a single specimen, Tsang 25656 from Kwangtung, the species seems to be confined to the province of Kwangsi. However, since several specimens were collected on or near the Kwangtung border, one may expect to find this species in future Kwangtung collections. 11. Adinandra Howii Merrill & Chun in Sunyatsenia 5: 131. 1940. DistRIBUTION: China (Hainan). eg Po-ting Distr., Hing-Lung, Sheon-Kai-Na, in forest, alt. 800 m., F. ow 73559 (TYPE, A), Aug. 30, 1935 (tree 7 m. high with brownish red bark: ee lustrous conti above, pale green beneath; stamens pure yellow). Small tree, 7 m. high; branches terete, glabrous, brownish red, the branchlets terete, grayish brown, glabrous except for a very fine appressed pubescence at the extreme tip of the new growth, the terminal bud densely sericeous. Leaves coriaceous, oblong-elliptic, 6-11 cm. long, 2.5—4 cm. wide, acuminate at the apex, tapering at the base, lustrous green, glabrous above, paler green, glabrous beneath, the margin serrulate, minutely glandular- apiculate, the veins 15-20 pairs, barely conspicuous on both surfaces, the petiole short, 3-5 mm. long, glabrous.: Flowers not seen. Fruit axillary, solitary; pedicel glabrous, 1.5-2 cm. long; bracteoles 2, caducous, nial fruiting calyx-lobes 5, imbricate, glabrous, broadly ovate, ca. 8 mm. , 5 mm. wide, the outer lobes larger with the margins entire, Rid inner ieee often glandular- margined. Fruit ovoid, glabrous, ca. 1 . diam., 3-celled, many-seeded; the placentae extending nearly ennai across the cell giving the appearance of six cells; style glabrous, entire, ca. 7 mm. long. Except for the terminal buds and the current year’s growth of branchlets, this species is entirely glabrous. Its closest relative is A. hatnanenstis Hayata, and it might be considered a glabrous variety of this species were it not for the three-celled ovary. The type specimen, the only specimen examined in this study, was gathered in the immature fruiting stage, hence 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 29 a description of flowers or mature fruit must wait for further collections. 12. Adinandra ryukyuensis Masamune in Trans. Nat. Hist. Soc. Formosa 24: 210. 1934. DistripuTion: Liu Kiu Islands. No SPECIMENS EXAMINED: Tree?; branches brown, the branchlets appressed- pilose at the ends, the buds densely ferrugineous- -pubesc cent. Leaves coriaceous, obovate, obovate- elliptic or elliptic, 3-6 cm. long, 1-2 cm. wide, acuminate at the apex, cuneate at the base, glabrous above, ferrugineous- -pubescent beneath, the margin entire, the petiole short, hirsute. Flowers axillary, solitary; pedicel ca. 2 recurved, ferrugineous- pubescent; bracteoles 2, "deltoid- lanceolate: fruiting calyx- lobes 5, imbricate, ovate, ca. 10 mm. long, 6 mm. wide, acuminate at the apex, sericeous- -pubescent. Fruit glabrous. Adinandra ryukyuensis and A. Zen-Tasiroi, the following species, must be linked together in this treatment. Both are from the Liu Kiu Island chain and may be identical. Translated descriptions are given here in an attempt to bring together as complete knowledge of the genus as possible. Although the description of the former is very incomplete, one cannot help but be impressed by the very close similarity between the two, even though varying terms have been used. Both have persistent bracteoles. In A. ryukyuensis the size is lacking and in A. Zen-Tasiroi the shape is not recorded. The calyx-lobes are almost identical in size, shape and pubes- cence. The fruit in both species is glabrous, a character which separates these two species from most others within their geographical range. In both species, the corolla and stamens are undescribed. 13. Adinandra Zen-Tasiroi Hatusima in Jour. Jap. Bot. 15: 133, 138, fig. 2. 1939, “Adinandra Millettii Bentham et Hooker f. sensu Maximowicz in Meél. (1886) 421, pro parte (plantas ex insl. Amami-ohsima).— Ito et Matumura, Tent, Fl. Lutch. 1 (1899) 324, pro parte. ee ae Pl. Formos. 2 (1911) 852. — Yamamoto et Mori in Sylvia 5 (1934) 31, et “Adinandra Millettii var. Matumura et Hayata, may on pine (1906) 45 DistrisuTiIon: Liu Kiu Islands. No SPECIMENS EXAMINED: Trees; branches brown-gray, subglabrous, the branchlets slender, densely appressed gray-hirsute. Leaves alternate, subcoriaceous, shortly petiolate, oblong-lanceolate, oblong or lanceolate, long-acuminate at the apex, cuneate at the base, 5- 8 cm. long, 1.5—2 cm. wide, shining glabrous above, appressed- -hirsute beneath, the margin entire or obscurely crenulate, the petiole 3—6 mm. long, densely hirsute. Flowers axillary, solitary; pedicels 1.5—2.2 cm. long, densely hirsute; bracteoles 2, opposite, immediately below the calyx-lobes, ca. 8 mm. long, 2 mm. wide, hirsute; en lobes 5, imbricate, oblong- -ovate or narrowly ovate- triangular, ca. | cm. mm. wide, acuminate at the apex, appressed gray-villous; oe aa stamens unknown; ovary glabrous, the style glabrous, ca. 6 m A discussion of the close relationship between 4. Zen. Tasiroi and the preceding species, A. ryukyuensis Masamune may be found under the latter named entity. Adinandra Zen-Tasiroi has the more complete description. According to the synonymy as recorded by Hatusima, this species has 30 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII been examined by several workers (Maximowicz, Ito and Matsumura, Hayata, and Yamamoto and Mori) and erroneously placed under 4. Millettii. This synonymy is quoted above merely for a record. 14. Adinandra yaeyamensis Ohwi in Act. Phytotax. Geobot. 7: 136, 1938. Distripution: Liu Kiu Islands. Riukiu Isl. sa Isl. (G. Koidzumi—type in Herb. Imper. Univ., Kyoto). — Same island, (G. Koidsumi; S. Sonohara).— Iriomote Isl., (G. Koidz at, ” Branches and branchlets terete, dark brown, the current year’s s branchlets — oblong to narrow obovate, 5-10 cm long, 2—3 cm. e, glabrous, shinin dark green above, paler, pilose at first beneath, ir subglabrescent, obtuse at the apex, contracted or abruptly narrowed at t ase, the margin lightly serrate, the veins somewhat conspicuous Seman, the midrib and petiole (4-5 mm. long) pubescent. Flowers axillary, solitary; — slender, terete, 1.5—3 cm. long, recurved, pilose; bracteoles 2, at t apex of the pedicel, caducous; calyx-lobes 5, broadly elliptic or i ae 3-5 mm. long, finally up to 7 mm, long, very obtuse at the apex, sparsely pilose on the external surface, ciliolate along the margin; ovary globose, sparsely long-pilose, the style ca. 10 mm. long, narrowed from the base, glabrous or nearly so, the seeds shining, ca. 1.5 mm. long, dark brown, punctulate. This species, like the others from the Liu Kiu chain of islands, is little known. No material was available for this study. The above description is merely a literal translation of Ohwi’s Latin and, along with the citation of specimens, is included here in order to bring together in one place all the literature of the genus to date. hwi states that the species can be separated from A. Millettii var. formosana (Hayata) Kob. by the longer pubescence, hardly appressed, the calyx-lobes obtuse at the apex, and the style, for the most part glabrous. : peru obtusissima Hayata in Sasaki, Cat. Governm. — rb. 347, 1930, nom. ud. - amamoto in Jour. Soc. Trop. Agric. 5: 347, fig. 9. 1933. — Kanehira, foes Trees, revised ed. 453, fig. 411. 1936. DISTRIBUTION: Formosa. No SPECIMENS EXAMINED: Branches and branchlets brown-gray, the branchlets of the current vear pubescent. Leaves alternate, petiolate, obovate, 4—7 cm. long, 2—3.5 cm. wide, very obtuse, rounded or mucronate at the apex, rarely retuse, cuneate or acute at the base, the margin entire or rarely lightly serrate near the apex, in the dried state strongly revolute, shining green above, pale, minutely pubescent beneath, the midrib sulcate above, prominent, minutely pubes- cent beneath, the primary nerves 10-12 pairs, elevated on both surfaces, extending from the midrib at an angle of 60°—70°, the petiole 5-7 mm. long, sulcate above, pubescent beneath. Fruit axillary, ae the pedi- cel slender, 2.5 cm. long, glabrous, swollen at the apex. Sepals 5, subequal, ovate, 8 mm. long, 5 mm. wide, obtuse at the apex, pubescent on the external surface. Fruit globose, ca. 11-12 mm. diam., sericeous-pubescent, 3-celled, the style persistent, 12 mm. long. Originally, in 1930, the name Adinandra obtusissima was listed as a nomen nudum by Sasaki, and attributed to Hayata. Later, in 1933, 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 31 Yamamoto gave a detailed description. He states that the leaves are very obtuse or rounded at the apex, 4—7 cm. long, 2—3.5 cm. wide, pubescent on the under surface. Both by Yamamoto and Kanehira the leaves are illustrated as nearly rotund at the apex and rather broad. The calyx- lobes are subequal, ovate, 8 mm. long, 5 mm. wide, obtuse at the apex, and pubescent. The pedicel is listed as 2.5 cm. long. The fruit is described as globose, 3-celled, ca. 11-12 mm. in diameter, sericeous-pubescent with a persistent style 12 mm. long. Although not mentioned in the description, evidence of pubescence on the style is depicted in the illustratious. Only three specimens of this species have been cited in literature, all of them collected by Matuda. The localities of collection are in southern Formosa near South Cape. They are Ariko-banti, Ako, Matuda 25443 and Mt. Daizyurin, Takao, Matuda 15544, 25545. I feel that this species is probably merely a variety of A. Millettii and may possibly be the same as A. Millettii var. obtusissima Kob., but since no opportunity has been available to compare the two entities, it is best to retain the specific status, even though dubious at the present time. Material for the following three species recently published by Gagnepain has not been available for this study. Except for the added citations of literature the descriptions are quoted as originally published. 16. ste etl ages Gagnep., n. sp. [in . Svst. Mus. Hist. Nat. Paris 10: 112. 42; i . Gén. Indo- Chine, Suppl. 1: ave 1943]. Arbor -- 12-15 m. alta. Ramuli floriferi 2 mm. crassi, pilosi, pilis rufis, appressis, dein glabrescentes. Folia oblongo-lanceolata, breviter acumi- nata, obtusiuscula, basi cuneata, infra pilosa, pilis appressis, supra viridiora glaberrima, 5—8.5 cm. longa, 2— 3. 5 lata, tenuia, firma, infra punctis sparsis notata, margine die serrata; nervi secundarii 12 in utroque latere, subtus magis conspicui, cum intermedio vix prominente; oe reticula- tim dispositae, ad margi inem conspicuores; petiolus 3-5 mm. longus, rufo pilosus, pilis appressis. Inflorescentiae pedicelli ery solitarii, arcuato- dependentes, dense et appresso-pilosi, ante anthesin 8-10 mm. longi; alabastra globosa, 8 mm. diam.; bracteae e calyce aie remotae, caducae, 4 mm. longae. — Sepala ovato- triangula, + 6 mm. longa lataque, dorso appresso-pilosa. Petala (haud evoluta) 5 mm. longa, dorso late sericeo. Staminum antherae connectivo acuminatae, dorso ventreque sericeae. Ovarium dense piloso- ae stylus sericeus, apice glabro, latitudine ovarli 14 longior. Fruc ANNAM: Dent du a prov. Ousme tr, ne st 382 (Poilane). TONKIN: réserve foresti¢re de Phong-du, prov. Tien-yen (A Che er). rs Cette espéce a ovaire velu, a style 2—3 fois cole long que l’ovaire se place aupres de A. rubropunctata Merrill et Chun. Elle en differe principale- ment: 1° par le connectif des étamines velu; 2° par le filet glabre. 17. Adinandra caudata Gagnep., n. sp. [in Not. Syst. Mus. Hist. Nat. Paris 10: 112. 1942; in Fl. Gén. Indo- Chine, Suppl. 1: 288. 1943]. Arbor elata, trunco 15 m. alto, cortice laevi rubescensque. Ramuli floriferi 2—3 mm. crassi, sericei mox glabri. Folia lanceolata, apice acuminato-caudata, basi cuneata, 8-17 cm. longa, 3-5 lata, supra, glabra 32 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII viridiaque, infra pilosa, pilis brevibus appressis, margine tenuiter serrata; laxum margine magis conspicuum efformantes; petiolus 7-12 mm. longus, piloso-sericeus, dein glabrescens. Inflorescentiae pedicelli axillares, 25-27 mm. post anthesin longi, ad calycem news incrassati, pilosi, pilis appressis; bracteae mox caducae vel nullae. — Se epala 5, appresso-pilosa, ovato-triangula, 15 mm. longa lataque, are glabra. Petala . . . Sta- mina... Ovarium east anthesin subglobosum, sericeo-pilosum, 15 mm. latum, 12 ‘altum, stylo apiculatum ; stylus columnaris, acuminatus, integer, 12 mm. longus, sericeus; semina immatura brunnea, nitida, tenuissime granulata, lenticulari-polyedra, 1.5-2 mm. diam.; embryo semiorbiculari. AnnAM: Dak-kiét, prov. Thanh-hoa, n° 1.840 (Poilane) ; Nui Bach-ma, pres Hué, n° 29.828 (Poilane); Nhatrang, n° 4.286 (Poilane). Ici les pétales, les étamines sont inconnus. Bien que les caractéres im- portants fournis par ces organes manquent, nous croyons que cette espece pourra assez facilement étre distinguée de tout autre. 18. Adinandra ce Gagne} . sp. [in Not. Syst. cae Hist. Nat. Paris 10: 113. 1942; in Fl. Gén. Indo. ‘Chine, Suppl. 1: 283. 1943]. Arbor 8-15 m. ie: trunco 15-25 cm. crasso. Ramuli floriferi 1—1.5 mm. diam., glaberrimi, grisei. Folia lanceolata, basi cuneata, apice acuminato-obtusiuscula, 6-10 cm. longa, 2.5—3 lata, utrinque ne. mar- gine leviter serrata; nervi secundarii 8—10 utroque latere, ad marginem arcuati, contluentesque; venulae tenues, reticulatim dispositae; petiolus gracilis, 7 mm. longus, glaber, supra canaliculatus. Inflorescentia ad ramulos 1—2-ennos inserta; pedicelli axillares, solitarii vel geminati vel terni, graciles, _ mm. longi, ad apicem gradatim incrassati; alabastra conica, 9-10 mm. longa; bracteae oppositae, e calyce vix remotae, ovato- obtusae, 2-5 mm. longae, ciliolatae.— Sepala 5, reniformia, 2 longa, 3-4 mm. lata, margine ciliolata, supra sericea. Petala 5, ‘triangula, mm. longa, supra basin 5 mm. lata, glaberrima. Stamina numerosa; filamento plano, brevi (1.5 mm.), breviter piloso; anthera_ triangulo- mucronata, 3 mm, longa, mucrone 1 mm. longo incluso; loculi praesertim m berrimum; stylus longe conicus, 4-5 mm. longus, glaberrimus; loculi 3 pluriovulati, imate ad parietes radiantibus. Fructus immaturus globosus, glaber, 7-8 mm. diam., maturus niger (Poilane). ANNA massif du Bi-doup, prov. Haut-Donnai, n°* 30.724 et 30.848 (Poilane). CambBopGce: Pu-chom, prov. Kompongchnang, n° 28.793 (Poilane) Differe de A. laotica Gagnep.: 1° ramules plus gréles; 2° feuilles 2 fois plus courtes et étroites; 3° a du limbe denticulée et veinules visibles; 4° pédicelles un tiers plus courts; 5° bractées opposées; 6° sépales réniformes; 7° connectif des anthéres égalant la moitié des loges, glabre. DOUBTFUL SPECIES 19. Adinandra kweichovensis Hu in Bull. Fan. Mem. Inst. Biol. 8: 145, 1938. Although no material of the species has been available for oe best I feel very dubious of its placement in the genus Adinandra. scrip- tion of the species leads one to believe that it belongs to the es aie 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 33 Hu states in his description, ‘‘arbor ad 4 m. alta, omnino glabra; folia . margine integra leviter revolutaque . . . petioli r robu usti ... 1.5 cm. ongi: . fructi 1-3, . . . sepala 5 ovata 5 mm. longa, marginibus ciliata; stylus pets 5 mm. ionease bacca globosa 8 mm. diametro; semina pauca, magna compressa ... 3 mm. longa, 2.5 mm. lata.” The completely glabrous plant, the entire margin of the leaves, the petiole, robust and 1.5 cm. long, the margin of the sepals ciliate, the globose fruit 8 mm. in diameter and finally the few seeds are all features of Cleyera japonica Thunb. The only variations, in my opinion, are the sepals (5 mm. long) and the style (5 mm, long). In Cleyera the sepals are seldom longer than 3 mm, and rounded and the style is usually, not always, two- or three-parted. EXCLUDED SPECIES ae we ae - in Jour. Arnold Arb. 26: 65. 1945=Ternstroemia ? insignis Wu ot. Jahrb. 71: 195. 1940. ice no flowers or fruit remain on the specimen, Wang 39626, designated by Li as the type of 4. bracteata, from the fragments of his dissection showing the many glabrous stamens arranged in series, the leaf-buds, and the sub- or pseudo-verticillate arrangement of leaves and branchlets, there seems little doubt that the species does not belong to the genus Adinandra, but probably to Ternstroemia. There is a close relationship between Li’s species and Ternstroemia insignis Wu, collected at the same locality Lao Shan in Kwangsi. How- ever, from the descriptions there can be no certainty that they are identical. There is variation in leaf-size and stamen measurements. Li stated that the stamens of A. bracteata were “1.1-1.3 cm.” long. The fragments of his dissections show stamens measuring 5—6 mm. in length, which compare favorably with those (5 mm. long) found in 7. insignis. However, these stamens may have been taken from an unopened flower-bud, since the filaments appear very short and undeveloped. Adinandra Chingii Metcalf in Lingnan Sci. Jour. 11:19. 1932=Cleyera japonica Thunberg emend. Sieb. & Zucc. Fl. Jap. 153. 1941. Adinandra_ obscurinervia nines & Chun in Sunyatsenia 2: 283, fig. 35. 1935= leyera japonica Thunberg emend. Sieb. & gr var. lipingensis (Handel- Mazzetti) Kobuski in Jour. Arnold Arb. 18: 127. 1937. ne stenosepala Hu in Bull. Fan. Mem. Inst. Bot. Ser. 8: 146. 1938= oxylon (Planchonella) atenoseenla Hu in Bull. Fan. Mem. Inst. Biol. Ser. 10: 129 1940. PHILIPPINE SPECIES Fifteen species have been described from the Philippine Islands under the genus Adinandra. The present study recognizes eight. Of these eight species six have been described by Merrill and one each by A. D. E. Elmer and C. B. Robinson. Three of the earlier described species have been transferred to other genera: A. coriacea Elmer to Ternstroemia gymnanthera 34 (W. & Kobuski, a (Merrill ) Kobuski. JOURNAL OF THE ARNOLD ARBORETUM the genus Cleyera in the Philippine Islands. The Philippine species of Adinandra form a distinct group among them- selves. species than with the Indo-Malaysian grou Below are listed the characters helpful in specific identification. (VoL. XXVIII A.) Sprague, A. urdanetensis Elmer to T. urdanetensis (Elmer) nd A. montana Merrill to Cleyera japonica Thunb. var. montana This last-named variety is the first representative of Their relationships are more clearly with the Eastern Asiatic Under these characters are recorded the species as far as present-day knowledge CHART FOR PHILIPPINE SPECIES OF ADINANDRA permits. Bracteoles PERSISTENT CADUCOUS apoensis elliptica levtensis luzonica he Vacgregorit Loheri l nigro-punctata maquilingensts Corolla-lobes PUBESCENT GLABROUS Cc alyx- lobes PUBESCENT elliptica ley see luzonica cess ae GLABROUS apoensis Loheri Macgregorii nigro-punctata Filaments PUBESCENT GLABROUS apoensis None apoensis elliptica elliptica leytensis Loheri leytensis luzonica Macgregorii Loheri nigro-punctata maquilingensis luzonic Macgregorii maquilingensis nigro-punctata Ovary Style PUBESCENT GLABROUS PUBESCENT GLABROUS ition apoensis ee ere ners luzo leytensis elliptica prabiae oheri leytensis es nigro-punctata Loheri luzonic Macgregor nigro-punctata Ovary THREE-CELLED FIVE-CELLED elliptica el ‘el leytensis luz Lohe nigro- oon Macgregorii maquilingensis 1947 | KOBUSKI, STUDIES IN THE THEACEAE, XV 39 KEY TO THE PHILIPPINE SPECIES A. Ovary and fruit pubescent; bracteoles caducous. B. Ovary and fruit 5-celled; calyx-lobes, ovary and fruit covered with a dense canescent osesaerag a nts pubescent; leaves conspicuously reticulate- r surfac 20. A. luzonica. . Ovary and fruit sally 3- Sila (rarely 4-celled) ; pubescence Ee present siraciine or appressed, not canescent; filaments glabrous; leaves not con- spicuously reticulate-veined on the kare surface. Leaves membrana ics) ow o s ceous or submembranaceous, narrowly elliptic, the apex long, tapering; pedicels cty slender, up to 3.5 cm. long; s 7 e always (Sohatdshaed 110) 40) Ce ee earn eR nrhaiin tO: 2 A, elliptica. CC. Leaves coriaceous; pedicels ergs seldom over 2-2.5 cm. long: style entire with a tendency to separate a e apex, or divided. D. Style straggling-pubescent, pe least the lower half; ue caducous (when ue resent broadly ovate, ca. 2 xX 2 mm.); 7x-lobes broadly ovate c DCR 0) v1 01 Cae area a ee emer mr eee pe Macgregorii. a DD. Style beeen bracteoles caducous _ present ag obovate ca. mm.) ; calyx-lobes narrow-ovate ca. 8 APS SINS eee tees ost SP a Sete gt eterna euceedh Waa wis ws acd bage w esas Su acey Serereranonene 23,248 re ec AA. Ovary and fruit glabrous; bracteoles persistent. . Ovary and fruit 3-celled. C. Ov ules and seeds few; seeds up to 5 mm. long, flat, seldom more than 20 ma Senet sire entre rsa n yey S76 Gr sites bisa b ar'g ad SoheneRe Reread e sian eam Got e 24. A. Loheri. CC. Ovules and seeds many; seeds seldom over 2 mm. long, flat, many maturing otten “more than: 100) ns oe hace a ese are 0 e.ieesarate 00s 25. A. leytensis. BB. Ovary and fruit 5-celled. C. Br ee rounded, ca. imbricate, broadly ovate, unequal, 6-7 mm. long, ca. 5 mm. wide; not as thick as in most species with less conspicuous variation at the margin, Peart -denticulate, the two outer lobes straggling-pubescent, the inner lobes quite glabrous; corolla-lobes 5, connate at the base, ovate, quite ce at the apex, densely pubescent in the middle of the exterior surface, ca. 10 mm. long, 4—5 mm. wide; ee ca. 30, appearing as a single series, the filaments glabrous, 2-3 mm. long, adnate to the base of the corolla, joined inconspicuously, ee almost free, the anthers 2-4 mm. long, densely pubescent, with an apicule ca. 1 mm. long; ovary subconical, 3-celled, densely pubescent, tapering at the apex into a glabrous style ca. 1 cm. long, the ovules many. Fruit ca. 1 cm. diam., 3-celled, straggling-pubescent, multi-ovulate, the seeds minute, dark brown-black The outstanding characters o this species are: (1) the submembrana- ceous elliptic leaves; (2) the generally long slender pedicels; (3) the 3-celled ovary and fruit; (4) the 1 cm. long entire glabrous style; (5) the thin calyx-lobes; (6) the glabrous, nearly free filaments and pubescent anthers; (7) the a always caducous bracteoles; and (8) the rather straggling pubesce Adinandra Robinsonii Elmer is included here as a synonym. After studying the species, I can find no characters suitable for separating them. Elmer, in introducing A. Robinsonii, remarked that it was very close to A. elliptica and offered no basis for separation. Later, Merrill suggested that the relationship seemed too close for separation, although he continued to list them as distinct entities. Questionably cited here is Ramos & Edano 45597. The style is glabrous and entire, but the leaves are thicker, more coriaceous than most specimens of A. elliptica and the base is less acute. Walliams 2533 from Mindanao 38 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIIL and Merritt 8834 from Mindoro have leaves large for the species. How- ever, in Williams 2533, the dissections of the flowers showed no differences from typical A. elliptica. My dissections of the ovary of the Williams number were unsatisfactory. Merritt 8834 is a fruiting specimen. Several dissections showed the fruit to be typically 3-celled. However, on the same specimen were fruits incompletely 3-celled appearing 5- or 6-celled. The leaves of Elmer 10185, 10185A from the island of Negros are smaller and the pubescence generally more dense, but typical of the species. This latter character is true especially of the pedicel and calyx-lobes. VERNACULAR NAME: Puyaka (Mangyan). . Adinandra anata vn in rea Jour. Sci. 9: 319. 1914; Enum. Philipp. Fl. Plts. 3: — Melchior in Nat. at es ed. 2, 21: 144. 1925. DISTRIBUTION: coe ae Lean) Luzon: Benguet Subprov.: Pauai, J. K. Santos [P.B.S.] 31924 (US), April- June 1918.— Mt. Pulog, M. Ramos & G. “Edaiio [P.B.S.] 40398 (A), Sept. 1921. — Mt. Pauai, E. Ouisumbing & M. Sulit [P.B.S.] 82356 (A, NY), March 1931. — Pauai, in forest, alt. 2100 m., R. C. MacGregor [P.B.S.] 8425 (isotypes, NY, US), June 1909. — Pauai, M. S. Clemens 9234 (A, G, US), Jan. mp —R. J. a [P.B.S.] 18372 (US), Jan. 1909. Lepantc Sub ov.: Mt. , M. Ramos & G. Edano [P.B.S.] 40276 (A, US), Sept. ven ~ Ball (Mt. aie E. dD. Merrill 4650 (NY, US), Nov. 1905 Tree 15 m.; branchlets terete, grayish, glabrous, the very young branch- lets invested with a dense, tawny, appressed-villous pubescence. Leaves coriaceous, elliptic to oblong-elliptic, 5-10 cm. long, 2—4 cm. wide, apex sually acuminate, acute to somewhat rounded at the base, the margins minutely but distinctly glandular-denticulate, the upper surface glabrous except occasionally near the very base, the lower surface with scattered, appressed tawny pubescence becoming glabrescent, the veins ca. 15 pairs, anastomosing, about equally conspicuous on both surfaces, the petiole 3—5 l eel pedicels stout, recurved, up to 2.5 cm. long, measuring as m as 3 mm. diam. near the calyx, pubescent; bracteoles 2, caducous; i. lobes S. pee unequal, up to 10 mm. long, ca. 7 mm. wide, ovate, acute at the apex, the outer lobes straggling-pubescent over the entire external cutee the inner lobes pubescent except along the margin, glandular-denticulate; corolla-lobes 5, unequal, connate at the base, obovate, up to 14 mm, long, 7-10 mm. wide, tawny-pubescent on the median portion of the external surface; stamens ca. 30, unequal in length, 8-10 mm. Jong, probably seriate, the filaments firmly adnate to the base of the corolla, glabrous, 3-5 mm. long, the anthers noticeably large, ca. 5 mm. long, usually equal or subequal, hirsute, prolonged at the apex into an apicule; ovary pubescent, 3- or 4-celled, tapering gradually at the apex into a broad gradually narrowed style (in type especially) which is split into 3 or 4 parts at the apex, pubescent except near the tip, the 3 or 4 stigmas appearing oblique. Fruit not seen. From the material examined, this species seems to be localized in the Pauai area of Benguet Subprovince. All the material cited above, save two sheets from the adjoining Subprovince of Lepanto, are from Benguet. These Lepanto sheets are Merrill 4650 and Ramos ¢» Edano 40276, collected at Mt. Data 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 39 The ovary is generally three-celled and the style three-parted, three- ridged or entire. The type specimen seems to be the exception. Only two dissections were possible from the type and both showed a four-celled ovary. This was to be expected as the style, in both cases, was four-parted. On the same specimen were styles that were three-parted. In the type, the style is very thick at the base, and although tapering somewhat towards the apex remains quite bulky. In other cases, where the style is three- parted, the tapering is more finely drawn out. “here is variation in the extent of pubescence of the style. It may be present on the lower half only or may extend up to the divided portion. Never is the style completely pubescent. The leaves are coriaceous, resembling in shape those of A. elliptica C. B. Robinson, the closest relative, the base, however, usually being more rounded. In A. elliptica, the floral parts are generally smaller. The calyx-lobes measure 6—7 mm. and the corolla-lobes 10 mm. in length. The style is glabrous and always entire. The leaves are membranaceous or submem- branaceous and taper at both ends. These two species are very difficult to separate, especially in the early stages of leaves and flower. VERNACULAR NAME: Batinai (Igorot). 23. Adinandra maquilingensis Merrill in Philipp. Jour. Sci. Bot. 9: 317, 1915; Enum. Philipp. Fl. Plts. 3: 73. 1923.— Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. — Sasaki, Cat. Governm. Herb. 347. 1930. DistrRIBUTION: Luzon (Laguna, Tayabas). Luzon: Laguna: Mt. Maquiling, M. Ramos [P.B.S.] 13650 (1sotyPEs, C, US), Sept. 1911.— Los Bafios, Mt. Maquiling, A. D. E. Elmer 17687 (A, C, G, M, NY, US), 18200 (A, C, G, M, NY, US), June-July 1917. Tayabas: Umiray, A. Loher 13999 (A) Small tree ca. 12 m. high; branchlets terete, brownish, generally glabrous, except ‘at the very tips where finely appressed-pubescent, Leaves coria- ca. 8 mm. long, 6—6.5 mm. wide, pubescent in the center, glabrous at the membranous margin; corolla-lobes 5, connate at the base, unequal, obovate, tapering quickly and abruptly at the base, obtuse at the apex, 11.5—12 mm. long, 5-7 mm. wide, pubescent on the median portion of the exterior sur- face; stamens ca. 35, 7-9 mm. long, unequal, the filaments 4—6 mm. long, generally glabrous, pubescent only where joined to the anthers, the anthers 40 JOURNAL OF THE ARNOLD ARBORETUM {[VOL. XXVIII ca. 3 mn. long, densely hirsute; ovary pilose, 3-celled, multi-ovulate, the style hirsute, 7 mm. long, 3-parted, the apices glabrous. Mature fruit not seen. From the material examined, it appears that this species is localized in the Mount Maquiling area in the Province of Laguna on the Island of Luzon. The bracteoles although caducous do not fall as quickly as those in the majority of species. When present, they are unusual in character, being long-obovate (6 < 3 mm.) and nearly equalling the calyx in length. The calyx-lobes themselves are characteristic, long and quite narrowly ovate, tapering to a distinct apicule, and covered with a straggling pubescence. The anthers are, as usual, hirsute with the pubescence extending down the filaments for a short distance which are glabrous for most of their length. The ovary is pilose, 3-celled and multi-ovulate, tapering into a style hirsute for almost its entire length, glabrous only at the very apex which is 3-parted. 24. seas Loheri Merrill in Philipp. Jour. Sci. Bot. 9: 318. 1914; rigs — Plts. 3: 72. 1923.— Melchior in Nat. Pflanzenfam. ed. 2, 21: Pr seksi rostrata Merrill, in op. cit. 316. — Melchior, loc. cit. Adinandra coriacea Merrill, in op. cit. 317, — Melchior, loc. cit. DistrRIBUTION: Luzon (Rizal, Laguna). Luzon: Rizal: Oriud, in primary forests at medium altitudes, A. Loher 5604 (1soTYPE, K), Dec. 1905. oe A. Loher 12642 ‘ee Oct. 1912.— A. Loher 14486 (A), Oct. 1913. Lagun Dahican River, M. Ramos [P.B.S.] 1567. (1s0- TYPES of A. rostrata, M, NY), Sept. 13, 1912. — F. L. penned [P.B.F.] 24663 (NY, US), Nov. 1915 Tree, generally glabrous except the terminal vegetative buds and very young branchlets: branchlets terete, grayish, the terminal buds tawny- villous. Leaves coriaceous, oblong-elliptic, glabrous or glabrescent, 8-14 cm. long, 3.5—6.5 cm. wide, broadly rounded to broadly or shortly acumi- nate at the apex, acute at the base tapering into the petiole, the margin crenulate to crenulate-denticulate, the veins 11-15 pairs, anastomosing near the margin, the petiole 5-7 mm. long. Flowers axillary, solitary; pedicels glabrous, curved, 2—4 cm. long; bracteoles 2, opposite, immediately below the calyx, persistent, unequal, broadly triangular to suborbicular, 2.5 x 3 mm. and 2.5 2.5 mm., scattered appressed-pubescent, the margins ciliolate; calyx-lobes 5, imbricate, thickened, glabrous, gui s rounded (often splitting ) at the apex, unequal, wider than long, 5.5—6 mm. long, 6-7 mm. wide; corolla-lobes 5, connate at the base, ‘obovate to broadly obovate, ca. 1.5 cm. long, 1.2 cm. wide, pubescent on the median portion of the dorsal surface; stamens 40-50, adnate to the oe of the corolla, unequal, 7-11 mm. long, the filaments 4— 8 mm. long, joined nearly their entire length, hirsute on the external surface, the anthers ca. 3 mm. long, oblong, acute, hirsute; ovary glabrous, conical-ovoid, 3-celled, multi- ovulate, tapering at the apex into a glabrous, entire style ca. 6 mm. long. Fruit glabrous, globose, ca. 1.5 mm. diam., 3-celled; seeds ca. 25, dark, shining, scrobiculate, 5-6 mm. long. The nearly glabrous character of this species is one of its unusual features. The terminal buds and very young branchlets, the dorsal surface 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 41 of the corolla-lobes and the stamens appear to be the only exceptions. The anthers are always hirsute, even in the most glabrous species. Another feature of note is the persistent, opposite bracteoles. Merrill, in his original description of this species, records the ovary as 5-celled. In all the material examined the ovary and fruit are consistently 3-celled. Undoubtedly Merrill had poor material at his disposal when he devised this species, since the fruit is not described and the ovary is recorded as 5-celled. An isotype in the Kew herbarium which has been available for this study possesses both fruit and flowers. Had Merrill had such a specimen he would have seen the relationship between A. rostrata and A. Loheri at the time of publication, since they were published simul- taneously. Adinandra rostrata was published from the fruit alone. The type material of this latter entity shows the very large 3-celled fruit with large (5-6 mm. long) seeds which are few (10-20) in number. The pedicel is unusually long, measuring up to 4 cm. Adinandra Loheri was described from flowering material in which the pedicels measured 2 cm. in length. Unfortunately the two specimens used as types of these two species represented the extremes in pedicel length, and since one was described from flowers and the other from the fruit they appeared to be different. Since then material has been collected which includes this variation in pedicel-length. In Loker 12642, a flowering specimen, the pedicels vary from 2—3.5 cm. in length on the same specimen. Although several Philippine species are known to have 3-celled fruit, this is the only species in which the fruit has been described. It agrees with those Indo-Malayan species with 3-celled fruit in having fewer but larger seeds than the species in which the fruit is 5-celled. VERNACULAR NAME: Malambot (Tagalog). 25. Adinandra leytensis Merrill in Philipp. Jour. Sci. Bot. 9: 377. 1914; Enum. Philipp. Fl. Plts. 3: 72. 1923.— Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 925. DIstTRIBUTION: Leyte. LEYTE: eres near eee in forests, alt. 600 m. C. A. Wenzel 760 (1sotypEs, A, C, M, US), May 27, Tree 17 m. high: branchlets tepete, glabrous, brown, the terminal buds appressed-tawny-pubescent. Leaves coriaceous, subelliptic, 5-8 cm. long, 2.5-4 cm. wide, obtuse to rounded at the apex, narrowly acute at the base, glabrous, somewhat shining, greenish above, roughened, darker, appressed- pubescent beneath, the margin obscurely denticulate, the veins ca. 12 pairs, prominent on both surfaces, anastomosing near the margin, the reticulations very distinct on both surfaces, the petioles ca. 5 mm. long. Flowers axillary, solitary; pedicels recurved, stout, ca. 1.3 cm. long, scattered- pubescent; bracteoles 2, opposite, immediately below the calyx, persistent, unequal, ca. 2 X 3 and 3 * 4 mm., broadly rounded, closely appressed- pubescent, the surface rough, thick, the margins ciliolate; calyx-lobes 5, imbricate, rounded, finely appressed-pubescent, unequal, the outer two Ca. 5 mm. long, 6-6.5 mm. wide, thick with little or no scarious margin, eciliate, the inner three gradually thinner toward the margin becoming scarious, somewhat ciliolate; corolla-lobes 5, connate at the base, obovate, 42 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII 13 mm. long, 8.5-10 mm. wide, rounded at the apex, tapering rapidly nee thickened at the base, finely appressed- pubescent on the externa! surface except along the scarious margin; stamens ca. 40, unequal, 6—10 mm. long; filaments extremely variable in length, 2-6 mm. long, joined the entire length of the shortest filaments, densely hirsute except for ca. 1 mm. at the base of the external surface where they are adnate to corolla, the anthers ca. 4 mm. long, approximately equal in length, hirsute, lanceo- late; ovary glabrous, conical-ovoid, 3-celled, multi- ovulate, tapering at the apex into an entire glabrous style ca. 7 mm. lon The characters helpful in distinguishing this species are: (1) the unequal persistent bracteoles; (2) the roughened, thick, pubescent calyx- lobes; (3) the long stamens with the filaments joined their entire length forming a tube and sufficiently adnate to the base of the corolla (1 mm.) to fall off intact with the corolla; (4) the glabrous 3-celled ovary; and (5) the glabrous entire style. The veining of the leaves of the species resembles very much that found in those of A. luzonica Merrill. However, the apex of the leaf of the present species is rounded and the pubescence of the under surface less distinct. The midrib also is more roughened. 26. Adinandra nigro-punctata Merrill in Philipp. Jour. Sci. Bot. 9: 320. 1914; Enum. ae Fl. Pits. 3:73. 1923.— Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. nisin TION: Leyte. Leyte: Dagami, in forests, M. Ramos [P.B.S.] 15355 (isotype, US), Aug. 13, 1912. Tree ca. 10 m. high; promenlets numerous, terete, glabrous, grayish brown, pubescent at very tip only and reddish brown when very young, the terminal buds pubescent. i coriaceous, glabrous except when unfold- ing, elliptic-oblong, 4—7 cm. long, 1.5—3 cm. wide, the apex broadly blunt- acuminate, minutely retuse, the base acute, the margin obscurely epee denticulate, the veins 7-8 pairs, equally evident on both surfaces, the petiole ca. 5 mm. long. Flowers axillary, solitary; pedicels glabrous, stout, ca. 2 cm. long; bracteoles 2, persistent, opposite, immediately below the calyx, gli ibrous, saeeuae suborbicular to obovate, ca. 3 mm. long, 3—4 mm. wide: calyx-lobes 5, imbricate, glabrous, subequal, rounded at the apex, the 2 outer Ae smaller, ca. 7-8 mm. long and broad, the margin lightly ciliolate, the inner 3 lobes longer, ca. 10 mm. long, 7-8 mm. wide, the margin more onan ciliolate; corolla-lobes [descrip. fide Merrill] 5, broadly rounded, slightly retuse, the median exposed parts (dorsal surface) very densely hirsute, otherwise glabrous; compos (fide Merrill) about 50, the filaments 5 mm. long, densely hirsute, the anthers ovate- lanceolate, somewhat acuminate, 4 mm. long, densely hirsute; ovary glabrous, ovoid, 5-celled, the ovules very numerous in each cell; style glabrous, 12 mm. long, the stigma minute, subcapitate The only authentic material available for this stays is a sterile represent- ative of the type, deposited in the United States National Herbarium. It has been necessary to rely, to a great extent, upon the original description by Merrill. Features of distinction appear to be: (1) the five-celled ovary; (2) the glabrous style, 12 mm. long; (3) the pubescent filaments, 4 mm. long; 1947 | KOBUSKI, STUDIES IN THE THEACEAE, XV 43 (4) the large corolla-lobes, 18 mm. long, pubescent on the external sur- faces; and (5) the glabrous, subequal calyx-lobes, 7-10 mm. long, 7-8 mm. wide. Very closely allied, and perhaps conspecific with it, is A. apoensis Elmer from Mindanao. However, until more material is available for study, I prefer to retain them both as species. The five-celled ovary is found only in two other species, namely A. luzonica Merrill and A. apoensis. In the former species, the ovary, style, ‘calyx-lobes, bracteoles and pedicel are covered with a dense, canescent pubescence which makes separation quite easy. 27. Adinandra apoensis Elmer, Leafl. Philipp. -: 8: 2833. 1915, “‘apoense.”’ — Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. Adinandra Merrillii Elmer, in op. cit. 2834. — me loc. cit. DisTRIBUTION: Mindanao (Davao, Agusan). Minpanao:. Davao: Todaya, Mt. Apo, A. D. E. Elmer 11265 (isotypes, A, C, G, M, NY, US), Aug. 1909.— Mt. McKinley, R. Kanehira 2704 (NY), Aug. 1930. Agusan: Cabadbaran, Mt. U angen he D. E. Elmer 14088 (isotypes of A. Merrill, A, C, G, M, NY, US), Oct. Tree ca. 15 m. high; branchlets terete, relatively short and numerous, grayish, the terminal bud appressed-pubescent. Leaves coriaceous, gla- brous, elliptic-oblong, 5-9 cm. long, 2.5—4 cm. wide, broadly rounded or obtuse at the apex, occasionally slightly emarginate, cuneate at the base, the margin crenately serrate towards the apex, entire towards the base, the veins 7-8 pairs, the primary veins evident on both surfaces, the retic- ulations prominent on the lower surface only, the petiole glabrous, ca. ong. Flowers axillary, solitary; pedicel 1-1.5 cm. long, stout, glabrous, recurved; bracteoles 2 ’ persistent, aaa opposite, immediately below the calyx, unequal, ey. mm. long, 2.5-3 mm. wide, broadly rounded, lightly glandular- -denticulate along the margin; calyx- lobes 5, imbricate, glabrous, unequal, thickened at the base, broadly ovate to rounded, 8-9 m m. long, 7-8 mm. wide, the margins quite entire; Feil lobes 5, connate - the base, obtuse at the apex, ca. 13 mm. long, 8 wide, pubescent on the median portion of the external surface; Erte ens ca. 35, unequal in length, 6-11 mm. long, probably seriate, the filaments 2-7 mm. long, joined at the base ag and adnate to the corolla, very hirsute, the anthers usually ca. 4 mm. long, occasionally shorter, apiculate, equally hirsute; ovary glabrous, elabose (not conical) , 5- celle d, multi- ovulate, the style terete, entire, glabrous, 1.5 cm. long, rising abruptly from the ovary, not tapering. Fruit not seen. This species appears very distinct from A. leytensis Merrill, with which it formerly was considered synonymous. In A. apoensis, the style is easily 1.5 cm. long, and rises abruptly from a globose, 5-celled glabrous ovary. In A. leytensis the style is only 7 mm. long, approximately one-half as long as that found in A. apoensis, and it tapers gradually from a conical 3-celled ovary. The leaves in A. leytensis have ca. 12 pairs of veins and the reticulations on the under surface are very distinct. Also the calyx-lobes are wider than long, very rounded at the apex and measure 5 & 6-6.5 mm. In A. apoensis there are only ca. 7 or 8 pairs of veins and the reticulations, 44 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII although obvious, are less conspicuous. The calyx-lobes are broadly ovate, longer than broad, and measure 8-9 X 7- ; Cited here as a synonym is A. Merrillii Elmer. Having studied the types of both species, I can see no differences separating the two. Elmer men- tions in the original description of A. Merrillii that the ovary was “‘very lightly woolly.” The material was very sparse, hence only a single ovary was studied. This appears to be glabrous and five-celled. No mention of the number of ovary-cells was made in the original description. Elmer also refers to the number of stamens as “3”. This undoubtedly is a typographical error. An accurate count could not be made from the very poor specimen I examined. However, the number appears to be 30 or more. The closest relative appears to be A. nigro-punctata Merrill. So far, material of A. apoensis has been collected only from Mindanao. Adinandra nigro-punctata seems to have been collected only once, from the type locality in Luzon. Future collections may show additional distinguishing characters. VERNACULAR NAMES: Malagsam (Bagobo), Sangnanan (Monobo). EXCLUDED SPECIES Cleyera japonica Thunberg emend, Sieb. & Zucc. var. montane Seite comb, nov. Adinandra montana Merrill in Philipp. Jour. Sci. Bot. 5: 363 ; Enum. Philipp. Fl. Plts. 3: 73. 1923.— Melchior in Nat. Pflanzenfam. ed. - ec 144. 1925. Not A. montana Ridley (1915) DistRIBUTION: Luzon rs Cagayan, Camarines, vaieey Norte, Laguna, Nueva Ecija, Nueva Viscaya, Rizal, Tayabas), enh (Bukidnon, Davao, Misamis). 1ZON: Prov. Benguet: Mt. Pulog, H. 4 pane M. L. Merritt & T. C. Zschokke [P.B.F.] 18149 (US), Jan. 1909. biased in the Oaks,” alt. 2300 m., E. A. Mearns [P.B.S.] 4420 (US), July 1907.— Pauai, J. K. Santos [P.B.S.] 31932 (NY), April-June 1918.— Mt. Baudan, M. Ramos & G. Edano [P.B.S.] 40314 (A, US), Sept. 1921.— Mt. Pulog, M. Ramos & G. Edano [P.B.S.] 44906 (A, NY), Feb.- March 1925. Prov. Cagayan; Mt. Dos Cuernos, along streams near the foot of the mt., alt. 10CO m., WM. Ramos [P.B.S.] 76995 (A), April 26, 1929, (3-6 m. high). — Mt. Tahun, on ake of slopes, alt. 2000 m., M. Ramos [P. ve 77107 (A), May 19, 1929 (4 m. high; flowers whitish yellow). Prov. Camarine M. Ramos 1587 (M, NY), Dec. 1913. Prov. Ilocos Norte: Mt. Palimbin, i. Ramos [P.B.S.] 33318 (A, US), Aug. 1918. ath Laguna: Mt. Cristobal, F. Canicosa [P.B.F.] 28980 (A), March 1922.— Mt. Bonahin, moist ridge, alt. a oT: ue [P.B.F.] 31116 (NY), April 4, 1929 oe io m. high; flowers white). rov. Nueva Ecija: R. J. Alvarez [P.B.F.] 22191 (US), Dec. 1910. — Mt. Umingan, Ramos & G. Edafio [P.B.S.] 26255 (A, US), Aug.-Sept. 1916. Prov. Nueva Viscaya: Mt. Alzapan, Ramos & G. Edano [P.B.S.] 45623 (NY), May-June 1925. Prov. Rizal: Montalban, A. Loher 12346 (A), Oct. 1909.— Mt. Lumutan, M. Ramos [P.B.S.] 42298 (M), April 1923. Prov. Tayabas: Mt. Camatis, Alcasir & G. Edano 4513 (A), 4487 (A), May 1939.— F. Canicosa [P.B.F.] 30017 (US), April 1925.— H. M. Curran & oe L. Merritt [P.B.F.] 7846 (US), Nov. 1907. Minpanao: Prov. Bukidnon: Mt. Lipa, M. Ramos & G. Edano [P.BS.] 38493 (A), ae -July 1920.— Mt. Candoon, M. Ramos & G. Edano [P.B.S.] 38933 (A, US), June-July on a of Tanculan, £. Fenix [P.B.S.] 26063 (A, ue July 1916. Prov. Davao: Mt. ee ewclayl ridge, alt. 1900 m., Elmer 10620 (A, C, G, M, NY, US), May 1909.— Todaya, Mt. Apo, macdaa cre 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 45 alt. 2300 m., A. D. E. Elmer 11465 (A, 7 M, NY, US), Aug. 1909. Prov. Misamis: Mt. Matiidang. E. A. Mearns & W. I. Hutchinson [P.B.F.] 4558 (isotypes of A. montana, M, NY, US), May 1906. Small tree 4-10 m. high; branchlets terete, reddish brown, glabrous even to the very young growth, including the terminal buds. Leaves coriaceous, glabrous, elliptic-ovate to elliptic-oblong, 5-8 (—10) cm. long, 2-3.5 (—4) cm. wide, usually more or less acuminate at the apex, sometimes blunt and shortly so, acute at the base, the margin entire, somewhat revolute in drying, the veins generally obscure, occasionally faintly visible on the upper surface, the petiole 5-10 mm. long, glabrous. Flowers axillary, up to 5 in an axil, sometimes solitary; pedicel stout, usually 1—2 cm. long, occasion- ally slightly shorter, glabrous; bracteoles 2, opposite or alternate, end caducous, when present minute, quite sca le- like, deltoi 5 m long; calyx- lobes 5, imbricate, glabrous, unequal ca. 3 mm. coe rene shorter, ca. 3 mm. wide , fimbriate- ciliolate along the margin, the outer lobes abruptly apiculate, the i inner lobes more rounded; corolla-lobes 5, glabrous, obovate to narrowly obovate, 9-12 mm. long, 4-5 mm. wide; stamens ca. 25, unequal in length, 4-7 mm. long, the filaments glabrous, adnate to the base of the corolla, sometimes appearing free, the anthers ca. 1 mm. long, very slightly retrorse-pilose; ovary glabrous, 3-celled, tapering into a glabrous style which is usually 3-parted, sometimes slightly so, occasionally incompletely so, topped by three stigmas. Fruit ovoid to globose, glabrous, three-celled, sometimes og aad so, ca. 1 cm. diam.; the seeds minute, brown, varying in num This is the first record of the genus Cleyera from the Philippine Islands. Described originally under Adinandra, this new variety of Cleyera japonica Thunb. has been collected profusely in many provinces of Luzon and Mindanao. .The axial placentation of the ovary, the retrorse pubescence of the anthers, the entire leaves and glabrous terminal buds, the 3-parted style and 3-celled ovary, and the very small, subrotund, conspicuously ciliolate calyx-lobes are characteristics of Cleyera rather than Adinandra. There is considerable variation in the parts of the flower but none appear consistent enough for specific delimitation. Ordinarily the calyx-lobes are approximately 3 mm. long and about as wide. In Curran & Merritt 7846 and Canicosa 30017 from Tayabas, Loher 12346 from Benguet and Cani- cosa 28980 from Laguna Province in Luzon the calyx-lobes are smaller, measuring only 2 mm. in length and width. However, in Mearns 4420 and Ramos & Edano 44906 from Mt. Pulog, Benguet Province, the calyx- lobes measure 2.7 mm. and 3.5—4 mm. respectively. There is also variation in the pedicel-length and the size of the leaf. Ordinarily, the pedicel measures 1.5—2 cm. long. In the first two specimens mentioned above, namely Curran & Merritt 7846 and Canicosa 30017, the pedicel measures only 7 mm. These two specimens are generally smaller in most characters. In Ramos & Edano 40314 from Mt. Baudan, Benguet Prov., the calyx-lobes are 2 mm. long but the pedicel measures 2 cm. Ramos & Edano 44906 from Mt. Pulog has pedicels measuring 2 cm. in length. 46 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII In the original description, no mention is made of the three- parted style and the fact that more than one flower is usually found in the axil. The degree of splitting in the style is variable. Entire styles and three-parted styles in varied stages may be found on the same specimen. However, the tendency to split is evident in all specimens. Also, the presence of more than one flower (up to five) is more prevalent than the solitary flower in the axil. These are all characters of typical Cleyera. VERNACULAR NAMES: Abu-abu, conn Ternstroemia urdanetensis (Elmer), comb. ~~ urdanetensis Elmer, Leafl. Philipp. Bot. 8: 2837, 1915 “urdanetense.” 192: » rill, Enum. Philipp. Fl. Plts. 3: 73. i aes epiphytica Elmer ex Merrill, Baan Philipp. Fl. Plts. 3: 73. 1923, in i cceemeniN: Mindanao, Luzon. Minpanao: Agusan: Cabadbaran, Urdaneta, A. D. E. Elmer 14078 (isotypes of A. urdanetensis, A, C, G, IS), Oct. Luzon: Prov. Sorsogon: Trosin, Mt. Bulusan, A. D. E. ome 15851 (A, ’, US), 17317 (A, NY, US), April & Sept. 1916.— Bulusan Volcano, M. Ramos oy . 23671 (A, C, NY, US), Sept. 1915 This species of Ternstroemia was first described as Adinandra urdane- tensis by Elmer. Later, Elmer gave it the herbarium name 7. epiphytica, which name Merrill, probably not realizing the relationship with 4. urdanetensis, published in his Enumeration INDO-MALAYSIAN SPECIES In this group are classified the species found in East Bengal, Ceylon, Siam, the Malay Peninsula, the Netherlands East Indies, Sarawak, and British North Borneo. Including the thirteen new species proposed here, there have been fifty-two species described under Adinandra which fall in this geographical group. Of these, twenty-four are recognized below, two are excluded from the genus, two are recorded as doubtful species and three are listed as little known It is with regret that I list species as “doubtful” or “little known.” Material of these species can surely be found in the herbaria of Malaya and the Netherlands East Indies, perhaps in abundance — but at present, under the strict shipping conditions in the Pacific area, the material is not available for this study. As a result of the war some of the material may never be available. Because of the lack of this material, my interpretations of some of the species of this group may be open to question. In too many instances it has been necessary to draw conclusions from a single specimen, often incomplete at that. In other cases decisions have been reached and the original iat ln only recorded when no material has been available at s information has been gathered, often very laboriously, from eet which appears annoyingly incomplete according to our present standards and which could apply easily to almost any species previously described. However precarious as it may be, I have attempted to bring 1947 ] KOBUSKI, STUDIES IN THE THEACEAE, XV 47 together the species of this geographical group which are presented below. In this geographical area is found the true center of distribution of the gen Species are much more varied in form than those of the Philippine Islands and Eastern Asia. Also, the number of species is considerably greater. As in the two former groups, I am recording a chart, compiled during this study, which brings together the characters most used in identification, and under these headings have listed the respective species. CHART FOR INDO-MALAYSIAN SPECIES OF ADINANDRA _Bracteoles Calyx-lobes PERSISTENT CADUCOUS PUBESCENT GLABROUS acuminata borneensis acuminata angulata angulata coarctata borneensis Brefeldii Brefeldii dasyantha caudatifolia celebica caudatifolia Griffith Clemensiae collina celebi vanic rctata dumosa Clemensiae lutescens colombonensis excelsa colombonensis myrioneura cordifolia Griffithii ina oblonga ordifolia impressa cordifolia phlebophylla strigosa polyneura cordifolia plagiobasis Corneriana quinque partita strigosa villosa dasyantha Sarosanther Corneriana Hullettii verrucosa dumosa integerrima excelsa avanica Hullettii lasiopetala pressa lutescens integerrima maculosa lasio petala magniflora maculosa yrioneura magniflora nunkokensis nunkokensis onga parvifolia parvifolia polyneur plagiobasis quinque partita phlebophylla Sarosanthera subsessilis subsessilis sa verrucosa Corolla-lobes Filaments PUBESCENT GLABROUS PUBESCENT GLABROUS acuminata angulata acuminata jeune borneensis Clemensiae borneensis dasyantha orneriana collina Clemensiae excelsa dasyantha colombonensis collina impressa 48 JOURNAL OF THE ARNOLD ARBORETUM Corolla. lobes (Continued) — Filaments (Continued) [voL. xxv PUBESCENT GLABROUS PUBESCENT GLABROUS ‘eteeasvind cordifolia colombonensis javanica javanic cordifolia cordifolia maculosa lasio petala stri cordifolia oblonga lutescens dumosa trig phlebophylla myrioneura excelsa Corneriana Sarosanthera nunkokensis riffit umosa oblonga Hullettii Hullettii parvifolia im pressa integerrima plagiobasi maculosa lasio petal phlebophylla agniflora lutescens olyneura se ahi magniflora subsessilis Sarosanthe yrioneura villosa verrucosa nunkokensis parvifolia plagiobasis polyneu sie dia subse verrucosa villosa Ovary Style PUBESCENT GLABROUS PUBESCENT GLABROUS acuminata aavciats acuminata angulata borneensis — caudatifolia borneensis caudatifolia celebic Clemensiae refeldii Clemen pCi CE coarctat celebica coarctata dumosa collina colombonensis collin excelsa orneriana pranerns cordifolia Griffithit integerrima cordifolia cordifolia press javanic strigosa trigosa lasiopetala lutescens dasyantha Corneriana polyneura agniflora dumosa dasyantha quinque partita nunkokensis excelsa Hullettii Sarosanthera oblonga iffit hii integerrima verrucosa phlebophylla Hullettii vanica villosa press lutescens lasio petala maculosa maculosa agniflora myrioneura myrioneura plagiobasis pre eal polyneura quinque partita diet etoils Sarosanthera parvifolia subsessilis plagiobasis verrucosa 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 49 Ovary FIVE-CELLED FOUR-CELLED THREE-CELLED TWO-CELLED borneensis myrioneura acuminata nunkokensis celebica oblonga angulata subsessilis Clemensiae caudatifolia coarctata Corneriana escen cordifolia Sarosanthera neura aia ver villosa KEY TO THE INDO-MALAYSIAN SPECIES A. Ovary and fruit glabrous 3- or 5-parted: style 3- or 5-parted or deeply sulcate. C. Ovary and fruit 5-celled; stigma and style 5-parted or sulcate. D. Bracteoles caducous or vestigial. (Bengal)........... 28. A. Griffithii. 917 cm. long, 4.5-7.5 cm. wide ; veins o mature leaves pressed above; anthers glabrous except for rare “rue setae. British North Borneo: Mt. Kinabalu)........... 30. A. impressa. CC. Ovary and fruit 3-celled; stigma and style 3-parted. See nm Serre BB. Soe rae style ent C. Ovary and fruit 5 were ovules many; seeds small (2-3 mm. long), many pa corte g. D. Terminal buds glabrous. E. Pedicels slendet, up to mm. long; leaves with midrib smooth Pe the blade not Ae -dotted (Malay slant East 1b Co Vc) arneeaP nee per eneicy iceye er aricre te tion ocr ca he ee ee . A. dumosa. EE. nary short and thick, 5 mm. long and ca. 3 mm. thick; leaves with midrib bisulcate beneath, es blade dark i sara -dotte d. (British North Borneo: Mt. Kinabalu)......... . A. verrucosa. DD. ba oe buds appressed-pubescent . Corolla-lobes glabrous; veins conspicuous up to 12 pairs, arcuate- ascending near the margin. F. ma lobes pubescent; filaments pubescent; leaves 6-10 cm. , 3-4 cm. wide, the base 2 ee North Borneo: Me ‘inabalt) Reng aT ae RCE Caters er ohe . A. colombonensis. 50 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII FF, — glabrous; filaments glabrous; leaves 9-15 cm. long, wide, the base rounded or obtuse. (British North 35. A “< . EE. Corolla lobes pubescent; veins conspicuous, more than 20 pairs, becoming evanescent towards the margin, not arcuate- ascending. 36. A. polyneura. CC. Ovary and fruit 3-celled; ovules few in number; seeds large (6-8 mm long), few (6-8) developing paenl heavily coriaceous, the veins conspicuously ashes on both surfaces with equally conspicuous reticulations; bracteoles minute, deltoid, alae branches always winged. (Pahang)..........--. sag taog ut tits higgtiee ewe 4 se'aa.4'o) RTS SL ak ne Re an ee Re ee 37. A. angulata. DD. — coriaceous, the veins inconspicuous above, more pronounced bene without obvious reticulations; bracteoles small, rounded at the apex ; ne branchlets usually terete, rarely winged or poi the branches always terete. (Malay Peninsula; Sumatra, Java, Borneo)... een wd Peg oe Dura wate u athena’: ao nap haa ee cen ase ee eee pare AA. Ovary pubescent; fruit pubescent or glabrescent, usually with a few hairs at the apex. B. Bracteoles persistent through fruiting stage. aves cordate at the base. ature leaves glabrous on the lower surface. pcan British North f 4. oe dee 0088 OS 88 ee Oke a Se ee Oe OM BOrmne@) 4c5 606 ok 4 och avadeere.s Sne Wie Wee HOF OS eee ninw a wae A. cordifolia. DD. ean leaves scabrous-pubescent on the ae surface. ae h North orneo: Mt. Kinabalu)............. A. cordifolia var. strigosa. CC. oe cuneate, obtuse or rounded at the ae not cordate D. Style glabrous E. Corolla pubes ent F. Leaves abel the petiole 1-2 mm, long; pedicel very short, ~2 mm. long; ovary 2-celled, pauci-ovulate; fruit with few (ca. 2) seed developing, seeds large up to 8 mm. long. (Sarawak).. sigeig dildo anda eae at 4G kas Cae eaE Re ea 40. A. subetssilis, FF. Leaves petiolate, the petiole 5-7 mm. long; pedicel ca. mm long; ovary 5-celled, multi-ovulate; fruit with many (over 100) seeds developing, seeds small (2-3 mm.). —— Peninsula).. cise A Slab ghd ohne date los up eG earn ane aay wong a eee Ras, ea . A; parvifolia. EE. Corolla glabrous. 3-celled, pauci- ins fruit 3-celled, few (10 or less) “gage the seeds large, up 8 mm. long. (Malay Peninsula; Sumatra, Java, Borneo) .: 23035. sesce cee ss 38. A. Sarosanthera. FF. Ovary 5-celled, multi-ovulate; fruit 5-celled, many (over 100) seeded, the seeds small (2-3 mm. long). G. Bracteoles densely pubescent, broadly ovate or deltoid, un- qual (5S x 5 mm. and 3 3 mm.); branchlets rusty-tomen- aeee terminal buds densely tawny-sericeous. (Malay Mla oa vices a sate eate- o's be eee ae ROS 42. A. Hullettii. GG. eral appressed-pubescent, subrotund, unequal (2 and 3 mm. long); branchlets appressed- agri near the apex terminal buds appressed-pubescent. (Malay Peninsula).. dy sig tence i etaentainte sated aateie ya wip ata yreiene asses ely oe are oa 3. A. aaculosi. olla pubescent. F. Ovary 3- or 2-celled, pauci-ovulate, the ovules large; fruit 3- or 2-celled, few seeded (less than 10), the seeds up to 8 mm. long. G. Pedicels short and thick, 5 mm. long, 4 mm. in diameter; ovary and fruit 2-celled. (British North rd Mt. Kina- Ball) focnee ac hee aeanpaateasrsaanes A, nunkokensis. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 51 GG. Pedicels usually over 10 mm. long, up to 20 mm.; ovary and fruit 3-celled. H. Branchlets, lower surface of leaves, pedicels, and bracteoles densely ferrugineous-sericeous; fruit yi single seed to each cell. (Malay Peninsula)......... girs . Branchlets, lower surface of the leaves ae pubescent) and other parts lightly appressed-pubescent; fruit with more than a single seed (ca. 3) to each cell. eo eo nearly flask-shaped, long-caudate at the apex; seeds dark, shining. (British North Borneo: Mt. Kinabalu). Bee ioctl Geasts wick “4 #910! 3.0) Speen “e fesaiagt hep II. Leaves coriaceous, up to 15 cm. lon cm. wide, acuminate at the apex; seeds light ba “aut (Malay Peninsula; Malacca; Sumatra)....... . A. acuminata. FF. etd 5-celled, multi-ovulate, the ere es fruit 5-celled, ny (10G or more) seeded, the seeds measuring ca. 3 mm. long. (Ma lay, Peninsula’): 0.000 se one ee eens 48. A. Seas. EE. Corolla glabrous. F. Terminal buds glabrous; midrib on the lower surface of the leaf bisulcate; calyx-lobes glabrous. (Sarawak, British North Borneo, I GENEOVcneetiselelars Siskeis: aie apis acai ueroerme es nene tenes 49. A. collina. FF. Terminal buds sericeous or appressed-pubescent; midrib on the lower surface of the leaf rounded, not bisulcate; calyx-lobes pubescen G. Small jewves, 3-5 cm, long, 2—2.5 cm. wide, occasionally larger ; bracteoles ca. 2 mm, long; corolla-lobes always five; seeds minute, ca. 2 mm. long, dark, shining. Sacre ee Borneo: Mit: Koinabalt)eee series ca claus scene Sans . Clemensiae. GG. — larger, 6-13 cm. long, 4-7 cm. | ae —. 5-6 . long, 6-8 mm. wide; corolla-lobes six in number; seeds ie 15 mm. or less, red in color. sie North Borneo). i ccs asec RRP, iS sa08 wis oie eceeoeis alee . A. magniflora. BB. Bracteoles quickly caducous, seldom lasting to onithesiee: C. Leaves obliquely cordate at base. (aaa Sacnecsarts s 52. A. plagiobasis. CC. Leaves acute or subrotund, never cordat Style glabrous E. Bracteoles persistent. (Southeast Borneo)....... 53. A. borneensis. EE. Bracteoles caducous. F. Filaments glabrous; veins up to 20 pairs, rather inconspicuous on the lower surface of the leaf, extending from the midrib at an angle: Of /6025 \USIIEALIY) ve crs sejs ee cpauecm es ones 10. 54. A. dasyantha. FF. Filaments pubescent; veins over 25 pairs, conspicuous on the lower surface of the leaf, approximately ie ree to the midrib. (British North Borneo)............. . myrioneura. DD. Style pubescent. Ovary and fruit 3-celled F. Bracteoles alternate (evident by scars when the bracteoles have fallen) with the lower bracteole as much as 5 mm calyx; apicule on the stamen not distinctive; leaves heavily corlaceous: ((lava)icwearcss te eesti le cies cc cae 56. A, javanica. FF. Bracteoles opposite (scars), immediately below calyx; apicule of he stamen 3 mm. long equalling the anther in length; leaves subcoriaceous: “(Sidi Aneaeeaereis o 0s. «Ses oe 57. A. lutescens. EE. Ovary and fruit 5-celled (or 4-celled in A. oblo F. Stigma 4-parted, oblique; ovary and fruit 4- celled. rapes ee oe Cn ro. Ce i ee 58. A. cones. 52 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII FF. Stigma entire; ovary and fruit 5-celled. G. Pedicels very short, ca. 5 mm. long. (Siam).............. sain a thgpaiai ung Rusate.4 Pienaar 4 ah owes aie pee Acer atee 59. GG. Pedicels 10-50 mm. long. H. Filaments glabrous; calyx-lobes ss acute at the apex. (Indo-China)..............-. 60. A. phlebophylla. HH. Filaments sericeous; sens lobes mace - the apex. (Malay Peninsula; Burma)..............-. A, villosa. 28. Adinandra Griffithii Dyer in Hooker f., Fl. Brit. Ind. 1: 282. 1874. — Szyszylowicz Pflanzenfam. III. 6: 189. 1893.— Melchior in Nat. Pflanzenfam, ed. 2, 21: 144, 1925 DistrRIBUTION: East Bengal. East BENGAL: W. Griffith 755 (asoTypE, G). Tree 15-20 m. with slightly angled branchlets, the branchlets een even to the younger growth, the terminal buds gla brous. Leaves coria- ceous, oblong-elliptic to oblong-oblanceolate, 12-15 cm. long, 4-5 cm. wide, acuminate at the apex, tapering at the base, glabrous on both surfaces, even in the very young leaves, the margin quite entire, the veins rather incon- spicuous, the petioles sturdy, glabrous, semi-terete, flattened above. Flowers axillary, solitary; pedicels ca. 2 cm. long, glabrous: ener 2; caducous or vestigial; calyx-lobes 5, imbricate, glabrous, rounded, m. long; corolla-lobes (fide Dyer) equalling the calyx-lobes in ain stamens (fide Dyer) sparingly setose, 1-seriate, adnate to the base of the corolla; ovary glabrous, tapering into the style, the style glabrous, distinctly 5-fid, the divisions of the style cylindrical, hardly exceeding the calyx-lobes. Fruit not seen. The above description of this species is not as complete as one might desire. Although an isotype of the species has been available for this study, little additional information can be given at present. The outstanding feature of the species is the deeply five-parted style. In the majority of species of Adinandra the style is entire for the whole length. The only other known exception is A. lasiopetala (Wight) Choisy in which the style is three-parted. In A. impressa Kob. and A. quinque- partita Kob. dissections made from buds show the style to be five-sulcate. The ovary of A. Griffithii is undoubtedly five-celled, although not men- tioned as such by Dyer. Also one may assume that the corolla-lobes are glabrous on the external surface, because, were they pubescent, Dyer would have drawn attention to the fact, since he does describe the corolla. Dyer states that the bracteoles are vestigial. Observing them, I cannot be certain whether they are vestigial or caducous. I am inclined toward the latter. A specimen deposited in the U. S. National Herbarium at Washington may be mentioned here. This specimen, collected by Dr. King’s Collector, no. 114, is incomplete. However, in an accompanying packet is material from previous dissections of buds showing the style definitely five-parted. The leaves are glabrous, as in the terminal bud, and the branchlets are slightly angled. The leaves are larger and less coriaceous than those of the type, measuring up to 20 cm. in length and 6.5 cm. in width. This additional specimen collected in the “‘Akha Hills” is recorded from 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 53 Duffla Hill as a “tree 50 to 60 ft. high, 3 to 4 ft. in dt.; leaves green, flowers white.” It was collected in Feb. 1890. The actual collector’s name is not clear but appears to be “Khan.” This locality, Akla Hills, or Aka Hills, as it is recorded in some gazetteers, is on the northeastern frontier of India, northeast of the Darrang District of Eastern Bengal. Duffla Hill is also recorded as Daffla Hill. Aka and Daffla are names of independent tribes. This locality agrees very well with that of the type, Griffith 755, which was collected in East Bengal 29. Adinandra quinquepartita, sp. nov. Arbor parva; ramis glabris, teretibus, flavo-brunneis, subrugulosis, ramulis glabris, teretibus brunneis, gemmis terminalibus conicis, glabris. Folia crasso-coriacea, elliptico- -oblonga vel ovata, 9-17 cm. longa, 4.5—7.5 cm, lata, apice acuta vel late acuminata, minute PS ne late cuneata vel subrotundata, undique glabra, subtus fusco- -punctata, margine subintegra vel minute glandulosa, venis lateralibus primariis 8-10 paribus, marginem versus inconspicuis, venis secundariis inter primarias frequenti- bus, petiolis crassis, glabris, 6-10 mm. longis, ad 4 mm. diametro. Flores axillares, solitarii; pedicellis 2-3.5 cm. longis, oe recurvatis: bracteolis ny oppositis, persistentibus, glabris, ca. 3 mm. longis, margine ciliolatis; sepalis 8, imbricatis, glabris, late ovatis vel subrotundatis, ca. 8 mm. longis, margine scariosis; petalis glabris; staminibus in numero Si. ?-seriatis, filamentis exteriore leviter pubescentibus, antheris oblongis, leviter hirsutis; ovario glabro, conico, 5-loculari, multi-ovulato; stylo insigno, glabro, apice leviter 5-sulcato; stigmatibus 5-partitis. Fructus ignotus. DistrIBUTION: British North ets British North Borneo: Mt. nabalu, Gurulau spur, in an open place, alt. 2100-2700 m., J. & M.S. Clemens oe (TYPE, A), Dec. 6, 1933 (small tree; flowers white, the stigma 5- parted). Unfortunately, complete material was unavailable for the study of this new species. In fact the material is so incomplete that, were it not for the unusually distinctive character of the pistil, I would have refrained from describing it as new The ovary is 5-celled, glabrous, and after swelling at the base tapers conically into an ill-defined style similar to those found in species of Gordonia. The style in the bud, which I examined, was undivided. However, the collectors, J. & M.S. Clemens, draw attention to the five divisions of the stigma and it is possible that, like A. Griffithii of India and A, lasiopetala of Ceylon the style in open flower and fruit may be divided. The stamens, although no accurate count could be obtained, were typically those of Adinandra in number and character. The anthers were oblong, elongated, and bore the characteristic pubescence of the genus. The filaments also were lightly pubescent, in series, and joined for part of their length. The corolla-lobes, very tightly folded in the bud, could not be separated. However, they were glabrous and of a more thickened nature than those found in Gordonia. The ovules were attached to a slight placenta 34 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII The specific name A. al refers to the 5-parted stigma. 30, Adinandra impressa, sp. Arbor ?; ramis griseis, : snes glabris, ramulis teretibus brunneis, Folia coriacea, glabra, oblongo-elliptica vel late elliptica, 4-7 cm. lon 2.5-4 cm. lata, apice late acuminata, basi obtusa vel late aa margine integra, supra nitida, subtus pallidiora, fusco-punctata, venis supra impressis, subtus subobscuris, primariis 5-8 paribus, intra marginem anastomosantibus arcuantibusque, venis secundariis inter primarias fre- quentibus, petiolis circiter 5 mm. longis, glabris. Flores axillares, solitarii; pedicellis glabris, teretibus, 1.5—2.5 cm. longis; bracteolis 2, ‘oppositis, deltoideis, 4 mm. aoe 6 mm. latis, ico ie tribus’ subrotundatis, circiter 5 mm. longis, 6 mm. latis; petalis 5, basi connatis, glabris, purpureis, circiter 12 mm. longis, apice acutis; staminibus circiter 25, 3-seriatis, m. longis, antheris 2—3 ongis, glabris, rarissime setis isolatis, ovario glabro, semi-globoso, circiter 4 mm. diametro, 5-loculari, multi-ovulato, stylo glabro, circiter 5 mm. longo, apice 5-sulcato, stigmatibus 5. Fructus elongato-globosus, ie. circiter 12 mm. longus, 8 mm. latus, 5-locularis, multi-seminatus, seminibus minutis, 1 mm. vel minus diametro, rubris. DistripuTion: British North Borneo. British Norta Borneo: Upper Mt. Kinabalu, alt. 2000-5000 m., J. & M. S. Clemens 30202 (typ, A), spr 1932 (flowers dark with purple petals and green calyx). — Upper Mt. Kinabalu, J. M. S. Clemens 29073 (A), April 8, 1932 (flower-buds dark, the petals purple, the ee green) quinquepartita Kob. this species has distinctly five stigmas. The style near the apex is sulcate with five distinct ridges. Dissections were made from the bud. It is possible that in maturing the style may become five-parted. The corolla is dark purple in color, and in the bud it is very elongate-conical, extending considerably beyond the calyx. The stamens, except for a rare evidence of pubescence on the anthers, are glabrous aaah he leaves are thick-coriaceous with impressed nerves on the upper surface. The terminal buds are generally glabrous. However, there are rare instances of pubescence on the buds and very young branchlets. In glabrosity this species rivals A. dumosa Jack. The mature seeds are minute, measuring a single millimeter or less in diameter, and are red in color rather than the traditional dark brown. wera lasiopetala eer oe in Mém. Soc. Phys. Hist. Nat. — Mem. Ternstr. 24). 5.— Dyer in Hooker f., Fl. Brit. Ind. 1: ai — Beddome, FI. Sylv fs planed Man. Bot. xxiv, t. iii, f. 3). ae - Szyszylowicz in Nat. Pflanzenfam. III. 6: 189. 1893. — Trimen, egies Fl. Ceyl. 1: 108, pl. 9. 1893. — Melchior in Nat. oo ed, 2, 21: 14 Cleyera aes Wight, Ill. Ind. Bot. 1: 99. Eurya lasiopetala (Wight) Gardner in Calcutta ae Nat. Hist. o 1847. Sarosanthera pean (Wight) Thwaites, Enum. Pl. Zeyl. 41. DistRIBUTION: Ceylo Ceyton: G.H. K. hous 775 (G). 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 55 Small tree ca. 10 m. high; branchlets terete, grayish brown, glabrous except when very young. Leaves coriaceous, oblanceolate, glabrous except when very young, 7-9 cm. long, 2—3 cm. wide, the margin serrulate, con- spicuously recurved especially along the lower ‘half of the leaf, the petiole semi-terete, flat on the upper surface, ca. 5 mm. long. Flowers axillary, solitary ; pedicels somewhat recurved, ca. 1.5 cm. long, appressed-pubescent, thickened towards the apex; bracteoles 2, unequal, opposite, immediately beneath the calyx, persistent, somewhat rotund, 3-5 mm. long, finely appressed-pubescent; calyx-lobes 5, yoneioniis: thickened, unequal, some- what rotund, ca. 8 mm. long, the outer lobes finely appressed-pubescent, the inner lobes sericeous on the el portion of the outer surface, the Trimen et al.) monadelphous, 18-25, filaments unequal, sparsely hairy, joined at the base, the anthers sparsely pubescent; ovary 3-celled, glabrous, attenuate at the apex into a filiform, glabrous style which, on the maturing of the fruit, becomes 3-parted, the stigma 3-parted. Fruit globose, 3-celled, fide auct., many-, small-seeded; seeds hippocrepiform. This species is exceptional " the genus in having a 3-parted style (in the maturing fruit), and small seeds. These characteristics are similar to those found in Cleyera. The persistent bracteoles, the single, axillary flowers and the pubescent filaments separate it, on the other hand, from Cleyera. The only specimen available for this study is Thwaites (C. P. 775), quite an authentic specimen since it was used by Thwaites in his treatment. It agrees perfectly, as far as the available characters are concerned, with the illustrations of Trimen. The stamens were not seen by the present author, but the statement by Dyer that they are unequal leads one to believe that some form of seriate arrangement exists. Only in a single series are the stamens usually of equal length. 32. Be ge dumosa Jack in Malay -Misc. 2 (7): 50. 1822 [repr. in Hooker, Comp. 53. 1835; in Calcutta Jour. Nat. Hist. 4: 207. 1843; et in Miscel. hi Sei China, II. 2: 271. 1887].— Choisy in Zollinger, Syst. Verz. Ind. Archip. 143. 1854; in Mém. Soc. Phys. Hist. Nat. Genéve, 1: 112 (Mém. Ternstr. 24). 1855. — Miquel, Fl. Ned. Ind. 1 (2): 477. 1859; Ann. Mus. Bot. Lugd.-Bat. 4 103. 1868. — Dyer in Hooker f., Fl. Brit. Ind. 1: 282. 1874. — King in Jour. As. Soc. Bengal, 59 (2): 188 (Mater. Fl. Malay Penin. 1: 128). 1890. — Szyszylowicz in Nat. Pflanzenfam. III. 6: 189. erp RCE IIa & Valeton, Meded. ’S Lands. Pl. 16: 223 (Bijdr. Boomsoort Java 3: 223). 6. — Merrill, Enum. Born. Pts. 391. 1921. — ere & Holttum in Gard. fin ae Settl. 3: 36. 1923. — Mel- Pflanzenfam. ed. 2, 21: 144. Adinandra a ae Korthals, Verh. Nat. Gesch. Bot: ed. Temminck, 106. 1840. Adinandra trichocoryvna Korthals, Verh. Nat. Gesch. Bot. ed. Temminck, 107. 1840. — Miquel, Fl. Ned. Ind. 1 (2): 477. a Ann. Mus. Bot. Lugd.-Bat. 4: 103. 1868. — Merrill, Enum. Born. Plts. 391. Adinandra cyrtopoda Miquel, Fl. Ned. ce scat 1: 478. Pe Adinandra stylosa Miquel, Fl. Ned. Ind. at 1: 478. as Adinandra glabra Miquel, Fl. Ned. Ind. Sup 79. Adinandra trichocoryna Korthals @ parvo: Miquel, | Mus. Bot. Lugd.-Bat. 4: 103. 1868. Adinandra trichocorvna Korthals 8 glabra Miquel, Ann. Mus. Bot. Lugd.-Bat. 4: 103. 1868 56 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII ee trichocoryna Korthals Y stylosa Miquel, Ann. Mus. Bot. Lugd.-Bat. 4: . 1868 Adinonira trichocoryn a Korthals 5 cyrtopoda Miquel, Ann. Mus. Bot. Lugd.-Bat. 1868 DISTRIBUTION: "Malay Peninsula, Netherlands East Indies, Sarawak, British North Borneo Matay PENINSULA: Singapore, Bot. Garden, J. & M. S. Clemens 22448 (NY), Nov. 7, 1929 (tree 10 m. high; flowers white). Rinses Bot. Gard. [Cult.], lawn 1. M. Nur 3410 (A), Oct. 12, 1925, 24975 (A), May 17, 1930.— Singapore, H. M. Curran s. n. (US), Aug. 1910; O. Kuntze 6108 (NY), Oct. | — Malacca, W. Grif- fith n. (G, M).— Perak, B. Scortechini s. n. (US). — Perak, near Ulu Sdangors, open oe forest, hilly locale, alt. 150-200 m.,.Dr. King’s pose pee 8698 (A), March 1886 (tree 15 m. high with spreading branches ; leaves dark metallic green; flowers white). — Penang, Penang Hill, G. Aing s. n. (US), Aug. 1879. — Pahang, Tasek Bera, at low altitude, M. R. Henderson pee ae Oct. 10, 1930. — Exact locality lacking, W. Griffith 756 (G); A. C. Maingay 184 (G uMaTRA: East Coast: Asahan, sein Forest Reserve, in primary forests, on “red” soil, B. A. Krukoff 4255 (A, M, NY, US), Nov. 3, 1932 20 m. high). — Asahan, between a and Pargambiran, alt. 270-360 m., H. H. Bartlett 6696 (C, NY, US), Feb. 11, 1927. — Koealoe, old jungle near the Ak oe Loendoet Concemian, H. H. og. 6989 (C, NY, US), March 19, 1927.— Asahan, between iPrapiae and Pargambiran, alt. 270-360 m., H. H. Bartlett 8061 (NY, US), May 22, Asahan, Bandar ota: H.S. Yates 1688 (A), 1917 (NY).— East southeast - Toba Lake, J. A. Lorzing 6563 (A), May 11, 1919. — Bilah, general region of Morbau, near Bilah PD chawn (Parbasiram), Rahmat Si Toroes 389 (NY), Feb. -March 1928. — Koealoe, Si Toenggir, 9 km. from Damoeli, Rahmat Si Toroes 1394 (NY), Sept. 20, 1928. ae sae vicinity of pees Parapat, Phar aiead Si Toroes 1577 (A), Rantou a. ‘Rahmat “s Torore:2 ies (A, US), M ay 28-June 26, 1932 reed Bilah, Sigamata, near Rantau Parapat, Rahmat Si Toroes 3125 (A, NY), June 27-July 23, 1932. — Bilah, Subdivision Laboehan Batoe, District Kota Pinang, Lannga Pajoeng, py ee Si Toroes 3434 (NY, US), March 7-30, 1933. — Subdivision Laboehan Batoe, District Kota Pinang, Si haa Angin (on the Soengei Kanan), Rahmat Si Toroes 4058 (A, NY, US), April-May 1933.— Subdivision Laboehan Batoe, District Bilah Hitean Haloban Pane of Concession Rantau Parapat), Rahmat Si Toroes 43035 (US), May 17-24, 1933. Riouw en Onderh: Indragirischebovenlanden, Batang Peranap, alt. 75 m., Neth. Ind. For. Serv. b.b. 30092 (A), Sept. 19, 1939. Tapa- noeli: Fishineitan, Panapparan to Maranti, H. H. Bartlett 8037 (NY, US), May 16-19, 1927.— Road from East Coast to Tapanoeli, ager of Toba Lake, new road near Peso Peso, virgin jungle, alt. 1350-1500 m., W. N. & C. M. pe ee 1162 (A, NY), Feb. 22, 1932 (tree 8 m. high; flowers — — Division Padang Si Dimpoean, Subdivision Padang Lawas, Sosopan on Ack Si Olip, Rahmat Si Toroes 5518 (A, NY), Sept. 1933. West Coast: “?Sankor,” J. E. Te eysmann S. N. —‘“?Sibogne,”’ J. E. Teysmann 5. n. ). Exact locality missing WwW Korthals s. n. (probable isotype of A. Jackiana, G, NY).— H. O. Forbes s.n. (G). ava: J. E. Teysmann s. n. (probable 1sotypes of A. rhe G, NY), 1868. — Herb. Hort. Buitenzorg [Cult.] 137 (VI-c-203, 205, 205A) (U oRNEO: J. E. Teysmann s. n. (probable 1soTyPE of ty soon G).—P. W. Korthals s. n. (probable isotype of A. trichocoryna, G, N British NortH BORNE Mt. Kinabalu, Mt. Nunkok, alt. 1650 m., J. & M. S. oe 32724 (A), April 15, 1933. — Mt. Kinabalu, Colombon River, alt. 1350 m., J. & M.S. Clemens 34005 (A), July 21, 1933. — Sandakan and vicinity, D. D. Wood 810 He PBS), 887 (A, PBS), — June 1920. — Sandakan and vicinity, M. Ramos 1549 (A, G, US), Sent, -Dec Sarawak: Mount es er 200 m., J. & M.S. Clemens 22321 (A, M).— Mt. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 57 Dulit, moss-forest, alt. ca. 1000 m., P. W. Richards 1620 (A), Sept. 8, 1932 (shrub 3 high; petals white, the calyx pinkish; leaves thick and leathery). Small trees with dark brown bark; branchlets terete, glabrous even to the current year’s growth, the terminal buds glabrous, only very rarely page pubescent. Leaves coriaceous, elliptic-oblong, 6-10 cm. long, ca. m. wide, usually obtuse at the apex with an obtuse acumen, tapering at he nies into a short petiole, the mature leaves glabrous on both surfaces, very rarely the under surface of the very young leaves lightly pubescent along the midrib, the margin entire or obsoletely serrulate, the veins obscure, the petiole ca. 5 mm. long, glabrous. Flowers axillary, solitary, rarely in twos; pedicel up to 1.5 cm. long, glabrous, accrescent towards the apex; bracteoles 2, persistent, opposite, immediately boar the calyx, coriaceous, glabrous, subrotund, unequal, 3-4 mm. long, 3 mm. wide, the margin entire or slightly -fimbriate: calyx-lobes 5, ‘hoes glabrous, broadly ovate, aneqial 7-8 mm. long, 6-7 mm. wide, thickened at the center, the margin scarious, usually entire, occasionally slightly fimbriate; ae lobes 5, connate at the base for 3-5 m am., glabrous, white, obovate, up t long, 6—7 mm. wide, bluntly acuminate at the apex; stamens Ca. iG seriate, 9-12 mm. long, the filaments 5—8 mm. long, adnate to the base of the corolla, accrescent towards the apex, densely pubescent, the anthers ca. 4 mm. long, hirsute, the apicule 0.5—1 mm. long, densely pubescent; ovary glabrous, subconical, ca. 4 mm. long, 5 mm. diam., occasionally furrowed, tapering at the apex into a glabrous style, 5- celled, multi-ovulate. Fruit globose, glabrous, 5-celled, many-se Adinandra dumosa, described by ree in 1822, is the original species of the genus and possesses the widest range of distribution of any of the Indo-Malayan species. It holds the same position in the Indo-Malayan distribution that 4. Millettii holds in the Eastern Asiatic distribution. The center of distribution for the species appears to be the island of Sumatra, the range extending northward into Malaya and down through the East Indian Islands into Java and Borneo. It is probably the most nearly glabrous species in the genus. Even the terminal buds and current year’s growth of the branchlets are glabrous. Species of Adinandra are often termed glabrous when close observation shows that the terminal buds and current year’s growth of branchlets are quite pubescent. As is usual in species of this family, one will find exceptions to the general rule. Specimens collected on the east coast of Sumatra near Bilah have a tendency toward pubescence. A fine scattered spreading pubescence, not too noticeable but enough to be exceptional, may be found along the midrib on the under surface of the leaves. How- ever, the unfolding buds are still quite glabrous since the dorsal surface of the leaf is exposed while unfolding. In the pubescent bud the obvious pubescence is on the dorsal surface of the leaf, which character is quickly dissipated and entirely lost in maturity. The leaves vary considerably in size and shape. Generally they are bluntly acuminate. Occasionally one finds leaves tapering into a fine acumen. ‘The stamens are numerous with both the anthers and filaments densely pubescent. In the latter case, the pubescence is so dense that it 58 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XXVIII is difficult to separate the filaments in dissection. This dense pubescence on the filaments in so glabrous a species is surprising. The style develops considerably in fruit, rivaling, at times, the pedicel in length. The bracteoles, always persistent, are small, rounded and spreading. Occasion- ally, in the very young flower-buds, the bracteoles are lightly appressed- pubescent. In this study, I have had opportunity to examine authentic material, perhaps isotypes, of A. Jackiana, A. trichocoryna, A. cyrtopoda, A. stylosa and A. glabra, and find no suitable characters worthy of specific separation. All these names have been listed above as synonyms of A. dumosa. VERNACULAR NAMES: Kajoe aling, kajoe si mar apiapi, kajoe apiapt, apiapi, daun saribu. 33. Adinandra verrucosa Stapf in Hooker’s Icon. Pl. 23: t. 2266. 1893; in Trans. Linn. Soc. Bot. Il. 4:133 (Fl. Mt. Kinabalu). 1894.— Merrill, Enum. Born. Pits. 391. 1921. Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925 DistripuTION: British North Borneo. British Nortu Borneo: Mt. Kinabalu: Paka Cave to Lobang, M.S. Clemens 10735 (ropotypEs, A, PBS), Nov. 15, 1915 (flowers pink).— Upper Kinabalu, alt. 2000-4000 m., J. & M. S. Clemens 29113 (A), March 27, 1932 (flowers pale pink suffused with white), — Head Kinatak, along margin of open place, alt. ca. 3000 m., J. & M.S. Clemens 33121 (A), Feb. 27, 1933 (small tree 8 m. high; flowers green and pink; fruit purplish green). Tree ca. 10 m. (fide Clemens) high; branchlets terete, glabrous, sturdy, noticeably thicker than those of most other species, verrucose-lenticellate, the terminal buds glabrous except for very light pubescence at the apex, the margin ciliate, otherwise glabrous. Leaves thick-coriaceous, oblong- elliptic, 12-15 cm. long, 5-6 cm. wide, obtusely emarginate at the apex, somewhat rounded at the base, glabrous on both surfaces, dark-punctate beneath, the midrib conspicuous beneath, bisulcate (in dried specimens), the veins inconspicuous on both surfaces, the margin quite entire with only occasional sharp denticulations, the petiole thick, sturdy, glabrous, usually 7-8 mm. long, sometimes only 4 mm. long. Flowers axillary, solitary or in twos or threes; pedicel sturdy, glabrous, seldom over 5 mm. long, often as thick as long; bracteoles 2, persistent, glabrous, opposite, immediately below the calyx, rounded, 3-4 mm. long and wide; calyx-lobes 5, imbricate, glabrous, rounded, 6-8 mm. long; corolla-lobes 5, glabrous, pale rose or pink suffused with white (Clemens), broadly ovate-elliptic, contracted no- ticeably at the base, 16-20 mm. long, 10-12 mm, wide; stamens apparently 4- or even 5-seriate, 20-24, apiculate, the filaments joined nearly their entire length, varying in length from the shorter or inner series, 7—8 mm. long, to the longer or outer series, ca. 13 mm. long, densely appressed- sericeous (silvery white) pubescent on the upper half of the dorsal surface, sparsely or occasionally pubescent on the lower half, the anthers oblong, uniformly ca. 5 mm. long, densely sericeous on the dorsal surface; ovary somewhat ovoid, ca. 4-5 mm. diam., glabrous, 5-celled, multi-ovulate, tapering rather abruptly into a thickened glabrous style, ca. 1.5 cm. long. Fruit globose, glabrous, ca. 1 cm. or more diam. Seeds many, minute, shining subreniform. From the specimens examined and from the literature concerning A. 1947 | KOBUSKI, STUDIES IN THE THEACEAE, XV 59 verrucosa, it appears that this species is confined to the ‘‘Upper Mountain- Zone” of Mt. Kinabalu, aie oo. According to Stapf the altitude range of this zone is «6000- 000” fee Except for the Shiai on ve fs this species is entirely glabrous. The sturdy character of all parts of the plant is the first feature which impresses one. However, the verrucose-lenticellate character of the young branchlets, the very short and thick pedicels, the bisulcate midrib (lower leaf surface) and the thick leaves and petioles are other features which tend to distinguish the species from most of the other near relatives. 34. Adinandra colombonensis, sp. nov. Arbor ?; ramis griseis vel griseo-brunneis, glabris, teretibus, ramulis brunneis, glabris, teretibus, innovationibus ago astra Folia subcoriacea vel chartacea, elliptica, 6-10 cm. longa, 3-4 cm. lata, apice gradatim attenuata vel rare abrupte acuminata, ae gradatim cuneata, supra glabra, subtus glabra vel leviter adpresso- pubescentia, margine integra vel undulata, venis supra obscuris, subtus primariis 8-10 paribus juxta marginem arcuantibus, venis secundariis inter primarias frequentibus, petiolis 2-3 mm. longis, glabris vel leviter pubescentibus. Flores axillares, solitarii; pedicellis teretibus, 2 cm. longis, adpresso-pubescentibus; bracteo- lis 2, oppositis, persistentibus, leviter adpresso-pubescentibus, subdeltoideis, circiter 1.5 mm, longis; sepalis 5, imbricatis, adpresso- Pubescentibus, semen ciliolatis, exterioribus duobus late ovatis circiter 5 mm 6 m. latis, interioribus tribus ovatis, 5-7 mm. longis, 4-5 mm, latis, cea scarlosis; petalis 5, albidis, glabris, basi 5 mm. connatis, circiter 15 mm. longis, 5-6 mm, latis, apice subrotundatis; staminibus circiter 40, 8-13 mm. longis, 4-seriatis, filamentis hirsutis, liberis, basi ad corollam adnatis, 5-9 mm. longis, antheris brunneis, elongatis , 3-5 mm. longis, hirsutis, ovario glabro, subgloboso, 5-sulcato, 5- eats ’ multi- ovulato, stylo glabro, circiter 12 mm. longo. Fructus non visi. Dismasetione. British North Borneo. BritisH NortH Borneo: Mt. Kinabalu, Colombon grr alt. 1500 m., J. & M Clemens 34237 (TypE, A), Aug. 2, 1933 ioe white with brown suber’. ag abalu, Colombon River, alt. 2000 m., J. & M.S. jae 33770 (A), June 28, 1933 (flower-buds cream-colored). This species is closely related to A. caudatifolia Kob., but is separated by the 5-celled ovary and many ovules. | It is safe to ee that in fruit A. colombonensis will produce many seeds. This assumption can be drawn from the size and number of ovules in the ovary. 35. ved excelsa Korthals, Verh. Nat. Gesch. Bot. ed. Temminck, 109. 1840. — Miqu Fl. Ned. Ind. 1 (2):477. 1859; Ann. Mus. Bot. Lugd.-Bat. 4: 103. 1868. a eh a in Nat. Pflanzenfam. III. 6: 189. 1893. — Stapf in Trans. Linn. Soc. London 4: 133 (Fl. Mt. Kinabalu). 1894. — Merrill, Enum, Born. Plts. 391. 1921. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. DistriBuTION: Borneo, British North Borneo. EO: Exact ay lacking, P. W. Korthals s. n. (probable isotype, G, NY). ee NortH Borneo: Mt. Kalawat, trail, M.S. Clemens 11168 (P.BS.), Dec. 10-11, 1915 (small tree with white flowers ad green fruit). Tree ca. 10 m. high; branches and branchlets brownish, terete, glabrous, the terminal buds elongate-conical, densely sericeous- pubescent. Leaves 60 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII coriaceous, oblong-ovate, 9-15 cm. long, 4-6.5 cm. wide, somewhat acute or shortly oblong- acuminate at the res rounded or obtuse at the base, cm. long. Flowers axillary, solitary; pedicels thin, 2-3 cm. long, glabrous, swelling noticeably above the longer bracteole into the calyx; bracteoles 2, persistent, alternate below the calyx, unequal, deltoid to subrotund, minute, 1.5-2 mm. long, glabrous; calyx-lobes 5, imbricate, coriaceous, subequal, subrotund, ca. 7 mm. long and wide, glabrous; corolla- lobes 5, ‘connate at the base, oblong, obtuse at the apex, fleshy, glabrous, ca. 12 mm. long, 6 mm. wide, white; stamens many, several-seriate, the filaments linear, glabrous, joined at the base, the anthers oblong, densely hirsute; ovary glabrous, somewhat conical, % celled, multi-ovulate, tapering at the apex into a glabrous style. Fruit globose, glabrous, 5- celled, many-seeded, attenuate at the apex into a persistent style According to Stapf, this species may be found in Sumatra. Only material from Borneo was available for study and this material contained little more than four probable isotypes. Adinandra excelsa appears very closely allied to the type-species A, dumosa, but can be distinguished by the sericeous buds, the alternate bracteoles, glabrous and subdeltoid, and the glabrous filaments. Distinguishing characters are: (1) the glabrous ovary and fruit; (2) leaves glabrous on both surfaces but sericeous in bud; (3) glabrous calyx- lobes and corolla-lobes; (4) glabrous, deltoid, alternate persistent bracteoles; and (5) glabrous filaments. Koorders describes the ovary as single-celled. However, dissections have shown it to be distinctly five-celled and multi-ovulate. The fruit is clearly five-celled also. 36. Adinandra polyneura, sp. nov. rbor ?; ramis teretibus, glabris, griseo-brunneis; ramulis teretibus, glabris, rubro-brunneis; gem mis terminalibus elongatis, luteo- adpresso- pubescentibus. Folia coriacea, oblongo-elliptica vel oblongo-obovata, apice gracili-acuminata, basi attenuata, supra glabra, subtus glabra vel glabres- centia, rariter leviter adpresso- pubescentia, fusco-punctata, margine subin- tegra vel crenulata, venis lateralibus primariis 20+ paribus, ad marginem non arcuantibus sed reticulatis, petiolis glabris, ut videtur circiter 1 cm. longis. Flores axillares, solitarii; pedicellis glabris, 1-1.5 cm. longis ni ae bracteolis 2, gegen. suboppositis, subdeltoideis vel late ovatis, glabris, circiter 3 mm. longis, 4 mm. latis, costatis; sepalis 3; imbricatis, glabris, subaequalibus, subrotundatis vel me ovatis, 8-9 m longis et tats margine cilidlatis vel glandulosis; petalis 5, basi nei oblongo-obovatis, 17-20 mm. longis, 9-10 mm. latis, apice rotundatis, exteriore adpresso-pubescentibus, margine scariosis; staminibus 30-35, 3-seriatis, 10-14 mm. longis, filamentis 5-8 mm. longis, basi connatis, ad corollam adnatis, dense hirsutis, antheris clongatis, 5—6 mm. longis, dense hirsutis; ovario subgloboso, glabro, 5-6 mm. diametro, 5-loculari, multi- ovulato, stylo glabro, circiter 13 mm. longo, basi circiter 3 mm. diametro. DISTRIBUTION: Netherlands East Indies (Sumatra). 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 61 ATRA: West Coast, Pajacombo Oeloe Air, alt. ca. 1100 m., Netherlands For. Expt. Station b.b. 5173 (type, K, fragment, A), July 7, 1922. The leaves of this new species resemble very much those of A. myrioneura Kob., and A. borneensis Kob. both of Borneo, in the large number of veins. In A. myrioneura, however, the corolla is glabrous, the bracteoles caducous, and the calyx, ovary and style pubescent. In A. polyneura the corolla is pubescent, the bracteoles persistent, and the calyx, ovary and style glabrous. In A. borneensis the bracteoles are caducous and the calyx is pubescent. The petioles of the leaves of A. polyneura appear to be rather cutstand- ing in that they seem to measure over a centimeter in length. This apparent length is caused by the gradual tapering of the blade along the petiole in a slight wing-formation. ‘The actual petiole hardly exceeds in length that of the majority of species in this genus Another interesting feature is the slight keel on the dorsal surface of the bracteole. Also, the pedicel appears 2-ridged, caused by the decurrence of the bracteole to the base, very much as a stem with decurrent leaves. VERNACULAR NAME: Madang limbe. Bi pra Sea Ridley in Jour. Linn. Soc. Bot. 38: 304. ne i. aes Penin. 1: 195. 1922. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 14 DISTRIBUTION: Malay Peninsula (Pahang). Pentnsuta: Pahang: Gunong Tahan, alt. 1700-1800 m., L. Wray & Hi. rie 5518 (TYPE, K), July 12, 1905 (tree 12-16 m. high). Tree 12-18 m. high; branches rough, sturdy, grayish brown, glabrous, 4-angled, lightly winged, the young branchlets conspicuously winged, gla- brous, thick, as much as 5 mm. in diameter, the terminal buds sericeous. Leaves thick-coriaceous, nee 10-15 cm. long, 5—7 cm. wide, obtusely acuminate to subrotund at the apex, acute at the base, glandular-apiculate along the margin from ce base to the apex, less conspicuously so near the base, glabrous on both surfaces, the veins conspicuous on both surfaces even to the fine reticulations, 8— 9 pairs arcuate-ascending near the margin, interspersed with an equal number of secondary veins, nearly as conspic- uous but not as long, the petiole thick, glabrous, ca. 1 cm. long, ca. 4 mm. in diameter. Flowers in pairs, presumably axillary (none attached) ; pedicels in pairs, sturdy, up to 5 cm. long, glabrous, flattened, ca. 1 mm. in diameter at the base, accrescent to 5 mm. in diameter at the apex, the narrow edges terminating in the bracteoles; bracteoles 2, small, almost vestigial, persistent, unequal, subdeltoid, subopposite, the lower bracteole 1+ mm. long, the upper bracteole 2+ mm. long; calyx-lobes 5, imbricate, rounded, glabrous, ca. 7 mm. long, 8 mm. wide; corolla-lobes (bud) 5, joined at the ae glabrous, the longer filaments equalling the anthers in length, the anthers hirsute; ovary glabrous, hemispherical, tapering at the apex into a glabrous style, 3-celled, two cells normal, one cell compressed, ovules few. Fruit globose, glabrous, ca. 1.5 cm. in diameter, 3-celled, few- seeded, 1-2 seeds per cell. Seeds few, large, 8-9 mm. long, 6-7 mm. across, flattened, dark, lustrous The branches and branchlets in this species are angled and winged. This character, along with the thick, conspicuously veined leaves, the long 62 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII thick pedicels (5 cm. long) in pairs, and the large seeds makes this species one of the most outstanding in the genus. Ridley, in his original description, referred to the flowers as ‘Flores non visi.” However, buds were available for dissection. Considering the size of the seeds, so few in number, and the irregularity in the cells of the fruit, I felt compelled to dissect one of the few buds accompanying the type to prove to myself that this species belonged to the genus Adinandra. In the fruit, I found three cells, two of which were larger than the third and occupied most of the fruit. The third cell was compressed along the wall of the fruit and was seedless. The two large cells contained only three seeds between them. These seeds were three times the size normally found in the genus, very uniformly scrobiculate and typical of Adinandra. In the ovary of the flower-bud, the same condition existed, namely two large cells and one smaller cell. The ovules, few in number, in the bud- stage were as large or larger than the mature seeds in most other species of the genus. Otherwise, characters in the flower-bud appeared to be typical for the ae The stamens numbered about 35, possessed hirsute anthers which e quite uniform in size and equalled the longest filaments in length. The corolla-lobes seemed thick in texture On paper, this species is difficult to separate from A. Sarosanthera. In most instances, A. Sarosanthera has terete branchlets. However, in rare cases in the species, one finds definitely angled branchlets, but not as pronounced as in A. angulata where it appears to be the dominant character, extending to the branches themselves oth species have 3-celled ovaries, not so well defined in A. angulata, and both have few but large seeds developing in the fruit. In both species the pedicels are pha somewhat flattened, and the bracteoles are quite minute. In A.-angulata the veins are very prominent on both surfaces with almost equally neaeniehe reticulations. The bracteoles are quite deltoid and occasionally apiculate. In A. Sarosanthera the veins are obscure on the upper surface, only lightly conspicuous beneath, and the bracteoles are rounded. A second related species is 4. /aotica Gagnepain from Laos, Indo-China. In the latter species the ovary and fruit are 5-celled and the bracteoles are caducous. 38. Adinandra Sarosanthera ae, Fl. Ned. Ind. 1 (2): 477. 1859; Ann, Mus. Bot. Lugd.-Bat. 4: 104. 1868. — Szyszylowicz in Nat. Pflanzenfam. THI. 6: 189. 1893. — Merrill, Enum. Born, Plts. 391. 1921; in Univ. Calif. Publ, Bot. 15: 198. 1929, s “sarcosanthera.”’ — Melchior in Nat. Pflanzenfam. ed. 2, 21: 144, 1925. Sarosanthera excelsa Korthals, Verh. Nat. Gesch. Bot. ed. Temminck 104. 1840. Adinandra lamponga Miquel, Fl. Ned. Ind. Suppl. 1: 479. 1862; Ann. Mus. Bot. Lugd.-Bat. 4: 104. 1868. — Szyszylowicz in Nat. Pflanzenfam. III. 6: 189. 1893, as “lampango.’ — Koorders & Valeton, Meded. ’S Lands. Pl. 16: oe (Bijdr. Boomsoort. Java 3: 220). 1896.— Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. Adinandra macrantha Teij smann & Binnendijk in Tijdschr. Ned. Ind. 25: 421. 1863. — King in Jour. As. Soc. Bengal 52 (2): 191 (Mater. Fl. Malay Penin. 1: 131 31). 1890. — Szy szylowicz i in or Pflanzenfam. III. 6: 189. 1893. — Ridley, Fl. Malay Penin. 1: 195, 1922. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 63 Adinandra leiopetala Miquel, Ann. Mus. Bot. Lugd.-Bat. 4: 104. 1868. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. Distripution: British North Borneo, Netherlands East Indies (Borneo, Java, Sumatra), Malay Peninsula (Johore, Perak, Pahang). British NortH Borneo: Elphinstone Prov., Tawao, A. D. E. Elmer 21629 (A, C, M), 21727 (C, G, M, NY), Oct. 1922-March 1923. Borneo: Exact locality lacking, P. W. Korthals s. n. (probable Isotype, G). Java: Cult. Bot. Gard. Buitenzorg, no. VI. c. 6. (G, NY, US), 1903. Sumatra: Atjeh en Ond. Gajo Loeéus, Penosan (Gn. Bohgane & Rangoet Galang), alt. ca. 1800-1900 m., Neth. Ind. For. Serv. b.b. 22381, 22391, 22400 (A) Feb. 1-3, 1937. — East Coast, vicinity of Ack Moente (Aer Moette), Asahan (northeast of Tomoean Dolok and west of Salabat), alt. ca. 500 m., Rahmat Si Boeea 9132 (A), Jan.-July 1936. — Palembang, Banjoesin en Koeboestreken, A. Thorenaar 111 (A), July 14, 1911. — Palembang, Lementang Ilir, Semangoes, Neth. Ind. For. Serv. b.b. 32125 (A), June 27, 1940.— Lampong, J. E. Teysmann s. n, (G).— Exact locality lacking, H. O. Forbes 3191 (G). tAy PentnsutaA: Perak: Sungei Krian Estate, at sea level, G. H. Spare 36003 (A). Johore: Bagan Liman, Sedili River, at low alt., E. J. H. Corner 23896 (NY), Feb. 18, 1931.— Sungai Sedili, E. J. H. Corner 25992 (NY), Aug. 28, 1932 (large tree 17 m. high). — Kangka Sedili Kechil, at low alt., E. J. H. Corner 28581 (A, NY), June 18, 1934.— Sungai Kayu, Mawai-Jemuluang Rd., in swampy forest, at low alt., E. J. H. Corner 28728 (A), Feb. 5, 1935. Pahang: Sungai Bera, near Tasek Bera, at low alt., M. R. Henderson 24113 (NY). Tree 25-30 m. high; branchlets brown, glabrous, appressed-pubescent when very young, the buds densely sericeous. Leaves coriaceous, oblong- elliptic, 9-12 cm. long, 4—5 cm. wide (often much larger, 15-18 cm. long, 6.5-7 cm. wide), obtusely acuminate at the apex, narrowed at the base, the lower portion of the margin entire, the upper part finely serrulate, glabrous when mature (thinly appressed-pubescent beneath when immature), the midrib conspicuous, the primary veins numerous, evenly spaced, con- i ; m. or less, somewhat winged on the upper surface, lower surface semi-terete. Flowers axillary, usually solitary, occasionally in twos; pedicel long, 3—-4.5 cm., glabrous, quite erect, often nodding at the apex, the bracteoles 2, minute, quite rounded, opposite or slightly alternate below the calyx; calyx-lobes 5, somewhat coriaceous, imbricate, glabrous, varying from 5 x 6 mm. (outer lobes) to 8 12 mm. (inner lobes), the outer lobes weakly glandular- denticulate along the margin, the inner lobes with a membranous margin but no denticulations; corolla-lobes 5, obovate, 1.5-1.7 cm. long, glabrous or thinly pubescent on the external surface near the apex; stamens 30-50, brownish, pubescent; ovary variable in pubescence, sometimes quite glabrous, pubescent at the apex, just below the style, or pubescent through- out, ‘tapering at the apex into the style, 3-celled, several-ovuled. Fruit globose, up to 2 cm. in diameter, glabrous or glabrescent, with a few straggling hairs at the apex, 3-celled, seldom over 6 seeds in a single locule, sometimes only 6 in the entire fruit. Seeds flat, hippocrepiform, up to 8 mm. long. This species was originally described by Korthals (1840) under the genus Sarosanthera as S. excelsa. Later, Miquel rightfully transferred the genus to Adinandra. Since the species A. excelsa Korthals already existed Miquel selected the name A. Sarosanthera as most suitable. 64 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII The distinguishing characters of this species are: (1) 3-celled pubescent ovary (usually completely pubescent, occasionally pubescent near the apex only and rarely completely glabrous); (2) large globose fruit (up to 2 cm, diam.), glabrescent, 3-celled or incompletely so, with large seeds up to 8 mm. in diameter (2 or 3 to a cell); (3) filaments glabrous or nearly so with occasional appressed hairs on the exterior surface; (4) styles glabrous; (5) calyx-lobes glabrous; (6) corolla-lobes glabrous; (7) primary véins many (ca. 20 pairs) evenly spaced and quite scalariform in appearance; and (8) pedicel 3-4.5 cm. long. This species is nearly as widespread as A. dumosa Jack, extending from Borneo through Sumatra and some of the smaller East Indian islands up into the Malayan Peninsula. In Borneo it has been referred to as A. Sarosanthera (and rightly so) by students of that island flora. In Sumatra, workers have called it A, lamponga and on the Malayan Peninsula it has been known as 4. macrantha. ‘The last two names have been combined before. Koorders and Valeton (1896) suggested the very close relationship among the three _ felt that they should be combined under a single species. a species as widespread as this, and cee in this family, one must eee considerable variation. The ovary, in authentic material from Sumatra, is extremely variable. Material. collected by J. FE. Teysmann s. n. from Lampong in southern Sumatra, the type-locality of A. damponga, presents a pubescent ovary. Rahmat Si Boeea 9132 collected on the east coast of Sumatra near Asahan also has a pubescent ovary. Thorenaar 111 and Neth. Ind. For. Serv. 6.6. 32125 collected near Palembang, Sumatra, both in fruit, show the ovary to have been pubescent because, although glabrescent in maturity, there are vestiges of pubescence at the apex. Forbes 3191 collected in Sumatra (the locality unknown to me), has a glabrous ovary. Cultivated in the Botanical Garden at Buitenzorg, Java, under the number (VI-c-6), is a specimen with a distinctly glabrous ovary. From British North Borneo Elmer 21629 in mature fruit seems to be quite glabrous and one may assume that the ovary was glabrous or nearly so in the flowering stage. . A Korthals specimen from Borneo labeled A. Sarosanthera has flowers with the ovary distinctly glabrous. propos of this discussion of pubescence, Koorders & Valeton in their discussion of A. lamponga (see reference above), stated that their descrip- tion was drawn up according to specimens in the herbarium of Koorders. They stated further that the living authentic specimen of A. macrantha | Hort. Bog. VI-c-6| examined by them was identical with A. lamponga and they united the two species under A. /amponga after comparison of A, macrantha with an authentic specimen [ 4. /amponga| in “Herb. Bogor.” and with the description by Miquel with which it appeared to them to agree fully. Only in the Sumatran specimen was the ovary truly pubes- cent. In Koorders’ specimen the ovary was pubescent only at the apex. Koorders & Valeton stated that some specimens from Djampang-Koelon (Herb. Kds. 81598) differed through an entirely glabrous ovary and much 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 65 larger fruit 2.5 cm. < 2 cm. with the fruit walls 4 mm. thick. As for the leaves, the description of A. leiopetala Miquel agreed better with their specimen than that of A. lamponga. In the specimen of Teysmann in Herb. Hort. Bog., however, the leaves agree with the description of A. leiopetala in the projecting base. Koorders & Valeton also feel that both species are very nearly related to A. Sarosanthera Miq. I have vacillated considerably in my attitude concerning this species, going so far as to attempt to write them up separately. However, my final conclusion is that the whole group should be combined under a single species, A. Sarosanthera Miq., even though the variation, especially in the pubescence of the ovary, is considerable. In a few of the specimens, particularly those collected by E. J. H. Corner in Johore, the young branchlets are quite noticeably angled. From the size of the leaves and branchlets, especially Corner 28581, I feel further that the specimens must have been collected from very vigorous young shoots. 39. Adinandra cordifolia Ridley in Kew Bull. Misc. Inform. 1938: 173. 1938. Sarawak: Kuching, G. D. Haviland (= b. p. r. c.) K. S. = 23 (isotype, G), Sept. 19, 1892. BritisH NortH Borneo: Mt. Kinabalu: ‘Tenompak, alt. 1650 m., J. & M.S. Clemens 29929 (A), June 8-11, 1932 (flower-buds opening white, the flowers with pure white petals and brown stamens). — Marai Parai, Upper Kinataki Gorge, alt. 1800 m., J. & M.S. Clemens 32434 (A), March 29, 1933 (tree 20 m. high; flowers cream-colored). — Marai Parai, head of the Sadikan River, alt. 2000 m., J. & M. S. Clemens 32863 (A), April 22, 1933 (tree 7 m. high; buds gray). Tree up to 20 m. high; branchlets terete, brown, glabrous, the very young branchlets appressed-pubescent, the terminal buds conical, appressed- pubescent. Leaves coriaceous, oblong-elliptic, 11-20 cm. long, 5-8 cm. wide, acuminate or obtusely acuminate at the apex, unequally cordate at the base, glabrous (the lower surface of the very young leaves pubescent), the margin recurved, entire or nearly so, the veins 10-12 pairs, conspicuous on the lower surface, the petiole thick, very brief, 2—3 mm. long, occasionally subsessile. Flowers axillary, solitary or in pairs; pedicel thick, sericeous, ca. 5 mm. long; bracteoles 2, persistent, coriaceous, opposite, immediately below the calyx, sericeous, unequal, the smaller one ca. 3 mm. long, 4 mm. wide, rounded, the larger one ca. 6 mm. long, 8 mm. wide; calyx-lobes 5, imbricate, sericeous, unequal, broader than long, the outer lobes larger, ca. 12 mm. long, mm. wide, ca. 2 mm. thick at the base, the inner lobes ca. 8 mm. long, 9 mm. wide, the margin scarious; corolla-lobes 5, connate at the base, oblong, glabrous, ca. 20 mm. long, mm. wide, rounded at the apex; stamens ca. 40, 2- or 3-seriate, 11-17 mm. long (in a single flower), the filaments 7-11 mm. long, connate at the base, adnate to the base of the corolla, lightly pubescent, at least on the upper half, the anthers oblong, ca. 5 mm. long, merely lightly pubescent, some appearing glabrous; ovary sericeous, subglobose, ca. 4 mm. diam., 5-celled, multi-ovulate, the style ca. 13 mm. long, sericeous at the base, otherwise glabrous. Fruit not seen. The large cordate leaves and the very large flowers are the outstanding characters in this species. The Clemens material cited above made 66 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII possible very essential additions to Ridley’s original description. The ovary is 5-celled and multi-ovulate; the style 13 mm. long and glabrous except at the base; the stamens are unusually long with the anthers only slightly and intermittently lightly pubescent; the corolla-lobes are glabrous and up to 20 mm. long; the calyx-lobes are broader than long, sericeous, the outer larger than the inner. Ordinarily, the inner lobes are the larger. 39a, Adinandra cordifolia var. strigosa, var. nov. Arbor; ramis teretibus, brunneis dense adpresso-pubescentibus, ramulis juventute dense flavo-sericeis, gemmis terminalibus, conicis, dense flavo- sericeis. Folia crasso-coriacea, elliptico-oblonga vel elliptica, 18-21 cm. longa, 11-12 cm. lata, apice abrupte acuminata, basi inaequilateralia, sub- cordata, supra es glabra, subtus dense pallido- pubescentia, margine ut videtur integra, venis primariis 18-20 ] paribus, subtus er cee naa intra marginem eden ie venis ote inter primarias fre quentibus, petiolis minutis, ca. 1-2 mm. longis, dense pubescentibus, vel rariter Salamone Flores non visi. So juvenes axillares, solitaril: pedicellis ca. 1 cm. longis, teretibus, dense fulvo-pubescentibus, recurvatis; bracteolis . 2. oppositis, ‘a asneger ag pubescent subrotundatis, in- aequalibus, circiter 3° X mm. 5 6 mm. ; sepalis 5, imbricatis, | subrotundatis, dense adpresso- ies yubescentibus, inaequalibus, circiter 8-10 mm. longis, 7—11 mm. latis; oat elongato-globoso, ee 5-sulcato, 5-loculari, multi-ovulato; stvlo circiter 1 cm. longo, glabro basi excepto. DisrripuTion: British North Borneo. British Norti Borneo: Mt. Kinabalu: Dallas, alt. 1000 m., J. Clemens 26229 (tTypr, A), Aug. 21, 1931. — Upper ae eer Penibukan, west ridge, jungle, alt. 1350 m., J. & M.S. Clemens 40774 (A . 18, 1933 (tree). The main difference between 4. cordifolia and " ee A. cordifolia var. strigosa lies in the permanence of the pubescence. Both entities are pubescent in the terminal bud, the very young unfolding leaves, and the young branchlets. In the species this pubescence appears to be very evanescent, the leaves and older branchlets shortly becoming quite glabrous or glabrescent. In the variety, the disappearance of the pubescence is less rapid. In branchlets showing growth of over a single year and in the mature leaves a dense strigose pubescence, Per rhaps permanent, is found. Because of this very apparent difference one is inclined at first to consider the two as separate species. Sean when an accurate appraisal of all parts has been made, the persistence of this pubescence seems to be the single basis of separation. In the type, J. & M.S. Clemens 26229, the young fruits are subtended by calyces similar in size to those of the species. However, in J. & Clemens 40774, the second specimen cited, the calyx-lobes are Sohiiat smaller, measuring ca. 8 mm. in length. 40. Adinandra subsessilis Airy-Shaw in Kew Bull. Misc. Inform. 1939: 505. 1939. DistrRIBUTION: Sarawak. SARAWAK: Mt. Dulit, on steep slope of rain-forest, alt. 700-900 m., Native Collec- tor (P.W. Richards no. 1958) (A, IsoTYPE), Sept. 19, 1932 (small tree with “whitish” petals). -1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 67 Small tree: branchlets terete, brownish, later glabrescent, the very young branchlets pilose, the buds sericeous. Leaves chartaceous or subcoriaceous, elliptic-oblong to oblong-ovate, 5-11 cm. long, 1.5-3.5 cm. wide, distinctly acuminate at the apex with the acumen nearly 2 cm. long, narrowed at the base, occasionally subrotund, the margin subentire, minutely glandular- serrulate, glabrous on the upper surface, appressed-pubescent beneath, later glabrescent, the midrib lightly impressed above, prominent beneath, pubescent, the veins 12-14 pairs prominent on both surfaces, arching upward and anastomosing, the petiole 2 mm. long, pubescent. Flowers axillary, solitary; pedicel very short, 1-2 mm. long, densely sericeous: bracteoles 2, persistent, sepaloid, ovate, opposite, immediately below the calyx, about as long as the pedicel, densely sericeous; calyx-lobes 5, im- bricate, unequal, thickened, deltoid-ovate, acute, sericeous on the exterior surface, glabrous within, the two outer lobes 3-5 mm. long, 2-3 mm. wide, 3 inner lobes larger, 4-7 mm. long, 33.5 mm. wide; corolla-lobes 5, white, broadly obovate, 8-9 mm. long, 7-8 mm. wide, rounded at the apex, sericeous along the median portion of the exterior surface, otherwise gla- brous: stamens (fide Airy-Shaw) about 30, uni-seriate, 4-6 mm. long, joined at the base and adnate to the base of the corolla, the filaments hairy along the upper portion, the lower part glabrous, the anthers lanceolate, 2-3 mm. long, sericeous on lower portion of dorsal surface, otherwise glabrous; ovary depressed-globose, 2.5 cm. diam., densely sericeous, 2-celled, pauci-ovulate, the style 5—6 mm. long. glabrous or slightly pubes- cent at the base. Fruit (Airy-Shaw) subglobose, ca. mm. diam., long-pilose, 2-celled, each cell one-seeded. Seeds discoid, 6-7 mm. diam., —2 mm. thick, shining. According to Airy-Shaw’s findings, the developed fruit of this species is two-celled with a single large seed in each cell. I have not had an oppor- tunity to observe this unusual feature but have examined an ovary from a bud of an isotype. A dissection showed the two cells in the ovary with a small number (approximately four) of ovules in each cell. This was the first species described with a two-celled ovary and fruit. The seeds are 6-7 mm. long, adding another species to the small group in which the seeds are few in number and develop to a size at least twice that which is found in the majority of species in this genus. In this paper a second species with a two-celled ovary, A. nunkokensis, is presented. This latter species has been collected on Mt. Nunkok, Mt. Kinabalu, in British North Borneo. The only evidence of relationship between these two entities appears to be the two-celled ovary. Besides the ovary, the features which are helpful in identification are: (1) nearly sessile leaves with the petiole not more than 2 mm. long, and the apex long-acuminate; (2) pedicels very brief (1-2 mm. long), or the flowers subsessile; and (3) ovary and fruit long-pilose, the style glabrous except at the base. 41. Adinandra parvifolia Ridley in Jour. As. Soc. Straits Settl. 54: 23. 1910; Fl. Malay enin. 195. 1922. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. Adinandra montana Ridley in Jour. Fed. Malay States Mus. 6: 44. 1915.— Non Merrill. joan Distrisution: Malay Peninsula (Pahang, Perak). 68 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Pauanc: Gunong Tahan, alt. 1650-1800 m., L. Wray & H.C. Robinson 5517 (K), July 12, 1905 es 10-15 m. high white flowers). PERAK: Gunong Kerbau, alt. 2200 m., H. C. Robinson s. n. (TYPE of A. montana, K), March 14, 1913. Tree 10-15 m. high; branches grayish brown, terete, pone the branchlets brown, glabrous, a terminal buds sericeous. ves coria- ceous, glabrous, elliptic, 6-8 cm. long, 3-4 cm. wide, obtuse : the apex, acute at the base, the margin somewhat recurved, lightly glandular- serrulate, the veins ca. 6-8 pairs, usually rather inconspicuous, occasionally mm. long. Flowers axillary, solitary, large, ca. 2 cm. in diameter; pedicel ca. | cm. long, glabrous; bracteoles 2, persistent, teres immediately below the calyx, glabrous, broadly ovate, unequal, ca. 2 * 2 mm.and3 X 3 mm.; calyx- lobes 5, imbricate, ca. 1 cm. long ee wide, unequal, the outer lobes broadly ovate, finely pubescent over ail the external surface, the inner lobes more nearly rounded, pubescent on the central portion of the external surface only, the margin scarious, glabrous; corolla-lobes 5, lightly connate at the very base, obovate, 15-16 mm. long, 9-12 mm. wide, pubescent on the upper half of the median portion of the external surface; stamens 40-45, ?-seriate, unequal 8-12 mm. long, the filaments variable in length 3—6.5 mm. long, pubescent on the upper third, — at the base, adnate to the base of the corolla, the anthers oblong, 5 mm. long, hirsute; ovary ca. 5 mm. across, densely pubescent, longitudinally fur- rowed, 5-celled, aed ovulate, tapering at the apex into a style ca. 1 cm. long. Fruit not see The outstanding sauae of this species to be used in determination are: (1) small oblong-elliptic leaves up to 8 cm. long, lightly veined with 6-8 pairs of veins; (2) glabrous pedicels (ca. 1 cm. long) and bracteoles, the latter persistent, unequal, and broadly ovate; (3) broadly spreading flowers, up to 2 cm. in diameter with long (15-16 mm.) pubescent corolla- lobes; (4) stamens long (up to 12 mm.), the anthers hirsute, ca. 5 mm. long, the filaments up to 6.5 mm. long and pubescent along the upper third, otherwise glabrous; and (5) ovary pubescent, longitudinally furrowed, 5-celled Ridley himself listed A. montana Ridley as a synonym of this species. Adinandra montana Ridley (1915) should not be confused with the earlier A, montana Merrill (1910) from the Philippine Islands. The latter species has been transferred to Cleyvera japonica Thunberg emend. Sieb. & Zucc. var. montana (Merrill) Kobuski. 42. ae israel Hullettii King in Jour. As. Soc. Bengal 59 (2): 191 larch ae Pept 1890. — _ in Nat. Pflanzenfam. III. Ridley, FI. Malay Penin. 1: 1922. — Melchior in Nat. oo poe 2, DistripuTion: Malay Peninsula (Singapore). SINGAPORE: Entrance to Botanic Garden, R. W. Hullett 103 (type, K), July 1885. ; branches terete, pubescent, furrowed, light brown, the youn branchlets densely and minutely rusty-tomentose, the terminal buds densely tawny-sericeous. Leaves coriaceous, oblong- elliptic, 9-14 cm. long, 4—5.5 cm. wide, shortly and plant acuminate at the apex, acute at the base, the 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 69 margin denticulate, slightly recurved when dry, the upper surface smooth, shining, glabrous, the lower surface opaque, densely and minutely tomen- tose, the midrib prominent, the nerves spreading, obscure, ca. 10 pairs, the petiole ca. 5 mm. long. Flowers axillary, solitary or in twos; pedicel tomentose, ca. 1 cm. long; bracteoles 2, opposite, ersten close to the calyx, broadly ovate or deltoid, ca. 5 X 5 mm. and 3 x 3 mm., densely pubescent; calyx-lobes 5, imbricate, fleshy, spreading, broadly ovate, unequal, the two outer lobes joined at the base appearing as though coalesced, ca. 10 mm. long, 8-9 mm. wide, pubescent over the entire exterior surface, the three inner lobes broader than long, ca. 8 mm. long, 9 mm. wide, pubescent on the central portion of the exterior surface, the margin oe scarious: corolla-lobes 5, ca. 12 mm. long, 5 mm. wide, connate for m m the base, rounded and recurved at the apex, glabrous; stamens ca. 45, seemingly 3- seriate, ve outer stamens erect, the second oblong, ca. 5 m io ong, densely hirsute, tapering into long, quite blunt apicules, the apices glabrous; ovary conical- penispheneal appressed- sericeous, ca. 4 mm. diameter, 5-celled, multi-ovulate, tapering into an entire style ca. - mm. long, glabrous except at the juncture with the ovary. Fruit not seen. Some of the outstanding features of this species are: (1) 5-celled sericeous ovary; (2) crowded stamens (ca. 45), with filaments glabrous, the inner two rows geniculate towards the apex, the anthers oblong hirsute, and the apicule long, blunt, glabrous; (3) corolla-lobes glabrous, recurve connate for nearly one-half their length; (4) style glabrous; (5) Pieipteates persistent, opposite, unequal, densely pubescent; and (6) both branches and branchlets densely pubescent. Melchior, in his key to the species, refers to the filaments as pubescent. My dissections show the filaments to be glabrous, and since King did not mention pubescence in describing the filaments one must assume that he also considered them glabrous. His description proved to be most detailed and accurate. 43. gee maculosa T. Anderson ex Dyer in Hooker f., Fl. Brit. Ind. 1: 282. a — King in Jour. As. Soc. Bengal 59 (2): 189 (Mater. FI. Malay tage 122129)): 1890. ere in Nat. Pflanzenfam. III. 6: 189. ea — Ridley, FI. ene Penin. 1: 195, 1922. — Melchior in Nat. Pflanzenfam. ed. 2, 21: vi 1925. Poricowmia ? integerrima Wallich, Cat. no. 1452 (in Herb. pee not in Herb. Linn. son. DistTRIBUTION: Malay ae (Pahang, Penang, Perak). PAHANG: Fraser’s Hill, E. J. H. Corner 33170 (A), Aug. 4, 1937. — Fraser’s Hill, upon the Selangor border, ae Pea 1350 m., M. Nur 11320 (A), Sept. 1, 1923. Penanc: N. Wallich 1452 (type, K). Perak: Larut, rocky locality, rich soil, alt. 800-1000 m., Dr. King's Collector 6331 (A), July 1884 (tree 15-30 m. high; flowers white, dark ge within, with glossy, waxy stamens). Exact locality lacking: A. C. Maingay 186 (G) Tree up to 25 m. high; branches terete, brown, glabrous, the branchlets brown, glabrous, appressed-pubescent near the apex, the terminal buds 70 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXVHI small, re aig A daa Leaves heavily coriaceous, elliptic to subro- tund, 10-15 cm. long, 4-5 cm. wide, shortly and abruptly acuminate at the apex, i at the base, glabrous, shining above, paler, dull, opaque, minutely roughened, generally glabrous beneath, occasionally lightly pubescent along the lower midrib, the margin entire, the veins very indis- tinct on both surfaces, the petiole ca. 5 mm. long. Flowers axillary, solitary; pedicels ca. 1 cm, long, generally appressed-pubescent;, bracteoles persistent, opposite or subopposite, immediately below the calyx-lobes, subrotund, unequal, 2-3 mm. long, appressed- eee calyx-lobes_ 5, imbricate, appressed-pubescent, unequal, ca. 7 mm. long, the inner lobes pubescent on the central portion of the exter cal surface, thinner and glabrous at the margin; corolla-lobes (fide King) “membranous, ovate, acute, glabrous, connivent”; stamens (fide King) “about 30; filaments attached to the petals, short, glabrous; anthers narrow, the cells elongate, lateral; the connective sericeous with short glabrous apiculus’; ovary depressed-hemispheric, Spires 5-celled, multi-ovulate, the style gla- brous. Fruit ca. 1 cm. in diameter, globular, at hrst pubescent, later glabrescent, the seeds numerous, small, shining, dar Anderson describes the leaves of this species as ‘‘finely mottled beneath with reddish-brown remotely denticulate glabrous.” The leaves of the type are finely but obscurely mottled beneath with reddish brown patches. However, on no other specimen examined did I find the coloration. King (and I depended much on his excellent description) refers to the under surface of the leaf as “pale brown, dull, opaque, minutely rugulose when dry.’ This species seems most closely related to A. vial rima 'T. Anders. which has silky pubescent terminal buds and thinner leaves. In 4. maculosa, the leaves are unusually thick, with the ae obscure, if at all obvious, on both surfaces. “haracters which are helpful in determination are: (1) thick glabrous leaves, opaque beneath and very abruptly acuminate; (2) corolla-lobes glabrous; (3) bracteoles persistent, subrotund and pubescent; (4) ovary densely pubescent, 5-celled with the style glabrous; and (5) filaments glabrous. — 44. Adinandra nunkokensis, sp. nov. Ramuli tereti, brunnei, pubescentes, gemmis ee conicis, dense sericeis. Folia coriacea, elliptico-obovata, 8-14 cm. longa, 2.5—4.5 cm. lata, apice acuminata, basi acuta, supra glabra nitida, subtus_ leviter maar pubescentia, margine leviter denticulata, nervis lateralibus pri- na 6-8 paribus, gracilibus intra ve: anastomosantibus, venis ures me inter primarias frequentibus, petiolis 4—7 mm. longis, supra planis la subtus semi-teretibus, dense eo Flores axillares, solitarii vel bini; pedicellis rugulosis, brevioribus crassioribusqu mm. eee 4 mm. crassis, recurvatis, leviter adpresso- -pubescentibus: bracteolis 2, alternatis, persistentibus, late ovatis, subaequalibus, ca. 2.5—3 mm. longis et 3 mm. latis, leviter adpresso-pubescentibus; sepalis_ 5, imbricatis, rugulosis, subrotundatis, ca. mm. longis, 5-6 mm. latis, pubescentibus, margine subscariosis glandulosis; petalis 5, basi connatis, cremeis, oblongis, 12-14 mm. longis, 5-7 mm. latis, medio dorso sericeis; staminibus 30-35, 4-seriatis, inaequalibus, 8-11 mm. longis; filamentis 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 71 4-7 mm. longis, in toto connatis, basi ad corollam adnatis, inferne nudis, superne hirsutis, antheris ca. 4 mm. Mahan wn hirsutis, vix apiculatis, ovario subgloboso dense sericeo, ca. 4-5 mm. metro, 2- ‘loculari, pauci- ovulato, ovulis magnis, ae gracile, ca. 11 mm. roneer basi hirsuto, superne glabro. Fructus non visu DistrtpuTion: British an Borneo. British Nortu Borneo: Mt. Kinabalu, Mt. Nunkok, alt. 800 m., J. & M. S. ae 32900 (TYPE, A), April 22, 1933. This species is related to Airy-Shaw’s A. subsessilis of Sarawak. Both have two-celled ovaries. Although no fruiting specimen was available for this study, I feel, from the small number and the large size of the ovules, that A. nunkokensis will probably have very few seeds. Adinandra sub- sessilis produces two seeds in fruit, one in each cell. The flowers in 4. subsessilis are nearly sessile, with pedicels only 1-2 mm. long, the bracteoles sepaloid, the calyx-lobes distinctly acute, the stamens uni-seriate and only 4—6 mm. long, and the style 5-6 mm. long. The leaves are more finely acuminate at the apex, the petiole but 2 mm. long, and there are 12-14 pairs of veins, prominent on both surfaces. In A. nunkokensis the flowers are brietly pedicellate, but the pedicels are longer (ca. 5 mm.) than those in A. swbsessilis and much thicker, the bracteoles are more rounded, the calyx-lobes subrotund, the stamens 4-seriate and 7-11 mm. long, the style 11 mm. long. The leaves are acuminate but not as finely attenuate as in A. subsessilis, the petiole 6-8 mm. long and there are only 6-8 pairs of veins. . Adinandra Corneriana, sp. nov. Arbor?; ramis teretibus, brunneis, anne ramulis crassis, tereti- bus, brunneis, dense luteo- ferrugineo-pubescentibus, gemmis terminalibus crasso-conicis dense ferrugineo- sericeis. Folia chastacea vel subcoriacea, oblongo-elliptica vel obovata, 14-20 cm. longa, 6-9 cm. lata, apice acuta vel obtusa, leviter acuminata, basi cuneata vel obovata, ar glabra, subtus 5, imbricatis, exteriore omnino dense leas ubesce ntibus, interiore glabris, binis exterioribus ee 6-8 mm. longis et latis, tribus interioribus maioribus, 8-10 longis, loti mm. latis; petalis basi connatis, apice acutis, Se pes: pubescentibus; stamina 30-35, 3- vel 4-seriatis, filamentis connatis, basi corollam adnatis, exteriore dense hirsutis, antheris oblongis, dense hirsutis, ovario semigloboso, hirsuto, 3- loculari, pauci-ovulato, stylo glabro, circiter 5 mm. longo. Fructus globosi, hirsuti, circiter 8 mm. diametro, 3-loculares, uno semine in quoque loculo, semini- bus subplanis, hippocrepiformibus, magnis, nitidis, circiter 5 mm. longis et 3 mm. lat Ct hee Malay Peninsula (Johore). Jouore: Mawai-Jemaluang road (13.5 miles), by stream on hillock, at low alti- 72 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XXVIII tude, E J. H. Corner 29010 (TYPE = fragment A), Feb. 8, 1935. — Mawai-Jemaluang road (13.5 miles), at low altitude, FE. J. H. Corner 29426 (K), May 13, Kayu Ara, Mawai-Jemaluang pet in dry Dryobalanops forest, at low altitude, E. J. H. Corner 29376 (K), May 5, Although three excellent specimens have been collected of this species, the material appears rather sparse. As a result few dissections were made. Since no open flowers were available, the buds were used. ence I have refrained from recording measurements of such parts as corolla-lobes and stamens which were obviously not fully developed. However, the charac- ters which might be used in a key, other than size, have been recorded above The three-celled ovary and fruit, with few ovules and seeds are important characters. The ovary of the bud (Corner 29010 type) exhibited few large ovules, hence one might anticipate fewer and larger seeds in the fruit. For the study of the fruit Corner 29376 was used. Here, in both fruits examined, it was found that a single large seed developed in each of the three cells of the fruit. Size is the distinctive element. Otherwise, the seeds were typical of the genus. All three specimens were collected by E. J. H. Corner in Johore at varying dates during the year 1935. It is a pleasure to name this species A. Corneriana in his honor. S. 46. Adinandra caudatifolia, sp. nov. Arbor ad 35 m. alta; ramis teretibus glabris, griseo-brunneis, ramulis teretibus, glabris, brunneis, ramulis juventutissimis adpresso-pubescentibus, gemmis terminalibus gracilibus, adpresso-sericeis. Folia chartacea, ellip- tica, 5-8 cm. longa, 2.5-3.5 cm. lata, apice abrupte caudata, basi acuta, obtusa vel subrotundata, supra nitida, glabra, subtus pallidiora glabra, rariter leviter pubescentibus, margine integra vel leviter crenulata, venis primariis ca. 8 paribus, intra marginem anastomosantibus arcuantibusque, petiolis minutis circiter 4 mm. longis, glabris. Flores non visi. Fructus axillares, ut videtur solitarii; pedicellis teretibus, glabris, ca. 2 cm. longis, verruculosis subferrugineis; bracteolis 2, oppositis, persistentibus, minutis circiter 1 mm. longis; sepalis 5, imbricatis, aaa creas recurvatis brescentibus, circiter 5-6 mm. longis, 4-5 mm. latis. Fructus a glabrescens, circiter 1.5 longus, 1 cm. pen 3-locularis, seminibus circiter 9, magnis, ahaa nitidis, circiter 7 x 5 mm Distripution: British North Borneo. British NortH Borneo: Upper Mt. arene! Penibukan, west ridge, jungle, alt. 1700 m., J. & M.S. Clemens 50377 (typr, A), Nov. 4, 1933 (tree 20 m.; fruit green — Upper Mt. Kinabalu, Penibukan, east ridge, jungle. alt. 1850 m., J. & M. o; Cloniens 50387 (A), Nov. 13, 1933 (tree 30 m.; fruit brown-purple). The species, A. caudatifolia, as the name signifies, is characterized by the tail- like apex of the leaf. Furthermore, the leaf is chartaceous, rather than coriaceous, as is the case in most species. The material lacks flowers. The fruit, with its evidence of pubescence both at the apex and base, shows that the ovary had been quite densely pubescent. This pubescence extends up the style. To what extent is not known, since all styles were broken near the base. The fruit is distinctly three-celled and the seeds large (ca. 7 X 5 mm.) and few in number (ca. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 73 3 to each cell). The pedicels are verruculose with subferrugineous swellings. 47. Adinandra pourcey Korthals, Verh. Nat. Gesch. Bot. ed. Temminck 109. 1840. — Miquel, Fl. Ned. Ind. 1 (2): 478. 1859; Ann. Mus. Bot. Lugd.-Bat. 4: 103. ee in Jour. Linn. Soc. Bot. 13: 330. 1873; in Hooker f., Fl. Brit. Ind. 1: 282. 1874.— King in Jour. As. Soc. Bengal 59 (2): 190 (Mater. Fl. Malay Penin. 1: 128). 1890. — Szyszylowicz in Nat. Pflanzenfam. III. 6: 189, 1893. — Ridley, - Malay Penin. 1: 194. 1922.— Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. Gordonia pane Wallich, Cat. no. 3664 in Pages Linn. Soc. Ternstroemia ? coriacea Wallich, Cat. no. 1 Camellia oie Wallich, Cat., p. 158 (not Roxb. ex Bot. Reg. 349, see Jour. ley . 13: 330. 1873). Polyspora aes. Choisy in Mém. Soc. Phys. Hist. Nat. Geneve 1: 179 (Mém. Ternstr. 91) 1855. Non Don DistriBuTION: Malay Pesca, Sumatra. Matay PENINSULA: ohore: Gunong Pulai, alt. 300 m., M. R. Henderson 28514 (K), April 29, 1934 (small slender tree 7-8 m. high; flowers white). — Kluang, Forest Ranger (Abdul Rahim) 5918 (K), May 3, 1925.—Sedanah, H. N. Ridley 13504 (K), Nov. 24, 1908. Perak: Larut, open top of mountain, alt. 1000-1300 m., Dr. King’s Collector 6289 (K), July 1884 (tree 12-15 m. high; leaves dark glossy green). — B. Scortechini 345 (K).— Gunong Butu, Puteh, Lower Camp, alt. 1130 m., L. Wray 1158 (K).— Exact locality lacking, L. Wray 2804 (US).— Larut, high mountains, alt. 1200-1300 m., Dr. King’s Collector 6352 (US), el 1884 (tree 20-30 m. high; er ae glossy. Breen fruit grayish brown). Penang: Western Hill, along path, alt. 800-900 m a R. Henderson 21358 (K), Sept. a. 1934 (small slender tree). can Hill, ._ C. Maingay 2267 (185) (G, US), Feb. 1867.— Open jungle, hilly locale, alt. 300- m., Dr. King's Collector 4842 (US), Aug. 1883 (tree 20-30 m. high; leaves dark glossy green; flowers white). Malacca: Garden jungle, H. N. Ridley 1952 (A), 11334 (K), 1892 and 1893. Singapore: Bukit Timah, _ J. H. Corner 34932 (A, K), April 12, 1938.— Common, H. J. Murton 111 (G, K), Dec. 1878. ite puden F. G. Mohamud 15518 (K), eed Bi; Small tree up to 25 m. high; branches brown-gray, terete, ai the branchlets glabrous except for the appressed pubescence on the current year’s growth, the terminal buds elongated, densely sericeous. Leaves coriaceous, oblong-elliptic to broadly ovate, 8-15 cm. long, 4-5 (—6) cm. wide, generally sharply and finely acuminate (up to 3 cm. long) at the apex, occasionally less prominent, acute at the base, glossy, dark green, glabrous above, somewhat paler, glabrous or scattered-pubescent beneath, the margin entire or obsoletely denticulate, the veins ca. 12 pairs, usually with double arching, not always, the petiole ca. 5 mm. long. Flowers axillary, solitary or in twos; pedicels terete, ca. 1 cm. long, finely appressed- pubesc cent at first, later glabrescent, smooth at anthesis, later developing wart-like protuberances becoming very rough and pronounced at the fruiting stage; bracteoles 2, persistent, alternate along the pedicel, the upper one not close to the calyx, small triangular, ca. 1 mm. long, usually pane -pubescent; calyx-lobes 5, imbricate, finely appressed-pubescent, unequal, ca. 6-7 mm. long, the outer two lobes narrower, ca. 3 mm. wide, the inner three lobes more rounded ca. 5.5 mm. wide; corolla-lobes 5, connate at the base, rounded at the apex, ca. 10 mm. long, 7 mm. wide, 74 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII lightly appressed-sericeous on the median portion of the external surface; stamens seemingly 4-seriate, unequal in length, up to 10 mm. long, the filaments adnate to the base of the corolla, connate the length of the shortest stamen, free from that point, lightly pubescent on the inner surface, ei anthers elongate, short-hirsute; ovary conical, densely pubescent, ca. mm. across, 3-celled, pauci-ovulate, tapering at the apex into an entire ae 7-8 mm. long, pubescent nearly to the apex. Fruit quite globose, up to 2 cm. diameter, glabrescent with evidence of pubescence near the apex, few-seeded, the pedicel at this stage rugged and warty, the seeds light in color, ca. 6 mm. long. Until borrowed material from Kew was obtained this species presented considerable difficulty. In the original description, Korthals stated that the sepals were glabrous and the fruit 4- or 2-celled. My study revealed the sepals to be distinctly appressed-pubescent and the ovary definitely 3-celled. However, these conclusions were drawn from Malayan material rather than Sumatran. Dyer, in Hooker f., Fl. Brit. India, 1: 282. 1874, stated also that the sepals and the leaves were glabrous. Later King (Jour. As. Soc. Bengal 59 (2): 190 |Mater. Fl. Malay Penin. 1: 128]. 1890) remarked that the species was glabrous in all parts except the stamens, and Ridley (Fl. Malay Penin. 1: 194. 1922) gave the same impression. From all the material examined I find the species to be quite pubescent. Korthals gave this impression in his description, and authentic Korthals specimens from Sumatra, although lacking both flowers and fruit, show the young branchlets, buds and leaves to be pubescent. At first I felt that the British botanists were misinterpreting Korthals’ species and perhaps were working with an entirely different species. However, the material from the Kew herbarium agreed with our own specimens in all instances, some actually being duplicates. The young branchlets, buds, under surface of leaves, pedicels, bracteoles, calyx-lobes, corolla-lobes, stamens, ovary and style are all pubescent in varying degrees. The leaves, calyx-lobes, pedicels, bracteoles and fruit tend to become glabrescent. The apex of the leaves of the Sumatran material is much more fine- and long-acuminate than the apex of those of the Malayan material. In some of the Korthals specimens the acumen is drawn out to as much as 3 cm. and appears somewhat caudate. In the Malayan material, although detinitely acuminate, the leaves are seldom as distinctive. Some workers have placed stress on the double arching of the veins of the leaves. This character varies considerably and is hardly stable. Also the wart-like condition of the fruiting pedicel has been stressed. As striking as this character happens to be, it is not found in the flowering material, for some reason not appearing until the developing fruit stage. Other characters which may be used in determination are (1) the alternate, small, persistent, triangular bracteoles; (2) the pubescent calyx- lobes and corolla-lobes; (3) the pubescent filaments joined the length of the shortest filament in series; (4) the pubescent, 3-celled ovary, pauci- ovulate, and the 3-celled fruit, few and large seeded; and (5) the entire style pubescent almost to the apex. 1947 | KOBUSKI, STUDIES IN THE THEACEAE, XV 75 The synonymy cited above is merely quoted from Dyer in Hooker f. ¥'1. Brit. India, 1: 282. 1874. It has not been possible for me to verify these names. However, I wish to have them quoted here as a record. VERNACULAR NAMES: Kelat, Membasah. 48. Adinandra integerrima Anderson ex Dy n Hooker f., Fl. Brit. Ind. 1: 282 1874. — King in Jour Soc. Bengal 5 Yay 190 tates FI. Fcogas Peni. 1: 130). 1890. * Says in Nat. Pflanzentam. III. . 93. — Ridley, Fl. Malay Penin. 1: 1922.— Melchior in Nat. noes : 2, 21: 144. 1925. — Craib, FI. ati Bean 1:125. 1925 Adinandra dasyantha Choisy in Mém. Soc. PAYS: Nat. Genéve, 1:112 (Mem. Ternstr. 24). 1855.— Non Korthals Ternstroemia integerrima Wallich, Cat. no. 1452, Ternstroemia reticulata Wallich, Cat. no, 2246 Gordonia reticulata Wallich, Cat. no. 3663. Camellia japonica Wallich, Cat. no. 3667, in herb. Kew, not in Herb. Linn. Soc. tala integerrima Pierre var. @ migro- pe Pierre, Fl. For. Cochinchine . 125. 1887. DistripuTION: Malay Peninsula. Penaxc: N. Wallich 3663 (typr, K). Jonore: H.N. Ridley 6336 (K). PAauanc: Sungei Yet., Bukit Fr ies alt. 1250 m., Native Collector 11149 (A), Sept. 29, 1923. Exact locality lacking, NW. Wallich 3667 (IX). in Herb. Linn. Soc. —_~ Small tree; branches terete, brown-gray, the branchlets appressed- fulvous-pubescent, very young branchlets of current year’s growth ensely so, the terminal bud densely fulvous-pubescent. Leaves subcoria- ceous or coriaceous, ovate to oblong-ovate to oblong-elliptic, 8-13 cm. long, 4—5 cm. wide, acuminate at the apex, usually acute at the base, occasionally somewhat rounded, the upper surface shining, glabrous, lower surface paler in color, sparsely pubescent (especially on the midrib) and covered with minute dark dots, the margin glandular-denticulate, the veins 10-12 pairs, slightly conspicuous on both surfaces, the petiole 5-7 mm. long, flat above, rounded beneath, pubescent. Flowers axillary, solitary; pedicel up to 1! cm. long, appressed- -pilose; bracteoles 2, persistent, ba anos immediately below the calyx, pubescent, unequal, the smaller ca. 2 mm. long, acute, the larger more rounded; calyx-lobes 5, yppnealy cae sericeous, un- m. long at anthesis, broadly on the upper median portion of the external surface, otherwise glabrous; stamens seriate, ca. 40, adnate to the base of the corolla, the hia united for the most part, ca. 2 mm. long, glabrous, the anthers ca. 3 m long, hirsute, the apicule 1+ mm. long; ovary depressed- hemiopherie, appressed-sericeous, 5-celled, multi-ovulate, the style sericeous. Frui appressed-pubescent, 1.75 cm. diam. (fide Dy er), the seeds small, many. Authentic material, Wallich 3663, 3667, original collections of the species loaned from Kew, were used in drawing up the above description. The synonymy cited above is borrowed from Anderson in Hooker f., Fl. British India. No opportunity has been afforded me to actually study the speci- mens cited as synonyms. They are listed here in compilation. Characters helpful in determination are: 1 ulvous-pubescent branchlets; (2) short, recurved pedicels, seldom over 1 cm. long; ) 76 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII persistent bracteoles, unequal and opposite; (4) calyx-lobes densely seri- ceous, the two outer lobes often appearing as a single huge bract; (5) short corolla-lobes, sericeous near the apex of the median portion of the external surface; (6) glabrous filaments and oblong short-hirsute anthers, the apicule over 1 cm. long; and (7) ovary and fruit 5-celled, densely sericeous. 49. Adinandra collina, sp. nov. Arbor parva; ramulis teretibus, glabris, griseo-brunneis, gemmis termin- alibus elongatis, glabris (costis marginibusque exceptis). Folia coriacea, elliptico-oblonga, 8-10 cm. longa, 3—4.5 cm. lata, apice subrotundata abrupte acuminata, basi cuneata, supra glabra, subtus glabra costa excepta, costa supra plana, subtus elevata, glabra vel leviter pubescente, ut videtur bisulcata, margine subintegra, minutissime serrulato-crenulata, nervis undique ane petiolis glabris, 6-8 mm. longis. Flores axillares, solitarii vel bini; pedicellis glabris, circa 1 cm. longis, apice recurvatis; bracteolis 2, persistentibus, glabris rare leviter pubescentibus ciliolatisque, phage eaten alternatis, inferiore minoreque late ovata, circa 2 mm. longa 2.5 mm. lata, superiore maioreque subrotundata, circa 3 mm. longa et 2. 5 mm. lata; sepalis 5, imbricatis, glabris, canary: inaequalibus, exteriori- bus 2 late ovatis, circa 4 mm. longis et 5 mm. latis, interioribus 3 subro- tundatis, 5-6 mm. longis et 5-7 mm. latis: petalis 5, basi connatis, oblongo- obovatis, glabris, inaequalibus, 16-18 mm. longis, 5—7 mm. latis, (aie rotundatis, abrupte apiculatis, apiculis circa 2 mm. longis; staminibus ca 30, ?-seriatis, inaequalibus, 12-15 mm. longis, filamentis 7—9 mm. ee basi breviter connatis et petalis brevissime vel vix adnatis, ad apicem accrescentibus, inferne nudis, superne pubescentibus, antheris oblongis, albido-hirsutis, 4-5 mm. longis, apiculis glabris, circa 0.5—1 mm. longis; ovario subgloboso, glabro, circa 4 mm. diametro, apice abrupte attenuato, 5-loculari, multi-ovulato, stylo glabro, elongato, circa 1 mm. longo, stigmate minuto. Capsula subglobosa, glabra, circa 1 cm. diametro, 5-locularis, multi-seminata; semina discoidea, castanea, nitidula, circa 2 mm. diametro. DistRIBUTION: Sarawak, British North Borneo, Borneo. SARAWAK: Kuching, G. D. & C. Haviland 852 (typr, G), Dec. 1892.— Mount Matang, near bungalow, J. & M.S. Clemens 20986 (7447) (A), ia 24, 1929 (tree 12 m. high; flower-buds cream- phage co oe (Native Collector) 136 (A, G, US), 1708 Meets US), 7 (A, G), 5135 (A , US). oe baie BorRNEO: 0. m., Gr eaters “(British - oeiey For. Dept. no. 3938) ‘(A 1934 (tree 10 m. high; fruit green). — H. Low s. n. (G).— Mt. Kinabalu, Dallas, alt. 1000 m., J. & M.S. Clemens 26853 (A), Oct. 23, 1931 (flowers cream-colored; fruit reen). Borneo: Sanggouw, H. Hallier 950 (NY), 1893-94. — Boorbeo aston Borneo), West Koetai, alt. 30 m., Neth. Ind. For. Serv. b.b. 16153 (A), .-Dec. 1931 In many characteristics, A. collina resembles A. verrucosa Stapf, to which it is closely related. oth have bisulcate midribs on the under surface of the leaf, the more noticeable in A. verrucosa. Both species are quite glabrous and resemble A. dumosa Jack in this respect. However, the two Bornean species have pubescence on the young terminal buds whereas in A. dumosa the terminal bud is always strictly glabrous. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 77 In A. verrucosa the leaves are larger, coarser, obtuse and bluntly emar- ginate at the apex, the midrib beneath is always glabrous. In A. collina the leaves are acuminate at the apex and a fine pubescence is found along the midrib on the under surface. The branchlets in the former species are very thick, measuring up to 4 and 5 mm. in diameter near the apex. In A. collina the branchlets are smooth, terete and measure only 1—2 mm. in diameter. The petiole in A. verrucosa is short and sturdy and when break- ing off from the branchlets leaves a large scar. The pedicels are very short and thick, seldom measuring over 5 mm. long, and often measuring the same in thickness. In A. collina the pedicels measure up to 2 cm. or more (1 cm. in type), and are comparatively slender. The corolla-lobes of A. verrucosa measure 10-12 mm. in width while in those of A. collina the measurement is 6-8 mm. in width. Adinandra verrucosa, to date, has been found only in the “upper-mountain” zone (6000-10500 ft.), of Mt. Kinabalu, as defined by Stapf, whereas 4. collina is found at low altitudes of less than 1000 ft. in the “hill-zone.” In the latter species, the majority of specimens cited were collected in Sarawak. In A. dumosa, as far as I have seen, there is no tendency toward pubes- cence. The terminal bud is strictly glabrous. The terminal bud seems to be a criterion of this variable tendency toward pubescence. In species where the terminal bud is pubescent, especially when the pubescence is of the appressed nature, there is a distinct tendency toward glabrescence. As a result, one may find varying degrees of pubescence on the leaves and floral parts, making it quite difficult to use the character of pubescence in a key or as a delimiting factor. This same situation occurs in A. collina. The terminal leaf-bud is quite glabrous, yet one can find pubescence along the bisulcate midrib (which character is very obvious in the unfolding leaves) and along the margin. In the type cited above the specimen is quite glabrous and one must look most carefully to see evidences of pubescence. In other collections of the species, one may find distinct pubescence on the lower surface of the leaves and on the pedicels, bracteoles and calyx-lobes. VERNACULAR NAME: Bangkao (Dusun). 50. Adinandra Clemensiae, sp. nov. Arbor magna; ramis ramulisque teretibus, glabris brunneis vel griseo- brunneis, gemmis terminalibus minutis, conicis, sericeis. Folia chartacea vel subcoriacea, subrotundata, obovata vel elliptico-oblonga, 3.5—5 2-2.5 cm. (6-7 X 3-4.5 cm.), apice subrotundata, obtusata vel late obovata, basi late cuneata vel acuta, glabra, subtus fusco-punctata, margine subintegra, cee aetlpeiee venis supra obscuris, subtus 6—8 lateralibus pri- mariis paribus intra inem anastomosantibus, petiolis glabris, 6-9 mm. longis. Flores er ees solitarii; pedicellis glabris, ca. 1 cm. longis, recur- vatis; bracteolis 2, oppositis, ‘persistentibus, glabris, juventute dense pubescentibus, inaequalibus, minore ca. 2 2.5 mm., apice rotundato, majore ca. 2 X 5 mm., apice rotundato; sepalis 5, imbricatis, pubescenti- gis ede subellipticis, ca. 7 mm. longis, 5-6 mm. latis, margine 78 JOURNAL OF THE ARNOLD ARBORETUM [voL. XXVIII scariosis, Ciliolatis; petalis 5, albis, basi ad 5 mm. connatis, glabris, obovato- oblongis, 15-16 mm. longis, ca. 5 mm. latis, apice obovatis en eget eats 25, 3-seriatis, 9-11 mm. longis, filamentis 5-7 m longis, basi ad corollam adnatis, integre connatis, iitsutis, “nterioribus geniculatis, antheris oblongis, ca. 4 mm. longis, hirsutis, non apiculatis; ovario subgloboso, dense sericeo, circiter 3 mm. diametro, ( loculari, loculis multi-ovulatis, stvlo circiter 14 mm. longo, glabro, basi sericeo. Fructus (J. & A. Clemens 31385) subglobosus, sericeus, ca. 10 mm. longus et 8 mm. diametro. 5-locularis, multi-seminatus; semina typica, nitida, ca. 2.5 mm. diametro. Distrrpution: British North Borneo (Mt. Britis Nortu Borneo: Mt. Kinabalu: Clemens 28453 (A), Feb. 7, 1932 Lee cream tingec tis: staminibus ca. Kinabalu), Tenompak, alt. 1650 m., J. & 1 with purple). - fe etioiees (buds green; flowers cream alt. 1650 m., J. & M.S. Clemens 28678 , March 2, 1932 tinged with pink, centers brown) Aaah alt. 1000 m., : c& M.S. Clemens 29359 (flower- ide green tinged with purple). — Tenompak, alt. 1650 (flowers cream tinged with purple). (A), April 28. 1932 J. & M.S. Clemens 30193 (A), Feb. 17, 1932 y M.S. Clemens 30194 (A), April 28, 1932.— Tenompak, \), May 3, 1932 (flowers green with cream- J. & M.S. Clemens 31385 (A), Fet 5 ts & M.S. Clemens 30201 (: &M. S. pe eta h Ne we 1650 m., -~ Penibukan, alt. 1350-1650 m., 1933 (epee: 7m. high). Colombon River, alt. 1850 m., J. & (A), June 16, 1933 (fower- bud pinkish). — Marai Parai, alt. vos m., J. & M.S. y 3, 193 -Colombon River, alt. 2350 m., J. & M.S. Clemens (ower- buds dull pink; fruit ee ae Colombon River, Yemens 33906 (A), July 12, 1933.— Colombon River, S. Clemens 34090 (A), July 19, 1933 (flowers with pink tinge, pistil white). — Colombon River, alt. 1850 m., J. & J 10, Sage ee white pecmming dull sain’, Colombon River basin, alt. 1500 m., J pediaaaed 40122 (A), Aug. 24, 1933 (flowers w ap , the sepals and sulbers Rares ers nibukan, jungle, alt. m., J. & M.S. Clemens 50516 (A), Nov. 3, 1933 (“big at style long; fruit gray). —~Gurulau, ae of spur, alt. 1650 m., J. & M.S. Clemens 50692 (A), Dec. 4, 1933 5 high; style elongated; fruit green).— Gurulau spur, mossy forest, alt. 7 Dec. 12, 1933 (small tree with gray fruit). (tree 25 m. hig 2650 m., J. & M.S. Clemens 50979 (A), From the collections made on Mt. Kinabalu in British North Borneo by J. & M.S. Clemens, this species seems to be the most widespread in that area. Collections were obtained from Dallas, at 3000 ft. altitude, Tenompak, Penibukan, Marai Parai and Colomb altitude and Gurulau spur at 5000 and 8000 ft. altitu were collected at approximately 5000 ft. altitude with the exception of tw from Dallas at 3000 ft. and one from Gurulau spur at 8000 ft. In most species of the genus the flower is white. In this species is found consid- erable variation in color ranging from a purple-tinge through salmon and pink to pure white. Color in the stamens is also mentione e name A. Clemensiae is selected in honor of Mrs. M. S. Cle mens who not only collected all the specimens cited for this species but many more specimens representative of various other species in this genus. he ovary in the type-specimen, J. & MZ. S. Clemens 34480 is more often four-celled than five-celled. At least, this was the case in the several dissections made for this study. However, a single dissection from Clemens 40122, collected in the same general jase showed a five-celled alt. 1850 m., J. & the anthers aes ee Clemens 34480 (TYPE, A), Aug. 1947 | KOBUSKI, STUDIES IN THE THEACEAE, XV 79 ovary. These two specimens, of all those cited above, are the only two in the flowering stage. The majority of specimens, oeale the same species, are in the bud stage. These buds, although sufficient for specific determination, are very unsatisfactory for ovary-dissections. Two speci- mens J. & M.S. Clemens 50156 and 50979 are in the fruiting stage. From no. 50156, three fruits have been laboriously sectioned from apex to base. Although to all appearances these fruits are normal, they proved to be solid and woody throughout with only occasional slits where cells might have been expected. In J. & M.S. Clemens 50979 the fruit is found to be distinctly three-celled with few developed seeds. In this specimen the walls between the cells are very thick. Considering the fact that in most fruits the cell-walls are so fragile that it is often difficult to ovserve them with certainty, this thick-walled condition is most unusual. One may assume from these observations that 4. Clemensiae may have an ovary with the number of cells varying from ae to five. The third fruiting specimen (J. & M. S. Clemens 31385) shows clearly a five-celled fruit with many seeds oo typical of the genus. In this specimen the cell-walls are thinner than those found in J. & M. S. Clemens 50979. The small, thin, somewhat rounded leaves with long petioles present a character which makes this species easily recognizable 51. Adinandra magniflora, sp. nov. Arbor parva; ramis et ramulis teretibus, glabris, brunneis, innovationibus glabrescentibus vel leviter pubesce entibus, gemmis terminalibus conicis, adpresso- -pubescentibus. Folia coriacea, elliptica, 6-13 cm. longa, 4—7 c lata, apice abrupte lateque acuminata, basi late cuneata vel ee. supra glabra, subtus fusco-punctata, glabra vel leviter pubescentia, margine integra, venis supra obscuris, subtus lateralibus primariis 8-10 paribus prominentibus, margine anastomosantibus arcuantibusque, venis secun- dariis inter primarias frequentibus, petiolis crassis, 2~4 mm. longis, glabris. Flores axillares, solitarii; pedicellis teretibus, crassis, 1-1.5 cm. longis, adpresso-pubescentibus; bracteolis 2, aaa subrotundatis, adpresso- pubescentibus, inaequalibus, circiter 5 x 6mm. et 6 X 8 mm.; sepalis 5, imbricatis, subrotundatis, inaequalibus, 10 mm. longis, 9-11 mm. latis, interioribus margine scariosis; petalis 6, albidis, imbricatis, basi connatis, glabris, circiter 20 mm. longis, 8-10 mm. latis, apice aces vel subro- tundatis; staminibus circiter 45, ut videtur 4-seriatis, 10-1 m. longis. filamentis ad basim liberis, basi ‘ad corollam adnatis, een 6-10 mm. longis, interioribus ad apicem accrescentibus, exterioribus gracilibus r subgloboso, sericeo, 5-sulcato, 5-loculari, multi- ovulato, stvlo gracile, glabro, circiter 15 mm. longo. Fructus elongato-globosus, sericeus, circiter 12 mm. longus, 8 mm. diametro, 5-locularis, multi-seminatus; semina rubra, minuta, 1.5 mm. vel minus longa et 1 mm. diametro. DistriBuTION: British North Borneo. British Norti Borneo: Mt. Kinabalu: Tenompak, alt. 1650 m. J. & M.S. Clemens 29596 (A), May 4, 1932 (flowers cream-colored, the buds green). — Peni- te canyon west of jungle, alt. 1350 m., J. & M.S. Clemens 32040 (A), March 10, 3. — Head of Dahobang, alt. 2000- 2300 m.,J.& M eee lly ae (tyPE, A), Apri 23, 1933 (flowers white). — Colombon River r, alt. Om., J. & M. S. Clemens 2 (A), Aug. 10, 1933 (flowers ile with bright ey ee: 80 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII When present, the open flowers of this species furnish the most outstand- ing character. The corolla is large, measuring 3 cm. across, white in color, glabrous, and the six individual lobes are 2 cm. in length. ‘This is the only instance of a corolla with six lobes. Only two flowers were observed in detail, one each from J. & M. S. Clemens es the type, and J. & M.S. Clemens 34482. In both cases six lobes were observed. Whether this number is uniform is questionable! According to the collectors, the stamens are “brick pink.” In this species and in A. impressa it is difficult to ascertain the number of cells in the fruit. In both species there is a similarity in the minute red seeds, which are so crowded in the fruit that they have penetrated the central axis, and the cell-walls appear to have been broken from the sides of the fruit. Also, in both cases, the placentae of the ovary (where it is easy to distinguish the five cells) appear blunt, with the many ovules nearly completely encircling the placentae. In keeping with the size of the other parts of the flower, the stamens are long. The filaments are free from each other their entire length, except at their base where they are adnate to the base of the corolla. The filaments of the inner or shorter stamens are somewhat geniculate and enlarged at their juncture with the anthers. owever, in the longer or outer stamens, the filaments are very straight and uniform in width. Cited here dubiously are J. & M.S. Clemens 32610 and 51673 collected on Mt. Kinabalu. Both specimens appear to belong to this species. However, both are specimens with smaller leaves and both are incomplete, making a true association uncertain. 52. Adinandra plagiobasis Airy-Shaw in Kew Bull. Misc. Inform. 1939: 504. 1939. DistRIBUTION: Sarawak. wak: Mt. Dulit (Ulu pales near Long Kapa, on side of steep ridge in primary rain-forest, alt. 500 m., P. W. naiige 1191 (isotype, A) Aug, 10, 1932 (tree 90 m. high, circumference 1.7 m. at 1.5 m. from ground, no buttresses, bark about 6 mm. thick, reddish, smooth with very numerous lenticels arranged in transverse rows, wood pale yellow; leaves gland-dotted). Tree 27 m. high; branchlets terete, somewhat rugulose, brown, with a thin but dense pubescence, the individual hairs measuring up to : 5 mm. in length, the terminal bud with the same pubescence. Leaves subcoria- long, 2-3.5 cm. wide, acuminate at the apex, oblique-cordate at the base, always glabrous above, spreading pubescent beneath, densely pubescent on the younger leaves (as on the young branchlets), concentrated on the rief, 3 mm. or ree long, the pubescence continued fro om the midrib Flowers axillary, solitary; se ae usually curved, 1.5—2 long, thickened toward the apex, the pubescence similar to that of the branchlets; bracteoles 2, caducous, alternate, 3-5 mm. distant from calyx; calyx- lobes 5, imbricate, ovate, ca. 10 mm. long, 6-7 mm. wide, the outer sepals laxly pubescent, the inner sepals more densely so; corolla-lobes 5, connate at the base, obovate, ca. 13 mm. long (2 cm., fide Airy-Shaw), ca. 10 mm. wide, 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 81 sericeous on the middle portion of the external surface, glabrous within; stamens nearly 50, in series, the filaments varying in length, 2-3 mm. long, sericeous on the external surface except on the lower portion where they are adnate to the corolla-lobes, the anthers 4-6 mm. long, sericeous; ovary at first truncate then tapering at the apex rather than conical, sericeous, ca. 3 mm. diam., 5-celled, multi-ovulate, the style 8-9 mm. long, sericeous at the ee otherwise glabrous. Fruit not seen The oblique, cordate base of the leaves is an unusual character in this genus and the most distinguishing feature of the species. In this character and in the leaf-texture is found a striking resemblance to the species of the American genus Freziera. The leaves of A. cordifolia Ridley, from Sarawak and British North Borneo, also have a cordate base. However, as Airy-Shaw states in his presentation, the true relationship of A. plaeto- basis is rather with A. villosa Choisy and A. coarctata Craib, which agree in indumentum and general leaf-outline. The indumentum in this species is unusual in that it is longer and more lax than the usual appressed type of pubescence found in the majority of species of the genus. My material of Richards 1191 is probably less copious than that from which Airy-Shaw drew his description. In the material examined here, the corolla-lobes measured only 13 mm. in length. Those described from the same number by Airy-Shaw measured 20 mm. long. I depended on a single flower for dissection, which, however, was in excellent condition. VERNACULAR NAME: Pungo (Laban). 53. Adinandra borneensis, sp. nov. Arbor ?; ramis teretibus, verrucosis, glabrescentibus, rubro-brunneis, ramulis teretibus, verrucosis, luteo-adpresso-pubescentibus, gemmis ter- minalibus dense luteo-adpresso-pubescentibus. Folia coriacea, elliptica vel obovata, 10-16 cm. longa, 4.5-6 cm. lata, apice abrupte acuminata vel subrotundata, basi cuneata, supra glabra, subtus adpresso-pubescentia, fusco- -punctata, margine integra, venis supra obscura, subtus lateralibus primariis 20-+- paribus, vix distinctis vel obscuris, petiolis 3-5 mm. longis, subtus pubescentibus. Flores pnote solitaril; pedicellis circiter 2 mm longis, adpresso-pubescentibus; bracteolis 2, alternatis, cito caducis; sepalis 5, imbricatis, circiter 10 mm. longis, cat bus exterioribus duobus subrotundatis, 9-10 mm. latis, omnino‘adpresso-pubescentibus, interioribus tribus ovatioribus, apice subrotundatis, 8— ‘9 mm. latis, adpresso-pubescenti- bus margine scariosis glabris; petalis 5, basi connatis, circiter 13 mm. longis, 9 mm. latis, omnino dense sericeis, apice rotundatis: staminibus 45-50, 4- vel 5- seriatis, 7-11 mm. longis, filamentis connatis, basi ad corollam adnatis, hirsutis, interioribus ad apicem geniculatis vel accrescen- tibus, exterioribus gracilibus, antheris elongatis, hirsutis, apiculatis, ovario sericeo, subgloboso, 5-loculari, multi-ovulato, stylo 5 mm. longo, glabro, stigmatibus 5-lobis. Fructus non visi. DistrIBUTION: Netherlands East Indies (Southeast Borneo). SOUTHEAST Borneo: Above oe Hitaja, at sea level, Neth. Ind. For. Serv. b.b. 20641 (TYPE, K), Feb. 1, — Above Makaham [sic], Long Tccknens: alt. 35 m., Neth. Ind. For. Sore eS ee (A), Jan. 23, 1936 This species is most closely related to another Bornean species, A. 82 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII myrioneura Kob., which has been collected from an approximate locality. Both have leaves with more than twenty pairs of primary veins. In the leaves of A. myrioneura the veins are very distinct on the under surface and rise at nearly right angles from the midrib. In A. borneensis the veins are quite indistinct on both surfaces and rise at a less erect angle from the midrib so, A. myrioneura has a four-celled ovary, the style is pubescent, the corolla glabrous and the leaves are more chartaceous. In A. borneensis, the ovary is five-celled, the style glabrous, the corolla pubescent, and the leaves are coriaceous. In the ovary of both species the placentae are split. In 4. borneensts the outer cell-wall projects in between the branched placentae and extends nearly to the central axis giving the appearance of ten cells, in some instances, Cited above is a sterile specimen collected by the Neth. Ind. For. Serv. (20601). The — are larger than those of the type. From general appearances this specimen agrees with the type and is mentioned in this paper for herbaria sehen perhaps the specimen may be in the flowering or fruiting stage. 54. a dasyantha Korthals, Verh. Nat. Gesch. Bot. ed. Temminck, 108. 1840. — Choisy in Mém. Soc. Phys. Hist. Nat. Geneve, 1]: 112 ees Ternstr. 24). 1855. oe Miquel, FI. Ned. Ind. 1 (2):477. 1859; Ann. Mus. Bot. Lugd.-Bat. 4: 103. 1868. — Szyszylowicz in Nat. ere Ill. 6: 189. 1893. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 192 DistRIBUTION: Netherlands East Indies Sie. Sumatra: Exact locality lacking, P. W. Korthals s. n. (probable isorypes, G, NY). — Tapanoeli, Angkola in Sipirok, aay alt. 500 m., Neth. Ind. For. Serv. b.b. 28194 (A), June 21, 1939, — ie = ast, Asahan, vicinity of Loemban Ria, Ra/mat Si Boeea 7607 (US), Feb.-April 1 Tree 8 m. high; branches erect, open, terete, pale gray, glabrous; branchlets terete, densely pubescent, terminal bud densely sericeous. Leaves coriaceous, pues oblong- sere occasionally obovate, somewhat acute at the apex, acute at the base, ca. 10-13 cm. long, 3-4 cm. wide, the margin entire or beste peg ‘glabrous above, densely pubescent along the midrib beneath, the young leaves pubescent beneath, the petiole ca. 5 mm. long, somewhat terete, flattened above, pubescent. Flowers axillary, solitary; peduncles terete, ca. 1 cm. long, thickened near the apex, pubescent; bracteoles 2, alternate near top of pedicel, caducous, pubescent; calyx-lobes 5, smbricate, subequal, rotund, ca. 6 mm. long and wide, pubescent on the external surface, membranaceous and glabrous at the margin, ciliolate; pene lobes 5, obovate, obtuse, densely sericeous on the external surface, glabrous within, fleshy, white, ca. 14 mm. long, mm. wide; stamens in series, numero us: filaments joined at the base, partly free, linear, glabrous, the anthers oblong, acute, densely hirsute; ovary hemispherical-conical, pubescent, attenuate at the apex into a glabrous terete style; stigma truncate. Fruit glabrous, somewhat meses attenuate at the apex into the persistent glabrous style, 5-celled, many-s d. Judging from the material available for study, this species appears very 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 83 rare and seems confined to the island of Sumatra. Three probable isotypes collected by Korthals in Sumatra, in addition to a single specimen collected by Rahmat Si Boeea in the East Coast region of Sumatra constitute the total material for this study. Characters helpful in determination of the species are: (1) glabrous ovary and style; (2) glabrous filaments; (3) dense sericeous pubescence of the corolla-lobes; (4) caducous bracteoles; and (5) pubescent terminal buds, branchlets and midrib of the under surface of the leaves. VERNACULAR NAME: Kajoe api-api. 55. Adinandra myrioneura, sp. nov. Arbor 12 m. alta; ramis teretibus, lenticellatis, griseo- -brunneis, glabres- centibus, ramulis juventutibus teretibus, sericeis, lenticellatis, gemmis terminalibus dense luteo-ferrugineis. — chartacea, oblongo-ovata vel oblongo-elliptica, 10-13 cm. longa, 3.5 m. lata, apice late acuta, abrupte obtuseque acuminata, basi late pa or subrotundata, supra glabra, subtus leviter pubescentia vel glabrescentia, margine integra vel subintegra, venis supra subobscuris, subtus prominentibus, lateralibus primariis circiter 25 paribus, a costa subperpendiculare extendentibus, ad marginem evanes- centibus, petiolis pubescentibus, circiter 5 mm. longis. Flores axillares, solitarii; pedicellis 6-7 mm. longis, adpresso-pubescentibus; bracteolis 2, alternatis, cito caducis; sepalis 5, imbricatis, subrotundatis, subaequalibus, exteriore adpresso-sericeis, interiore glabris, exterioribus 5-6 mm. longis, 6—7 mm. latis, interioribus leviter latioribus, circiter 6 mm. longis, 8-9 mm latis, margine scariosis; petalis 5, basi connatis, albidis, late obovatis, circiter 10 mm. longis, 9 mm. latis, apice rotundatis, exteriore omnino adpresso-sericeis; staminibus circiter 45, 4- vel 5-seriatis, linearibus, 6-10 mm. longis, filamentis basi ad corollam adnatis, omnino connati is, 3- 6 mm, longis, dense hirsutis, medio accrescentibus, antheris exteriore dense hirsutis, ie a circiter 4 mm. longis, ovario subgloboso, sulcato, sericeo, circiter 4 diametro, 4-loculari, multi-ovulato, placentis ramificatis, stylo gene 5 mm. longo, glabro basi excepto. Fructus non visi. DistrRIBuTION: British North Borneo. tish Nortn Borneo: Tambato, Tambunan, plain, alt. 400 m., Puasa-Angian mar a Borneo Forestry Dept. no. 3885) (Type, A, K), Feb. 28, 1934 (tree 12 m. high with white flowers). Adinandra myrioneura is outstanding because of the many primary veins on the under surface of the leaf extending nearly at right angles from the. midrib, at first very distinct but gradually fading towards the margin until nearly obscure. The bracteoles are alternate and quickly caducous, dropping before anthesis. The corolla-lobes are distinct in that the sericeous pubescence is found over the entire exterior surface rather than confined to a median portion as in the majority of the species in this genus. The stamens are many, the filaments closely joined for their entire length, densely pubescent on the exterior surface, and swollen to the width of the anther at the apex. The four-celled ovary is also an unusual feature. Only three dissections were made, but all showed true 4-celled ovaries, in which the placentae were clearly branched close to the axis. The cell-wall was deeply indented opposite the point of branching. VERNACULAR NAME: Bangkau (Dusun). 84 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII 56. Adinandra javanica Choisy in Zollinger, Syst. Verz. Ind. Archip. 143, 146. 1854; in ( 855. — . Soc. Phys. Hist. Nat. Genéve 1:112 (Mém. Ternstr. Miquel, Fl. Ned. Ind. 1 (2): 477. 1859. — Seyszvlowiex in Nat. Pflanzenfam., III. 6: 189. 1893. — Koorders & Valeton, Meded. ’S Lands. P1. 16: 224 (Bijdr. Boom- soort. pnts 3: 224). 1896. — Melchior in Nat. Pflanzenfam. ed. 2, 21: 144. 1925. itil tion: Netherlands East Indies (Java, Lesser Sunda Islands). JAVA 3esooki, Jang Plateau West, Taman Hidoep., alt. 1900 m., C. G. G. J. van Steenis ae (A), July 13, 1938. Lesser SuNDA ISLANDs: ins. De Voogd Small tree 10-12 m., branching low, near the ground (fide Koorders) ; branchlets terete, sericeous, becoming glabrous: the terminal buds densely surface, along the midrib when mature, often glabrous, the petiole semi- terete, 4-5 mm. long, flat on the upper surface, pubescent beneath. Flowers axillary; pedicels up to 2 cm. long, recurved, pubescent, often lightly so and near the calyx-lobes; bracteoles 2, caducous, ovate-lanceolate, 4—5 mm. long when present (fide Koorders) , alternate, the lower bracteole as much as 5 mm. below the calyx-lobes; calyx-lobes 5, imbricate, subequal, very obtuse, 8—9 mm. long, densely appressed- sericeous, the narrow margin me embranaceous; corolla-lobes not seen, densely sericeous on the outer surface (fide Koorders); stamens (not seen) ca. 40, connate at the base, the filaments glabrous, the anthers sericeous; ovary nde nsely sericeous, somewhat conical, 3-celled, tapering into a densely sericeous style. Mature fruit not seen, immature fruit (ca. 1 cm. diam.) densely sericeous, 3-celled, multi-ovulate, the style densely sericeous, neither style nor fruit showing evidence of glabrescence. The closest relative to A. javanica Choisy is A. dasyantha Korthals. Both are characterized by pubescent calyx-lobes, corolla-lobes and ovary. Also both have alternate, deciduous bracteoles. However, A. javanica has a densely sericeous style, and the ovary and fruit are distinctly three-celled. On the other re A. dasyantha differs in the glabrous style, the five-celled ovary and frui In most alae and Netherlands East Indian species the presence of a three-celled ovary and fruit presages few ovules and even fewer seeds. Adinandra javanica appears to be one of the exceptions. The immature fruit examined shows that the developing seeds are minute and apparently many in number. 57. Adinandra lutescens Craib in Kew Bull. Misc. Inform, 1925: 19. 1925; Fl. Siam, Enum, | 1925 DIstTRIBUTION: Siam, so Malay States (Kedah). S1AM: Pattani, alt. 50 m., . G. Kerr 7828 (type, K), Sept. 12, 1923 (shrub or — tree up to 5 m. hi i Patan Banang Sta., evergreen ne alt. 50 m . G. Kerr 7274 (K), July 22, 1923 (small tree, 2 ne high). yaa Gunong Jerai, M. “th 0044 (K), Dec. 1 Shrub or small tree 5—8 m. high; branches terete, gray-brown, glabrous, the young branchlets appressed-pubescent, the terminal buds fulvous- sericeous. Leaves coriaceous or subcoriaceous, oblong-oblanceolate, 5—11 m. long, 2.5—4 cm. wide, occasionally asymmetrical, obtusely acuminate 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 85 at the apex, cuneate at the base, glabrous, darker above, paler, lightly pubescent beneath, eventually glabrous, the margin lightly denticulate, the veins 10-12 pairs, anastomosing near the margin, the petiole 3-5 mm. long, pubescent beneath, eventually glabrescent. Flowers axillary, solitary; pedicel slender, recurved, 1.5—2.5 cm. long, thickened towards the apex, appressed- -pubescent; bracteoles 2, quickly caducous, opposite, immediately below the calyx; calyx-lobes 5, imbricate, ovate, ca. 1 cm wide, unequal, appressed- -pubescent, the outer lobes broadly acute, glandular- -denticulate along the margin, the inner lobes more nearly rounded, the margin membranaceous, entire; corolla-lobes 5, connate at the base, lanceolate, ca. 1.5 cm. long, 0.5 em. wide, long- attenuate, tapering into an acuminate apex, sericeous on the median portion of the exterior surface; stamens in series, the filaments ca. 3 mm. long, adnate to the base of the corolla, joined at the base, sericeous on the apical half, the basal half glabrous, the anthers ca. 3 mm. long, densely hirsute, the apicule glabrous, 2-3 mm. long, equalling the anther-cells in length; ovary sericeous, 3-celled, multi-ovulate, tapering into the style, the style entire, sericeous diminishing in density toward the apex. Mature fruit not seen. ma- ture fruit somewhat globose, densely sericeous, 3-celled. The outstanding characters of this species are: (1) long-attenuate corolla-lobes, 1.5 cm. long, 0.5 cm. wide, tapering from near the base to an acuminate apex, the sericeous area tapering also; (2) stamens ca. 9 mm. long, with the filament, anther and apicule of about equal length, the apicule unusually long for the genus; (3) bracteoles quickly caducous; and (4) ovary 3-celled, sericeous. The closest relative is A. phlebophylla Hance. In the latter species the corolla-lobes are only 7-9 mm. long and, although they are acute at the apex, they do not taper gradually and continuously from the base to the apex; the stamens measure only 4—5 mm. long; the filaments are glabrous and less than a millimeter in length, the apicule rather inconspicuous; and the veins on the lower surface of the leaves tend to branch midway between the midrib and the margin. Craib records the ovary as two-celled. Only a single dissection was made in this study on a partially developed fruit. The results showed the young fruit to be clearly three-celled VERNACULAR NAME: Kandis Burong. 58. so beas oblonga Craib in Kew Bull. Misc. Inf. 1924: 88. 1924; Fl. Siam. Enum. 1925.— Gagnepain in Fl. Gén. Indo-Chine, Suppl. 1: 284. 1943. nine: Siam Stam: Pitsanulok, Sukotai, oe luang, open grassy Aare alt. 1100 m., A. F. G. Kerr 5923 (TyprE, K), May 2, 1922 (small tree about 4 h). Small tree ca. 4 m. high: young branchlets neice nee brown, lightly and finely appressed-pubescent, the terminal bud short, conical, appressed- pubescent. Leaves coriaceous, elliptic to oblong-elliptic, 5-9 cm. long, 2.5—4.5 cm. wide, obtuse or abruptly acuminate at the apex, rounded or cuneate at the base, glabrous above, finely appressed-pubescent beneath, especially along the midrib, the margin lightly revolute when dried, glandular-denticulate, the veins 15-20 pairs, prominent on both surfaces, anastomosing freely, the petiole ca. 5 mm. long, lightly appressed-pubescent. 86 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII Flowers axillary, solitary; pedicels ca. 2 cm. long, recurved, lightly appressed-pubescent; bracteoles 2, alternate or subopposite, quickly caducous; calyx-lobes 5, imbricate, 7-8 mm. long, 5-8 mm. wide, finely appressed-pubescent, unequal, the outer lobes smaller, ovate, the inner lobes more obovate, wider; corolla-lobes 5, connate at the base, ovate, acute at the apex, 10-11 mm. long, 5-6° mm. wide, pubescent over the entire exterior surface except along the margin; stamens ca. 35, seriate, 6-7 mm. long, the filaments adnate to the base of the corolla and joined to each other, unequal, 1-2 mm. long, pubescent on the upper third, otherwise glabrous, the anthers ca. 4-5 mm. long, oblong, hirsute, the apicule ca. | mm. long, often as long as the shorter filaments; ovary densely appressed- pubescent, tapering into a style densely pubescent at the base, less so near the apex, the stigma 4-parted, oblique. Fruit globose, ca. 1 cm, in diameter, appressed-pubescent, 4-celled, many-seeded. The distinguishing characters of this species useful in determination are: (1) ovary appressed-pubescent, 4-celled; (2) style densely pubescent at the base, less so towards the apex; (3) stigma 4-parted, oblique; (4) corolla-lobes ca. 1 cm. long, ovate, acute at the apex, pubescent over the entire exterior surface except along the margin; (5) stamens 6-7 mm. long, the filaments very short, 1-2 mm. long; and (6) leaves elliptic, small (up to 8 cm. long), the veins 15-20 pairs, conspicuous on both surfaces. This species, along with A. myrioneura Kob. of Borneo, has a 4-celled ovary. Just how consistent this character may prove to be will depend upon the examination of future collections. In other species, such as A. Macgregorii Merrill from the Philippine Islands and A. Clemensiae Kob. from British North Borneo, where a 4-celled ovary has been noted, the character was found to be inconsistent. In A. Clemensiae, where a 4-celled ovary occurred quite consistently in the type-specimen, more material was available for study, and the prevailing number of cells for the ovary appeared to be five. In A. oblonga the stigma is 4-parted and oblique. In A. Macgregorii the style was also 4-parted. Still, on the same specimen were found flowers in which the ovary was 3-celled. In its gross characters, this species resembles 4. parvifolia Ridley very closely, the leaves appearing almost identical. However, it can be separated by the longer pedicels, the smaller acute corolla-lobes which are pubescent over the entire external surface, the very short filaments of the stamens, and the 4-celled fruit and 4-parted stigma. 59. Adinandra coarctata Craib in Kew Bull. Misc. Inform, 1925: 18. 1925; Fl. Siam. E . 1:124. 1925. DistRIBUTION: Siam. Stam: Betong, Gunong Ina, evergreen forest, alt. 1200 m., A. F. G. Kerr 7559 (type, K), Aug. 10, 1923 (small tree ca. 4 m. high). Small tree ca. 4 m. high; branchlets gray-brown, terete, eventually gla- brous, younger branchlets tomentulose, very young branchlets of current year’s growth fulvous-pilose, the terminal buds densely fulvous-pilose. Leaves coriaceous, oblong-elliptic, 7-10 cm. long, 3-4 cm. wide, obtusely or abruptly acuminate at the apex, cuneate to somewhat rounded at the 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 87 base, glabrous above, lightly pubescent beneath, the pubescence more densely concentrated on the midrib and along the margin, conspicuously dark-dotted beneath, the margin revolute, lightly glandular- apiculate, this long, oe pubescent on aap Bae Flowers axillary, solitary; pedicels stout, recurved, ca. 5 long, fulvous-pilose; bracteoles 2, opposite, immediat ely below the fe quickly caducous, when present ovate, ca. 3.5 mm. long, 3 mm. wide, pilose, ciliolate; calyx-lobes 5, imbri- cate, unequal, the outer lobes broadly ovate, 9-10 mm. long, 8-9 mm. wide, obtuse, pilose, the inner lobes ca. 13 mm. long, pilose, the margins dentic- ulate; the ovary subconical, ca. 4 mm. long, densely sericeous, 5-celled, multi-ovulate, tapering at the apex into a sericeous style glabrous only at the apex Craib in his excellent description of A. coarctata mentions the ovary as being four-celled. I felt dubious about this character and sacrificed a developing fruit on the type-specimen to verify his finding. My single dissection showed four large well-developed cells with a fifth smaller, almost abortive cell, this latter containing developing ovules, however. Outstanding characters in this species are: (1) bracteoles quickly caducous, when present ovate, pilose, ciliolate, ca. 3.5 mm. long and 3 mm. wide; (2) inner calyx-lobes larger (ca. 13 mm. long) than the outer lobes, pilose, denticulate along the margin; (3) ovary densely sericeous, 5-celled, with one cell smaller than the others, multi-ovulate; (4) style entire, gen- erally sericeous, glabrous at the apex; (5) terminal buds densely fulvous- pilose; and (6) leaves with a generally thin pubescence on the under surface, the pubescence densely concentrated on the midrib and margin obscuring the glandular-apiculations along the margin. Although the corolla-lobes and stamens have not been described for this species, from the type and distribution of the pubescence one may expect the former to be pubescent. This species is so closely associated with A. glischroloma Handel- Mazzetti and its varieties of China and Indo-China that later collections may prove it to be no more than another variety in this complex. How- ever, until these further collections are made, it is best to continue the species in its present status. 60. oe ia i ag crear i" ae Bot. 14: 240. 1876. — Szyszylowicz in at nzenfam. II]. 6: 189. —Craib, Fl. Siam. Enum. 1: 124. 1925. — ain in FI. Gén Indo- Chine, Suppl. 1: 286. 1943. Adinandra integerrima ee Fl. F Cochinchine 2: t. 125. 1887. Non T. An- derson. — Pitard in Lecomte, FI. ‘Gén, Indo-Chine 1: 334. 1910. Adinandra_ integerrima a oe Y phlebophylla (Hance) Pierre, Fl. For. Cochinchine 2: t. 125 DISTRIBUTION: a a. ee Campopta: Prov. Tpong, Knang ite alt. 1500 m., LZ. Pierre 608 (ISOTYPES, A, M, NY), May ee (tree 8-15 m. high) Tree 8-15 m. high; branchlets terete, grayish brown, cle emg! ana cent when very young, later glabrous, the terminal buds sericeous. Leav coriaceous, oblong-elliptic, 8-12 cm. long, 3-4 cm. wide, obtusely eile 88 JOURNAL OF THE ARNOLD ARBORETUM [voL, XXVIII at the apex, acute at the base, glabrous above, sparsely appressed- pubescent beneath, especially along the midrib, the margin serrulate, the veins raised on both surfaces, aati midway between the midrib and the margin, the petiole ca. 5 mm. long, rounded, pubescent beneath, flat, glabrous above. Flowers oie. oe pedicels 2-5 cm. long, slender, graceful, swelling in diameter near the apex, appressed-pubescent at first, later glabrescent; bracteoles 2, caducous, oblong-lanceolate when present, 4 mm. long, 2 mm. wide, densely sericeous, usually close to the calyx; calyx-lobes 5, imbricate, minute at anthesis, unequal, ca. 7-8 mm. long, 5—6 mm. wide, appressed- -sericeous, especially the inner lobes, increasing at maturation of fruit to ca. 12 mm. long, 7 mm. wide, ovate, distinctly acute at the apex; corolla-lobes 5, cake unequal, 7-9 mm. long, 2-3 mm. wide, densely sericeous, acute at t th Ne apex; ee ca. 30, 2- or 3-seriate, minute, 4—5 mm. long, the filaments only ca. 0.5—0.8 mm. long, joined at the base, glabrous, the anthers oblong, very aay sericeous; ovary and style elongated, only 5 mm. long over all, densely sericeous at anthesis. Fruit globose, ca. 1 cm. diameter, densely sericeous, topped by a persistent sericeous style 1.5 cm. long. Seeds minute, dark brown. The outstanding characters of this species are: (1) corolla-lobes lightly sericeous, shorter (6-8 mm. long) than the cane (7-9 mm. long); (2) stamens minute, ca. 3-4 mm. long, the filaments glabrous, very short 0.5—0.8 mm. long), the anthers so densely sericeous that their outline is difficult to discern; (3) pedicels sturdy, up to 5 cm. long in fruit; (4) bracteoles caducous, when present 4 mm. long; and (5) ovary, fruit and style densely sericeous. From 1887 to 1925, this species had been considered synonymous with A, integerrima T. Anders. Pierre felt the two entities were identical and in his treatment of A. integerrima probably used the type of A. phlebophylla (Pierre 608) in describing and illustrating A. integerrima for his FI. For. Cochinchine (2: t. 125. 1887). It is safe to assume that this specimen provided the material for both his description and illustrations, since they agree perfectly. Craib (1925), was the first to draw attention again to A. phlebophylla and he returned it to its specific status. Adinandra phlebophylla is more closely allied to A. lutescens Craib than to A. integerrima. In leaf-characters these two species are almost iden- tical. However, Craib’s species has a much longer corolla (1.7 cm. long) and the filaments are approximately four times longer and more in propor- tion with the anthers than those in A. phlebophylla. The number of cells of the ovary is not known. Even though isotypes of the species were available for this study, the material was so poor that I was not able to make dissections which would add to this knowledge. The species is treated in this portion of the study rather than with the Chinese species since its relationships are with this group and also because of the former confusion of its association with A. integerrima. Adinandra villosa Choisy in Mém. Soc. Phys. Hist. Nat. Genéve 1:112 (Mém. rnstr. 24). 1855.— Dyer in Hooker f., Fl. Brit. Ind. 1: 283. 1874.— Kurz, For. Fl. Brit. Burma 1: 100. 1877. — King in Jour. As. Soc. Bengal 59 (2): 190 (Mater. Fl. Malay Penin. 1: 130). 1890.— Szyszylowicz in Nat. Pflanzenfam. III. 6: 189. 1893. — Ridley in Jour. Linn. Soc. Bot. 38: 304. 1908; Fl. Malay 1947 | KOBUSKI, STUDIES IN THE THEACEAE, XV 89 Penin. 1: 196. 1922. — Melchior in ae See ed. 2, 21: 144. 1925. Ternstroemia ? sericea Wallich, Cat. no Schima Wallichii Choisy in Mém. See. oe Hist. Nat. Geneve 1:179 (Mém. Ternstr. 91). 1855 Adinandra integerrima T. Anderson $ villosa (Choisy) Pierre, Fl. For. Cochinchine 22.125. 1887. DistripuTion: Malay Peninsula, Burma. ALAY PENINSULA: Penang, Penang Hill, alt. 670 m., E. J. H. Corner 31595 (A, K), July 21, 1936. — Perak, Larut, rocky localities in dense jungle, Dr. King’s Collector 6256 (K), July 1884 (tree 12-16 m. high; leaves glossy deep green; flowers light ae white with a silvery gloss; fruit silvery white). — Selangor, Sempang Mines, H. N. Ridley 15594 (K), April 1894 URMA: Tenasserim, Tavoy, NV. ‘Wallich 1454 (K). Tree 12-15 m. high; branchlets pilose, pale brown, the leaf-buds sericeous. Leaves coriaceous, oblong-lanceolate or elliptic, 10-14 cm. long, 3-5 cm. wide, shortly acuminate at the apex, rounded or broadly cuneate at the base, rarely acute, glabrous above, sparsely pubescent beneath, the margin entire or seage crenulate, the nerves 7—9 pairs, acai arching within the margin, not prominent, the petiole ca. 5 mm, long, pilose. Flowers axillary, serlly solitary; pedicel pilose ca. 1 cm. long, recurved, pilose; bracteoles 2, caducous, ovate, when present occurring from the middle of the pedicel upward; calyx-lobes 5, imbricate, nearly equal, broadly ovate to somewhat rounded, up to 10 mm. long and 8 mm. wide, pilose on the external surface; corolla-lobes 5, connate at the base, ovate, obtuse or rounded at the apex, densely sericeous on median portion of the external surface, glabrous at the margins; stamens ca. 30, adnate to the base of the corolla, the filaments short, sericeous; ovary depressed- hemispherical, ee sericeous, 5-celled, tapering at the apex into a densely sericeous style. Fruit 10-12 mm. diam., densely sericeous, 5-celled. Seeds numerous, minute, shiny, dark brown, ca. 1.5 mm. long. This species has been little understood and frequently misinterpreted. the Arnold Arboretum. However, I found myself unwilling to accept my own interpretation and made a inte appeal to Kew for a secondary loan of representative authentic material of this species. ‘To my surprise and satisfaction Corner 31595 was included in the loan along with Weallich 1454, the latter specimen probably the most renowned in this species. I was embarrassed when I realized that a psychological element had played an important part in my reluctance to accept the species at first. I mention this unscientific incident, since it may help to explain the dubious position the species has held in the past. Considering some of the very pubescent species such as A. glischroloma Handel-Mazzetti, I was very disappointed in the amount of pubescence on A. villosa when confronted with the actual type. I had expected a much more pubescent species. Although the name is appropriate, the species is hardly more villous than many other species in the same geographical locality. All the above cited specimens are in the fruiting stage; hence complete descriptions of corolla-lobes and stamens are still wanting. King in his “Material for a Malayan Flora” gives rather a complete description of the 90 JOURNAL OF THE ARNOLD ARBORETUM [voL, Xxvm species and this was used in my own description of corolla-lobes and stamens. Pierre, in his Fl. For. Cochinchine 2: t. 125. 1887, reduced this species to synonymy under A. integerrima T. Anders. along with 4. phlebophylla Hance. Pitard, in Lecomte, Fl. Gén. Indo-Chine 1: 334. 1910, followed Pierre in his treatment. However, Craib, Fl. Siam. Enum, 1: 125. 1931 retained all as separate entities. Adinandra bicuspidata, sp. nov. Arbor parva ad 10 m. alta; ramulis teretibus, brunneis, leviter lenticella- tis, glabrescentibus, hornotinis pubescentibus adpressis, Paka dense ferrugineo- pines Folia chartacea, subrotundata vel ovata, 7-12 cm. longa, 3—6 cm. lata, apice abrupte acuminata, basi late eu vel cuneata, supra ae abn glabrescentia et leviter punctata, margine integerrima, venis undique prominentibus reticulatisque, primartiis 10-12 paribus, intra marginem anastomosantibus arcuantibusque, venis secundariis brevioribus inter primarios frequentibus, petiolis ca. 1 cm. longis, supra planis glabrisque, subtus semi-teretibus, pubescentibus vel areas Flores axillares, solitarii; pedicellis adpresso-pubescentibus, 1-1. longis, apice subrecurvatis; bracteolis 2, oppositis, probabiliter si anthesin caducis, ovatis, 6-7 mm. longis, 3-4 mm, latis, sericeis; sepalis 5, imbricatis, inaequalibus, late ovatis, exterioribus ca. 9 mm. longis et 8 mm. latis, apice grees dense sericeis, interioribus minoribus; petalis (immaturis) 5, ad basim liberis vel subliberis, oblongo-ovatis, ca. 5 mm. longis et 2.5 mm. latis, dense sericeis margine exceptis: staminibus Cia: ca. 35, uniseriatis, ca. 2.5 mm. longis, filamentis ad basim liberis, ca. 0.75 mm. longis, glabris, antheris dense ferrugineo-sericeis, ob- longis, ok bicuspidatis, loculis nceiiiee ultra connectivum projectis; ovario (immaturo) subplano, densissime ferrugineo-sericeo, loculorum numero ignoto, stvlo (immaturo) integro, ca. 4 mm. longo, densissime omnino apice incluso ferrugineo-sericeo. Fructus ignotus DistripuTION: Malay Peninsula (Penang). PENANG: Government. Hill, alt. 650-700 m., C. Curtis 2241 (K, Type), March 1890 fecal ve 25 ft. high). = Because of the incomplete material in the type specimen I was at first unwilling to describe this new species. However, after careful study it appears to be distinct from the other species in its geographical area. The material, even though incomplete, is sufficient to permit a technical description, and hence I finally have decided to include it in this study. The name A. bicuspidata is derived from the bicuspidate appearance of the anthers. In some species studied, the connective projects beyond the anther cells, causing anthers to appear apiculate. In other species, the anther cells are flush with the connective, causing the anthers to appear truncate. In this new species, the anther cells project sharply above the connective, giving a bicuspidate appearance to the anthers Other characteristics which lend distinction to this species are: (1) the chartaceous leaves with the veins distinctly reticulate on both surfaces; and (2) the long ferrugineous pubescence of the anthers, ovary and style. The pubescence of the style is further unusual in that it extends the entire length of the style, obscuring the stigma. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 91 In the type specimen the flowers were immature, and hence the measure- ments of most floral parts cited above are subject to change when more complete material is available. It appears that this species may be very rare, since it has been collected only once, and that nearly sixty years ago. Unfortunately, the true number of cells in the ovary could not be ascertained. Although two dissections had been made, in both cases the cell-walls of the ovary were too indistinct to permit one to say with assurance just what their number might be. am inclined to feel that the ovary is five-celled, but of that I am not certain. Because of the uncertainty of the cell number in the ovary, this species cannot be included in the key and is placed at the end of the Indo-Malaysian species. At Kew, this species had been associated with A. villosa Choisy. How- ever, the chartaceous, reticulate leaves, the bicuspidate anthers, the glabrous filaments and the opposite bracteoles all distinguish A. bicuspidata from the older species. LITTLE-KNOWN SPECIES eines Brefeldii pail in ens 3 5 ands. Plant. 19: 349, 642. 1898. — Melchior in Nat. Pflanzenfam. ed. 2, 21: ete aio heen ba pie i rae CrLeBes: [Cited here dubiously] En Ond. Makale-Rantepao, Paniki, alt. 300 m., Neth. Ind. For. Serv. b.b. 20271 (A), Feb. 12, 1938. Tree 43 m. high, the trunk 105 cm. in diameter, the bark brown-gray. Leaves elliptic-oblong, decurrent at the base into a petiole 2 cm. long, the lower 15 of the margin entire, the remainder crenate-serrate, acute at the apex, glabrous on both surfaces, ca. 14 cm. long, 6 cm. wide (the leaves of the young shoots 30 cm. long and 11.5 cm. wide). Young parts minute appressed-puberulent. Branchlets glabrous. Peduncles 2.5—3.5 cm. long, glabrous, bibracteolate. Sepals glabrous, the margin subciliate, semi- orbicular, 5 mm. long. Fruit depressed-globose, 2.5 cm. diam. glabrous. Adinandra Brefeldii and the following species, A. celebica, were offered by Koorders as new species with accompanying descriptions far too brief for satisfactory study without type-specimens. It may be that the types were also incomplete. Although both species appear to have been described from fruiting material, very little information concerning the fruit is presented. The interesting feature in the above description (which is merely a translation of the original), is the petiole of the leaf. Koorders mentions the leaf as being decurrent into the petiole 2 cm. long. In the above-cited sterile specimen, the petiole is hardly much over 1 cm. long but the leaf- blade is decurrent into the petiole and this decurrent condition persists along the branchlet. The specimen, although sterile, evidently has been collected from a vigorous young shoot, the leaves being of considerable size. Adinandra celebica Koorders in Meded. : rarer Plant. 19: 350, 642. 1898. — Melchior in Nat. Pflanzenfam. ed. 2, 21: DistripuTION: Netherlands East Indies pee CeLepes: Kandari, O. Beccari s.n. (K), Feb. 1883. 92 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII spicuous, the petiole ca. 5 mm. long, glabrous, occasionally lightly appressed-pubescent (on newly unfolded leaves). Fruit axillary, solitary ; pedicel 2-2.5 cm. long (3-4 cm., Koorders), glabrous, terete, recurved imbricate, quite glabrous, subrotund, ca. 5-6 mm. long, 6-7 mm. wide, rugulose. Young fruit glabrous, subglobose, tapering at the apex into a glabrous, entire style, 5-celled, many developing seeds The information offered by Koorders in his description of this species is too brief, and offers no details that might not be applied to any but the most exceptional species in the genus. He does state, however, that in the Kew Herbarium there is a specimen collected by Beccari in central Celebes which is supposedly identical with his type. From the Kew Herbarium was obtained a specimen collected by Beccari which may or may not be the specimen in question; hence it is cited here dubiously. The pedicels in the Beccari specimen measure not more than 2.5 cm., whereas Koorders records the measurements as 3-4 cm. He also states that the leaves are glabrous to the naked eye but laxly pilose under the hand-lens. The Beccari specimen is quite glabrous. However, the ter- minal leaf-buds are appressed-pubescent which would indicate that the very young leaves are pubescent, on the under surface at least. Only by an examination of the actual type can one expect to gain a concept of this species. 64. Adinandra crenulata T. Anderson ex Dyer in Hooker f., Fl. Brit. Ind. 1: 283. 1874. — Szyszylowicz in Nat. Pflanzenfam., III. 6: 189. 1893. — Melchior in op. Terasiroewia crenulata Wallich, Cat. no. 3723, herb. Madras. A discussion of this species is quite difficult since the only available reference is the sketchy original description by T. Anderson. A note, perhaps the original manuscript of Anderson, on an otherwise blank her- barium sheet in Kew states, “Possibly a native of the Malayan peninsula. Wallich’s specimen is the only one that exists and was obtained from the Madras Herbarium and bears no locality. It is probably one of those plants collected by Konig during his visit to the settlements in the Malayan peninsula.” Considerable doubt as to the origin of the type existed at the time of the description, and, from the literature examined, it appears that no worker has associated any material collected later with the species A. crenulata. Both Szyszylowicz and Melchior in their treatments of the Theaceae in “Die Natiirliche Pflanzenfamilien” incorporated this species in their work. Unfortunately, the original description of the species is such that it 1947 | KOBUSKI, STUDIES IN THE THEACEAE, XV 93 might be applied to any number of species in the Malayan region. Because of this lack of a single distinctive feature, and the fact that the type is in Madras rather than in one of the Malayan herbaria, its true identity has been lost and will oe to be until opportunity can be afforded some worker to study the typ Ridley 4801, acsaated in the U. S. National Herbarium and collected in the Singapore Garden Jungle, may belong to this species. The distinguish- ing feature of this specimen is the decided crenulation of the leaf-margin. It is not a good specimen. There are no terminal buds and the partially developed fruits are in rather poor condition. It was originally identified as A. maculosa T. Anders. DOUBTFUL SPECIES 65. Adinandra cape ‘ad in Mém. Soc. Phys. Hist. Nat. Geneve 1: 112 (Mém. Ternstr. 24). This species, aba only once and then doubtfully by Choisy, is based upon a reported species by Gaudichaud from the Hawaiian Islands. There has been no other record of Adinandra from the Hawaiian Islands since that time, hence one is led to believe that Choisy erred in his generic designation. The subsessile cordate leaves lead one to believe that Choisy may have had in mind Eurya sandwicensis A. Gray. This is merely a supposition, since no material has been available for this study. However, Choisy’s brief description does apply for the most part to specimens examined of E. sandwicensts. 66. saute sylvestris Jack in Malay Misc. 2 (7): 3. 1822 [repr. ] ae Jour. Hist, 4: 208. 1843; et in Miscel. Papers Indo-China II. 2: , 302. 1887]. The Nes of this species is very limited. In Misc. ath Indo- China is reproduced the original description. There, on p. 295, Adinandra sylvestris is described as “Bacca trilocularibus. Suka beranak. Malay. A large forest tree at Moco Moco.” On p. 302 of the same volume, Sir J. D. Hooker, in a note to the editor states, ““Adinandra sylvestris. Not taken up by later authors.—Stka béranak—i. e., fond of having children.” EXCLUDED SPECIES as Miquelii King in Jour. As. Soc. Bengal 59 (2):192 (Mater. Fl. Malay _ 1: 132). 1890=Ternstroemia baneana Miquel, Fl. Ned. Ind. Suppl. 1: 477. Adnan shibics aa King ex Durand & Jackson, Index Kew. Suppl. 1 (Add. ae” 01. 1906=Ternstroemia baneana Miquel, Fl. Ned. Ind. Suppl. I: pre This ae was originally described as Ternstroemia bancana by Miquel in 1862. In 1890, King, thinking the species an Adinandra, transferred it to the latter genus under the name A. Miquelii. Later, Ridley (Fl. Malay Penin. 1: 197. 1922) treated the species under its original genus and name, Ternstroemia bancana Miquel. In the meantime, Durand and 94 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII Jackson (1906) had made still another combination, Adinandra bancana, erroneously attributing the binomial to Kin AFRICAN SPECIES Two species of Adinandra have been described from Africa. The first species, A. Mannii Oliver (1868), was collected on St. Thomas Island off the west coast of Africa in the Bay of Guinea, and the second species, 4. Schliebenii Melchior (1934), was collected in Tanganyika Territory in eastern Africa, a span of nearly the whole African continent separating the two entities. It appears from the literature that both species have been collected only once. Although these two species differ considerably in many features from the Indo-Malaysian and Papuan groups, on the other hand they possess many interesting characteristics in common with the latter and are closely related. Among the outstanding characters in which they differ from the other groups are: ) stamens in a single series, the filaments glabrous, free except at the very base, the anthers glabrous, subsagittate; and (2) corolla-lobes three times longer (35-50 mm.) than the calyx-lobes, connate at the base only, otherwise free. disposed in a pseudo-tube, the apices of the lobes arching over the pistil and stamens. Szyszylowicz (1893) created a section, Eleutherandra, for A. Mannii based on the free filaments and supposedly separating this species from all other species of Adinandra. Melchior (1925) continued this section of Szyszylowicz and in 1934 added to the section the species A. Schliebenii. This character, if meant to separate these two species from all others, is not well chosen, since nearly every species in China must be included. The most outstanding characters to my mind are the glabrous subsagittate anthers and the corolla-lobes arranged in a tube. No formal key is necessary to separate here these two species. Their differences are mentioned in detail in the discussion following the aaa . Adinandra Mannii Oliver in Fl. Trop. Afr. 1: 170. 1868. — Hocker, Icon. Pl, 11: 039. 1867. — Szvszvlowicz in Nat. Pflanzenfam. IIT. 6: 189, 1893. — Mele - in Notizbl. Bot. Gart. Mus. Berlin 8: 657. 1924; 11: 1100. 1934; in Nat. Pflanzentam. ed. 2, 21: 144. 1925. DisrriputTron: West Africa (Island of St. Thomas). Gutr or Gutnea: Island of St. Thomas, summit of the peak of the island, G. Mann 1066 (isotype, G) Tree 10 m. high: branches gray, terete, glabrous, the young branchlets red-brown, glabrous, terete, minutely tuberculate-punctate, the terminal buds glabrous. Leaves chartaceous to subcoriaceous, oblong-elliptic to oblong-obovate. quite uniform in size, 6-7 (—9) cm. long, 2-3 cm. wide, acuminate at the apex, asymmetrical, subrotund at the base, glabrous on both surfaces, the midrib reddish near the base, tuberculate- -punctate on the lower surface, the margin glandular-denticulate, the s 15-18 pairs, conspicuous but not prominent, anastomosing midway : the margin, the petiole 1-2 mm. long. Flowers axillary, ? solitary; pedicel ca. 2 cm. long 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 95 (fide Oliver), glabrous; bracteoles 2, persistent, glabrous, opposite, imme- diately below the calyx, broadly ovate, 7-8 mm. long, 5-6 mm. wide, acute at the apex, carinate; calyx-lobes 5, imbricate, glabrous on the exterior surface, very lightly appressed- -pubescent (with binocular) on the interigr surface, broadly ovate, acute at the apex, unequal, the two outer lobes 12-14 mm. long, ca. 10 mm. wide, the three inner lobes 17-19 mm. long, ca. 10 mm. wide, their ees . scarious, entire; corolla-lobes 5. slightly connate only at the base, not spreading, somewhat tubuli- form, oblong, ca. 25 (35-45) mm. ne 7-10 min. wide, obtuse to subrotund at the apex; stamens ca. 30 in a single series, equal, ca. 8 mm. jong, the filaments glabrous, ca. 4 mm. long, adnate to the base of the corolla, otherwise free, the anthers oblong-linear, equalling the filaments in length, strictly glabrous, the apicule less than 0.5 mm. long, truncate to emarginate; ovary conical-ovoid, glabrous, ca. 4 mm. diam., tapering into the style, 4-celled, multi- ovulate, the placentae bifid, the style glabrous, entire, ca. 22 mm. long, the stigma 4-lobed. Fruit not seen. The outstanding characters which separate this and the following species, A. Schliebenii Melchior, from the rest of the genus are: (1) anthers gla- brous; (2) filaments glabrous, free their entire length except at the very base, at which point they are joined together as well as adnate to the base of the corolla; and (3) corolla-lobes linear-oblong, connate at the base only, but connivent their entire length presenting a tube-like appearance. The characters which separate A. Mannii from A. Schliebenii are: (1) inner calyx-lobes glabrous and entire; (2) corolla-lobes subrotund at the apex; (3) bracteoles longer, measuring 7-8 mm. in length; (4) ovary and style glabrous; and (5) stamens only about half as long (8 mm.) and more numerous (30). 68. Adinandra Schliebenii Melchior in Notizbl. Bot. Gart. Mus. Berlin 11: 1076, 097. 1934 Distripution: East Africa (Tanganyika Territory). No SPECIMEN SEEN, Tree 20-30 m. high; very young branchlets smooth, compressed, gla- brous, red to gray-brown. Leaves ane ous to coriaceous, obovate- oblong, 5-10 cm. long, 2—3.5 cm. wide, acute at the apex or shortly and obtusely acuminate, cuneate at the base, eee on both surfaces, the midrib red beneath, the margin serrulate-dentate, glandular, the veins conspicuous beneath because of the red color, the petiole ca. 5 mm. long. Flowers axillary, solitary, ca. 3 cm. long; pedicels 2.5—3.5 cm. long, recurved, glabrous; bracteoles 2, persistent, opposite, immediately below the calvx, glabrous, broadly ovate, acute at the apex, carinate, the outer : , mm. (outer lobe) to 17 mm. (inner lobe), ca. 12 mm. wide, the two outer lobes glabrous on the exterior surface, the three inner lobes sericeous on the exterior surface except along the scarious margins, ciliolate; corolla- lobes 5, free, twice as long as the innermost calyx-lobe, linear-oblong, 32—45 mm. long, 5-7 mm. wide, acute at the apex, somewhat narrowed and connate at the base, not spreading, somewhat tubuliform; stamens 15—20, uni-seriate, ca. 15 mm. long, glabrous, the filaments free (inter se), adnate 96 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVHI to the base of the corolla, filiform, 8-9 mm. long, the anthers linear, ca. 5 mm. long, cordate-sagittate at the base, the connective projected at the apex into a small subulate-lanceolate apicule 1 mm. long; ovary conical, ca. 6 mm. long, 4 mm. wide, sericeous-tomentose, imperfectly 5-celled, multi-ovulate, attenuate at the apex into a filiform style 24-30 mm. long, sericeous along the lower portion, glabrous along the upper portion, lightly 7 sulcate, the stigma minute, lightly 5-sulcate. Fruit indehiscent, conical- void, ca. 2 cm. long, 1 cm. wide, the seeds many, small, reniform, 1.2-1.5 mm. diam., brown, shining, reticulate. This species has been described so minutely that it can be separated easily from the preceding species, even though the type has not been seen. It is fortunate, since the type may not be in existence any longer. Adinandra Schliebenii Melchior was collected by H. J. Schlieben in Tanganyika Territory in eastern Africa, far removed from the Island of St. Thomas in the Bay of Guinea where A. Mannii was collected. According to Melchior, the type specimen is numbered 3175 and was found growing on the northwest side of the Uluguru Forest in Morogoro (Tanganyika Terr.), at an altitude of 1900 m. The flowers are orange and red in color. On December 28, 1932, the plant was in both flower and fruit. The characters separating this species from A. Mannii are: (1) inner calyx-lobes pubescent and ciliolate; (2) corolla-lobes acute at the apex; (3) bracteoles unequal and shorter, measuring 3 and 5 mm. in length; (4) ovary and lower half of the style sericeous; and (5) stamens twice as long, measuring 15 mm. in length, and less in number (15-20) VERNACULAR NAME: Msungu (Kiliguru). PAPUASIAN SPECIES To a three species have been described from New Guinea, A. Forbesii Baker, A. Brassii Kobuski, and A. calosericea Diels. The last named ger has . transfered to the new genus Archboldiodendron, which midway between Gordonia and Adinandra in characters. The styles in Archbold nro are free, very short, and recurved, resembling those of Gordon ood-structure the genus resembles Adiandri. It is not peat as swat stele the fruit is dehiscent or indehiscent. From the appearance of the ovules one would not expect the seeds to be winged. The first two species mentioned above belong, according to the collector, L. J. Brass, among the very large trees of New Guinea. They are closely related to each other and are separated from the Indo-Malaysian species by their large, fleshy purple corolla-lobes and their very numerous (50-60), long, densely pubescent stamens, in very definite series. The character- istics separating the two species from each other are noted in the discussions following their descriptions. 69, ig ean: ee hast in Jour. Bot. 61: Suppl. 4. 1923. — Kobuski in Jour. ol ay ee ae sive — Guinea. BritisH New Guinea: Sogeri region, Mt. Gawada, H. O. Forbes 545 (TYPE, Brit. Mus. Nat. Hist.; isorypr, K), 1885-1886. 1947] KOBUSKI, STUDIES IN THE THEACEAE, XV 97 Large tree; branchlets gray, terete, glabrous, the terminal buds appressed- pubesce nt. Leaves coriaceous, ovate-oblong, 5-8 cm. long, 2.8-3.5 cm. wide, acute or broadly acute at ‘the apex, cuneate at the base, glabrous on both "surfaces, the margin flat, lightly serrulate, the veins 8— 10 pairs, conspicuous on the lower surface, less so on the upper surface, anastomosing and arching upward near the margin, the reticulations quite conspicuous, the petiole up to 10 mm. long, glabrous. Flowers axillary, solitary; pedicel up to 4.5 cm. long, glabrous, conspicuously lenticellate; bracteoles 2, persistent, opposite, immediately below the calyx, broadly ovate or deltoid, ca. 1.5 mm. long, lightly appressed-pubescent; calyx-lobes 5, imbricate, recurved even in the bud, semi-orbicular, ca. 5 mm. long and wide, lightly appressed-pubescent or glabrescent; corolla-lobes 5, connate at the base, fleshy, glabrous, 15-20 mm. long, 12—14 mm. wide, scarious at the margin; stamens ca. 56, 5-seriate, 10-14 mm. long, filaments only lightly adnate to the base of the corolla, connate, densely sericeous on the exterior surface, and the upper half of the interior surface, the anthers lightly hirsute, oblong, subequal, 4-5 mm. long; ovary globose, glabrous, ca. 5 mm. long, 6 mm. wide, attenuate at the apex into an entire style, 5-celled, multi-ovulate. Fruit not seen. Adinandra Forbesii can be separated from A. Brassii by the following characters: (1) leaves oblong-elliptic, acute at both ends, conspicuously veined on the under surface; (2) pedicels up to 4.5 cm. long, glabrous, covered with many conspicuous lenticels; (3) bracteoles opposite; (4) corolla-lobes glabrous; and (5) calyx-lobes recurved, even in the bud. This species is closely related to A. Brassit and has many characteristics in common with it, especially in the floral parts. In both species the corolla-lobes are fleshy, the calyx-lobes hardly more than 5 mm. long and thinner than the corolla-lobes, the stamens are conspicuously 5-seriate, long (up to 15 mm.), and densely sericeous. According to Baker, Szyszylowicz, and later, Melchior, the stamens are described as “pentadelphous” or in “five- bundles. . True, they vary in hey appear joined in a single tube. Perhaps, in drying or in boiling they may be broken up into groups, but my study showed them more characteristically in an unbroken tube-formation. 70. Adinandra Brassii Kobuski in Jour. Arnold Arb. 21: 150. 1940. DistripuTIon: New Guinea, British New Guinea: Lake Daviumbu, Middle Fly River, rain-forest, L. J. Brass bark anil toaen brown wood; leaf apex subretuse and margin much r ceed: especially near the base; flowers purple, pendent beneath leaves). NetTHERLANDS NEw GuInEA: Two kilometers s. w. of Bernhard Camp, Idenburg River, frequent on slopes of primary rain forest, alt. 850 m., L. J. Brass & C. Versteegh 13176 (A), March 19, thick, black, scaly, fairly rough, the sap-wood red-brown, the heart-wood violet). — Six kilometers s. w. of Bernhard Camp, Idenburg River, frequent tree of primary forest, on slope of ridge, alt. 1200 m., L. J. Brass & C. Versteegh 12519 (A), Feb. 15, 1939 (tree 28 m. high, diameter 55 cm., crown not wide-spreading, the bark 8 mm. 98 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XXVIII thick, brown fairly smooth, the sap-wood light brown, the heart-wood violet; flowers dark red; fruit green when young, red when mature). Very large trees up to 30 m. high; branchlets gray, terete, glabrous, appressed-pubescent at the very tip, the terminal bud a ppressed-pubescent. Leaves thick-coriaceous, obovate to subrotund, (4—) T-ti cm. long, (2.5—) 4-6 cm. wide, subrotund and subemarginate at the apex, cuneate to rounded at the base, the upper surface glabrous, shining, the lower surface glabrescent, the margin subrevolute with a few inconspicuous glands, the veins 15—20 pairs, inconspicuous on the lower surface, the petiole glabrous, 5—7 mm. long. Flowers axillary, solitary or in pairs; pedicel appressed- pubescent or glabrescent, free from lenticels, 3-3.5 cm. long; bracteoles 2, persistent, below the calyx, subopposite or alternate, broadly deltoid, ca. 1.5 mm. long, 2.5 mm. wide, puberulent; hegeer lobes ‘5, imbricate, broadly ovate to subrotund, 4—5 mm. long, 5—6 mm. wide, obtuse or rounded at the apex, lightly appressed-pubescent ; pena Hive 5, lightly connate at the base, purple, fleshy, 15-20 mm. ate 14-16 mm. wide, appressed-pubescent on the external surface, scariose at the margin; stamens 55-60, 5-seriate, 10-15 mm. long; filaments onl lightly adnate to the base of the corolla, 6-11 mm. long, connate, de nsely sericeous on the exterior surface, the vg a hirsute, oblong, subequal, 3-4 mm. long; ovary globose, glabrous, . 6 mm. long, 7 mm. wide, gradually attenuate into the entire style, 5. celled, the placentae prominently 2-branched, recurved with many ovules attached. Fruit globose, red, glabrous, up to 3 cm. long. Adinandra Brassti is listed by Brass as one of the large trees of New Guinea. It is a very distinctive species characterized by thick, dark, scaly bark, light brown or red-brown sap-wood and purple heart-wood. The corolla-lobes are purple, and much more fleshy than the Malayan and Chinese species; the calyx-lobes, on the other hand, are smaller and thinner. Adinandra Brassii can be separated from A. Forbesii, its closest relative and the only other species known at present from New Guinea, by its obovate or subrotund leaves which are not oblong, the much less con- spicuous veining, and the fine appressed-pubescent corolla-lobes. When first described this species appeared very outstanding because of the densely hirsute anthers and filaments. Since Baker, in describing 4. Forbesii, did not mention pubescence on either the corolla-lobes or the stamens, I assumed them both to be glabrous. An examination of the type of A. Forbesii showed, as far as the corolla-lobes were concerned, this assumption to be correct. However, the stamens (both anthers and filaments) are so densely sericeous that they appear to be the most noticeable characteristic of the flower ARNOLD ARBORETUM, Harvarp UNIVERSITY. 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 99 THE CYPERACEAE COLLECTED IN NEW GUINEA BY L. J. BRASS, I* S. T. BLAKE With four text-figures Turs is the first contribution to an account of the specimens of Cypera- ceae collected in New Guinea chiefly by Mr. L. J. Brass during the Arch- bold Expeditions between 1933 and 1939. Included also in the treatment are some other specimens from New Guinea and the Solomon Islands from various collections. I am particularly grateful to Proj. Elmer D. Merrill, of the Arnold Arboretum, for the opportunity to study the Brass collection, and to Dr. Lily M. Perry for copies of much literature unavailable to me. To Mr. A. J. Jessep and Mr. J. H. Willis I am indebted for the loan of specimens and literature from the National Herbarium of Victoria, Mel- bourne (cited as herb. Melbourne), and to Mr. W. Hartley for the loan of a set of the specimens collected by Mr. C. E. Carr, now in the herbarium of the Council for Scientific and Industrial Research, Canberra (cited as herb. Canberra). Unless otherwise indicated, the type-specimens of new species are in the Queensland Herbarium, Brisbane (herb. Brisbane) and duplicate-types in the Gray Herbarium of Harvard University, Cambridge, Mass. The present paper deals with the genus Carex, and in this genus I have received valuable help from Mr. E. Nelmes, of the Herbarium, Royal Botanic Gardens, Kew, England, who has been studying the Malaysian and Australasian species of the genus for some years. The arrangement of species here follows that in Kiikenthal’s well-known monograph in Das Pflanzenreich IV.20. 1909, but some variations of his treatment have been found necessary at this later date. The specimens have been critically compared with copious Australian material, but the ‘comparison with Malaysian material has been much less extensive and considerable reliance has had to be placed on descriptions alone. Carex L. Subgen. Primocarex Kikenth. Carex eremostachya sp. nov. (Sect. Circinatae Meinsh. sensu Kikenth.). Fic. 1, A. Rhizoma repens circiter 1.5 mm. crassum squamis tenuibus brunneis opacis tenuiter striatis obtectum. Culmi approximati vel laxe caespitosi, * Botanical Results of the Richard Archbold Expeditions. 100 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVUI circiter 20-25 cm. alti, graciles flaccidulique, acute triquetri a compressi, scabri. Folia pauca culmum superantia, inferiora ad vaginas brunneas integras redacta, 3—4 superiora laminifera; laminae acutae, usqu ad 20 cm. longae, 2—2.5 mm. latae, planae, glabrae, inferne laeves sursum nervis marginibusque scabrae, ca arinatae ceterum obscure nervosae. Spica unica, erecta, androgynaecea, ebracteata, 10-12 mm. longa, 5 mm. crassa, oblonga, pluri- et densiflora, variegata, pars mascula linearis circiter 4— 5 mm. longa. Glumae femineae ovatae acute acuminatae, 3.5—4 mm. longae glabrae, dorso obtuse car ge ish -nerves, lateribus rubro- brunneae, Utriculi suberecti, 4.5 mm. longi, . lati, membranacei, pallide viridescentes, elliptici, rostrati, ad cine glabri laevesque, prominule circiter 15-nerves, basin versus angustati brevissime stipitati, sursum in rostrum circiter 1.5 mm. longum gracile lineare triquetrum glabrum laeve ore firmo bidentatum (dentibus strictis acutis) subsensim contracti. Nux utriculo S.T BLAKE — A. Carex eremostachya S. T. Blake: a. part of plant with fruiting culm, x “a pistillate glume, flattened out, x 10; c. utricle, dorsal view, 0; ut, x i. Carex perileia S. T. Blake: a. part of plant with fruiting culm, pistillate glume, flattened out, X 10; c. utricle, dorsal view, x 10. All figures from type-specimens. ’ 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 101 brevior et subaequilata, 1.8 mm. longa, 1.2 mm. lata, oblonga subobovoidea, obtusissima sed brevissime acuminato-apiculata, basi subacuta, valde tri- quetra angulis prominulis, lateribus fere plana. Stylus basi haud in- crassatus; stigmata 3. NETHERLANDS NEw GuINEA: 9 km. northeast of Lake Habbema, Brass 10255 (ryPE), October 1938, alt. 2800 m., massed on open beaches of small stream in forest. This species is most nearly allied to C. rhizopoda Maxim., from Japan, differing in the more densely-flowered female part of the spike with broader reddish brown (not green) glumes, the nut more obtuse at each end, and perhaps also in the shorter male part of the spike. Ohwi, in Mem. Coll. Sci. Kyoto Imper. Univ. ser. B, 11(5): 443. 1936, has erected a new section Rhizopodae Ohwi for C. Pe iude, but C. eremostachye appears to agree better with the characters of the pice: under which Kikenthal arranged C. rhizopoda in Pflanzenr. 1V.20: 98. 1909. C. eremostachya also appears to be closely allied to C. ee Gross, but to judge from the description the latter has a narrower spike but longer utricles (5-6 mm.). No mention is made of a rhizome. Carex capillacea Boott, Illustr. 1: 44, t. 770 Carex rara Boott subsp. capillacea (Boott) See in Pflanzenr. 1V.20: 102. 1909. NETHERLANDS New GuINnEA: Lake Habbema, Brass 9085, August 1938, alt. 32 scattered along grassy shores of lake; 11 km. northeast of top of Mt. = alr Brass & Mvyer-Drees 9751, Sept. 1938, alt. 3400 m., in wet grassy valley. Not previously recorded from New Guinea; known to extend as far north as Sakhalin and Manchuria and west to Sikkim, and known to occur in southeastern Australia. Brass 9085 has a short setaceous bract half as long as the spike. Subgen. Vignea (Beauv.) Nees Carex aes R. Br. Prodr. 242. 1810; Kiikenth. in Bot. Jahrb. 69: 261. 1938; Ohwi in Bot. Mag. Tokyo 56: 214. 1942 NETHERLANDS NEW GuINeEa: Lake Habbe ema, Brass 9248, August 1938, alt. 3225 ., gregarious on wet low shores of lake, erect in clumps to 1. 5 m. high ee recorded for Netherlands New Guinea by Ohwi, ]. c., and for Northeast New Guinea by Kiikenthal, 1. c.; elsewhere known from Australia, New Zealand and its dependencies, and New Caledonia. Carex curta Gooden. in Trans. Linn, Soc. Lond. 2: 145. 1794. NETHERLANDS New Gurnea: Lake Habbema, Brass 9037, August 1938, alt. 3225 m., common on marshy flats; Lake Habbema, Brass 9119, August 1938, alt. 3225 m moist burnt-over ground in a native camp; Lake Habbema, Brass 9539, August 1938, alt. 3225 m., open grassland, plentiful on wet sandy soil. Not previously recorded for New Guinea, but widely spread in Europe, Asia, North and South America, and southeastern Australia. This species has commonly been called C. canescens L., but E. 5. Mar- shall has shown, in Jour. Bot. 45: 365. 1907, that the type of Linnaeus is a specimen of quite a different plant, Sonmcnly known as C. Buxbaumi Wahlenb., and that the present plant has to be called C. curta Gooden. This view is followed by Nelmes in Proc. Linn. Soc. Lond. 155 sess.: 284. 102 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXVIII 1944, ‘Ihe confusion seems to have arisen with the misidentification by Lightfoot, Fl. Scotica 550. 1777, and has been perpetuated by such authori- ties as Bentham (e.g. in Handb. Brit. Fl. 2: 913. 1865, and in FI. Austral. 7: 439. 1878), Boeckeler (in Linnaea 39: 122. 1875), C. B. Clarke (in Hook. f. Fl. Brit. Ind. 6: 706. 1894), Kiikenthal (in Pflanzenr. [V.20: 216. 1909), Ohwi (in Mem. Coll. Sci. Kyoto Imper. Univ. ser. B, 11(5): 268. 1936), Mackenzie (in N, Am, FI. 18: 374. 1935), and Duman (in Catholic Univ. Am. Biol. Ser. 36: 44. 1941). F. N. Williams, when discussing the European varieties of Carex canescens in Jour. Bot. 46: 369-376. 1909, agreed with Marshall’s interpretation, but continued to use the name C. canescens in the traditional sense with the remark that “I think that all doubt is cleared away by citing Lightfoot as the authority.” But this practice is contrary to the International Rules of Nomenclature, partic- ularly Arts. 16, 18, 46 and 47 Carex perileia sp. nov. (Sect. Elongatae Kunth). Fic. 1, B. Rhizoma stoloniforme tenue. Culmi caespitosi, gracillimi, compresse trigoni, striati, glabri, laeves, usque ad medium foliati, quoad visi 20 cm. alti. Folia pauca culmum superantia, flexuosa, angustissima, conduplicata (explanata circiter 1 mm. lata), marginibus superioribus scaberula, ima a vaginas persistentes pallidas redacta. Inflorescentia subcapitata, ovata, densa, brunnea, circa 1.5 cm. longa lataque; bracteae squamiformes, ima breviter aristata. Spicae 3 (quoad visae) approximatae, sessiles, gynaecan- drae parte masculina perbrevi, obovatae vel oblongae, squarrosae, 7—10 mm. longae, 7-8 mm. latae. Glumae femineae ovatae, acutae, muticae vel fere muticae, carina angusta 1—3-nervi virides, lateribus enervibus brun- neae vel rufo-brunneae nitidae, marginibus anguste albae subscariosae, 3-3.5 mm. longae. Utriculi patuli, 5 mm. longi, 1.4 mm. lati, fulvescentes viridi-marginati, omnino glabri aie Pe rnuintl elliptici, membrana- cei, basi rotundata incrassati, plano-(subconcavo-) convexi, nervis 2 mar- ginalibus conspicuis, nervis tenuioribus ie plano circiter 7 latere convexo circiter 10 notati, sursum in rostrum latiusculum complanatum leviter incurvum breviter bidentatum omnino laeve glabrumque purpureo-tinctum cetero utriculo subaequilongum dorso anguste sulcato profunde fissum subsensim contracti. Nux utriculum implens, 2 mm. longa, 1.2 mm. lata, oblongo-ovata, plano-convexa. Stigmata NETHERLANDS NEW GuINnEA: Lake Habbema, Brass 9583 (type), August 1938, alt. $225 m., marshy shores of lake The specimen seen has a single fruiting culm. The species is very close to some forms of C. echinata Murr. (C. stellulata Good.), but has narrower leaves and longer utricles with a relatively longer beak deeply split on the back and with entirely smooth margins. The split on the back usually appears as a narrow brown furrow, as the result of the overlapping of the edges. Subgen. Indocarex Baill. Carex Dietrichiae Boeck. in Flora 58: 122, 1875. SotromMon Istanps: San Cristobal: Waimamura, Brass 2660, August 1932, lowlands on pathways and in open swampy places in the rain-forests, plentiful (inflorescence brown 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 103 Previously known to me with certainty only from Queensland. Kuken- thal, in Pflanzenr. 1V.20: 264. 1909, regards C. Dietrichiae Boeck., C. fissilis Boott, and C. indica L. var. Milnei C. B. Clarke as synonymous, treating the group as C. indica var. fissilis (Boott) Kukenth. Carex Dietrichiae differs from C. indica L. in having shorter awns to the brac- teoles, bright brown (not pallid) male glumes, and less spreading, less turgid utricles less abruptly constricted into their beaks. Mr. Nelmes informs me that C. fissilis has much longer pistillate glumes than C. Dietrichiae. | I have not seen the specimens referred to C. indica var. Milnei by Ridley in Trans. Linn, Soc. II. Bot. 9: 247. 1916, nor those referred to C. indica ee _ Carex lamprochlamys S. T. Blake: a. part of middle region of inflorescence, X 1; b. pistillate glume, flattened out; x 10; c, d. utricle, dorsal and lateral views, X 10. B. Carex melanophora S. T. Blake: a. upper part of inflores- cence, X 1; b. pistillate glume, flattened out, x 10; ¢, d. utricle, dorsal and lateral views, X 10; e. nut with style, x 10. All figures from type-specimens. 104 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVUI var, fissilis by Kikenthal in Bot. Jahrb. 59: 59. 1924. The specimen from New Guinea referred to C. fissilis by F. Mueller in Trans. Roy. Soc. Vict. 1(2): 36. 1889 belongs to C. continua C. B. Clarke Carex lamprochlamys sp. nov. (Sect. Indicae Tuckerm.). Fic. 2, A. Herba robusta circa 75 cm. alta. Culmi subvalidi, acute Sgt uninodes, glabri laevesque. Folia summam inflorescentiam d attin- striata, carinata, marginibus revoluta, supra scaberrima, subtus scabra, omnia laminifera, inferiora vaginis sanguineo-fuscis plus minusve persis- tentibus. Panicula plus quam dimidium culmum occupans, 7—8-nodis, composita, linearis, fere continua. Bracteae longe vaginantes vaginis pal- lidis ampliatis prope os antice puberulis, inflorescentiam superantes, folii- formes, summae filiformes. Paniculae secundariae 12, singulae vel binae, striati, scaberuli. Rhachis scaberrima. Bracteolae emarginatae, arista scabra longe aristatae. Spicae pallidae ferrugineo-tinctae, androgynaeceae, oblongae vel ovoideae, plerumque 6-8 (raro 5) mm. longae, circa 5 mm, latae, parte mascula partem femineam adaequante vel ea breviore. Glumae femineae fulvae, ovatae, acutae admodum acuminatae, aristulatae, pluris- triatae, circiter 2 mm. longae. Utriculi pallidi, nitidi, membranacei, 3-3.5 mm. longi, 1 mm. lati, trigoni haud inflati, nervis prominulis circiter 15 notati, inferne elliptici vel obovati basi contracti, glabri laevesque, sursum in rostrum gracile planum fere rectum vel leviter excurvum ore integrum oblique sectum marginibus admodum scabrum cetero utriculo aequilongum subabrupte contracti. Nux 2 mm. longa, 0.9 mm. lata, brunnea, ellipsoidea utrinque acuta, triquetra, apice indistincte annulata, lateribus concava, basi breviter stipitata. Stylus basi incrassatus; stigmat a 3. Papua: Central Division: Mafulu, Brass 5323 (type), Sept—Nov. 1933, alt. 1250 m., common erect ground plant in Castanopsis forest; Boridi, Carr 13160, Sept. 1935, alt. ca. 4700 ft., young secondary forest (herb. ca. 2 ft. tall) (herb. Canberra). Kukenthal cited Brass’s collection under C. Rafflesiana Boott var. con- tinua (C, B. Clarke) Kiikenth., in Bull. Jard. Bot. Buitenz. III. 16: 315. 1940, and Carr’s collection under C. indica L., in Bot. Jahrb. 69: 262. 1938, The species differs from C. indica L. and its immediate allies (subsect. Turgidulae Kiikenth.) in that the utricle is not olivaceous, is by no means inflated, and is gradually attenuated into its beak, and differs further from C. indica itself in that the pistillate glumes are only shortly aristate. From C. Rafflestana and its allies (subsect. Hispidulae Kikenth.) it differs in the stouter more rigid panicle and in that the utricle is but sparsely scabrous and then only on the narrow beak. By reason of the non-inflated utricle not abruptly narrowed into its slender only slightly clr beak, this species seems best placed in subsect. Gracilirostres Kikenth., but it is readily distinguished by its pallid shining mature utricle. ae Ss specimen is a somewhat immature plant with a culm about 50 cm. high and leaves about 7 mm. wide se ish continua . Clarke in Hook. f. Fl. Brit. Ind. 6: 717. 1894. © Ralston ee var. continua (C. B. Clarke) Sahar in Philip. re ar ee 1911, in Bot. Jahrb. 59: 59, 1924, in Bull. Jard. Bot. Buitenz. III. 16 1940. 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 105 NETHERLANDS NEw GurInEA: 4 km. southwest of Bernhard Camp, Idenburg River, Brass 13211, March 1939, alt. 850 m., rain-forest, one clump on river-bank; 4 k southwest of Bernhard Camp, Idenburg River, Brass 13722, March 1939, alt. ne oe Camp, Idenburg River, Brass 12371, Jan. 1939, alt. 1500 m., rain-forest, one clump in sandy stream-bed; 18 ni m, northeast of Lake Habbema, Bele River, Brass 11225, Nov. pua: Central Division: Near summit of Owen Stanley Range at 9200 ft., W. spelt (herb. Melbourne) ; Mt. Tafa, Brass 4842, May-September 1933, alt. 2400 plentiful on roadside in forest; Mt. Tafa, Brass 5007, ep Rae 1933, alt. s 400 m., common roadside species, also found occasionally on forest floor, Mt. Taf Brass 5088, alt. 2400 m., May—September 1933, scattered on thinly se face a an old landslip. Not previously recorded for Netherlands New Guinea. Brass 5007 was ae by Kiikenthal, in Bull. Jard. Bot. Buitenz. III. 16: 315. 1940, under C. Rafflesiana Boott var. continua (C. B. Clarke) Kiikenth. Kukenthal also cites Brass 5323, 4391, and 4623 as belonging to this form; of these Brass 5323 serves as the type of the abundantly distinct C. damprochlamys S. T. Blake, hong nos. 4391 and 4623 are referred below to C. neoguineen- sis C. B. Clarke. MacGregor’s specimen was referred to C. fissilis Boott by F F. Mueller, in Trans. Roy. Soc. Vict. 1(2): 36. 1889. Kiikenthal’ s earlier opinion that C. continua C. B. Clarke is specifically distinct from C. Rafflesiana Boott seems preferable to treating it as a variety of the latter. The specimens seen from New Guinea and the Philippine Islands seem to form a homogeneous series. The species extends to the eastern Himalayas. Carex neoguineensis C. B. Clarke in Jour. Linn. Soc. Bot. 37: 1904, in Kew Bull. Add. Ser. 8: 73. 1908; Kiikenth. in Pflanzenr. IV.20: iy ine in Bot. Jahrb. 59: 59. 1924; Suring. in Nova Guinea Bot. 8: 713. 1912; Ridl. in Trans. Linn. Soc. II. Bot. 9: 246. 1916 NETHERLANDS NEW ere Lake Habbema, Brass 9210, August 1938, alt. 3225 m., associated with Sphagnum and grasses in shrubby edges of forest; 9 km. northeast of Lake Habbema, Brass 10882, October 1938, alt. 2650 m., wet bank of an open stream 2 forest; 9 km. northeast of Lake Habbema, Brass 10925, October 1938, alt, 2800 m., on on wet sandy soil of openings in forest. Papua: Central Division: Mt. Albert Edward, Brass 4391, May-July 1933, alt. 3600 m., small colony on an open toss te slope (inflorescence brown); Mt. Victoria, W. ‘MacCresor in 1889 (herb. or rne); Murray Pass, Wharton Range, Brass 4623, June— —September 1933, alt. oe, m., common on ney and recently burnt areas (leaves pale green, infloresence oo lower regions of British New Guinea, W. MacGregor in 1894 (herb. Melbourne). The specimens cited show some variation between one another, par- ticularly as to the size of the utricle. Brass 9210 and 4623 have small utricles 2.5-2.8 mm. long, while on Brass 10882 and MacGregor’s specimens the utricles are 2.7-3.5 mm. long with longer more slender beaks. The e-specimen is described as having utricles 3 mm. long with the beak emarginate tip of the utricle. Mr. E. Nelmes does not mention this character, but tells me that apart from the size and rather stouter beak the utricle of Brass 9210 is rather similar to that of the type and that the glumes are much alike. In view of the variation on Brass’s specimens I 106 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVUI now have no hesitation in referring all the above-cited specimens to Clarke’s species. MacGregor’s specimens have coarser leaves up to 10 mm. wide and the partial panicles sometimes paired, but the material is rather scrappy. As mentioned in the discussion following C. continua above, Brass 4623 and 4391 were determined by Kiikenthal as C. Rafflesiana Boott var. con- tinua (C. B. Clarke) Kiikenth. Carex neoguineensis is undoubtedly closely allied to C. continua, but it has more loosely flowered spikes with the utricles more abruptly contracted into a more slender trigonous promi- nently decurved beak, while the shorter stiffer leaves and bracts (the former usually not exceeding the base of the inflorescence, the latter only about as long as the partial panicles) and the shortly aristulate to muticous bracteoles give a different facies to the species. Carex melanophora sp. nov. (Sect. Indicae Tuckerm. subsect. Hispidulae Kiikenth.). Fic. 2, B. Herba stolonifera usque ad 4 dm. alta; stolones graciles usque ad 2.5 mm. crassi, squamis brunneis vel fuscis persistentibus obtecti. Culmi solitarii vel caespitosi, graciles, enodes, obtuse tr rigoni, striati ceterum laeves vel fere laeves, glabri. Folia prope da culmi conferta, numerosa, recta, summam inflorescentiam fere attingentia, 2-3 mm. lata, apicem triquetrum versus longe attenuata, _ majore ser plana vel marginibus recurvula, conspicue carinata, crebre nervosa, costa supra appresse scabro- -pubescentia ceterum glabra, ahah at ta (mediis exceptis) scaberula, carina sursum scabra, emortua persistentia; vaginae atropurpureae vel fuscae tandem in fibras dissolutae. Panicula composita, 12-24 plantae occupans, linearis, densa, basi interrupta. Paniculae secundariae 14, plerumque binae, fuscae, oblongae vel lar oblongae, dense spiciformes, plerumque 1—2.5 cm. longae, 5—7 mm. latae, sup erlores subsessiles approximatae, inferiores prominule pedunculatae ima plus minusve distans solitaria. Bracteae ps vaginis arctis antice tenuiter breviterque pubescentibus. Bracteolae squamiformes, aristatae, plus minusve hispidae. Pedunculi filiformes, erecti, scabri, imus usque ad 8 cm. longus. Rhachis scabra. Spicae androgynae, Bee eerie appressae, oblongae, 4—6 mm. longae, 2—2.5 mm. latae; pars mascula partem femineam circa adaequans sed (utriculis circumdata) inconspicua. Glumae femineae sanguineo-fuscae, c acutae, muticae vel plus minusve cuspidatae, 2.5-3 mm. longae, tenuiter S-nerves, Carina tenui straminea. Utriculi fusci vel sanguineo- fusci, erecti, 3.5-3.8 mm. longi, 0.9-1 mm. lati, sesh utrinque angustati, suba aeque subacuteque trigoni, nervis 6—8 tenu ibus (quorum interdum 2 solis con- spicuis) notati, appresse scabro-pubescentes, marginibus hispiduli, in ros- trum sensim nga S; rostrum rectum cetero utriculo aequilongum vel fere aequilongum, subgracile, compressum, ore albo-marginato profunde bidentatum ents strictis), marginibus scabrum. Nux arcta, elliptica, stipitata, conspicue apiculata, acute triquetra lateribus concava, stramineo- brunnea, 1.6 mm. longa, 0.8 mm. lata. Stylus basi haud incrassatus; stigmata 3, NETHERLANDS NEw GurNeA: North slopes of Mt. Wilhelmina, Brass & Myer- Drees 10046, Sept. 1938, alt. 3950 m., wet places in alpine grassland, common; northern 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 107 slopes of Mt. Wilhelmina, Brass & Myer-Drees 10080, Sept. 1938, alt. 3950 m., alpine grassland, matted on a rock; Mt. Wilhelmina, 7 km. northeast of top, Brass & Myer Drees 9828 (TYPE), Sept. 1938, alt. 3560 m., abundant on moist grassy cliffs. Carex melanophora is evidently close to C. sarawaketensis Kukenth., but differs in the nodeless smooth stems, smaller partial panicles, smaller spikes, longer and perhaps more distinctly nerved utricles which are scab- rous rather than hispid except on the margins of the beak, and perhaps also in the development of stolons. e three collections cited above are unlike each other at first glance, but oe are indistinguishable as to their spikelets and appear to reflect differences of habitat. All specimens show at least a piece of a stolon. The tvpe-sheet consists of two pieces 3 and 4 dm. high with narrow partial panicles the lower of which are distant and long-peduncled. Brass & Mvyer-Drees 10080 consists of tiny plants up to 5 cm. high; they are immature but evidently represent starved plants. The third collection is intermediate with plants 7-13 cm. high, not quite mature, with fairly dense inflorescences with the lower partial panicles shortly pedunculate and thicker than in the type. The utricles are some- what larger than in the type, 4-4.5 mm. long and 1.1—1.2 mm. wide. It Mes appear that the usual length of the utricle of the species varies from 3.5 to 4.2 mm. Carex rhizomatosa ae in Zoll. Verz. Ind. Archip. 60: 1854; Kiikenth. in Bot. Jahrb. 59: 60. 1924, 69: 264. 1938, in Bull. Jard. Bot. Buitenz. III. 16: 316. 1940. va: Central Division: Mafulu, Brass 5318, Sept.-Nov. 1933, alt. 1250 m., common amongst tall grass of deforested spurs. This collection differs from the general run of the species in the somewhat smaller nearly smooth utricles 2.5 mm. long. It was previously cited by Kiikenthal in Bull. Jard. Bot. Buitenz., 1. c. The species ranges from Assam and Burma to New Guinea. Subgen. Eucarex Coss. & Germ. Carex Gaudichaudiana Kunth, Enum, 2: 417. 1837; Ohwi in Bot. Mag. Tokyo 56: 215. 1942. NETHERLANDS New GuINEA: Lake Habbema, Brass 9234, August 1938, alt. 3225 m., common on marshy banks of streams; Mt. Wilhelmina, 7 km. northeast of top, Brass & Myer-Drees 9923, Sept. 1938, alt. 3700 m., few plants on boggy grasslands The specimens are small and immature, corresponding to var. humilior. Kiikenth. in Pflanzenr. [V.20:; 313. 1909, which was recorded for Northeast New Guinea in Bot. Jahrb. 69: 264. 1938. To judge from the range of variation seen in Australian plants this small state is not worth taxonomic recognition. The species occurs in southeastern Australia (including Tasmania) and New le and was recently recorded for Netherlands New Guinea by Ohwi, | Carex phacota Spreng. Syst. 3: 826. 1826; Ohwi in Bot. Mag. Tokyo 56: 214. 1942. - Eastern Division: Summit of Mt. Dayman (herb. Melbourne, without further data). New for Papua; recorded by Ohwi, 1. c., for Netherlands New Guinea, otherwise known from Nepal and Japan to Java. The sheet carries an old 108 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII herbarium: label marked in the handwriting of F. Mueller ‘““Carex Summit of Mt. Dayman,” to which J. H. Willis has added “New Guinea (ca. 1894).” Carex euphlebia sp. nov. (Sect. Acutae Fries). . 3, C. r dense caespitosa rhizomate bre “ Culmi caespitosi, cum inflorescentia 30-65 cm. alti, graciles, acute triquetri, sursum scabridi, usque ad medium foliati, basi vaginis nigricantibus nitidulis obtecti. Folia coriacea, inflorescentiam longe superantia, 2.5—-5.5 (pro more 3—5.5) mm, lata, in apicem acutum fere planum scabrum longe attenuata, plerumque plana sed marginibus scaberrimis plus minusve recurva vel revoluta, nervis evaginantes, 2—3 inferiores foliiformes inflorescentiam superantes, ceterae breviores, summae squamiformes. PR gcventing capillares, tenaci-flexiles, angulis scabri, imus usque ad 5 cm. longus. Spicae 10-14, singulatim dispositae, vix distantes, lineari- Ease atro-fuscae, densiflorae, inferi- ores pedunculatae plus minusve nutantes, superiores subsessiles patulae; terminalis mascula, interdum basi breviter feminea, 5-12 mm. longa; ceterae androgynae, 2—6 (plerumque 3—5.5) cm. longae, 4.5—5 mm. crassae, in %4—) partem superiorem masculam, circiter 1.5 mm, crassam, abrupte angustatae. Glumae femineae violaceo- -nigrae, dorso viridulo es -nerves, margine angustissime albo-hyalinae, 1.5-1.8 mm. longae, oblongo-ovatae, plus minusve obtusae, cuspide plana scabra glumam fere adaequante vel ea breviore praeditae. Utriculi pallide brunnei, glumam superantes vel multo superantes, 1.8-2.2 mm. longi, 0.9-1 mm. lati, tandem divaricati, biconvexi, elliptici utrinque acuti vel apice breviter subacuminati, ore minute truncati vel emarginati, 14-16-costati, sursum punctulati ceterum laeves, glabri. Nux utriculo brevior aequilata, elliptica, biconvexa, brunnea, circiter 1.4 mm. longa. Stylus exsertus; stigmata 2. NETHERLANDS NEW GuINEA: 9 km. northeast of Lake Habbema, Brass 10877 (TYPE), October 1938, alt. 2650 m., common on open banks of a stream; 9 km. north- east of Lake Habbema, Brass 10559, October 1938, alt. 2800 m., few scattered clumps 1933, alt. 3800 m., amongst coarse tussock grass on deforested slope (stiff sedge in broad spreading tussocks). Brass 4386 was determined by Kiikenthal, in Bot. Jahrb. 69: 264. 1938, as C. Graeffeana Boott, from which it differs in the narrower leaves, fewer shorter coarser spikes, and rather longer utricles acute at the base and with more numerous nerves. It appears to be closely allied to C. Pandanus Ohwi, in Bot. Mag. Tokyo 56: 214. 1942, from Netherlands New Guinea, but has much longer leaves with more persistent sheaths, commonly cuspi- date glumes, and darker coloured, rather smaller, elliptical (not obovate) utricles punctulate in the upper part. It is also close to C. philip pinensis Nelmes, in Kew Bull. 1938: 109. 1938, differing in that the terminal spike is usually male, in the rather more slender spikes, in the ovate rather than obovate glumes, and in the smaller elliptic (not obovate) smooth paler coloured utricle which is less obtuse at the base On the type specimen, selected as such because of its full maturity, 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 109 there are 13 spikes and the glumes are muticous or shortly pointed. Brass 10734 measures 22 cm. to the top of the inflorescence and has 10 spikes rather shorter than those on the type (up to 3.5 cm. long), with which it otherwise agrees. Brass 4386 is similar to the type in appearance, but is not quite so mature; it carries 10 spikes and the glumes are rather prominently awned. Brass 10559, also with 10 spikes, is immature, and the uppermost spike has a few female flowers at the base Carex montivaga nom. Carex ae Kiikenth, in Bot. Jahrb. 69: 264. 1938; non Mackenzie in Bull. ot. Club 42: 617. 1915, Peace New Guinea: Lake Habbema, Brass 9032, August 1938, alt. 3225 m., common on wet sandy soil of open ridges; Lake Habbema, Brass 9339, August 1938, alt. 5 m., common on sandy patches on wet grassy Saues: Lake Habbema, Brass ee ae 1938, alt. 3225 m., alpine grassland, tufted in a sandy pool; 11 Carex tricholoma S. T. Blake: a. apex of culm with inflorescence, x T Seo. Zs " pistillate glume, flattened out, x 10; c. nut, xX 10. B. Carex brevis S. ; Blake: portion of plant, x 1; b. pistillate glume, flattened out, x 10; c. utricle, dorsal view, X 10. C. Carex euphlebia S. T. Blake: a. portion of middle part of inflorescence, X 1; 0. pistillate glume, flattened out, x 10; c. utricle, dorsal view, x 10; d. nut with style, x 10. Fig. A, a from Brass 4274; remainder from type-specimens. 110 JOURNAL OF THE ARNOLD ARBORETLM [VOL. XXVIII km. northeast of top of Mt. Wilhelmina, Brass & Mvyer-Drees 9759, Sept. 1938, alt. 3450 m., in sma Kiikenthal based his species on part of Clemens 6069, from Sarawaket in Northeast New Guinea. I have not seen the material, but the specimens cited above, particularly Brass 9339, agree so well with the description that there seems no doubt they are conspecific. The following notes on these specimens may be regarded as an amplification of the original description: Culms 18-42 cm. high; spikes 3-4, the terminal one male and 8-12 mm. ong, the others sinus and up to 12 mm. long (8-12 mm. on Brass’s specimens, 6-10 mm. on the type); the lowermost bract occasionally has a very short neat and the utricles vary from 3 to 3.6 mm. long. Carex tricholoma sp. nov. (Sect. agitfioi Kikenth.). Fro; 3; A; soi Ran aes Kiikenth. var, ciliato-marginata eed in Bull. Jard. Bot. . HT. 16: 318. 1940, in Bot. Jahrb. 70: 464. 1940; non C. ciliato-marginata ne Rhizoma breve. Culmi caespitosi, circa 15-30 cm. alti, graciles, trigon1, striati, prope apicem subtiliter scabri vel laeves, prope basin foliati. Folia omnia laminifera, culmo multo breviora, rigida, suberecta, sursum ee at angustata sed apice ipsa anguste praemorsa, conspicue carinata, marginibus recurva ceterum plana, prope apicem trique pie, carina et marginibus et facie superiore scabra; vaginae brunneae, hirtellae, emortuae in fibras reticulatas dissolutae. Bractea ima inflorescentiam adaequans, angusta sed vix setacea, breviter vaginans, ceterae breves vel squamiformes. Spicae 3-4, brunneae, contiguae vel ima breviter distans, sessiles vel breviter pedunculatae, interdum etiam spica basali longissime pedunculata addita; terminalis gynaecandra, clavata, circiter 10-12 mm. longa, inferne circiter 1.5 mm. sursum 3—4 mm. crassa; laterales femineae, oblongo-cylindricae, ore | mm. ¢ siflor 2-2.3 mm. longae, late ovatae, obtusae vel _fere cheat pln vel anguste hee Utriculi glumam a aa Se cieerantiog, 23-2 longi, 0.95 mm. lati, membranacei, straminei nervis valde obscuris, mar- ginibus lateralibus manifeste ciliati, faciebus glabri laevesque, elliptici, trigoni angulis lateralibus conspicuis, in rostrum breve conicum brunneum ore slings minute bidentulum marginibus ciliatum sensim contracti. Nux cinerea, 1.5 mm. longa, 0.9 mm. lata, o longo- -ovalis obtuse trigona, anguste oc. lateribus convexa, sub apice annulata constricta. Stylus brevissimus, basi pyramidalis; ins ata 3. Papua: Central Division: Mt. Albert Edward, Brass 4274, May-July 1933, alt. 3680 m., few plants amongst grass on banks of a grassland pond; Mt. Albert sonehens Brass 4316 (tyre), May-July 1933, 3680 m., sporadic in forest glades, not ¢ Carex tricholoma differs from C. montivaga S. T. Blake (C. Sides Kukenth. non saeneertae : the broader more scabrous leaves, prominently sheathing lowermost bra vnaecandrous terminal spike, and in the smaller very shortly nae utricle prominently ciliate on the lateral margins nearly throughout its length. Kiikenthal founded his variety solely on Brass 4316 and I have accepted a specimen of this number as the type of the species described above, but the excellent specimen Brass 4274 was found useful in elucidating some of the details. 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 111 A small plant in herb. Brisbane collected by M. S. Clemens on Mt. Sarawaket, at 9-13000 ft. in April 1937, may belong to this species, but the specimen is not sufficiently satisfactory for definite determination. Carex brevis sp. nov. (Sect. Mitratae Kiikent 2 oa Be Carex ses gd Kiikenth. var. oe ieee in Bull. Jard. Bot. ra 18. 1940, in Bot. Jahrb. 70: 464. 1940; non C. hispidula Gau Herba oa ea aapeth rhizoma ee. descendens a. ceealnst (inflorescentia inclusa) circiter 2-5 cm. alti, sub inflorescentia vaginis foliorum obtecti, basi Snot incrassati. Folia multa, rigida, patula admodum decurva, stricta vel plus minusve curva culmis multo longiora, nsim acutata, transversim secta plerumque V-formia, apice triquetra angustissime obtusa, 2—5.5 cm. longa, 3-4 mm. lata, a striata, inferne carina marginibusque dense subtiliterque scabra, sursum undique scabra; vaginae brunneae tandem in fibras plus minusve etiadiiias dissolutae. Bracteae foliiformes oe superantes, conspicue vaginantes, vaginis antice admodum pubescentes, superiores angustiores. Spicae 4 approximatae, breviter re atate, erectae, pallide brunneae vel fulves- ntes; terminalis mascula linearis 7-8 mm. longa, 1.5 mm. lata; laterales femineae, singulae, oblongae, 5.5-8 mm. longae, circiter 3 mm. latae, subdensiflorae, interdum spicis basalibus longepedunculatis additis. Pedunculi graciles scaberuli. Glumae femineae tenuiter membranaceae, .4-2.5 mm. longae, 1.8 mm. latae, ovatae, obtusae vel acutae, muticae, concavae, dorso lutescenti tenuiter 3-nerves vix carinatae, lateribus hyalinae fulvo-tinctae; glumae masculae plus minusve acutae, inferiores carina ae Utriculi fulvi, coriacei, 2. 5-2.6 mm ongi, 1. Sm m. la ti, obscure setts undique dense hispiduli. elli ptici, obtuse trigoni, basi leviter contracti. Nux straminea, 1.5 mm. longa, 0.9 mm. lata, ovoidea, obtusa, sub apice annulata abrupte constricta, Pa 3-costata, lateribus fere plana. Stylus basi pyramidalis; stigm oe Central Division: Mt. Albert saa Brass 4418 (tyPE), May-July 1933, alt. 3680 m., common in small tufts on a barren rocky ridge crest (leaves very stiff). Described by Kiikenthal as a variety of C. bulbostylis Kiikenth. non Mackenzie (C. montivaga S. T. Blake), but it differs greatly from that species in appearance, in the stiffer broad leaves, included culms, sheathing bracts, pedunculate female spikes with muticous glumes (usually not muticous in C. montivaga), and rather shorter very shortly beaked densely hispidulous yellowish utricles. The rhachilla is well developed and is sometimes so long as to be exserted from the utricle when it bears another utricle, sometimes well developed, sometimes rudimentary. At times the glumes appear cuspidate due to the inrolling of their upper margins. In Bull. Jard. Bot. Buitenz. IIT. 16: 318. 1940, Kiikenthal cites three collections under his C. bulbostylis var. hispidula, namely J. & M. S. Clemens 29005 and Clemens 32344 from Borneo and Brass 4418, without designating the type. In the same year, in Bot. Jahrb. 70: 464, the same description is given, but the collections are cited as follows: Polen Neu-Guinea: Mt. Albert Edward (Brass n. 4418!). rea: Brit. Nordborneo, Mt. Kinabalu (Clemens n. 29005!; 32344!).” 112 JOURNAL OF THE ARNOLD ARBORETUM (VOL. XXVIII Under these circumstances we may consider that in the latter publication Kiikenthal emphasised that Brass 4418 represented his variety, so that this may be regarded as the type of his trinomial. I have not seen the Clemens specimens. In any case the type of the binomial C. brevis is the sheet of specimens of Brass 4418 in herb. Brisbane. Carex breviculmis R. Br. Prodr. 242. 1810; Ridl. in Trans, Linn. Soc. II. Bot. 9: 247. Carex oo R. Br. subsp. Royleana Nees ex Wight sensu aia in Bull. Jard. . Buitenz. IIT, 16: 318. 1940, in Bot. Jahrb. 70: 464. PAPUA: cae Division: Murray Pass, Wharton i Brass 4697, June—Sept. 1933, alt. 2840 m., pale species eens on grassland slop I do not see how these specimens differ from te general trend of the species in Eastern Australia, unless the utricles are somewhat less pubescent. The length of culm varies considerably, sometimes even on the same plant, and evidently has no taxonomic importance in this species. I have seen no authentic specimen of subsp. Royleana, which is said to have somewhat scabrous culms; the other characters described are to be found in characteristic C. breviculmtis. Carex constricta sp. nov. (Sect. Frigidae Fries subsect. Decorae Ktkenth.). Fic. 4, A. erba pygmaea caespitosa stolonifera phyllopoda. Culmus 4 cm. altus, intra folia absconditus et horum vaginis obtectus. Folia numerosa culmum superantia, usque ad 5 cm. longa, 2-3 mm. lata, plana, sensim acuta, perrigida, carinata, sursum carina marginibusque scabra: vagin purpurascentes diu integrae sed tandem in fibras fissae. it ea e spicis 6 erectis constructa; spica terminalis mascula linearis circiter 7 mm. longa et 1.2 mm. crassa, breviter pedunculata; spicae laterales femineae, i l m. vel vix exsertis. Bracteae inflorescentiam superantes, foliiformes, promi- nule vaginantes. Glumae femineae oblongo-ovatae, truncatae vel bilobae, aristulatae, carina angusta basi 3-nervi virides, lateribus rufescentes aut ferrugineae, marginibus anguste hyalinae, 33-95 mm. longae Sp igihang usque ad 0.3 mm. longa inclusa). Utriculi oblique patentes, 2.3-2.6 m longi, circiter 1 mm. lati, membranacei, acute triquetri, leviter ounre lateribus fere plani, 2-nerves, glabri, minute rugulosi nitidule fusco-olivacei, inferne elliptici basin versus attenuati, sursum in rostrum obliquum circiter 0.7 mm. longum marginibus scabrum ore hyalinum obliquum integrum vel bidentulum subabrupte contracti. oe arcte inclusa, fere atra, ellipsoideo- ovoidea, tricostata, lateribus conve sub apicem ipsum dilatatum annula tum constricta. Stylus basi eee ee: stigmata 3. Papua: Central Division: Mt. Victoria, Owen Stanley Range, W. MacGregor in 1889 (type in herb. apie photographs of specimen with analytical drawings in herbs. Gray, Brisbane, and Kew). A single piece seen with one fruiting culm. In habit it appears to be close to C. celebica Kiikenth., differing (so far as can be judged from the description) in the truncate ‘mucronate (not acute) glumes and coarser spikes with more spreading utricles. Kiikenthal describes the spike of C. celebica as 1% mm. thick, but it is not clear whether this refers to the 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 113 flowering or the fruiting spike. The nut of C. constricta is similar to that of the Mitratae, but the geminate spikes and shape of the glumes do not accord with this section. Carex tricuspidata Kiikenth, in Bot. Jahrb. 70: 466. 1940. NETHERLANDS NEW GuINEA: Mt. Wilhelmina, 4 km. northeast of top, Brass & M yer-Drees 9970, Sept. 1938, alt. 3650 m., tufted on grassy seepages; Mt. Wilhelmina, 11 km. northeast of nae nee & Myer-Drees 9803, Sept. 1938, alt. 3400 m., covering the saucy bottom of a stream. ously known ae from Northeast New Guinea. The specimens cited ee belong to Kiikenthal’s species, though collectively they ed ST BLAKE Fic . Carex constricta 5. T. Blake: a. plant, x 1; 0b. pistillate one ee one x 10; c, d. utricle, dorsal and lateral views, x 10; e. nut with style, x . f. diagram of in Aonescenes ear in solid black, bracts stippled). B. Car rex 10; c. utricle, dorsal view, ; d. nut with style, x 10. All figures from a S. T. Blake: a. upper part - nay Ie xX 1; b. pistillate glume, flattened type-specimens. 114 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXVIII differ from the description in that the spikes are up to 6 cm. long, sometimes all distinct from one another (not up to 4 cm. long with the two uppermost female ones fastigiate), in that the utricle is only about 3.5 mm, (3.3-3.6 mm.) and not 4 mm. long, and in that the basal leaf-sheaths ultimately break up into fibres, while an elongated rhizome appears sometimes to be present. The inflorescence of Brass & Myer-Drees 9970 agrees pretty well with Kiikenthal’s description, but on this specimen the basal sheaths are broken up into fibres. In the other collection the plants are more elongated with longer and more distant spikes. The basal leaf-sheaths are but little disintegrated and evidence of an elongated rhizome is present. Carex brunnea Thunb. Fl. Japon. 38. 1784; Kiikenth. in Bot. Jahrb. 59: 60. 1924, 69: 265. 1938 : Anns New Gutnea: 9 km. northeast of Lake Habbema, Brass 10924, October oe alt. 2800 m., shrubby edges of a land-slip, rare; Bele River, 18 km northeast of Lake Habbema, Brass 11592, Nov. 1938, alt. 2200 m., occasional eee clumps on grassy, formerly cultivated slopes. New for Netherlands New Guinea, previously recorded from Northeast New Guinea. Known from ‘he Mascarene Islands, Ceylon, E. Asia and Japan, Indo-Malaya, Queensland, N. S. Wales, and New Caledonia. Carex acrophila sp. nov. (Sect. Hymenochlaenae Drejer subsect. Graciles Tuckerm.). Fic. 4, B Rhizoma brevissimum. Culmi caespitosi, erecti, cum inflorescentia 25-60 cm. alti, gracillimi (usque ad 0.9 mm. crassi), obtuse trigoni, sursum angulis admodum scaberuli, glabri, enodes, basi solum foliati. Folia prope ba : usque ad 1.6 m m. lata, in apicem longe filiformem sensim atten uata, ad superficiem Senet atque ad margines atque ad carinam superam debiliter laxiusculeque scabro-pilosula; vaginae fusco-brunneae vel fuscae, sg plerumque integrae, imae elaminatae. Panicula simplex, 7-15 cm. lon vaginantes, vaginis clare brunneis. Spicae 12-20, saturate brunneae, plerumque ternae vel binae (interdum singula), breviter pedunculatae, androgynae (pars mascula parte feminea brevior angustiorque), lineari- cylindricae, plerumque 1.5—3.5 cm. longae, 3 mm. crassae, subdensiflorae. vel aristulatae, ferrugineae marginibus anguste hyalinae, dorso crebre 3-nerves anguste carinatae, lateribus striatae. Utriculi canine glumas superantes, saturate rufo -brunnei, 5—5.5 mm. longi, 1-1.2 mm. lati, mem- branacei, nervis circiter 15 notati, praecipue parte superiore mers: Asal me ae angustius ovoidei, plano-c onvexi, inferne in stipitem graciliorem 1—1.4 mm. longum conspicue angustati, sursum in rostrum gracile fuscum bidentulum circiter 1.3 mm. longum subsensim contracti. Nux utriculo multo brevior, late elliptico-oblonga, plano-convexa, 1.75—2 mm. longa, 1.2 mm. lata, brunnea. Stylus tenuis, basi admodum incrassatus; stigmata 2, circiter 5 mm " NerHERLANDS New Gunga: Lake Habbema, Brass 9515 (Type), August 1938, alt. 3225 m., common about native camps; northern slopes of Mt. Wilhelmina, Brass & Myer-Drees 10072, Sept. 1938, alt. 4050 m., common on old grassy screes; Mt. 1947] BLAKE, CYPERACEAE COLLECTED IN NEW GUINEA, I 115 Wilhelmina, 7 km. northeast of top, Brass & Myer-Drees 9829, Sept. 1938, alt. 3560 m., alpine grassland, occasional erect tufts on wet ground; Mt. Wilhelmina, 7 km. northeast of top, Brass & Myer-Drees 9926, Sept. 1938, alt. 3560 m., plentiful amongst long grasses of a marshy slope, slender tufts 70-90 cm. high; Mt. Wilhelmina, 5 miles east of top, Brass 9409, August 1938, alt. 3440 m., common in marshy hollows This may be the form described by Kikenthal as C. brunnea Thunb. var. subteiogyna Kikenth., in Rep. Sp. Nov. 8: 8. 1910, from the Philippine Islands (not C. subteinogyna Ohwi in Mem. Coll. Sci. Kyoto Imper. Univ. ser. B, 11(5): 469, pl. 15, text-fig. 17. 1936), but it differs from C. brunnea in the nearly smooth rigid less flattened narrower leaves, the short peduncles, the longer glumes, the longer long-stipitate utricles less suddenly contracted into the beak and in the nut much shorter than the utricle. Carex acrophila is closely allied to C. teinogyna Boott, from India and Burma, but differs in the denser inflorescence, the shorter peduncles, the utricles with a longer stipe and more gradually narrowed to a shorter beak with shorter teeth, the shorter nut and rather shorter stigmas. It is also evidently close to C. spathaceo-bracteata Kikenth., in Bot. Jahrb. 70: 466. 1940, from Northeast New Guinea, differing in the less rough stems and leaves, 2—3-nate (not 3—5-nate) spikes, longer not conspicuously dilated sheaths to the bracts, paler glumes, and larger utricles hispid on the sides as well as on the margins. The cited collections form a very homogeneous series, varying from one another almost exclusively in the height of the plant. The nut is occasionally subtrigonous. Carex Brownii Tuckerm. Enum. a. 21. 1843. HERLANDS NEW GUINEA: m River, Brass 11791, Dec. 1938, alt. 1600 m., Aen in shallow grassy pools, ne tufts 50-60 cm. hi oh, New for New Guinea; known elsewhere from Japan, Korea, New South Wales, and New Zealand. Carex subtransversa C. B. Clarke in Philip. Jour. Sci. 2(C): 108. 1907. Papua: Central Division: Murray Pass, Wharton Range, Brass 4676, June—Sept. 1933, alt. 2840 m., a few plants at edge of forest (leaves pale, inflorescence erect). New for New Guinea; previously known from the Philippine Islands and Japan. Carex Doniana Spreng. Syst. 3: 825. 1826. Papua: Central Division: The Gap, Carr 13731, Dec. 1935, alt. 7200 ft., open ridge top (herb, 1 ft. tall) (herb. Canberra) New for New Guinea. This collection was cited along with Clemens 4451 and Clemens 5551 under the original description of C. japonica Thunb. var. mesogyna Kukenth., in Bot. Jahrb. 69: 265. 1938, but it differs from the description in that the terminal spike is wholly male, not female in the middle. It matches one piece of Ramos & Edano in Philip. Bur. Sci. 40359. There are two pieces under this number in herb. Brisbane, appar- ently representing the same form, the larger one agreeing with Kiikenthal’s description of C. japonica Thunb. var. chlorostachys Kiikenth., the second piece, matched by the Carr sheet, being smaller and with narrower leaves. Carex Doniana appears to be sufficiently distinct from C. japonica Thunb. 116 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII by reason of the longer cylindrical more shortly pedunculate always erect spikes and usually broader leaves. Carex subtransversa C. B. Clarke is very similar in appearance to C. Doniana, but has less spreading rather broader less distinctly nerved utricles with rather longer beaks, and the style is oblique and scarcely swollen at the base. Carex fascicularis Soland. ex Hook. f. Fl. Nov. Zel. 1: 283. 1853. NETHERLANDS New GuINEA: Lake Habbema, Brass 9211, August 1938, alt. 3225 m., lining open lake-shores, on hummocked marshy ground, large erect clumps 60-100 cm. hig New for New Guinea; previously known only from eastern Australia and New Zealand. There remain three collections of which the specimens are not sufficiently complete for satisfactory determination. Brass 9235, from sunny bogs, Lake Habbema, 3225 m., consists of small plants evidently allied to C. stenophylla Wahlenb., of the Northern Hemisphere, but the utricles are far too undeveloped for certain determination. Brass 11540, from the open banks of the Bele River, 2200 m., is an /ndocarex allied to C. filicina Nees and C, saturata C. B. Clarke. It almost certainly represents an unde- scribed species, but as the most advanced utricles are not quite mature it seems inadvisable to attempt to describe it until more mature material is available. And finally, Brass 4407, from a broken cliff face on Mt. Albert Edward at 3680 m., appears to represent an undescribed species of the Mitratae, but only very immature flowers and one overmature utricle are present on the specimen. QUEENSLAND HERBARIUM, BRISBANE, QUEENSLAND, AUSTRALIA. 1947] HOWARD, SOME PLANTS OF CUBA neg NOTES ON SOME PLANTS OF CUBA RicHArD A. HowARD With two plates THE PLANTS described in this paper were collected in Cuba in the sum- mer of 1941 by the author. Headquarters during this collecting trip was the Atkins Institution of the Arnold Arboretum, at Soledad in southern Las Villas province. I am grateful to Mr. David Sturrock and the admin- istration of the Soledad gardens for their many courtesies and for the privileges of using the facilities of the station. Most of my collections during that summer were made in the San Juan and Trinidad Mountains and in southern Las Villas province. One short trip was made to the Oriente province with Brother Leon, of the Colegio de la Salle of Havana, and many of the new entities described in this paper were collected then. The trip to the Oriente and the territories in which collections were made were described by Brother Leon under the title, “Excursion Botanica a las Tierras de Moa” (De La Salle No. 225: 23-29. 1941). The collection numbers cited are my own. I am grateful to Brother Leon for the pleasure of accompanying him on that trip and also to Mr. George Bucher and Mr. J. R. Grist for the hospitality and assistance offered during our stay at Moa. The description of a new species of Neobracea was supplied by Dr. Robert E. Woodson, Jr., of the Missouri Botanical Garden, and I am grateful to him for examining this material and supplying the name. Rajania nipensis sp. nov. PraTE I, Fics. 1-8. Planta volubilis glabra; foliis 5-7 cm. inter se remotis; petiolis 3—4 cm. ied teretibus; ace late triangularibus vel ovato- -triquetris vel trilobis, m. | atis, basi rotundatis vel truncatis, lobo medio lineari- oe e basi cm. longo, 1 cm. lato, 3-nervio, nervis lateralibus medio parallelibus margine supra medium confluentibus; ‘lobis lateralibus oblongis vel rotundis, cum nervis 2 (3) arcuatis; inflorescentiis 2 vel 3 axillaribus, racemosis vel spicatis, 2.5—3.5 cm. longis; floribus congestis, cymis brevibus 0.5 mm. longis, 1—3-floris; perianthii segmentis ovatis, 6; staminibus 6, usque 0.3 mm. longis, antheris ovatis, introrse dehiscentibus. ORIENTE: als de Nipe near Woodfred, in flower July 28-29, 1941, Howard 6136a (GH, TYPE). Rajania eee is similar to R. prestoniensis Knuth. It differs in hav- ing leaves which are rounded or truncate at the base and without a sinus. In addition, the middle lobe is 3-nerved and attenuate, not mucronate as in R. prestoniensis In the collection cited above is a single pistillate specimen tangled with 118 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII the staminate shoots selected as the type. ‘This pistillate specimen has the mature leaves elongate-ovate, with the basal lobes rounded or poorly defined. The size and venation are approximately the same as in the type specimen. The pistillate inflorescences are fascicled in the leaf-axils, 3—5 being present per axil. The branches are racemose and 7—10 cm. long, although one or two measure 19 cm, in length. The individual flowers are solitary on pedicels nearly 2 mm. long which bear a lanceolate bract near the base. The flowers are 1—-1.5 mm. high, with a 6-parted perianth. The perianth-segments are ovate and average 0.5 mm. in length. ‘There are 6 staminodes present in the flower and these are minute and bifid at the apex. The three styles are well developed and also bifid. The ovary is strongly 6-lobed, the alternate lobes being large and small. Dissection of the small lobes reveals the locules, which are almost lost in the mass of sterile tissue. One locule is usually larger than the others and the ovules in this appear turgid. The other locules also have two superposed ovules but these are shrunken. Only a faint indication of a lateral gibbosity is developed in the specimen at hand; however, the condition of the ovules indicates that but one locule will develop and that the specimen is a true Rajania. The young leaves are distinctly punctate-glandular, but the glands are fairly well masked in the mature lamina. Rajania linearis (Griseb.) [Uline in msc, ex Knuth in syn.] Howard, comb. nov. PiaTeE I, Fic. 9. Dioscorea linearis Griseb. Cat. Pl. Cub. 251. 1866. IENTE: Sierra de Nipe, 15 km. south of Woodfred, Howard 6096 &. Collected in flower July 28-29, 1941. The specimens cited above are staminate, but they appear to be identical with a pistillate specimen collected by Wright (3254, type number of D. linearis Griseb.) in the Gray Herbarium. ‘right’s specimen bears an annotation, ‘“‘Rajania linearis (Griseb.) Uline det. Uline.” In Knuth’s monograph of the Dioscoreaceae (Pflanzenreich IV.43: 168. 1924) this Wright number is divided. One pistillate portion is cited as Dioscorea linearis Griseb. and another pistillate specimen is referred to Rajania microphylla Knuth. The species of Rajania, however, is typified by staminate material collected by Poeppig, and no description of the cited pistillate specimen is given. I have not seen the Poeppig specimen nor any fragments of the Wright material that could possibly be referred to the description of that staminate plant. The specimen of Wright 3254 at the Gray Herbarium consists of a shoot with a series of female flowers and young fruits attached. These are definitely Rajania fruits, having one carpel developed into a wing, with the stylar portions asymmetrically placed. On the same sheet are two infruc- tescences bearing fruits that are Dioscorea fruits and which have been annotated by Uline as D. tamoidea Griseb. These agree with the description given by Knuth for the fruits of D. linearis. The specimen selected by Knuth as the type of D. linearis was in the 1947 | HOWARD, SOME PLANTS OF CUBA 119 Berlin herbarium and is presumably destroyed. It seems necessary to conclude that his description of mature fruits of D. linearis was based on the unattached fragments of this Wright collection, which are fruits of D. tamoidea Griseb., as Uline note The pistillate inflorescence of Rajania linearis is short and bears few flowers. Mature inflorescences of the specimen on hand are 1—1.5 cm long and bear 3—5 flowers, of which 2 or 3 mature on the average. The one-winged samaras are smaller than those of the majority of species, being about 4-5 mm. wide and 8-10 mm. long. A solitary small seed is developed. Dioscorea nipensis sp. nov, (Sect. Lychnostemon Uline). Prate: IT, Fre. 1. Planta volubilis, 4-6 m. alta, glabra; caulibus 1 mm. crassis, striatis; foliis 8-10 cm. inter se remotis, alternis; petiolis teretibus 3 cm. longis; laminis circumscriptione angusto- ovatis, veriter elongato- triquetris, 6-7 cm longis, 2.5—3 cm. latis, apice attenuatis, basi rotundatis, sinu nullo, auriculis rotundatis, nervis 5, arcuatis; inflorescentiis ¢ 23-27 cm. longis, spicatis, floribus in capitulis ‘dispositis, 2—4-glomerulatis, glomerulis 0.5—1 cm. inter se remotiusculis, bracteis 0.7 mm. longis, lanceolatis, perianthio 1.5 mm. diametro, lobis ovatis: staminibus fertilibus 3, usque 0.7 mm. longis, fila- mentis carnosis, antheris oblongis, thecis late distinctis, extrorse dehiscenti- bus; antheris sterilibus 3, usque 0.4 mm. longis, quam fertilibus brevioribus, perianthii lobis interioribus oppositis, antheris apice acutis, basi hastatis; rudimento stylino 3-lo ORIENTE: 15 km. the of Compania de Moa mill, Moa, in flower July 25, 1941, Howard 5873 (GH E). This species is se to D. tamoidea Griseb. It may be distinguished from that species by the leaves, which have rounded bases without sinuses, and by the longer narrow blades Like those of most specimens of Dioscorea, the leaves turn black on drying. The outer pair of lateral veins frequently fork very close to the e Rajania tenella sp. nov. Pirate II, Fic. 2 Planta volubilis glabra; petiolis teretibus, gracilibus, 1.5—2 cm. longis; laminis elongato- -triquetris, 5.5-7 cm. longis, 3.3—4 cm. latis, apice attenu- atis, basi truncatis, sinu nullo, lobis obtusis, exs. papyraceis, nervis 3, medio a lateralibus parum distincto, lateralibus arcuatis margine confluentibus: inflorescentiis solitariis vel binis axillaribus, 2.5—4 cm. longis, leviter flex- uosis, cymis lateralibus 3-4 mm. pian 4—9-floris, floribus minutis, perianthio 1 mm. diametro, lobis 0.4—0.5 mm. longis, staminibus 6, usque 0.2 mm. longis, ene dehiscentibus; a ranean stylino subnullo. OrIENTE: Sierra de Nipe south of Woodfred, in dense woods, with yellow flowers, July i 29, 1941, pees 6134 (GH, TYPE). Rajania tenella is similar to R. theresensis Uline and R. prestoniensis Knuth, which I know only from the descriptions. It differs from R. prestoniensis in the shape of the blade, the mature leaves being without sinuses and attenuate rather than mucronate. It is unlike R. theresensis in having the leaves 3-nerved, the inflorescence much shorter than the 120 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII leaves, and the flowers smaller. It is easily distinguished from R. nipensis by the truncate base of the mature leaves. Considerable variation occurs between the juvenile leaves and the mature leaves in most species of Rajania (Plate IJ, fig. 2). For example, the juvenile leaves of the present species are elongate-ovate with the basal lobes more or less developed. The laminae are 2.5—3 cm. long and 0.3-1 cm. wide. There are 3 primary veins, with the lateral veins arcuate. The apex is long-attenuate and the base is truncate or rounded. The petioles are terete and from 0.6-1.2 cm. long Because of this variation with age, care must be taken to collect and describe only mature specimens. Frequently the basal portion of a vine may be mature, but near the apex long axillary shoots, perhaps 4—9 feet long, may appear with only juvenile leaves. It is very easy to snatch immature material when such plants grow commonly in dense thickets. Platygyne volubilis sp. nov. PuaTeE II, Fics. 3-5. Frutex monoicus scandens; ramis molliter ferrugineo- -pubescentibus; foliis alternis, petiolatis (petiolis teretibus 4-6 mm. longis ferrugineo- pubescentibus) stipulatis ovatis, pubescentibus; laminis obovatis vel ob- lanceolatis, 5—6.5 cm. longis, i-- 2.5 cm. latis, rigide membranaceis, basi cuneatis, apice acutis vel rotundatis, margine sinuatis, sparse dentatis, supra et margine urenti-pilosis, subtus sparse hispidis, nervis lateralibus 5, arcuatis, anastomosantibus; inflorescentiis foliis oppositis, ¢ racemosis, ramis usque 3-4 mm. longis, 3—5-floris, pedicellis 4—5 mm. longis; sepalis 4, usque 3.5 mm. longis, triangularibus vel lanceolato-ovatis, reflexis, extus dense strigosis, intus sparse strigosis; corollis nullis; staminibus 5-8, mentis crassis, usque 1.5 mm. longis; rudimento pistilli see inflorescenti 2 glomeratis vel breviter racemosis, ramis ad 2 mn _ longi 3—5-floris, pedicellis brevibus, 2 mm. longis; perianthiis 6—9- lobatis, lobis variabilibus, lanceolatis vel ovato-oblongis, 4—5 mm. longis, urenti-ciliatis, extus strigo- sis; staminium rudimentis nullis; stylis 3, usque e 5 mm. longis, oblongis apice bifidis carnosis, ferrugineo-strigosis; ovario globoso trilobato, sparse urenti-piloso; fructu tricocco 5 mm. longo, apice sparse piloso, pedunculo usque 4 mm. longo; columella 3 mm. alta; seminibus rotundis, 3 mm sae ORIENTE: 20 km. west of Moa, in open pine woods, in flower and fruit, July 26, 1941, Howard 5961 (GH, TYPE). Most of the plants of Platygyne that I had seen growing in the central and western provinces of Cuba had been low shrubs, semi-scrambling over low bushes and in fence rows. It was quite a surprise, therefore, to find in the pine woods west of Moa a specimen of Platygyne climbing in the branches of the tall pine trees. The specimen also had distinctive obovate leaves with cuneate bases, short inflorescences, and the fruits were almost without the stinging hairs so common in the only other species of Platygyne recognized at present, P. Aexandra (Jacq.) Muell. Arg. The indument in specimens of Platygyne is of interest because of the several types of hairs found on different parts of the plant. Of most interest to the collector are the short white stinging hairs found only on the leaves, perianth, and on the mature fruits of P. hexandra. This is quite 1947] HOWARD, SOME PLANTS OF CUBA 121 in contrast to the other common “ortiga” of Cuba, Tragia spp., in which stinging hairs are abundant on the stems as well as the leaves and inflores- cences. In Platygyne volubilis the stinging pubescence is found only on the upper surface of the leaves and along the margins as cilia. Th hairs on the surface of the leaf are usually single, while the ciliate margin consists of clusters of hairs. The pistillate perianth also has a ciliate margin of stinging hairs, but this is lacking on the staminate perianth. Finally, the fruits of P. kexandra are usually densely covered with the white pubescence of stinging hairs. In contrast the fruits of P. volubilis have only a few, if any, apical stinging hairs, and the rest of the capsule is covered with a very short ferrugineous indument. The pubescence on the rest of the plant, including the perianth, staminal disk, and styles, consists of ferrugineous short pointed stiff hairs, best described as strigos The pistillate aie is extremely variable. Baillon, Euphor. 453. 1858, believed the perianth consisted of two whorls, while Mueller, in DC. Prodr. 15: 913. 1864, called them all calyx-segments of one whorl. The outer segments are usually linear-lanceolate and uniform, The inner segments vary greatly in size and shape and are frequently fused and 3—4-lobed at the apex. The seeds are 3 mm. in diameter and almost spheroid. Mature seeds of P. volubilis are gray and are lightly mottled with a dark brown color. The seeds of P. Aexandra in all the fruiting specimens I have seen are tan, mottled with dark brown. The seeds of Platygyne do not have an aril, although a linear protuberance or hilum is present for about one-third the circumference of the seed. Phyllanthus chryseus sp. nov. Frutex parvus monoicus; ramis paenultimis teretibus robustis usque 5 2.5—-3 mm. latis, castaneis, nitidis; ramulis foliigeris 20-28 cm. longis teretibus, 2 mm. crassis, 16—26-phyllis; stipulis minutis usque 1 mm. Ap triangularibus vel lanceolato- subulatis; foliis distichis, petiolis 2.5-3 mm longis, 5 mm. latis, carnosis, staminibus coalescentibus vel 1 mm. longis, sporanglis 4, antheris late ovati is, 0.7 mm. longis, extrorse et transverse dehiscentibus; floribus 9 solitariis axillaribus, pedicellis 1.3 cm. longis, sepalis 4, ellipticis, 5 mm. latis, 4 mm. longis, apice rotundis; disco obscuro, ovario 5 mm. nee. stylis . fimbriato-laceratis, 4—9 partitis; fructibus tricoccis usque 5 mm. longis; ae oblongis, hebeti-nigris, profunde et crasse reticulatis, 7 mm. lon OrIENTE: 2’ shrub, common in —— along ravine 15 km. southwest of Moa, leaves golden color, in flower and fruit July 26, 1941, Howard 5829 (GH, tyre); Moa, Bucher 75 (NY); Rio Cabanas, Moa, May 27-31, 1943, Marie-Victorin & Clement 21755 This was a very striking plant in the field, for the brilliant golden leaves 122 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII were in contrast with the red iron soil on which it was growing. On drying the leaves turn light green above and pale brown beneath. This species is distinct from others in the Orbicularia section of Phyllanthus by its large orbicular leaves, which are numerous for their size. The bright golden color and strongly revolute margins of the leaves, the small stipules, and the short triangular bracts distinguish this species from others of the cinctus group as established by Carabia, in Ecological Monographs 15: 331. 1945, In gross appearance Phyllanthus chryseus appears similar to P. sub- carnosus Griseb., which is based on fragmentary material. An isotype of the latter species at the Gray Herbarium has a single fruit with unusually heavy capsules for PAyllanthus, and the seeds are smooth and brown. The dull black, coarsely and deeply reticulated seeds of the present species are distinct from any I have seen in this genus, but so few seeds have been described that I can not evaluate this character. Eugenia moaensis sp. nov. Frutex glaber; ramis hornotinis teretibus vel subcompressis, ferrugineis; petiolis 2-4 mm. longis, 1.5 mm. crassis, sulcatis; laminis lanceolato-ovatis vel ellipticis, 5.5-6.5 cm. longis, 2—2.5 cm. latis, basi et apice acutis, apice nacre coriaceis, nervo medio supra leviter impresso, subtus promi- nte, nervis lateralibus eset aoa. margin harper inflorescentiis Sibi, pedicellis solitariis vel 2—3-fasciculatis, 6-10 mm. longis, brac- teis minutis, 2, ciliatis; calvcis lobis 4, ovatis, 2 mm. longis si ‘latis, ciliatis; petalis 4, oblongis, 4 mm. longis, 1-1.2 mm. latis, ciliatis; staminibus numerosis 2 mm. longis; stylo 3 mm. longo; ovario 3-loculato, ovulis numerosis; baccis 5-7 mm. diametro, pericarpio crasso succulento, dense glanduloso; seminibus 1 vel 2, rotundatis vel subreniformibus 3—5 mm. longis, testa brunnea membranacea nitida. ORIENTE: ba shrub, flowers yellow, fruits green, along creek bank at Moa, Howard 5866 (GH, TYPE). The ne of the present species are thick and coriaceous. The upper surface is shiny when dry and darker in color than the lower surface. Punctate glands are present in the lamina but are well hidden in the tissues of the leaf. Approximately 6 strong lateral veins are revealed on the upper leaf-surface, and these anastomose near the margin. ‘The veins are obscure below. Eugenia moaensis appears to be similar to E. (?) Ossaeana Urb., which was described on very incomplete material collected by Shafer in the same locality as the present specimen. It is different in being completely glabrous, except for a ciliate perianth, and the venation of the leaves does not agree with the description as given by Urban. ‘The pedicels of the present species are much longer than those of E. (?) cupuligera Urb., to which this plant is also similar. Eugenia moaensis may also be related to EF. cristalensis Urb.; however, it differs from this in the leaf-characters and the leaf-size. Calyptranthes oblongifolia sp. nov. Frutex; ramis hornotinis compressis, sub interstitiis foliorum plus 1947] HOWARD, SOME PLANTS OF CUBA 123 minusve sulcatis, pilis dibrachiatis adpressis ferrugineis vel pallescentibus vestitis, annotinis glabrescentibus teretibus vel lineato-striatis, dichotomis: petiolis 4-5 mm. longis subteretibus supra s sulcatis; laminis oblongis vel oblongo-lanceolatis, 4.5—7 cm. longis, 1.5—2 cm. lati is, basi rotundatis, antice acutis vel rotundatis, nervo medio supra per totam longitudinem anguste impresso, nervis lateralibus ca. 15 inconspicuis; inflorescentiis ad basin ramorum hornotinorum binis, subcapitatis; pedunculis 2-4, 2~5 mm. longis, capitulis 3—5-Hloris; baccis nigro-purpurascentibus, depresse sphaeroideis, 5-7 mm. qaneee loculis 2; seminibus reniformibus 0 in quoque loculo, 3 mm. longis ORIENTE: ee g south coast of Bahia de Moa near Moa, 4’ shrub, fruits green je purple on maturity, Howard 5948 (GH, type). Calyptranthes oblong rifolia is characterized by the oblong or lanceolate- oblong leaves and the few-flowered, very short peduncles of the seemingly capitate, axillary inflorescence. In general appearance this species is similar to C. rupicola Urb. and C. heterochroa Ur The occurrence of malpighiaceous or two-armed hairs on the plant is characteristic of several species of Calyptranthes. In the present species these hairs are very dense on the young leaves, the stems, peduncles, flowers and fruits, giving the appearance of a ferrugineous tomentum. In the older leaves the hairs are broken off even with the epidermis and leave the sunken hair-bases commonly found in the other species and which frequently seem to be mistaken for epidermal glands. The upper surface of the mature leaves of C. oblongifolia is glabrous except for the impressed midrib, while the hairs on the lower surface persist. These hairs usually remain ferrugineous in color, although many turn gray or white. The inflorescence consists of short axillary peduncles with a very few flowers and appears to be capitate. The bracts subtending the flowers are minute and fall away very early, none appearing in the mature infruc- tescences. ‘The flowers had matured on the specimens studied, but several young fruits still possessed 3 or 4 stamens attached to the hypanthium at the line of dehiscence of the calyptra. The style was 2 mm. long and bifid at the apex. The calyptra was conical and about 1.2 mm. in diameter. The drupes are a dark purple color when fresh, drying black or brown. Because of the dense pubescence, many appear to be silvery in reflected light. There are 2 cells to the ovary, each containing 2 seeds. The seeds are reniform and tan on the convex surface and black or dark brown on the sides and in the concave surface. The embryo is curved. Psidium confertum sp. nov. Frutex vel arbor parva; ramis hornotinis plus minusve compressis tenui- bus vix 1.5 mm. latis, tomentosis, pilis crispis, vetustioribus ferrugineis; ase 2—2.5 mm. longis, laminis ovatis vel ellipticis, 2—2.8 cm. longis 1.4-1.6 cm. latis, basi rotundatis, apice rotundis vel parce emarginatis, margine recurvo, nervo medio supra impresso, subtus prominente, nervis 4—6 mm. longa, pedunculis 0.8—1.1 cm. longis; calycibus 4-lobatis, leviter imbricatis, tomentosis, lobis ovatis 1-2 mm. longis; petalis 4 vel 5, ovatis, 124 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII extus sparse pubescentibus, dense glandulosis; staminibus numerosis; stylis 2 mm. longis; ovario 2-loculari, ovulis in quoque loculo ca. 10; seminibus reniformibus 2 mm. longis; fructibus immaturis, stylis usque 9 mm. longis. Or1ENTE: Wet hillside near the airfield at Moa, 12’ shrub collected in flower and young fruit July 26, 1941, Howard 5901 (GH, Type). Psidium confertum is characterized by the small ovate to elliptic leaves, the persistent ferrugineous tomentum, the cymose inflorescence, and the 2-celled ovary. In these characters it differs from P. bullatum Brit., P. ophiticola Brit. & Wils., and P. leiophloeum Urb., to which it is related. The species is named for the crowded condition of the flowering shoots. The internodes are from 1-1.5 cm. long with the leaves opposite or in whorls of three at the nodes. The compact inflorescence adds to the density of the shoots. The leaves are densely glandular; however, the glands are hidden by the persistent tomentum on the lower surface and are not visible from the dorsal surface. The very young leaves are densely pubescent on both surfaces. The lower surface remains persistently pubescent, while the upper surface is glabrous and becomes a bright yellow-green color on maturity. Two small bracts subtend the flower. The buds are globose above the ovary. Mature fruits are 5-6 mm. in diameter, and the numerous seeds are reniform, about 2 mm. long. On the concave surface of each seed extending over the sides are areas which are dark brown to black in color. The remaining surface of the seed is a yellow-brown. The embryo is only slightly curved. Eugenia Sturrockii sp. nov. Arbor parva vel frutex, glaber; ramis hornotinis compressis, sub inter- stitiis foliorum plus minusve sulcatis; petiolis 3-4 mm. longis, subteretibus supra sulcatis; laminis anguste lanceolatis 4.5—5 cm. longis, 0.8—1 cm. latis, basi et apice acuminatis, coriaceis glabris, supra nitidis, marginibus revo- lutis, nervis lateralibus inconspicuis; inflorescentiis axillaribus, racemosis, racemis usque 2 mm. longis, bracteis ovato-lanceolatis, minutis: 4-5-floris, quoque — 2.5 mm, cca testa membranacea, levi, brunnea, nitida. ORIENTE: shrub or small tree in dense woods of ravine, 15 km. southwest of Compania de ee mill, Men. tas July 26, on: Howard 6034 (GH, TYPE) This species is named in honor of Mr. David Sturrock, who was Super- intendent of the Atkins Institution of the Arnold Arboretum at Soledad, Cuba, for many years and who was so very helpful in arranging my trip to the Oriente. This species at first glance is similar in appearance to E. psiloclada Urb. However, the leaves of the new species are coriaceous, the veins obscure, the upper surface shiny, and the margin strongly recurved. Eugenia psiloclada, by contrast, has thin leaves with veins and glands prominent, 1947] HOWARD, SOME PLANTS OF CUBA 125 dull surfaces, and perfectly flat margins. The leaves of E. Sturrockii have the characteristic glands, but these are buried in the leaf-tissue and are visible only when the leaf is held to the light. Several of the fruits examined had only one mature seed. Closer examination revealed an aborted cell and ovule at the apex of the fruit. The flowers are articulated to the pedicel, the fruits dehisce at this point very readily, and the peduncles persist even on the old shoots. Neobracea Howardii Woodson, sp. nov. Frutex ramosus ca. 2 m. altus, ramis teretibus cortice luteo, ramulis puberulis tandem glabratis. F olia opposita petiolata, lamina late elliptica apice obtusa basi latiuscule acuta 4—7 cm. longa 2—3.5 cm. lata firmiter subterminalis pluriflora, ec tenui ca. 3-4 cm. longo puberulo, pedicellis ca. 0.5 cm. longis puberulis, bracteis lneari- abldneeolatis usque 0.5 cm. longis nee Flores albi, calycis laciniis anguste lanceolatis acuminatis 0.5 cm. longis subfoliaceis ut in pedicello vestitis intus in sinibus 0-—2-glandulosis, corollae subinfundibuliformis tubo ca. 0.9 cm. longo basi ca. 0.1 cm. diam. faucibus usque 0.25 cm. diam. ampliatis extus sparse irregulariterque eee uns prope basim staminigero, lobis late oblongo- dolabriformibus ca. 0.8 cm. longis. Antherae subsessiles lineari-lanceolatae basi anguste ee ca. aa 3 cm. longae dorso dense puberulae. Ovaria oblongoidea glabra ca. 0.15 cm. longa, nectaria 5 compresse subquadrata ca. 0.05 cm. longa, stylo crassiusculo ca. 0.2 cm. longo, stigmate capitato 5-maniculato ca. 0.05 cm. longo apice papillato. Folliculi immaturi tenues remote moniliformes. Las Vittas: Limestone Loma Ventana, Trinidad Mountains, San Blas—Buenos Aires, Aug., 1941, Howard 6495 (Herb. Missouri Bot. Gard., TYPE). This species is very well marked by its abundant Gives ace and broad, elliptic foliage, as well as the several-flowered inflorescence. Since the publication of my revision of Neobracea (in Ann. Missouri Bot. Gard. 23: 169. 1936), I have examined a flowering specimen of NV. Ekmanii Urb., the type and heretofore the only known specimen of which was in fruit, and that apparently mislaid when I visited the herbarium at Dahlem in 1930. The flowering specimen was collected by Mrs. G. C. Bucher during the summer of 1939 near Moa, Oriente, Cuba. The corolla, apparently og is very broadly ffendibulitornt the proper tube 6 mm. long, 1 mm. in diameter at the orifice, the lobes 10 mm. long. The calyx- lobes are Te lanceolate, 2 mm. jong. Bonnetia cubensis (Brit.) comb. Kieseria cubensis Brit. in Bull. Torr. Bot. Club 41: 19. 1914. Bonnetia Mart. & Zucc. has been conserved (Kew Bull. 1940: 113. 1940) and Kieseria Nees has been rejected. A new combination is needed for this Cuban species. The species is apparently endemic along the coast of Bahia de Moa on serpentine soil. The type collection was made in 1914 and only two more recent collections are known, Howard 5842, a 15’ tree with pink flowers 126 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII collected July 26, 1941, in dense woods 15 km. southwest of Compania de Moa mill, and Marie-Victorin, Clement & Alain 21436, collected in flower April 1943 at the mouth of the Punta Gorda River at Moa. Buxus flaviramea (Brit.) comb. Tricera flaviramea Brit. in Bull Torr. Bot. Club 42: 499. 1915. A second record of this species is Howard 5100, collected June 15, 1941, in open grassy savannah 10 km. west of Santa Clara city, Las Villas Prov. Xyris Jupicai Rich. in Act. Soc. Hist. Nat. Paris 1: 106. 1792. This is a very widely distributed species through the Antilles, but the present collection (Howard 6011, in flower, July 26, 1941, airfield at Moa, Oriente) is of interest since it is the first representative of the genus known to me from the Oriente. The many species of Xyris in Cuba are centered in Pinar del Rio province, with two species reaching the Cienaga swamps in southwestern Las Villas province. The present collection consists of larger and more tuberculate plants than the average, but the specimens still fit within the limits of the species as defined by Malme in N. Am. FI. 19: 11. 1937. Cassia clarensis (Brit.) comb. nov Chamaecrista clarensis Brit. in N. Am. Fl. 23: 278. 1930. Cassia insularis (Brit. & Rose) comb. Chamaefistula algee Brit. & sata in N. Am. FI. 23: 239. 1930. In addition to those from the type localities, specimens of these two species are known from the Castillo de Jagua and the coast of Cienfuegos Bay. Society or FELLOws, HARVARD UNIVERSITY. EXPLANATION OF PLATES PLATE I Fics. 1-8. Rajania nipensis Howard (Howard 6136a): _ habit, @ single flower from portion of 9 inflorescence, x 6; sagittal section ap flower showing two superposed fertile ovules, x 8; 6. habit, ¢ x 2; 7. portion of ¢ inflorescence, x 3; 8. @ flower showing pistil- vadinent and 6 introrse fertile stamens, x 15. Fic. 9. Pua linearis (Griseb.) Howard (Howard 6096): habit, x WY. Pirate II Fic. 1. oscorea nipensis Howard (Howard eee habit, x Y%. Fic. 2. Rajania ers Howard (Howard 6134): habit Y%, Fics. 3-5. Platygyne volubilis Howard (Howard 5961): 3. habit, x %; < ames view of fruit, x 1.5; 5 lateral view of fruit, Jour. Arnotp Ars. VoL. XXVIII Prate I SomME PLANtTs oF CUBA Jour. Arnotp Ars. VoL. XXVIII Prare II SoME PLANts OF CUBA 1947] LI, THE GENUS BRANDISIA 127 RELATIONSHIP AND TAXONOMY OF THE GENUS BRANDISIA Hutr-Lin Li Tue cEeNus Brandisia is confined to the warmer parts of continental eastern Asia from eastern India, Burma, and Indo-China to southwestern and southern China. It has been included in the Scrophulariaceae, usually in the tribe Cheloneae, together with other anomalous genera like Paulownia and Wightia. Its taxonomic position has been doubted ever since it was first described, but no extensive and intensive studies have ever been made on the genus. The following study, based on the Chinese species of the genus, is made on materials assembled from various herbaria in America. ‘To the curators of these herbaria I am indebted for their kindness in furnishing material for the study. This study was made at the Academy of Natural Sciences of Philadelphia during the tenure of a Harrison Fellowship for Research of the University of Pennsylvania. To Mr. Charles M. B. Cadwalader, President of the Academy, I wish to express my thanks for permitting me access to the splendid library and herbarium facilities. To Dr. Francis W. Pennell, Curator of Plants, my thanks are due for his suggestions and his kindness in reading the manuscript. Abbreviations for the various herbaria cited are as follows: AA Arnold Arboretum of Harvard University ANSP Academy of Natural Sciences of Philadelphia B Bonati Herbarium, to be deposited at the herbarium of the University of California at Los Angeles MBG Missouri Botanical Garden NYBG New York Botanical Garden RBGE Royal Botanic Gardens, Edinburgh uC University of Calitcuain USNH United States National Herbarium RELATIONSHIP The tribe Cheloneae of the Scrophulariaceae is very artificial and heterogenous. To this tribe have been referred several of the woody genera of various localities, while the family Scrophulariaceae as a whole is herbaceous. The anomalous positions of Paulownia and Wightia have been questioned by some authors. Hallier (in Bull. Herb. Boiss. II. 3: 181-207. 1903), in making a review of the taxonomic relationships of the various genera of the Scrophulariaceae, concluded that these two genera should be assigned to the Bignoniaceae. Campbell (in Bull. Torrey Bot. Club 57: 47-50. 1930) reached the same conclusion for Paulownia. He 128 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII indicated that the small amount of endosperm present in the seed is not sufficient reason to exclude it from the Bignoniaceae, with which it agrees in many other characters. Hallier, however, still considered Brandisia as belonging to the Cheloneae and evidently had not made any special study of the genus. When Hooker and Thomson first described the genus no fruiting material was available. They noted the general appearance of the plant to be very different from the Scrophulariaceae, and rather verbenaceous. However, they placed it in the Scrophulariaceae because of the many-ovuled capsule. They remarked: “The order can, in fact, only be approximately determined, as the dehiscence of the fruit and the structure of the seed are unknown. If the seeds be exalbuminous, it will go to Bignoniaceae; but the ovary and placentation are so much like Scrophulariaceae, that in all probability albumen will be found in the seeds. With regard to its position in the Scrophulariaceae, they said: “The position of Brandisia among Scrophularineae, as Mr. Bentham has kindly pointed out, is probably among Cheloneae, the only tribe in which large woody climbing plants occur. e flowers though solitary, have two opposite bracteoles on the pedicel, and the curiously elongate ovules probably become winged seeds like those of Wightia.” In the Genera Plantarum, the fruits and seeds were described and the genus was placed in the tribe Cheloneae together with Wightia and Paulownia. This has been followed subsequently by other authors. In 1873, Kurz (in Jour. As. Soc. Beng. 42(2): 39-141. 1874), in listing Brandisia discolor for the Burmese flora, roa as the status of the genus and made the following statement: “Wightia, Wall., arcte affinis. Etiam Buddleiae generi affinis, sed differt corolla a are etc. et certissime inter Sesameas recipienda est. Gardneria, a cl. Benthamio Loganiaceis adnu- merata, Solanea esse videtur.”” He placed the genus under ‘“Pedalineae.” In 1877 Kurz (For, Fl. Brit. Burma) listed Gelsemium and Buddleia under Sica ” in addition to Brandisia. Both these genera belong to the ee ea Hance (in Jour. Bot. 18: 299. 1880) mentioned that Brandisia “has a curious resemblance to the genus Eremophila, in Myoporaceae.”” This is an Australian genus. The Myoporaceae is a small family, with the genus Myoporum extending from the tropics of the Old World to India and eastern Asia. Thus we have suggested relationship for the genus Brandisia to the following families: Loganiaceae, Solanaceae, Scrophulariaceae, Bignoni- aceae, Pedaliaceae, Myoporaceae, and Verbenaceae. Being strongly zygomorphic in its corolla, Brandisia cannot be associated with either Loganiaceae or Solanaceae, in which the corolla is actinomorphic. The place of the genus is evidently with the other families which have distinct zygomorphy. In spite of the general resemblance of the plant in appearance to some genera of the Myoporaceae and the Verbenaceae, its many-ovuled ovaries and capsular fruits will not permit it to be included 1947 | LI, THE GENUS BRANDISIA 129 in either of these two. Being a shrubby genus and without staminodes, it is not referable to the Pedaliaceae. The families Scrophulariaceae, Bignoniaceae, Gesneriaceae, and Oroban- chaceae are closely related and it is sometimes difficult to draw sharp distinctions between them. The Gesneriaceae and Orobanchaceae are predominantly herbaceous families, with unilocular ovaries, the latter being also parasitic. The choice for placing Brandisia is therefore left to the Scrophulariaceae and the Bignoniaceae. e Bignoniaceae are nearly all woody plants, many of them climbers. The Scrophulariaceae are predominantly herbaceous. ‘The chief distinction between the two families is the presence of endosperm in the seeds in the Scrophulariaceae and the absence of the same in the Bignoniaceae. The fruit of the Scrophulariaceae is usually a capsule or sometimes a berry. The seeds are usually numerous and small. In the Bignoniaceae, the fruit is most often a two-valved capsule, usually very long and silique-like. The seeds are numerous, large, flattened, and membranaceous-winged. In the i. the Gibe Cheloneae is one of the few that contains woody genera. In this tribe, which is artificial and ill-defined, there have been included some anomalous genera including Paulownia, Wightia, and Brandisia. Paulownia has been included in this family instead of the Bignoniaceae because of the presence of endosperm in the seeds. Hallier (1. c.) has shown that endosperm is only scantily present in Paulownia and is absent in Wightia, and that the seeds of both are winged. He concluded that both genera, in which all species are trees, should be transferred to the Bignoniaceae. The ovoid loculicidal capsular fruit of Brandisia is very similar to those of the Scrophulariaceae, but the seeds, although very small, are winged and are without endosperm. ‘They are small, numerous, linear, with an elongated membranaceous winged testa which is reticulated. The fruit of Wightia is an oblong or ovoid capsule, but is septicidal, and the seeds are membranaceous and broadly winged all around. ‘The fruit of Paulownia is a woody capsule, loculicidally dehiscent, with numerous small winged seeds. There is an apparent relationship between the three genera mentioned. The transference of Paulownia and Wightia to the Bignoniaceae, although they still appear somewhat anomalous in that family, seems to be a better arrangement than to have retained them in the Scrophulariaceae. Brandisia, although seeming even more isolated if placed in the Bignoni- aceae, certainly should be associated with Wightia, whatever the latter’s disposition. Here we have a situation in which there must be additional knowledge in anatomy, pollen structure, cytology, etc., before a satisfactory disposition can be made of these genera as to their taxonomic position, as well as delimitation of the various families concerned. DISTRIBUTION To the eight species recognized in this study, there should be added 130 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII three more from Annam, described by Bonati, of which no material is available. The genus ranges from Assam (India) and southern Burma to western China, north to western Hupeh, and southern China, east to Kwangtung. It is also found in Annam, Indo-China. So far it has not been recorded from Formosa, Hainan, or Tonkin. The species are found mostly on mountains up to an altitude of 3000 meters, those in the south ascending usually to higher levels than those in the north. Most of the species are local in distribution and rare in occurrence. The most common and most widely distributed species is Brandisia Hancei, which is found in Yunnan, Szechuan, Hupeh, Kweichow, and Kwangsi. Brandisia rosea is found in northwestern Yunnan, Assam, and Bhutan. Brandisia discolor is found in southern Burma and southern Yunnan. Brandisia Swinglei occurs in Kwangtung and eastern Kwangsi, while B. kwangsiensis is known only from Kwangsi; the two species B. laetevirens and B. glabrescens are confined to southern Yunnan. Brandisia racemosa is found in Yunnan from the south to the northeast. Three species, B. Chevalieri, B. annamitica, and B. scandens, are known from Annam only. The genus is primarily of the subtropical regions of eastern continental Asia, particularly of the mountains at medium levels. The center of distribution is apparently in Yunnan, where six of the eleven species occur. Only two, B. Hancei and B. rosea, occur in the Sino-Himalayan region in northwestern Yunnan, and these are not endemic. More species, although mostly of rather limited distribution, are found in the warmer regions of the province in the south as well as in the east. TAXONOMY Brandisia Hooker f. & Thomson in Jour. Linn. Soc. Bot. 8:11, 1865, Shrubs, erect, scandent, or straggling, sometimes parasitic, often with stellate tomentum; leaves opposite or rarely subopposite, oblong-lanceolate times several together or short- or long- racemose, pedicellate, the pedicels bibracteolate; calyx campanulate, with or without distinct veins, subequally 5-dentate or ‘sometimes 4—7-dentate, rarely bilobed; corolla infundibular, bilabiate, usually densely tomentose, the tube usually incurved, the lips spreading, the upper lip larger and longer, concave, 2-lobed, the lobes road, obtuse, the lower lip shorter, 3-lobed, the lobes smaller, subequal or unequal, acute, attached at the same level or in the middle and much — stamens 4, didynamous, ascending, subexserted or included, attached the base of the corolla-tube; staminodes absent; filaments glabrous; te rounded, bilocular, the cells confluent above, long-pilose especially at the tip and the margins; ovary ovoid, hairy, bilocular, many- ane styles elongate, filiform, glabrous, the stigma simple, entire; ovules umer- ous, linear-oblong; fruits capsular, chartaceous, ovoid, acute, ieeulieidally dehiscent into 2 valves; seeds small, numerous, linear, with thin winged membranaceous elongated a, reticulated. Tyre species: B. discolor Hook. f. & Thoms. 1947 ] LI, THE GENUS BRANDISIA 131 About 11 species, in the subtropical and warmer parts of continental eastern Asia. With the addition of many other species to the original, it is necessary to redefine the generic concept of the genus and to form three subgenera, of which the last two contain each but a single species. In addition to the eight species recognized for China, India, and Burma, there are three species described by Bonati from Annam, of which no material is available to me. Two of these, B. Chevalieri Bonati and B. scandens Bonati, evidently belong to Eubrandisia. Brandisia annamitica Bonati, described as having the flowers in groups of 3-5 and having two staminodes, may require the erection of another subgeneric group. SUBGENERA OF BRANDISIA Suscenus I. Eubrandisia, subgen. nov. Frutex erectus vel scandens; foliis oppositis, basi acutis vel cordatis; floribus axillaribus, solitariis vel binis; corolla flava vel rosea, labio antico breviore, lobis subaequalibus: calyce campanulato, plerumque 5-dentato. SuscEeNnus II. Rhodobrandisia, subgen. nov. Frutex erectus; foliis oppositis vel suboppositis, basi longe attenuatis; nee axillaribus, solitariis vel binis; corolla rosea vel raro flava, labio antico fere breviore, lobis subaequalibus; calyce bilobato, lobis integris vel oe bifidis. Suscenus III. Cocecineabotrys, subgen. nov. Frutex scandens, parasiticus; foliis oppositis, basi rotundatis; floribus racemosis; corolla rubra, labio antico valde breviore, lobis inaequalibus; calyce campanulato, 5-dentato. Key TO THE CHINESE SPECIES A. Flowers axillary, single, rarely 2 together, scattered throughout the whole length the stems; calyx usually long-pilose within, sometimes tomentose; lower lip of corolla about as long as or slightly shorter than the upper lip, the lobes subequal and attached at about the same level; capsules densely tomentose to rarel B. Calyx more or less prominently 5-costate, 5- ee? usually long-pilose within ; leaves strictly opposite, the base acute to cordate (SuscEeNus I. Eubrandisia). alyx broadly campanulate, about as fie as broad, much enlarged above, the teeth divided for 4 to 4 the length. D. Leaves rien le to a petiolate, the petiole to 4 or : pe long ; leaf=bases ‘strongly cordate: a.2 ceca at oss csiee sais create . Hance. DD. Leaves audacity and comet onions bg petioles 5 mm. or more long; leaf-bases acute to rounded or subcordat Petioles short, pane mm. nee erie large, 10 mm. long, 8-9 mm. wide F. Leaves strongly discolored on the two surfaces, the upper dark brown to almost black in the dried state; leaf-bases rounded to subcordate. G. Leaves broadly ovate, to 4 cm. wide and 9 cm. long; indumentum grayish; calyx-teeth long-acute, 7 mm. long. 2 GG. Leaves narrowly ovate, to 3 cm. wide and 9 cm. eae indumentum brownish; calyx-teeth nee acute, to m LONG i035 cee are te eae PS OKs wee D4 ste ha 3. B. Preuo. 132 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XXVIII FF. Leaves subconcolorous on the two surfaces, the upper brownish in the dried state; leaf-bases acute to subrounded. .4. B. Rkwanegsiensis. EE. oe long, 1 cm. or more in length; calyx small, 6-7 mm. long and WE i nyia7s'8 s 494 65-44 va cen VR ESE bie nada inde 5. B. discolor. CC. Calyx cylindric: -campanulate, 13 to 2 times as long as wide, narrowed or very slightly expanded above, the teeth divided to about a quarter of the PONGEN GE The COR ss oso 6k odes bhi eK ee RSS RCE RRC 6. B. glabrescens. BB. Calyx without prominent nerves, 2-lobed, tomentose to glabrous within, the lobes entire slightly bifid; aeis opposite to subopposite, the bas ases long- attenuate (Suro ENuS II. Rhodobrandisia).. 0.0.0.0. ce ce ee rosea AA. fecs, racemose, many together on long or short branches or toward is end of the stems; calyx pubescent within; lower lip of corolla twice or more shorter than the upper lip, ie middle lobe smaller than and attached half as low as the lateral ones in a deep notch; capsules glabrous (SuBcENUs III. oe ). Sa Rie, 6 ORV nies 18 9raijev aes wis ee, FM GE GHTK 5-Fi4E Gide AoGoe aE GoD ee ale dsr . B. racemosa. 1. Brandisia Hancei Hook. f., Fl. Brit. Ind. 4: 257. 1884, in nota; Forbes & Hemsl. in Jour. Linn. Soc. Bot. 26: 179. 1890; Diels in Bot. Jahrb. 29: 565, 1900; Rehder in Sargent, Pl. Wils. 1: 573. 1913; Litipricht in Rep. Sp. Nov. Beih. 12: 480, 1922; Hu in Jour. Arnold Arb. 5: 233. 1924; Pai in Contr. Inst. Bot. Nat. Acad. Peiping 2: 186. 1934; Hand—Maz., Symb. Sin. 7: 831. 1936. Brandisia discolor sensu Hance in Jour. Bot. 18: 299. 1880, non Hook. f{. & Thoms. Huren: No _ precise Pete Henry 1009 (NYBG), 3007 (USNH); se Hupeh, Wilson 147 (NYBG, USNH), 147a (NYBG); Ichang, Wilson 2404 (A USNH); Liang Sung Sab W. Y. Chun 3840 (AA), 4185 (USNH); En-shih asta H.C. Chow 1885 (AA, USNH) ; Yangtze Gorges, N. C. Nelson s.n. (UC). Szec 7 Tchen-keou-tin, Farges s.n. (ANSP, B); Mount Omei, Chiao & Fan 362 (AA), Yu 289 (AA); Nan-chuan District, W. P. Fang 5675 (AA, ANSP, NYBG), 5676 ve ANSP, NYBG) ; Kwang-yun District, F. 7. Wang 22616 (AA); Wanhsien, Yen-ching- kou, Mrs. W. Granger 3 (UC). Yu ae No precise localities, HE. E. Maire 1344 UC), Forrest 9370 (AA), 9672 (AA), T. T. Yai 14812 (AA) ; Hee-chan-men, Delavay 1934 (ANSP, B, NYBG) ; Mo-so-yn, ned 4626 (ANSP, B); Kiao-kia, F. Ducloux 1237 (UC); Yoniive sen, F. Ducivan 159 (ANSP), 1027 Doria, UC) ; Siao-ou-long, E. E. Maire 183 (AA); Pe-yen-tsin, S. Ten 23 (AA), 34 (AA); Yungpe, S. Ten 403 (UC); Mengtze, A. Henry 9013 (AA, ANSP, MBG, NYBG, USNH) ; between Muang ae and Szemao, J. F. Rock 2821 (AA, UC, USNH) ; between Tengyueh and Likiang, . F. Rock 8115 (AA, UC, ape Kun-ming, C. Schneider 55 (AA), = (AA), H. hie Mazzetti 13053 (AA), W. Wang 62624 (AA), 62050 (AA), F. C. How 74244 (AA), 74247 (AA); ss 9a chen. near Tali, C. Schneider 3246 (AA); oe shuei District, H. T. Tsai 53027 (AA); Likiang, C. W. Wang 71414 (AA), K. M. Feng 2547 ; Hua- ning; Y. Tsiang & H. ‘an : > ~ Lis) Xr Se ~~ ~ e-s - lo a > > oa Do a S N we ~ > > ~~ Do hn a Co N ~~ > > ~~ trict, Steward, Chiao, & Cheo 948 (AA, NYBG, USNH); Ying- hive District, Steward, eee sce SOS ‘(USNH), Y. Tsiang 7795 (NYBG); Chengfeng, Y. Tsien 4456 (AA, BG, USNH); Tungtze, Y. Tsiang 4944 (NYBG), 5134 (NYBG, UC); An- ae Y. Tsiang 7416 (NYBG, UC), 9321 (NYBG); Chen-lin-shan, Kweiyang, S. W. Teng 90019 (AA). KWANGSsI: nett District, Steward & Cheo 49 (AA). This species was first identified as Brandisia discolor Hook f, & Thoms. by Hance. Hooker (1. c.), however, under that Burmese pe saan mentioned: ‘The Chinese plant referred to B. discolor by Hance, differs the broad calyx-lobes and corolla; it is B. Hancei, H. f.” Ha ance ’s ee is a Mesney collection from Kweiyang, Kweichow Province. Forbes and Hemsley cited Henry s, n. from Hupeh, Mesney s. n. from Kweichow, and also Maire s. n. and Bourne s. n. of unclarified locations. Rehder cited Henry 1150 and 3007 from western Hupeh, and Henry 9013 from Mengtze, 1947] LI, THE GENUS BRANDISIA 133 Yunnan. He said: “No good has of this species has been published so far, but according to Dr. A. B. Rendle of the British Museum who kindly compared Wilson’s oe 3404 with the type of B. Hancei Hook. f from Kweichou, there is no difference between the two specimens. Wilson’s specimen also agrees perfectly with Henry’s specimen from Hupeh referred by Hemsley to B. Hancei. Henry’s No. 9013 from Yunnan differs slightly in its smaller and narrower leaves.” Hu gave a detailed description for the species but cited no specimens. Brandisia Hancei is saan a common shrub at medium levels (500-2600 meters) i mountains of western China. It is readily distinguished from ae ee . by its smaller, very compactly arranged leaves, which are nearly sessile and strongly cordate at base. ‘The leaves are densely stellate-ferruginous-tomentose on the lower surfaces and are not strongly discolored on the two surfaces as in some other species. The flowers (March—November) are red or yellow on the inside 2. Brandisia Swinglei Merrill in Philip. Jour. Sci. Bot. 13: 157, 1918. Kwanctunc: Lofaushan, £. D. Merrill 10851 (tyPE coLty., NYBG, UC); Lung- men District, Nan Kwan Shan, W. 7. Tsang 25401 (AA), 25438 (AA). Kwancsi paar District, Tong Shan, W. T. Tsang 22862 (AA); Yao Shan, C. Wang 39502 (AA), 40192 ( A). eae shrubs, on mountain slopes at altitudes of about 500-1000 meters, in Kwangtung and adjacent parts of Kwangsi. Flowers pale yellow. meee from June to September. The type of the species was from: “Kwangtung Province, Loh Fau Mountain icc Merrill 10851, August 22, 1917, in thickets along small streams, altitude about 950 meters; very rare, a single plant observed.” JIsotypes in the herbaria of the New York Botanical Garden and the University of California have been seen. This species is characterized by its large broad leaves with grayish indumentum on the lower surface and the relatively longer and more acuminate calyx-teeth. It has the easternmost range of any species of the genus. 3. Brandisia laetevirens Rehder in Sargent, Pl. Wils. 1: 573, 1913. YUNNAN: Szem A. Henry 12605 (AA TYPE, MBG) ; rae shan District, H. T. Tsai 51495 (AA); Shih -ping District, H. T. Tsai 53434 (A A shrub in forests on mountains at altitudes of co to 1800 meters, in southern Yunnan. Flowers yellow. ne in January. The type of the species was from “Yunnan: Szemao, eastern mountains, alt. 1300 A, Henry (No, 12605).” The aida in cite herbarium of the Arnold Arboretum e been seen This species is a close relative of B. Hancei and B, discolor. It is dis- tinguished from the former by its large, less cordate, and more distinctly petiolate leaves. It differs from the latter notably in the larger calyx and shorter petioles. The corolla of B. discolor is also narrower and more densely tomentose and the leaves are darker on the upper surfaces, which become almost black in dried specimens. 4. Brandisia kwangsiensis sp. nov. Frutex scandens, ramulis hornotinis fulvo-tomentosis; foliis chartaceis, 134. JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII ovato-lanceolatis vel sy fon a acuminatis, basi acutis vel subrotun- datis, 6-8 cm. longis, 2—3.7 cm. latis, margine integris, supra initio sparse stellato-tomentosis, mox ee et glabris, in sicco luteo-viridibus, subtus dense flavescentibus stellato-tomentosis, reticulatis, nervis lateralibus utrinsecus 5—7, adscendentibus conjunctis, supra costa impressa_ subtus elevata; petiolis 4-8 mm. longis, fulvo-tomentosis; floribus axillaribus, solitariis, pedicellis gracilibus, fulvo- tomentosis, 8-9 mm. longis, medio triangularibus, acutis, 2-3 mm. longis; corolla late campanulata, leviter curvata, bilabiata, circiter 2.4 cm. longa, rubra, extus dense stellato-tomen- tosa, limbo intus tomentoso, tubo intus glabro, labio postico magno truncato emarginato, lobo antico 3-lobato, fere dimidio breviore, lobis ovatis acutis; staminibus stylisque corolla subaequilongis; antheris dense ad capsulis ovoideis, dense tomentosis, calyce incluso circiter 1 cm. lon Kwancsi: Ling-wun District, in ae woods, S. K. Lau pai ae AA); Pin Lam, S. P. Ko 55685 This species is siionuly characterized by its relatively narrow leaves with acute to subrounded bases. From other related species, such as B. Swinglet, B. laetevirens, and B. discolor, this new species is also readily distinguished by the nearly concolorous leaves. The upper surfaces of the leaves are only slightly more brownish than the lower, while in the other species mentioned the upper surfaces are much darker and often turn blackish in the dried state. The reddish flowers appear in January. Besides the two collections mentioned above, another specimen, R. C. Ching 6300 (NYBG), collected from Nee Bai, Kwangsi, may be referable to this: species. On it the leaf-bases are more rounded. The specimen bears only incomplete ae 5. Brandisia haa Hook. f. & Thoms. in Jour. Linn. Soc. Bot. 8:11. t. 4. 1865; As. Soc. | 42: 236. 1873, For. Fl. Brit. Ind. 2: 250. 1877; Hook. , Fl. Brit. “Ind 4: 257. 1865; Brandis, Ind. Trees 491. 1906; Rehder in Sargent, a Wils. 1: 573. 1913. YUNNAN: Szemao, A. Henry 12605B (AA, ANSP, MBG, NYBG), J. F. Rock 2769 (AA, USNH), C. W. Wang 80987 (AA, USNH) A shrub on ridges in bushes at an altitude of about 1300 meters, from lower ie to southern Yunnan. Flowers yellow. Flowering in March. The type of the speci s a collection made by Brandis in aaa Burma, ae no collection is “ui cited with the original descriptio This species was originally found in lower Burma a ‘the southern Yunnan plants are apparently referable to it. The species is characterized by the slender distinct petioles, strongly darkened upper leaf-surfaces, and the relatively small calyces and narrow corollas 6. seeui cpaininigy Rehder in Sargent, Pl. Wils. 1: 574. 1913. A. Henry 9716 ae USNH), 9716A (AA Type); Ping-pien moti H. ‘T. Tsai poe (AA), 62798 (A A slender shrub, in forests at ade of about 2000-2300 meters, in southern Yunnan. Flowers yellow. Flowering in July and August. Rehder cited two collections: “Yunnan: Mengtze, forest, alt. 2000 m., A. Henry 1947] LI, THE GENUS BRANDISIA 135 (No. 91764); south of Red River from Manmei, alt. 2300 m., A. Henry (No. 9716).” Both are in the herbarium of the Arnold Arboretum, the first being designated by Rehder as the type. The number, however, should read 97 This species is characterized by the long calyx. Tate this distinct calyx- ag it can ee be readily distinguished from the related B. discolor, B. Hancet, , by the — pubescent leaves, which are eed to broadly aie at the 7 Brandisia rosea W. W. Smith in Notes Bot. Gard. Edinb. 10:10. 1918. SIKANG: Tsarong, Salween-Chingkiang Divide, G. Forrest 20320 (AA. USNB). Yunnan: Western flank of the Tali Range, G. Forrest 11565 (AA, RBGE type, UC) ; Shang-pa District, H. T. Tsai 54774 (AA), 54859 (AA), 56644 (AA); Che-tse-lo, H. T. Tsai 54172 (AA), 58548 (AA); Chungtien, Wu-tso on the Yangtze bank, K. M. Feng 3344 (AA). In woods or open thickets on mountain slopes at altitudes of 2100 to 3000 meters, in northwestern Yunnan and adjacent Sikang and also in eastern India. Flowers deep rose to reddish orange. Flowering from July to September. The type of the species is from western Yunnarr: “China, western flank of the Tali Range, Yunnan. Lat. 25° 40’ N. Alt. 10,000 ft. Shrub of 3-5 ft. Flowers deep rose. Open dry situations amongst scrub. July 1913. G. Forrest. No. The holotype in the Edinburgh Herbarium has been seen. Brandisia rosea is distinctly characterized by the bilobed calyx. In this respect, it deviates from the original generic limits. It also differs from other species in its general appearance, especially in the relatively narrow long-attenuate leaves which are sometimes oppositely arranged. In view of these differences, it is here segregated as representing a different subgenus. ‘The fo and seed characters are the same as in the other species. A variety with yellow flowers was described by Fischer, from Assam, India, as Brandisia rosea W. W. Smith var. flava C. E. C. Fischer, in Kew Bull. 1934: 93. 1934. In the Genera Plantarum, it is stated that, besides the original B. discolor, there is another species in India. Hooker, in the Flora of British India, said: ““The supposed second species of Brandisia, alluded to in the ‘Genera Plantarum’ as a native of Bhotan, has stellate romentun, lanceolate leaves and a 2-lipped calyx, and cannot be con- generic.” Fischer identified this plant as belonging to B. rosea. He said: “Griffith’s specimen [Kew Dist. 3748] from Bhutan, referred to in the Gen. Pl. and in the FI. Brit. Ind., and a sheet collected by Nuttall also in Bhutan (both in the Kew Herb.) are this species, but being in fruit only it cannot be determined whether they represent the typical species or the colour variety.” 8. Brander racemosa Hemsl. in Kew Bull. 1895: 114. 1895; Oliver in Hook. Ic. PI. 24: ¢. 2383, 1895; Schneider, Ill. Handb. Laubholzk. 2: 618. fig. 399, c-h. 1911; Rehder in Sargent, Pl. Wils. 1: 574. 1913; Hand.—Mazz., Suni . Sin. 7: 831. 1936. YuNNAN: Mengtze, Hancock s. n. aotae Henry 9973 (AA, MBG, NYBG, USNH), Leduc s. n. (ANSP); Siao-ho, E. E. ed 3986 (ANSP, B); Mongkou, E. E. Maire 10199 (NYBG, UC); Kiang- hoe E. E. Maire s. n. (ANSP); Kiao Kia, F. Ducloux 989 (NYBG, UC); near Kun-ming, i: Handel-Mazzetti 5051 (AA, USNH). 136 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII A spreading shrub, at altitudes of 800 to 2800 meters, in Yunnan. Flowers scarlet. Flowering in September. Hemsley’s type was from “Western China, in shady copses, Mongtze, Yunnan, Hancock, 143,” It has not been seen. The description and original illustration by Oliver are distinctive. Two specimens collected by Hancock in the type locality in 1896 are in the herbarium of the University of California. Brandisia racemosa is a most unique species of the genus, distinguished not only by the peculiar corolla-structure but also by the inflorescence. The flowers are arranged in short or long racemes. The calyx is shortly and inconspicuously 5-dentate. The corolla is strongly bilobed, the upper lip being much longer than the lower. The lower lip is 3-lobed, with the two lateral lobes adnate to the sides of the upper lip midway in the corolla and the middle lobe attached much lower in a deep notch between the two lateral lobes. The plant also differs from other species of the genus in general appear- ance, with its long straggling branches and relatively smaller and more distinctly serrate leaves. The whole plant is blackened in the dried state, indicating its possible parasitic or semi-parasitic habit. Rehder mentioned that: “It was introduced by Mr. Wilson into cultivation, but it could not be.grown successfully, as it is apparently parasitic and its proper host plant is not known,” Although the plant differs radically from other species of the genus in many respects, its basic characters in the stamens, fruits, and seeds agree with the genus in general. It is thought best to retain it in the genus but to designate a subgenus for its sole accommodation. DOUBTFUL AND EXCLUDED SPECIES Brandisia praticola W. W. Smith in Notes Bot. Gard. Edinb. 10: 10. 1918. This species is based on Forrest 13350 from northwestern Yunnan. The type, in the Edinburgh Herbarium, is available for study. It proves to be Pterygiella nigrescens Oliver. Brandisia Souliei Bonati in Bull. Soc. Bot. France 56: 467. 1909. This species is based on Soulié 5199 from Yargong, Sikang Province. It has not been seen. According to the description, this species is dis- tinguished by the 3-verticillate leaves, which are concolorous on the two surfaces, the flowers’ in axillary cymes, the membranaceous calyx with its 5 lobes subulate and splitting at the tips, etc. It does not represent the genus Brandisia but belongs in the Labiatae. A short discussion of this misplaced species will be issued in this Journal in the near future by Prof. E. D. Merrill. Soocuow UNIversITy, SoocHow, CuHINa. 1947] SAX & SAX, GENERIC HYBRIDS OF SORBUS 137 THE CYTOGENETICS OF GENERIC HYBRIDS OF SORBUS Hatiy J. Sax AND Kari Sax With one plate THE POMOIDEAE constitute a group of rather closely related genera of common origin (Sax, 1933). All genera have a basic chromosome number of 17 and generic hybrids are relatively common. Natural hybridization is most frequent between Sorbus and the related genera Aronia, Amelan- chier, and Pyrus. Rehder (1940, 1947) has described or enumerated seven hybrids between Sorbus and Aronia species, three hybrids of Sorbus and Amelanchier, and one hybrid between Sorbus and Pyrus. These genera are closely related and have been included in a single genus Pyrus by some taxonomists. e nomenclature in the following descriptions is that of Rehder (1940). The cytological analysis is based upon acetocarmine smear preparations, following fixation in alcohol-acetic acid. Hybrids between Sorbus and Aronia include Sorbaronia Jackii, S. Arsenit, S. hybrida, S. fallax, S. heterophylla, S. alpina, and S. Dippelii. Both the pinnate- and entire-leaved species of Sorbus have produced hybrids with Aronia species, all of which have entire leaves. Sorbaronia Dippelit has been crossed with Sorbus alnifolia in the Arnold Arboretum. Three of the Sorbaronia hybrids, S. Jackii, S. alpina, and S. Dippelii, have been examined cytologically and attempts have been made to grow Fy. generations. Sorbaronia Jackii is a naturally recurring hybrid between Sorbus ameri- cana and Aronia prunifolia. Specimens of the Sorbus parent in the Arboretum, both mature trees and 2-year old seedlings, have from 5 to 8 pairs of leaflets with an average number of slightly less than 7. The Fy hybrid has variable leaves, ranging from entire, as in Aronia, to pinnate leaves with as many as 3 pairs of leaflets in addition to the terminal leaflet. Often the leaves are partially pinnate or lobed. The average number of leaflets is 1.4. The open pollinated progeny of S. Jackii are extremely variable in size and viability. Of 50 seedlings selected from the survivors grown in the greenhouse, seven have died in the nursery at the end of the second year, seven are extreme dwarfs, and 36 have made more or less normal growth. All surviving seedlings have at least partially pinnate leaves, ranging from 3 to 7 leaflets with an average of 4.5. No segregates had any entire leaves, but these progeny may be back cross hybrids with adjacent pinnate-leaved species of Sorbus. Seedlings of Sorbus americana grown in the nursery showed much greater uniformity than those of the hybrid. Of 50 seed- 138 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII lings, only one had died at the end of the second year. Sorbaronia Jackii produces abundant fruits, but much of the seed is defective or produces weak seedlings — pollen sterility is variable, ranging from O in some anthers to about 50 per cent in others. The meiotic divisions were generally irregular, but in some pollen mother cells the divisions appeared to be relatively normal. At first meiotic metaphase there were from 13 to 15 bivalents and from 4 to 8 univalents in most of the figures examined, although the number of univalents ranged from 0 to 12. Lagging univalents were generally found at both first and second anaphase (PL. I, Fic. 1). The tetrads resulting from the two meiotic divisions were often irregular with some micronuclei, but as a rule nearly 50 per cent of the pollen appeared to be morphologically normal. Sorbaronia alpina is a cross between Sorbus Aria, a species with simple leaves, and Aronia arbutifolia. This hybrid, as represented by the variety super-aria, is more sterile than Sorbaronia Jackii. Occasionally there appeared to be 17 bivalent chromosomes at the first meiotic metaphase, but the divisions were generally quite irregular, with as many as 14 lagging univalents at anaphase. For some anaphase figures the lagging chromo- somes appeared to include an occasional bivalent (PL. I, ric. 2). The second meiotic divisions also were irregular and the pollen fertility ranged from 0 to 24 per cent. For both of these Sorbaronia hybrids the pollen sterility varied greatly in different flowers and some anthers contained no good pollen. Some fruits are produced by S. alpina, but we have not been able to germinate the seeds. Sorbaronia Dip peli is a hybrid between Sorbus Aria and Aronia melano- carpa. This hybrid is relatively fertile, as indicated by both pollen sterility and by fruit development. Yet some anthers produce no good pollen and we have not been able to germinate open pollinated seed. Seventeen bivalent chromosomes were found at meiotic metaphase (PL. I, FIG. 3), and there was little irregularity at meiotic anaphase. Pollen development was variable, but some counts showed as high as 58 per cent normal pollen. The pollen fertility of the Aronia parent was only 76 per cent and that of the Sorbus parent about the same. We have succeeded in crossing Sorbaronia Dippelii with the distinct Sorbus species S. alnifolia. This hybrid has the characteristic foliage of S. alnifolia, the male parent. It is a slow growing shrub and has not yet flowered. Several hybrids between Sorbus and Amelanchier have been described. The three known hybrids are growing in the Arnold Arboretum, but only Amelasorbus Jackii is large enough to provide meiotic material. This hybrid was found among seedlings grown from seeds collected in the Moscow mountains of Idaho by Professor J. G. Jack. It is described by Rehder as a hybrid between Sorbus scopulina and Amelanchier florida. According to Professor Jack there were two types of the hybrid, one with partially pinnate leaves and one with only an occasional cleft or lobed leaf. Only the latter type is now available in our collections. 1947] SAX & SAX, GENERIC HYBRIDS OF SORBUS 139 Amelasorbus has been difficult to analyse cytologically. The microspore mother cells appear to disintegrate at any stage of development from prophase to microspores. In favorable prophase preparations 32-34 chromosomes were counted at diakinesis, indicating a very loose association of paired chromosomes. At meiotic metaphase there were from 12 to 15 bivalents and from 4 to 10 univalents (Pu. I, ric. 4). At anaphase the number of lagging univalents are variable, but as many as ten have been observed (Pu. I, Fic. 5). Although the meiotic behavior of the chromo- somes of Amelasorbus is similar to that of two of the Sorbaronia hybrids, the sterility is much greater. Practically no normal pollen is formed and very little fruit is produced. Over a period of years only two viable seeds have been obtained and the resulting seedlings, which have now flowered for several years, are very similar to the parental hybrid in both appear- ance and sterility. Apparently they are of parthenogenetic origin. Two other forms of Amelasorbus are listed by Rehder (1947). Amela- sorbus Hoseri is thought to be a hybrid between Sorbus hybrida and Amelanchier sp. It has partially pinnate leaves. Amelasorbus Racibor- skiana is listed as a cross between Sorbus sp. and Amelanchier asiatica. It too has partially pinnate leaves. Both of these hybrids were raised at the Kornik Botanical Garden in Poland and named by Wroblewski. Our specimens of these hybrids are not sufficiently mature for meiotic analysis. The known crosses between Sorbus and Pyrus are limited to Sorbopyrus auricularis, a natural hybrid between S. Aria and P. communis. We have in the Arnold Arboretum only the variety bulbiformis, a form which re- sembles Pyrus in most characters. This hybrid is a triploid with pre- sumably two sets of pear chromosomes. Its cytological behavior has been described earlier (Sax 1932). We have obtained some progeny from open pollinated seeds of Sorbopyrus. The seedlings are variable but tend to segregate into two groups—some which are vigorous and resemble the mother tree, and a series of dwarf segregates which survived for one to three years. In all of these generic hybrids involving Sorbus there is considerable sterility, and when segregates are obtained many are weak and do not survive. The genera Sorbus, Aronia, Amelanchier, and Pyrus can be crossed only in certain combinations. One might suppose that Aronia and Amelanchier could be crossed, since both have produced at least several hybrids with Sorbus, but repeated attempts to cross these genera have failed. However, we also have been unable to produce any artificial hy- brids between Aronia and Sorbus, and yet seven such spontaneous hybrids are known and some are recurrent in nature The fact that Sorbus, Aronia, Amelanchier, and Pyrus can be crossed in certain combination indicates a certain degree of similarity. The behavior of the hybrids does indicate cytogenetic differentiation usually found in diverse species. Whether such incompatibility merits more than specific differentiation is a matter of taxonomic convenience. 140 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVUI SUMMARY Natural hybrids between Sorbus and Aronia and between Amelanchier and Pyrus are partially sterile and show considerable meiotic irregularity. Sorbaronia produces a variable progeny both in morphological characters and viability. Amelasorbus produces few viable seeds and the progeny appear to be of parthenogenetic origin. Sorbopyrus sets few seeds and the progeny include many dwarf segregates which survive for only a few years. The cytogenetic analysis indicates a common origin of these genera, but a degree of differentiation sufficient to maintain generic isolation. LITERATURE CITED Renper, A. (1940) Manual of Cultivated Trees and Shrubs. MacMillan, New York. REHDER, - (1947) eee of Cultivated Trees and Shrubs. In press, Harvard Univ. Press, Cambridge Sax, K. Gace Phecwaisens relationships in the Pomoideae. Jour, Arnold Arb. 13: 363-367. Sax, K. (1933) The origin of the Pomoideae. Proc. Amer. Soc. Hort. Sci. 30: 147-150. EXPLANATION OF THE PLATE Drawn from acetocarmine preparations. Magnification approximately x 2000. Fic. 1. Sorbaronia Jackii. First meiotic anaphase. Fic. 2. Sorbaronia alpina super-aria. First meiotic anaphase. Fic. 3. Sorbaronia Dippelii. First meiotic metaphase. Fic. 4. Amelasorbus Jackii. First meiotic metaphase. Fic. 5. Amela- sorbus Jackii. First meiotic anaphase. ARNOLD ARBORETUM, Harvarp UNIvERSITY. Jour. Arnotp Ars. VoL. XXVIII Puare | GeENneERIc Hysrips oF SORBUS 1947] BUCHHOLZ & GRAY, FIJIAN ACMOPYLE 141 A FIJIAN ACMOPYLE JouHn T. BucHHoiz AND NettTa E. Gray With one plate Amonc the collections of John W. Gillespie from the Fiji Islands is a single specimen (no. 3273, from Namosi Province, Viti Levu) which rep- resents a species of Acmopyle. This specimen had been included in a loan of material of Podocarpus from the herbarium of the Arnold Arboretum and had been inadvertently identified as Podocarpus vitiensis Seem. It differs in many important features, both externally and internally, from P. vitien- sis, and it agrees so generally in the anatomy of the leaves with Acmopyle Pilger that there can be no doubt about the generic identity. Specifically it differs somewhat from Acmopyle Pancheri (Brongn. & Gris) Pilger in several details, such as the external shape of the leaves and their texture, width, and apex. This is not the first report of Acmopyle in the Fiji Islands. Sahni (in Phil. Trans. Roy. Soc. London B, 210: 253-310. 1920), in his treatise on the structure and affinities of Acmopyle, mentions an earlier collection by J. Horne in 1877-78, found in the herbarium of the Royal Botanic Gardens, Kew, which he examined in considerable detail. Sahni concluded that this specimen represents a species of Acmopyle. He also records the opinion of Dr. Stapf, who examined this material and who “does not con- sider it likely that the Fiji specimen is specifically identical with Acmopyle Pancheri, but he agrees to the generic identity.’ We have examined the leaves of the Gillespie specimen histologically, and also those of Acmopyle Pancheri, numerous specimens of Dacrydium falciforme (Parl.) Pilger, and some of Dacrydium taxoides Brongn. and Gris. Details given by Sahni concerning the flattening of the leaf in these species in the vertical plane, with the vascular bundle remaining in an un- changed orientation, can be substantiated in all essentials. The vascular bundles therefore appear as if turned sidewise through 90° with reference to the leaf surface. The above mentioned two species of Dacrydium, representing a special section of that genus, and the two species of Acmopyle share in this unusual type of leaf. Leaves of the Dacrydium spp. possess hypoderm and are bifacial, having palisade tissue and stomates on both sides, while the leaves of Acmopyle found on the pectinate twigs of the last order of branching have very little or no hypoderm, palisade tissue only on the upper side which is exposed to light, and the stomates confined almost entirely to the lower surface of the leaf. The unique type of leaf anatomy therefore serves to segregate Acmopyle from other genera of the Podocarpaceae. 142 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Acmopyle Sahniana sp. nov.) Arbor parva; foliis spiraliter nen in ramulis ultimis biseriatim in una planitie expansis, basi in ramulum longe decurrentibus; laminis falcatis, 16-19 mm. longis, 3.5-4 mm. latis, supra nitido- viridibus, marginibus foliorum subparallelibus sed in apicem oblique acutum terminantibus, costa supra non conspicua; floribus masculis seminibusque ignotis. Leaves with two bands of stomates in 14—18 irregular rows each beneath, interrupted only slightly at midvein and leaving wide non- stomatiferous bands at borders; single vascular bundle within leaf turned through 90°, the xylem toward the adaxial margin, the phloem and single resin canal toward the abaxial margin of the leaf FIJI ISLANDS: Vitt Levu: Namosi: Mt. Vakarongasiu, alt. 900 m., on an exposed ridge, John W. Gillespie 3273 (Typr in herb. Arnold Arb.), Oct. 2, 1927. Acmopyle Sahniana may be distinguished from A. Pancheri (Brongn. & Gris) Pilger by the wider leaves, the more nearly parallel leaf-margins, and the obliquely acute apex. The leaves of A. Pancheri are usually 2.5—3.5 mm. wide, are gradually narrowed toward the tip, have a narrow rounded apex, and are thicker. Both species have the vascular bundle turned through 90° and are without hypoderm or have only scattered fibers of hypodermal tissue near the midvein of the leaves. he midvein is not centrally placed in A. Sahniana. The wider margin is in the adaxial posi- tion and this convex adaxial edge near the base of leaf is much thinner than in A. Pancheri. OTHER SPECIMENS EXAMINED FOR COMPARISON2 Acmopyle Pancheri (Brongn. & Gris) Pilger ALEDONIA: Mt. Mou, J. Franc 170 (A, US), Le Rat 2594 (A), Louisa Clark Williams 149 (Bish) Dacrydium taxoides Brongn. & Gris NEW CALEDONIA: Cougui, Pancher 379 (NY). Dacrydium falciforme (Parl.) Pilger vel aff. Group A. Specimens resembling the original descriptions: MALAY PENINSULA: R. E. Holttum (Singapore field no. 20626) (UC); Pawanc: Fraser Hill, 4000 ft., Aug. 25, 1923, M. Nur 10507 (A). BORNEO: British NortH Borneo: Mt. Kina- balu, Marai Parai, M. S. Clemens 10962 with ee (UC), J. & M.S. Clemens 33078 (A, UC); Mt. Pg nitie Penibukan, J. & M.S. Clemens s. n. (A); SARAWAK: t Poi, J. & M.S. Clemens 20263 (A); Sania BornEo: Res. Wester-Afdeeling, Bengkajang ai Neth. Ind. For. Service 24778 (A); (locality?), Batu Gajah Lugga, B. W. Hullett 5695 (A). CELEBES: Res. Manado, Paleo, 1400 m., Neth, Ind. For. Service 17544 (A). PHILIPPINES: Mowpons: Mt. Halcon, £. D. med in honor of Professor Birbal Sahni, F. R. S., of Lucknow, India, in recognition of his comprehensive study of the morphology of this genus. Prof. Sahni, and was educated at Punjab and at Cambridge (Eng.) Universities. He is now Professor of Botany and Dean of the Faculty of Science at the University of Lucknow. 2 Place of deposit of specimens is shown by the parenthaticn) letters, as follows: (A) Arnold Arboretum; (Bish) Bernice P. Bishop Museum; (NY) New York Botani- cal Garden; (UC) University of California ; (US) United States National Herbarium. Jour. Arnotp Ars. VoL. XXVIII PLATE I A Fry1an AcCMOPYLE 1947] BUCHHOLZ & GRAY, FIJIAN ACMOPYLE 143 Merrill 5744 (NY); Luzon: Mt. Umingan, M. Ramos & G. Edano 26394 (A, NY); G. Edano, Philip. Nat. Herb. no. 4508 (A). Group B. ra with leaves smaller than those of Group A: NEW CALE- DONIA: Mt. Mou, C. J. Pitard 2091 (A, NY). British New Guinea: C. E. Lane-Poole 397 Gua leaves) (A). EXPLANATION OF THE PLATE A. Acmopyle Sahniana, from Viti Levu, Fiji Islands, Gillespie 3273 (A). B. Acmopyle Pancheri, a twig from Williams 149 (Bish), for comparison. UNIVERSITY’ OF ILLINOIS an HARVARD UNIVERSITY. JOURNAL OF THE ARNOLD ARBORETUM VoL. XXVIII APRIL 1947 NuMBER 2 A MONOGRAPH OF THE GENUS STREBLOSA KORTHALS (RUBIACEAE) C. E. B. BREMEKAMP With two text-figures INTRODUCTION KORTHALS DESCRIBED in his “Overzicht der Rubiaceén van de Neder- landsch-Oostindische Kolonien” (in Ned. Kruidk. Arch. 2 (2): 245. 1851) i.a. a new genus Stredblosa, to which he referred Blume’s Psychotria tortilis and two species based on specimens which he had collected himself in Sumatra and Borneo. Although he gave no explanation of the name, we may safely assume that it refers to the peculiar arrangement of the flowers along the branchlets of the inflorescence which they encircle in the same way in which the shoot of a winding plant twines round its support. Blume’s epithet ¢ortilis too was apparently inspired by the arrangement of the flowers. Korthals’ generic description contains several inaccuracies; the calyx is not 5-dentate but 5-fid or, more rarely, 5-partite, the corolla-tube is not infundibuliform but cylindrical or, sometimes, in the upper half narrowly campanulate, and it is not glabrous inside but provided with a ring of hairs or with five hair-bundles, the aestivation of the lobes is not subvalvate but quite simply valvate, the ovules are neither peltate nor attached to the centre of the septum but ascending from the inner angle of the ovary- cells, the pyrenes are not plane on the inside but provided with two quite conspicuous contiguous excavations, and the embryo is not amphitropous but straight; the plants moreover are not fruticose but herbaceous, and their inflorescences are not axillary but at first terminal and subsequently pseudo-axillary. The generic description is followed by a note in which he gives an exposition of the reasons which induced him to separate this genus from Psychotria. It is reproduced here in English translation: ‘‘On account of 146 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII their habit it seemed to me that the generic identity of the species here referred to Streblosa with species belonging either to Psychotria or to Grumilea, already looked dubious, for they differ from them in their membranous, reticulately veined leaves and in the peculiar corymbose inflorescence. The analysis of the flower revealed, in addition to a number of identical features, in the structure of the ovary a diagnostic character of sufficient importance: whereas in Psychotria the ovules arise from the bottom of the cells, they are in Streblosa attached to the centre of the septum and peltate.” The argument derived from the position of the ovules, by which Korthals apparently set great store, and which he thought would justify the creation of the new genus before the tribunal of his fellow botanists, unfortunately was based on inaccurate observation: in reality the ovules of Streblosa arise just like those of the other Psychotrieae from the inner angle of the ovary-cells. Miquel already had some doubts with regard to Korthals’ description of the position of the ovules. In his “Flora Indiae Batavae” (2: 294. 1857) he reduced the genus to a subdivision of Psychotria for which he retained the name Streblosa and which he characterized with the aid of the data provided by Korthals. An interrogation mark, however, is put behind the words with which the attachment of the ovules to the middle of the septum is described. Afterwards Miquel’s valuation of the distinguishing characters underwent a change, for in his ‘““Ecloge Rubiacearum Archipelagi Indici” (in Ann. Mus. Bot. Lugd.-Bat. 4: 211. 1869) the genus is reéstablished. As no generic description is given, and as no comment is made on this change of attitude, it is impossible to say how the latter was brought about. It is noteworthy that Miquel corrected in this work Korthals’ assertion with regard to the habit of Str. tortilis (Bl.) Khs: he himself describes it as an herbaceous plant. Of the two new species described by Korthals only one, the Sumatran Str. polyantha, was found by Miquel in the Leiden herbarium (l.c. 262); it seems that the specimen on which the other species, the Bornean Str. undulata, was based, had disappeared already, and the type of Str. tortilis var. 8, collected by Korthals on G. Singalang in West Sumatra, which is said to differ from the form described by Blume in its sessile inflorescences and elliptic leaves, was not met with either; in contradistinction with Str. undulata (v. infra) it has never again been mentioned in literature. It may have have been conspecific with one of the Sumatran species dealt with below, but nothing definite can be said with regard to its identity, Hooker f. reduced Streblosa in Benth. et Hook. f., Gen. Pl. 2:,124, 125. 1873, to Psychotria: as this author attached great importance to the position of the ovules, it is unlikely that he would have made this reduction if he had not convinced himself of the inaccuracy of Korthals’ description. The type-species was referred by Hooker to the section Mapouria (a ebracteatae), where it occupies on account of the persistent bifid stipules, the cincinnate arrangement of the flowers, the small green fruits and the two contiguous excavations on the ventral side of the pyrenes, a most 1947 | BREMEKAMP, THE GENUS STREBLOSA 147 anomalous position, Korthals’ other species are not mentioned, but a plant collected in Penang and in the Wallich herbarium, inserted under the provisional name Psychotria microcarpa Wall., is declared conspecific with Ps. tortilis Bl., an error which was repeated by ets and Gamble in their ‘Materials for a Flora of the Malay Peninsula” (in Jour. As. Soc. Beng. 72 (2): 11. 1906) and by Ridley in his short survey of the genus Streblosa in Jour. As. Soc. Straits 57: 62. 1911. It was corrected by the latter in his ‘“‘Flora of the Malay Peninsula” (2: 148. 1923). Baillon referred Streblosa in Adansonia 12: 325. 1879 and in his “Histoire des Plantes” (7: 285. 1880) to Uragoga L., which in his delimitation is an even more unwieldy genus than Psychotria L. in that of Hooker, and it is therefore not necessary to discuss this reduction. It is noteworthy, however, that he corrected Korthals’ assertion with regard to the aestivation of the corolla-lobes. Writing of Str. tortilis he stated in Adansonia: “Sur les échantillons que nous avons pu examiner, toutes les corolles étaient simplement valvaires.”’ Schumann made no mention of the genus Stredlosa in his monograph of the family in Nat. Pflanzenfam. 4: 4. 1891; and in the “Nachtrage”’ to this work too I have vainly looked for any reference to it. In Dalla Torre & Harms, Genera Siphonogamarum, 503, 1905, however, it is included among the synonyms of Psychotria. Stapf was, after Miquel, the first to recognize the generic distinctness of Korthals’ genus. In his paper on the flora of Mt. Kinabalu (in Trans. Linn. Soc., Bot., II. 4: 182. 1894) he described a new species, Str. urticina, and expressed himself with regard to its generic position in the following terms: “I think Streblosa should be kept up as a distinct genus on account of its peculiar habit, which is produced chiefly by the herbaceous growth, by the very thin leaves, the axillary inflorescences (which in the present species resemble very much those of a common nettle), and by the minute flowers and fruits. Miquel says of Streblosa ‘frutices’; but the specimens of S. tortilis (Bl.) Korth., the only species seen by him, are decidedly herbs. The genus Stredlosa is limited to Malaya.” The statement that the inflorescences are axillary, is incorrect: in reality they are terminal but soon pushed aside by an axillary branch developing in the axil of one of the two highest leaves. The reference to Miquel is incomplete: in the “Flora Indiae Batavae” the latter merely repeated Korthals’ statement, but in his paper in the “Annales” he described S¢ér. tortilis correctly as “herba basi radicans.”’ The next author who recognized Streblosa as a distinct genus, was Valeton. In Bot. Jahrb. 44: 568. 1910, he described a new species, Str. glabra, based on a specimen collected by Winkler (Breslau) in South-east Borneo; it appears very close to Str. urticina Stapf. At the same time he gave a more detailed description of Str. undulata Khs., based on some other specimens collected by Winkler in the same region as Str. glabra and as Korthals’ type. Whether these specimens really are conspecific with the latter, is difficult to sav, because of Korthals’ incomplete description, but 148 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII it should be noted that the leaves are elliptic according to Korthals, whereas Valeton states that they are 900-120-150 mm. x 35-50 mm., which means that they are lanceolate in his specimens. However, as there is little chance that Korthals’ type will ever be recovered, and as there is otherwise nothing in his description which positively excludes the possibility that his type was conspecific with the specimens described by Valeton, the latter’s interpretation may perhaps be accepted. At about the same time Ridley (in Jour. As. Soc. Straits 57: 62. 1911) devoted a few pages to Streblosa. He agreed with Stapf that the genus should be restored, but no new arguments were brought forward to support this view. He recognized seven species, i.e. apart from the three for which the genus was created and the one described by Stapf, three new ones. Of these three the first, Str. hirta, is probably identical with Str. polyantha Khs., of which Ridley had seen no material. The only difference which I can find in comparing Korthals’ specimens with Ridley’s description, lies in the slightly narrower leaves (10-15 cm. & 3.7-5.0 cm. instead of 9-12 cm. X 4.5-5.7 cm.). In the description of Str. pubescens Ridl. there are a few points which make the position of this species altogether dubious: the stipules are said to be lanceolate-acuminate, whereas in this genus they are always two-lobed; the stamens should be inserted near the base of the tube instead of in the ring of hairs in the middle or somewhat above the middle, the stamens as well as the style should be included, whereas the flowers of Streblosa are always heterostylous with either the stamens or the style exserted, and the disc is said to be composed of 5 renjform bodies, whereas I found it in Streblosa always semi-globose or annular with a single, more or less distinct, transverse impression on the top. For the time being, therefore, I consider Str. pubescens a ‘species incertae sedis.” The third species, Str. bracteata, on the other hand, is a true representative of the genus, probably nearly related to the plant afterwards described by Merrill under the name Str. myriocarpa. Like the latter it is found in West Borneo. Ridley remarked with regard to Str. urticina Stapf that he saw in the herbarium of the Botanic Gardens, Singapore, a specimen collected by Haviland and bearing the number of the type, which was “decidedly woody,” whereas Stapf described it as “herbaceous and scandent.” Ridley was right in so far that the plant is certainly no climber, and that it is woody at the base: as secondary thickening, how- ever, is by no means uncommon in the basal parts of the stems of dicotyledonous herbs, this is no valuable argument against its herbaceous character. In the following year Elmer (Leafl. Philipp. Bot. 4: 1356. 1912) published some remarks on a plant which he had collected in the island Palawan, and which he referred to Str. glabra Val. It differs, however, conspicuously from the latter by the nature of the stipules, whose lobes are separated from each other by a rather wide gap instead of being contiguous. His remarks contain several mistakes: the valvate corolla- lobes are described as imbricate and slightly twisted from left to right; 1947] BREMEKAMP, THE GENUS STREBLOSA 149 the stamens are said to be inserted in the basal part of the corolla-tube, and the anthers should be more or less united, the stigma ‘‘submitraform,” the ovary “superior or nearly so, surrounded by 5 lobular disk append- ages,’ and the ovules should be pendulous. That the genus is said to be monotypic and previously known from Borneo only, may be regarded as a slip of the pen. In the same year Valeton (in Icones Bogorienses 4: 139, t. 343. 1912) published a detailed description of Str. tortilis (Bl.) Khs., accompanied by a plate which unfortunately is not entirely satisfactory. His ideas on the generic characters are summarized in the following sentence: ‘Le port de la plante, les inflorescences axillaires, l’insertion scorpioide des fleurs sur les rameaux (voir la figure 2 dans la table citée de Stapf) et la structure du fruit sont autant de caractéres de valeur générique et qui séparent la Streblosa du genre Psychotria.”’ Valeton therefore was the first to recognize the taxonomic importance of the monochasial structure shown by the ultimate ramifications of the inflorescence, and although he calls the latter here for sake of convenience axillary, the remarks in his paper in Bot. Jahrb. 44: 568. 1910, show that he knew very well that they are in reality at first terminal and afterwards pseudo-axillary. A few years later Merrill (in Philipp. Jour. Sci. Bot. 10: 141. 1915) described under the name Str. axilliflora, a new species from the oo In a short note attached to the description he remarks: ‘Ia agree- ment with Valeton and with Ridley in considering Streblosa Korthals to be a valid genus. In aspect the plants are entirely different from Psychotria and Grumilea, and the differential characters appear to be constant. From all our representatives of Psychotria and Grumilea the present species is distinguished by its minute fruits and axillary inflores- cences; surely, if Grumilea is to be separated from Psychotria, there ee as great or greater reasons for distinguishing Streblosa as a valid genus.” Merrill afterwards (Enum. Philipp. Fl. Pl. 3: 564. 1923) sabe the Philippine species with Str. glabra Val., but this was an error: it differs from the Bornean species in the same way as the plant collected - Elmer in Palawan, the stipular lobes namely are not contiguous but separated from each other by a rather wide gap, and the internodes are ecostate. Since then one new species was described from the Malay Peninsula by Ridley and eight from Sumatra and Borneo by Merrill. With regard to our knowledge of the generic characters however, no further progress was made. In this paper a detailed analysis of the generic characters is given. Twenty-five species are described, and an attempt is made to arrange them in a more or less natural sequence. Eleven of the twenty-five species are new, whereas three of the old ones, namely Sér. hirta Ridl., Str. platyphylla Merr. and Str. puberula Merr. are reduced to synonymy, and one, Str. pubescens Ridl., set apart as a “‘species incertae sedis.” Of Str. urticina Stapf and of Str. palawanensis Brem. two varieties are recognized, and of Str. axilliflora Merr. three. 150 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVHI GENERIC DESCRIPTION Streblosa Khs. in Ned. Kruidk. Arch. 2 (2): 245. 1851; Miq. in Ann. Mus. Bot. Lugd.-Bat. 4: 211, 262. 1869; Stapf in Trans. Linn. Soc. oe II, 4: 182. nh Val. in Engl., Bot. Jahrb. 44: 567. 1910; Ridl. in Jour. As. Soc. Straits 57: 6 1911; Val. in Ic. Bog. 4: 141. 1912; Elmer, Leafl. oa Bot. 4: 1356. ae Merr, in Philipp. Jour. Sci. Bot. at 141. 1915, Enum. Philipp. Fl. Pl. 3: 564. 1923; Ridl., Fl. Mal. Pen. 2: 148. 1923; Merr. in Univ. Calif. Publ. Bot. 15: 288. 1929: Lemée, Dict. Pl. Phan. 6: sa 1935; Merr. in Mitt. Inst. Allg. Bot. Hamburg 7: 295. 1937, in oe Mich. Acad. Sci. 22: 194. 193! Psychotria - Bl., Bijdr. Fl. Ned. Ind. 958. 1826; DC., Prodr. 4: 520. 1830; Miaq., Fl. Ind. Bat. 2: 294. 1857, eel 223, 1860; Hook. f. in Benth. et Hook. f., Gen. Fl. “ae 125. 1873: Boerl., Handl. Fl. Ned. Ind. 2 (1): 139. 1899; Dalla Torre & Harms, Gen. Siph. 503. 1905; King & Gamble in Jour. he Soc. Beng. 72 (2): 11. 1906. Uragoga spec. Baill. in Adansonia 12: 325. 1879; Hist. Pl. 7: 285. 1880. Genus Psychotriearum caule sympodiali esis simplici, inflores- centiis primum terminalibus, deinde pseudo-axillaribus et ergo semper solitariis ad nodos, ramulis ultimis inflorescentiae quoque nodo floribus duobus instructis, florum paribus cincinnaliter dispositis, fructibus parvis et viridibus, globosis vel didymis, in mericarpia duo discindentibus, meri- carplis drupaceis, pyrenis dorso 3-costatis s, facie ventrali excavatis et excavatione a carina mediana plus minusve distincta in sakiog duas divisa a generibus aliis ad tribum hance pertinentibus distinguendun rbae_ plerumque simplices, raro semel vel bis as -dichotome furcatae, erectae vel saepius ascendentes, casu quo parte decumbente haud raro radicante et interdum ex axl illis innovationes emittente. Caulis 9 paniculatim dispositi, ramulis ceterum haud raro semel dichasialibus, ramificationibus monochasialibus, raro inflorescentia tota semel vel bis dichasialis, ramificationibus monochasialibus, vel tota monochasialis. Monochasia quoque nodo floribus duobus instructa, florum paribus cincin- naliter dispositis. Bracteae forma et magnitu udine diversae sempe evolutae, interdum ad anthesin deciduae. Flores pedicellati vel subsessiles, semper 5-meri et heterostyli. Ovarium biloculare, loculis ovulo singulo ex angulo interiore ascendente instructis. Calyx plerumque 5-fidus, lobis triangularibus, ovato-triangularibus vel Send hay 5-partitus, casu quo lobis interdum imbricatis. Corolla alba, maxime 8 mm. longa sed plerumque multo brevior, breviter eee ela ees interdum dimi- dio superiore paulum inflato, lobis valvatis, tubo intus ad medium vel supra medium annulo vel fasciculis 5 pilorum icici. gees in annulo vel 1947] BREMEKAMP, THE GENUS STREBLOSA 151 inter fasciculos pilorum inserta, in flore longistylo subinclusa, in flore brevistylo exserta; filamenta glabra; antherae oblongae, dorsifixae. Granula pollinis globosa, minute reticulata, 3-pora. Discus semi- -globosus el annuiaris, apice interdum sulco instructus, glaber. Stylus glaber, in flore brevistvlo inclusus, in flore longistylo exsertus; stigmata ‘iliformia vel linearia, parole in flore longistylo apice recurvata. Fructus parvus, minus quam 3 m iam., viridis, globosus vel didymus, calyce persistente et disco plus minusve exsiccato coronatus, in Map una duo discindens: mericarpia drupacea; pyrenae dorso 3-costatae, facie ventrali excavatae, excavatione a carina si ga plus minusve ieee in parice duas divisa, endocarpio tenui. Semen endocarpio conforme, testa rubra vel rubro- maculata, albumine ane corneo, embryone recto parvo, radicula inferiore. Species adhuc notae 25, distributae in Peninsula Malayana, Sumatra, Java, Borneo, Insulis Filippinis, ubi in umbris nemorum crescunt SPECIES TYPICA: Str. tortilis (Bl.) Khs. The stems are either erect or, more often, ascending, and there is doubt- less a certain amount of vegetative propagation, young plants developing sometimes from the axils of the fallen leaves in the decumbent and rooting basal part. Even in those species that are provided with erect stems, for instance in Str. Johannis-Winkleri Merr., in which this kind of growth is perhaps most pronounced, I have sometimes found remains of old shoots at the base of the young ones. I suppose that in these plants the whole stem sinks down after the last fruits have ripened, and that here too new stems develop from axillary buds. That these innovations are but rarely met with in herbarium specimens, probably means that they are not regularly formed. In some of the species with ascending stems, “especially in those in which the leaves do not show much difference in shape and size, e.g. in Str. tortilis (Bl.) Khs. and in Str. bullata Merr., development might even be continuous, the oldest part of the stem gradually decaying and the bent at the same time shifting in the direction of the younger parts: in such species innovations are perhaps formed only when the normal development in some way or other is checked. amification is extremely rare: so far it has been observed in three species only, in Str. lampongensis Brem., Str. chlamydantha Brem. and Str. longiscapa ees It is always brought about in the same way, namely by the development of two axillary shoots instead of one at the base of an inflorescence. This is rather interesting from a morphological point of view, for it definitely proves the terminal nature of the inflorescence: wfihiout the evidence of the forked plants the latter would always remain somewhat dubious, for the terminal position of the young inflorescence might be spurious, and the presence of but one inflorescence at each node is in itself no sufficient proof. In habit the Streblosa species therefore are not unlike those belonging to the genus Votopleura Brem. (Psychotria L. sect. Notopleura Hook f. in Benth. & Hook. f., Gen. Pl. 2: 124. 1873, cf. Brem. in Rec. d. trav. bot. Néerl. 31: 289. 1934), for the species of that genus too are herbaceous and 152 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII provided with usually simple, sympodial stems, the terminal inflorescence being pushed aside by an axillary shoot Suelinnine from the axil of one of the leaves at its base. In other respects, however, there are important differences. In Notopleura the leaves are succulent, the stipules undivided, the inflorescences paniculate or corymbose, the flowers ebracteate, the corolla-lobes imbricate, and the pyrenes dorsiventrally compressed and on the inner side flat or provided with a prominent keel; the genus, moreover, is confined to Tropical America. The affinity between Streblosa and Notopleura, therefore, must be regarded as rather remote. Although the stems probably do not last for more than a few months, there is often a considerable amount of secondary thickening in their basal parts. As a rule the increase is almost entirely due to the formation of secondary xylem, but in the species described below under the numbers 1-11, there is also a considerable development of secondary bast. e formation of cork is, as a rule, but weak, and in most species the stem, therefore, remains green. As in most of the plants belonging to the undergrowth of the forest, the leaves are, as a rule, large, flat, thin and discolorous. In the majority of the species belonging to the subgenus Eu-streblosa (Str. tortilis, Str. myriocarpa, Str. maxima, Str. Johannis-Winkleri, Str. bullata, Str. brac- teolata, Str. anambasica, Str. polyantha, and to some extent in Str. deliensis and Str. scabridula) the underside of the leaves is covered with greyish scales, which under the microscope prove to be dome-shaped excrescences crowned by a stoma, an arrangement which apparently is not uncommon in shade plants. Taxonomically of more importance is the structure of the epidermis on the upper side. As in Psyvchotria the latter consists of straight-walled cells, but in the midst of these some much smaller ones are found which are produced above the common surface into lenticular or conical excrescences (fig. 1). The latter are often difficult to detect with the naked eye, but appear to be present in all species. In the other genera of the Psychotrieae I have looked, so far, in vain for them. Far more conspicuous are the reddish or black dots with which the upper side of the leaves of Str. tortilis, Str. rsa ae Str. polyantha, and probably those of Str. multiglandulosa, which I could not investigate myself, are marked. These dots are caused by the aie of a dark-coloured resin in some of the epidermis cells. As a rule they are in these species a very conspicuous feature, but as a differential character they are of little value, for their development varies considerably, even in leaves of the same specimen; in some instances the cells contain little or no resin, and as the cells themselves are indistinguishable from the rest of the epidermis cells in shape and size, microscopical investigation too is in such cases of no avail. Of Str. tortilis and Str. microcarpa I have seen specimens in which the dots appeared to be completely absent. esin may also be present in some of the cells of the mesophyll and in the mesocarp of the fruit, and even in other parts of the plant, for instance in the cortex and in the secondary bast of the shoots; in the latter these 1947] BREMEKAMP, THE GENUS STREBLOSA 153 cells are arranged in short longitudinal rows. Their number varies con- siderably, and they are not always easy to detect, but it is almost certain that they are nowhere completely absent. One of the most suitable places to look for them, is the connective. In the species 1—6 they are a con- spicuous feature of the secondary bast. When they are present in the mesophyll, the leaves are, as a rule, more or less distinctly marbled. It is noteworthy that such marbled leaves occur also in some related plants, e.g. in the Madagascar species of Psychotria. Raphide-cells are in some species a conspicuous feature, whereas in others they are difficult to detect. This depends apparently on the consistency of the leaves, for microscopical investigation reveals that they are nowhere absent. As a matter of fact, they are a general feature of the whole tribe, and are also found in some of the latter’s allies, for instance in the Hedyotideae, whereas they are constantly absent in the Ixoreae and their allies, which, as I have pointed out in my ‘“‘Monograph of the genus o 100 150 200m n i n ry eee eerie | Fic. 1. Epidermis of the upper side of the leaf of Streblosa myriocarpa Merr. 154 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Pavetta” (in Fedde’s Repert. 37: 11, 12. 1934), form another circle of affinity. The nervation of the leaves is very uniform; the meshes formed by the lateral nerves and the venules are large and more or less irregular in outline. Acaridomatia are always absent The stipules are always interpetiolar and produced into two lobes, but they vary considerably in size and shape. In the species belonging to the subgenus Eu-streblosa they are much larger than in the subgenus Para- streblosa, and in the first-named subgenus the basal part is stronger developed than the lobes, whereas in the species belonging to the latter the lobes are better developed than the basal part; in Eu-streblosa they are moreover more or less persistently ciliate or ciliolate, whereas in Para-streblosa they are eciliate. In part of the species belonging to the latter the basal part shows a prominent midrib which runs down to the preceding node. In these species, which form the series Costatae, the lobes are moreover contiguous, whereas in the two other species, which are brought together in the series Ecostatae, the midrib is invisible and the internodes accordingly ecostate, and the lobes are separated from each other by a wide gap. Costate stipules are found also in some of the species belonging to the subgenus Eu-streblosa, but only in one of them, Str. Johannes-Winkleri Merr., the rib runs down along the stem. In the other species all the nerves arise from the angles of the stipule and converge towards the top of the undivided part, where they enter the lobes. As in many other genera belonging to this family, a fairly large number of colleters is found in the axils of the stipules. ‘They are in this genus rather large. The inflorescences are always of terminal origin, but by the development of a branch from the axil of one of the leaves at their base, they are, as stated above, soon pushed aside. ‘They are, as a rule, shortly pedunculate, but in Str. tortilis the peduncle is about as long as the petiole of the leaf at its base, and in Str. longiscapa it is even about as long as the whole leaf. In these two species the peduncles are moreover erect, whereas they are in the other species of the subgenus Eu-streblosa recurved, and in those belonging to Para-streblosa patent. The inflorescences themselves show various degrees of complexity. The simplest form is that found in Str. bracteolata Merr.: it consists of a single axis bearing at each node a sterile bract and a pair of flowers, one provided with a bract and the other ebracteate. The ebracteate flower might be regarded as a superposed one, and the whole inflorescence as a raceme, but it seems more plausible to regard the ebracteate tlower as a terminal one, and the bracteate flower as representing one of the branchlets of a dichasium, the other branchlet forming the continuation of the axis (fig. 2). This interpretation is supported by two weighty arguments: (1) the ultimate ramifications of the inflorescences of the Psychotrieae are everywhere of the dichasial type, and there is therefore every reason to assume that those of Streblosa will be of the same or at least of a 1947 | BREMEKAMP, THE GENUS STREBLOSA 155 related kind; (2) in almost all the other species of Streblosa the branchlets are at least partly arranged in pairs with a single flower in the fork between them, which means that they themselves are parts of a dichasium, and it is a well-known fact that the branchlets of a dichasium often develop into monochasia, whereas there are no fully authorized examples of dichasia developing into racemes or spikes. The flower-pairs are arranged in such a way that a line connecting the successive pairs would twine round the axis like the shoot of a winding plant. The arrangement of the flower-pairs, therefore, is cincinnate. The inflorescence, however, is not a simple cincinnus, for in a cincinnus we find but one flower at each node. For the same reason it is not exactly a monochasium, although we will use this term here for the sake of convenience, but more or less intermediate between the latter and the dichasium. In Str. bullata Merr. the inflorescence is twice dichasially branched with rather long-pedicellate flowers in the forks, but after the second bifurcation G d € Fic. 2. Diagrams illustrating the structure of the aaa branchle a. vertical projection; 6. the same, but the terminal flowers connec by a ot line representing the axis; this figure shows perhaps better than a a ae lateral flowers o not stand before the terminal ones, but more or less beside them; c. the lateral shoot which is to form the continuation [ the inflorescence axis in the position which theoretically it ought to have; d. the condition found in Para-streblosa; th condition found in Eu- sipehione: at * the | renter beacteal” node. 156 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII the branchlets develop in the same way as the inflorescence of Sér. bracteolata. In the other species we find three or five subumbellately or, rarely, racemosely arranged branchlets, of which the lowest pair are sometimes in their turn provided with a pair of lateral branchlets; the main branchlets as well as those of the second order may be once dichasially ramified, but the prongs of these dichasia, or the branchlets themselves if they are unforked, are monochasial. The monochasial structure of at least a part of the inflorescence, is a character of great taxonomic importance, for this arrangement is found nowhere else in the Psychotrieae. It is in cymous inflorescences always somewhat difficult to make out how the terms bract, bracteole and pedicel should be used, but if we stick to the principle that the bud in the axil of a bracteole remains rudimentary, we may find a way out of the difficulty, for in this case the only hypsophylls to which the name could be applied, would be the ultimate ones in whose axils the flowers remain rudimentary, but as these hypsophylls are indubi- tably homologous with the other ones, there is obviously no reason to give them another name: in the cymous inflorescences there are therefore no bracteoles. When this is conceded, the determination of what should be called pedicel, offers no difficulty: in a dichasium it is obviously the axis which begins at the insertion of the dichasial prongs or at the insertion of the bracts in whose axils the latter develop. In a monochasium the base of the pedicel is accordingly recognizable at the insertion of the bract or of the bract-pair. In Streblosa the matter is somewhat complicated by the circumstance that the two bracts which one would expect to find at the same height, are in reality often inserted at different heights. This means that between the two of them an extra internode has developed, which we will call the interbracteal internode. ‘The shoot which forms the continuation of the axis of the monochasium develops always from the lower bract, which, as a rule, is somewhat larger than the other one, in whose axil the lateral flower develops. In the subgenus Eu-streblosa the interbracteal internode is always easily recognizable, whereas in Para- streblosa it is either absent or indistinct. In Para-streblosa the flowers, moreover, are alwavs subsessile, whereas in Eu-streblosa subsessile flowers are met with in Str. ¢tortilis and in Str. microcarpa only; in the other species they are distinctly pedicellate. Because of the presence of the interbracteate internode the flower-pairs of Eu-streblosa give one the impression of being stipitate, and for the sake of convenience, this term has been used in the descriptions of the various species. The flowers are always 5-merous and heterostylous. It is true that those of Str. urticina were described by Stapf as 4-merous, but this is a mistake: in the drawings on the plate accompanying his description they are correctly shown as 5-merous. ‘They are as a rule very small. The largest ones which I could investigate, are those of Str. bracteolata, which reach a length of 9 mm.; in several other species they vary between 2.5 1947 | BREMEKAMP, THE GENUS STREBLOSA 157 and 3.5 mm. Heterostyly is very often met with in the Psychotrieae and their allies; it occurs, for instance, also in the Hedyotideae. It is, however, no general feature of these tribes. The ovary is always bilocular and, as everywhere else in the Psychotrieae, each cell contains a single ovule ascending from the angle between the bottom of the cell and the septum. The calyx consists of a short tube and five usually short, broadly triangular, ovate-triangular or suborbicular, rarely somewhat longer and then oblong, lobes, alternating with baculiform glands. ‘The latter are like the previously mentioned raphides and the tendency to heterostyly a general feature of the Psychotrieae and of some of the related tribes, e.g. of the Hedyotideae. The aestivation of the calyx-lobes is usually more or less open or subvalvate; the large calyx-lobes of Str. chlamydantha Brem., however, are distinctly imbricat The corolla is always white, and seasiais of a short cylindrical or, in the upper half, above the insertion of the stamens, slightly widened tube and five spreading lobes, valvate in the bud. In the middle or somewhat above the middle the tube is provided with fairly long hairs. In Eu-streblosa the latter form a complete ring, whereas in Para-streblosa they are arranged in five bundles alternating with the stamens. The stamens are inserted at the same height as the hairs. In the long-styled flower they are more or less included, and in the short-styled one exserted. The filaments are always glabrous, and the anthers oblong and dorsifixed. The pollen grains are small, globose, 3-porous and minutely reticulated. The disk is, as a rule, fairly large, semi-globose or annular, and glabrous. At the top it shows a transverse impression corresponding in position with the septum in the ovary. The style is thin and glabrous, and ends in two filiform or narrowly linear stigma-lobes. In the short-styled flower it is included, and the stigma-lobes are but slightly spreading, whereas in the long-styled flower it is exserted, and the stigma-lobes are more distinctly spreading and at the top sometimes recurved. The fruits are very small, smaller, in fact, than anywhere else in this tribe, and in size and shape more like those of Hedyotis, Lerchea or Xanthophytum than like the drupes of the other Psychotrieae. They are either globose or, more often, didymous, crowned by the persistent calyx and the somewhat shrunken disk, and they remain always green. When fully ripe, they split into two deaaceius mericarps, which in the herbarium are distinctly costate, but in the fresh condition smooth. The mesocarp is but weakly developed, and the endocarp too is thin. The pyrenes vary in colour between yellowish, red and almost black, and they are usually 5-ribbed, three of the ribs ornamenting the convex side, and the two others marking the border of the latter; the flat inner side shows in the centre an excavation, which a more or less distinct keel divides into two parallel or upwards somewhat divergent hollows. The keels may be 158 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXVIII acute or obtuse, and the excavation on the flat side varies in the various species somewhat in size and shape, but the differences are difficult to describe, and they are therefore not mentioned in the descriptions. The seed shows exactly the same sculpture as the pyrene, and consists of a thin, red or red-dotted testa and a uniform horny endosperm. The embryo is small and nearly straight, and the radicle points downwards, The shape and sculpture of the pyrene, and the way in which the testa surrounds and eventually penetrates the endosperm, are important diag- nostics for the genera belonging to the Psychotria group. Streblosa is for instance easily distinguishable from Mapouria Aubl., to which the type- species was referred by Hooker in Benth. & Hook. f., Gen. Pl., by the presence of the two contiguous hollows on the ventral side of the pyrene, and from Psychotria L. in the delimitation in which it is taken by me (cf. Brem. in Jour. Bot. 71: 273. Oct. 1933, and in Pulle’s “Flora of Suriname” 4: 261. 1934) by the absence of the median groove on the ventral side of the pyrene and the corresponding intrusion of the testa into the seed. Apart from the difference in size, the pyrenes of Streblosa, however, are very similar to those of Geophila Don; the fruits of the latter, on the other hand, are not only larger but also more fleshy, and coloured instead of green. The pyrenes of Notopleura, the genus which resembles Streblosa in habit, are dorsiventrally compressed, and their ventral side is either entirely flat or provided with a protruding ridge, but the latter does not arise from the bottom of a cavity. Of these plants too the fruits are coloured and larger and more fleshy than those of Stredblosa. Within the genus Séredblosa itself two slightly different kinds of fruit are met with. In the subgenus Eu-streblosa and in Str. chlamydantha and Str. longiscapa the mericarps are nearly globose, and the fruits there- fore didymous, but in the other species of Para-streblosa the mericarps are semi-globose, and the fruits accordingly globose: in herbarium material the latter are 8-ribbed, whereas the didymous fruits are either 6- or 10-ribbed. In the species which are normally provided with didymous fruits, occasionally some globose ones are found; the latter are 5-ribbed and owe their aberrant shape to the abortion of one of the mericarps. CLASSIFICATION OF THE SPECIES! The genus has here been divided into two subgenera. In Ey-streblosa the stipules are wider than the stem, and their basal part is more strongly developed than the lobes, the flower-pairs are distinctly stipitate, i.e. the interbracteate internode is always developed, the hairs in the corolla-tube form a complete ring, and the fruits are always didymous. In Para- streblosa, on the contrary, the stipules are not wider than the stem, and their lobes are better developed than the basal part, the flower-pairs are 1 The abbreviations for the herbaria of alk institutions cited in this paper are as follows: AA = Arnold Acborein BZ Buitenzorg Botanic Garden; HBG = Inst. Allg. Bot., Siete K = Ro waht — Gardens, Kew; NY = New York Botanical Garden; U = Bot. Mus. Utrec 1947] BREMEKAMP, THE GENUS STREBLOSA 159 subsessile, i.e. the interbracteate internode is either absent or but indis- tinctly developed, the hairs in the corolla-tube are arranged in five bundles, and the fruits are usually globose; didymous fruits, however, are found in Str. chlamydantha and in Str. longiscapa. A further subdivision of Eu-streblosa does not lead to results which are of practical value, the groups becoming too small. The first two species, Str. tortilis (Bl.) Khs. and Str. microcarpa Ridley with their subsessile flowers, are doubtless nearly related, but Str. lampongensis Brem., a species with distinctly pedicellate flowers, shows in the structure of its stipules with their comparatively long lobes, an unmistakable likeness with these two species. The next three species, Str. bracteata Ridl., Str. myriocarpa Merr. and Str. maxima Brem., form another natural group, characterized by large leaves and very large, oblong or obovate stipules. Ovate or ovate-lanceolate stipules provided with a prominulous or promi- nent midrib characterize a group of Bornean species, to wit Str. Johannis- Winkleri Merr., Str. bullata Merr. and Str. bracteolata Merr., to which the less known species Str. multiglandulosa Merr. and Str. undulata Khs. interpr. Val. probably also belong. A somewhat isolated position is occupied by Str. anambasica Brem.; in the small size of the flowers it resembles the Sumatran species which form the next group, but it differs conspicuously from this group by the more strongly ramified inflorescences. The five Sumatran species, Str. deliensis Brem., Str. hypomalaca Brem., Str. leiophylla Brem., Str. polyantha Khs. and Sir. scabridula Brem., are provided with ecostate stipules, and possess infiorescences with three undivided branchlets and very small flowers. In Para-streblosa two series may.be distinguished: the Costatae with conspicuously costate stipules and bicostate internodes, and the Ecostatae with ecostate stipules and smooth internodes; in the first series the stipular lobes are contiguous, whereas in the second they are separated from each other by a wide gap. e first are confined to Borneo, whereas the second group is found in Palawan and other islands of the Philippines. The latter comprises at present but two species, one consisting of two, and the other of three varieties, but it is not impossible that further study will show that these varieties are sufficiently distinct to be regarded as species. Among the Costatae, Str. chlamydantha Brem. and Str. longiscapa Brem. occupy a more or less isolated position. They are provided with didymous fruits, and their calyx-lobes are larger than in the other species. KEY TO THE SPECIES a. The undivided part of the stipules wider than the stem and at least 2 mm. high, but, as a rule, much higher. Hairs in the corolla-tube in a complete ring. Fruits always didymous. — Species from the Malay Peninsula, Sumatra, Java and Borneo. Eu-streblosa b. Stipules with an ovate base and two slightly spreading lobes, which are longer than the base. Flowers either sessile or pedicellate, but the pedicels never more than 0.6 mm. long. 160 ao a : Stipules ovate, obovate or oblong in outline; JOURNAL OF THE ARNOLD ARBORETUM {VOL. XXVIII Cc. rt of the flowers of each pair sessile; the pedicel of the other one not mo long (pedicels of the flowers in the forks of the dichasia sare 8 longer). d. Leaves ovate or ovate-elliptic, less than twice as long as wide, and rarely cm. long; on the underside lepidote. Peduncle about as long The larger floral bracts about 1 mm. Str. tortilis. more than 10 as the petiole of the seg mcd _ long. —Java.and West. Sumatra... i.csica ducuie ccs cumsacinnas Leaves lanceolate- or Mame nara more than twice as lone. as wide, and nearly always more than ong; not lepidote on the underside. Peduncle much shorter than the petiole of the preceding leaf. The larger floral bracts at least 1.2 mm. long. — Malay Peninsula - Sumatra....... 9:05 ae iedrsapvasts do chcanawale Goad sia ker Area ONDA adie Gib Brapy dey Som ESAS . Str. microcar pa. e All flowers distinctly pedicellate; the pedicels about 0.6 mm. long. — South 3. Str. lampongensis. the incision between the two lobes ntre. Flowers always pedicellate, and the pedicels d: not reaching beyond the cen g. e. Leaves at least 15 cm. long. Stipules oblong or obovate. — Bornean species. m = Leaves slightly pubescent on the midrib beaneath. — West Psat evans esac ow nope ade Alea eevee cont aeess eae eppaytent Str. bracteata. f: Stem aaa rufous-pubescent. Leaves on the underside on midrib, nerves and venules densely pubescen g. Petioles 1-2 cm. long; Weds 15-26 c 7-13 cm., subobtuse, — with about 20 pairs of nerves. Midrib of the stipules indistinct. — West BiOTME cw ceurcatrcs.ckee thr ornb bracts 8. BOTS fers us ath arsa-a ole ards ia Gos Ae AA we CLE ee 2. Involucral bracts either ASOLO. occ aca io oS ate GH 6 chs $0 sss 5.8 Se en eee 3, Involucral’ DractS 5.6.65 4650555 bee oi oe cl eo aE tha Bde RES 7 w _ = - sate ~~ co = =) ~ — = o = > a oo wn os — oO r=C) wr oO A 3 i ro) o 3 o oO o =a pf =} n °° —_ io) =] Q . Shoots decumbent. Leaves ovate-elliptic or elliptic, acute or suddenly contracted at the base, on the upper side sparsely hirsute and not more than 4 cm. long. Inflorescences subsessile. Ovary and _ capsule clinica — South Sumatra and West Java repta acker ex Brem. : Shoots erect. Small, not more than 7 cm. high herb. ‘Leaves lanceolate-elliptic, cuneate at the base, on the upper side scabrid, 4-6 cm. long. Inflorescence pedunculate. Ovary and capsule pilose. — East Sumatra —.................. ieee gee wean Ga ese Gas saw een Aa pau Se Sea aan ee a ‘anleewweni Brem. 5. ni tices ag: branched plant. Tip of ny stipule yee Involucral bracts free, the arger ones decussate, each of them subtending a cyme; the two pairs of ik er ones representing the oat pairs of sees of the cymes subtended by the lower pair of involucral Bracte.-=Simnal lur (Simeulowe) — S55 ey Se sees 8 ake EERE PRE CORP ERES LS AS a pT ea Os malurensis Brem. : Sv — plants. Tip of the stipule undivided. Involucral betots united at the 6 > oa ase a flat receptacle, all pear CYIMIGS i 4.55664 ek bogs ea pee hay ores ee 6. Bracts sending the branchlets of the cymes at the base on each side with patent lobe. Calyx-lobes jie more than 3 mm. long, covered with ferrugineous hairs. — Sumatra and the islands to the west of it—...... . P. assimilis Brem. : Bracts n never with two and rarely with a single lateral lobe. Calyx lobes at least 5 . long, covered with grey hairs. — Northwestern Sumatra—............ raid Gus hauei ork aioe nies is Ate ah eee casane te gta-e inter Pause ar ete s ae ue nae a (Miq) Brem. . All flowers cymose. Bracts all ovate; those of the inner prea shorter than the corolla.— Malay Peninsula and Sumatra—.............. . sylvestris Ridl. : Flowers in the axils of the outer bracts in triads; in the sin of the inner ones solitary. Bracts oblong to spathulate, those on the main axis always longer than the corolla. — West Java—.................. 7. P. pilosa (Miq.) Loin 8. Involucral bracts 4. Flowers pedicellate.— Malay Peninsula, a atra, SG aie wikia ia ace seg sata Eta ty-a Gan Ryan aua'estal 4s w AEG Da etavelonele te ais dice sara 8. P. rivularis Hen : Involucral bracts 5 “wr 1947 | BREMEKAMP, THE GENUS POMAZOTA 195 9. Leaves on the upper side glabrous. Tip of stipule bipartite. Inner bracts spathulate. — Sumatra — .....----- sees eee ects eres 9. P. scabiosiflora Brem. - Leaves on the upper side sparsely hirsute. Tip of stipule undivided. Inner bracts linear. — Batu Islands — .........-.- eee eee ee cere eres 10. P. batuensis Brem. TAXONOMY! Pomazota Ridl. in Trans. Linn. Soc. II, 3: 308. 1893; K. Schumann in Engl. & Prantl, Nat. Pflanzenfam., Nachtr. 1: 310. 1897; Ridl. in Fl. Mal. Pen. 2:43, 1924; Lemée, Dict. Pl. Phan. 5: 502. 1934. Coptophyllum Khs. in Ned. Kruidk. Arch. 2 (2): 161. 1851; Miq., Fl. Ned. Ind. : 175, 348. 1857, Suppl. 226. 1860, in Ann. Mus. Bot. Lugd.-Bat. 4: 230. 1869; Benth. & Hook. f., Gen. Pl. 2:68. 1871; Boerl., Handl. Fl. Ned. Ind. 2 (1): 57. 1891; Ridl., Fl. Mal. Pen. 2: 43. 1923; Lemée, Dict. P]. Phan. 2: 297. 1930; non Coptophyllum Gardner in Hook,, Ic. PI. 5: t. 477, 478. 1842. Genus Rubiacearum Hedvotidearum, corollae tubo intus paulum infra orem annulo pilorum instructo, stylo antheras excedente, capsula globosa ramulos oppositos breves, ex internodio singulo constantes et foliis ad stipulas redactis instructos terminantes, nunc dichasialiter ramificatus vel sympodialis, casu quo inJorescentiis primum terminalibus, deinde in ct atae, 1- pulatae, axilla glandulis aliquibus baculiformibus instructae; exteriores 4. 5 vel 8 aliis latiores, capitulum involucrantes; omnes fertiles et nunc cymas bi- vel i i bibracteolatos < Oo o, a =“ —_ sob) Oy Oo wn | cS -Q 2 2} (oo) = fq?) if) ae = ga om c © 1d 2] Corolla alba, luteola vel carneola, hypocrateriformis, parva, tubo interdum paulum inflato, infra orem plus minusve annulatim constricto et ibidem intus annulo pilorum patentium instructo, ceterum glabro, lobis ovatis acutis, flore aperto patentibus, costa extus ciliatis. Stamina ad medium 1 The abbreviations for the herbaria of the institutions cited in this paper are as follows: AA = Arnold Arboretum; BZ = Buitenzorg Bo anic Garden; HBG = Inst. Allg. Bot., Hamburg; K = Royal Botanic Gardens, Kew; NY = New York Botanical Garden; U = Bot. Mus. Utrecht. 196 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII vel infra medium tubum inserta, filamentis glabris, es lineari- oblongis, subbasifixis, apiculatis, plerumque inclusis. Discus annularis, conicus vel cylindricus, glaber vel farinosus. Stylus ai versus vix conspicue papillosus, ceterum glaber, tubo subaequilongus vel eo paulo longior; stigmata crassa et brevia, basi patentissima, apice paulum re- curvata. Fructus capsula globosa, disco conspicue rostrata et calyce coronata, pariete tenui, ultimo dimidio superiore tabescens et partem apicalem cum calyce et disco operculi instar rejectans. Semina numerosa, runnea, angulosa, alveolata, alveolis fundo granulatis. Distributum speciebus adhuc notis 10 in Peninsula Malayana, Sumatra, Java, terra Borneénsi, Lectotypus: P. capitata (Miq.) Brem. n. comb. (Coptophyllum Miq.). Series A. Sympodiales florescentiae primum terminales, plerumque mox a ramo axillari in positionem lateralem coactae, interdum, cum rami axillares duo evolvuntur, rca ramorum_ remanentes. Pedunculus age numquam _squamis stipulaceis instructus. Flores axillis bractearum pro parte minime in cymas vel triades dispositi (in P. reptante aia omnes singuli). Species 1- Bracteae involucrantes 4. a. Herba decumbens, foliis ovato-ellipticis vel ellipticis parvis, in- florescentia subsessili parva, ovario capsulaque glabris. 1. Pomazota reptans Backer in sched.; typus: Backer 31142 (BZ). Herba ramosa, partibus apicalibus primum aura usque ad 12 cm. altis, post anthesin decumbentibus et radicantibus. Cau les 1.2 mm. diam., densius fusco-pubescentes, vivo ru ri, Folia petiolo pubescente, 3-15 mm. longo, vivo ay instructa; lamina elliptica vel ovato-elliptica, 2.5-4 cm. longa et 1.3—2.2 cm. lata, apice acuta vel breviter acuminata, apice indivisae, margine ciliatae, ceterum glabrae. Inflorescentia sub- sessilis, a ramo axillari mox in positionem lateralem coacta, 8-12 mm. diam. Bracteae involucrantes 4, decussatae, ellipticae vel obovatae, 6-9 n mm. longae et 4-6 1 latae, margine ciliatae; bracteae interiores 2 vel 4, spathulatae, exterioribus aequilongae; omnes extus intusque fusco- hirsutae. Bracteae papa ntes plerumque p tades tr ades, 1 pemptadum tamen i bao rudimentariis; cymae in axillis bractearum Bracteae florum lateralium pemptadum vel triadum spathulatae, basi conduplicatae, 4 mm. longae; bracteae florum sequentium et bracteolae flor'um singulorum eis angustiores, 2—2.5 mm, longae. Flores 5-meri. Ovarium glabrum. Calyx lobis ovatis, 0.7-1.3 mm. longis et 0.4—0.5 mm. latis, subobtusis, margine ciliatis sed ceterum glabris. Corolla alba tubo 3.5 mm. longo, 1.2 mm. infra orem annulo pilorum instructo, lobis 1—1.5 mm. longis. Stamina 0.5 mm. supra basin fbi inserta, filamentis 2.3 mm., 1947] BREMEKAMP, THE GENUS POMAZOTA 197 antheris 0.8 mm. longis. Discus ii 0.5 mm. altus. Stylus 4mm. longus. Capsula glabra, 1.5 mm. dian Hab. Sumatram Australem et i Occidentalem. SumatrA: Lampongs: G. Rate Telangaran, Jbut 24 (BZ); G. Tanggamus, alt. 700 m., de Voogd 161 (BZ). Java: Res. Buitenzorg: Lewiliang, alt. 500 m., v. Steenis 2717 (BZ); G. Parungpung near Tjampea, alt. 750 m., Bakhuizen v. d. Brink 4174 (BZ); ibid., Bakhuizen v. d. Brink Jr. 610 (U); Kalapa Nunggal, alt. 200 m., Backer 23381 (BZ); ibid., alt. 400 m., Backer 5992 (BZ); ibid., alt. 500 m., Backer 31142 (BZ, TYPUS) Specimina sumatrana a me nondum visa fide Backer (in litteris) citata sunt. B. Herba erecta, foliis lanceolato-ellipticis, ae scabridis, inflores- centia pedunculata, ovario capsulaque pilo 2. Pomazota Vanleeuweni Brem. n. spec.; TYPUS: Docters v. Leeuwen 3276 (BZ). Herba erecta, simplex, 5-7 cm. alta. Caulis densius pubescens, 2—2.5 mm. diam. Folia in petiolum dense pubescentem, 2—5 mm. longum con- aaa lamina lanceolato-elliptica, 4-6 cm. longa et 1.6-2.7 cm, lata, apice molliter pubescens, costa utrimque et nervis subtus dense pubescentibus, margine ciliolata, nervis utroque latere costae 10-13, venulis inconspicuis. Stipulae 8 mm. longae et 7 mm. latae, apice indivisae, extus dense pilosae, margine ciliatae. Inflorescentia distincte pedunculata: gg ele ae dense pubescens, post anthesin usque ad 3.5 cm. accrescens. Bracteae extus dense, intus sparse pubescentes, margine pilis similibus ciliatae; sres involucran tes 4, grees ee Aaa ovatae, 12 mm. longae et 11 m » BE praecedentibus dimidio breviores. Flores 5-meri. varium pilosum. Calyx lobis ovatis, 0.5 mm. longis et 0.7 mm. latis, subacutis, extus pilosis, margine pilis similibus ciliatis. Corolla alba dicta a a me non visa. Discus cylindricus, 1 mm. altus. Capsula pilosa, 2 mm. Hab. Sumatram Orientalem SuMATRA: East Coast Gouvt: wie, in the forest between stones on the bank of a stream, Docters v. Leeuwen 3276 (BZ, Typus, dupl. L.). b. Bracteae involucrantes 8, liberae. Cymae decussatae. . Pomazota simalurensis Brem. n. spec.; TYPUS: Achmad 516 (BZ). Herba e basi breviter repente ascendens, pseudo-dichotome ramificata, 15-40 cm. alta. Caulis primum densiu s pubescens, sania bisulcatis, deinde plus minusve glabrescens et sahies 2.5-4 mm. diam. Folia in petiolum pubescentem, usque ad 4 cm. longum hace: ‘petiolus foliorum superiorum tamen usque ad 8 mm. redactus; lamina lanceolata vel oblanceolata, 6-12 cm. longa et 2.5—4.5 cm. lata, apice acuminata, basi cuneata, supra sparse hirsuta, costa tamen densius pubescens, subtus costa nervis venulis pubescens, margine ciliata, sicc. supra saturate, subtus dilute brunnea, nervis utroque ie costae circ. 13, venulis paucis. Stipulae 14-16 mm. longae et 6— m. latae, apice plerumque in dentes duas parallelas, usque ad 4 mm Ae exeuntes, margine ciliatae, ceterum glabrae. Inflorescentiae pedunculatae; pedunculus dense hirsutus, ad anthesin circ. 1 cm. longus, postea usque ad 3 cm. accrescens; capitulum 198 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXVIII 2.5 cm. diam., post anthesin usque ad 3.5 cm. diam. accrescens. Bracteae involucrantes 8, quarum 4 majores triades breviter stipitatas suffulciunt; bracteae laterales triadum infimarum basi conduplicatae et torsae et inde in peg idem ut bractea triadem suffulciens expansae involucrum sup- p a involucrantes majores decussatae, flores singulos bibracteolatos sufful- cientes; bracteolae ovatae, ovario paulo longiores. Flores 5-meri. Ovarium pilosum. Calyx lobis ovato-lanceolatis, 1.1 mm. longis et 0.5 mm. latis, extus sparse pilosis, margine ciliatis. Corolla matura nondum visa, colore ignoto. Stamina prope basin tubi inserta. Discus conicus, 0.5 mm. altus. Capsula pilosa Hab. insulam a Sumatra ad occasum Simalur sive Simeulowe dictam. SIMALUR: = s.l., Achmad 516 (BZ, Typus, dupl. typi L). c. Bracteae involucrantes 8, basi connatae. Cymae quaternae. a. Bracteae cymarum primigenae basi utrimque lobo patente munitae. Calvcis lobi minus quam 3 mm. longi, ferrugineo-hirsuti. 4. Pomazota assimilis Brem. n. spec.; TypuS: Loerzing 5880 (U). Coptophyllum capitatum Miq. in errore apud Ridl. in Kew Bull. 1926: 68. 1926. Herba e basi breviter repente ascendens, 13-20 cm. alta. Caulis simplex, primum dense ferrugineo-hirsutus, internodiis bisulcatis, deinde glabrescens et teres, 2.5-4 mm. diam. Folia in petiolum primum dense, deinde sparse hirsutum, 0.5-3.0 cm. longum contracta; lamina lanceolata vel oblanceolata, 7-13 cm. longa et 3.2—4.3 cm. lata, apice acuminata, basi acuta vel cuneata, ee sparse hirsuta, subtus costa nervis venulis princi- palibus densius ferrugineo-hirsuta, margine ciliata, sicc. supra olivaceo- brunnea, subtus dilute brunnea, nervis arte latere costae 10-1 venulis paucis. Stipulae 10 mm. longae et 9 mm. latae, apice indivisae, margine ciliatae, ceterum glabrae. ar Rae te pedunculatae: peduncu- lus subglaber, post anthesin circ. 3 cm. longus. Bracteae involucrantes ’ 4—5 mm. latae, ad medium interdum plus minusve constrictae; minores 10-12 mm. longae sed majoribus subaequilatae ; omnes basi in receptacu- lum 7-10 mm. diam. connatae, apice subacutae, primum utrimque densius pubescentes. Flores omnes in dichasia densa dispositi; dichasia extus bracteis majoribus instructa; bracteae primigenae basi utroque latere lobo yatente munitae, 13 mm. longae, lobo mediano 4 mm. lato, lobis lateralibus 3 mm. longis et 3 mm. latis; bracteae sequentes sensim minores, plerumque sine lobis lateralibus; bracteolae florum ultimorum flori aequilongae. Flores S-meri, Ovarium dense pilosum. Calyx lobis lineari-oblongis, 2.7 am. 0.3-0.4 mm. latis, extus intusque pilis ferrugineis dense hirsutis. Corolla alba tubo 3 mm. longo, paulum inflato, 0.3 mm. infra orem annulo pilorum instructo, lobis 1 mm. longis. Stamina 0.7 mm. supra basin tubi inserta; filamenta 0.9 mm. longa; antherae 0.9 mm. longae. Discus annularis, 0.3 mm. altus. Stylus 3.5 mm. longus. Cap- sula 2 mm. diam., pilosa. Hab. Sumatram et insulas a Sumatra ad occasum., Sumatra: East Coast Gouvt: S.E. of Sibolangit, alt. 650 m., Loerzing 5880 (U, ere Lan Betimus, N.W. of Sibolangit, alt. 350 m., Loerzing 5318 (U); Bukit 1947 | BREMEKAMP, THE GENUS POMAZOTA 199 Pasang near Sibolangit, Mandur ta 7246 (K); Res. pore Padang si Dempuan, syleve 4854 (NY). Batu Istanps: P. Batu, Raap 655 (BZ); P. Masa, as 9 (BZ). MentTaweEt ISLANDS. ue Ibut 194 (BZ); Be Kloss 14517 (KK 8. Bracteae cymarum primigenae basi interdum uno sed numquam duobus lobis patentibus instructae. Calycis lobi plus quam 5 mm. longi, griseo-hirsuti. 5. Pomazota capitata (Miq.) Brem. n. comb. Coptophyllum? capitatum Migq., Fl. Ind. Bat. 2: 348. 1857, pie 216. 1860. Coptophyllum capitatum See in Ann. Mus. Bot. Lugd.-Bat. 4: 230. 1869; Boerl., Handl. Fl. Ned. Ind. 2 (1): 128. 1891; vix Ridl. in FI. el Pen. 2: 43. 1923, quae probabiliter est P. poulaers. Hend., non Ridl. in Kew Bull. 1925: 89. 1925, quae est P. rivularis Hend., nec op. cit. 68. 1926, quae est P. assimilis Brem. Herba e basi breviter repente ascendens, circ. 20 cm. alta. Caulis simplex, primum dense cae hirsutus, internodiis bisulcatis, deinde glabrescens et teres, 2.5—5 diam. Folia in petiolum primum dense, ae sparse hirsutum, 1.5- "3. "5 cm. longum contracta; lamina lanceolata, 10-16 cm. longa et 3. S— 5-2 CI, at, apice acuminata, basi acuta ve n se ferrugineo-pubescens, One ciliata, sicc. supra saturatius, che dilute brunnea, nervis utroque latere costae 11-13, venulis paucis. Sti- pulae 12 mm. longae et usque ad 15 mm. latae, apice indivisae, margine ciliatae, ceterum glabrae. Inflorescentiae pedunculatae; pedunculus sub- | glaber, ante anthesin 1.5 cm. longus, postea usque cm. accrescens Bracteae involucrantes 8, quarum 4 majores, omnes oblongae; majores 20 mm, long 7 atae; minores 15 mm. longae sed majoribus sub- aequilatae; omnes basi in receptaculum 10 mm, diam. connatae, apice subacutae, primum utrimque dense, deinde sparsius pubescentes. Flores omnes in dichasia densa dispositi; dichasia extus bracteis majoribus in- structa; bracteae primigenae basi interdum lobo singulo munitaé; bracteae sequentes floribus breviores; bracteolae ‘eke ultimorum minimae. Flores nt vel 6-meri. ee dense pilosu Calyx lobis linearibus, 5—5.5 longis et 0.6-0.7 mm. latis, aibauitnsis. seat intusque pilis griseis fee hivsutis Corolla Walre ignoto, tubo 3 mm. longo, 0.5 mm infra orem annulo pilorum instructo, lobis 1.5 mm, longis. Stamina 0.3 mm. supra ae tubi inserta, filamentis 1 mm., antheris 0.9 mm. longis. Discus conicus, 0.8 mm. altus. Stylus 3.0 mm. longus. Capsula 2.5 mm. ace pilosa. . Sumatram Occidentalem. a MATRA. Res. Tapanuli: mountain forest near Lumut, Junghuhn sn. (L, U, typi). d. Bracteae involucrantes 5. a. Indumentum luteo-brunneum, Bracteae omnes ovatae; interiores floribus breviores. 6. Pomazota sylvestris Rid]. in Trans. Linn. Soc. II. 3: 308, f. 61. 1893, Fl. Mal. Pen 923. Herba suberecta vel e basi breviter repente ascendens, 8-15 cm. alta. Caulis simplex, primum dense, postea sparsius pilis luteo- brunneis hirsutus, 2.5-4 mm. diam. Folia inferiora in petiolum dense hirsutum, usque ad 12 mm. longum contracta; superiora brevius petiolata vel subsessilia: lamina ie ancenc 6-11 cm. longa et 2—4.7 cm. lata, apice acuta 200 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII vel breviter acuminata, basi acuta vel cuneata, utrimque sed deneaigiyan margine, costa utrimque et nervis subtus pilis luteo-brunneis hirsuta, supra saturate et subtus dilute olivaceo-brunnea, nervis utroque en costae 11-13, venulis paucis. Stipulae 12 mm. longae et 10 mm. latae, apice ae extus hirsutae, margine ciliatae. cae oun peduncu- Jatae; pedunculus hirsutus, ad anthesin circ. 1 cm. longus, deinde usque ad 5.5 cm. accrescens. Bracteae involucrales 5, nero usque ad 15 mm. longae et 11 mm. latae, supra medium contractae, apice obtusae, sub- breviores. Flores in axillis bractearum involucr ae in cymas ‘bipa aras, in axillis bractearum aliarum in cymas uniparas bs aoa Flores 5-meri, basi bracteae inserti. Ovarium dense pilosum. Calyx extus dense pilosus, tubo 0.3 mm. longo, lobis ovatis, | mm. longis et 0.7 mm. latis, margine ciliatis. Corolla alba, tubo 3.5 mm. longo, 0.5 mm. infra orem annulo pilorum munito, lobis 1.2 mm. longis. Stamina 0.8 mm. supra basin tubi inserta, filamentis 1.7 mm., antheris 1.2 mm. longis. Discus cylindricus, 1 mm. altus, farinosus. Stylus 3.5 mm. longus. Capsula pilosa. Hab. Peninsulam Malayanam et Sumatram greets. eae Matay PENINSULA: Pahang: Kota Tongkat, Evans sn. (K, ex Fed. Mal. States Mus.), exemplum typi. SumatTrRA: East Coast Gouvt: Asahan, between H. Pa vrs bys Kopas, Bartlett 6759 ee near Rantau Parapat, Bila, Rahmat 2582, 2715, 3169 (NY); Tapanuli g si Dempuan, S. Manaun, Rahmat 4516 4 ) ; near . 76. anno 1813,’ which seemed to confuse the issue. Checking the reference, however, indicated an error on Bertolini’s part, for the Primula suaveolens entry is in Desvaux’s Jour. Bot. 2: 76. 1813, not in volume four, again confirmation that volume two was issued in 1813. As Fernald notes, the signatures (every 16 pages) are dated at the bottom of the first page of each beginning with p. 49, February, 1813, and continuing to June, 1813 (p. 273). It is not claimed that these monthly dates are actually the dates of issue, but this might well be the case. This system of dating the signatures was continued through volume 2, p. 1, July, 1813, to p. 257, December, 1813, but was abandoned in volumes three and four. Volumes three and four are dated 1814. On page 285 of volume four Desvaux says: ‘Les trois derniers Numéros du 3.° volume pour 1814, paraitront avec les premiers Numéros du 5." volume commengant I’ année 1815.” This explains, in part, the incomplete volume three but throws no light on the difficulties that the editor may have had with the printers. In all but one of the sets of this periodical that I have seen, volume three closes abruptly with page 192, and I judge that it was 250 JOURNAL OF THE ARNOLD ARBORETUM [ VoL. XXVIII Desvaux’s plan to issue four numbers of 48 pages each. It is worthy of note that the set in the library of the New York Botanical Garden contains pages 193 to 240 of the volume under discussion, this indicated as no. 5, May, 1814. But a footnote on page 197 clearly indicates that it could not have been published before February, 1816; see Hazen, Bull. Torr. Bot. Club 56: 531. 1929. The fifth volume promised by Desvaux for 1815 never appeared, and apparently also only certain fugitive issues of the last part or parts of volume three for 1814 were actually distributed. I know of only the New York copy of vol. 3, no. 5, never having seen a reference to an existing copy of no. 6. I have seen no fascicle covers of any of the numbers of this second run of Desvaux’s Journal de Botanique. Apparently the unsettled conditions in Paris associated with the fall of the first empire in 1814, and the first year or two of the restoration were not conducive to regularity in publication of technical periodicals. ARNOLD ARBORETUM, HARVARD UNIVERSITY. 1947 | MERRILL, ON THE IDENTITY OF BRANDISIA SOULIET 25] ON THE IDENTITY OF BRANDISIA SOULIEL BONATI E. D. MERRILL With one plate ON EXAMINING the proofs of Dr. Li’s paper on Brandisia, this Journal, p. 136 his correct elimination of Bonati’s species as a representative of this genus was noted. At the time Li’s paper was prepared it was impossible to secure special information from European centers. With the improve- ment in communications following the termination of hostilities it im- pressed me as desirable to see if the excluded Bonati species of Brandisia could be more definitely placed. Suspecting that the type was preserved in the Paris herbarium I appealed to Dr. H. Humbert of the Muséum (histoire naturelle, who courteously sent me an excellent photograph of the holotype. An examination of this rather clearly indicated that the family Scrophulariaceae was not involved but that probably Brandisia Souliei Bonati actually belonged in the Labiatae. Turning to Bonati’s description it is evident that he did not examine the structure of the ovary, for all that he says regarding the innermost set of floral organs is ‘‘stigmate bifido. lobis brevissimis. Capsula ac semina?”’ Feeling that it was essential that the structure of the ovary be checked I appealed to Dr. F. Gagnepain who reexamined the type and reported: “Je me hate de vous informer: Que cette espece est une Labiacée; 1° parce qu'elle a le stigmate particulier a cette famille; 2° parce que j’ai vu trés nettement au moins 2 nucules sur le podogyne au fond du calice. La chose est donc jugée; ce n’est pas une Scrophularicée. Je n’ai pas eu le temps de faire de plus amples recherches et de trouver le genre de Labiacées auquel appartient cette espéce litigieuse.”’ On the basis of the additional information courteously supplied by Dr. Gagnepain, the conclusion was reached that the genus Chelonopsis Miquel (Labiatae), as currently interpreted, was represented by Bonati’s species, and after a consideration of the published descriptions of the proposed species that further Brandisia Souliei Bonati (1909) was identical with Chelonopsis albiflora Pax & K. Hoffm. (1922). So convinced am I as to the specific identity of the two, one erroneously placed in the Scro- phulariaceae, the other correctly in the Labiatae, that on the basis of the original descriptions and without having seen specimens representing either of the species, the Pax and Hoffman one is reduced to synonymy and the following adjustment in nomenclature is made: 252 JOURNAL OF THE ARNOLD ARBORETUM [voL, XXVIII Chelonopsis Souliei (Bonati) comb. Brandisia Souliei Bonati, Bull. Soc, ‘Bot. France 56: 467. ic ai sia albiflora Pax & K. Hoffmann ex Limpricht in Repert. Sp. Nov. Beih. 12: The ee is apparently known only from the collections on which the two descriptions were based, and all three of the cited specimens came from Batang [Pa -an|, Province, Eastern Tibet |=Sikang Province, China], Soulie 5199, NBTEONE, collected in 1904 and Limpricht 2221, 2230 “Bejui-Batang” and ‘Batang-Litang,” collected in 1914, the last two numbers being the basis of Chelonopsis albiflora Pax & K. Hoffm. “Batang” is Pa-an, and what Soulie and — designated as eastern Tibet, at least as to the ‘“Batang”’ plants, is now Sikang Province, China. The published descriptions agree with each other closely. One of the striking characters even in CAelonopsis, is that most of the leaves are in whorls of threes. The three collections came from the same general region, but unless duplicates of the Limpricht collections were distributed to other centers previous to the destruction of the Berlin herbarium, perhaps the Bonati collection is the sole extant representative of the species. Naturally a botanist studying the Labiatae as did Dunn* in his consid- eration of the Labiatae of China cannot be censured for having overlooked a species erroneously described some years earlier and misplaced in the system not only as to the genus but also as to the family; Dunn recognized four species of Chelonopsis as occurring in China. Nor can Pax and Hoffman be blamed for overlooking an ample earlier description of a species that they described as new and for the same reason. This is an excellent illustration of the fact that to describe a new species is a very simple procedure, but to place a suspected new species in its proper = and to determine whether or not a proposed species is actually ‘new,’ (i.e., previously unnamed and undescribed) is another matter. EXPLANATION OF PLATE PLATE I Photograph of the holotype of Brandisia Souliei Bonati = Chelonopsis vial (Bonati) Merrill. Courtesy of Dr. H. Humbert, Muséum d’histoire naturelle, Pari ARNOLD ARBORETUM, Harvarp UNIVERSITY. * Dunn, S.T. A Key to the Labiatae of China. Notes Bot. Gard. Edinb. 6: 127-208. 1915. Jour. Arnoip Ars. VoL. XXVIII PLaTE I HERB. MUS. PARis CHELONOPSIS SOULIEL (BonaT1) MERRILL 1947] REHDER, CULTIVATED TREES AND SHRUBS, V 253 NOTES ON SOME CULTIVATED TREES AND SHRUBS, V ALFRED REHDER Chamaecyparis obtusa f. Sanderi (Sander), comb. nov. sista Hides Sanderi Sander ex Masters in Gard. Chron. ser. 3, 25: 287 (1889), nom. ubnud. — Beissner in Mitt. Deutsch. Dendr. Ges. 1899(8) ; 116 (1899), pro syn.— Unger in Mitt. Deutsch. Dendr. Ges. 1900(9):69 (1900), pro syn. — Anon, in Moller’s Deutsch. Giartn.-Zeit. 15: 589, fig. (1900), nom. subnud.; _p. 246, 428. Chamaecyparis obtusa ericoides hort. Jap. ex Boehmer, Cat. 1899-1900 (suppl.) : 2 (1899), nom. nud.— Beissner in Mitt. Deutsch. Dendr. Ges. 1991(10): 77 Dec. Girtn.-Zeit. 18: 291, fig. (1903, June 2C); Handb. Nadelh. ed. 2, 556, fig. 142 (1909). sa Hioenibrools Dwarf Conif. 41, fig. (1923) ‘‘var.’’ — Rehder in Bailey, Cult. Evergr. 216, fig. 41 (1923).— Non Retinispora obtusa var. ericoides Hoopes (1868). ssi Sanderi (Hort.) Sander in Gard. Chron. ser. 3, 33: 266, fig. 111 (1903) ; (Suppl.), fig. 107 (p. ii) (1903, April 25), nom. subnud. bee pisifera var. Sanderi Dallimore & Jackson, Handb. Conif. 219 (1923), nom. tenta Juniperus Sabine Ungeri Anon. in Gartenwelt, 33: 290, fig. (1929). Juniperus sabina Sanderi Anon. in op. cit. 291 (1929), pro syn. This juvenile form has been listed by most recent authors as Chamaecyparis obtusa ericoides Boehmer, a nomen nudum first validated in 1909 by Beissner (l.c.). This name, however, should be considered a later homonym of Retinispora obtusa var. ericoides Hoopes (1868), since Retinispora obtusa Sieb. & Zucc. and C. obtusa Endl. are synonymous. Retinispora obtusa var. ericoides Hoopes is based chiefly on Chamaecyp- aris ericoides Carr. (1855); there can, however, be no doubt that C. ericoides Carr. does not belong to C. obtusa, but represents a juvenile form of C. pisifera, namely C. pisifera f. squarrosa |Zucc.| Beiss. and partly C. thyoides f. ericoides (Carr.) Rehd. As the epithet ericoides has been applied to forms under three different species in the genus Chamaecyparis, and may therefore cause confusion, its rejection in favor of Sanderi, about which there can be no doubt as to the plant meant by it, is in accordance with the spirit of the Rules of Botanical Nomenclature (see Art. 4), even if Retinispora obtusa var. ericoides Hoopes and Chamaecyparis obtusa var. ericoides Beissner are not homonyms in the strict sense of the word; moreover, the first two figures of this plant were published under the names Juniperus Sanderi and Retinispora Sanderi. In the note in Gard. Chron. ser. 3, 25: 287 (1899) on Juniperus Sanderi, it is stated that J. Sanderi, a Japanese species, was introduced by F. Sander & Co., about 1896, but I have not been able to verify this statement. 254 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVHI Carya sect. I. Pacania (Raf.), comb. nov. Hicoria subgen. Pacania Rafinesque, Alsogr. pee - (1838). Hicoria subgen. praeaiig: srg sane lc. (1838), p. Carya sect. Il. Apocarya C. de Candolle in De Oat, Prodr. 16,2: 144 (1864). Hicoria sect. Apokicori ae Handb. Laubh. 2: 336 (1892). Hicoria [sect.]. Apocarya Sargent, Silva N. Am. 7: 135 (1895). The oldest subdivisional name for this section has been generally over- looked, but as it was validly published with a description and reference to the species belonging to it, it must replace the name Apocarya C. de Candolle. For the second group, I have retained the name Eucarya C. de Candolle, since its circumscription agrees exactly with that adopted here, while it seems doubtful which of the names of the three subgenera into which this group was split by Rafinesque should have preference. x Malus purpurea (Barbier) Rehder f. pendula (Bean) Rehder, comb. nov. - oe ae ” SS AS v2 - ips x } LD r y = a) = 33 . oS SS y a ‘ IN Ss ‘ os DG Re % SS JES, <== x ZAM NNN = ee MVE N BE B* wom Be Sanday wy ES SER ONG SOE N > 2" NM ae as nek > Saptaae aoa Bog: BRAN Sabey Sh i) i Pon neh ARS ’ 1} A qi; Ay) if FS if : ” Lue... 4 aT | 4 A 4 4 1 Aonnte J Fic, 1. Testa-cells of Acranthera longipes Merr. (a) and A. siliquosa Brem. (b). The cells consist almost entirely of the strongly thickened bottom-wall, which in (a) has the shape of a plano-concave, in (b) of a plano-convex lens; (a!) and (b!) show the deeper layers, in which the peripheral pit-canals converge towards the centre; (a) and (b*) a surface view of the same cells; in (b*) the pit-canals in the central boss are seen to diverge a little. the latter are, at least in the female flowers, always linear and spreading; the placentas are peltate, and the testa-cells provided with a few very large pits (Fig. 2). These differences are, of course, not all equally important. The value of the difference in the attachment of the placentas to the dissepiment, has already been discussed; it certainly justifies some doubt with regard to the propriety of leaving the genus Acranthera in this tribe. The peculiar 1947 | BREMEKAMP, ACRANTHERA 273 kind of floral mechanism is an even more weighty argument against the view that it should belong to this group, for in none of the general related either to Hedyotis L., Urophyllum Wall., Sabicea Aubl. or Mussaenda L., a style functioning as a temporary depository for the pollen is found. As I have pointed out in my “Monograph of the genus Pavetta L.” (in Fedde’s Repert. 38: 11. 1934) a style of this nature characterizes one of the main | ay eT 200M ; Testa-cells of Mussaenda frondosa L. (a) and M. cylindrocarpa Burck (b). The side-walls too are strongly thickened, but pits are confined to the bottom-wall. The size of the cells and the number of pits differ considerably in the two species, but the pits themselves vary but little. groups of the family, comprising the tribes Ixoreae, Gardenieae, Alberteae. Vanguerieae, the isolated genera Crossopteryx Fenzl and Coptosapelta Khs., and perhaps part of the present tribe Naucleae. In all these plants, however, the anthers wither after the pollen has been deposited on the style, whereas the receptaculum pollinis of Acranthera remains included between the empty anthers, and can be reached only through the windows between the projecting tips of the connectives. For this reason, and also because the stamens are not, as in the Ixoreae and their allies, inserted in the corolla-throat, but at the base of the tube, it seems to me that Acranthera can not be related to this group either. Moreover, the connec- tion of the stamens by means of the projecting tips of the connectives with the top of the style is a feature so entirely unparalleled in the Rubiaceae, that one might feel inclined to regard the position of the genus with regard to the rest of the family, as similar to that of the Asclepiadaceae with regard to the Apocynaceae, although in this way the importance of the connection between the anthers and the style, is doubtless over-emphasized. It is perhaps more readily comparable to the way in which in the genera Ceropegia L. and Dichaelia Harv. (Asclepiadaceae) the corolla-tips cohere. 274 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII There is one point more on which I should like to expatiate a little, although it does not concern a difference between Acranthera and Mussaenda, for in the point I mean, the absence of raphide-cells, there is complete agreement between the two genera. With regard to the presence or absence of raphide-cells there is no conformity between the various groups included in my enlarged tribe Hedyotideae, and as all other large groups are in this respect entirely consistent, this is doubtless a weak point. It might be argued that Hedyotts L. and its nearest allies, which are all provided with raphide- cells, are perhaps more closely related to the Spermacoceae and Psycho- trieae, which they resemble in the aestivation of the corolla-lobes, but from which they differ in the pluri-ovular ovary-cells, than to Sabicea Aubl., Urophyllum Wall. and Mussaenda L. and their respective allies, which they resemble, apart from the aestivation of the corolla-lobes, to some extent in the nature of the placentation, but in which raphides are apparently always wanting. A satisfactory solution of this question can not yet be given. For the determination of the position to be assigned to Acranthera, this is not of paramount importance, as the relations between this genus and the various groups in which raphides are wanting, are hardly more pronounced than those with Hedvyotis. THE SUBDIVISION OF THE GENUS Up to now no attempt has been made to arrange the species of Acranthera in natural groups. Valeton, who at one time (in Ic. Bog. 4: fasc. 4. 1914) described no less than nine new species, contented himself with an alphabetical arrangement! ‘This apparently means that he had not been able to distinguish well-defined groups, and was of opinion that the best way to make the species distinguishable, was to figure them, which he accordingly did with as many as he could lay hands on. The result is that of no other rubiaceous genus of any extent has such a large percentage of the species been figured (of the 20 species known in 1914 e.g. no less than 14, i.e. 70 per cent). These efforts, however, were, I am afraid, of little avail, for even with the aid of the pictures the species are by no means easily recognizable. It was to be expected, therefore, that a more thorough analysis would lead to better results. On account of the fact that not all species were available, and also because on some points, for instance on the habit and on the flower-colour, reliable information could not always be obtained, the results are not yet fully satisfactory; but I have no doubt that as soon as more material is forthcoming, the defects of my classifica- tion will easily be amende It is here perhaps the most suitable place to make a few remarks on the two points just mentioned, the habit and the flower-colour. A good deal of misapprehension has existed up to now with regard to the habit of these plants. On the labels they are not rarely described as shrubs and in one case even as a small tree (4. longipes Merr.), but these 1947 | BREMEKAMP, ACRANTHERA 215 indications should not be taken at their face value. It is a well-known fact that European collectors, accustomed to the soft shoots of the herbs of their native countries, often take tropical herbs for shrubs, even though there is no sign of branching, because they are misled by the woodiness of the stems, The indication “small tree” on the label of some of the specimens of A. longipes Merr., on the other hand, is doubtless a faulty translation of the “pohon ketjil” of the Malay collector. It is often overlooked that in Malay each plant provided with an erect stem is called “pnohon,” no matter whether it is an herb a few centimeters high and without a singe branch or a tree with a large crown. Stapf was doubtless right when he described the habit of the Acranthera species as herbaceous. Among all the specimens investigated by me, but very few were branched, and in these few exceptions the ramification was always of the pseudo-dichotomous kind. They belong to species in which the stems normally are sympodial. This means that occasionally at the base of the inflorescence instead of one bud both buds develop. Even those species which, like A. frutescens Val., reach a fairly considerable height (1.5—2 m.), apparently remain unbranched. It is not impossible that the larger plants require more time to complete their life-cycle than those that remain lower, but no reliable information with regard to this point seems to be available. In some species there are indications that the stems in the end sink down, and that innovations are produced from their basal parts. Our information with regard to the flower-colour too is far from complete. This is all the more unfortunate as it looks as if this might be of real taxonomic importance. The three western species are blue-flowered. The flower-colour is un- known to me, of the Assam A. tomentosa R.Br. ex Hook.f. but in the nearly related A. siamensis (Kerr) Brem. the flowers are white. The first of the species described from Borneo, A. atropella Stapf, was said to be provided with a dark blue corolla, but, as stated above, this is probably a mistake. The majority of the Bornean species and the only Philippine one possess white or slightly tinted flowers; in the rest of the Bornean species the corolla is yellow, orange or red. In these plants the calyx too is often coloured, not rarely in a different shade. With regard to the flower-colour of the Sumatran species we are insufficiently informed, but those of Brooks’s Benkulen species, the one that was identified by Ridley as Psilobium nutans Jack, are stated to be greenish white. The characters with which we will mainly have to be content, are: the shape and size of the stipules, the position and form of the inflorescences, the shape of the corolla-tube, the fusion or complete freedom of the basal parts of the stamens, the presence or absence of a fringe of cilia between the outer rim of the thecae and the connective, the shape of the latter, the presence or absence of a disk, the form of the receptaculum pollinis, the structure of the fruit and that of the testa. By the aid of these characters I have divided the genus in nine subgenera, of which the first, 276 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Eu-acranthera, is confined to Ceylon and the Indian Peninsula, the second, Androtro pis, to Assam and Peninsular Siam, the fourth, Amphoterosanthus, to Sumatra, and the others to Borneo, except the eighth, Mitracme, which contains, besides some Bornean species, the only representative of the genus found in the Philippines. Eu-acranthera and Androtropis differ from all the other subgenera by the presence of a well-developed disk. In Eu-acranthera, moreover, the corolla is blue and its tube begins with a cylindrical part, which widens in the upper half to funnel-shape, whereas in the other subgenera the corolla is presumably never blue, and its tube is either entirely funnel-shaped or, more often, narrowly campanulate. Androtropis differs from the other subgenera in the presence of a staminal tube formed by the basal halves of the filaments; in all other subgenera the filaments are entirely free. The third subgenus, Cleomocarpus, differs from the other subgenera in the structure of the fruits, which are narrowly cylindrical and marginate, the hardened margin remaining after the seeds have been shed, in the form of a “replum.” The latter, however, is morphologically not equivalent to the replum of the herbaceous Capparidaceae, the Cruciferae and Cheli- donium, for in these plants the rim is formed by the fused margins of the carpels, whereas in the subgenus Cleomocarpus of the genus Acranthera it represents the midribs of the carpels. The dissepiment, accordingly, is not, as in the Cruciferae, attached to the rim, but stands at right angles with it; in the end the dissepiment disappears with the valves. The stipules of Cleomocarpus are smaller than those found in any other sub- genus; in shape, however, they are not unlike those of the subgenera Eu-acranthera, Androtropis and Amphoterosanthus. Its two species are nearly glabrous plants with densely reticulated leaves. The fourth subgenus, Amphoterosanthus, is confined to Sumatra and the neighbouring island Simalur. It resembles the three preceding sub- genera in the triangular shape of the stipules, but differs from them in the position of the inflorescences, which are borne on opposite brachyblasts provided with a pair of rudimentary leaves. The inflorescences themselves are few-flowered. In position and structure they are similar to those found in some species belonging to the subgenus Dichroanthes and to those of Ablepharidesma. The seeds are reticulate, whereas in almost all other species of which ripe fruits were available, the seeds were found to be either distinctly carunculate or nearly smooth. There is, however, no difference of fundamental importance between the various kinds of seed. The two species of Psilobium, described by Jack, and Gonyanera glauca Khs. belong probably to this subgenus. As no specimens are extant, and as the plants are not identifiable from the descriptions, it did not seem advisable to retain either of these names for the denomination of the subgenus. In the other subgenera the stipules are wider and longer than in the preceding ones, and of an entirely different shape, namely ovate, elliptic or oblong. The differences between these subgenera are not so striking as to “a ~) 1947] BREMEKAMP, ACRANTHERA those between the former. They are mainly confined to the position and structure of the inflorescence, the shape of the receptaculum pollinis and the presence or absence of a fringe of cilia between the thecae and the connective. Athroophleps, however, is well characterized by the peculiar nature of the reticulation. It is noteworthy that with regard to the shape of the receptaculum pollinis and to the presence or absence of a fringe of cilia along the connective, the first four subgenera, Eu-acranthera. Androtropis, Cleomocarpus and Amphoterosanthus, show a uniform char- acter: their anthers are always eciliolate, and the receptaculum pollinis is everywhere fusiform, and of about the same length as the thecae. In the three subgenera Phanerochiton, Dichroanthes and Ablepharidesma too the receptaculum pollinis is of about the same length as the thecae. and either fusiform or cylindrical. In Phanerochiton and Dichroanthes the anthers are ciliolate, in Ablepharidesma eciliolate. The monotypic subgenus Phanerochiton differs from the two others in the long, scarious stipules, the trichotomous inflorescence with its large bracts and the thick-walled fruit, and from all other representatives of the genus in the presence of dark-coloured resin-cells on the upper side of the leaves. In Dichroanthes the filaments are about as long as the anthers, the recepta- culum pollinis cylindrical, and the flowers often, perhaps even always. gaily coloured, the corolla yellow, orange or red, and the calyx white or in 4 different shade of orange or red. In the other Bornean species the filaments are always much shorter than the anthers, and the corolla is apparently everywhere either white or but slightly tinted. Dichroanthes is divided in two series, one with terminal inflorescences and cohering stigmata, the other with inflorescences borne on opposite brachyblasts and with free stigmata. In Ablepharidesma the inflorescences are few-flowered and borne on opposite brachyblasts, the anthers eciliolate, the shoots thin, and the leaves and stipules small. In the subgenus Mitracme the receptaculum pollinis is much shorter than the thecae and mitriform, i.e. the papillae are reclinate and increase from the top towards the base of the receptaculum gradually in length. In this subgenus I distinguish four series. In the first series the anthers are ciliolate, whereas in the other three they are always eciliolate; the species belonging to this series are robust plants with large trichotomous inflorescences provided with large bracts. The second series is monotypic, the only species being the Philippine 4. philippinensis Merr.; it is a low plant with a trichotomous inflorescence provided with well-developed bracts. The third series is also monotypic, the only species being a narrow-leaved plant with few-flowered inflorescence and small bracts. The fourth series comprises plants with umbelliform inflorescence, in habit not unlike some of the species belonging to the subgenus Dichroanthes. The last subgenus, Athroophleps, is a small but very natural group. confined to North Borneo and easily recognizable by the peculiar arrange- ment of the thick-set prominulous venules, which form narrow meshes, 278 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII stretched in a transverse direction. In A. capitata Val. this curious ae was noticed already by Valeton (cf. Ic. Bog. 4: 276 (line 6), 391, fig. 12. 1914). In none of the other subgenera are the meshes sels: in this direction, and except in Cleomocarpus, where the reticula- tion, however, is very faint, their number is always much smaller. Other peculiarities of this subgenus are the annular shape of the receptaculum pollinis and the great length of the points in which the connectives are drawn out. DESCRIPTIO GENERIS* — Arn. ex Meisn., Pl. Vase. Gen, 1: 162, 2: 115. 1838; Arn. in Ann Hist. 3: 20. 1839; Endl., Gen. Pl. 1394. 1839; non Kurz in Jour. As. Soc. Se a Peto 1872, nam specimen citatum ad Aphaenandram Mic, pertinet ; Hook.f. in Benth. & wenn Gen. Pl. 2: 64. 1873; Bedd., Ic. Pl. Ind. Or. 1:5 bal t. 23-25. 1874; Hook.f., Brit. Ind. 3:92. 1880, A. Griffithit Hook.f. et A. Maingoyi barges quae genus Asemanthiam Ridl, ducendae sunt exclusis; non Hemsl. in Jour. Bot. 25: 204. 1887, nam specimen citatum ad Asemanthiam Maiogayi tes Rid]. pertinet; K. Sch. in Engl. & Prantl, Nat. Pflanzenfam. I\ 1891; Trimen, Handb. Fl. Ceylon 2: 324. 1894; Stapf in Trans. Linn. Soc. se II, 4: 173. kG Val. in Bot. Jahrb. 44: 550. 1910, 48: 111. 1912; Merr. in Philip. Jour. Sci. Bot. 8:32. 1913; Val. in Ic. Bog. 4 (3): 181, t. 355, 356. 1913, 4 (4): 275, t. 391-398. 1914, A. strigosa Val. p. 391, et t. 399, excl.; Gamble, Fl. Pres. Madras 2: 611. 1921; Lemée, Dict. Pl. Phan. 1:45. 1929; Merr. in Papers Mich. Acad. Sci. 19: 194. 1934 et in Mitt. Inst. Allg. Bot. 937. Mussaenda spec., Baill., Hist. Pl. 7: 319, 449. 1880. geotognd — Mal. Misc. 2: 84. 1822 (etiam in Calc. Jour, Nat. Hist. 4: 27. 1843) ; . 4: 618. 1830; Roxb., Fl. Ind. ed. Wall. 2: 320. 1832; Miq., Fl. Ind. Bat. 2: 199. 1857; Hook.f. in Benth. & Hook.f., Gen. Pl. 2: 75. 1873; Boerl., Handl. Fl. Ned. Ind. 2: 64. 1891; K. Sch. in Bnei. & Prantl, Nat. Pflanzenfam. IV. 4: 70. 1891; Lemée, Dict. Pl. Phan. 5: 629. 1934, nomen vix usum et typo perdito haud facile veeeprrlaenty igitur melius rejectandum; non K . As. Soc. Bengal 41 (2) . 1872, nam specimen citatum ad genus ee Hook.f. pertinet ; Ridl. in ae Bull. 1925: 84. 1925; Kerr in Hook., Pl. t 3332. 1937, nomine generico recte usi sunt Gonyanera Khs. in Ned. Kruidk. Arch, 2 (2): 183. age aes Fl. Ind, Bat. 2: 200. 1857; Hook.f. in Benth & Hook.f., Gen. Pl. 2:75. ; Boerl., raphe Fl. Ned. Ind. 2: 64. hata K. Sch. in Engl. & Prantl, Nat. a IV. 4: 70. 1891; Lemée, Dict. Pl. Phan. 3: 316. 1932, nomen typo perdito haud facile ore en et igitur aa Herbae plerumque simplices, rarius pseudo-dichotome ramificatae, sub- rosulares, ascendentes vel suberectae, basi lignescentes. Caulis plerumque sympodialis, obtuse quadrangularis, internodiis bisulcatis. Folia opposita et aequalia, petiolata; lamina plerumque oblanceolata vel obovata, apice acuminata, margine anguste membranacea ciliata, substrigosa vel strigosa, facie ventrali sub lente albido-granulata, costa nervisque subtus plerumque pilosis, raphidibus et acaridomatiis nullis. Stipulae interpetiolares, sim- plices, extus haud raro colletris subulatis sparsae, intus glabrae. In- * The abbreviations for the herbaria of the institutions cited in this paper are as follows: AA=Arnold Arboretum; BD=—Bot. Mus., Berlin-Dahlem; BZ—Buitenzorg Botanic Garden; HGB=Inst. Allg. Bot., Hamburg; L=Rijksherbarium, Leiden; U=Bot. Mus., Utrecht. 1947] BREMEKAMP, ACRANTHERA 279 blastos) oppositos, foliis duobus rudimentariis instructos terminantes; flores paniculati, cymosi vel cymoso-umbellati, rarius solitarii, pedicellati vel subsessiles; bracteae variae. Flores hermaphr roditi, plerumque 5-, raro 4-meri, aliqui interdum 6-meri. Ovarium plerumque elongatum, rarius turbinatum, biloculare, dissepimento plerumque tenui_ et faciliter dis- cindente; placentae utroque loculo duae, prope axem orientes et ei paral- lelae, rarius confluentes, per totam longitudinem dissepimento affixae; ovula numerosissima. Calyx usque ad basin partitus, corollae tubo plerumque subaequilongus; lobi haud raro paulum inaequilati; glandulae interlobulares plerumque conspicuae. Corolla colore variabili, extus plerumque pilosa, intus semper glaberrima; tubus nunc e basi cylindrica infundibuliformis, nunc totus infundibuliformis vel anguste campanulatus; lobi patentes, aestivatione reduplicato- -valvata. Stamina basi corollae in- serta, tubo plerumque inclusa, rarius breviter exserta; filamenta glabra vel rarius papillosa, plerumque libera, raro usque ad medium connata; antherae lineares, conjuncte stylum includentes, introrsae, thecis facie ventrali tas ee connectivo faciem dorsalem totam complente, haud raro gibboso vel carinato, margine interdum ciliolato, apice in apiculam vel aristam apici eer incumbentem et cum eo cohaerentem producto. Discus nunc bus aequilongus, parte superiore papillis plerumque 10-seriatim dispositis obtecta pro receptaculo pollinis agente, apice in stigmata subulata vel semi-conica parallela vel confluentes exeuns. Fructus ovoideus, turbinatus Distributum speciebus adhue certe notis 35 in umbra nemorum Cey- laniae, Peninsulae Indicae, Assamiae, Siamiae Peninsularis, Sumatrae, terrae Borneénsis, insulae Filippinae Mindanao dictae. Species typica: A. ceylanica Arn. ex Meisn. KEY TO THE SUBGENERA Disk shortly cylindrical or semi-globose. Inflorescences always at the end of the m.— Species from Ceylon, the Western Ghats, Assam and Peninsular Siam. Corolla blue; basal part of the tube cylindrical, upper part infundibuliform. Stipules triangular, at the most 1.5 cm. en usually much short Stipules not more than 3 mm. long. Inflorescences at the pe a the stem. Fruits bicostate; the ribs after the shedding of the seeds agit behind in the form of a replum. Seeds carunculate.— Bornean species...........-.-- ee ee ser ca Ain “dra eee bite orsay si Hienahe dh ace Spar tope aeslgh a mug Cleomocar pus an 7- _ mm. long. Inflorescences at the end of oenecne brachyblasts rudimentary leaves, few-flowered. Fruits S- or 6-costate. on ae — Species from Sumatra and Simalur...D. Amphoterosanthus 280 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXVIII Stipules ovate, elliptic or oblong, usually more than 1.5 cm. long, if somewhat shorter (e.g. subgenus Ablepharidesma), then at least with the greatest width near the dle and several times longer than the diameter of the stem Receptaculum pollinis as sy as or but slightly shorter than the thecae; and sr Sceaisnaiyge or fusiforn nective ciliolate. se eaves on the upper side dotted with dark-coloured resin-cells. ee either trichotomous with fairly long primary branchlets or, by the suppression of the lateral branchlets, long-pedunculate. Bracts large. Corolla reddish.—A_ single Bornean ee Bip cusceveses si areca Sa Gee G a PoP aw Saree waiara Oras ton pie cad wipe 4a Gua ee lea ee Phanerochiton Inflorescences either umbelliform or reduced to a . flower, always subsessile. Bracts small. Corolla yellow, orange or red. — Bornean pecie ichroanthes Connective eciliolate. Inflorescences borne on opposite acta pro- — with rudimentary leaves, few-flowered. Corolla white. — nean SDOCIOSsa-3 six o-d aaa hia. vb Bina BA aca wd Spe die dial wd waa raw ae Able deseo Roan pollinis much shorter ery aan either mitriform or an nular. Inflorescences always at the end of the stem. Retic ulation lax; the smaller oat a or but slightly prominent. Receptaculum pollinis mitriform.— Species from Borneo and Mindanao. Siete bee Sere MAW ALR ONE GS 64,8 ASE Does ee wae Hales ERS H. Mitracme Meshes of the reticulation small and stretched in a transverse direction; the venules themselves prominulous. Receptaculum pollinis annular. — Confined to North Borneo.......... 0.0.00. c cee ee eee I. Athroophle ps Suspcenus A. EU-ACRANTHERA BreM. SUBGEN. NOV. Herbae humiles, caule petiolisque pilosis. Folia laxe reticulata. Stipulae ovato-triangulares, internodiis breviores. Inflorescentia termina- lis, trichotome corymbosa. Corolla coerulea vel hey tubo e basi cvlindrica infundibuliformi, quam calyce multo longiore. Stamina fila- mentis papillosis, liberis, quam antheris longioribus, ee eciliolatis. Discus breviter cylindricus. Receptaculum pollinis fusiforme. Fructus ovoideus. Semina carunculata ye tres, Ceylaniam et Peninsulam Indicam habitantes, ubi in altitude 900-1500 m. crescunt. Subgeneris tvpus: 4. ceylanica Arn. ex Me KEY TO THE SPECIES Calyx-lobes obovate, spreading; anthers shortly apiculate. — Indian Peninsula....... 0: 58 gf 8 pK ws § yh Aa BAR BCs ARORA 8G, Gick natcb ie Wh TB oe ae. ws AGG! Raglan ate Wane a Ne A, hasocesiinic Calyx-lobes linear-lanceolate or narrowly triangular, erect; anthers arist: Leaves dese ia us on the upper side; calyx-lobes linear-lanceolate; calle tube 2. oh mg 4 cm. long. — Indian Peninsula............. 0.0.0. ce eee eee ee eee . grandiflora i. asia ate 2 ve upper side; calyx-lobes narrowly ne corolla- tube 2 cm. long. — Ceylon....... 0.0... 0c cece cece cece eee een 3. A. ceylanica i. — anamallica Bedd., Ic. Pl. Ind. Or. 1: 5, t. 23. 1874; Hook.f., Fl. Brit. Ind. 1880; Gamble, Fl. Pres. Madras 2: 611. Pet Indian alae (Wes- ees ras s 2. Acranthera grandiflora Bedd., Ic. Pl. Ind. Or. 1:5, t. 25. 1874; Hook.f., Fl. Brit. Ind. 3:93. 1880; Gamble, iy Pres. Madras 2: ‘611. 1921. +- Indian Peninsula (Western Ghats). 1947 | BREMEKAMP. ACRANTHERA 281 ex Meisn., Pl. Vasc. Gen. 2: 115. 1838. Acranthera zeylanica Arn. in Ann, Nat. Hist. 3: 21. 1839; Walpers, Repert. 6: 1846: Thwaites, Enum. Pl. Zeyl. 138. 1860; Bedd., Ic. Pl. Ind. Or. 1:5, t. 1874: Hook.f., Fl. Brit. Ind. 3:92. 1880, Trimen, Handb. Fl. of Ceylon 2:3 1894. — Ceylon As no material of 4. anamallica and of A. grandiflora was available to me, the key was based on data gathered from the literature. ~s h - ° 3 = = =} = — > a") = <<) ° o Me is =] = ye ~ o - + ~ =) SusceNus B. ANDROTROPIS (R.Br. in Wall. Cat. n. 8398, gen. inedit.) Herbae altiores. Folia laxe reticulata. Stipulae ovato-triangulares, internodiis multo breviores. Inflorescentia terminalis, corymbosa. Cor- olla ubi color notus viridula, tubo infundibuliformi, quam calyce longiore. Stamina filamentis glabris, usque ad medium connatis, quam antheris longioribus, antheris eciliolatis. Discus breviter cylindricus vel semiglo- bosus. Receptaculum pollinis fusiforme. Fructus cylindricus, 5-costatus. Semina carunculata. Species duae, Assamiam et Siamiam Peninsularem habitantes. Sub- generis typus: 4. tomentosa R.Br. ex Hook.f. KEY TO THE SPECIES re ee a 0S 1 Ux cers ee Cw Pp LS os tomentosa Leaves glabrous on the upper side or near the margin and towards the base strigose. Corolla about 2 cm. long. — Peninsular i ee ier ore rene sits siamensis 4. Acranthera tomentosa R.Br. ex Hook.f., Fl. Brit. Ind. 3:92. 1880 (Androtro pis tomentosa R.Br. in Wall. Cat. n. 8398, nomen).— Assam 3. Acranthera siamensis (Kerr) Brem. n. comb. Psilobium siamense Kerr in Hook., Ic. Pl. t. 3332, 1937. — Peninsular Siam. Of A. tomentosa | had but a single, not very good, specimen, whose flowers were much larger than those of the specimens described by Hooker. Fruits of this species were not available to me. For A. siamensis I relied on the detailed description given by Kerr and on the excellent plate by which the latter is accompanied. Kerr referred this species to Psilobium Jack, because it resembles the specimen collected by Brooks in Benkulen, which Ridley (in Kew Bull. 1925: 84. 1925) had identified as Ps. nutans Jack, but which is probably a different species (see above). The near affinity between the Siamese plant and A. tomen- ) Kerr ends his description with the following remark: ‘It seems probable, judging from a cursory examination, that some plants assigned to Gardenia, Section Gardeniella, should rather be referred to Psilobium.” The section Gardeniella was created by Ridley (FI. Mal. Pen. 2: 80. 1923), who referred to it four species from the Malay Peninsula, which show but little resemblance to the typical representatives of the genus Gardenia. The possibility that they might belong to Acranthera, of which representatives are found both to the North and to the South of the Malay Peninsula, certainly deserves our attention. Judging from the somewhat 282 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXVIII meagre descriptions, I do not believe however that anyone of them can be transferred to this genus. The first, G. tentaculata Hook.f. is described by Ridley as a “bush,” a habit which is entirely unknown in Acranthera; its stipules, apparently, are similar to those of A. strigosa Val., a species which in my opinion does not belong to this genus. The three iia ones seem to be unbranched, but they differ from all species of Acranthera known so far in the position of the inflorescences: the latter, namely, are found on the defoliated part of the stem. In two of them, G. virescens Ridl. and G. pulchella Ridl., the stipules, moreover, are like those of G. tentaculata ‘“‘fringed with long points.” The last one, G. didymocarpus Ridl., had provisionally been referred by H. H. W. Pearson to Acranthera. Its removal by Ridley to Gardenia evidently means that it does not fit the description of Acranthera. It is possible, however, that Ridley was in- fluenced by the lateral position of the inflorescences. Suscenus C. CLEOMOCARPUS Brem. suBGEN. Nov. Herbae ascendentes vel suberectae, pr heware! Folia dense sed vix conspicue reticulata. Stipulae breviter et late triangulares, in bea brevem connatae, apiculatae, usque ad : mm. altae, persistentes. In- florescentia terminalis. Corolla alba, extus glabra, tubo aeciiae. Stamina filamentis glabris et liberis, quam antheris multo sigh tee antheris connectivo carinato, eciliolato. Discus inconspicuu ecep- taculum pollinis fusiforme. Fructus cylindricus, sitedalnioy sin costis — e repli instar remanentibus. Semina caruncula s duae terrae Borneénsis partes i aniisaaiane et orientalem fastens Subgeneris typus: A. siliquosa Brem. n. spec. v. infra. KEY TO THE SPECIES Leaves with 10-12 pairs of nerves. Inflorescence laxly paniculate, many-flowered. — East Borneo... joc ieececsy cee ince abo ac es be ees pene ene sncneaesescOe J . siliquosa Leaves with 6-8 pairs of nerves. Inflorescence cymose, 2- to 7-flowered. — or BOVNG0 53.550 obi 5 a sy tes bain oe ae Oe eas Veh eae Ae ee nS 7. A. ophiorrhizoides 6. Acranthera siliquosa Brem. n. spec.; typus: Endert 3368 (BZ). Herba suberecta, usque ad 2.5 m. alta. Caulis sympodialis, diametro ad apicem 2 mm. basin versus usque ad 5 mm. aucto, glaberrimus, sicc. fuscescens, internodiis 2—7.5 cm. longis. Folia petiolo 2—6 cm. longo, glaberrimo instr ucta; lamina oblonga vel obovata, 12-24 cm. longa et gine pilis brevibus sparse et Vix sagen tein strigosa, ceterum je) & Sal bo Ss : 3 = a) 5 ve 2 x O-S — =] 5 ® } =. = Series D,. UMBELLIFLORAE erbae ascendentes vel subrosulares. Inflorescentia umbelliformis; ea parvae. Connectivum eciliolatum. — Species 30-32. 30. Aeranthera Nieuwenhuisii Val. in sched.; typus: Jaheri (Exped. Nieuwwenhuis) 754 (L). erba ascendens, circ. 15 cm. alta. Caulis primum breviter sed densis- sime, deinde sparsius cue neve diametro ad apicem 1.7 mm., basin versus usque ad 2.5 mm. aucto, internodiis 8-12 mm. longis. Folia petiolo 7-10 mm. longo, ioe sed densissime fulvo- -strigoso instructa; lamina lanceolata, 7-11 cm. longa et 2—3.3 cm. lata, apice acuta vel acuminata, basi acuta vel subacuta, mar oe dense strigosa , supra costa quae basin versus est dense fulvo- strigosa excepta glaberrima, apo costa ne 1947] BREMEKAMP, ACRANTHERA 303 densissime, venulis principalibus sparsius strigosa, utrimque opaca, sicc. supra saturate brunnea, subtus fusca, nervis utroque latere costae 5 vel 6. Stipulae oblongae, 1.7 cm. longae et 5.5 mm. latae, apice acutae, extus pilis brevibus sparsae, margine ciliatae, sicc. pte costa praesertim circ. 6-flora, basi bracteis linearibus 5 mm. longis et 0.5 mm. latis instructa. Flores pedicello circ. 10 mm. longo, breviter sed densissime fulvo-strigoso elati. Ovarium cylindricum, 8 mm. longum et 1 mm. diam., ut pedicellus densissime fulvo-strigosum. Calycis lobi lineares, 11 mm. longi et t 1.5-1.8 mm. lati, caudato-acuminati, extus intusque pilis brevibus sparse, margine dense strigosi, 3- vel 5-nervii. Corolla colore ignoto, ae cm. longa, extus breviter strigosa; tubus 14 mm. longus et 4.5 mm. diam.; lobi ovati, 5-5 mm. longi et 3.5 mm. lati, acuti. Stamina 9 mm. longa; pe 3 mm antherae 6 mm. longae, connectivo in appendicem anguste triangularem, 1.3 mm. longam exeunte. Stylus basi aa incrassatus; a eae pollinis 2.5 mm. longum et sesh 0.8 mm. diam.; stigmata subulata 1.5 mm. longa. Fructus nondum n Habitat terrae Beniensis partem centralem. CENTRAL Borneo: - s.l., Jaheri (Exped. Nieuwenhuis) 754 (L, typus). Acranthera Nieuwenhuisii comes very near to A. Hallierti Val., from which it differs in the greater length of the internodes, the lanceolate, entirely green, less conspicuously ciliated leaves, the oblong stipules, the linear calyx-lobes and the greater length of the points in which the con- nectives are drawn out. Both species are in general aspect not unlike some species belonging to the subgenus Dichroanthes, namely A. variegata Merr. and A. aurantiaca Val. ex Brem., from which they are, of course easily distinguishable by the eciliolate anthers and the mitriform recep- taculum pollinis, and probably also by the colour of the corolla. It is true that the colour of the corolla in A. Nieuwenhuisii is unknown, but as that of A. Hallierti is said to be white, there is good reason to assume that the flowers of A. Nieuwwenhuisti too will be white 31. Acranthera Hallierii Val. in Ic. Bog. 4: 183, t. 356. 1913. Herba subrosularis. Caulis ad apicem pilis satis longis strigosus, basin versus glabrescens, diametro ad apicem 2 mm., basin versus usque ad 3.5 mm. aucto, internodiis circ. 5 mm. longis. Folia petiolo 0.5—2.5 cm. longo, primum dense, einde sparsius strigoso instructa; lamina elliptico- lanceo- lata vel obovata, 6.5-11 cm. longa et 2.5—4.5 cm. lata, apice acuta vel subacuminata, basi acuta vel subtruncata et prope petiolum subito con- densissime, venulis sparsius et brevius strigosa, utrimque opaca, variegata, sicc. supra saturate brunnea, subtus fusca, nervis utroque latere costae 6 vel 7. sinilee ovatae, 13 mm. longae et 6 mm. latae, obtusae, extus sparse strigosae, costa prominula, deciduae. Inflorescentia subsessilis, e floribus 2—6 composita; bracteae angustae, vix conspicuae. Flores pedi- cello usque ad 1.5 cm. longo, pilis brevibus dense griseo-strigoso elati. Ovarium cylindricum, 5 mm. longum et | mm. diam., ut pedicellus dense griseo-strigosum. Calycis lobi lanceolati, 8.5 mm. longi et 1.7—2 mm. lati, 304 JOURNAL OF THE ARNOLD ARBORETUM (VoL. XXVHI acuti, extus sparse, costa et margine densius strigosi. Corolla alba, 17 mm. longa, extus pilis brevibus dense strigosa; tubus 11 mm. longus et 5 mm. diam.: lobi ovato-triangulares, 6-7 mm. longi et 3 mm. lati. Stamina 10 mm. longa; filamenta 3 mm.; antherae 7 mm. longae, connectivo in appendicem vix 0.5 mm. longam producto. Receptaculum pollinis 2 mm. longum et basi 0.8 mm. diam.; stigmata subulata 0.5 mm. longa. Fructus nondum notus Habitat terrae Borneénsis partem centralem. Centra Borneo: s.l., Amdjah (Exped. Nieuwenhuis) 310 (L, exemplum typi). The figure and description given by Valeton of the receptaculum pollinis are wrong, and the white patches of the leaves are neither mentioned nor shown. 32. Acranthera atropella Stapf in Trans. Linn. Soc. Bot. II, 4: 173. 1894. — British Yorth : Specimens of this species were collected by Low & Haviland on the slopes of Mt. Kinabalu at an altitude of 1500-1800 m, It reaches, there- fore, a greater height than any of its allies. Its position in the subgenus Mitracme is not yet fully assured, for Stapf describes the receptaculum pollinis, which he too confuses with the stigma, as capitate, but as his description in other respects closely resembles those given above of A. Nieuwenhuisii and A. Hallierii, 1 have little doubt that its receptaculum pollinis is in reality mitriform. The corolla has been described as ‘obscure cyaneum,” but this may be the colour it assumes in the herbarium: the flowers of the other Bornean species are, at least in the living state, never blue. The similarity between this species and A. Hallierii has been noticed already by Valeton, who, however, gave an erroneous interpretation of the structure of their inflorescences (cf. Ic. Bog. 4: 182. 1913). He thought that the flowers were in reality axillary, the umbellate arrangement being simulated by the shortness of the upper internodes, but in A. Hallierii the flowers are by no means subtended by ordinary leaves, but by minute bracts, and the inflorescence is doubtless cymoso-umbellate. Stapf de- scribes the flowers of A. atropella as ‘‘pseudo-umbellate,” but this means probably that the flowers do not open centripetally, as in a true umbel, but centrifugally. As he says that the inflorescence is sessile and at the base surrounded by the two upper pairs of leaves with their stipules, the possibility that the flowers might be axillary, seems excluded. True um- bels are probably unknown in the whole family; when the flowers are fascicled at the top of a common peduncle, the arrangement always seems to be entirely or partly cymose. Stapf mentions a minute white punctation of the upper side of the leaves caused by collapsed cells. This punctation has also been observed in A. Nieuwenhuisii and in the three species belonging to the next subgenus. SusceNnus I. ATHROOPHLEPS BreM. SUBGEN. NOV. Herbae ascendentes vel suberectae. Folia supra sub lente cellulis collapsis albo-punctata, venulis ad nervos plus minusve perpendicularibus 1947] BREMEKAMP, ACRANTHERA 305 et valde approximatis utrimque transverse striata. Stipulae oblongae vel obovatae, obtusae, internodiis superioribus subaequilongae, dense reticu- latae, plerumque deciduae. selene terminalis. Flores sessiles vel subsessiles. Ovarium turbinatum. Corolla alba, tubo anguste campanu- ato. Stamina filamentis amie et liberis, quam antheris multo cate bus, antheris eciliolatis, connectivo in appendicem filiformem produc Discus inconspicuu s. Stylus cylindricus; receptaculum pollinis ates. Fructus cylindricus. Species adhuc notae tres terrae Borneénsis partem septentrionalem habitantes. Subgeneris typus: A. athroophlebia Brem. n. spec. v. infra. KEY TO THE SPECIES Flowers laxly paniculate; bracts deciduous. Leaves, calyx and corolla, apart fro the ciliate margin, entirely glabrous. — North Borneo......... SosnA® etiroophbi Flowers subcapitate; bracts persistent. Leaves on the underside, at least h nerves, as well as calyx-lobes and outside of the corolla, pubescen Leaves with 6-10 pairs of nerves; midrib and nerves hirto- pubescent on the underside. anes early deciduous. — North Borneo............ : capitata Leaves wit —5 pairs of nerves; rae and nerves strigose on the uae Stipules ae cae — North Borneo.........................8 . Ruttenii 33. Acranthera athroophlebia Brem. n. spec.; typus: J. & M.S. Clemens 30616 (L). Herba suberecta, usque ad 90 cm. alta. Caulis sulcis interdum primum pulverulentus, mox totus glabrescens, diametro ad apicem 5 mm., basin versus usque ad 10 mm. aucto, internodiis superioribus ee 7 cm. longis, sicc. nigrescens. Folia petiolo 4—5 cm. longo, ale sicc. nigrescente instructa; lamina oblanceolata vel sagan 16-29 cm. longa et 4.7—10 cm. lata, apice acuta vel sensim et vix conspicue senate basi cuneata, peas parce sed longius ciliata, teens glabra, utrimque opaca, sicc. supra nigrescens, subtus fuscescens, nervis utroque latere costae 6-8. Stipulae oblongae, 2-3 cm. longae et 1-1.5 cm. latae, apice rotundatae, longo elata, laxe paniculiformis; rachis ad anthesin circ. 7 cm. longa, postea usque ad 18 cm. elongata; rachis ramulique glabri, sicc. nigres- centes; ramuli ultimi monochasiales, post anthesin usque ad 5 cm. elongati, multiflori; bracteae ovatae, concavae et basi saccatae, plerumque circ. | cm. longae, margine dense ciliatae, ceterum glabrae, sicc. fuscescentes, ad anthesin deciduae. Flores subsessiles. Ovarium 5 mm. longum et 2.7 mm. diam. » glabrum. Calycis lobi oblongi, 7.5 mm. longi et 2.5 mm. lati, acuti, margine ciliati, ceterum glabri, nervis circ. 7 quorum 3 fortiores ] ap ieaes Corolla margine ciliata, ceteru um glabra; tubus 9 mm. longus et . diam.; lobi ovati, 4 mm. longi et 3 mm. lati, cee dimidio Se reflexo et reduplicato. Stamina 7 mm. longa; filamenta m antherae 5.2 mm. longae, connectivo carinato in appendicem contorta am, 1.7 mm. longam producto. Stylus supra receptaculum pollinis obtusus. Fructus late cylindricus, 15 mm. longus et 4 mm. diam., glaber. Semina reticula Habitat terrae Borneensis partem ee RITISH NortH Borneo: Mt. Kinabalu, Penibukan, a M. S. a. 30616 (L, typus, AA, dupl. typi, BZ, tripl. ae ri 31308 ee 32136 306 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII In the colour which it assumes in drying, and in the presence of collapsed cells on the upper side of the leaves, this species resembles A. atropella Stapf, from which it differs, however, conspicuously in its almost complete glabrescence, in the larger size of the leaves and stipules, and above all in the paniculately arranged, subsessile flowers. As Stapf states that in his species the venules are indistinct, the possibility that it might belong to the subgenus Athroophleps, need not be taken seriously. The differences between this species and the two next ones have been summarized in the key. 34. Acranthera capitata Val. in Ic. Bog. 4: 275, t. 391. 1914. Herba ascendens, circ. 30 cm. alta, post anthesin decumbens et ex axillis inferioribus innovationes emittens. Caulis primum pilis rufo-fuscis brev- iter pubescens, deinde glabrescens, diametro ad apicem 2.5 mm., basin versus usque ad 7 mm. aucto, internodiis superioribus 0.5—3 cm. longis, inferioribus usque ad 7 cm. longis, sicc. fuscescens. Folia petiolo 2.5—7 cm, longo, dense pubescente, sicc. olivaceo-brunneo instructa; lamina elliptica vel saepius obovata, 13-25 cm. longa et 6.5—9 cm. lata, subobtusa vel vix conspicue acuminata, basi acuta vel saepius cuneata, margine dense ciliata, supra primum pubescens, mox glabrescens, subtus costa nervis nec non venulis fortioribus dense hirto-pubescens, ceterum subglabra, utrimque opaca, sicc. supra saturate olivacea, subtus fusca, nervis utroque latere costae 6-10. Stipulae ellipticae vel obovatae, 12-15 mm. longae et 8-12 mm. latae, apice rotundatae, costa basin versus prominente, extus primum satis dense, costa densius pubescentes, margine dense ciliatae, deciduae. Inflorescentia breviter pedunculata, plus minusve capituliformis, re vera pentachotome corymbosa, ramulis brevissimis; pedunculus dense_hirto- pubescens, 1-2 cm. longus, recurvatus; bracteae exteriores ovatae, 12 mm. longae et 7 mm. latae; aliae gradatim minores et praesertim angustiores; omnes sicc. fuscescentes, extus sparse, intus densius pubescentes, margine dense ciliatae, ad anthesin persistentes. Flores sessiles. Ovarium 3 mm. longum et 1.2 mm. diam., dense griseo-pubescens. Calycis lobi lanceolati, 5 mm. longi et 1.8 mm, lati, acuti, margine ciliati, extus intusque pubes- centes, nervis 3 vel 5 quorum 3 fortiores instructi. Corolla extus parce pubescens, margine ciliata; tubus 7 mm. longus et 4 mm. diam.; lobi 2.8 mm. longi et 1.9 mm. lati, acuti. Stamina 6.7 mm. longa; filamenta 1.6 mm.; antherae 5.2 mm. longae, connectivo vix conspicue carinato in appendicem rectam, 0.9 mm. longam producto, Fructus glaber dictus, nondum plene maturus 10 mm. longus et 2 mm. diam. Habitat terrae Borneénsis partem septentrionalem. Borneo: Eastern and Southern Division, Tidoong: Ulu Sebulu, Amdjah 634 (L, dupl. typi); S. Tulit, id. 664 (L). According to Valeton l.c. the flowers of Amdjah 664 should be violet, those of the other specimens white; maybe the colour changes before the corolla is shed. The differences between this species and A. Ruttenii Brem. are given in the key. Although rather different in aspect, they are doubtless nearly related. 1947] BREMEKAMP, ACRANTHERA 307 35. Acranthera Ruttenii Brem. n. spec.; typus: Rutten 642 (U). Herba suberecta, circ. 60 cm. alta, post anthesin decumbens et ex axillis inferioribus innovationes emittens. Caulis primum pilis longis strigosus, deinde glabrescens, diametro ad apicem 1.5 mm., basin versus usque ad 3 mm. aucto, internodiis superioribus 0.8—1.5 cm. longis, inferioribus usque ad 11.5 cm. longis, sicc. saturate olivaceus. Folia petiolo 1.5—2.5 cm. longo, dense oe -villoso, sicc. olivaceo instructa; lamina lanceolata vel obovata, 8-11 cm. longa et ‘4 5 cm. lata, apice acuminata, basi contracta, margine dense ae supra primum sparse villosa, deinde glabrescens, subtus costa nervis nec non venulis fortioribus dense strigosa, ceterum glabra, utrimque opaca, sicc. supra olivacea, subtus dilute fusca, nervis utroque latere costae 4 vel 5. Stipulae oblongo-ellipticae, 13-15 mm. ta basi extus primum praesertim costa parce villosae, margi um densius ciliatae, subpersistentes. Inflorescentia pedunculata, capituliformis: pe- dunculus dense griseo- -pubescens, circ. cm. longus, patens; eae exteriores ovatae, 13 mm. longae et 7.5 mm. latae: aliae angustiores omnes sicc. fuscescentes, extus sparse et intus densius pubescentes, margine dense ciliatae, ad anthesin persistentes. Flores sessiles. Ovarium 4 mm. longum et 2. 2 mm. diam., dense griseo- seat Calycis lobi lineari- lanceolati vel lineari- -oblongi, 7.5 mm. longi et 1.8-2.2 mm. lati, acuti, margine ciliati, extus vix conspicue, intus densius pubescentes, nervis circ. 5 quorum 3 fortiores instructi. Corolla extus sericeo-villosa, margine ciliata; tubus 7 mm. longus et 3 mm. diam.; lobi 2.7 mm. longi et 2 mm. lati, acuti. Stamina 6.5 mm. longa; filamenta 1 mm.; antherae 5.5 mm. Jongae, connectivo vix conspicue carinato in appendicem rectam, 0.7 mm. a oe Fructus nondum notu ae Borneeénsis partem scot nate oe Eastern and Southern Division: Bulongan, S. Sadjau, 117° 40’ E.long., 2° 40’ Nilat., Rutten 642 (U, typus) INDEX SPECIERUM 21. abbreviata Val. in Ic. Bog. 4: 181, t. 355. 1913 — West Borneo. anamallica Bedd., Ic. Pl. Ind. Or. 1:5, t. 23. 1874 — Indian Peninsula. 33. athroophlebia Brem. n. spec. — North Borneo. 32. atropella Stapf in Trans. Linn. Soc. Bot. II, 4: 173. 1894 -— North Borneo. 16. aurantiaca Val. ex Brem. n. spec. — West Borneo. 20. axilliflora Val. in Bot. Jahrb, 44: 550. 1910 — South-east Borneo. 12. bullata Merr. in Mitt. Inst. Allg. Bot. Hamburg 7: 284. 1937 — West Borneo. 34. capitata Val. in Ic. Bog. 4: 275, t. 391. 1914 — North Borneo. 3. *ceylanica Arn. Meisn., Pl. hips Gen. 2: 115. 1838 — Cey] didymocarpus H. H. W. pen rson ex Ridl. in Jou Fed. Mal. “Sialee Mus. 4: 1909, in syn.: ae didy SE Rid 24. Endertii Brem. n. spec. — East Borneo. 7. frutescens Val. in Bot. Jahrb. 44: 551. 1910 — Borneo, North Borneo ae 2. grandiflora Bedd., Ic. Pl. Ind. Or. 1:5, t. 25, 1874— Indian Penin : Griffithii Hook.f., Fl. Brit. Ind. 3:92. 1880—Asemanthia eae "(Hook f.) Brem. n mb _— 308 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII 31, Hallierii Val. in Ic. Bog. 4: 183, t. 356. 1913 — Central Borneo. 22. hirtistipula Val. in Ic. Bog. 4: 277, t. 392. 1914— Central Borneo. 11. involucrata Val. in Ic. Bog. 4: 279, t. 393. 1914 — West Borneo. 13. Johannis-Winkleri Merr. in Mitt. Inst. Allg. Bot. Hamburg 7: 285. 1937 — West Borneo. 29. lanceolata Val. in Ic. Bog. 4: 281, t. 394. 1914 — Central and West Borneo. 8. longipes Merr. in Papers Mich. Acad. Sci. 19: 194. 1934 — East Sumatra. longipes Merr. in Mitt. Inst. Allg. Bot. Hamburg 7: 284. 1937, Sone illeg.= longipetiolata. 19. oe Merr. ex Brem. n. nom. (longipes Merr. 1937, non 1934) — West 23. Peon Val, in Ic. Bog. 4: 283, t. 395. 1914 — Central Born faingayi Hook.f., Fl. Brit. nan ee 92. 1880—A semanthia Maingayi (Hook.f.) Ridl. in Kew Bull. 1939: 600. 18. megaphylla Brem. n. spec. an Tse 14. monantha Val. in Bot. Jahrb, 48: 111. 1912 — North (?) Bor multiflora Val, in Ic. Bog. 4: 285, t. 396. 1914, ith eeallene Seam multinervia Val. in sched. cf. bullata. mutabilis Hem a in Jour. Bot. 25: 204. 1887 (Mussaenda mutabilis Hemsl. in Hook., Ic. Pl. t. 1718. 1887)—=Asemanthia Maingayi (Hook.f.) Ridl. in Kew Bull. 1939: nes 1939 mutica Val. in sched. cf. simalurensis 30. Nieuwenhuisii Val. ex Brem. n. spec. — Central Borneo. 7. ophiorrhizoides Val. in Ic. Bog. 4: 287, t. 397. 1914— North Borneo. 25. parviflora Val. in Ic. Bog. 4: 289, t. 398. 1914 — West Borneo. 28. philippinensis Merr. in Philip. Jour. Sci. Bot. 8: 32. 1913 — Mindanao. 35. Ruttenii Brem. n. spec. — North Borneo. 17. salmonea Brem. n. spec. — East Borneo. 5. siamensis pena Brem. n. aN (Psilobium siamense Kerr in Hook., Ic. Pl. t. 3332. 1937) — Peninsular Sia 6. siliquosa eae n. spec. — East Spies 10. siorpntediges Brem, n. spec. — Simalur strigo ge in Ic. Bos. 4: 291, t. 399. 1914, species incertae sedis nondum visa — Cen ral Borneo. 4. tomentosa R.Br. ex Hook.f., Fl. Brit. Ind. 3: ae 1880 — Ass uniflora lee x G. Don) Kurz in Jour. Soc. Bengal “41 (2): 312. paar (Mussaenda nerve Wall. ex G. Do n=-Aphaenandra uniflora (Wall. G. Don) "Brem, in Blumea, Suppl. 1: 121. 193 15. variegata Merr. in Mitt. Inst. Allg. Bot. cee 7: 285. 1937 — West Borneo. 26. velutinervia Brem. n. spec. — North sais 9. Yatesii Merr. in Papers Mich. Acad. 194. 1934 — East Sumatra. zevlanica Arn. in Ann, Nat. Hist. 3: ee a ee ZEIsT, HOLLAND. 1947 | FAULL, TROPICAL FERN HOSTS OF RUST FUNGI 309 TROPICAL FERN HOSTS OF RUST FUNGI J. H. FAutyi RusTS ON FERNS are referred in current literature on mycology and plant pathology to the definitive genera Hyalopsora, Milesia, Uredinopsis, Desmella and Puccinia and in a few instances to the imperfect genus Uredo. The complete life-histories of many species of the first three have already been determined experimentally; in all cases they have proved to be heteroecious, with species of Adbies serving exclusively as their aecial hosts. So it may quite safely be assumed that the same pattern is poten- tially true of all the other species of Hyalopsora, Milesia and Uredinopsis. Regarding Desmella, uredia and telia only are known; and as they are so unlike those of the three foregoing genera, the identities of hosts that can carry the aecial stages of Desmella rusts are not even conjectured. It is a strange genus, taxonomically standing quite apart from the other fern- restricted genera. Thus far it has been reported from the American tropics only. Among the remaining fern rusts, a single species of Puccinia has been described, though solely with respect to its uredia and telia. Finally, as to the ved fern rusts, they can with reasonable certainty be recognized as uredo stages of one or other of the five definitive genera. Indeed, not a few of the named species of these genera have been described from the uredo stage and without knowledge of telia. Although technically open to some objections, this practice does offer advantages if used with discretion. With possible exceptions of Australia and Tasmania, fern rusts are world-wide in distribution. Of course, locally within any extensive region there may be limiting factors, such, for example, as continuously high day and night temperatures. Otherwise they are likely to be found wherever ferns grow and on a surprisingly large number of specific hosts. Natur- ally, where those that are Adies-infecting occur beyond the ranges of Abies, they can be perpetuated solely by seasonal transmission from affected ferns to plants of identical kind or of species that are likewise susceptible. Indeed, as the southern distributional limits of Abies are approached, this method becomes increasingly frequent, even in those rust species that lack amphispores. Of course this method is possible for amphispore-producing rusts in any latitude, but otherwise only wherever there is a sufficiently close overlapping of successive seasonal crops of fronds, that is, a succession within the vital life-span of ordinary urediospores. It may not be super- fluous to recall to mind here that the most southerly ranges of existing Abies, though well within the tropics in the western hemisphere and reaching to the tropics in the eastern, fall far short of the equator. - Ww Ne = mn a ~ — JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVHI TABLE I PUBLISHED TROPICAL FERN HOSTS, THEIR RUSTS AND REGIONAL ORIGINS Fern hosts Rusts Regional origins . Adiantum andicola Liebm. Uredinopsis investita Faull Guatemalajy . Anemia fulva (Cav.) Sw. Desmella Aneimiae (Henn.) Syd. Brazil . Anemia Phyllitidis (L.) Sw. Desmella Brazil . Anemia tomentosa (Sav.) Sw. Desmella Aneimiae (Henn.) Syd. Brazil? (A. cheilanthoides Kaulf.) Anirophyum lanceolatum (L.) Hyalopsora Polytaenti (KCT) Dominican a > J . Denns enp.) 1e1s (K JOURNAL OF THE ARNOLD ARBORETUM . Dennstaedtia ordinata (Kau aulf.) Moo Fern hosts Blechnum occidentale L.* . Blechnum orientale L. . Blechnum unilaterale Sw. . Cheilanthes membranacea (Dav axo . Cheilanthes microphylla Sw. . Cheilanthes pyramidalis Fée . Cheilanthes tenuifolia (Burm.) Sw. . Coniogramme fraxinea (Don) Diel . Cyclopeltis semicordata (Sw.) J. Sm. . Cystopteris fragilis (L.) Bernh. . Dennstaedtia adiantoides (H. & B.) Moore . Dennstaedtia cicutaria (Sw.) Moore . Dennstaedtia dissecta (Sw.) N . Dennstaedtia exaltata (Kze.) Hieron re Moore taedtia rubiginosa (Kaulf.) Moore* . Dennstaedtia rubi aulf.) viNOSA or . Diplazium sp. (?) . Diplazium sp. (?) . Diplazium crenulatum O. Liebm. . Diplazium expansum Willd. . Dryopteris sp. . Dryopteris, n. sp. (2?) acc. to Maxon 7. Dryopteris boqueronensis i ) TABLE II (Continued) Milesia Hyalopsora Milesia Uredino psis Milesia Hyalopsora Milesia Hyalopsora Desmella Hyalopsora Milesia Milesia Milesia Milesia Milesia Desmella Milesia H yalopsora Hyalopsora Milesia Desmella Milesia Milesia Milesia Rusts [VOL. XXVIII Regional origins Panama; Venezuela Borneo Mexico Mexico Jamaica Mexico New Guinea China (Lat. 25° N) Trinidad Mexico Cuba Mexico Jamaica Mexico Jamaica Jamaica Guatemala; Jamaica ; Guatemala Cuba New Guinea Jamaica Ecuador 1947] TABLE II (Continued) FAULL, TROPICAL FERN HOSTS OF RUST FUNGI Fern hosts Rusts Regional origins : Drsope . Drvopteris concinna (Willd.) Kuntze . Dryopteris dentata (Forsk.) C. -Chi,. . Dryopteris diplazioides(Desv.) Urban . Drvopteris effusa (Sw.) Irban . Dryopteris equestris (Kunze) C; Chr . Dryopteris firma (Baker) C. Chr Drvopteris heteroclita (Desv.) CG. Chr . Dryopteris melanochlaena Cc Chr. . Dryopteris navarrensis Christ . Dryopteris Nockiana (Jenm.) C:’Chr : Devotee oligocarpa (H.& B Kuntze . Drvyopteris oligocar pa (H. & B.) Kuntze . Dryopteris opposita (Vahl) Irban (?) . Drvyopteris paleacea (Sw.) Ceehr. ; ea patens (Sw.) Kuntze* : Drveptr ares (Sw. ) Un- derw. V tC. Chr. . Dryopteris plat (KI. Karst.) Hieron. (approaches D. navarrensis Christ) . Dryopteris resinifera (Desv.) Weatherby : wi clare aie (Desv.) Weathe ee ee Maxon & Morton : a la rudis (Kze.) rm Sloanii (Bak.) ntze (D. oligophylla on) Milesia Milesia Milesia Milesia Milesia Milesia Milesia Milesia Milesia Milesia Desmella Milesia Milesia Milesia Milesia Milesia Hyalopsora Desmella Milesia Milesia Milesia Milesia Guatemala; Panama Panama Ecuador Jamaica Guatemala Jamaica Jamaica Guatemala Panama Jamaica Jamaica Guatemala; a Mexic ee ena ; Mexic ae Mexico Guatemala Mexico Guatemala Guatemala Guatemala Jamaica [VOL. XxXvIi 316 JOURNAL OF THE ARNOLD ARBORETUM TABLE II (Continued) Fern hosts Rusts Regional origins 70. Dryopteris Sprengelii (Kaulf.) Milesia Jamaica Kuntze 71. Dryopteris tetragona (Sw.) Desmella Jamaica Urban* 72. Elaphoglossum lingua (Rad- Hyalopsora Jamaica i ack. 73. Elaphoglossum Pringlei (Da- Milesia Mexico venp.) C. Chr. 74. Hemionitis palmata L. Milesia Jamaica 75. Odontosoria Jenmanii Maxon Milesia Jamaica 76. Pellaea cardiomorpha Weath- Hyalopsora Mexico erby 77. Pityrogramma sulphurea Desmella Jamaica (Sw.) Maxon 78. Polybotrya osmundacea HBK Milesia Jamaica 79. Polypodium aureum L Desmella Panama 80. Polypodium ellipsoideum Fée Hyalopsora Guatemala; Mexico 81. Polypodium fissidens Maxon Hyalopsora Guatemala 82. Polypodium fissidens Maxon Milesia Guatemala 83. Polypodium loriceum L. Desmella Jamaica 84. Polypodium Martensii Mett. Hyalopsora Guatemala; KiCO 85. Polypodium plesiosorum Uredino psis Guatemala; Kunze exico 86. Polypodium Veitchii Bak. var. Hyalopsora China (Lat. 26° glaucopsis (Franch.) Ching N) 87. Polystichum rachichlaena Fée Milesia Guatemala 88. Pteris longifolia L Milesia Jamaica 89. Pteris longifolia L. Desmella Jamaica 90. Pteris quadriaurita Retz. Milesia Panama 91. Pteris quadriaurita Retz. Desmella Jamaica 92. Tectaria irregularis (Pr.) Milesia New Guinea Copeland (?) 93 Hyalopsora Mexico . Woodsia mollis (Kaulf.) .J.Sm. SUMMARY AND COMMENTS 1. Rusts are recorded for the first time as occurring in the Ophioglos- these may possibly be added the saceae and Cyatheaceae. Gleicheniaceae. 2. Hyalopsora rusts are recorded for the first time as occurring in the To Ophioglossaceae and Cyatheaceae. 1947 | FAULL, TROPICAL FERN HOSTS OF RUST FUNGI 317 3. Milesia rusts are correctly recorded for the first time as occurring in the Schizaeaceae and Cyatheaceae. To these may possibly be added the Gleicheniaceae. 4. Species of the following fern genera, found growing within the tropics, are recorded in this paper as: (a) hosts for Desmetra rusts: Adiantum (1), Anemia (4), Blechnum (1), Cyclopeltis (1), Dennstaedtia (1), Diplazium (1), Dryopteris (5), Pityrogramma (2), ‘“Polypodiacea” (1), Polypodium (2), Pteris (2); (b) hosts for Hyaropsora rusts: Adiantum (2), Anogramma (1), Antrophyum (1), Asplenium (1), Athyrium (2), Blechnum (1), Botry- chium (1), Cheilanthes (1), Coniogramme (1), Cyathea (1), Cystop- teris (1), Diplazium (2), Dryopteris (1), Elaphoglossum (2), Pellaea (2), Polypodium (4), Woodsia (1); (c) hosts for Miresia rusts: Anemia (1), Asplenium (1), Athyrium (1), Blechnum (4), Cheilanthes (2), Coniogramme (1), Cyathea (3), Dennstaedtia (6), Diplazium (1), Dryopteris (26), Elaphoglossum (1), Gleichenia ? (1), Hemionitis (1), Nephrolepis (3), Onychium (1), Odontosoria (1), Pellaea (1), Polybotryum (1), Polypodium (2), Polystichum (2), Pteris (2), Tectaria (1); (d) hosts for Ureprnopsis rusts: Adiantum (1), Anogramma (1), Asplenium (1), Athyrium (1), Blechnum (1), Cheilanthes (2), Cystop- teris (1), Pellaea (1), Polypodium (1), Pteridium (including under P. aquilinum its varieties and marginal species). The names in black face are genera within the tropics recorded for the first time as hosts for rust fungi. The numbers of species reported for each genus are indicated in parentheses. 5. Desmella is reported in this paper on 21 fern host species collected within the tropics, Hyalopsora on 24 species, Milesia on 63. species, Uredinopsis on 10 species, Puccinia on one species, and Uredo on 4 species. The paper lists a total of 109 tropical fern species (exclusive of varieties, etc.) known to be hosts of rust fungi; of these 73 are recorded for the first time. I have found 25 rusted fern host species in Guatemala, 27 in Jamaica, 27 in tropical Mexico and 9 in Panama. ashioka, in collections made both north and south of the Tropic of Cancer, reported 16 from Formosa (as published by Hiratsuka & Hashioka in their ‘Uredinales collected in Formosa”). Varieties are not included in any of these num- bers, nor are the submarginal species of Pteridium aquilinum. 7. According to my experience, rusted ferns in the tropics are rarely found below an elevation of about 2000 feet above sea level. Uredo rust on Lygodium is exceptional; I collected it at sea level around Puerto Barrios in Guatemala. I could find no fern rusts on Barro Colorado Island, Panama Canal Zone; the greatest elevation on that island is said to be under 550 feet. Much is to be expected from the vast mountainous regions extend- ing eastward from Iran (Persia) to the South China Sea, from which have 318 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII come scant collections only, and these from very limited northern areas, probably in reality extratropical, such as around Kunming, Yunnan, China. Comprehensive data based on old and new records embodied in fern rust collections from the tropics are summarized in Table III. This table designates all the relevant fern genera hosts, the numbers of their affected species, the involved rust genera for each fern genus, and the regions in which the collections were made. TABLE III FERN RUST HOST GENERA IN THE TROPICS AND THEIR DISTRIBUTION —Desmella H yalopsord Milesia Uredinopsis Adiantum 4° Trinidad 1* Mexico 1; Brazil 1 Guatemala 1 Anemia 4 Brazil 3; Jam. ———— Jamaica 1: ae 1; Mex. 1 Mexico 1 Anogramma 1 —- Mexico l ——— Mexico | Antrophyum 1 ——— Dom. Rep. 1; Porto Rico 1 Asplenium 2 —_—_—_——_ Guatemala 1; New Guinea 1 Mexico 1 Athyrium 4 ——- Mexico 2; Mexico 1 Guatemala 1; New Guinea 1 Mexico 1 Blechnum 6 Brazil 1 Borneo | New Guinea 1”; Colombia 1 Trop. Am.* 1, 1, 3 Cheilanthes 4 Mexico 1 Jamaica 1; Mexico 1, 1 New Guinea 1 Coniogramme 1 —_—— wo Formosa 1 Cyathea 4 ———— New Guinea 1 Guatemala 1; amaica 1; Mexico 1 Cyclopeltis 1 Trinidad 1 —— Cysto,teris 1 ——-_ Mexico l ——— Mexico 1 Dennstaedtia 6 Jamaica 1; — Colombia 1; Cuba Porto Rico 1 1; Guatemala 1; Jamaica 3; Mexico 2; Panama 1 Diplacium 4 Cuba 1 New Guinea 2. Guatemala 1 * The seein indicate he alee of host species niga: Those in bia face are new records. » This aah be a Hyalopsora. © Colombia 2, Costa Rica 1, Cuba 1, Ecuador 1, Guatemala 1, Jamaica 1, Mexico 2, Panama 1, Porto Rico 1, Venezuela 2. 1947] FAULL, TROPICAL FERN HOSTS OF RUST FUNGI 319 TABLE III (Continued) Desmella Hyalopsora Milesia Uredinopsis Dryopteris 29 Jamaica 2; Guatemala 1 Ecuador 1, 2; Mexico 1; Formosa 2; Porto Rico3; Guatemala 8; Venezuela 2 Jamaica 1, 9; Mexico 3; New Guinea 1; Panama Elaphoglossum 4 —— Colombia 1; Mexico 1 Jamaica 1, Gleichenia (?) 1 — New Guinea | —__—___ Hemionitis 1 ——__—. —__— Jamaica 1 —_—_—— Lygodium 4° —_—_—_ —_—— Ne phrole pis 3 —_—_—_——__ —__—_—_ Colombia 1 Phil. Islands 1; Porto Rico 1 Odontosoria 1 on —— Jamaica 1 SEE Onychium 1 as —_—_— Phil. Islands 1 or Pellaea 3 —_—— Ecuador 1; Madagascar 1 Mexico 1 Mexico 1 Pityrogramma 2 Ecuador 1; maica 1; Venezuela 1 Polybotrya 1 a —_ Jamaica 1 “Polypodiacea” 1 Ecuador 1 eee 2a oe Polypodium 7 Jamaica 1; Guatemala 3; Formosa 1 Guatemala 1; Panama 1 Mexico 2 Mexico Polystichum 2 —_—_——- —-- Formosa 1; Guatemala 1 Pteridium ost world-wide Pteris 2 Jamaica 2 —_—_ a 2 Pana 1 Tectaria 1 ~S oe or New nae 1 W oodsia 1 = Mexico 1 4 Brazil 2; Br. Guiana 1; Guatemala 1; San Salvador 1; Trinidad 1; Venezuela 1. The rusts involved are sefatied to the genera Uredo and Puccinia in Tables I and II. ARNOLD ARBORETUM, Harvarp UNIVERSITY. 320 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIIE SPECIES NOVAE FAGACEARUM SINICARUM Woon-YouNG CHUN AND YING TSIANG THE MATERIAL on which this paper is based is part of an extensive collection made between the winter of 1942 and the summer of 1945 under the direction of the junior author. When the situation in Hong- kong, where the Botanical Institute was temporarily stationed, became critical, Professor Tsiang led a part of our staff to northern Kwangtung to contact the main body of the University. The attack on Hongkong commenced while the party was enroute. In the winter of 1942 a Branch Institute was established at Li Yuan Po in southern Hunan, and from this new center, botanical explorations in southern Hunan, northern Kwang- tung, and the Shih-Wan-Ta Shan region bordering Kwangtung and Kwangsi continued throughout the four years of the Pacific War. As a result of unremittent labor under difficulties needless to detail, nearly 10,000 numbers totalling over 100,000 specimens, have been col- lected. This large pongenal is now available to specialists and botanical institutions throughout the Pending the reissue of our oat Sunyatsenia, this paper is submitted to the Journal of the Arnold Arboretum for publication to mark the re- sumption of normal relations between our two institutions. The types of the new species here-in described are preserved in the herbarium of the Botanical Institute, Sun Yatsen University; duplicate types in the herbarium of the Arnold Arboretum. Lithocarpus Chifui Chun et Tsiang, sp. nov. rbor 12 m. alta et trunco (e notis collectoris) ad 75 cm. diametro; ramuli mediocriter robusti teretes glabrati, in sicco fusco-nigri lenticellis paucis magnis discoloribus notati anno secundo vix dissimiles; gemmae terminales solitariae vel aggregatae, proportione parvae, 3-4 mm. longae, subglobosae vel ovoideo-obtusae perulis ovatis breviter acuminatis rubes- centibus lucidis dorso carinatis margine ciliolatis obtectae. Folia ut videtur per duos annos persistentia, firme coriacea glaberrima oblongo- lanceolata vel oblongo-oblanceolata, apice late acuta, basi saepe oblique cuneata in petiolum robustum paulo producta rarius fere rotunda et in petiolum nna contracta, supra atro-viridia subtus cinereo-viridia (teste collectore) siccitate flavescentia, margine integra saepe late undulata, magnitudine in eodem ramulo satis variabilia, minora plerumque ad 15 m, longa et 5 cm. lata, majora 25 cm. longa et 12 cm. lata, costa media supra valde elevata infra medium canaliculata, basin versus “circiter 1.2 mm. lata, costa nervisque subtus distincte elevatis lateralibus a costa angulo 25°-30° ab illa divergentibus utrinsecus 10 sea 11 inter se satis distantibus patenti-curvatis ante marginem vix prominente arcuato-ana- stomosantibus trabeculis transversis subparallelis ae sub lente crebre 1947] CHUN & TSIANG, SPECIES NOVAE FAGACEARUM 32] foveolato- reticulatis conjunctis; petiolus ge turgidus supra late sulcatus glaber, vix ultra 1.5 cm. lon ngus. Flores ¢ desunt. Flores @ (post anthesin) quini-aggregati, unicus ran maturans, 4 alii abortivi, glomeris inter se distantibus plerisque haud maturatis; lamellae 4—5 con- centrices tenues erosae obscure griseo- -puberulae. Fructus biennes secus pedunculum validum erectum nigrum parce lenticellatum ad 17 cm ongum proventu in examplo viso 7-8 sessiles contigui unilaterales bifariam divergentes; cupula globosa circiter 2 cm. diametro, stylopodio excepto totam glandem includens, tenuis extus undique griseo-lepidotis, zonis 3—4 tegris superioribus lacerato-dentatis triangularibus tenuibus_ ornata. Glans apice excepta cum cupula concrescens depresse globosa_ vertice complanata basi rotundata glabra pallida subnitida, longitudinaliter rugulosa Daas lignoso 2-2.5 mm. crasso; intus eseptata, semen penta gon Kwanctunc: Yu Yuen Hsien, Mo-Fung Shan, tree . m. tall, % m. in diameter, leaves dull green above, gray green beneath, Nov. 19, 1933, S. y Ko This new species may be compared with piece ste Hick. et A. Camus from which it is amply distinct by isolated, not ternately coalescent acorns. The pistillate inflorescence of our new species, as educed from the fruiting state, shows some interesting features. The flowers are clustered in 5’s in more or less pentagonal groups, but all those on the under side of the rhachis remain abortive, and of those more advantageously placed in relation to light, only one in each cluster reaches maturity. The individual flowers are contiguous but not confluent. The ripe acorns are alternately juxtaposed 1 in two ranks all facing one direction. This outstanding species is named after Dr. W. P. Fang, Professor of Botany in the National Gy acoaed of Szechuan in recognition of his exhaustive researches on the flora of Omei Shan, and in appreciation of the cooperative spirit dominating his associations with the authors. Lithocarpus chrysocoma Chun et Tsiang, sp. nov. Arbor usque ad 12 m. alta, trunco 50 cm. diametro coma late ovoidea. partibus juvenilibus inflorescentiisque ac fructibus dense et molliter to- mentosis, cortice in lamellas irregulares profunde fisso, ramis vix patentibus ramulis robustis angulatis junioribus molliter fulvido-tomentosis cito sor- dide flavescenti-tomentosis eo eae glabrescentibus. Gemmae ovoi- deae vel rotundatae obtusae, circiter 2 mm. longae, perulis ovalis rubro- brunneis lucidis dorso ee ae pilosis. Folia triennia coriacea, farinosa vel pu en lepidotula, elliptico- vel lanceolato-oblonga apice abrupte falcatim breviter caudato-acuminata, basi shite cuneata, margine integerrima subtiliter revoluta, 7-10 cm. longa et 3—4 cm. lata, maxima 16 X 6.5 cm., costa nervisque supra Schnee ae elevatis lateralibus in utroque latere 9-10 interdum 12 sursum leviter curvatis ante marginem evanescentibus venulis a petiolis molliter sordide to- mentosis, supra leviter sulcatis, 1-1.5 mm. longis. Amenta mascula apicibus ramulorum hornotinorum nine et ex axillis foliorum sum- 322 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXVIII morum paniculata rhachidibus angulatis dense fulvido-tomentosis 3—9 cm. longis, floribus fasciculatis, perianthii lobis membranaceis rubro-purpuras- centibus extus tomentosis ovatis concavis circiter 1 mm. longis, staminibus longe exsertis °1.5-2 mm. ioe filamentis glabris filiformibus, antheris globosis; spicae femineae solitariae in axillis foliorum superiorum aggre atae, 3.5 cm. longae; florum fasciculi ad axem crassiusculam dense to- mentosam conspersi sessiles 3—6-flori; flores (juvenales non visi) mens numerosis arcte imbricatis tomentosis; styli 3 glabri divaricati. Fructus biennes 2-3 aggregati inferne confluentes sessiles turbinato- de ager cupula tenuiter lignosa, circiter 1.5 cm. alta et 2 cm. diametro, extus fulvido-tomentosa intus adpresse sericeo-pilosa; squamae multiseriatae deorsum confluentes apicibus acuminatis liberis falcatis; glans praeter apicem planam vel leviter curvatam medio umbonatam cee depresse globosa vel ssi apie sericeo-puberula, basi cupula solum concrescens, pericarpio usque 1.8 cm. crasso; stigmatibus deciduis; cicatrix alta convexa circiter 18 mm. lata. Kwanctunc: Yu-Yuen Hsien, Ching-Chi Tung, tree 6-8 m. tall, July 1933, S. P. Ko $2839, 52842, 52992, 52933; same district, Shang-Yuen Shan, July 1933, S. Ko a Mo-Fung, Nov. 1933, S$. P. Ko 53686; Yu-Yuen Hsien, Chine. Chi Pune. Aug. . 1935, Z. S. Chung 10957 (TYPE); same locality, Oct. Y. Li 2059. eee Without data, No. 5977 (B. I. tert. No. 38776). Hunan hang, Mang Shan, tree 7 m. tall, Oct. 1942, P. H. Liang 83704, pis 83853; same locality, Oct. 1942, S. H. Chun 2720. This new species is easily distinguished from its nearest allies, L. amygdalifolia (Skan) Schottky and L. dealbata (Hook.f. et Th) Rehd, by the golden to reddish yellow pulverulent indumentun: oa we ander surface of the leaves. Lithocarpus Paihengii Chun et Tsiang, sp. nov. Arbor ad 12 m. alta trunco 22 cm. diametro cortice cinereo-brunnei in lamellas irregulares fisso; ramuli annotini validi angulati profunde sulcati in sicco nigrescentes hornotini crassissimi cylindrici sordide nigrescentes lenticellis satis conspicuis es conspersi; gemmae magnae obtuse ovatae, 10-12 mm. longae, perulis exterioribus late ovato-rotundatis flavido- vel fulvo-brunneis glabris interioribus albo-sericeis obtectae Folia persistentia valde coriacea rigida plana oblongo-lanceolata, 15-20 cm. longa, 5-8 em. lata, apice subito in acumen latum falcatum acutum breviter acuminata, basi late cuneata plerumque in petiolum quasi-alatum decurrentia, supra intense viridia lucida subtus pallidiora vel albescentia (e collectore) initio dense indumento fulvo-furfuraceo-pulverulento anno secundo saepe plus minus evanescente obtecta, costa media utrinque elevata supra dimidio inferiore sulcata subtus et nervi ochrascentes, nervis utrinse- us 8-10 sub angulo 50° divergentibus obliquis curvatis nonnulla furcatis versus marginem indistincte anastomosantibus; ale transversae subtus tantum tenuiter prominulae; petiolus complanatus, 2.5—-3 cm. longus. Fructus biennes in rhachidibus validis nigris lenticellatis 2.5—-5 cm. longis 2-vel 3-ni glomeratis glomerulis inferne coalescentibus sessilibus; cupula depresse globosa, tenuiter lignosa, 2—2.5 cm. diametro, ore 10-14 mm. diam., fere totam pier Sie obvoluta sed basin tantum glandem adhaerens, intus densissime et brevissime albo-tomentosa, squamis conspicuis crassis imbricatis deltoideis apicibus liberis, inferne flavo-cinereo-tomentosis 1947] CHUN & TSIANG, SPECIES NOVAE FAGACEARUM 323 superne glabrescentibus; glans depresse globosa, 18 mm. lata, 15 mm. alta, castanea nitida minute albido-furfuracea parte superiore detergente plana vel subrotundata apiculata; cicatrix basilaris valde convexa albida rugu- losa, 15 mm. diametro; nucula subpentagona, obtuse 5-costata. Hunan: I-Chang Hsien, Chin-Chuan Village, Cha Shan, alt. 930 m., tree 6 m. tall, on slope of sparsely wooded ravine, leaves whitish green beneath Sept. 25, 1942, P. H. Liang 63698; same locality, tree 12 m. tsll, 22 cm. diam., bark gray brown peeling off in plates, young leaves brownish beneath, Sept. 26, 1942, P. H. Liang 83729, This new species is most closely related to Lithocarpus cleistocar pa (Seemen) Rehd. et Wilson from which it differs in thicker coriaceous leaves with a fulvous furfuraceous indumentum on the undersurface which persists more or less into the second year, and especially in the very different fruits. In L. cleistocarpa the acorn is completely adnate to the cupule excepting the exposed apex, while in this new species, it adheres to the involucre only at the base. The wood is hard and durable and is highly valued at least locally for making carrying poles. It affords us unusual pleasure to associate this handsome and useful tree with the name of Professor W. C. Cheng, a dendrologist of outstanding achievements. Quercus dispar Chun et Tsiang, sp. nov. Arbor 8-16 m. alta partibus juvenilibus dense griseo-flavido tomentosis; ramuli tenuisculi teretes grisei serius glabrescentes lenticellis sparsis Vix conspicue notati. Folia persistentia biennia tenuiter coriacea elliptico- oblonga vel obovato-lanceolata, absque petiolo 8-10 cm. longa = poem lata, apice breviter lateque falcato-acuminata vel -acuta, basi obtusa vel cuneata, margine e tertio infero raro e medio crenato-dentata dentibus us mm, longus sordide fusco-pilosulus; stipulae chartaceae ellipticae acutae, 5 mm. longae, 1.5 mm. latae, extus puberulae. Inflorescentiae non satis evolutae. Fructus sessiles solitarii in autumno secundo maturi. Cupula pateriformis, basi plana, tertiam partem glandis amplectens, ad 3 cm. ia 8 cm. alto, intus minute sericeo-pilosula extus flavido-griseo tomentosa, concentrice zonata annulis circiter 10 tenuibus infimis grosse erosis supremis integris. Glans depresse globosa vel disciformi-subhemi- sphaerica, 2.8-3 cm. lata 1.2-1.5 cm, alta, plus minusve dense puberula, sn plana umbonata, basi truncata; cicatrix plana rugulosa, 16 mm. iam. Kwancst: Shang-Si Hsien, Shih-Wan-Ta Shan, Feng-Hwang Hsiang, shrub 8 m. tall, in sparsely wooded ravine on wet sites beside stream, branchlets pale gray, leaves deep green above, pale green beneath, young inflorescences white, March 3, 1944, S. H. Chun 4659. 324 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVII This species is related to Quercus vestita Rehder et Wilson from which it differs in indumentum, smaller differently shaped leaves with much more regular venations, and sessile deeper cup with more numerous annular rings. Quercus Dussaudii Hick. et A. Camus which has similar fruits, besides being glabrous, differs in having an acorn with a deeply intruded basal scar. Quercus delicatula Chun et Tsiang, sp. no Arbor 13 m. alta fructu excepto ciesinn cortice ramulorum fere lenticellis parvis sparsis obtecti. Gemmae ovatae obtusae brunneae, vix ultra 2 mm. longae. Folia chartacea, per duos annos persistentia petiolata elliptica vel elliptico-lanceolata, apice breviter caudato-acuminata, basi nitidula, subtus leviter tantum pallidiora opaca, costa nervisque tenuibus supra cum trabeculis crebre reticulatis planiusculis vix conspicuis inferne manifeste elevatis haud venulosis nervis utrinsecus 7—8 patenti-ascendenti- bus ad marginem arcuatis obscure confluentibus; petioli graciles superne sulcati, circiter 1 cm. longi. Flores ignoti. Fructus biennes in axillis foliorum superiorum solitarii pedicellis cum pedunculis 1-1.5 cm. longis stipitati; cupula cupuliformis, 8-11 mm. alta, 14-18 mm. diametro, tenuis sed firma, intus adpresse fulvo-sericeo-pilosa; annulis 6—7 tenuibus erosulis sursum atrobrunneis striatis glabrescentibus deorsum dense griseo-tomen- metrica vel saepe sursum plus minusve oblique curvata, atrocastanea lucida, 22.5 cm. longa et 1-1.5 cm. diametro, apice basique adpresse puberula: cicatrix plana, 5—6 mm. diam < TUNG: Cheng Hsien, Shih-Wan-Ta Shan, tree 13 m. tall, scattered in mixed forest in dense shade, fruits light yellow, Dec. 19, 1943, S. H. Chun 3990, This new species is characterized by delicate, glabrous vegetative parts, thin, relatively small, entire, caudate-acuminate subconcolorous leaves, and a small cupule with the upper half of the annular rings glabrescent and the lower half densely tomentose. It is not likely to be confused with any of the described Chinese or Indo-Chinese species. Quercus disciformis Chun et Tsiang, sp. nov. Arbor vulgo 10-14 m. alta (usque ad 40 m. alta et 1 m. diametro fide Wang), cortice irregulariter fisso, ramulis hornotinis sordide flavescenti- pubescentibus annotinis vetustioribusque glabrescentibus atro-corticatis minute striatis. Folia ut videtur per duos annos persistentia subcoriacea in ambitu magnitudineque satis variabilia elliptico-oblonga obovato-ob- longa, vel plus minusve peacune. minima 6 cm. longa et 2.5 cm. lata, maxima 13 cm, longa et 4.5 cm. lata, basi obtusa late cuneata vel subro- tunda, apice in caudam angustam ec acutam 1-1.5 cm. longam + subito attenuata margine supra basin vel dimidio superiore sinuato-dentata dentibus grossis incurvis calloso-apiculatis, supra olivacea subtus pallidiora in sicco utrinque brunnescentia, costa supra leviter impressa subtus elevata 1947] CHUN & TSIANG, SPECIES NOVAE FAGACEARUM 325 nervis utrinsecus 11-13 tenuibus sensim curvatis in dentes excurrentibus venis subtransversis crebris subtus tantum prominulis, petiolis teretibus gracilibus calvis circiter 2 cm. longis. Gemmae oblongo-ovoideae acutae, 5-8 mm. longae, perulis ovatis obtusis rubro-brunneis, interioribus ad- pressis puberulis. Flores masculi desunt; feminei (juvenilibus ignotis) foliorum superiorum axillis 2-5 breviter spicati sessiles, pedunculo erecto tomentoso 1—2.5 cm. longo, ovario ovoideo, stigmatibus alte 3-fidis divari- catis glabris. Fructus biennes sessiles, in rhachidibus abbreviatis singuli vel raro gemini et inter se aversi cupula basi glandem tantum sustinens nullo modo amplectens plano-disciformis, 3-4 cm. diametro margine plano vel passim decurvo, intus dense molliter fulvo-sericeo-velutina, extus lamellis concentricis 8-9 tenuibus fulvo-tomentosis argute eroso-dentatis supremis integerrimis ornata; glans (e typo) 0.8 cm. alta, 3 cm. diametro, discoidea, apice impressa dense adpresse sericeo-tomentosa stylopodio valido conico truncato coronata; cicatrix concava glabra, circiter 2 cm. diametro. Kwanctunc: Hsin-I Hsien, Ling-Tung Pao, Chung-Tung, tree in mixed woods, leaves pale green beneath, Aug. 3, 1931, C. Wang 31087 (TYPE) ; same locality, tree 10 m. tall in edge of woods, fruit immature, Aug. 3, 1931, S. P. Ko 51729; Yun-Fou Hsien, without precise locality, Sept. 12, 1928, S. S. Sin No. 5160 (Bot. Inst. Herb. No. 136531); same district, Yun-Mu Shan, Lan-Kang, tree 40 m. tall, 1 m. diam. in community woods, cupule and acorn green, Nov. 17, 1934, C. Wang 37392. HAINAN: Pai-Sha Hsien, Fon-Ta Chuen, tree 9 m. tall, in dense woods, April 19, 1936, S. K. Lau 26333. Kwanosit: Shang-Hsi Hsien, Shi-Wan-Ta Shan, Tung-An Hsiang, tree 19 m, tall, in forested ravine side of stream, branchlets blackish brown, fruits pale brown, Feb. 3, 1944, S. H. Chun 4492. This new species is distinguished from all other species known to us by the thin flat broad involucre, velvety tomentose inside (excepting the basal scar) supporting, but not at all embracing, a slightly narrower more or less depressed acorn. The acorns as described are evidently immature, those collected later in the season are fully 1.5 cm. long and uniformly hemispheric. In most specimens the fruits are solitary but in others they are borne, back to back, in pairs from the same level on opposite sides of a short spike. Quercus Paohanii Chun et Tsiang, sp. nov. Arbor ramosissima usque ad 15 m. alta fructu et partibus junioribus utrinsecus 10-12 ante marginem furcatis petioli 4-6 mm. raro usque mm. longi. Stipulae caducissimae subulatae longe pilosae circiter 1 cm. 326 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVHI longae. Amenta mascula subpendula in axillis foliorum vel bractearum solitaria 5-6 raro-7 cm. longa rhachis parce pilosa angulata; flores singuli spe perianthii lobis 5 membranaceis brunneis ovatis albo-pilosis ciliatis 1 mm. longis; stamina exserta antheris ovoideis filamenta sub- cin He circiter 0.5 mm. longis. Flores feminei spicati in axillis foliorum solitarii, pedunculo 3-7 mm. longo, stigmata ad 1 mm. longa crassiuscula divaricata. Fructus biennes sessiles subglobosi vel oblongo- ovoidei; cupula semiglobosa 1 cm. diametro intus dense cinereo-tomentosa, squamis permultis imbricatis pti cinereo-tomentosis apicibus glandulo- sis; glans subglobosa 1 cm, longa et 6 mm. diametro apice puberula, stigmatibus subcapitatis sue. Kwantunc: Yu-Yuen, Hsien, Wu-Yang Shih, tree in woods, Nov. 19, 1940, P. H. Liang 532 (type) ; Tung-Shui Ts’ung, tree 5 m. tall, Jan. 8, 1941, P. H. Liang 83015. Hunan: I-chang Hsien, Li-Yuan-Pao, in light wands, April 8, 1942, S. H. Chun 2099; same locality, Nov. 10, 1943, S, H. Chun 2946; tree 12 m. tall, flowers greenish, March 25, 1943, S. H. Chun 2977. Kiancst: Without precise locality, April- May 1942, S. M. Mu 21138. This evergreen oak with unusually small acorns is very common in the vicinity of Li-Yuan Pao, I-chang Hsien, S. Hunan where a branch of the Botanical Institute was established throughout the war. It thrives from level-land up to an altitude of 1000 m., intermixed with deciduous species and conifers. Its range extends from the northern part of Yu-Yuan District in Kwangtung province to the foot of Mang Shan, the famous region in southern Hunan where virgin forests still exist. The edible acorns of this tree, colloquially called Chui-li, are sold in local markets, The wood is fine and durable, much used by the natives for making furniture and handles of various kinds of tools. It is especially good for preparing charcoal. This graceful oak which no doubt will become a common ornament in local and foreign gardens is named after the collector, Mr. Pao-Han Liang, Lecturer in Botany of this Institute. Quercus bella Chun et Tsiang, sp. nov. igi 9 m. alta dense ramosa confertifolia (e collectore) fere omnino glabra ramis ramulisque gracillimis minute lenticellatis ultimis angulatis fuscis eas subteretibus nigro-corticatis. Folia per duos annos persistentia crasse membranacea lanceolata absque petiolo 7-13 cm. longa 2.5—3.5 cm. lata, apice sensim acuteque acuminata, basi attenuata e medio primariis 14-16 ascendentibus fere rectis inferioribus versus marginem curvatis superioribus in dentibus exeuntibus atque costa elevata prominen- tibus oe transversis tenuissimis crebre connexis, petiolis tenuibus 5—10 mm. longis. Flos masculinus ignotus. Flores feminei ad apicem pedunculi gemini ae pedunculo in axillas foliorum superiorum solitario erecto quam petiolo breviore circ. 12 mm. longo; bracteae sub flore 3 valde inaequales, majora semi-circularis, omnes dorso pilosulae; perianthium urceolatum extus deorsum glabrum vertice annulatim adpresse denseque sericeo-pilosum; styli 4 robusti apice in stigmatem crassum capitatum 1947 | CHUN & TSIANG, SPECIES NOVAE FAGACEARUM 327 lobulatum abrupte dilatati. Fructus ut videtur triennis solitarius sessilis ; cupula pateriformis fere plana tantum tenuiter marginata 5 mm. alta 3 cm. lata glande paulo latior, intus adpresse fulvo-pilosa centro umbilico con- vexo circ. 16 mm. lato conspicue notata, extus atrata glabrescens concen- trice annulata annulis 7 laxis levibus planis inferioribus latis grosse irregu- lariter erosis marginalibus 2 vel 3 angustissimis multo proprioribus integris; glans depresse hemisphaerica basi cupulae solum adnata, 1.6—-1.8 cm. alta 2.5-2.8 cm. lata, atrocastanea adulta glabrata apice umbone conico circumcirca annulato obscure puberulo coronata; cicatrix margine gradatim curvata concava usque ad 16 mm. lata Kwanctunc: Fang-Cheng Hsien, Shi-Wan-Ta Shan, alt. 200-250 m., shrubby tree in sparsely wooded ravine along stream on moist sites, branchlets blackish, leaves deep green above, light reen beneath, dry fruits brown, Mar. 24, 1944, S. H, Chun 4772. Differs from Q. Dussaudii Hick. et A. Camus in the dentate leaves with more numerous lateral veins and the shallow cup only supporting, not at all enclosing the acorn. According to the collector, this oak is suitable for carpentry. The felled logs left im situ are used for the culture of edible mushrooms. This tree has ornamental qualities which recommend it for landscape planting. BoTANICAL INSTITUTE, SuN YATSEN UNIVERSITY, CaNnTON, CHINA. 328 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII NOTES ON CHINESE STERCULIACEAE Woon-YOUNG CHUN AND HstaANc-Hao HsukE Sterculia subracemosa Chun et Hsue, sp. nov Frutex parvus ramulis juvenilibus gracilibus sparse stellato- ager sicco striatis rubro-brunneis. Folia subcoriacea obovato-oblong cm. longa, 4—6.5 cm, lata, apice me usa cailivedlc basi ik vel cuneata, utraque pagina glabra, nervis primariis lateralibus utrinsecus 8-10 ascendentibus inter se anastomosantibus cum venulis utrinque valde prominentibus; petiolus 1.5—2.5 cm. longus, apice basique incrassatus, initio dense stellato-tomentosus deinde glabrescens. Inflorescentia sub- racemosa, plus minusve pendula, gracillima, usque ad 10 cm. longa, dénse rufo-stellato-tomentosa, pedicellis ultimis 8-10 mm. longis, articulatis, bracteolis lineari-lanceolatis, acutis, 2 mm. longis, tomentosis. Flores ignoti; perfecti albi, roseo-tincti; calyx 13 mm. longus 5-lobatus extus pubescens intus praeter partem superiorem loborum glaber, lobis ovato- lanceolatis tubo campanulato subaequantibus vel leviter longioribus, margine ciliolatis; antherae 17 (an semper? ) prope basin ovarii aggregatae, biloculares, connectivo incrassato; ovarium globosum circiter 1.5 mm longus pubescens, stigma 5-lobata. Fructus ignotus TUNG: Hsin-I District, pune hank: small shrub along stream, flowers white, tinged pink, March 21, 1932, C. Wang 31738 This species is allied to Sterculia ladcegetosie Roxb., differing in the obovate leaves and pale flowers. In certain characters it approaches S. hainanensis Merr. et Chun, but the obovate-oblong leaves and larger flowers are obviously different from those of the latter species. Sterculia micrantha Chun et Hsue, sp. nov Arbor 7 metralis, ramis rues medulla ampla alba spongiosa repletis; ramuli validi ery foliorum delapsorum cicatricibus magnis lenticellisque sparsis brunneis notati, siccitate rubro-brunnei valde ae longa, 3.5—6.5 cm. lata, apice obtusa vel abrupte acuta, basi late ee supra glabrescentia, subtus ad costam venasque sparse stellato-tomentosa, costa valida nervis primariis lateralibus utrinque 12, patenti- -ascendentibus subtus prominentibus; petiolus 3-8 cm. longus, sparsissime stellato- tomentosus; stipulae lineari-lanceolatae, 10-12 mm. longae, tomentosae. Inflorescentia paniculata, gracilis, usque ad 26 cm. longa, multiflora; pedicelli ultimi filiformes 3-4 mm. longi, articulati, leviter tomentosi: calyx albus, tenuis, campanulatus, 5-6 mm. longus, 5-lobatus, extus parce puberulus intus praeter lobos glaber sed manifeste papillosus lobis_ tri- * Type specimens listed in this paper are deposited in the Herbarium of the Botanical Institute, Sun Yatsen a Canton, China; duplicate types will be sent abroad as soon as conditions permit 1947] CHUN & HSUE, CHINESE STERCULIACEAE 329 aa ltophios um 3 mm. longum gracile, eae antherae 10, ad apicem androphori sessiles, biloculares, connectivo crasso. Flores ® : antherae circiter 15, circa basin ovarii aggregatae; ovarium cum gynophoro 1.2 mm. longo dense tomentosum, globosum, — 5-loculare; ae gynophoro aequilongus pubescens; stigma 5, globose. Fructus igno YuNNAN: King-Tung District, Tsu-Kai, alt. O m., tree 7 m. ne in light woods, leaves lustrous green above, pale green beneath, coe white, Oct. 5, 1933, Y. Tstang 12308. This very distinct species is not closely allied to any of the known Chinese species. is characterized by stout branchlets with broad spongy pith and relatively small white flowers in slender panicles. Sterculia Henryi Hemsley var. cuneata Chun et Hsue, var. nov. Planta habitu omnino typi et fortasse cum congruens sed _foliis deorsum manifeste attenuatis basi cuneatis aes haud_ rotundatis, floribus minoribus. r vel arbuscula 3-5 m. alta. Folia obovato-oblonga 15-21 cm. longa, c 5—6.5 cm. lata, apice plus minusve longe subitoque acuminata, infra medium versus = distincte attenuata, basi cuneata vel acuta; petiolus 4-8 cm. lon Racemi usque ad 7 cm. longi. Flores (e nota ad vivum) pallide rosei; calycis segmenta vix ultra 10 mm. longa. Yunnan: Ma-Kuan Hsien, N. W. of Ta-Wei Shan, alt. about 1000 m., in dense wooded Tavine along stream, shrub 3-5 m. tall, flowers pale red, March 15, 1940, Wang, Ko and Lau 10032; same locality, Ching-ya, alt. 800 m., tree on sparsely wooded slope, flowers pink, April 4, 1940, Wang, Ko and Lau 10496 (TYPE). Reevesia lofouensis Chun et Hsue, sp. nov. Arbor vel frutex, 4-13 m. alta, ramulis novellis dense stellato-tomentosis, fere levibus, siccitate albo-flavidis. Folia coriacea elliptico-oblonga 10-15 cm, longa, a 5—6 cm. lata, apice in acumen latum obtusum vel acutiusculum contracta, basi inaequilateraliter cuneata, utrinque glabra, margine undu- supra fere obsoletis; petiolus 1-3.5 cm. longus teres utrinque vix dilatatus, novellus stellate- tomentosus, mox glabrescens. Inflorescentia non satis evoluta paniculaia, terminalia, circiter 4 cm. longa lataque, nultaions dense ee stellato-tomentosa. Gemmae florales obovoideo- globosae, circ..2 1 divs... Capsula oblongo- pyriformis, 5-angulata, rugosa, dense stellato- een pedicellus circ. 2 cm. longus; semina 12 mm, longa 6 mm. lata, oe Bch alis oblongis obtusis hyalinis nitidis brun . latis. ee a= a ae eae -Hsou Tai, tree 13 m. tall in woods, branchlets green, leaves deep green above, pale green beneath, May 3, 1930, flowers in bud, N. K. Chun 40822 , same locality, shrub 4 m. tall, fruit green, brownish tinged, July 24, 1930, N. K. Chun 41342 (type). This species is related to Reevesia thyrsoidea Lindley and Reevesia longipetiolata Merr. et Chun, but differing from both in its densely stellate- tomentose brancnlets and peduncles, thicker, fewer veined, cuneate leaves, and tomentose petiole not dilated at both ends. 330 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXVIII Reevesia taiwanensis Chun et Hsue, nom. nov. Reevesia eemneneg Hayata Ic. Pl. Formos. 9: 8. 1919, non Sprague in Kew Bull. 1914: . 1914. ForMOSA: nee leg. Hayata et Sasaki s. n.. July 1912. Although the two specimens are not exactly comparable, the one collected by A. Henry on which Sprague’s species is based is in flower, and the other, described by Hayata, is in fruit, we are confident that two distinct species are involved judging by vegetative characters alone. Reevesia formosana Sprague is easily distinguished by the rounded to cordatulate leaf-bases. Helicteres viscida Blume Bijdr. 1: 79. 1825; Kurz Fl. Brit. Burma 1: . 1877; ry Fl. Cochinch. pl. 209A. 1888-91; te p. in pennant Fl. Gen. se Chine 1; 1911; Ridley Fl. Malay Peninsula 1: 281 oe in Mem. Sci. Soc. ven 1: 168, 1924; Craib Fl. Siam. Enum. 1: ie 5 Hainan: Po-Ting District, vicinity of Tung-Ch on in ravine, subshrub, flowers white, Dec. 4, 1936, L. Tang 430; same locality, Oct. 10, 1936, S. K. Lau 28343; Ting-An District, Dec. 4, 1933, C. Wane 35181; Ling-Shui District, Seven-Finger Mountain, Jan. 11, 1934, C. Wang 36352; Ling-Shui District, Tung-Chiah, Nov. 18, 1932, Tso & Chun 44283. DistRIBUTION: Java, Malay Peninsula, Burma, Indo-China, Siam, China: Yunnan (fide Gagnepain, l.c.). New to Hainan Eriolaena spectabilis (DC.) Planchon ex Mast. in Hook.f. Fl. Brit. Ind. 1: 371. 1874. Wallichia spectabilis DC. Mem. Mus. 10: 104. t.6. 1823 Microlaena spectabilis Wall. List No. 1173, 1828 Gumsia tiliacea Wall. List No. 1173D, 1828. G. chloroxylon Ham. in Wall. List p.157. 1173E. 1828. Eriolaena sp. Chung in Mem. Sci. Soc. China 1: 167. 1924. Yu : Szemao, A. Henry 12506, ex Herb. Mi-Le Hsien, Dec. 2, 1932, H. T. "Tead 51393, “Kw wancsi: Hsi-Lin District, ge ay 800 m., tree 7 m. high, bark gray, fruits silvery, dehiscent, July 2, 1936, W. H. Soo 67905 DistRIBUTION: India, Central Himalaya; China: Yunnan, Kwangsi. Henry’s 12506 which the senior author examined at Kew undoubtedly belongs to this species. Presumably this is the first authentic record of the species for China. BoTtaNicaL INSTITUTE, Sun Yatsen UNIversity, CANTON, CHINA. 1947] MERRILL & PERRY, KAJEWSKIELLA 331 KAJEWSKIELLA, A NEW RUBIACEOUS GENUS FROM THE SOLOMON ISLANDS E. D. Merritt AND L. M. PERRY With one plate KAJEWSKIELLA GEN. NOV. Calycis tubus cylindricus, gracilis; limbi cupularis lobi 5, lineares, elongati, persistentes. Corolla tubularis, in alabastro oblongo- sactniey extus dense villosa, tubo intus glabro; limbi lobi 5, erecti, valva Stamina 5, basin paullo supra tubi corollae inserta, filamentis oleate glabris; antherae dorso affixae, non versatiles, paullo exsertae, lineares, obtusae. Discus annularis. Ovarium 2-loculare; stylus teres, exsertus, stigmate punctiformi; ovula numerosa, placentis septo adnatis inserta. Capsula erecta, teres, siliquaeformis, septicide 2-valvis, vetusta epicarpio (demum saepe fibroso) ab endocarpio cartilagineo margine placentifero vel nudo secedente, polysperma. Semina numerosa, adscendentia, minuta, in sicco angulata, testa reticulato- irs albumine carnoso; embryo clavato-oblongus. Frutices ramosi, pubescentes. Folia opposita, petio- lata, membranacea, lanceolata vel elliptico-lanceolata. Stipulae inter- petiolares, triangulari-ovatae, subpersistentes. _Racemi simplices, axillares, pauciflori, sub anthesin abbreviati, in. fructu longissime pedunculati. Flores majusculi, pubescentes. Kajewskiella trichantha sp. nov. Frutex usque ad 5 m. altus; ramulis teretibus vel internodiis superioribus ia aa villosulis; uh membranaceis, lanceolatis vel elliptico-lanceo- latis, 12—25 cm. longis, 3 .5—-8 cm. latis, utrinque sensim ee be attenuatis in petiolum decurrentibus, apice acute acuminatis, acumin 1.5 cm. longo, margine integris, supra costa villosula lamina sacs et sparsim pilosa, subtus novellis dense adpresse ferrugineo-villosis vetustiori- bus costa venisque dense et lamina sparsim adpresse villosis, nervis lateralibus utrinsecus 15—20 oblique arcuatis prope marginem adscendenti- bus, utrinque perspicuis, venulis manifestis vel inconspicuis; petiolo 1. longo, gracili, villoso; stipulis triangulari-ovatis, 1-1.5 cm longis, 5-7 mm. latis, longiuscule acuminatis, costa et basin versus -+ axillaribus, ad anthesin 5—7 cm. longis (floribus inclusis), in fructu 6—17 cm. longis, omnino villosulis; floribus 3—7, pedicellatis, pedicellis brevibus basi bracteatis; calycis limbo cupulari, 2. 5 mm. longo, lobis linearibus, acutiusculis, 5-7 mm. longis, 1 mm. latis: ae tubulari vel in alabastro oblongo-clavata, tubo 2.5-3 cm. longo, 6-7 mm. diametro, intus glabro, lobis circiter 7 mm. longis, acutis, erectis vel ie cadentinns intus glabris; staminibus + 5—7 mm. supra basin corollae tubi insertis, paullo exsertis, 332 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII filamentis compressis, glabris, antheris linearibus, 4 mm. longis, dorso affixis, non versatilibus, rimis introrsum dehiscentibus; disco annulari; stvlo tereti, 3.9 cm. longo, stigmate punctiformi, non papilloso; ovario onga, 2 mm. diametro, calyce persistente coronata; seminibus numerosis, minutis, reticulato- -punctatis. SOLOMON IsLANps: Bougainville: Kupei Gold Field, i 1659, 16067, ne 1930, alt. 950 and 900 m., rain forest (plant up to 5 m, , but often found m smaller; buds 4.5 cm. long, 1.3 cm. diameter, beautiful ira semi- iponeeoac nai same locality, Kajewski 1734 (Type), April 1930, alt. 1000 m., rain forest (flowers pink, covered with white hairs). While working over the Rubiaceae of Papuasia as represented in the Richard Archbold Expeditions collections from New Guinea, and the Brass and Kajewski collections from the Solomon Islands, we found the above cited specimens which we were unable to name to genus, although they clearly belong in the family. For this material we here propose the new genus Aajewskiella. At present we are unable to suggest an affinity for it. According to the classification of the family in Engler and Prantl, Pflanzenfam. IV. 4: 16. 1891, it seems to belong to the Oldenlandieae. In its ferrugineous-villous indumentum, in its stipules, in the form and texture of the leaves as well as in the venation, and in the minute reticulate- punctate seeds, Kajewskiella strongly suggests Xanthophytum Reinw., but in its technical characters it is not allied to that genus. The very slender terete capsule is somewhat like an abbreviated fruit of Dolicholobium A. Gray; both have septicidal dehiscence, and the exocarp becomes more or less shredded in age as in that genus, and in Bikkia Reinw.; the valves of the fruit persist as in the latter genus, and the placentae are either broken free or attached along the margins of the valves. The similarity of these characters, however, is probably only superficial. The fairly large tubular or very narrowly elongate obconical corolla entirely glabrous within, the glabrous stamens attached a little above the base of the corolla (about where it emerges from the calyx), the non-papillate stigma, and the short slender terete capsule seem to offer a combination of characters which we believe worthy of generic recognition. EXPLANATION OF PLATE PiaTeE I Fic. a. Flowering branchlet +; b. Flower, X 1; c. Flower laid open, x 1; d. Fruit, 1; e. Old fruit er ee x 1; f. aa of placenta to show attachment of ovules, X 10; g. Seed, x 10. ARNOLD ARBORETUM, Harvarp UNIVERSITY. Jour. Arnotp Ars. Vou. XXVIII PiatE I Su %, * ‘ \ \ < SS we \ \ \ \ \ ' x *%\ \ \ Ss \ Si Ps Sicel isha a a oe hts ore een Sas ate oe EP ‘ ‘. r. SS ay, i E — — ry f Sa y ii Be ie i p a Pe eh a Ee GRSeSe ESS f Ceaven Ww Dien” KAJEWSKIELLA TRICHANTHA MERRILL & PERRY 1947] SMITH & GLEASON, PLANTS FROM ST. LUCIA 333 TWO NEW FLOWERING PLANTS FROM ST. LUCIA A. C. SmitH AND H. A. GLEASON AMONG the West Indian specimens which Dr. John S. Beard, formerly of the Forest Department, Trinidad, has sent to the Arnold Arboretum in recent years, the two species from St. Lucia discussed below appear to be hitherto undescribed. ROSACEAE Licania oligantha A. C. Smith, sp. nov Arbor 5-15 m. alta, ramulis gracilibus, hornotinis complanatis sericeo- puberulis, annotinis teretibus glabrescentibus lenticellatis; stipulis lineari- lanceolatis 4-8 mm: longis sericeis mox caducis; petiolis subteretibus rugulosis 2-5 mm. longis primo pallide sericeis' mox glabris; Tenia chartaceis utrinque in sicco olivaceis oblongo-ellipticis, (3.5—) 5 cm, longis, (1.8—) 2.5-4.5 cm. latis, basi obtusis vel acutis et in eee breviter decurrentibus, in apicem calloso-apiculatum cuspidatis, margine integris et leviter recurvatis, juventute subtus et costa supra sparse strigillosis mox glabrescentibus, costa supra leviter elevata subtus promi- nente, nervis secundariis utrinsecus 6-10 erecto-patentibus leviter curvatis anastomosantibus et rete venularum copioso utrinque prominulis; in- florescentiis plerumque axillaribus interdum terminalibus compacte cy- moso-paniculatis sub anthesi 1—2.5 cm. longis et plerumque 1 —5-floris, pedunculo leviter complanato brevissimo et ramulis_ sericeo-puberulis, bracteis oblongis puberulis 3—3.5 mm. longis, ramulis lateralibus paucis 2 sitis membranaceis ovatis 2—2.5 mm. longis minute sericeis; floribus subsessilibus, ior ad 1.5 mm. longis; calyce cupuliformi sub anthesi t apice diametro, extus arcte sericeo, tubo 2—2.5 mm. longo basi subacuto ite copiose et br reviter sericeo, limbo erecto-patente pro- funde 5-lobato, lobis papyraceis deltoideis 3-3.5 > 2-3 mm. apice acutis intus puberulis basi anguste imbricatis; petalis 5 membranaceis obovato- vel ovato-ellipticis, 2-3 mm. longis, 1.3-1.5 mm. latis, basi rotundatis, apice obtusis, intus hispidulis extus puberulis mox glabrescentibus et caducis; staminibus 15—20 sub anthesi leviter exsertis, filamentis ligulatis 2.5-3.5 mm. longis utrinque copiose pallido-hispidulis superne angustatis glabrisque, antheris ellipsoideis 0.5—0.6 mm, longis; ovario ellipsoideo copiose hispidulo, stylo tereti circiter 5 mm. longo praeter apicem glabrum copiose hispidulo, stigmate minute peltato. St. Lucia: Quilesse, alt. 300 , J. S. Beard 492 (type, Arnold Arb.), June 5, 1945 (tree 15 m. high, rare in uae pte rey of rain-forest; trunk 50 cm. diam., fluted; flowers creamy white; fruit blue, attractive to birds) ; La Sorciére, alt. 700 m., J. S. Beard 488, June 1, 1945 (tree 5 m. high, in elfin woodland; trunk 30 cm. gene: flowers creamy white). 334 JOURNAL OF THE ARNOLD ARBORETUM [voL. XXVIII Licania oligantha is characterized by its very compact few-flowered in- florescences and its concolorous essentially glabrous leaf-blades. From L. leucosepala Griseb., originally described from Guadeloupe but also recorded from Dominica, St. Vincent, and St. Lucia (by R. O. Williams, Fl. Trinidad & Tobago 1: 315. 1932, as Moquilea leucosepala), the new species differs not only in its inflorescence, but also in its slightly smaller leaf-blades with fewer secondaries and its obscurely sericeous rather than cano-tomentose calyx. In referring plants of this alliance to Licanta rather than Mogquilea, | follow the generally accepted reduction of the latter genus. Concerning this species Dr. Beard writes: “It is a small tree with a trunk up to 30 cm. in diameter; the wood is hard and of no economic value. It is almost confined to the exposure sub-type of the rain-forest in the interior of St. Lucia, occurring on the crests of all but the highest ridges and Pitons. I have noted it on La Sorciére, Piton Flore, Piton St. Esprit, Grand Magazin, Morne Tabac, and the slopes of Morne Gimie, at elevations of 1600 to 2200 ft. [Note: The type-collection was found at a lower elevation]. The species is not found in the elfin woodland at the top of Morne Gimie (3000 ft.), and it is very rare in the normal rain-forest of lower elevations. In its particular association it is one of the principal dominants, forming approximately 30% of the stand. The associate dominants are Micropholis chrysophylloides, Myrcia sp. (Beard no. #478), and Tovomita Plumieri. Together with a number of species of lesser abundance, these plants form a low forest of small trees some 30 ft. high, a reduced type due to exposure to wind. In the other islands the principal dominants of the equivalent forest type are Richeria grandis, Micropholis chrysophylloides, and Podocarpus coriaceus. The first of these has not been found by me on St. Lucia, and the last is very rare, found only on Morne Tabac. The Licania oligantha—Micropholis ihidehd Sie data ta association of St. Lucia is, therefore, an interesting special type. MELASTOMATACEAE Miconia luciana Gleason, sp. nov. Sect. Octomeris. Arbor 8 m. alta, ramis juvenilibus sparsissime furfuraceis mox glabris. Petioli usque 7.5 cm. longi, sparse furfuracei praesertim supra atque sparse glanduloso- oe Laminae tenues virides, ovato-lanceolatae vel fere ellipticae, maxima qua suppetit 20 cm. longa 10 cm. lata, acuminatae crenato-serratae, garner 5-nerviae, utrinque glabrae. Inflorescentia pedunculata, late pyramidalis, ca. 1 dm. longa, trichotoma. Flores 5-meri, omnes sessiles, saepe solitarii ad apices ramulorum. Fyaaith ium obconi- cum, 2.7 mm. longum ad torum, minutissime furfuraceum. Calycis tubus 0.8-0.9 mm. longus, truncatus, ad marginem fere scariosus; sepala ob- soleta; dentes exteriores conici, ad tubum adnati, ultra tubum 0.3 mm. projecti. Petala rosea, obovato- oblonga, 5.5-6 mm. longa a lata, fere equilatera, vix retusa. Stamina isomorpha; filamenta oa 3.3 mm. longa; antherae lineares, 3.2 mm. longae, 2-loculares, poro 1947 | SMITH & GLEASON, PLANTS FROM ST. LUCIA 335 terminali dehiscentes; a infra thecas 0.2 mm. productum, simplex, leviter dilatat m. OQOvarium fere inferum, 4-loculare; Rae rectus, glaber, 6.7 mm. iepais inferne dilatatus; stigma eae ucIA: Quilesse, alt. 300 m., J. S. Beard 477 (type in herb. N. Y. Be i dupl. ‘Armold Arb.), May 23, 1945 (tree 8 m. high, in understorey of rain- aie trunk 10 cm. diam.; flowers pink; local name: bois cendre). From its general aspect and the nature of the inflorescence and the structure of the flowers, M7. luciana appears related to the Puerto Rican M.Thomasiana DC. It differs from that species in its thinner leaves and various other structural characters. Of Miconia luciana Dr. Beard writes: ‘It is a small tree, frequently only 3 m. high, with rather conspicuous pinkish flowers. I have found it occasionally in the understorey of the Dacryodes—Sloanea rain-forest throughout St. Lucia.” ARNOLD ARBORETUM, HARVARD UNIVERSITY, and New York BotTANIcAL GARDEN. 336 JOURNAL OF THE ARNOLD ARBORETUM [voL. XXVIII ASTRAGALUS IN ARGENTINA, BOLIVIA AND CHILE IvAN M. JOHNSTON LY IN SouTH AMERICA does the great genus Astragalus have an extensive development in the southern hemisphere. There are ten species reported from the mountains and plateaus of tropical Africa. In the temperate southern parts of that continent the genus is represented only by one species in Natal. In South America, however, there are nearly ninety species. The genus is represented along the cordilleras from Ecuador to the Straits of Magellan, and has two important centers of speciation, one in the high country of northwestern Argentina and adjacent Bolivia and another in adjoining Chile and Argentina along the cordilleras south of the peak of Aconcagua. Though the present paper treats the species known from south of Peru, all the South American species of Astragalus were studied during its preparation. Accounts of the two Ecuadorean species, and of seven of the Peruvian species have already been published, Jour. Arnold Arb. 19: 88-96 (1938). The present paper treats the remaining seventy-seven species of the genus known to me from South America. In general structure and appearance the South American Astragali are rather similar to those found in North America. Their relations are with the North American and not with the Old World members of the genus. They are probably the products of an ancient southern migration and subsequent differentiation of various types originating in North America where the genus has one of its major evolutionary centers. In North America, Astragalus is represented from Alaska and Baffinland south to northern Guatemala, with most of its more than three hundred species concentrated in western United States. The areas occupied by North and South American species are, accordingly, separated by Colombia and nearly all of Central America, and their major areas of speciation by the great distance between Argentina and the United States. The close relationship between North American species and others in Argentina and Chile is frequently very clear. Indeed, some of the southern plants seem to be more closely related to the northern ones than they are to any found on their own continent. None, however, seems to be specifically identical with North American plants. Though there are many close similarities between northern and southern species, evolution in South America has not closely paralleled that in North America. This becomes very clear if an attempt is made to apply to the South American species the system of segregate genera set up for the North American species by Rydberg, No. Am. FI]. 24: 251-462 (1929), and Am. Jour. Bot. 15: 584-595 (1928), 16: 197-206 (1929), and 17: 231-238 (1930). Some of the South American species fall readily into the Rydbergian 1947 | JOHNSTON, ASTRAGALUS 337 segregates, but many straddle the arbitrary lines between these so-called genera, and others simply do not fit into any of them No satisfactory subgeneric classification of the South American species has been worked out. Rydberg’s classification seems artificial at many places and furthermore is to be condemned as provincial, since it is based almost exclusively on species of one region and without full consideration of those in other parts of the world. It would have to be modified con- siderably, as well as amplified by the addition of comparable new groups, before it could accommodate the South American species satisfactorily. That, however, is a task that should be left to a monographer with a wide acquaintance of the genus as developed in all parts of the world. Astragalus has only one clean line of cleavage in South America. This breaks the genus into two natural groups probably worthy of generic recognition. One, the section Phaca, with fifteen species, is readily distinguished by its well separated, distinct, herbaceous stipules. The other, the section Euastragalus, has more or less united, membranous, ochraceous stipules and includes over seventy species. In the present paper the synonymy given for the species is complete except for certain names published by Otto Kuntze, Rev. Gen. 2: 940-949 (1891), and 32:73 (1898). Substituting the name Tragacantha for Astragalus, Kuntze published hundreds of new combinations under the former invalid name. It has seemed a waste of space to reprint the scores and scores of useless binomials he applied to South American species. Only when he coined new specific epithets for South American species, are Kuntze’s names of any taxonomic significance and only the latter are listed in the synonymy given in this paper. Among the large number of specimens studied, only about half are cited. Under well understood species, of which I have seen numerous specimens, I have cited only selected collections to show geographic range. Only under new, rare, or poorly understood species has all the examined material been cited. The particular herbaria in which the examined specimens are located are indicated by abbreviations, in parentheses, following the collector’s name During the preparation of this paper the specimens in many herbaria have been examined. he basic herbarium and library work was carried on at the Gray Herbarium. During a visit to Europe the genus was studied at Kew, London, Hamburg, Berlin, Munich, Geneva, and Paris, and loans of selected material from these botanical centers were later studied at the Gray Herbarium in conjunction with loans of South American Astragali from Stockholm, Upsala, New York, Washington, and Chicago. From South America much interesting material has been received on loan and for identification. Dr. Angel Cabrera of La Plata, sent specimens and arranged the loan of Spegazzini’s Astragalus. To Prof. L. R. Parodi, Dr. Arturo Burkart, Prof. Alberto Castellanos, Sr. A. Ruiz Leal, and Sr. Juan V. Monticelli I am indebted for many interest- ing specimens. Prof. Marcial R. Espinosa sent me some of the collections 338 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII and very obligingly loaned me fragments of certain critical specimens in the Philippi Herbarium. Other very interesting and useful Chilean collec- tions have been received from Sr. Gualterio Looser, Sr. Gilberto Montero, and Dr. C. Grandjot. I have had help from many sources and I am grateful to the many botanical friends in Europe and America who have aided in this work on the genus Astragalus. The abbreviations for the herbaria cited in this paper are as follows: BD=—Herb. Bot. Mus., Berlin-Dahlem, Germany; Behn=Herb. Dr. Behn, Valparaiso, Chile; BM=Herb. British Mus. Nat. Hist., London, England: Boiss=Boissier Herb., Geneva, Switzerland; Burkart—Herb. Inst. Bot. Darwinion. Buenos Aires, eres Calif—Herb. Univ. Cali- fornia, Berkeley, California; Del—Delessert Herb., Geneva, Switzerland; —=Herb. Chicago Nat. Hist Mus. | Field Mus. ], Chicago, Illinois; G—Gray Herb., Harvard Univ., Cambridge, Massachusetts; Gottenburg— Herb. Hort. Bot. Gottenburg, Gottenburg, Sweden; Grandjot—Herb. Dr. C. Grandjot, Santiago, Chile; Hamb=—Herb. Inst. Allg. Bot., Hamburg. Germany; K=Herb. Roy. Bot. Gard., Kew-Surrey, England: LP =Herb. Mus. La Plata, La Plata, Argentina; antes Herb. Bot. Mus., Munich, Germany ; NY=He rb. New York Bot. Gard., New York, New York; Paris—Herb. Mus. Hist. Nat., Paris, France; Parodi—Herb. Dr. L. R. Parodi, Buenos Aires, Argentina; Santiago—Herb. Mus. Nac., Santiago, Chile; Speg—Herb. Mus. La Plata, La Plata, Argentina; Stock—Herb. Riksmus., Stockholm, Sweden; US=United States Nat. Herb., Washing- ton, D. Cc. Upsala Herb. Upsala Univ., Upsala, Sweden. KEY TO SPECIES A. Stipules herbaceous, not united, separated by a quarter of the circumference of the ste em. Legumes compressed, scarcely if at all inflated, rather closely ee the lon 1. seeds, 5-13 mi, JONG. . «6 disse esi oo Ss 0b 0 SA sooo Oe ale'Se wads _ Chamissonis Legumes distinctly inflated, 8-40 mm. long; seeds more numerous ne at maturit rattling about within the mature po Pods evidently ie pisos he stipe 3- 10 mm. long; pedicels becoming 5—7 mm. long; standard y Fruit distinctly ioe aye Saw went sd araa ig ween area Aiea ate 648 2. A. Pissisi Fruit strict or ascendin Pod oadest ace middle; leaves flat; flowering inflorescence A, Gan eee CONSOxctis inet a pale sip slo aini ole wenn 3 Edmonstonei Pods beshetnch below middle; eka conduplicate; flowering inflore cence becoming loose and elongate............... 4. A. p epee? Pods sessile or ae a — es ss than 3 mm. ie a small, 8-15 . long, usually strict or ascending. Pedicels 3-5 mm yee at anthesis, becoming 5—7 mm. long in fruit; standard sallow: leaflets frequently in iictlets ees ona aaa eae os pivewinae ea ahes de 4515 4 aes pa Riese as os eee aie 'a'sa ects 5. A. te Pedicels 0.5-1.5 mm. long at anthesis, becoming 1-2.5 mm. long in fruit; standard purple or violet; leaflets in oe bed pairs. Herbage glabrous; leaflets thickish; plant montane............ he glace apeciee ewe Uw Age dudslee gale gee bee een ad es A. Looserii Herbage on leaflets not very thick nor pie plants of low AMGQUGES 0s ae sindy doses ode oie dive ones eee . A. Berterianus 1947] JOHNSTON, ASTRAGALUS 339 Legumes medium to large, 15-40 mm g ts annual or at least short- lived: corolla pinkish; arid coastal regions of northern Chile. Herbage pallid with a dense velvety indument...8. A. paposanus Herbage green or somewhat cinereous, strigose............... Aa ain a etre oo a, os bok oe ee ee 9. A. coquimbensis Plant strongly perennial, mon g s strictly erect, oe tough and rigid, opaque, lustrous ce AGsMACULILY: 6355) .Acducsetecee ave een ars 10, At 5, eee eee spreading and becoming pendulous, ena or ellipso a , translucent, pale or mottled at maturi ae and fruit glabrous Leaves glabrous or with the hairs eran to the gins and midrib of the leaflets... .11. ielgpickes Leaves sparsely but evidently siriuuee Meiudsisa dest fies Z sf die asa ib 8 fee g 0 he A eee valerianensis Ovary distinctly strigose and fruit at least sparingly so. Corolla pinkish; flowers dense, 15-— pods 15-20 (—25) mm. long, numerous and Saad in globose clusters; northern Chile............... 13. A. vagus Corolla with at least the rae conspicuously purple or violet; pods not in dense clusters. Leaflets smooth, ee with closely appressed silky hairs; pods 2-4.5 mm. long............. Jad Raydo tad edd shee ee Darumbium Leaflets somewhat shaggy ae short ascending to spreading hairs; pods 2-3 cm. long............ Wire FO Haw Shona sa a Sea 15. A. monticola AA. Stipules mebranous, more or less united into an amplexicaul sheath B. Fruit with the sutures not inflexed nor recessive, or with the lower one more ae than the upper. Peduncles 2-15 cm. long, equalling or surpassing the subtending leaf; flowers rae Patagonia and the andes of middle and southern Chile and Argen oe ‘ot compressed about the margin of the valves, the sutures not sali Plant Sea short villous, the hairs more or less curly and spreading ; obcordate, plicate; pods short villous...16. A. vesiculosus ee pees se to glabrous, hairs appressed ; eaters acute to weakly retuse, on to lanceolate, revolute; pods mottled, prevailingly glabr Lea ets more or less strigose above; wings conspicuously sur- pass the k G@Ldiscc-ncc sestareee tere sastanens 30° ee 0s 17. A. Arnottianus ree Se above; wings shorter than the keel " of about Equal) lemetiiestn.ras aloe eae oa saosin 18. A. palenae Fruit compressed at least about the margin of the valves, ee and leathery, the sutures prominent, elevated on the compressed rim of the fruit Pod somewhat compressed dorso- Bieesierek more or less sean in cross-section, lower suture most ab ominent ; corolla ca. 12 m long, petals all broad; eastern ges Fe sae uae 19. A. aorta Pod flattened laterally, ee to aan in cross-section, sutures equally en corolla 15-20 mm. long, the petals narrow and elon Plant los pallid; calyx with pale hairs; fruit tr wider than lon . [lini Oe OE a 8) See we ele 6 wa 6 Wie abe 6 6.48 6. 8:9 © 6 Oe 340 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Plant short strigose or glabrous; calyx usually with black hairs; fruit valves longer than broad. Leaflets elliptic to oblanceolate, apex rounded or emarginate Ue uis Seaweed keersaNewaeesss 21 Cruckshanksil Leaflets iene to linear, apex elongate, 1-5 dm. inne tipicen oe well developed, on scattered leaves; peduncles obviously axil- V.c5 a nedee wae phen unk ep ween 22. A. Amunategui Stems short, usually 0.5-5 cm. long, slender, with very much shortened internodes; leaves appearing to be tufted and the peduncle scapose...23. A Peduncles se short, 0-1.5 c ing Argentin vary auaie fruit glabrous, 10-15 mm. long; plant orate with green 2—6-jugate leaflets; peduncles is flowered, becom long; northwestern Argentina.................2.000- 24, ei Burkartii Ovary strigose; fruit pallid, strigose, 4-7 mm. long; plants with dense appressed pale indument; leaflets 5—9-jugate. Plant densely caespitose, the leaf- et. usually springing from an ough flavescent stipules; fruit alana soli- tary, remaining hidden among stipular en mm Ives 2-3 mm. broad; northern Argentin Plant with elongating leafy stems, 1- ae cm. evident, 6—7 mm. long, valves 5-6 mm, wid Fruit usually solitary in the lea oa cle exposed along oe older parts of the coarse trailing stems; northern ina and adjacent Chile 26. A sults iprteeh to many on evident axillary peduncles ; peduncles mm. lon patagonicus m. long, conspicuously shorter than the subtend- —7- flowered ; high altitudes in Bolivia and northern Chile and ong, ores a A. cryoticns lon fruit becoming ong. Plant coarse; leaflets lanceolate, 3-10 mm. long, apex acute; a 14 mm. long; valves broadest below middle; 10) tg +: Seen ee oe ee 27. A, se Plant slender; leaflets oblong or elliptic, 3-4 mm. long, a ro obtuse; corolla ca. 6 mm. long; ue broadest at or shove the middle; Andes of Coquimbo and Sa MANS oy esacean a siens owas eee eee 28. A. bellus BB. Fruit a the lower suture inflexed, or at least less prominent than the upper C, Fruit distinctly lesa sa or reflexed. —— of Arg eaflets oe 7 deltoid, broadest at apex and very deeply and conspicuously emarginate; fruit 25-30 mm. long, falcate, upper dge convex in outline.......... 0.0... eee eee 29. A. tehuelches Leaflets linear to elliptic, broadest near middle, the apex acute to inconspicuously retuse; fruit with upper edge straight or concave in outline, except in A. Ruiz-Lealii. Plant abundantly appressed villous, pallid, caespitose, the pods apparently short and not much surpassing the calyx; Chubut. 30. A _ Earn short strigose, cinereous g d and slender, loosely branched; pods elongate. bedi distinctly convex, lower suture depressed in a hallow narrow groove; corolla ochroleucous and merely sp spotted with blue or purple. 1947] JOHNSTON, ASTRAGALUS 341 Valves swollen on either side of the aie suture, the suture appearing to be inflexed, s becoming pinnae or somewhat woody; pane a northern Ar eentilia’.::..ce cs tciser acts ocane este atl ets 31. A. distinens Valves 7 ee on either side of the dorsal suture, the suture prominent its whole length, pods less mee and more elongate; northern aera nia. er rn ay OOS ro oe . A: Bergii Fruit angular and more or less laterally oma in cross section more or less sagittate or cordate to triangular, valves plane on the sides, lower suture inflexed. Corolla 8-14 mm. long, bright colored; ovary and fruit glabrous. Plant annual; fruit more than 6 ae as long as b CRNO.ce cee eae 33: ioe: oad; Ri Plant pevontel: fruit 2-4 times as on as broad. Pod 10-18 mm. long, strongly and acutely triangular, lateral angles much more promi- th nent t e lower suture; valves readily and completely separating at aturity ; Mendoza and Cordoba.. . carinatus Pod 5-9 mm. long, lateral pape ee and less an than the lower suture; valves much tougher than in the last, very cas Corolla 4-7 mm. long, coloring pale. Inflorescence wit numerous’ rather crowded flowers; stems hi aii 1-4 dm. long; san 3 ester ge . A, Joergen iioraeenee i. few (2-7) distinctly connie flowers; stems 1.5 dm. long or less. Ovary glabrous; fruit oblong, valves 3-4 times as long as broad, upper suture straight concave in lateral outline; leaflets glabrous above; Rio Negro and Chubut sel aise ieegtacs 9h od SO 37. A. pauranthus Ovary strigose; fruit falcate, bees 6-10 times as long as broad, upper suture convex in lateral outline; leaflets Ce above...... sidhslacd us’ ood aaa Oat ee 38. A. Ruiz. Leali DD. Plants of Chile ings much aenece than the keel. Fruit and ovary glabrous; plant prostrate, green; a 6-9- JUSAates CaUlINy ss .04 sa 40 deleree ee eran 9. Monteroi ae aod es strigose ; ca plak te or erect; stats 6-16- ate; ncepcion — oquimbo Wings an eae the Leaflets 3; plant densely ‘iilky strigose....41. A. valparadisiensis Leaflets 12-31. Plant prostrate, stems slender and widely spreading; edu Mensa. & atus a . Plant eee or decumbent; peduncles usually much sur- assing the subtending leaf; ae smaller, standard usually less than 12 mm. : 32 JOURNAL OF THE ARNOLD ARBORETLM [ VOL, XXVIII Pod curved, lateral outline lunate with upper edge concave and the lower edge strongly convex, their sweeping curves continuing into the style, hence apex of pod gradually attenuate; Coquimbo to Anto- a Fruit loosely and abundantly tomentose; plant perennial, stems 2-6 dm. long; rea 8-14- jug Coquimbo Fruit strigose to subglabrous; plant prevail ual; stems 1 dm. long; leaflets 5- jugate; Atacama and Antofagasta............ ee eee era ae 44. A. Dodti Pod eerie in lateral outline more or less oblong, with the er edge straight or convex and the lower ce more or less convex; style terminating the ruit. Fruit laterally much compressed, the lower suture pex; cordiller i? ee eee A. Germaini CC. Fruit strict or ascending, not pendulous. E. with elongating stems, not caespitose nor pulvinat Plant coarse; stipules very large and loosely sheathing. . long; leaflets usually flat; fruit with well developed false septum, Fruit evidently villous- otinae corolla over 1 cm. long; plant usually erect; flowers usually projected beyond ‘ihe leaves.. sb Re waa da paw alskeea.4 ako adiersd ae Kea e BA 47. A. Garbancill Fruit ruc tae strigose or glabrous; corolla 6—9 mm. lon lant spreading; flowers and fruit much surpassed on leav a gray-green, evidently strigose; widely distributed.. jain eo aes eG ao so aa eed ee 8 Ka 48. Ala dreesnbensis Plant bright green, nearly glabrous, rare...49. A. cryptanthus Plant slender; stipules small and not very conspicuous, usually less n 5 mm. long; leaflets usually folded or revolute. Plants of Patagonia and the cordilleras south of Cerro Aconcagua (lat. 33° S). Fruit slender and elongate, 1-2 cm. long, 2-2.5 mm. in diameter, with a nearly complete 4 septum and hence, practically 2-celled............... A. Sanctae-Crucis Fruit short and stout, 5-15 mm. long, ae half as wide as long, 1-celled, the false septum absent or weakly Leaves linear-lanceolate to narrowly oblong, apex acute to rounded; fruit becoming tough and rigid; Chile at the west base of the cordilleras............... a SA Ki Ee whe dea eka ace eae 1. A. orthocarpus Leaves obcordate, strongly retuse at apex; ae charta- srt see leaflets glabrous or with only a few scatter eel mie the midrib; south of esate ee Tal .....52. A. Domevkoanus 1647] JOHNSTON, ASTRAGALUS 343 Herbage cinereous, younger leaflets evidently stri- gose beneath; Mendoza... 33. 4. complicatus Plants from north of Cerro rene wit ith the exception of _ triflorus, a coastal annual of Peru and no. Chile, all from Bigh altitudes in northwestern ae or the Bolivian plateau. Fruit densely white villous-strigose, 4-5 mm. long, seeded, false septum completely absent; corolla more or less marcescent and ie the ovary at maturity BE easkio a ledeein bh th 2x CRE ee piniae Fruit not pallid, sane or somewhat strigose, usually 10 mm. long or more, 4—12-seeded; corollas not marcescent, uous. Flowers small, standard 3-5 mm. long; fruit with the false septum absent or inconspicuous. Inflorescence congested even in fruit ; broad; prostrate perennials of high altitudes in Bolivia and Argentina.....55. A. micranthellus eens becoming racemose; pods 5-6 mm. road; erect annual of coastal Peru and no. Chile. Dora ie! a ase rets ena odode ee 56. A. triflorus Flowers larger, standard 5-9 mm. long; fruit iti a narrow incomplete ore septu Plant cinereous, conspicuously vheete strigose ; leaves usually dhoete? than the internodes; leaflets thickish, rounded or obtuse at apex.........-. sig eM ioe Sal gitar a Se ee ee 57. A. Famatinae Plant green, glabrous or only spars ely strigose ; leaves ane much see the internodes; Onary and fruit distinctly naar north- aig ’ A Ovary and fruit cain or practically so Arge ntina ‘tO. Peru, Tare. sc... s cee Biss cece eeeeeseeves..-59, A. Weddellianus EE. Plant caespitose or decidedly pulvinate, leaf-bearing stems very short and not much elongating, usually less than 1 cm. long or if longer then compacted into very dense pulvinate masses (cf. alpine forms of A. micranthellus and A. tarijensis Plant rhizomatose; fruit 7-13 mm. long, 4-7 mm wide: strongly flattened laterally, intrusion of iow er suture and the false septum largely confined to the attachment-end of pod...........-..--- Plant without rhizomes. Flowers pedicellate, the pedicels 1-4 mm. long, at anthesis ele- vating the flower above its subtending bract; peduncles usually evident at anthesis, up to 4 cm. long. Herbage glabrous or practically SO; flower solitary ; fruit completely 2-celled by a iueh false septum 1 A 61 Herbage evidently hairy; flowers clustered; fruit 1-celled, the false septum membranous and incomplete or absent. Plant yellowish green; calyx-lobes with black hairs; leaflets 3-4-jugate; peduncles 1- eae ovary glabrous; caudex rather coarse....... 62. A. Venturi JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Plant pallid, frequently mene silvery strigose; calyx usually with only pallid hai Ovary and fruit glabrous; lea lets 2-4-jugate; stems with internodes 2—5 mm. long; — with an incomplete false septum......... . confinis Ovary and fruit strigose; stems a. shorter in- ternodes. e Leaflets lanceolate to narrowly ovate, 2-3 mm. broad, 4—6-jugate; fruit somewhat lunate, 10-15 mm. long and 3-5 mm. high; false septum complenely —_— peta cciepaialeaatatns Soha wha seuees goueunayt . A. hypsogenus — linear, 1-2 mm. ae 2-3-jugate ; oblong, 6-8 mm. long and 2-4 mm. Gel, false septum narrow but eviden Sas Aes Esa ach en ele OTA 65. A. cr winiislallass Flowers sessile or subsessile, at anthesis base of calyx closely ensheathed by its bract; peduncles pane age short or none (in fruit rarely paca 3-7 mm. long in A. minus and A. Werdermanni). Flowers 4-8, in sessile subumbellate clusters in the leaf-axils, inte with large bracts. Leaflets ceca to suborbicular, loosely clothed with non-lustrous hairs, usually glabrescent in age, apex oo to deeply iene stipules white, papery, very broad and loose, usually ores er ae jeroeeitane siease vans een pees terepe pusillus Leaflets ovate to lanceolate, densely and permanent clothed with lustrous hairs, apex acute obtuse. — shat the hairs ca. 1 mm "aes corolla . 6mm. long; Bolivia and Avceniina taal apie Re ne ala glade ealeme wea 67. A. Urbanianus Herbage neg villous, the hairs ca. 2 mm. long; corolla ca. 8 . long; Central Peru......... ea ieig + aie Mateg ete a gp kp 8. A. Dillinghami Flowers seme or geminate in leaf-axils; bracts relatively inconspicuou Fruit with a more or less incomplete o_o false septum; valves rather firm and t Leaves canescent or silvery, scuaaaie gs ae villous even in age, not Egriocserral stipule broad and coarse, 5—7 . long; leaf-bearing stems coarse, usually 2- ‘ mm. thick ieiemee ees SS e586 565 Om ree Oe Seep Leese 69. A. peruvianus Leaves green or cinereous, hairs scattered or dis- appearing in age and frequently ered on the upper surface; stipules smaller, 2-4 mm. long; leaf-bearing stems slender, 1—1.5 mm. vee Leaflets very broadly obcordate with a deeply emarginate apex, strongly folded, sparsely ae as with the hairs lying oblique to the Sidevageea vada 70. A. Werdermanni boas bic to elliptic or lanceolate, in- volute- , apex rounded or very weakly and oa emarginate, hairs paralleling midrib. 1947] JOHNSTON, ASTRAGALUS 345 Leaves appressed short- villous, the hairs s short and congested, frequently sub- ulvinate.....<..54 1. A. minutissimus Leaves strigose with shorter more ap- pressed hairs which scarcely, if at all, protrude beyond the leaflet-apex; stems elongating, forming prostrate mats.. i.2le on eae eee Ze minimus Fruit without any false septum, valves rather thin and ragile Plant distinctly pulvinate, the leaves aaa ma cescent and persisting on even the oldest stems; young stems slender and dee internodes 1-10 mm. long; stipules small, not imbricate; pods 4—5 mm. long, several- ae Argentina... Wd Se gute a ld Qidite need 21: eee 73. A. pulviniformis Plant caespitose, the leaves very much less persis- short, internodes shorter than the stipules; stipular sheaths im- bricate and clothing the stem; pods usually 2 mm. long and 1-2-seeded. Stipules with ciliate margins, back glabrous or practically so even when young; calyx-tube o p> — TQ a ° cI o a ° ts wn wn e. Caudex dense, a mass of crowded strict stems and branches intermixed with Sa leaf-remnants; stipules al- DESCENE..o. dee eaea ew ene 74. A. Reichei Caudex a loose mass of spreading stems and branches, these more or less r like and covered with tightly cathine stramineous stipules, old leaves rather promptly deciduous 5-66 nase ees Speer er rene ant 75. A. flavocreatus Stipules broader and more loosely sheathing, ith numerous appressed hairs at least when young; calyx- nies cylindric, densely Herbage with non-lustrous indument; stipules papery, albescent; standard with obovate blade; Bolivia and NST eR ais Fede esac A. deminutivus Herbage with lustrous indument; stipules firmer in texture and darkening in age; standard with a suborbicular blade; Centra WPenie ase cd. +-4)- 77. A. Dielsii 1. Astragalus Chamissonis (Vogel) Reiche, Anal. Univ. Chile 97: 553 (1897), and FI. Chile 2: 96 (1898). Phaca ochroleuca ney & Arnott, Bot. Misc. 3: 186 (1832).— Type collected at Valparaiso, Cuming 3 Astragalus ochroleucus a & A.) Gray, Bot. U. S. Explor. Exped. 1: 413 (1854). Not Gilibert (1782). 346 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVUI Phaca Chamissonis Vogel, Linnaea 10: 592 (1836).— Type collected between Tal- cahuano and Concepcion, Chamisso. Phaca chilensis Nees, Del. Sem. Hort. Vratislav. (1833), and Linnaea 10: Lit (1836). — Described from garden material; perhaps from seeds originally pnts to Europe from Chile by Bertero. Astragalus chilensis (Nees) Reiche, Anal. Univ. Chile 97: 111 (1897), and FI. Chile 1 (1898), not Sheldon (1894). Asteaeahas placens Clos in Gay, FI. Chile 2: 109 (1846).— Type collected near ‘alparaiso, Gay 407. Astragalus ovallensis Clos in Gay, FI. Chile 2: Aa (1846).— Type collected in “los arroyos del departmento de Ovalle, ” Gay 4 Astragalus Volckmanni eet an, Univ. er 18: 46 (1861), and Linnaea (1864); Reiche, Anal. Univ. Chile 97: 547 (1897), and Fl. Chile 2: 90 (1898). — Type ree at es ace southern Atacama, Volckmann. Phaca araucana ¥. Philippi, Anal. Univ. Chile 84: 14 (1893).— Type from sands near sea north of Rio ao prov. Cautin, fe Philip pi. Astragalus 0 araucanus (F. 1.) Reiche, Anal. Univ. Chile 97: 557 (1897), and FI. ile 2: (1898). she, ed monospermus Philippi, Anal. Univ, Chile 84: . (1893); Reiche, re Ini “hile 97: 552 (1897), and FI. Chile 2:95 (1898).— Type from Las rancas near Angol, 1881. Astragalus laetevirens Philippi, Anal. Univ. Chile 84: 30 (1893); Reiche, Anal. Univ. Chile 97: 553 (1897), and Fl. Chile 2:96 (1898).— Type collected at Laja, Bio Bio, Rahmer. Rance: Widely distributed in central Chile, extending from the southern parts of Atacama south to Cautin, on hills near the coast and at lower altitudes in the Cordillera. CHILE: Atac : Camarones, herb. Reed as A. Volckmanni (K). Co- imbo: Coquimbo, ex we Philippi as P. bahay (US, K); La Serena, 1930, Claude Joseph 5446 (US); El Molle, 30 km. west of Vicufa, marsh land along Coquimbo River, 400 m., alt., 10- 16 dm. tall, fl. white-yellow or pink, Wagenknecht 18495 (G); Queb. Arrayan, 25 km. south of Estacion Pelicana, 5 dm. tall, fl. yellow-white and violet, Wagenknecht 18406 (G); ravines of the ic Ovalle, Gay 491 (type of A. aati. Paris). Aconcagua: Valparaiso, Cuming 389 (type of P. ochroleuca, K, BM); Valparaiso, Wilkes Exped. (US); hills near es ane org 824 (G, Del, Paris) ; Popaico near Valparaiso, Feb. 1829, Gay 467 (type of placens, Paris). Santiago: Cord. de Santiago, ex Pippi as P. ochroleuca ae Boiss). Col- chagua: Cord. de Tinguiririca, 2300 m. alt., 1929, Pirion 73 (G). Curico: Nacimiento del Teno, Feb. 1896, Resear (G). Maule: Banos de Longavi, Jan. 1888, Schoenemann (G); Valle del Maule superior, Cord. Linares, Jan. _ Reiche as A. laetevirens (FM, BD). Nuble: trail to Volcan Antuco, herb. Reed as P. ochroleuca (K); Cord. de Chillan, Germain (K, BM, ie Boiss) ; i — Renexado, Jan, 1877, ex Philippi as A. Chamissonis (Speg). Gon epcion: a Vega de Con cepcion, Poeppig 97/70 (BD); Concepcion, D’Urville (BD), and ae As (KK) ; between Talchaguano and Concepcion, Chamisso (type of A. Chamissonis, BD) ; Coronel, 1860, Ochsenius 444 (BD); Itata, Jan. 1878, ex Philippi as P. chilensis (Speg). Bio-Bio: Valle Renaico, 1897, Neger (Munich); San Ignacio de Peme- hue, Jan. 1894, Philippi (G); Ercilla, Feb. 1892, Kuntze as T. canescens (NY). Cautin: Budi, Jan. 1923, Claude-Joseph 1879 (US, NY) This slender erect-growing fruticose species varies from glabrescent to very densely strigose. The plants with the densest indument come from the province of Concepcion and are the typical form of A. Chamissonis. Philippi’s A. monospermus appears to be this densely strigose form. Plants from beyond the province of Concepcion are much more sparingly strigose and some of these become glabrescent in age. I do not believe 72 1947 | JOHNSTON, ASTRAGALUS 347 that practicable segregates of this species can be found upon degree of pubescence. The older writers often used greek letters to distinguish this variation. Phaca ochroleuca var. 8 of Hooker & Arnott, L.c., is based upon Cuming 810 from Concepcion. It is typical A. Chamissonis. At Paris Clos has labeled Gay 586 h. Ch., from “Chile,” as the var. 8 and has appended a note which probably indicates the source of his var. 8. “Les deux echantillons de Phaca ochroleuca H. et Arn. dans lherbier Webb sont differents en ce qu’un est glabre et l’autre est tomenteus.” The var. £ of A. placens, Clos. 1.c. 109, is based upon Bertero 824 from Valparaiso. The inadequately described P. chilensis Nees, appears to belong to the present species. At Munich there is a good specimen of our present ae associated with a label reading ‘“‘Phaca chilensis — hort Monoc. — 1838 A similar specimen at Berlin has a label, in three different scripts, re may be indicated as follows, his — Dalea? e Chile— var. dubia OK — HORT. BOT. BEROL. — . Astragalus Pissisi (Phil.), Phaca Pissist ERED Anal. ae Chile 18: 46 (1861), and Linnaea 33: 45 (1864). — Given as based upon material from Been in prov. Coquimbo, Volckmann; radix borealis montis de Chacabuca, ipse said canescens var. Pissisi (Phil.) Reicho, Anal. Univ. Chile 97: 555 (1897), hile 2: 98 (1898 “Hills south of the Rio Aconcagua and north of Santiago; reported from soriweniers Coquimbo. CHILE. Coquimbo: Chanaral, oi Reed as P. Pissist (K). Aconcagua: Montenegro, 1884, Borchers (BM M). Santiago: Chacabuco, fl. white, ex Philippi as P. Pissist (BD, Del, ne Stock) ; Cuesta de Chacabuca, 1500 m. alt., 1935-36, ne (G); near Colina, fl. white, Sept. 1861, ex Philippi as P. Pica (BD, D One a the collection upon which Philippi originally based this very distinct species is labeled as having come from Chanaral, in northwestern Coquimbo. Subsequently, however, the species has been found only in the region about Chacabuca Pass, to the north of Santiago. I suspect that the specimen said to have come radi Coquimbo may have been mislabeled. e coarse flower, the elongate pedicels, and the large papery long- stiped pendulous pods are outstanding characteristics of this species. It is very distinct but probably has its closest relative in C. limariensis from which it differs in its coarser flowers, more papery pendulous pods and broader flat closely strigose leaflets. 3. Astragalus Edmonstonei (Hook.) Robinson, Proc. Am. Acad. 38: 148 (1902). Phaca flava Hooker & Arnott, Bot. Misc. 3: 186 (Aug. 1832).— Type collected at Valparaiso, Cuming 611. Astragalus flavus (H. & A.) Sheldon, Minn. Bot. Studies 1: 158 (1894) ; Reiche, nal. Univ. Chile 97: 564 (1897), and Fl. Chile 2: 107 (1898). Not Nuttall (1840). Phaca chrysantha Moris, Mem. Accad. Torino 25: 104, t. 4 ake Can upon plants cultivated at Turin, evidently grown from seed collected by Bertero. Astragalus pain (Moris) Reiche, Anal. Univ. Chile 97: 565 eon. and FI. Chile 2: 108 (1898). Not Boiss. & Hoten! (1849). 348 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Astragalus affinis Steudel, Nom. ed, 2, 1: 159 (1840), nomen. — Based upon material f Phaca Edmonstonei Hooker, Trans. Linn. Soc. London 20: 227 (1847).— Based upon collections incorrectly labeled as from the Galapagos Islands, Romans Ne. Phaca — Philippi, Anal. Univ. Chile 84: 22 (1893), nom. provis.; Reiche, Anal. Univ. Chile 97: 564 (1897), and Fl. Chile 2: 107 (1898). ). — Based upon era from Concon. Not Meyer (1831). Phaca acutidens Philippi, Anal. Univ. Chile 84:16 (1883); Reiche, Anal. Univ. Chile 97: 564 (1897), and FI. Chile 2: 107 (1898).— Type from near mouth of the Rio Maule, P. Ortega. Astragalus Hohenacheri Spegazzini, Anal. Mus. Nac. Buenos Aires II, 4: 265 (19C2) —A renaming flavus (H. & A.) Reiche. Rance: Central Chile along the coast. CHILE. Aconcagua: in pascuis arenosis maritimis, Quintero, fl. flavi, “Yerba loca,” Jan. 1830, pine §21 (Paris, type of A. affinis; NY); Quintero, Werdermann 15 (G, US, BM, BD); in arenosis maritimis ad Quintero et Concon, Poeppig 199/161 (BM, BD, Geneva, Paris) ; Concon, Oct. 1884, Philippi as P. podocarpa (G); Renaca, March 1916, Jaffuel 666 (G); Valparaiso, Buchtien (BD, US); Valparaiso, 1831, Cuming 611 (K, type of P. flava). Maule: Constitucion, Oct. 1892, Reiche (G). Indefinite: Chile, in arenosi, frequens, Sept. 1831, Gay 438 (Paris) ; Chile, ies 42 and 43 (K), 44 (BD); “Charles Island, Galapagos,” Edmonstone (K, Edmonstonei; G). I have seen specimens of this readily recognizable species from Quintero, Concon, Renaca, Valparaiso and Constitucién. Of the synonyms listed I have seen authentic material of all but P. acutidens.: All clearly belong to the unmistakable species treated here. Though originally given as from the Galapagos Islands, the collection by Edmonstone which is the type of P. Edmonstonei is unquestionably the present Chilean plant. Edmonstone was the botanist on the cruise of the Herald during which visits were made in Chile (Valparaiso and Concep- cion) and Peru as well as on the Galapagos Island. His collections became confused after his untimely death in Ecuador in 1846. It may be noted that the other Galapagean records for Astragalus are also based upon falsely labeled specimens. The collection by Du Petit-Thouars, mentioned by Hooker when he described P. Edmonstonei, represents flowering ma- terial of A. Menziesii Gray, probably from Monterey, California. Astra- galus brevidentatus Wright, Kew Bull. 200 (1906), also attributed to the Galapagos, is certainly this Californian plant. Cf. Johnston Jour. Arnold Arb. 19: 95-96 (1938). 4. Astragalus limariensis Munoz, Jour. Arnold Arb. 20: 246, t. 1 (1939).— Type m Fray Jorge, Munoz 2. RANGE: Known only from the valley of the Rio Limari, prov. oe pia Chile. . Coquimbo Fray Jorge, Sept. 26, 1939, Munoz 2 (G, type); Fray Jorge, low hills near the fustidings of the estancia, Sept. 26, 1939, Miiin "B. 04 (G) west of Ovalle, Sept. 19, 1917, Baeza (G). A well marked species probably most closely related to A. verticillatus of south central Chile. It agrees with its relative in gross habit, flower shape and size, inflorescence, and narrow elongate leaflets, but differs in its much larger, differently shaped and distinctly stiped pod. 1947] JOHNSTON, ASTRAGALUS 349 5. Astragalus ore (Phil.) Reiche, Anal. Univ. Chile 97: 543 (1897), and FI. Chile 2: Phaca Sate Piper Anal, Univ. Chile 84: 15 (1893).— Based upon material from Araucania, Philippi in 1887, and from Constitucion, Reiche Phaca brachu oise Philippi, Anal. Univ. Chile 84: 15 (1893). — Type from San Javier, prov. Maule, P. Ortega. Astragalus brachytropis (Phil.) Reiche, Anal. Univ. Chile 97: 544 (1897), and FI. Chile 2: 87 (1898). Not Meyer (1831). Astragalus sei Spegazzini, Anal. Mus. Nac. Buenos Aires ser. 2, 4: 264 (1902). — Based upon A. brachytro pis Ruiche. not Meyer : %Inan 2 along the coastal hills of Chile from eee and Talca south- ward ae northern parts of the province of Concepcion. CHILE: Talca: Talca, ex Philippi as P. ne (BD). Maule: Em- pedrado, Sept. 18, 1892, Reiche ta Villa Alegre de Loncomilla, Oct. 24, 1931, Espinosa (G). Coneenee Penco near Concepcion, Oct. nie Neger (Munich) ; Araucania, Nov. 1887, Philippi as P. aiedlisia (isotype, G, BM.) The type of P. verticillata is given as Bie in Nov. 1887, in “Araucania.”’ It is very similar to the material from Penco collected by Neger. Both specimens have flower but lack fruit. The leaflets are geminate and very slender. During his ‘‘Botanische Excursion in das Araukanerland” in Nov. 1887, cf. Bericht. Ver. Naturk. Kassel vol. 41 (1896), Philippi visited San Javier, Concepcion, Angol, Traiguen and Temuco. Accordingly his type collection of A. verticillatus could have originated in the vicinity of Cancepcion where Neger collected specimens very much resembling it. The material I have seen from the watershed of the Rio Maule is coarser and has less slender leaflets than represented in the type of A. verticillatus and in Neger’s plant from near Concepcion. This more northern material is that described as P. brachytropis upon the basis of a plant from San Javier, prov. Maule. Its leaflets are usually in pairs along both sides of the leaf-rhachis but in one collection from Talca, agreeing in all other details with the species, they are solitary along the rhachis. The species is evidently related to A. Pissisi with which it agrees in such distinctive details as color, size, and structure of the corolla and unusually elongate pedicels. It differs from its relative in having much smaller (5-10 mm. long), globose or ovoid pods, that are sessile in the calyx (not stiped) and borne on erect pedicels. The pods bear scattered appressed usually pale hairs. They are usually abruptly contracted at the base. Philippi describes P. verticillata as having corollas in which the wings are short and only half the length of the keel. This is evidently a mistake. Authentic material of the species shows the corollas to be like those of P. brachytropis which he properly described as having wings almost twice as long as the keel. Reiche so described them in his Flora de Chile. 6. oa aa Looserii, sp. nov. Planta perennis; caulibus pluribus erectis sparse stricteque ramosis 3—10 dm. altis fistulosis (basim versus 3—5 mm. crassis) apicem versus sparse inconspicueque fusco-strigosis alibi glabrati; foliis glaberrimis concoloribus viridibus, rhachibus 7-15 cm. longis ascendentibus, foliolis crassiusculis 350 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII 10—12-jugatis 1-2.5 cm. longis 2—4 mm, latis paulo supra medium latioribus apice obtusis vel subtruncatis basi attenuatis; stipulis liberis triangulari- bus; pedunculis (partibus floriferis inclusis) 15-30 cm. longis foliis duplo longioribus; floribus laxe racemosis (racemis saepe 10-15 cm. longis) stricte ascendentibus; hig: vag pilis nigris brevibus abundantibus vestitis, tubo cupulato 2—2.5 mm. longo, lobis subulatis ca. 2 mm. longis, pedicellis a. 1 mm. longis; anil 1 cm. longa purpurea vel Regen leguminibus ascendentibus vel rariter plus minusve recurvatis ca. 1 cm _ longis inflatis leviter lateraliterque compressis sparse nigro- ie apice acutis basi rotundis vel obtusis infra medium crassioribus. Rance: Cordilleras of Santiago, Aconcagua and Coquimbo, 0G Santiago: Fierro Carrera, Las Condes, much branched, up to 1 m. tall, ‘reddish, 2800 m. alt., Jan. 1930, Gualterio Looser 1132 (type, Gray Herb.) ; ae ‘Caeta, albino form, Tit. 1930, Looser 1133 (G); Mina de la Disputada, Cord. de las Condes, 3500 m. alt., Dec. 1930, Father Denys Le Manchee 256 and 257 (G); Las Condes, herb. Reed (BM); Cord. de Santiago, Feb. 1870, Reed as P. elata (K) ; Cord. de Santiago, ex Philippi as P. Berteriana (K); Cord. de Santiago, ex Philip as P. elata ? (BD, Del); Cord. de Santiago, Jan. 1873, ex Philippi as P. elata (Speg). Aconcagua: Rio Sobrante, above Chincolco, in vega, 2700 m. alt., fl. blue-white, Morrison 17323 (G). Coquimbo: Fraguitas, Dec. 1874, Germain (G); La Mol- loca, Cord. de Hlapel, Jan. 1888, ex Philippi as P. elata (BM); Rio Rapel, Higuera, Agua Amarilla, herb, Reed as P. elata (BM); Rio Rapel, Reed (G); bed of the river of San Isidro [| Vicuna], stems 5—7.5 dm. tall, 600 m. alt., 1836, Gay 363 in pt. (Paris) ; Chile, Gay 876 in pt. (Paris) ; Chile, Gay as P. elata (K, BD, Del). A species evidently related to A. Berterianus and one resembling that species in habit of growth, and size, shape and structure of fruit. The new species is a plant of the cordilleras. Its relative belongs to the valleys and coastal hills and is quickly distinguished from the montane plant by a less robust habit, conspicuously hairy herbage, and smaller, less juicy leaflets. Astragalus Looserti is best known from the mountains east of Santiago. Philippi and Reiche seem to have confused it with the very different A, curvicaulis. I have been unable to separate the plants of the cordilleras above Santiago from some that come from the cordilleras further north in Coquimbo. ‘The four sheets at Paris collected by Gay and determined by Clos as P. elata contain varying mixtures of three species. The pre- dominating plant, however, is the northern form of A. Looserii. 7. Astragalus Berterianus (Moris) Reiche, Anal. Univ. Chile 97: 555 (1897), and FI. Chile 2: 98 8). Phaca canescens Hooker & Arnott, Bot. Misc. 3:185 (1832).—Type from Val- paraiso, Cuming 735. Astragalus canescens (H. A.) Gray, Bot. U. S. Explor. Exped. 1: 412 (1854) ; Reiche, Anal. Univ. a 97:555 (1897), and Fl. Chile 2:98 (1898). Not DeCandolle (1802). Phaca Berte ertana Moris, Mem. Accad. Torino 37: 105, t. 26 Said — Type grown in the Bot 2 Gaaien at Turin from seeds collected by Berter Astragalus blot Clos in Gay, Fl. Chile (2: 109 (1846)" «ty pe collected along t o Cachapoal, Oct. 1828, Bertero 6 a Astragalus piven Clos in Gay, Fl. Chile 2: 119 (1846). Not aa 1840).— Type collected in the streets of Santiago, Sept. 1829, Gay 4 Astragalus placens var. oblongifolius (Clos) Reiche, Anal. Univ. fi WE 552 (1897), and Fl. Chile 2: 95 (1898) 1947] JOHNSTON, ASTRAGALUS 351 joer Phaca penile Philippi, Anal. Univ. Chile 84:17 (1893).— Type collected near Los ; PaO oa (Phil.) Reiche, — Univ. Chile 97: 556 (1897), and FI. Chile 2: 99 (1898). Not Fischer (185 Phaca hiatal Philippi, Anal. a Chile 84: 21 (1893).— Type from sandy places near Concon, Oct. 1884, F. Philip pi. Dieaiics dolichostachys (Phil.) Reiche, Anal. Univ. Chile 97: 556 Chile 2: 99 (1898). Phaca dissitiflora Philippi, Anal. Univ. Chile 84: 25 oe collected at 1700 m. in the valley of the Rio Colorado in Jan. pede laxiflorus var. dissitiflora (Phil.) Reiche, ee ane Chile 97: 556 (1897), Fl. Chile 2:99 (1898). Not ae Pgs ne aconcaguensis Spegazzini, Anal. Nac . Buenos Aires ser. 2, 4: 264 (1902). A new name from A. laxiflorus rane - Fischer Coquimbo: Caimanes, Oct. 1935, Espinosa (G) ; Andocollo, Oct. 1926, oo Josh 4532 (US). Aconcagua: Cord. de Norte Ligua, 1902, Castillo ; Quintero, Werdermann 24 (G, US, BM, BD); Concon, dunes, Oct. 12, 1884, ere (G, isotype of P. dolichostachya) ; between Concon and Quintero, Oct. 1928, Guenther & Buchtien ia ehaca, sand hills, 1832, Bridges 46 (K); Valparaiso, — 1897), and FI. 1831, ek (35 WK At of P. canescens); Valparaiso, Cuming 734 (K) and 736 eM) King (K) and Bridees (BM) ; age Felipe, 1926, Claude-Joseph 3846 and 3850 (US); Santa Rita, 1879, ex Philippi a Berteriana (Speg). Santiago: Colina, 1825, Maca (KD; Colina, Oct. 1887, ex on as P. Berteriana (BM) ; Chacabuco, ex Philippi as P. Berteriana (BD); Lampa, Reed (K); Lampa, Nov. 1861, ex Philippi as P. eitsdle (Speg) ; Tiltil, dry sunny slope, 700 m. alt., Looser 728, 737 and 785 ae Montero 149 (G); between Tiltil and Cuesta da La Dormida, 700-1200 m. alt., Oct. 1927, Looser 784 (G); streets in Santiago, Sept. 1829, Gay 498 (type of A, oy hence ae Paris) ; Cajon del Maipo, Philippi as P. canescens (BD); near the Rio Maipo, May 1828, Bertero 66 (Paris); Rio Colorado, Jan. 1888, Philippi (G, isotype of e ET ee Renca, Oct. 1876, ex Philippi as P. Pissisi (Speg). Col- chagua: ear the ae Cachapoal, Oct. 1828, Bertero 66 (Paris, type of A. oblongi- folius ; Del). Nuble: Cord. de Chillan, Philippi as P. canescens (BD). Talca: Curico, Jan. 1928, Cis. Joseph 5230 (US). The type of A. Berterianus was grown in Italy from seeds sent by Bertero from Chile. Since Bertero seems to have made only two collections of this species, one near the Rio Maipo in May, 1828 and another near Rancagua along the Rio Cachapoal in Oct. 1828, one of these, probably the former, is the source of the seed he sent to Turin. The latter collection supplied the type of A. oblongifolius Clos. This species varies in the amount of dark hairs on the inflorescence, calyx and fruit. Some plants have a minimum of dark hairs and are light in color, as in the type of P. canescens. The color of indument shows no geographical correlation. 8. Astragalus paposanus Johnston, Contr. Gray Herb. 85: 51 (1929).— Type from quebrada above Agua Perales, near Paposo, Johnston 5586. Rance: known only from the vicinity of Paposo and Taltal. CHILE. Taltal: Quebrada Anchuna, Sept. 1936, Montero 2982 (G); aereds Guanillo above Agua Perales, near sale Dec. 8, 1925, Johnston 5586 (type, G) between Paposo and Punta del Rincon, Nov. 30, 1925, Johnston 5564 (G); dope: above El Rincon, dryish ridge above fertile belt, Dec. 7, 1925, Johnston 5498 (G). Very closely related to A. coquimbensis and perhaps simply an outlying northern form characterized by a velvety pallid indument. 352 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XXvuI 9. Astragalus peepee er (H. & A.) Reiche, Anal. Univ. Chile 97: 566 (1897), and Chile 2: 109 (189 Phaca ig ehiaeg Hooke : as Bot. Misc. 3: 184 (1832).— Type given as rom Coquimbo, Cumin Phaca atacamensis Philippi ree Univ. pin 84:16 (1893).— Type collected at La Brea, Atacama Desert. 1884, San Rom Phaca et tes Philippi, aeons Univ. Chile "84: 21 (1893).— Type collected near Carrizal Bajo, Atacama, T. Astragalus vasticola Johnston, Contr, Gray Herb. 85: . (1929). — Based upon Phaca atacamensis Phil., not A. atacamensis Fries (1905). RANGE: Deserts along the coast from Coquimbo north towards Taltal. CHILE. Coquimbo: Coquimbo, Cuming 864 (type of P. ecient K); Coquimbo, ex Philippi, no. 221, as P. coquimbensis (BD). Atacama: Carrizal Bajo, herb. Reed as P. carrizalensis (isotype of P. carrizalensis, K); near Pique aie Totoral, 200 m. alt., Werdermann 468 (G, K, BD); between Monte Amargo and ) ; Playa Caldera, cerca Morro, Nov. 1936, — (G); near Caldera, Gigoux (G) ; haiween Caldera and Queb. de Leon, 10 alt., Werdermann 1638 (G, BD); Chanaral, Sept. 1909, Reiche (G); Aguada Grande. Johnston 5824 (G, K); indefinite locality, Geisse 122 (NY); and Morong 1283 (NY). Antofagasta: Aguada Cachina, Johnston 5739 (G, K). An annual species of the coastal desert of northern Chile. It has evident relations only in the closely related A. paposanus. OQ a 2 oO “3 & an ° “o p = = be | & a — = i~} 3 1S be a wa rN ies) oO 10. Astragalus Salierrae (Clos) Reiche, Anal. Univ. Chile 97: 559 (1897), and FI. Chile 2: 10 oo 8). Phaca elata Hooker & Arnott, Bot. Misc. 3: 185 (1832).— Type from eae “Cor- dillera of te alee from the upper Aconcagua Valley), Cuming . Astragalus elatus (H. & A.) Reiche, Anal. Univ. Chile 97: 557 (1897), and Chile 2: 100 (1898). Not Boiss. & Bal. (1859). Phaca curvicaulis Clos in Gay, Fl. Chile 2: 6 o_o — Type from “Central Chile” (probably cordilleras of Coquimbo), Higher a of baa and ne and the high coast-ranges east of Limache in the prov. Santiago. HILE. Santiago: Las Vizcachas, coast ranges about 8 km. east of La Dormida, 1400-1500 m. alt., open slope, Morrison 16819 (G). Aconcagua: Ojos de Agua, Cruckshank 92 (K); Juncal, 2200 m. alt., Buchtien (G, US, BM, BD); Juncal, 300-2400 m. alt., 1925, Pennell 12975 (G); Rio Blanco, 1950 m. alt., Nov. 1927, 224 (type of P. elata, K); locality indefinite, Bridges 40 (K, BM) ; locality oe 1839, John Style (Del). Coquimbo: Caren, Queb. Luncuman, 40 km. n, e. o Illapel, ca. 1000 m. alt., in gravel, fl. lavender, Worth & Morrison 16496 (G); pr Coquimbo, Gay 876 in part (Del) ; locality indefinite, Gay as P. curvicaulis (type of P. curvicaulis, Paris; G, K, BD, Del). The type of P. curvicaulis at Paris is given as collected by Gay, but has no number, collection-date, nor locality. It seems very similar to part of Gay 876 which consists of a mixture of A. Looserii and this species. No locality is given for Gay 876 at Paris, but at Geneva it is labeled as from the province of Coquimbo. Since Gay does not appear to have collected in the upper Aconcagua valley, which seems to be the southern limit of the species, it seems probable that he must have obtained his material in the cordillera east of Ovalle or Coquimbo. The type of P. elata is simply given as from the Cordilleras of Chile. Cuming, its collector, is known to have collected in the upper Aconcagua valley where the species is best 1947] JOHNSTON, ASTRAGALUS 353 known, and it is sierra certain that his material must have come from that area. The var. 8 mentioned by Hooker & Arnott, l.c., based upon Cuming 734 ee A. Berterianus. The present species is a very distinct one, unique in its firm strict elongate lancoid pods. It has been confused with A. Looserii but is readily distinguished from that species by its much larger more inflated and rounded fruit, smaller flowers, scanty short pale, rather than dense conspicuous black indument on the ovary, fruit and calyx, and cinereous rather than green juvenile leaves. 11. Astragalus pehuenches Niederlein in Roca, Exped. Rio Negro 2: 206, t. 4, (1881). Ee collected between Pampa de Tilqui and Pampa de Trili, Neuquen, Nie der Phaca sss Gillies ex Hooker & Arnott, Bot. Misc. 3: 183 (1832).— Type from between Mendoza and Uspallata on Hee road to Canota, Gillies. Astragalus inflatus Gillies ex Steudel, Nom. ed. 2, 1: 161 (1840). Not DeCandolle (1802). Phaca striata Clos in Gay, FI]. Chile 3 ra (1846). — Type from Andes de la Dehesa, prov. Santiago, Nov. 1829, Gay 9 Astragalus striatus (Clos) Reiche, Univ. Chile 97: 570 (1897), and FI. Chile 2: 113 (1898). Not Nuttall (1840). sas macrocarpa Philippi, Linnaea 28: 619 (1857).— Type from the Cordillera e San Fernando, Feb. 1843, Bustillos Prodan macrocarpus (Phil.) Reiche, aaa Univ. Chile 97: 563 (1897), and FI. Chile 2: 106 (1898). Not Pallas (1776), nor DeCandolle (1802 Tragacantha bisinflata Kuntze, Rev. Gen. 2: 940 (1891).— Based upon P. inflata Gillies Tragacanthe grandis Kuntze, Rev. Gen. 2: 941 (1891).— Based upon P. macrocarpa Phil. Astragalus grandis (Kuntze) Spegazzini, Rev. Argent. Bot. 1: 214 (1926). Astragalus megalocarpus Spegazzini, Anal. Mus. Nac. Buenos Aires ser. 2, 4: 265 ).— Based upon A. macrocarpus Reiche, not Pallas. eae! macrocarpus var. petiolatus Hauman, ‘Ana 1]. Soc. Cien. Argentina 86: 280 918-19).— Type from Puente del Inca, Sanzin 349, :: Chilean cordilleras from southern Coquimbo to Talca; Argentine cordil- leras from southernmost San Juan south to northern Neuquen and eastward on the desert lowlands into central La Pampa and northern Rio Negro CHILE. Coquimbo: Cuncumén, Rio Choapa, ex Philippi (BD). Santiago: Andes de ie Dehesa, Nov. 1829, Gay 926 (type of P. striata, Paris) ; Las Condes, herb. Reed (G, K). Colchagua: Banos del Flaco, ewe Begoen (G).; Val Tinguiririca, 1600 m. alt., 1897, Wilczek 87 (Boiss). Talca Cuesta Vergara, ca. 2300 m. alt., Jan. 1933, Grandjot ee ae Los Ontos peek m. alt., 1936, Mexia 7859 (G); Cord. de Curico, 1800 m. , Jan. 1897, Reiche (B ARGENTINA. San a a de la Rio on aren Caleta Bicrak, March 1883, Giissfeldt (BD). Mendoza: Valle de Canota, Gillies (K); between Uspallata and Paramillo de Canota, gravelly soil in valley, Gillies (K) ; valleys between Mendoza and Uspallata by road of Canota, Gillies (type of P. inflata, K); La Loma del Agua Kurtz 7618 (NY); Los Molles, upper Rio Salado, Jan. 5, 1893, Kurtz 7487 (NY); near La Quebrada, Dec. 1933, Ruiz Leal 1836 (G); Paso de la Cruz de Piedra, 1500 m. alt., Kuntze (NY, US, BD); Malargue, Monticelli E 71 (G). La Pampa: Puelén, Durando (G); Gob. de la Pampa, 1932, un cura (Burkart). Neuquen: between the pampas of Tilqui and Trili, rare in gypsum-soil, ca. lat. 37°, long. 70°, June 26, 1879, Niederlein (type of A. pehuenches, BD). Rio Negro: Chocle-choek Dec. 1907, herb. Spegazzini (Speg); campo between Rio Colorado and Rio Negro, Jan. 1903, herb. Sli (Speg). Indefinite: Patagonia, lat. 50°-53°, Moreno & Tonini 438 (N 354 JOURNAL OF THE ARNOLD ARBORETUM (VOL, XXVIII A coarse spreading herbaceous plant with glabrous fruits and ovary. In its large spreading or pendulous papery pods it suggests A. Darumbium, but that plant has smaller flowers, a strigose ovary and fruit and a more erect habit of growth 12. Astragalus valerianensis pier Contr. Gray Herb. 85: 166 (1929).— Type from Rio Valeriana near La Cueva, 1926, Johnston 6033 Rance: Cordilleras of sae ee CHILE. Atacama: Rio Valeriano near La Cueva, cordillera east of Vallenar, 2800 m. alt., loamy talus slope, Johnston 6033 aa G). A plant somewhat suggesting the montane forms of A. Darumbium but quickly distinguished by its glabrous ovary and fruit, strict stems, pedicels 4—6 mm. long, and detached more northerly occurrence. It is probably most closely related to A. pehuenches, but differs from that plant in its evidently strigose herbage, more slender and strict stems, and larger pods. 13, Astragalus vagus (Clos) Reiche, Anal. Univ. Chile 97: 563 (1897), and Fl. Chile : 106 (1898). Phaca vaga Clos in Gay, FI. Chile 99 (1846).— Type from Pasto Blanco, cordilleras east of Coquimbo, Gay 3 — Rep sende + Hieronymus, ae Acad. - Cordoba 4:22 (1881).— m Paramillo, west of San Juan, Echega Astragalus tit var. Echegarayi (Hieron.) ane Bol. Acad. Nac. Cordoba 26: Phaca :. at Philippi, Anal. ja Chile - ig SS — Type from Queb. Seca, cordillera de Atacama, lat. 27° 55’, long. 69° , San Roman Astragalus San Romani (Phil.) Aes Anal. eed on 97: 559 (1897), and FI. Chile 2: 102 (1898) Rance: Higher cordilleras in the provinces of Atacama, ppeowrind and San Juan. CHILE. Ataca Cerro Cadillal, Cord. Rio Turbido, 3200 m , Werder- mann 962 (G, K, BD); ero parts of Queb. de San Miguel, ‘about car flats and somewhat purplish at base, 3250 m. alt., Johnston 5918 (G); Laguna Chica, in gravel, fl. white to flesh-colored, 3200 m. sit, Johnston 5971 (G); Rio Sancarron below Rucas, Arvada bench, fl. pink to magenta, 3200 m. alt., Johnston 6205 (G, K). Coquimbo: Pasto Blanco, Cord. de Elqui, 2837 m. alt., Gay 364 (type of P. vega, Paris; ‘G, K, BD, Del). ARGENTINA. San Juan: Paramillo, Jan. 1876, Echegaray (type of A. Echegarayi, BD). This appears to be the most common Astragalus with inflated pods found in the cordilleras of Atacama and Coquimbo, It may be related to A. coquimbensis, which occurs at low altitudes west of the cordilleras, but differs from that annual in its strong perennial root and somewhat smaller capitately congested pods. When I examined the type of A. Echegarayi I identified it with A. vagus, but subsequently as my studies of the genus progressed I have come to wonder if it might not really be a form of A. monticola. Phaca San Romani, however, is unquestionably a synonym of the present species. 14. Astragalus Darumbium (Bertero) Clos in Gay, Fl. Chile 2: 112 (1846) Sutherlandia Darumbium Bertero ex Colla, Mem. Accad. Torino 37:55 (1834). — pe from near Rancagua, Bertero. 1947] JOHNSTON, ASTRAGALUS 355 Phaca ponte rey Philippi, marae 28: 620 (1857).— Type from Las Aranas ine, C e Santiago, Oct. 1853, Philippi. ses i ae (Phil.) ae Anal. Univ. Chile 97: 554 (1897), and FI. 4). Chile 106 (1898). Not Somm. & Levier (1893-9 Phaca hae Philippi, Anal. Univ. Chile 41: 690 (1872).— Type from the Cord. de antiago Astragalus macrapies var. robustus (Phil.) Reiche, Anal. Univ. Chile 97: 554 (1897), and FI. Chile 2: 106 (1898). Tragacantha firma Kuntze, Rev. Gen. 2: 941 (1891).— Based on s ‘aie Phil. ges chi atuelii Chodat & Wilczek, Bull. Herb. Boiss. ser. 2, 2: 477 (1902).— Type from Arroyo Manga, Valle del Atuel, 1897, Wilczek 86. RANGE: cataiiine of Santiago and Colchagua and adjacent Mendoza. . Santiago: Cerro Provencia, 2100-250Q m. alt., 1932-33, Grandjot (Grandjot); Valle de Ramon, 2200 m. alt., Feb. 1933, Grandjot (Grandjot); Valle de la Hierba Loca, 2300 m. alt., 1932, Grandjot (Grandjot) ; Manquehue, Oct. 11, 1908, Rudolpho (G); Las Arafias, Cord. de Santiago, Oct. 1853, Philippi (isotype of P. poal near San Joaquin, perhaps fai poe waste xen from the cordilleras, 1829, Bertero 679 (isotype of S. Darumbium, Paris, Del, BD, BM); Rancagua, Dec. 1828, Bertero 679 (Del) GENTINA. Mendoza: Arroyo Manga, Valle del Atuel, in sand and about rocks, 1900 m. alt., 1897, Wilczek 86 (type of A. atuelii, Boiss). The type of A. Darumbium is a very rank plant with coarse decumbent stems becoming 1 m. long and 7 mm. thick. It was discovered along a stream at the base of i cordilleras and thought to be a possible waif washed down from higher altitudes in the mountains (‘“‘fortasse semina ex cordilliera provenant”). Similar very coarse plants have not been re- discovered. A very closely related plant, however, has been found in the cordilleras and described as P. macrophysa, P. robusta and A. atuelii. This latter has shorter more slender stems, smaller leaves, perhaps shorter pedicels and larger black-hairy pods. The most reduced form is macrophysa. Apparently intermediate between this and the type of A. Darumbium is P. robusta. The indument on all these forms varies from sparsely to very densely strigose, some plants appearing glabrous at first sight while others are obviously silky with abundant appressed hairs. When A. atuelii was originally described the type was given as from 900 m. altitude and as representing Wilczek’s collection no, 372. Th label on the type at Geneva, however, gives the altitude as 1900 m. and the collection number as 86. e species is related to A. monticola and, in fact, grows with it in the cordilleras east of Santiago. It is distinguished from this relative by its usually larger fruit, slightly more elongate calyx, and silky appressed indument. Some of the montane forms have a root that is much more slender than in other cordilleran species of this general relationship, and in gross habit frequently suggest forms of A. coquimbensis and A. paposanus. The two latter species, however, are not closely related to our plant and are readily distinguished by having pink rather than violet or purple corollas. 356 JOURNAL OF THE ARNOLD ARBORETUM [voL. XXVIII 15. Astragalus monticola Philippi, Linnaea 33: 47 (1864); Reiche, Anal. Univ. Chile 97: 549 (1897), and FI. Chile 2:92 (1898). — Type collected near the silver mines of Las Aranas, Cord. de Santiago, Philippi. Astragalus ? Barceloi Philippi, Anal. Univ. Chile 84: 27 (1893) ; Reiche, Anal. Uni Chile 97: 566 (1897), and FI. Chile 2: 109 (1898). — “Ex Andibus ailianenabus, ni fallor, provenit.” Rance: Cordilleras of Santiago and San Juan. CHILE. Santiago: Las Arafias, Nov. 1861, Philippi (type of A. monticola, Santiago); Mina Disputada, Cord. de los Condes, 3500 m. alt., Father Deny Le Manchee 258 (G); Mina Disputada, 2500-2600 m. alt., April 1933, Grandjot (Grand- jot); Valle de la Hierba Loca, 2000-2300 m. alt., Nov. 1932, Grandjot (Grandjot) ; Valle Largo de las Condes, 2600 m. alt., Jan. 1934, Grandjot (Grandjot) ; rea Provencia, 2100 m. alt., Nov. 1933, Grondiat (Grandjot); Potrero Grande, 216 ARGENTINA. San Juan: between Tudeum and Queb. de Conconta, Jan. 1, 1930, Perez Moreau 30/105 (G). A species evidently related to the montane form of A. Darumbium and differing from it chiefly in having dull spreading, rather than silky closely appressed hairs on the leaves, stems and fruit. Though A. monticola and A. Darumbium sometimes grow together (Dr. Grandjot collected them both at Valle de Hierba Loca and at Cerro Provencia) they maintain their characters and are readily separable. The type of A. Barceloi, as Philippi suspected, is almost certainly mislabeled. The label on the type at Santiago reads “Astragalus Barceloi, Cord. de Chillan?,” but the specimen agrees Closely with the material I have cited from the cordilleras east of Santiago and I believe probably originated there. 16. Astragalus vesiculosus Clos in Gay, FI. Chile, t. 17 (1844), and Fl. Chile 2: 120 (1846) ; Weddell, Chlor. Andina 2: 261 (1861); Reiche, Anal. Univ. Chile 97: 559 (1897), and Fl. Chile 2: 102 (1898).— Type from Cord. de los Patos, Gay 488. pg. 16 (1843).— Type from 2400 m. alt. in the Cord. de San Fernando, Meyen. Astragalus nubigenus (Meyen) Taubert in Engler & Prantl, Nat. Pflanzenf. iii. Abt. 3: 303 (1894); agrees — Univ. Chile 97: 560 (1897), and FI. Chile 2: 103 (1898). Not Don (18 Phaca a Philippi, am 28: 680 (1857).— Type from the Cord. de San Fernando, Bustillos. Tragacantha andina Kuntze, Rev. Gen. 2: 940 (1891).— Based upon P. Bustillosi Phil. Astragalus — var. Arnottianus Meyen ex Reiche, Anal. Univ. Chile 97: 561 (1897), a . Chile 2: 104 (1898).— Based upon (Phen Arnothiana Hook.” sensu Mey ee ise 1: 356 (1834); i.e., Rio Maipo, 11000 ft. alt., Meyen. Astragalus apenas Reiche, Anal. Univ. Chile 97: 562 (1897), and FI. Chile 2: 105 (1898). — Based upon Phaca Bustillosi Phil.; not Astragalus Bustillosii Clos. Astragalus Meyenianus Spegazzini, Anal. Mus. Nac. Buenos ae Ba 262 (1902).— Based upon Astragalus nubigenus Taubert; not A. nubigen Astragalus Ameghinoi Spegazzini, Anal. Mus. Nac. Buenos ren es oe (1902). — Astragalus Meyenianus var. paucifoliolata Hicken, Darwiniana 1:56 (1923).— = from oe del Burro, Rio Atuel, Gerth 102 and 106a. RAN the provinces of San Juan and Coquimbo southward along the higher ae a Madbe and Talca, and with outlving stations in Chubut and Santa Cruz. 1947 | JOHNSTON, ASTRAGALUS 357 CHILE. Coquimbo: crest of the high cordilleras of Los Patos, 3340 m. alt., Gay 488 (Paris, type of A. vesiculosus); Cerro la Yerba Loca, east of La Vega Escondido, dept. Illapel, 2800-3450 m. alt., fl. blue and white, Dec. 1938, Morrison 16932 (G). Aconcagua: near Junta de Pinquenes, Rio Sobrante, dept. Petorca, rocky slopes, 3500-3700 m. alt., fl. blue-purple, Morrison 17284 (G). Santiago: Cord. de Santiago, 3500 m. alt., Feb. 1899, Reiche as A. Bustillosti (BD); Cord. de Santiago, 1899, ex Philippi as A. rupestris (BM); Cajon del Cepo, Cord. Santiago, 900 m. alt., Feb. 22, 1894, Dessawer (Munich); above Laguna Negra, 3300 m. alt., 1902, Hasttaes 498 (US) ; Rio cay [upper Rio del Volcan!] 3300 m. alt., Feb. 1831, Meyen as P. Arnottiana (BD, type of A. nubig. var. Arnottianus Reiche) ; Paso de Maipo, Jan. 1883, Giissfeldt pn Paso ie la Cruz Piedra, 2300 m. alt., Jan. 1892, Kuntze (NY, BD, US). Colchagua: Cord. Tinguiririca, east of Reno, 2200 m. alt., 1930, Hee 702 (G); El Teniente, Rio Coya, 2800 m. alt., 1925, Pennell 12317 (G); Cord. de San Fernando, Feb. 1843, herb. Philippi as P. Bustillosii (? type of P. Bustillosi, aa. Cord. de San Fernando, Meyen (BD, type of P. nubigena) ; Cord. del Rio Tinguiririca, 2400 m. alt., Feb. 1831, Meyen as P. nubigena (BD Talca: near Volcan Peteroa, Bridges 1129 (K); Santa Elena, Cord. de Curico, Feb. 1902, Flaminio Ruiz (G). ARGENTINA. San Juan: crest east of Carrisito (Rio Blanco), ca. 3200 m. alt., Nov. 17, 1915, ae 1448 (BD). Mendoza: Punta de Vacas, March 1901, Sontorai (Speg) ; “La Cumbre, Las Cuevas,” Dec. 1908, Spegazzini (Speg) ; Paso del Portillo, 3500-4000 m. alt., Rais Leal 1985 and 2110 (G); Paso Hondo, Valle Rio Tunuyan, Ruiz Leal 2064 (G); Rio Tordillo, Piedra del Burrero, ca. 2900 m. alt., 1897, Wilczek 105 (Boiss, US); Cajon del Burro, ca. 2900 m. alt., Wilczek 104 (Boiss, WS): mae de los Guanacos, 2600 m. alt., Jan. 1921, Carette 227 (G); Mala Dormida, 2700 m. aan «L921 Coane 238 (LP). Chubut: Valle de Laguna Blanca, ene Te (K). Santa Cruz: between San Julian and Rio Deseado, 1899, Ameghino (Speg, type of 4. Ameghinoz). This species is best known from the cordilleras about the headwaters of the Rio Maipo. Though apparently with a rather disrupted distribu- tion it is a very distinct and readily recognizable species, characterized by its grayish loose indument, broad crowded obcordate leaflets, small fre- quently pale corollas and firm rounded inflated hairy pods Among the synonyms cited above, only A. Ameghinoi deserves special mention. The type of this species comes from southern Patagonia, far to the south of the principal range of A. vesiculosus. In appearance and all vegetative and habit characters it is similar to A. vesiculosus. The fruit, however, is small (ca. 1 cm. long), rigid, and very densely and conspicu- ously hairy. It is, however, not thoroughly ripened, and I suspect that some or all of these differences may be associated with immaturity. Con- cerning A. Ameghinoi it may be noted that of the two collections cited by Spegazzini the one from the Golfo de San Gorge is a very immature specimen of A. Dusenii showing flower-buds only. 17. Astragalus Arnottianus (Gillies) Reiche, Anal. Univ. Chile 97: 561 (1897), and 898). Fl. Chile 2: 104 Phaca Arnottiana Gillies ex Hooker & ae Bot. Misc. 3: 184 (1832).— Type material from El Cerro de la Polcura and Las Lefas, Mendoza, Gillies. Phaca uspallatensis Philippi, Anal. Univ. Chile 36: 168 (1870).— Type collected between Mendoza and Uspallata. Phaca Reedi Philippi, Anal. Univ. Chile 36: 168 (1870).— Type from the eastern side of Paso del Portillo, Mendoza, Ree Astragalus Reedii ie Hauman, ive eg. Hautes Cord. 127] Anal. Soc. Cien. Ar- gentina 86: 281 (1918 358 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII Phaca nana Philippi, Anal. Univ. Chile 84: 19 (1893).— Type from Cordillera del Peuco, prov. Colchagua sine satay nanus (Phil.) Riche: Anal. Univ. Chile 97: 569 (1897), and Fl. Chile 2112 (1898). Not DC (1802 sa ira Philippi Spegazzini, Anal: Mus. Nac. Buenos Aires iv. 4: 265 (1902).— Phil. ANGE: Along the os cordilleras from Coquimbo and La Rioja southward to Mendoza and Colchagu CHILE. wane nak between Guanta oo Banos del ee Reed (K); Estero de Guanta, 3500 m. alt., Johnston 6252 (G, US). Aconcagu upper Rio Acon- cagua, 3600 m. alt., Gosse (K); Caracoles, ae 3508 (G); aie 2300 m. alt., 1913, Buchtien (G, US, BM, BD). Santiago: Cajon de Esmeralda, 1900, J Philippi (G); Valle del Yeso, Reed (K); Laguna. Negra, Feb. 1901, Ayarzum (BM). olchagua: El Teniente, Rio Coya, 2800 m. alt., Pennell 12337 (G); Cord. del Peuco, 1886, herb. Philippi (type of P. nana, Santiago). oja: Queb. Descubrimento Nuevo, Hosseus 1398 (BD). Mendoza: Cord. del Tigre, 2500 m. alt., King 323 (BM); Cumbre de Uspallata, 3600 m. alt., Wilczek 103 (Boiss) ; Siente del Inca, King 12 (K); Las Cuevas, Hosseus 2152, 2168 and 2230 (BD); Paramillo de las Cuevas, 3800 m. alt., Bettfreund 261 (BD); Punta de Vacas, King 690 (BM); Queb. Benjamin Matienzo, Las Cuevas, Perez Moreau 12632 (G); Las Lenias, Gillies (K); Las Lenfas and El Cerro de la Polcura, Gillies (type of P. Arnottiana, K); Queb. de las Cauteras, Las Heras, 1200-1600 m. alt., Semper 4222 (G); Valle de las Relinchos, 3400-3600 m. alt., Semper 4921 and 4922 (G); Tres Quebradas, Tupungato, 2840 m. alt., Ruiz Leal 3613 (G); precord. near Rincon Colorado, Tunuyan, 2500-3000 m. alt., Ruiz Leal 1307 (G); near Cuesta de los Aflijidos, Cord. del Portillo, Ruiz Leal 1876 and 1876 A-B (G). A well known and readily recognized species of the central cordilleras of Argentina and Chile. In its area it is the only species with mottled, glabrous, inflated pods. It appears to be common in the Andes of Mendoza below Uspallata Pass. From this area came the type of 4. Arnottianus. Also from this region came the plants discussed and well =e i = A. oreophilus) by Hauman, | Veg. Hautes Cord, 126] Anal. Soc. Cie Argentin a 86: 280, t. 24, f. 4-5 (1918). The oa is ails related to A. palenae which occurs along the cordilleras further to the south. From this relative A. Arnottianus is distinguished by having the leaflets more or less evidently strigose on the upper surface and the wings of the corolla always conspicuously longer than the keel. The fruit is always glabrous. At high altitudes the species becomes reduced in size and very compact in habit. One of these dwarfed extremes appears to be that described by Philippi as Phaca Reedi. 18. Astragalus palenae (Phil.) Reiche, Anal. Univ. Chile 97: 542 (May 1897), and 2:85 (1898); Spegazzini, Rev. Fac. Agron. y Vet. La Plata 3: 600 (Sept. 1897). ? Phaca any Philippi, Linnaea 28: 681 (1857).— Type collected by Germain in the cordilleras east of Linares, prov. Maule. ? Astragalus oreophilus (Phil.) Reiche, Anal. Univ. Chile 97: 561 (1897), and sha hand Philippi, Anal. Univ. ma 84: 23 (1893).— Type from the Valley o Palena, ca. lat. 44° S., Delfin oe me Philippi, Anal: Univ, chi 84: 23 (1893).— Type from Laguna de Gualletué, prov. Gautin, 1887, Rahm Astragalus Rahmeri (Phil.) Reiche, pee Univ. Chile 97: 542 (1897), and FI. Chile 2: 85 (1898). 1947] JOHNSTON, ASTRAGALUS 359 Astragalus pallens Reiche ex K. Schum, in Just, Bot. Jahresber. 261: 352 (1900), ] calami Astragalus esis var. grandiflora Spegazzini, Anal. Mus. Nac. Buenos oy ser. 2, 4: 267 (1902).— Type collected by Spegazzini near oe Nahuel-hua Rance: Along the cordilleras from Neuquen and rea (and _ possibly ar southward to the drainage of the Rio Palena and Chub CHILE. ?Maule: cord. Maule, Germain as A. ae (K, BM, Del). utin: “Araucania” [ ?Laguna a Gualletué, upper Bio Bio] ex Philippi as P. Rahmeri (K, BD ARGENTINA. Weraae San Martin de los Andes, 720 m. alt., Comber 791 (K); Zapala district, Opozo sub Comber 1160 (K); Pino Hachado, 2000 m. alt., 1920, Parodi 2038 in pt. (Parodi). Rio beige Puerto Americano near Bariloche, 1928, Spegazzini (Speg, type of var. grandiflora). Chubut: near Carren-leofu, 1901, Illin (Speg); near Carren-leofu, 1889, Moyano (Speg); Pampa Chica, Nov. 12, 1908, Skottsberg 567 (Stock); Chubut, Dec. 1897, collector? (Speg); Patagonia, Moreno & Tonini 439 as Trag. Cruckshanksii (NY). 18a. Astragalus palenae var. Duseni (Macloskie), comb. Astragalus brevicaulis Dusen, Sven. Exped. Ma any 3: no. 5, pg. 158 (1900). — Type from Cabo Domingo, eastern Fuegia, Duse Astragalus Duseni Macloskie, FI. slay pot 2s 6 (1906).— Based upon A. brevicaulis Dusen (1900), not Nelson (189 Rance: Eastern Fuegia northward to ear Chubut. ARGENTINA. Santa Cruz: Killik Aike, Rio Gallegos, 1900, Brown 21 and 54 (NY); Cape Fairweather, mouth of Rio Gallegos, Capt. King (K, BM); Rio Coyle, Estab. Las Vegas, 1916, Dauber 130 (Parodi) ; Puerto Santa Cruz, Dec. 1904, Dusen 5481 aed: Rio Santa Cruz, 1882, Spegazzini (Speg); Lago Argentino, Furlong 81 (G, NY, K); Burmeister Peninsula, Lago Argentino, Prichard (BM); Lago San Mar- tin, el collector ? (Speg); San Julian, 1833, Henslow 80 (K); San Julian, 1904, Dusen (Stock); San Julian, 1931, Blake 2, 172 and 180 (K); Golfo de San Jorge, Feb. 1896, Ameghino (Speg, as A. Meiea bind Chubut: Rio Aysen, Dec. 1900, collector ? (Speg); Lago Blanco, Dec. 1903, collector ? (Speg); Chubut, Jan. 1898, collector ? (Speg Among the species with inflated fruit the present species is characterized by its austral distribution, its folded leaflets that are glabrous above, an its corolla-wings which are shorter than the keel or barely surpass it. Typical A. palenae comes from the valley of the Rio Palena (or the Rio Carrén-leofu as it is called within Argentine). This form has the wings of the corolla evidently surpassed by the keel. The corolla is about 11 mm. long. Further north, about Lake Nahuel-huapi and the headwaters of the Rio Bio Bio, a coarser form is common, This has larger leaflets and coarser (13-16 mm. long) corollas and is represented by the type of the var. grandiflora Speg. I have accepted this variety as a luxuriant form of the species. Typical A. palenae is known from the same general region in which the variety has been found. Beginning in southwestern Chubut and extending to Fuegia there are plants that agree well with typical A. palenae except that the wings and keel are subequal. These plants represent the variety Duseni. In the collection cited from Lago Blanco half of the plants in it have corollas of typical A. palenae and the other half those of the var. Duseni. They are 360 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII otherwise indistinguishable. The ovary and fruit of A. palenae and its var. Duseni are usually glabrous. As exceptions to this general rule, however, there is to be noted a collection from San Julian (Blake 180) in which some of the plants have glabrous and the others have densely strigose ovaries. A collection (Chubut, Jan. 1898) in the Spegazzini Herbarium has the ovaries and also mature fruit conspicuously strigose. I have seen no authentic material of Phaca oreophila but I suspect that it may represent what is here treated as A. palenae. The descriptions of A. oreophilus given by Philippi and by Reiche seem to apply equally well to both A. Arnottianus and A. palenae. The type of A. oreophilus unfor- tunately lacks corollas. It came from the cordilleras above Linares in what is now the province of Maule. Should A. oreophilus prove actually conspecific with A. palenae then the former name, being older, must be taken up as the correct appellation for the concept here treated. It is possible that the species may have another old name in Phaca quindecimjuga Phil. (1862), which was based on collections said to be from the cordilleras of Coquimbo though perhaps actually from the moun- tains of southern Chile. Further details regarding P. quindecimjuga will be found in the discussion of unplaced species. . Astragalus argentinus Manganaro, Anal. Soc. Cien. Argentina 87: fig. 14 (1919). — Original collections from Sierra del Tandil, Sierra erie ie Sierra de Curamalal. RANGE: “oii only from the mountains in the southern half of the province of an Air NT Buenos Aires: Sierra de la Ventana, Nov. 1904, Dusen 6279 rks oa Tornquisto, Dec. 21, 1899, collector not given (Speg). Though this very well marked local species was compared with A. Bergii by Manganaro, it certainly cannot be closely related to that plant. While showing evident differences in the proportions of its flowers, A. argentinus reveals its obvious relations with A. Cruckshanksii in its tough leathery fruit and salient pod-sutures. 20. onary Illinii, nom. nov agalus Moyanoi var. villosul Spegazzini, Anal. Mus. Nac. Buenos Aires 7: 267 (1902) .— ‘Type from along the Rio Carren- -leofu, 1900, Illin RANGE: Known only from Ae : ARG NA. Chubut: Carren-leofu, 1900, .V. Jilin (Speg, type); Chubut, 1903, collector ? (Speg). A species apparently most closely related to A. Cruckshankswu but differ- ing conspicuously in its indument and fruit. The leaves, fruit, an younger parts of the plant are canescent with a subtomentose indument of very slender loosely appressed hairs 1-2 mm. long. The angularly subovate or subglobose fruit is evidently higher (dorso-ventrally) than long or thick. It usually measures 10 X 6 X 6mm. The valves though highly convex are compressed at their very margin. This consequently elevates the keel which encircles the fruit. The fruit is one-celled and has no false septum. When mature it is tough, rigid and filled with seeds. 1947] JOHNSTON, ASTRAGALUS 361 21. eta pee Cruckshanksii (H. & A.) Grisebach, Abhand. K. Ges. Wiss. Goettingen 3.018793 ase Ae aaa Reiche, Anal. Univ. Chile 97: 558 (1897), t Chile 2: 101 (18 Phaca Cruckshanksii Ls & Arnott, Bot. Misc. 3: 184 (1832); Philippi, Linnaea 33:41 (1864).— Type collected by Cruckshank between Tambillos and Puente aA oti cordillera of Mendoza, not in the “cordilleras of Chile” as originally Phaca Landbech PtP, Linnaea 33: 42 (1864).— “In Andibus prov. Colchaguae legit o andbec ee Londbecki ‘(Phil.) Reiche, Anal. Univ. Chile 97: 562 (1897), and FI. Chile 2: 105 (1898). Cordilleras from Atacama and Mendoza south to Colchagua and Neuquen. Queb. Alfalfa, dept. Vallenar, 3200 m. alt., Johnston 5991 (G, K); Rio Valeriano at Juntas del Encierro, 3200 m wilt. FoRaion 6015 (G, US); Los Cuartitos, 3420, Johnston 6228 (G). Banos del Toro, 3500 m D); high cordilleras of Los Patos, 2537 m. alt., Sas ( concagua: El Penon, Poeppig 99/18 (BD, Paris); Juncal, 2100 m. alt., Elliott 164 (K); Las Calaveras, 3100 m. alt., Buchtien 1195 (G, US, BM, BD) ; ; . Santiago: Cone de la Dehesa, Nov. 1929, Gay 533 (Paris); Valle Ramon, 2800 m. alt., Feb. 1933, Grandjot (Grandjot); Potrero Grande, 2200 Mipcalt., ahi 1933, Grandict (G); Valle Engorda, 2400 m. alt., near San Gabriel, Maipo valley, 3000 m. alt., Grandjot 3586 (G). Colchagu San “hee 1800 m . alt., be 1936, Cord. d euieeyi Tambillos and Puente del Inca, Cruckshank 94 (type of P. aaa Kew) ie Perez Moreau 12628 (G); Puente del Inca, Malme 2817 (US); Las Cuevas, »2 Moreau 12633 (G); Cerro del Diamante and Cerro de la Polcura, Gillies (K) ; Pian de los Aflijidos, Ruiz Leal 1877 (G); Real de Contreras, Valle Tunuyan Ruiz Leal 2120 (G); Valle Tunuyan, Ruiz Leal 1925 (G); Los Guanacos, 2000 m. alt., Carette 3000 (LP); Mala Dormida, 2700 m. alt., Carette 240 (G); Cajon del Burro, Valle Atuel, 2900 m. alt., Wilczek 115 (Boiss, US); Piedra del Burrero, 2900 m. alt., Wilczek 118 (Boiss, US); Paso de la Cruz Piedra, 2500 m. alt., Kuntze as T. Philip- piana (NY). Neuquen: Liu Cullin, 1350, Comber 309 (K); Pino Hachado, Parodi 2198 (Parodi); Vega Lolog, 810 m. alt., Comber 824 (K). As here defined the name A. Cruckshanksii is restricted to the well known and characteristic plant of middle altitudes in the central Andes of Chile and Argentina. It is closely related to A. Amunategui and, in the southern parts of its area, perhaps intergrades with that more southerly and easterly ranging species. It differs from its relative in its more dis- tinctly montane habitats, its prevailingly prostrate stems, and its elliptic or oblanceolate obtuse or emarginate leaflets. species was first col- lected and is best known along the road to Uspallata Pass in Mendoza. Plants from the area have been illustrated and discussed by Hauman, [Veg, Hautes Cord. 125] Anal. Soc. Cien. Argentina 86: 289. t. 21, f. 5 (1918). Chilean specimens and those from Mendoza are rather uniform in appearance and usually have evidently strigose leaflets and fruit. The specimens cited from Neuquen, however, have the ovary and fruit glabrous and two of the collections an Cullin and Pino Hachado) have practically glabrous leaflets. These southern forms have the leaflets of A. Cruck- 362 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII shanksii and seem more closely related to that species than to A. Amuna- tegui. I suspect that they may prove to represent a variation worthy of some taxonomic recognition. It will be noted that they come from an alitude distinctly lower than that affected by good A. Cruckshanksii. The fruit of typical A. Cruckshanksii is very strongly compressed later- ally, the elliptic leathery valves being almost flat and parallel. It is noteworthy, therefore, that in one of the collections from Neuquen (Parodi 2198), the valves are not flat but convex and the pod evidently inflated, being lenticular in cross-section and compressed only in a narrow rim about the margin of the valves. The fruit, in fact, is very similar to that found in some specimens (Carette 242 and 243) of A. Amunategut. 22. Astragalus Amunategui Philippi, Anal. Univ. Chile 84: 31 (1893).— Type from Las Choicas a. Amnatieidonas Philippi ex Reiche, ace Univ. Chile 97: 560 (1897), Chile 2: 103 (1898).—A variant spellin meni Cruckshanksii var. glabrescens Kuntze, "Rev, Gen. 3:73 (1898).— Type from Chilean side of the Paso de la Cruz Piedra, 2500 m. alt., Kuntze. Astragalus Moyanoi Spegazzini, Rev. Fac. Agron. y Vet. La Plata 3: 601 (1897). — Type from Teka-choique, western Chubut, ca. lat. 43° 30’, Moya RANGE: Southern Mendoza and adjoining Santiago south into ae. CHILE. Santiago: Paso de la Cruz Piedra, 2500 m. alt., Kuntze (type of var. glabrescens, NY; isotypes US, BD). Colchagua: “Cord. de San Fernando,” Philippi as A. Amunateguianus (US, K, BD). ARGENTINA. Mendoza: Las Choleas: Jan. 1872, ex Philippi as A. Amuna- teguianus (Speg); Cuesta de los Aflijidos, Ruiz Leal 1877 (G); Cerro de Guanacos, 2 m. alt., Carette 242 and 243 (LP); Cerro Nevado, dept. S. Rafael, Carette 2552 (G). Neuquen: Mangrullo, 900 m. alt., Ammann 76 (FM); Laguna Llancanelo, Nov. 1902, herb. Spegazzini 2145 (Speg). Chubut: Rio Teka-choique, 1889, (type of A. Moyanoi, Speg); Comodoro Rivadavia, Renard 13804 (G), Ferruglio 30/1859 (G), and Parodi 289 (Parodi) ; Puerto Madryn, Dusen 5374 (Stock). The plants here assembled under A. Amunategui come from an area lying between the ranges of A. patagonicus and A. Cruckshanksii and perhaps intergrade with both species. ‘Their closest relationship seems to be with A. patagonicus. Indeed they may prove to be simply a luxuriant northern form of that southern species. Typical A. Amunategui is a plant having the gross habit of A. Cruck- shanksii but differing in its lanceolate leaflets and somewhat inflated strigose pods. The flowers and fruit are bunched in a subumbellate cluster on peduncles that are distinctly recurved at maturity. Specimens very similar to the type have been collected at Cerro Guanacos and Cerro Nevado, Mendoza, by Carette. The other collections referred to the species differ from those distributed by Philippi and Carette in being more erect plants with elongate loosely flowered racemes on erect peduncles. but differ in some details. The type of A. Movyanoi has the ovary and the leaflets glabrous whereas these are strigose and the leaflets less definitely acute in the var. glabrescens. The collections from Comodoro Rivadavia and Puerto Madryn much resemble one another and are very slender plants 1947 | JOHNSTON, ASTRAGALUS 363 with very few erect stems and very slender possibly annual roots. The racemes are loosely flowered and elongate and borne on erect peduncles. This particular form is very similar to some of the luxuriant forms of A. patagonicus. When Philippi published 4. Amunategui he stated that it was named in memory of Michal L. Amunatégus and gave its source as “Habitat in Andium loco Las Choicas dicto.” I have seen a number of collections, scattered in various herbaria, distributed by Philippi as “A. Amunategui- anus Phil.” These are mostly labeled as from the “Cordillera de San Fernando,” but one is given as from ‘‘Las Choicas, Jan. 1872.” All these specimens are extremely similar in details of size, maturity, discoloration etc., and I belreve them to be parts of one collection. They agree closely with the original description and almost certainly are duplicates of the type. Their source is probably a short distance east of the Chile- Argentina boundary in Mendoza. There is a well known route from San Fernando, Colchagua, up the Rio Tinguiririca to the Paso de las Damas on the continental divide. Descending eastward from the pass the road crosses Arroyo de Las Choicas and passes close to the Mina de Las Choicas. The type of the Astragalus was probably obtained in this vicinity. In Jan. 1872 Paul Ortega traveled this route to Valle Hermosa (about 25 km, south of Choicas) and collected the types of Gayophytum robustum Phil., Verbena ulicina Phil., Boopsis breviscapa Phil., Carex vallis-pulchrae Phil., etc., and very likely the original material of our plant also. As now known the species is practically confined to Argentina. The only Chilean material seen is the collection by Kuntze, the type of the var. eas labeled “Chile, Paso Cruz, 2500 m., Jan. 1892.” If the data on the label are trustworthy the plant came from 1000 m. below the summit of Paso de la Cruz Piedra in the headwaters of the Rio Maipo, in the province of Santiago. Interestingly, among all the specimens referred to A, Amunategut, this Chilean one has characters most closely approaching those found in A. Cruckshanksii. 23. Astragalus patagonicus (Phil.) Spegazzini, Rev. Fac. Agron. y Vet. La Plata 5 97). ape patagonica Philippi, Anal. Univ. Chile 84: 20 (1893).— “In Patagonia haud cul a colonia chilensi Decembri 1878 legit orn. Henricus Ibar Beas Southern Patagonia. CHILE. Magallanes: Patagonia, ex Philippi as P. patagonica (isotype, K, peg). ARGENTINA. Santa Cruz: Killik Aike, Rio Gallegos, Brown 27 (NY); Estab. Las Vegas, Rio Coyle, Dauber 129 (G); upper Rio Santa Cruz, Lago Argentino, Hauthal 8786 (Speg); near Lago Argentino, Furlong 82 (G, NY); Rio Santa Cruz, Spegazzini (Speg); Puerto San Julian, Dusen 6332 (K, Stock) and Blake 2A (K, BM); Caleta Olivia, Donat 141 (G in pt.; K, NY); Cafiadon de las Vacas, 1891, Beaufils (G); Patagonia, lat. 50°-53°, Moreno & Tonini 188 and 206 (NY). hubut: Valle de la Laguna Blanca, Koslowsky 178 and 183 (K); Pampa Chica, Skottsberg 539 (Stock); Lago Musters, 1899, Illin (Speg) ; Lago Colhue Huapi, 540 . alt., Riggs 44 (G, FM); Corcovado, 1901, Jllin (Speg); along Rio Carren-leofu, 1900, /llin (Speg). Neuquen: Cord. del Viento, subito de Atrenco, Rengonese 95 (G) 364 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII This species appears to be most common in Santa Cruz and southern Chubut. The plants are dwarfed with the leaf-bearing seasonal stems only a few centimeters long and with the internodes very much shortened or scarcely developed. A multicipital caudex is formed from which arise tufts of leaves and the scapose peduncles. The flowers are usually few and borne in umbellate clusters. The flattened pod may be strigose, but frequently is more or less glabrous. The leaflets are narrow, 1-3 mm. broad, and linear. Most of the collections cited are readily placed and obviously referable to A. patagonicus. A few specimens from Santa Cruz (Furlong 82; Dusen 6332) and Chubut (Lago Musters and Carren-leofu, Illin) are more vigorous plants with more elongate stems than the other specimens cited and make embarrassing approaches to the various forms of A. Amunategut. 24. Astragalus Burkartii, sp. nov. Herba caespitosa; ramulis caudicis Nas congeste breviterque ra- mosis 3—5 mm. crassis, stipulis laxe ochraceis 3-5 mm. longis glabris vel breviter ciliatis marcescentibus chartaceis pallidis dense vestitis; foliis apice caudicis stipulosi congestis; rhachibus 10-30 O) mn longi : presseque villosis apice rotundis vel retusis; floribus 2—4 in racemos brevis- onge pedunculatos dispositis: calycibus sparse adpresse villosis 8-10 mm. longis tubo 2-3 mm. crasso 1-2 mm, longe pedicellato dentibus gracilibus 2—5 mm. longis tubo brevioribus; corolla conspicua violacea vexillo 15 mm. longo, lamina late elliptica 8— 10 mm. longa ca. 7 mm, lata, alis 3-4 mm. latis vexillo 2 mm. brevioribus carina 2-3 mm. superantibus; ovario glabro; leguminibus evidentibus ascendentibus bicon- vexis glabris Dileniedns basim versus crassioribus lateraliter compressis margine acutis prominentibus; valvulis 10-15 mm, longis 5-7 mm. latis medium versus vel paulo infra medium latioribus apice acutis recte api- culatis basi obtusis vel rotundis intus sparse dade vee as 10-15 ccna maculatis compressis ca. 2.5 mm. diam AR NA. Tucuman: - Penas Azules, ger at ules, oak among dry ies fl. rather mea blue, 3400 m. alt., Jan. 1 * ata 5403 (G). Cata- arca: Cerro de la Tambilla, low and sorsading: ‘i ple blue, March 25, 1917, Joergensen 1337 (Gray Herb., type); Sierra del Anconquija, dept. Santa Maria, rocky places, fl. bright violet, 4600 m. alt., March 1, 1925, Venturi 6693 (US). Salta: Cerro de Cachi, dept. San Carlos, rocky places, fl. bright violet, 4500 m. alt., Marc 12, 1927, Venturi 6692 (US); between Cuesta del Arca and Trancas, Jan. 1897, Spegazzini 2100 and 2102 (Speg); Pampa Grande, Jan. 1897, Spegazzini 2103 (Speg). The nearest relative of A. Burkartii is probably A. Venturit, a poorly ‘understood and perplexing species also occurring at high altitudes in north- western Argentina. From that species it is readily separated by having flowers almost twice as large, usually more numerous leaflets, a sparser pale rather than yellowish indument, a more compacted caudex with less elongate stems, and looser pale rather than yellowish sheathing stipules. The fruit of A. Burkartii appears to be much larger and its sutures are both salient. They are laterally compressed and ovate-elliptic in outline. No traces of a false septum are present. 1947] JOHNSTON, ASTRAGALUS 365 25. Astragalus erypticus, sp. nov. Planta humilis caespitosa; caudice multicepitali denso e radice valida rofunda oriente, ramulis stipulis ochreaceis imbricatis obtectis apice fasciculos foliorum dense argenteo-strigosorum proferentibus; stipulis lobis angustis 2-3 mm. fener ok flavescente, vexillo ca. 6-7 mm longo cum lamina 3.5-4 mm. lata, alis 5-6 mm. longis angustis quam carina 1.5 mm. longioribus, leguminibus inter stipulis occultis 3-4 mm. longis 1-1.7 mm. crassis 2—3 mm. altis dense appressi-villosis, margine vix ee suturis prominentibus donatis, valvulis cartilagineis ellipticis vel sae ovatis; seminibus saepe ie INA, Tucuman: aoe Calchaquies, dept. Tafi, en la puna tucu- mana, 4400 m. alt., or he 1933, Burkart 5407 (type, Gray Herb.) ; La Puerta, Sierra Calchaquies, 4000 m , Jan. 30, 1933, Burkart 5405 (G); Sierra del Cajon, dept. , mountain en among rocks, fl. yellowish, 3500 m. alt., March 12, 1927, Venturi 9435 (US). A very distinct species with its closest relations in A. cryptobotrys of northwestern Argentina and adjacent Chile. The two species agree in fruit-structure. Their pods are more or less lenticular with both sutures salient. The valves are cartilaginous and abundantly covered with similar appressed silky hairs. There is no false septum developed. The two species differ, however, in size of fruit, habit of growth, and shape of leaflets. In gross habit our present plant much simulates that of 4. peruvianus and its allies. 26. Astragalus cryptobotrys, nom. no Phaca clandestina Philippi, F1. oa 14 (1860).— Type from Valle Sandon, 2700 m. alt., Philip pi. Adievalis clandestinus (Phil.) Hierony mus, one Acad. Nat. Cien. Cordoba 4: [Sert. Sanjuan.] 22 (1881); Reiche, Anal. Univ. Chile 97: 537 (1897), and FI. Chile Phaca cryptantha Philippi, Fl. Itin. Tarapaca. 14 (1891).— Type from between guas Calientes and Socaire (prov. Antofagasta, east of Salar de Atacama), 3000 ?Astragalus clandestinus var. Flossdorfii Hicken, Darwiniana 1: 28 (1922). — Type from Queb. de Encrucijada, Sierra Famatina, La Rioja, 4000 m. alt., March 1913, A In Chile known only from the puna in eastern Antofagasta. In apawes ranging at high altitudes from Salta (and Los Andes) south to Tucuman and San Jua CHILE. Antofagasta: Atacama Desert (betw. Ag. Calientes and as, Jan, 1885, ex Philippi as P. nen (BD) ; Aguada vere 4000 m. alt., Werdermann 1034 (G, K, US, BD); Cord. Domeyko, 1925, Pe 487 (BD); Atacama Desert (Valle Sandoni) Philippi as P. clandestina (K, 7 Boiss) ; Guanaqueros to Agua Delgada, Feb. 5, 1901, visas? as A. bolivianus in ARGENTINA. Salta: Cerro de Cachi, 4000- 4500. m. alt., 1927, Venturi 6703 (US). Catamarca: ee o Yutuyaco, north of Andalgala, 3500 m. alt., Joergensen 1627 in pt. (US); Sierra renee ja, 4600 m. alt., 1924, Venturi 6624 (US) : Junta de los Reales Blancos, 1930, ome 39/498 (G). Tucuman: Las Lagunas, dept. Tafi, 4400 m. alt., 1933, Parodi 10829 (G) ; Cumbres Calchaquies, dept. Tafi, 4000-44CO 366 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII m. alt., 1933, Burkart 5399, 5400 and 5401 (G); Est. Santa Rosa, dept. Chicligasta, 4600 m. alt., 1925, Venturi 4008 (US); Est. Las Pavas, dept. Chicligasta, 3500 m 1926, Venturi 4711 (US); El Pelado, 1912, Rodriguez 450 (G). La Rioja (Sierra Famatin a): Alto Blanco, rege cick ieeee 28/188 (G); Mina Jareta, 1879, Hieronymus & Niederlein 806 (K, BD); between Mina Jareta and Alto de Espiritu Santo, 1879, Hieronymus & send 795 (BD) ; Cueva de Perez, 1879, Hieronymus & Niederlein 358 (BD); Tocino io 4320 m. alt., 1915, nae 1911 (BD); below Tocino Pass, 3900-4320 m. alt., , Hosseus 1903 (BD San Juan: Paramillo, southwest of San Juan, Jan. 7 oe (BD) ; tee ‘Colangiiil, Queb. del Salto, 1930, Perez-Moreau 30/104 (G). A plant much resembling A. geminiflorus of Ecuador in gross habit. That northern plant, however, has an incompletely 2-celled fruit with the lower suture inflexed, smaller more crowded distinctly silvery silky leaflets, short petioles, and a somewhat more trailing habit of growth. I do not believe that there is any close relationship between A. geminiflorus and A. cryptobotrys. The latter, in any case, has its closest relations in A. crypticus and A. Asplundii. Reiche has cited Phaca diminutiva Phil., which was collected with the type of P. cryptantha Phil., as a synonym of the present species. The plants described as A. diminutiva, however, seem so very much smaller than the most reduced forms of the present species known to me that I hesitate to accept it as a synonym. In any case the name cannot be transferred to Astragalus since the resulting binomial is preoccupied by another valid species. 27. oo Asplundii, sp. nov. Herba perennis incana; caulibus numerosis e caudice humili laxe ramoso orientibus 5-10 cm. longis simplicibus tomentosis, internodiis 3-20 m longis; rhachi folii 4-8 mm. longo; foliolis 5—8-jugis lanceolatis vel eee lato-ovatis 3-10 mm. longis 2—4 mm. latis basim versus latioribus apicem versus gradatim attenuatis concoloribus planis utrinque villoso-tomentosis; stipulis laxe ochreaceis grandis chartaceis 4-6 mm. longis; pedunculis axillaribus 5-15 mm. longis quam foliis conspicue brevioribus laxe ascen- dentibus ad apicem flores 3-7 capitato-congestos productis; calycibus subsessilibus ca. 5 mm. longis, lobis subulatis ca. 2 mm. longis; corolla ca. 14 mm. longa; leguminibus sublenticularibus dense villoso- -strigosis unilo- cularibus 6-7 mm. longis 5—6 mm. altis 3-4 mm. crassis sub medium crassioribus et altioribus basi oblique rotundis apice obtusis margine angulatis vix inflexis. BOLIVIA: Uyuni, Potosi, 3700 m. alt., March 24, 1921, Asplund 6132 (type, U. S. Nat. Herb.) and 3093 (Upsala); Uyuni, 3700 m. alt., March 25, 1921, Asplund 6131 (US) and 3161 (Upsala); Uyuni, 3660 m. alt., Feb. 23, 1903, Hicken 12 (Stock). A very distinct species with evident relations in A. cryptobotrys. From the latter it differs in its larger more elongate acute leaflets and its clustered flowers borne on obvious axillary peduncles. The species is known only from one locality in southern Bolivia, lying several hundred kilometers to the north of the range of its relative. 28. Astragalus bellus (Kuntze) Fries, Nova Acta R. Soc. Sci. Upsala IV, 1!: 135 (1905). Phaca Angie Clos in Gay, Fl. Chile 2: 92 (1846); — Chlor. Andina 2: 263 (186 - Type from Cordillera de los Patos, Gay 1947 | JOHNSTON, ASTRAGALUS 367 ee fala (Clos) Reiche, Anal. Univ. ee 97: 539 (1897), and FI. Chile . Not Salisb. (1796), nor Boiss. (1843). Pragacantha bella Kuntze, Rev. Gen. 2: 940 (1891). Based upon P. pulchella Clos. RAN Known only near the Chile- ae boundary, between the provinces of Coauimb and San Juan CHILE-/ ARGENTINE BOUNDARY: Cordillera de los Patos, rare about the base of scattered rocks, 3100 m. alt., Gay 490 (type, Paris) ; Chile, Gay (G, K, BD). A small depressed plant with silky strigose stems and leaves. The strigose fruit is several-seeded, compressed about the margins, and nearly as broad as long. The peduncles are 1—3 mm. long and bear loose clusters of several flowers. The pedicels are about 1 mm. long. ‘The species seems to be most closely related to A. cryptobotrys but has much more slender stems, smaller me and less compressed legumes which are compressed about the margin 29. ee kee os Spegazzini, Anal. Mus. Nac. Buenos Aires 7: 268 (1902). m Rio Chico, Chubut, Ameghino. RANGE: eid only from southern Patagonia (Chubut). ARGENTINA. Chubut: Rio Chico, 1900, Ameghino (type, Speg.); Chubut, Dec. 1900, collector not given (Speg.). A very distinct species and one very readily recognizable by its coarse, rather fleshy, conspicuously emarginate leaflets and its elongate fruit 2.5—3 cm. long. The pod is narrowed at both‘ends and is somewhat falcate, the lower margin being concave and the upper one convex in lateral outline. The lower suture is strongly inflexed. There is a very narrow incomplete hyaline false septum. The corollas of the species have not been seen. The coarse calyx, however, suggests that the corollas are probably 10-15 mm. long. The species may have its closest relative in the very different A. Ruiz-Lealii of Mendoza 30. Astragalus chubutensis Spegazzini, Anal. Mus. Nac. Buenos Aires 7: 266 (1907). ed upon collections made by Illin on the Rio Chubut (near Lago Musters) and Rio Carren-leofu RANGE: Known only a Chubut and northern Santa Cruz, Patagonia. ARGENTINA. Chubut: Rio Chubut near Lago Musters, “campo alto cerca del primer manantial,”’ 1899, IJllin (type, Speg.); Carren-leofu, 1900, Jilin (Speg.). Santa Cruz: Caleta Oliva, 325 m. alt., Oct. 23, 1929, Donat 178 (G, NY); Min- erales, 300 m. alt., Oct. 23, 1929, Donat 178 (K, FM, Stock) ; Lago Argentino, upper Rio Santa Cruz, in 1902, Heuthal (Speg.). A low subcaespitose plant with the numerous leaf-bearing shoots 1-2 cm. long. The leaves, stems, peduncles and calyces are densely appressed villous. This indument i is white except on the calyces and stipular sheathes where black hairs may be numerous. The peduncles are 3—5 cm. long and evidently surpass the numerous more or less basal leaves. They are terminated by a capitate cluster of 3-8 blue flowers which become reflexed after anthesis. No mature fruit has been seen. MHauthal’s collection, however, has immature fruits which suggest that when mature they would be about 1 cm. long, straight, and only incompletely exserted from the calyx, 368 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII . Astragalus distinens oo Fl. Patagonia 2: 505 (1905).—A renaming of A. distans Gray, not er. Astragalus distans Gray, Bot S. Explor. Exped. 1: 412 (1854). Not Fischer 3).— Type collected by saa on the dunes at the mouth of the Rio Negro, northern Patagonia RANG From the Rio de la ‘Plata westward and northwestward to San Luis and Jujuy, on southward, probably in the interior, to the lower Rio Negro ARGENTINA. Rio Negro: dunes near mouth of Rio Negro, U. S. Explor. Exped. aie US, G). Buenos Aires: Barrancas del Cazador near Escobar, Nov. 23, 1924, Doello-Jurado 24/1891 (G) ; Campana, limy barrancas along the Rio Parana, fl. ‘bluish, Parodi 4736 (Parodi) and 8609 (G); Burkart 3080 and 5661 (G); Hauman 13807 (G); Castellanos 26/1912 (G). Santa Fe: Arroyo Frias, south of Rosario, fl. blue, Oct. 1929, Cabrera O3t (Stock, G). Corrientes: without locality, 1821, Bonpland (Paris). Cordoba: Achiras, fl. bluish white, 800 m. alt., 1930, King 735 (BM); Capilla por Los Cocos, Nov. 1915, Hauman 13789 (G). San Luis: Cerro del Portezuelo, Nov. 1820, Gillies (K) ; San Luis, Jan. 1910, appa (Speg); La Guardia, March 23, 1882, Galander (BD). La Rioja: near Los Corrales, Sierra Famatina, 1879, Hisense & Niederlein (BD). Catamarca: Sierra del An- conquija, dept. Santa Maria, fl. yellowish, in rocky places, 4200 m. alt., Jan. 1926, Venturi 6628 sage {?] Los Hoyitos, Feb. 7, 1930, Castellanos 30/491 (G). Tucuman: betw. Amaicha and El Molle, dept. Tafi, dry place, fl. dark blue, Feb. 3, 1933, Burkart ne (G); between Amaicha and El Molle, Loree m. alt., fl. pale blue, Feb., 1933, Burkart 5391 (G); Amaicha del Valle, dept. Tafi, 2300 m. alt., dry rocky valley, fl. violaceous white, Jan. 31, 1933, Burkavt 5392 “e oa Tileara, “en los pajonales del cerro,” fl. white, 3000 m. alt., Feb. 19, 1927, Venturi 7422 (US); Cerro La Solidad, dept. Humahuaca, “en los pastigales,” fl. cathy blue, 3000 m. alt., Jan. 23, 1929, Venturi 8894 (G UAY. Carasco, Montevideo, Dec. 1869, Fruchard (Paris); Montevideo, sandy place, Gibert 391 (K); Banados de Santiago a 1937, Rosengurtt 554 (G) ; Arenal Grande, April 1876, Arechavaleta (BD); on face of very dry clay banks at Cape Fray Bentos, Rio Uruguay, Tweedie (K); Fray Bentos, 1933, pte 1833 (G). The plant here treated is a very distinct one which is quickly recogniz- able by its unique fruit. It is most closely related to the more southerly ranging, coastal A. Bergii with which it agrees very closely in all structures save its fruit. The type of A. distinens is given by Gray, l.c., as from the “Rio Negro, North Patagonia; on sand-hills” where it is said to have been collected with A. Bergii (=A. distans var. B Gray, l.c.). Pickering, Geogr. Distr. Animals and Plants 77 (1870), who was one of the collec- tors, indicates that Astragalus was collected on the “tract of sand-hillocks” at the mouth of the Rio Negro and presumably those to the southward of the anchorage at the river mouth. Except for the type, I have seen no material of A. distinens labeled as coming from south of northernmost Buenos Aires and northern San Luis. Hauman, Anal. Mus. Nac. Buenos Aires 24: 396 (1913), and Manganaro, Anal. Soc, Cien. Argentina 87: 149 (1919), however, report “A. Bergii” in western parts of the province of Buenos Aires (Nuevo Plata, and Rivera) towards the La Pampa boundary. ‘There is a possibility that these records may refer to material of A. distinens and so indicate an extension of that species from San Luis and Cordoba southward in the interior towards the lower Rio Negro. It is difficult to believe that the types of two such closely related species as A. distinens and A. Bergii could have both come from the mouth of the 1947] JOHNSTON, ASTRAGALUS 369 Rio Negro. The species concerned are so closely allied that hybridization could be expected if they grew in the same area. The type of A. distinens is thoroughly typical of the species as here accepted and is said to have been collected with plants that are equally typical of A. Bergi. It was collected by members of the U.S. Exploring Expeditions, under Capt. Wilkes, while the ships anchored at the mouth of the Rio Negro during the final week of Jan. 1839. No other landing was made by the expedition between Rio de Janeiro and the Straits of Magellan. The material of 4. Bergii, collected by the expedition, probably did come from near the anchorage, since that species has been found there subsequently by various botanists. The material of A. distinens, however, was probably collected further north, possibly during some overland excursion made while the ships were at anchor. The material collected of the two species is very mature and has ripened pods. That of A. distinens is slightly less matured and not so dried and weathered as that of A. Bergii. 32. Astragalus Bergii Hieronymus, [Sert. Patagon. 17] Bol. . Nac. Cordoba 3: 343 1 ).— Type from near mouth of the Rio Negro, ao distans var. B Gray, Bot. U. S. Explor. Exped, “412 (1854). — Type { ear mouth of the Rio Negro. Rance: Known only from the southern sane . the province of Buenos Aires where it seems to be confined to sands near the c ARGENTINA. Buenos Aires: mouth of i“ = Negro, Oct. 2, 1874, Berg 79 (type of A. Bergii, BD) ; near mouth of the Rio Negro, U. S. Explor. Exped. (type, US, G) ; El Carbon, Carmen de Patagones, Feb. 1898, Spegazzini (Speg) ; Bahia Blanca por San Blas, 1903, Ameghino 13806 (G); Baterias, Bahia Blanca, Dec. 1901, Spegazzinz Speg) ; Monte Hermoso, 1916, Carette (G); Sierra de la Ventana, Nov. 1904, Dusen 6272 (Stock); Sierra de Curamalal, Arroyo Cochenleuft, fl. blue, Nov. 1939, Cabrera 5490 (G). The plant treated here has been confused with that which I have called A. distinens. The notes by Spegazzini, Anal. Soc. Cien. Argent. 47: 233 (1899), and by Hauman, Anal. Mus. Nac. Buenos Aires 24: 396 (1913), apply to both species. Manganaro, Anal. Soc. Cien. Argent. 87: 149, fig. 15 (1919), however, while giving a general range that includes A. distinens, seems to have figured and described plants referable to A. Bergii as here accepted. She states, however, that she examined specimens from Rio Negro, Bahia Blanca, Sierra Ventata and Nueva Plata. I have seen no material from as far north as Nueva Plata (Partido Pehuajo, prov. Buenos Aires). Monticelli, Lilloa 3: 345, fig. 22 (1938), reporting A. Berga from La Pampa, illustrates a plant which seems to represent that species. A specimen from him, collected at Telén, La Pampa, showing flowers but no fruit, bears the annotation “fragmento del ilustrado en mi trabajo.” The acceptance of this record would extend the known range of 4. Bergi from southwestern Buenos Aires northwestward into La Pampa The type of A. Bergii, which I examined at Berlin, is ecied as collected from the shade of shrubs of the Patagonian campo near the Rio Negro on Oct. 2, 1874. According to Berg, Peterm. Geogr. Mittheilungen 21: 367 (1875), he was on the brig ‘‘Rosales” which sailed down the Rio Negro from Carmen de Patagones on Sept. 29, 1874. After a halt of three days 370 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII at the mouth of the Rio Negro the vessel arrived at the Rio Santa Cruz on Oct. 8th. On Oct. 2nd, the date on which the type was collected, Berg accordingly must have been at the mouth of the Rio Negro. I am unable to distinguish flowering, non-fruiting plants of the two species I am calling A. distinens and A. Bergii. Since Berg’s type is at anthesis, and entirely lacks fruit, I have arbitrarily associated the name 4. Bergii” with what appears to be the common species about the mouth of the Rio Negro, This application of the name, however, must remain a practical expedient, designed to avoid the coining of a new botanical name, until more collecting is done about the lower Rio Negro or until someone discovers floral or vegetative characters whereby the type of A. Bergii can be positively assigned to one or the other of the two species concerned here, for, as noted in my discussion of A. distinens, the type of this latter species also came from the mouth of the Rio Negro and possibly may represent the species collected by Berg. 33. Astragalus Spegazzinii, nom. nov. Astragalus Rengifoi var. lejocarpa Spegazzini, Anal. Mus. Nac. Buenos Aires 7: 19 from near the confluence of the Rio Limay and Rio ete Spegazzini. Rance: Known only from the upper parts of the watershed of the Rio Negro. ARGENTINA. Rio Negro: ran the confluence of the Rio Limay and Rio Neuquen in dry sandy places, Dec. 1897, Spegazzini bit ae near General Roca, valley of the Rio Negro, in clay soil, ie, Fischer 145 (G, NY, US, BK). In the flowering condition this plant very ie hee some of the intermediates connecting A. Amunategui and A. patagonicus. Fischer’s plant has been reported as A. Cruckshanksii by Hicken, Physis 2: 15 47: 237 (1899), as A. Arnottianus and later, as his proposed new variety, A. Rengifoi var. lejocarpa. The relations of the species are certainly not with A. Cruckshanksii and its allies, nor with A. Arnottianus. The fruit is linear, has an inflexed lower suture, and becomes deflexed in age. Astragalus wane is probably its closest relative. Fischer’s plants, collected at anthesis, lack even partially matured fruit. Spegazzini lection has young fruit. The largest of these are falcate and linear, becoming 2 cm. long, ca. 1.7 mm. eae and strongly flattened laterally. The upper edge is concave and the lower is convex in lateral outline. 34. Astragalus carinatus (H. ) Reiche, Anal. Univ. Chile 97: 557 (May, 1897), nd Fl. Chile 2: 100 aie Kurtz, Bol. Acad. Nac. Cordoba 15: 511 (July 1897). Phaca carinata Hooker & Arnott, eo Misc. 3: 185 (1832).— Type from below Horni en Me nGOrs, Cruckshank 8 Phaca carinata var. 8 Hooker & peer Bot. Misc. 3: 185 (1832). Phaca nas Gillies ex Hooker & Kenont, Bot. Misc. 3: 185 (1832), in synonymy. RANGE: Base of the cordilleras of Mendoza; Cordoba; and perhaps Catamarca. ARGENTINA. Mendoza: below the Hornillos, Cruckshank 89 (type, “eG Portez. del Pantanillo and Portez. de la Casa de Piedra, Gillies (type of var. P, abundant along the descent from Portez. de la Casa de Piedra, Gillies (K) ; aie a los Chacayos, Nov. 3, 1834, Gillies (K); near baths of Villa Vicencio, Nov. 15, 1822, 1947 | JOHNSTON, ASTRAGALUS 371 thes (K); hills above baths of Villa Vicencio, Nov. 15, Gillies (G); Potrerillos. . “El Salto,” fl. blue, 1900 m. alt., Nov. 1933, Ragonese 97 (G); Mendoza, Jan. oe Spegazzini (Speg.); El Alambre Atravezado, Estan. La Vila, Tupungato, Ruiz ae 2799 (G); Puesto del Manzano, Tunuyan, 1700-2700 m. alt., Ruiz Leal 1745 and 2318 (G); Cerro Nevado, dept. San Rafael, Carette 2555 (G). Cordoba: Capital, herb. Poa 2140 (Burkart). In gross habit this plant frequently suggests forms of the very different A. Arnottianus, but may be readily separated by its fruit and by its leaflets which are glabrous on the upper surface. It is a plant of the base of the cordilleras, and not of their higher parts. The fruit is glabrous, deltoid in cross-section, and has its lower surfaces broadly inflexed and flattened. The corolla varies in size from 9-12 mm. in length. The type represents the large-flowered form Perhaps referable to A. carinatus are some collections from Catamarca (Yacutula, Schickendantz 61; Granadillas, Schickendantz 98 and 104; Catamarca, Lorentz 41) which have a form of corolla slightly different from that in typical plants of Mendoza and Cordoba. Also questionably referred to the present species is a collection made by Perez Moreau (no. 30/97) near Las Vicunitas, Queb. de Conconta, in the cordilleras of San Juan. This latter is a caespitose plant with minute leaflets, about 1 mm. long, and reduced scapes only 5-15 mm. high. It ma ¥ be a compacted depauperate montane form, comparable to depauperate nee phases of A. Arnottianus that have been collected near El Portillo, Mendoza. 35. Astragalus Parodii, sp. nov. Herba perennis e radice crassa valida profunda oriens; Se aage 3- - (saepe 5-15) cm. longis foliosis angulatis numerosis gracillim S:Ca...3 Mm crassis decumbentibus onge sparseque ramosis sparse ee iiecuc 5-10 cm. diametr i latis costatis sed enervatis subtus sparse inconspicueque strigosis, supra saepe ae apice retusis, infra medium basim versus contractis; racemis 4—7 floris 1-2 cm. longis laxifloris: pedunculis 1-4 cm, longis foliis 1—2-plo Ty eee calycibus 3—4 mm. longis sparse et pallide nigreque strigosis, lobis linearibus quam tubo paulo. brevioribus; corolla caerulea; vexillo 8-10 mm. longo, lamina obovata 5—6 mm. lata 5—7 mm. longa apice ae alis 7-8 mm. longis quam carina 1-2 mm. longioribus; legu- s subovoideis vel oblongis lateraliter compressis angulatis unilo- aes 5-9 mm. longis 2-3 mm. crassis 2—3.5 mm. altis induratis tarde dehiscentibus pendulis, apice acutis et apiculatis, basi rotundis, infra medium the basim versus crassioribus et latioribus, dorse acutis et summam ad marginem incrassatis et paulo compressis, subtus planis vel leviter une suturis prominentibus; seminibus congestis ca. 10. ANGE: Known only from the Sierra ae — Cordoba. All collections seem to be from the Dept. Punillo near Los Giga ARGENTINA. Cordoba: base ae ies Dec. 2, 1878, Hieronymus (BD); Quebrada at base of Los Gigantes, Jan. 22, 1880, Ealandes (BD); near 372 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII sageene in quebrada at base of Los Gigantes, Dec. 3, 1878, Hieronymus (BD) Sage, Jan. 1, Hieronymus (BD); valley of the Rio Catre below the potrerillos, Feb. 4, 1877, eon 761 (BD, Paris); Cuesta del Gaucho, Jan. 9, 1876, Hieronymus (BD); Copina, moist alpine meadows, fl. blue, 1500 m. alt., Dec. 29, 1935, Burkart 7360 (type, pene erb.) ; between Cuesta de Copina and Las Encenadas, Jan. 8, 1876, Hieronymus (BD); near Pampa de Achala, iia soil, 2000 m. alt., “ig 1926, Parodi 7490 (BD); Pampa de Achala, moist meadows, 2200 m. alt., Dec. 29, 1935, Burkart 7355 (G); Pampa de Achala, 10 km. below Hotel El Condor, ‘rocky meee fl. bluish white, 1000 m. alt., Dec. 16, 1938, Goodspeed 17228 (G). A very distinct species having its closest affinities probably with 4. carinatus. It differs from the latter in its smaller, indurate, very much less angulate pods and fewer, somewhat smaller, more congested flowers, Its tough pods also suggest A. Bergii and A. distinens but these species have much more numerous smaller pale flowers in elongate racemes and curved pods with at least one suture recessive. In lateral outline the pods of A. Parodii are strongly convex above and nearly straight below. In cross-section they are roughly triangular. Their upper edge is prominent being compressed laterally and slightly thickened along the suture. The pods are flattened or weakly and broadly intruded dorsally. The legumes of this species dehisce only very tardily if at all. 36, Astragalus Joergensenii, sp. nov. Herba perennis gracilis laete viridis erecta vel decumbens; caulibus 1.5—7 dm. longis sulcatis sparse strigosis longe ascendenter ramosis foliosis; foliolis 8—-11-jugatis ellipticis vel lanceo-oblongis 3-11 mm. longis latis apice obtusis vel retusis subtus strigosis supra glabris margine plus inusve involutis; pedunculis 4—7 cm. longis foliis aequilongis vel paulo se en apicem versus flores congestos gerentibus; floribus violaceis tate detlexis; calycibus cupulatis 1—-1.5 mm. profundis nigro-strigosis ca. re mm. longe pedicellatis, lobis angustis tubo subaequilongis; vexillo 5—7 mm, longo; alis carina evidenter longioribus; ovario strigoso; legumini- bus unilocularibus refractis 1.5-2 cm. longis 2-4 mm. altis sparse strigosis falcatis chartaceis valde angulatis supra carinatis subtus conspicue lateque inflexis 1-3 mm, crassis; seminibus numerosis Rance: Northwestern Argentina at middle altitudes, ARGENTINA. Catamarca: a Dec. 1909, Spegaszini (Speg); El Can- dado, common, Feb. 10, 1917 and Jan. 20, 1916, Joergensen 1123 (US) and 1126 (type, Gray Herb.). Tucuman: sierras near Tafi del Valle, fl. blue-violet, 2400 m. alt., Feb. 3, 1933, Burkart 5404 (G); Estan. Santa Rosa, Pueblo Viejo, dept. Chicligasta. among grass on ridge, fl. violet, 4000 m. alt., Dec. 22, 1925, Venturi 7513 (US). Salta: Cerro de Cachi, dept. S Carlos, among grass, fl. bluish, 3500 m. alt., March 11, 1917, Venturi 6697 (US); Cerro de Cachi, fl. violet, 3500 m. alt., March 10, 1917, Venturi 6696 (US); Pampa Grande, Jan. 1897, — (Speg) ; Amblaio to Cachi, Jan. 1897, cages Paget Jujuy: Tilca en los pajonales del cerro, fl. violet, 3000 m. alt., Feb. 18, 1927, Venturi 7419 (US): Maimora, dept. Tilcara, fl. neg EArt in gravel, | m. alt., Feb. 15, 1927, Venturi 10411 (US); Volcan, mbaya, grassy place, fl. bluish, 2300 m. alt., Jan. 20, 1926, Venturi 9180 (US). in ne mountainous regions of northern Argentina this plant has been confused with the form of A. distinens found in the same area. It is not closely related to that species, however, differing conspicuously in the 1947 | JOHNSTON, ASTRAGALUS 373 structure of the fruit and usually in the color as well as form of the corolla. The pods, though averaging somewhat smaller, are similar in structure, texture and shape to those of 4. carinatus. It is probably in the evidently different and more southerly ranging 4. carinatus and A. pauranthus that our present species finds its closest relatives. 37. Astragalus pauranthus, sp. nov. Planta humilis ascendens e radice hea profundi perenni erumpens; caulibus pluribus gracillimis 1-2 os yngis strigosis sparse ascendenter ramosis foliosis; rhachibus folii 2-5 cm. longis; foliolis distantibus 5—7- jugatis 3-7 mm. lon ngis 1-3 mm. ae linearibus vel oblongis supra glabris subtus strigosis apice saepe retusis; pedunculis 5-10 cm. longis foliis subduplo longioribus supra medium laxifloris; floribus 3-8 distantibus pallidis; vexillo 5-7 mm. longo; alis carina conspicue longioribus; calyci- bus 2 mm. longis campanulatis nigro-strigosis, lobis gracilibus tubo paulo Nevioniige ovario glabro; leguminibus refractis glabris 12-15 mm. longis 3-4 mm. altis basim versus rotundis apice obtusis supra carinatis saepe rectis vel (a latere viso) plus minusve concavis subtus late inflexis 2—3 mm, crassis Rance: Patagonia. ARGENTINA. Rio Negro: General Roca, on dunes at foot of the barrancas on the north side of the ef of the Rio Negro, Oct. — Nov. 1914, Fischer 99 (type, Gray Herb.; US, NY, FM, K, BD). Chubut: Caen leofu, 1900, Zllin (Speg) ; Chubut, 1899, Illin (Speg) ; can 1900, Basaldua (Speg) The collections made by [lin and Basaldua were hase cited as A. Bergii by Spegazzini, Anal. Mus. Nac. Buenos Aires 7: 265 (1902). That species, which is known only to the east and north of the range of A. pauranthus, is a very different plant. Our species may be separated from it by the few differently shaped flowers, the more slender more lowly habit of growth, and the chartaceous on with broadly and strongly inflexed lower suture. The affinities of A. pauranthus are with A. Bergiit, A. saan and possibly A. Ruiz- nies . Astragalus Ruiz-Lealii sp. nov. gee gracillima annua strigosa cinerea; caulibus basi sparse ramosis erectis 5-15 cm. altis ca. 1 mm. crassis; foliis caulinis, rhachibus folii 2—5 cm. longis sursum 4—12 mm. supra basim efoliolatis: foliolis distantibus $—7-jugatis 2-9 mm. longis 1-2 mm. latis linearibus vel oblongis, apice m plus minusve reflexis pallide strigosis 2—2.5 c cm. longis utroque attenuatis gracilibus 6-10-plo longioribus quam latis falcatis, margine superiore (a latere viso) convexis secus sutura prominente p aulo compressis, ee inferiore (a latere viso) concavis valde inflexis; eee numeros 374 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XXVIII Rance: Known only from Mendoza. ARGENTINA. Mendoza: Mendoza, Jan. 1910, sears zint (Speg); dry rocky places near Mina La Atala, dept. Las Heras, Nov. 1, 1936, Ruiz Leal 4200 (G); bed (type, Gray Herb.) ; near Oueb. de la Mina La Atala, Sept. 20, 1937, Ruiz Leal 4676 (G); Km. no. 59, Paso de la Cumbre, dept. Lavalle, Sept. 20, 1938, Ruiz Leal 5297 (G). A very distinct species characterized by its very slender stems, its annual root, few flowers, and elongate extrorsely falcate pods. In having ex- trorsely falcate pods which become deflexed, the species can be compared only with the very different A. tehuelches of southern Patagonia. The seminiferous (superior) suture of the fruit is strongly convex in lateral outline and not concave nor straight as is prevalent in this genus. e valves are laterally compressed just below the superior suture and then somewhat swollen. The lower suture is abruptly and distinctly inflexed. There is a very narrow hyaline incomplete false septum inside the pod. The species is probably most closely related to A. tehuelches and A. pauranthus, but has smaller and fewer flowers, more slender stems, and a distinctly annual root. The collection by Spegazzini, cited above, was mixed with material of A. carinatus and is part of collection no, 2146 in his herbarium at La Plata. To BE CONTINUED ARNOLD ARBORETUM, Harvard UNIVERSITY. JOURNAL OF THE ARNOLD ARBORETUM Vor. XXVIII OCTOBER 1947 NuMBER 4 ASTRAGALUS IN ARGENTINA, BOLIVIA AND CHILE IvAN M. JOHNSTON Continued from p. 374 39. Astragalus Monteroi, sp. nov. Herba prostrata; caulibus 2—4 dm. longis 1—2 mm. crassis sparse incon- spicueque strigosis sparse longeque ramosis foliosis, internodiis 3-5 cm longis; foliis superioribus paulo reductis; rhachibus foliorum 3-5 cm. longis sparse strigosis partibus basalibus efoliatis 0-15 mm. longis; stipulis chartaceis ocreatis 2.5—-5 mm. longis, lobis 1-2.5 mm. longis; foliolis 6—9 jugatis carnosulis obcordatis vel oblongo-obcordatis supra medium vel edium versus latioribus apice retusis basi acutis supra glabris subtus glabratis vel praesertim in margine et costa strigosis costatis sed enervatis ngis 4-7 mm. latis 0.5—-1 mm. longe petiolulatis ; pedunculis 3-6 cm. longis cum floribus 8—20 congestis terminatis; calycibus 4-5 mm longis sparsissime strigosis tubo 2—2.5 mm. longo lobi is 1-2 mm. longis angustis pedicello 1-2 mm. longo; vexillo 9— 10 mm, longo 8 mm. lato caerulea 2—2.5 mm. brevioribus; ovario sees leguminibus Nabe pendulis pene 11-14 mm. ‘longis oblongis acuminatis supra acutis subtus planis 3-4 mm. crassis, a latere viso supra leviter concavis et subtus valde concavis. CHILE. Cautin: Puerto Saavedra, 3 m. , Feb. 932, G. Montero 1161 (type, Gray Herb.) ; Llaima, pe 1925, Claude pee an ae Budi, Feb. 1928, Claude ae 4878 (US). In gross habit and in corolla structure the present species much suggests the coarse forms of A. palenae (i.e. the var. grandiflora Speg.). The fruit of A. Monteroi, however, is very different from that of A. palenae and very much more like that found in A. amatus of the coast of central Chile. The closest relative of A. Monteroi is evidently A. amatus. From this species it is quickly distinguished by its glabrous ovary and fruit. Itisa coarser and much larger plant with much larger more fleshy leaves. The species seems to be very well marked. 376 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII 40. Astragalus amatus — in Gay, Fl. Chile 2: 115 (1846).— Type ia as from rfiro, on road from La Serena to ee 1433 m. alt., Gay Astragalus poresnenr oan "Hooker & Arnott, Bot. Beechey Voy. 18 a: and Bot. Misc. 3: 186 (1832); Hooker, Bot. ag. 60: t. 3263 (1835); Reiche, Anal. Univ. Chile 97: 540 (1897), and FI. Chile 2:83 (1898).— Not Miller (1768). — Given as collected by Lay and Collie at “Conception,” Chile. Phaca tricolor Clos in Gay, Fl. Chile 2: 105 (1846).— Based upon collections (devoid . geographical date) said to have come from the coast of Coquimbo and Aconcagua. Astragalus mee (Clos) Reiche, pe Univ. Chile 97: 550 (1897), and FI. Chile 2:93 (1898). Not Bunge (1868-69). ai es ferrugineus Clos in Gay: FI. Chile 2: 114 (1846); Weddell, Chlor. na 2: 262 (1861); Reiche, Anal. Univ. Chile 97: 567 (1897), and FI. Chile 2: ere (1898).— Based upon material (mixed and probably mislabeled) given Cauquenes, Gay 237. Astragalus leucomallus Philippi, Linnaea 30: 186 (1859).— Based upon material from coast near Topocalma, prov. Colchagua, Gay 527 and Volckmann. ei brachyptera Philippi, Linnaea 33:44 (1864).— Type collected near Catemu, . Aconcagua, Philip pi. prey brachypterus (Phil.) Reiche, Anal. Univ. Chile 97: 541 (1897), and FI. Chile 2:84 (1898).— Not Fischer (1853). Astragalus litoreus Philippi, Linnaea 33:46 (1864); Reiche, Anal. Univ. Chile 97: 540 (1897), and FI. Chile 2: 83 (1898).— Type collected on the coast near Coronel, 1861, F. Philippi. Tragacantha brevialata Kuntze, Rev. Gen. 2: 940 (1891).—A new name based achyptera Phil. (1864); A. brachyptera Reiche (1897), not Fischer (1853). Astragalus chilensis Sheldon, Minnesota Bot. Studies 1: 157 (1894).— A new name based upon A. procumbens H. & A. (1830), not Miller (1768). Astragalus subandinus Spegazzini, Anal. Mus. Nac. Buenos Aires ser. : 264 (1902).—A new name based upon P. brachyptera Phil.; A. brachypterus Reiche (1897), not Fischer (1853). RANGE: West-central Chile, from the province of Concepcion north to the province of Coquimbo, along the ocean and east into the coastal mountain . Coquimbo: hills of decomposed basanite near Porfiro, road between La Serena and Arqueros, 1433 m. alt., Gay 269 (Paris, type of A. amatus) ; Coquimbo, Cuming 881 (K); Andocollo, Oct. 1926, Claude Joseph 4533 (US); Est. Fray Jorge, 215 m. alt., Skottsberg 786 (Stock, Gottenburg); Fray Jorge, 1935, Munoz 137 an 165 (G); Punta del Viento, Fray Jorge, 400 m. alt., 1925, Werdermann 916 (G, BD); Cuesta de Cavilolén, ca. 30 km. s. w. of Illapel, snes banks, standard blue-purple, wings white, 500 m, alt., Worth & Morrison 16465 (G). Aconcagua: Zapallar, Sept. 17, 1917, Johow (G); Cerro de la Cruz, Zapallar, Behn (Behn); Punta de Quintero, cliffs, Poeppig 189/307 (BM, BD); Quintero, Werdermann 1 (G, BD); mountains [above Catemu] Philippi (BD, isotype he P. brachyptera) ; Concon, Poeppig 190/257 and 188/211 (BD, vay Playa de Concon, sand, 1937, Jaffuel & Pirion 3946 (G); Vifia del Mar, dunes, 1830, Asn 1824 (G, NY, BD, Paris) ; Valle de Marga Marga, 40 km. inland, 250 m. alt., 1929, Jaffuel 247 and 3021 (G); Las Zorras, 1917, pos le 1038 (Stock, Gottenburg) ; Playa Ancha, 1829, Bertero 823 (Paris); hillside, Queb. Jaime, Valpo., Nov. 1895, Brenning 129 (BD); El Alto, hillside, 1830, steele 1821 (G, Paris); Laguna Verde. 1930, Garaventa 2122 and 2129 (G). Santiago: San ‘Antonio, dunes, Sept. 1829, Gay 469 (Paris). Colchagua: ae March 1894, Philippi (G, LP); near stream, Cauquenes, Feb. 1831, Gay 237 (Paris, type of A. ferrugineus). Talca: Llico, Dec. 1861, ex Philippi (LP); Moca, Feb. 1931, Fuentes (G). Concepcion: Coronel, sea- shore, Feb. 1861, F. Philippi (G, isotype of A. litoreus); Coronel, ex Philippi as A, litoreus (BD); Talcahuano, Dec. 1861, herb. Reed (K); Quiriquina, Nov. 1924, 1947 | JOHNSTON, ASTRAGALUS 377 Claude Joseph 2855 (US); San Miguel, Nov. 19353, Barros 137 (G); Concepcion, Cuming 137 (K); Concepcion, Sept. 1895, Neger (BD); prov. Concepcion, ex Philippi as A. litoreus (Boiss, Stock); San Vicente, Dec. 1894, Philippi (G). In- definite: “Chile,” Gay sine no. and Gay 584 (Paris, type of P. tricolor) ; “Chile,” Gay as P. tricolor (G, K, BD). Assembled here under the name 4. amatus is a very variable group of plants of western Chile, found from the provinces of Concepcion to Coquimbo, and characterized by having the wings of the corolla con- spicuously shorter than the keel. The group of forms is very readily recognizable and has a close relative only in A. Monteroi of the coast of southern Chile. Possibly it may be a complex of local species. The material available to me, however, is inadequate for an attempted analysis of it at this time. Certain of the phases of A. amatus deserve special comment. There is an erect plant of the coastal mountains, somewhat back from the ocean in the province of Aconcagua, which has small ovate or oblong leaflets which are markedly and gradually diminished in size upwardly along each leaf- rhachis. Its corollas appear to be a darker blue than in the forms from other regions. This plant is typical A. amatus and is that described as P. brachyptera by Philippi. I have seen collections of it from Porfiro, Andacollo and Catemu. The other forms of A. amatus, sensu lat., are decumbent or prostrate and appear to grow near the ocean or in the coastal hills under more marked maritime influence. The most striking of these spreading forms is a white-tomentose plant with coarse stems and thickish broad ovate leaflets which appears to be restricted to sands and dunes along the ocean. This extreme form has been described as A. leucomallus Phil. Though very conspicuous in its extreme development the tomentose plant seems to intergrade with the common, excessively variable and more mesophytic forms of the fields and slopes back from the sea. These latter vary in coarseness of stems and in size and shape of the leaves. Their leaflets may be subcordate to linear or narrowly elliptic, appressed villous to nearly glabrous, 3-15 mm. long, 2-6 mm. road, and their apices from obtuse to retuse. They present, accordingly, considerable variation in gross aspect. To this complex of forms belong the names, 4. procumbens & A., A. litoreus Phil. and A. tricolor Clos, the first to the much col- lected form north and south of Valparaiso, the two latter names to the small-leaved glabrescent slender-stemmed laxly decumbent phase which prevails in the region about Concepcion. The type of A. procumbens H. & A., which I could not find at Kew, is said to have been collected at Concepcion by Lay and Collie. It is said to have elliptic retuse leaflets which are ‘“tomentoso-hirsutis” or “‘sericeo- villosus.” I have seen no such plant from near Concepcion, where the glabrescent forms such as A. tricolor and A. litoreus seem to predominate or, at least, have been most collected. The original description of 4. procumbens fits best the plants found on the slopes about Valparaiso. 378 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII The type of A. amatus Clos is a mixture of two very different forms. The label on the type collection (Gay 269) gives the source of the plants as Porfiro, 1000 m. alt., in the mountains northwest of Coquimbo. Of the two forms mixed under Gay 269, one is indistinguishable from the spreading densely white-tomentose plant of the coastal sands which Philippi later called A. leucomallus. The other plant, erect, cinereous and appressed villous (rather than tomentose) and bearing more numerous (9-15) smaller strongly truncate or retuse leaflets, is most like the type of A. brachypterus (Phil.) Reiche. This latter is the form which has the numerous leaflets noticeably reduced in size up each leaf-rhachis and the one known only away from the ocean in the coastal hills of Coquimbo and Aconcagua. In his discussion of 4. amatus Clos distinguishes these two forms, mentioning first the erect plant of the interior and second the spreading white-tomentose plant of the coastal sands. In his description of the species, however, the tomentose dune-form seems to be somewhat emphasized. In this connection it may be noted that, at Paris, Clos has identified Bertero 1821 as “A. amatus.” This particular collection by Bertero is the tomentose dune plant. Since, however, Clos identified other very similar collections of the white-tomentose dune-form as “A. procum- bens,” I cannot attach any importance to his identification of the collec- tion of Bertero mentioned. Not only are the collections of ‘A. amatus” mixed in the herbarium at Paris, but so are the duplicate collections of the species distributed to other herbaria. For example, at Berlin there are two sheets of “A. amatus,’’ one is entirely the tomentose dune plant and the other is a mixture of A. mudus and A. vesiculosus. The sheet of “A. amatus” at New York is the tomentose dune plant. At the Gray Herbarium the sheet of “A. amatus” is a mixture of the erect plant of the interior (Le., A, brachypterus) and A. nudus. Similar mixtures are found in other herbaria. It may be noted here that the material of the dune-form forming a part of the mixed type of A. ferruginosus, in appearance, dis- coloration, maturity, manner of preparation, etc., is so very much like that mixed in the type collection of A. amatus, that I am confident that they were collected and prepared at the same time and are parts of a single dissociated collection. This would mean that the material of this particular collection of the coastal dune-form was not only divorced from its proper label but was also separated and improperly associated with collections of other forms of A. amatus from different localities. There is plenty of evidence that some of Gay’s collections did get mixed. In the present case of “A. amatus” it seems clear that the label on the type at Paris could apply only to the plant of the interior (ie., A. brachyp- terus) collected at Porfiro, and that the material of the white-tomentose dune form must be extraneous. Hence, because in his publication Clos discussed the erect plant of the interior before mentioning the spreading plant of the dunes, and since there is every reason for believing that the plant of the interior, rather than the coastal dune form, will be found at 1947 | JOHNSTON, ASTRAGALUS 379 Porfiro, the locality given on the associated label, I believe that the erect plant of the interior should be accepted as the typical element in the complex type of A. amatus. The type collection of A. ferrugineus, at Paris, is a mislabeled mixture. The type-sheet contains a plant of the tomentose dune form (i.e., 4. leucomallus) and one of the slender glabrate forms similar to those found about Concepcion (i.e., A. tricolor and A. litoreus). These are obviously parts of two different collections made at different places. The label is, with the exception of the plant-name added by Clos, entirely written by Gay. It reads, “237, Provincia Colchagua, Astragalus ferrugineus Clos ad rivulos, Cauquenes, februario 1831, fl. septembri, etc., frequens, t. basaltique, (Chili).” During February 1831 Gay was exploring along the Rio Cachopoal and visited Banos de Cauquenes and the Estancia de Cauquenes in the province of Colchagua. This area, at the base of the cordilleras, however, is far beyond the known range of the coastal A. amatus and its forms. They are not known, nor to be expected, east of the coastal mountains, within 100 km. of Estancia de Cauquenes. The material on the type-sheet, hence, represents parts of two different collec- tions (probably from the dunes near Valparaiso and from the coast near Concepcion), which are improperly associated with the label belonging to an unknown third plant. Though Clos, in his description of A. ferru- gineus, seems to be concerned only with the tomentose dune form, the type collection is so badly mixed that I believe the species had best be discarded as a nomen confusum. The type material of P. tricolor, at Paris, consists of two sheets devoid of any precise geographical data. This type-material and the original description agree well with plants collected in the coastal area of the province of Concepcion. When 4. tricolor was described, however, it was given as growing near the coast in the provinces of Coquimbo and Aconcagua. Among the many collections which I have seen from the coast of Aconcagua and Coquimbo none are similar to the original material of A. tricolor. I am confident that the type-material of A. tricolor really came from the region about Concepcion. 41. Astragalus valparadisiensis Spegazzini, Anal. Mus. Nac. Buenos Aires ser. 2, 264 (190 renaming of A. trifoliatus Phil. Astragalus trifoliatus Philippi, Linnaea 28: 681 (1857); Reiche, Anal. Univ. Chile 97: 543 (1897), and FI. Chile 2: 86 (1898). Not Boiss. (1843).— Type orig- inally said to be from the Andes of the province of Santiago. Reiche, how- ever, gives it as from “Algarrobo,” on the coast of Aconcagua. CHILE. Aconcagua: Valparaiso, ex Philippi dedit. 1876, as A. trifoliatus (BD, Del); Valparaiso, no collector given, as A. trifoliatus (Speg.). I know this plant only from descriptions and from the collections, all evidently duplicates of one gathering, above cited. It is a very distinctive plant with its trifoliate leaves, coarse loose stipular sheathes and very dense abundant long silky golden indument. It is to be noted that in the literature Philippi and Reiche differ as to the origin of this species, the 380 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII former giving it as from the cordilleras, the latter as from the coast near Valparaiso. The specimens distributed by Philippi, however, are labeled as from Valparaiso, and this is probably correct. The general habit of the plant suggests an inhabitant of dunes and a close examination of the specimens reveals the presence of clean white sand in the stipular sheathes and among the hairs of the indument. No fruit of the species is known. The original collection shows flowers at anthesis. These much suggest those of A. nudus and I suspect that is the closest relative of A. valparadisiensis. 42. ea oriay cachinalensis Philippi, Fl. Atac. 15 (1860); Reiche, Anal. Univ. Dra 97: 547 (1897), and FI. Chile 2: 90 (1898); Johnston, Contr. Gray Herb. (1929).— Type collected at Cachinal de la Costa, Philippi. Rance: Fog-bathed hills and cliffs along the coast of prov. Antofagasta, northern Chile. CHILE. Antofagasta: Cachinal de la Costa, Philippi as A. cachinalensis (BD, isotype); near Aguada Grande (‘“Cachinal de la Costa” of Philippi), 1925, Johnston 5825 (G, US, K); near Aguada Cachina, 1925 Johnston 5740 (G); El Rincon near Paposo, 1925, Johnston 5497 (G, US); Aguada Panul, 1925, Johnston 5431 (G); near Aguada Miguel Diaz, 1925, Johusion 5349 (G, US) A plant with slender elongate trailing leafy stems and very large flowers. The bluish standard becomes 12-18 mm. long. The keel is cream-colored and tipped with bluish. The petals become yellowish when dry. The species is a very distinct one and is restricted to the fog-bathed slopes along the southern half of the coast of the province of Antofagasta. It is obviously an isolated outlying relative of A. amatus of central Chile. 43. Astragalus nudus Clos in Gay, Fl. Chile — 115 (1846); Reiche, Anal. Univ. Chile 97: 546 (1897), and FI. Chile 2: 89 (1898).— Based upon two collections, dunes near La Serena, Gay 15, and between La Serena and Arqueros, 1300 m alt., Gay 27 RANGE: ican only from the region about Coquimbo and along the coast in southern Atacama. CHILE. Coquimbo: Rare, hills of decomposed basanite (near Porfiro), road between La Serena and Arqueros, 1300 m. alt., Gay 270 (Paris, type); Herradura, 1917, Skottsberg 1057 (Stock); scattered on dunes near ocean, La Serena, Sept. 1836, Gay 15 (Paris); Isla Guacolda, Huasco, on sand of island, perennial, standard laven- der and white, wings white, Oct. 26, 1938, Worth & Morrison 16224 (G); Chile, Gay as A. nudus and A. procumbens (K); Chile, Gay as A, amatus, in pt. (BD) ; Chile, Gay as A. procumbens (G, NY). The corolla of A. nudus dries yellowish except for the distal half of the keel which is purplish or blue. When fresh the standard is probably pale bluish and the wings cream-colored. The whole plant is densely strigose and canescent. The stems appear to be erect or ascending. The rela- tionships of the species are with A. valparadisiensis and, particularly, with Dodti. I have not seen the completely mature fruit of this species. 44. Astragalus Dodti Philippi, Linnaea 33:48 (1864), Anal. Univ. Chile 27: 343 865), and Anal. Univ. Chile 84: 28 (1893); Reiche, Anal. Univ. Chile 97: 545 (1897), and Fl. Chile 2:88 (1898).— This species was based upon flowering plants from near Copiapo, Dodt, and upon fruiting plants from Arqueros, Gay. 1947] JOHNSTON, ASTRAGALUS 381 Astragalus Rengifoi Philippi, Anal. Univ. Chile 84: 25 (1893); Reiche, Anal. Univ. Chile 97: 544 (1897), and Fl. Chile 2:87 (1898).— Based upon material col- lected near Chanarcillo, Sept. 1876. Astragalus melanogonatus Johnston, Contr. Gray Herb. 85:52 (1929).— Type from Reyes, prov. Antofagasta, Johnston 5353 Rance: Gravelly and sandy soils on the coastal desert “of Atacama and Antofa- CHILE. Antofagasta: Antofagasta, 1925, Johnston 3632 (G, US); Punta Reyes near Miguel Diaz, 1925, Johnston 5353 (G, type A. melanogonatus). Atacam of Caldera, nar washes above beach, tay rake white, Oct. 21, 1938, Worth & Morrison 16183 (G); Caldera, Gigoux ; Caldera, 1885, Borchers (BM); Monte Amargo, 200 m. alt., 1924, Werdermann - (G, K, BD); near Copiapo, Sept. 1854, Lechler 2805 (K); Chanarcillo, Sept. 1876, ex Philippi (G, LP, isotypes of A. Rengifoi). Indefinite: Desert of Atacama, Geisse, sub Morong 98 (NY) and Morong 1165 (G, NY, US, K); Atacama Desert, ex Philippi as A. Rengifoi (BD, US); ? Vicuna, Oct. 1926, Chiude Joseph 4597 (US) os jo) = tad jon Oo Q i) Ko) 3° wm fo} oy =, = = re) wn Le — 2 This species is prevailingly annual and erect. It is very closely related to A. nudus and possibly may not be separable from that more southern species. The legumes are strongly compressed and have nearly parallel sides. They may be permanently strigose or become glabrescent. he corollas of A. Dodti and A. nudus are very similar in size and form, but the coloration in the latter species is more intense. Astragalus Dodti and A, Rengifoi are evidently conspecific. The more northern form described as A. melanogonatus is essentially similar differing chiefly in the broader, oblong rather than linear, leaflets. 45, ea a Berteri Colla, Mem. R. Accad. Sci. Torino 37:55, t. 9 (1832), and P . Bertero. 19, t. 9 (18312); Reiche, Anal. Univ. Chile 97: 546 (1897), and a “Ch ile 2:89 (1898).— Based upon material ees as from “locis aridis de la Leona secus rivum Cachapual,” collected by Bertero. Astragalus prostratus Hooker & Arnott, Bot. Voy. Beye hy 18 (1830), and Bot isc seen ). Not Scop. (1786-88).— Based upon collections from C Collie ie alberta et Nom. os . 1: 159 (1840), nomen. — Given as based pon Bertero 65 from prov. Colch “ grata Clos in Gay, FI. Chile 2: a (1846).— Type collected in “Chile” by Astragalus gratus (Clos) Reiche, Anal. Univ. Chile 87: 570 (1897), and FI. Chile 113 (1898). Astragalus filifolius Clos in Gay, FI. Chile 2: oa (1846). — Type collected between a Serena and Arqueros, Oct. 1833, Ga Pia elongata Philippi, Linnaea 33: 43 Beis —“In Andibus de Chillan dictus Astras dencsies (Phil.) Reiche, Anal. Univ. Chile 97: 568 (1897), and FI. Chile 111 (1898). Not Willd. (1803). Phaca concinna auete Linnaea 33: 44 (1864).—‘“In montibus elatus praedii atemu prov. Aconcagua inveni.” Astragalus concinnus (Phil.) Reiche, Anal. Univ. Chile 97: 568 (1897), and FI. Chile 2: 111 (1897). Astragalus cauquenensis Philippi, Anal. Univ. Chile 41: 691 (1872) .— Type from Banos de Cauquenes, proy. Colchagua. Astragalus pencanus Philippi, Anal. Univ. Chile 84:30 (1893).— Based upon collections from Cerro de Caracol near Concepcion, Jan. 1887, Philippi, and from Hac. de Almendoro, Araucania. 382 JOURNAL OF THE ARNOLD ARBORETUM [voL. XXvI Tragacantha longior Kuntze, Rev. Gen. 2: 941 (1891).— Based upon P, elongata Phil. Astragalus elongatus (Phil.) Reiche, not Willd. (1803). Astragalus ae he gle Sess gpesin Anal. Mus. Nac. ser. 2, 4: 265 (1902).— Based upon P. elongata Phil., not Astragalus elongatus Willd. (1803). RANGE: Central cai along the coast anal east to the base of the cordilleras, from Coquimbo south to Concepci CHILE. Atacama: Caldera, Oct. 1894, Gigoux 10 (G). Coquimbo Rivadavia, 900 m., Sept. 17, 1936, Montero 2849 (G); rare on dry hills aoa jaad between La Sorens and Arqueros, Oct. 1833, Gay 141 (Paris, type of A. filifolius) ; prov. are Gay as A. filifolius (G); Loma Fray Jorge, 1917, Skottsberg 1467 (Stock). Aconcagua: Las Palmas de Pedegua, 1300 m. alt., Sept. 1933, Grandjot (Grandjot) ; indefinite [Hac. Catemu], ex Philippi as P. concinna (BD, Stock); Valparaiso, Cuming 736 (K). Santiago: Colina, 1825, Macrae (K); Banos de Colina, Bridges (K) ; Curacavi, Oct. 1898, Castillo 14 (G). Colchagua: near Rio Cachapoal, Sept. 1928, Bertero 65 (NY, Del); along Rio Cachapoal near San Joa- quin, Oct. 1829, Bertero 65 (Paris, type of A. wget nore indefinite, Bertero as A. Berteri (BD); Rancagua, ex Philippi, no. 212, as A. amatus (BD); Banos de Cauquenes, 1867, Philippi as A. cauquenensis (BD); i ls Oct. 1922, Claude Joseph 1824 (US). Talca: Sie Oct. 1921, Claude Joseph 1670 (US). “Nuble: > co N p od ao s S. oO wn Qu 2s onl _ po co pP _ Os — a o — ‘Oo w =) = = a) = ° ~ ‘© Nh an Concepcion: Tome, Nov. 1925, Claude Joseph 4035 (US); Concepcion, Beechey Voy. (K, type of A. prostratus); Concepcion, Reed (K); San Miguel, Nov. 1933, Barros 124 (G) ; open pasture land, West 5093 (Univ. Calif.). Indefinite: “Chile,” Gay (Paris, ype of A. gratus); Loncoche, Sept. 1926, Claude Joseph 4320 (U As here treated A. Berteri is almost certainly an aggregate of several species or good geographical varieties. The specimens available to me have been not only too few in number but also too deficient in fruiting structures for me to propose a segregation of this complex. The type of A. Berteri was collected by Bertero on Cerro de la Leon near the Rio Cachapoal in 1830, and is no doubt similar to other collec- tions at this locality made in 1828 and 1829 and distributed under his number 65. This number, however, is an aggregate consisting of A. orthocarpus and two forms of A. Berteri, one of the latter with strigos the other with glabrous ovary. Colla describes the fruit as ‘‘glabriuscula,” rather than glaberrima, and we may assume he had the fruiting plant from Bertero, found most widely in herbaria under Bertero 65, in which, though the fruit appears to be glabrous, it is actually sparsely strigose when viewed under a lens. Colla described the fruit as triquetrous, but his illustration and Bertero’s specimens show the lower suture to be. distinctly inflexed and the lateral faces of the pod plane and almost parallel. The herbage is silky strigose. This typical form of the species, ranging along the base of the cordilleras, appears to have synonyms in P, grata, A. alberjilla, P. concinna, P. cauquenensis and A. elongatus. rowing in the coastal area, from Concepcion to Coquimbo and Atacama, is another form of this aggregate. This is usually ylabrescent. It has well developed very slender stems bearing elongate peduncles frequently twice as long as the subtending leaf. The leaflets are usually 1947] JOHNSTON, ASTRAGALUS 383 very slender. In southern plants they may become emarginate at the apex. The following names apply to this coastal form: A. prostratus, A. filifolius, and A. pencanus. These two principal forms, that along the coast and that in the drier interior, vary much in the form, color, proportions and size of the corolla- parts. I suspect that a thorough study of the corollas will aid in delimiting segregates from this complex. I have made a_ prolonged attempt to find characters which would sharply separate the coastal and interior plants and so reduce some of the obvious complexity of the species. My inability to achieve this I am now convinced is due to the fact that not two but probably four or five species or varieties are involved. However, the following attempt to contrast the coastal and interior forms may be of interest: Plant erect or the short stems decumbent only at the base, cinereous to canescent with abundant pee hairs; pods usually deltoid in cross-section, broadly inflexed: -below;- Montaine...:6. 60. s6sbs beeen see ss sees . Berteri typicus Plant loosely decumbent with slender elongate spreading stems, green or cinereous with sparse appressed hairs or subglabrous; pods laterally much Cee seG, narrowly sagittate in cross-section, deeply and narrowly intruded below OAS CA ep cee ens erensete eeteee crs & m5: ie 6 er 0 al arse. ge 9 ar ons gee ent eo “A, filifolius” — 46. ee re Germaini Philippi, Linnaea 28: 621 (1857); Reiche, Anal. Univ. Chile 97: 547 (1897), and Fl. Chile 2:90 (1898).—Type from Las Aranas, Cord. a iago, Philippi. gsr ee Segethi Philippi, Linnaea 33: 47 (1864).— Type from Las Aranas, Cord. Santi ip pi. Pees ects Philippi, Anal. Univ. Chile 41: 690 (1872).— Type from Andes de Talcarehue, prov. Colchagua. Astragalus Closianus Philippi, Anal. Univ. Chile 41: 691 (1872).— Type from near Bafos de Cauquenes, prov. Colchagua, Oct. 1867, Philippi. Astragalus Dessaueri Philippi, Anal. Univ. Chile 84:26 (1893). aie from de las Viscachas, Est. Cauquene, prov. Colchagua, de Dess Astragalus alfalfalis Philippi, Anal. Univ. Chile 84:29 (1893). — Type —_ Valle del Rio Colorado, prov. Santiago, 1700 m. alt., Jan. 1888, Phili Astragalus azureus Philippi ex Reiche, Anal. Univ. Chile ah 548 (eon), and Fi. Astragalus Germaini var. azureus Philippi ex Fries, Nova en R. Soc. Sci. Upsala ser. IV, 14: 135 (1905). Rance: Cordilleras of Santiago, Colchagua and Curico. CHILE. Santiago: Cord. de Santiago, ex Philippi as A. Germaini (BD); mountains east of Santiago, 900 m. alt., Nov. 23, 1900, Hastings 191 (US); Quelte- hues, 1700 m. alt., 1927, Montero 324 (K); Queb. del Manzano, Valle del Maipo, Nov. 4, 1930, Pirion 1242 and 1243 (G); El Manzano, Nov. 1937, Espinosa 17 (G); Divisadero de Lagunillos near San Jose de Maipo, 1936, Looser 3612 (G). Col- chagua: Valle del Clarillo, 1200 m. alt., Nov. 1933, Grandjot (Grandjot). A species evidently related to A. Berteri but readily distinguished by its firm glabrous triquetrous fruit, its usually obtuse or retuse leaflets, and its montane distribution. In having a triquetrous fruit with a broad flat lower face A. Germaini is similar to A. carinatus of the eastern slopes of the Andes. 384 JOURNAL OF THE ARNOLD ARBORETUM [vVOL. XXVIII 47. Astragalus baslctcaey Cavanilles, Icon. et Descr. Pl. 1:59, t. 85 (1791).— ollections from the puna near Tarma and from the prov. Huaro- chir material cultivated at Madrid. Astral amelie L’Heritier, Stirp. Nov. 6: 168 (1791-92).— Type from Peru, mbe Predict Benthawianes Gillies ex Hooker & Arnott, ie Misc. 3: 187 (1832).— Type from El Alto de las Manantiales, Mendoza, Gillie Arq \ Astragalus Mandoni Rusby, Mem. Torrey Bot. Club 3:19 (1893). — Based’ upon Solivian collections, Sorata, Mandon 709 and Mt. Tunari, Bang 1022. Astragalus Garbancillo var. Mandoni (Rusby) Macbride, Field Mus. Pub. Bot. 9 , This is the most widely distributed and the most commonly collected species of South American Astragalus. It ranges at middle altitudes from central Peru southward to the Tucuman and Mendoza areas in Argentina. In central Peru it has been collected in the departments of Libertad, Ancash, Junin, Huanuco, Huancavalica and Lima at 2000 to 3500 m. altitude. In southern Peru it is known from the departments of Puna and Cuzco. In Bolivia it is best known and has been much collected about La Paz and Lake Titicaca. From there it ranges southward through eastern Oruro and southern Cochabamba and on through central and eastern Potosi, Camargo and Tarija into Argentina. Much of its range in Bolivia is hence in the Yungas. ‘The plant evidently affects the more mesophytic and more dissected country east of the puna. Within Argen- tina it extends through Los Andes, Jujuy and Salta reaching to Tucuman (2000-3400 m. alt.), and further southward through Catamarca, La Rioja (Sierra Famatina) and San Juan (Leoncito, Echegary; Cord. Colamgiiil, 3150 m. alt., Perez Moreau 30/109) to its southern limit in northern Mendoza (Portez. de los Chilcas, 2000 m. alt., Semper 4261, Alto de los Manantiales, Gillies). With the exception of one very questionable specimen, there is no record of the species in Chile. It reaches the Pacific Slope only in central Peru. Within our area A. Garbancillo is very distinct and has no close relative. The difficulties in delimiting the species are confined to central Peru where it appears to have given rise to the evidently related A. Pickeringii Gray, A, alienus Gray and possibly also A. romasanus Ulbrich, cf. Johnston, Jour. Arnold Arb. 19: 92-95 (1938). Throughout its wide range 4. Garbancillo is rather variable. Some of its variations seem to be vaguely correlated with geography, but none of them definitely enough to warrant any nomenclatorial recognition at the present time. he type collection of A. unifultus is a plant with the stems pallid and strigose. It much resembles the common form found in the department of Junin. I suspect that it probably came from that section of central Peru. Judging from the original description and _ illustration of 4. Garbancillo, it must be a very similar plant from the same region. Plants very similar to the type of A. unifultus are the common and usual form of the species in western Bolivia and Argentina. 1947] JOHNSTON, ASTRAGALUS 385 The name A. Garbancillo Cav. appears to have six months or even a year priority over A. unifultus L’Herit. The former was published in volume one of Cavanilles’ Icones, which is dated in the preface Jan. 1791, and was apparently issued the same year. This date has never been questioned. L’Heritier’s species, however, appeared in the sixth fascicle of his Stirpes Novae, which, though bearing the title-page date 1785, was actually issued much later. Britten & Woodward, Jour. Bot. 43: 267-68 (1902), accept the date of the sixth fascicle as being doubtfully December 1891. A contemporary reviewer, Usteri, Neue Annalen der Botanik, pt. 2, p. 11 (1794), however, states that this fascicle “quanquam titulus annum editionis 1785 prae se ferat, non tamen nisi medio anno 1792 evulgatus est.” Hence there seems good reason for accepting Cavanilles’ binomial as the prior name. The commonest form of A. Garbancillo in northern Bolivia and southern Peru is a rather slender more or less fruticulose plant with the stems and herbage green and rather sparsely strigose. One of the extremes of this form has been described as A. Mandoni. I have seen similar forms from various parts of the range of the species. When it was published, A. minor was given as having been collected by Gay at Arqueros, a mining region at the base of the cordilleras northeast of Coquimbo, Chile. The type at Paris, however, is associated merely with a printed label indicating that the plant was collected by Gay in Chile. The single specimen is in flower. The stems are slender, erect and thinly strigose. The leaves are also sparsely strigose. The whole aspect of the plant is that of the forms of the species found in southern Peru, a region where Gay is known to have also collected. I believe that the type of A. minor is another example of the many cases of mixed and confused collections to be found among Gay’s Chilean plants. I do not believe that the type was collected in Chile. The type of A. Benthamianus is one of the two collections of A. Gar- bancillo which I have seen from Mendoza, at the southern limit of the species. The specimens are mature ones with ripe fruit and withered corollas. The coarse stems appear to have been white and _strigose when young. Besides the slender green form and the stouter white-strigose typical form of A. Garbancillo there is another variation that deserves some comment. The flowers of the species usually average about 15 mm. in length. In northern Argentina and southern Bolivia there are certain excessively vigorous luxuriant forms which produce flowers 20 mm. long or more. These plants may be green or white-strigose. Characteristic specimens of this vigorous plant have been seen from the mountains of Tucuman (Lorentz & Hieronymus 694; Venturi 4661), from the province of Jujuy (Hofsten 754) and from the department of Tarija (Fiebrig 2469a). This coarse, large-flowered plant may possibly merit a name. The poisonous qualities of this species have long been known. Awscha- lom, Lab. Quimica Analitica, Univ. Tucuman, Pub. no. 1: 1-29 (1928), has studied the species and has identified the saponin. 386 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XXVIII 48. Astragalus arequipensis Vogel, Verhandl. K. Leop.-Carol. ree Naturf. 19: suppl. 17 (1843).— Type from near Arequipa, 3000 m. alt., Mey Astragalus Orbignyanus hina: Chlor. Andina 2: 260 (1861). re. pe the prov. Carangas, d’Orbigny 1437 Astragalus yi ee tao ous. Abh. K. Gess. Wiss. Gottingen 24: 102 (1879). m La Cienega, sarap, 1874, Hieronymus & Lorentz. deans bolivianus Philippi, Cat. Pl. Itin. Tarapaca 15 (1891); Reiche, Anal. Univ. Chile 97:539 (1897), and a Chile 2:82 (1898).— Described from “collections from Quebrada de Calalaste, Minique, and Sitani, 3500-3700 m. alt. Astragalus let tes Rusby, Mem. Torrey Bot. Club 3:19 (1893).— Type from 1890, Bang 174. Ra h qantas of northern Argentina northward over the Bolivian mre re southern Peru PERU. Arequipa: heerlen. Meven (BD, type of A. arequipensis); near Arequipa 2000-2500 m. alt., 1923, Guenther & Buchtien 1107, 1109, 1113 and 1115 (Hamb); between Arequipa and Puna Lagunillas, 4300 m, alt., 1934, Hammar- lund 73 (Stock); between Arequipa and Puna Sumbay, 4060 m. alt., 1933, Hammar- lund 79 (Stock). Puno: Chuquibambillo, 3950 m. alt., 1925, Pevnall 13351 (G, US, FM); Azangaro, June 1854, Lechler 3243 (K); Cisne Pampa, Huancané, 3250 m. alt., 1919, Shepard 83 (G, US); Pucard, 1902, Weberbauer 438A (BD); Puno, Martenet 575 (Paris); Juliaca, May 16, 1902, Williams 2513 (US) BOLIVIA. La Paz: Guarina, 1865, Raimondi 10515 (BD); prov. Omasuyos, 3900-4000 m. alt., Mandon 710 (G, K, BM, Del); near tne Paz, 3000 m. alt., 1890, Bang 174 (NY, type of A. sinocarpus; G, US, K, BM, BD); between Palca and La Paz, 4000 m. alt., 1910, Pflanz 373 (BD); Valle de Pape d meee east of La Paz, 3500-4000 m. alt., 1906, Hauthal 224 (BD); Isla del Sol, Titicaca, 3840 m. alt., Buchtien 2870 (US); ragenseay 1903, Hill 120 (K); near Viacha, 3900 m. alt., 1907, Buchtien 625 (U ere Paz de Azacucho, Corocoro, 1846, Weddell 4338 (Paris). Oruro: Bohs 4100 m. alt., 1908, Buchtien 1812 (US, BD); Prov. Carangas, west of Oruro, LOriiens 1437 (Paris, type of A. Orbignyanus). Potosi: Quechisla, prov. Nor Chichas, 3400 m. alt., 1932, Curdenas 305 (US); Lagunillas, 3800 m. alt., 1933, Cardenas 381 (US); at Oct. 3, 1894, Kuntze (NY, US, Chiguana, 3700 m. alt., 1921, Asplund 6127 (US) and 3077 (Upsala). Chuquisaca: mountains between Pomabaribs and Comargo, Weddell 3981B (Paris). Tarija: Calderillo, 3100 m. alt., Jan. 6, 1904, Fiebrig 2469 in pt. (BD). Arica: oe Cord. Volcan Tacora, ca. 4300 m. alt., April 1926, Werdermann 1119 (G, K, ARGENTINA. Los een Calalaste, Jan. 1885, Philippi (US, K, BD, isotypes of A. bolivianus); Mina Concordia, Feb. 26, 1927, Castellanos 28/824 (G). Jujuy puna near Yavi, May 1873, Lorentz & Hieronymus 702 (BD); Moreno, 3500 m. alt., 1901, Fries 759 (Stock). Salta: Cerro de Cachi, 3500-4000 m. alt., Venturi 6700 and 6706 (US); Cuesta de Arca to Trancas, Jan. 1897, Spegazzini (LP). ata- marca: Yutuyaco, dept. Andalgala, Jorgensen 1629 (G); Chucula, Feb. 2, 1930, Castellanos 30/520 (G). Tuc ee an: La Cienega, 1874, Hieronymus ¢& Lorentz 663 (US, FM, BD, isotype ot A. drepanophorus) and 693 (FM, K, BD); La Cienega, dept. Tafi, 2800 m, alt., 1933, eats 5390 (G); La Quenua, eck, Tafi, 3000 m. alt., Jan. 1930, Parodi 10769 (G) and Burkart 5389 (G); Lara, Jan. 1912, Rodriguez 261 (G); Est. Las Pavas, dept. Chicligasta, 3500 m. alt., Venturi 4710 (G, US). Corral Colorado, Sierra Famatina, Feb. 12, 1879, Hieronymus & Niederlein (BD). San Juan: Las Cabeceras de Leoncito, Jan. 1876, Echegaray (BD). ~— The leaves of this species are large and coarse and much surpass the inflorescence. The flowers are small, congested and short-pedunculate. The stems have short internodes and are usually less than a decimeter long, and usually spreading. The calyx is black hairy. The triquetrous 1947] JOHNSTON, ASTRAGALUS 387 fruit is characteristically lunate in outline and inside has a well developed membranous false septum. The synonyms I have listed all evidently belong to trivial phases of this readily recognized species. The species has only two close relatives, A. Brackenridgei Gray (cf. Johnston, Jour. Arnold Arb. 19: 95. 1938) of central Peru and A. cryptanthus Wedd the Bolivian Plateau. It is quite possible that the former is only a geographical variant and that the latter is only a freak glabrescent form of the present species. 49. ere cryptanthus Weddell, Chlor. a Bi (1861).— Type from la uebrada de las Lagunas de Potosi, d’Orbigny 14 bo ay Potosi: near snow-line, Queb. id - ee de Potosi, oo = Rese type). La Paz: La Paz, Aug. 24, 1901, Williams 2351 (US, NY). Arica: oe Cord. de Volcan Tacora, ca. 4500 m. alt., aan mann age (Gy KS This peculiar si is evidently related to A. arequipensis and may be only a very peculiar form of it. It is apparently rare for I have seen only three different collections, all remarkably similar though from widely separated stations, and all from within the range of A. arequipensis. The plants are distinguished from A. arequipensis by their bright green color and very sparse indument. ‘The stems also tend to be more elongate and the small leaves somewhat more succulent. The fruit is glabrous and may be lunate or may be short and oblong as in the related A. Brackenridgei of central Peru. 50. Astragalus Sanctae-Crucis staan Rev. Fac. Agron. y Vet. La Plata 3: 505 1897).— Type from Mon Leon near the Rio Santa Cruz, Patagonia, Spegazzini. Rance: Along the coast of Patagonia in sands and gravel. ARGENTINA. Santa Cruz: Minerales, 200 m. alt., Oct. 28 and Nov. 20, 1929, Donat 196 (G, NY, K); Monte Leon near Santa Cruz, in gravel, 1882, Spegazzini (type, La Plata). Chubut: campo near Cabo Raso, Fischer (LP). io Negro: Viedma, medanos del Faro, Be 14, 1928, Castellanos 28/1139 (G). Buenos Aires: Necochea, dunes, Jan. 4, , Cabrera 1303 (G); Punta Negra, 15 km. s. of Necochea, shifting sand near ae corolla blue, base of petals white, Dec. 12, 1938, Eyerdam, Beetle & Grondona 23728 (G) Without any close affinities in South America, this plant probably has its closest relatives in the group of A. Nuttallianus of North America. It is prostrate and has very elongate straight subcylindrical strict pods. -The false septum is well developed and nearly divides the fruit into two cells. 51. Astragalus orthocarpus, nom. nov Phaca oligantha Philippi, Tanase: 33: 42 (1864).— Type from the base of the cordillera in the prov. Colchagua, Chile. Astragalus oliganthus (Phil.) Reiche, Anal. Univ. Chile 97: 567 (1897), and FI. Chile 2: 110 (1898). Not Kar. & Kir. (1842). : Known only from the province of Colchagua, Chile, at the base of the cordilleras. HILE. Colchagua: Monte de la Leon, Bertero 65 (BM, Boiss, Paris) ; with- out locality, Bridges (G, K, BM, BD) and ex Philippi as P. oligantha (BD, Del, Boiss, Stock) ; Centinela, 400 m. alt., 1928, Montero 1382 (G). 388 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XXVIII A species readily confused with A. Bertert and perhaps most closely related to it. It may be distinguished from A. Berteri by its erect or ascending, short and stout, rather plump legumes. In habit the plant seems to be more slender than its relative, having elongate leafy prostrate or decumbent stems and apparently slightly smaller flowers. The fruit has convex valves which become tough and hard. The lower suture is situated in a broad shallow depression. Flowering plants of A. orthocarpus and A. Berteri are distinguished with difficulty. Among the species listed as synonyms of A. Berteri there are several founded on flowering plants which conceivably may be proved to belong to the present plant and so provide it with an older name. 52. Astragalus Domeykoanus (Phil.) Reiche, Anal. Univ. Chile 97: 550 (1897), and Fl. Chile 2: 93 (1898). Phaca Domeykoana Philippi, og Univ. Chile 84:18 1893).— Type from Tur- rieta, prov. Talca, lat. 35° 30’ S., Feb. 1879, F. Philippi. Phaca orites Philippi, Anal. cae Chile 84: 24 (1893).—Type from Gualletué, cord. prov. Cautin, Feb. 1887, Rahmer. oe — (Phil.) Reiche, Anal. Univ. Chile 97: 551 (1897), and FI. Chile 2:94 Pte procumbens var. aber: Kuntze, Rev. aa i 73 (1898). — Based upon specimen from “Patagonia,” Moreno & Tonin . Rance: Along the cordilleras an seh and cele secs to southern Patagonia. CHILE. Talca: Turrieta, rane? east of Talca, Feb. 1897, F. Philippi (LP, isotype of A. Domeykoanus). ule: Cord. de Maule, Germain (K, BM); Andes de Antuco, 2400 m. alt., 1828, ae (Munich) ; Cord. de Linares, ex Philippi 215 (BD). Cautin: Araucania, ex Philippi as P. araucanus (K, BD; ?Pisotypes of P. orite ARGENTINA. Neuquen: Pino Hachado, 1920, Parodi 2284 (Parodi); Cajon de Pino Hachado, Sept. 1900, Spegazzini (LP); Cerro Colohuincul, moist sandy hillside, 1500-1800 m. alt., Dec. 17, 1926, Comber 878 (K); Polcahue, 1200 m. alt. Jan. 16, 1926, Comber 458 (K); Valle Escondida, iron-bog in valley, 1800 m. alt., Comber 279 (K); Lagunas Lui Cullin, Comber 1169 (K); Sierra Mamuil Malal, 1800 m. alt., Jan. 3, 1927, Comber 926 (K). Rio Negro: Bariloche, Cerro Gutierrez, 1500 m. alt., Feb. 28, 1905, Buchtien 23 (US, K). Chubut: mountains near Carren-leofu, 1901, Jilin (LP); Valle de la Laguna Blanca, Dec. 20, 1901, Koslowsky 187 (K); Valle de la Lago Blanco, Dec. 1903, no collector given (LP). Santa Cruz: near ~_ yearn 1000 m. alt., Jan. 30, 1905, Dusen 5765 (Stock); Lago Posadas alt., Dec. 20, 1908, ‘Skottsbere 760 (Stock). In- definite: Patagonia, Psi = Tonini 727 (NY, type of var. glabrescens). This species is very closely related to the poorly understood A. compli- catus of the andes of Mendoza. It may be no more than a variety of that species with sparsely and inconspicuously strigose or glabrous leaflets. It has obcordate or retuse obovate-oblong leaflets that are normally folded along the midrib. In texture they are rather thick and fleshy. The fruit is apparently rather variable. Most specimens have sparsely black- strigose pods about 8 mm. long which are either distinctly triangular in cross-section or are more or less compressed. They have a distinctly in- truded lower suture. In some specimens the pods are not only larger (10-14 mm. long) and of more fragile texture, but are also less angular 1947 ] JOHNSTON, ASTRAGALUS 389 and more or less inflated. The fruit is then broadly lenticular in cross- section with the lower margin only weakly flattened. I am unable to correlate these surprising variations in fruit with any other structural variations of the plant or with geography. The type-collection of A. Domeykoanus is in the flowering condition and has no fruit. It is very similar to other collections from the cordilleras of south-central Chile which have short angulate pods that are contracted apically into a short beak. These plants are all low and spreading and come from near the continental divide. The type of P. orites comes from the headwaters of the Rio Bio Bio. The fruit is of the short angulate type. The leaflets though rather large are characteristically glabrous, deeply notched and evidently folded as in true A. Domeykoanus, but the stems are about 2 dm. long and appear to be erect. These differences in leaf-size and growth-habit may well be ecological. The type sheet of T. procumbens var. glabrescens, labeled merely as from Patagonia, is a mixture of A. Domeykoanus and A. Cruckshanksii. The specimen of the former is practically glabrous; that of the latter is evidently strigose. Since the material of A. Domeykoanus is more ap- propriately covered by the epithet ‘‘g/abrescens,’ I am of the opinion that Kuntze’s variety had best be treated as a synonym of the present species. 53. seca complicatus Gillies ex Hooker & Arnott, Bot. Misc. 3: 187 (1832). El Cerro de la Polcura, Gillies RANGE: Known only from the type locality. ARGENTINA. Mendoza: El Cerro de la Polcura, andes of Mendoza, Gillies, ‘, type This species is known only from a single collection made over a hundred years ago. It is evidently related to A. Domeykoanus, which ranges widely along the Andes further south. It differs from the southern plants only in its distinctly strigose foliage. The type came from a locality 200 to 300 km. north of the most northern known station of its relative. The relationship between the two species is so clear that I shall not be surprised if new collections from Mendoza prove that the species cannot be separated by the characters given. The name A. complicatus, being the oldest, must then be applied to the larger specific concept. 54. Astragalus punae, sp. nov. Planta cinerea; caulibus pluribus decumbentibus 6—8 cm. longis simpli- cibus vel basim versus longi-ramosis sparse plus minusve adpresse villosis ca. 1.5 mm. crassis; foliis pilis 0.5—1 mm. longis sinuatis leviter adpressis griseo-vestitis; rhachibus 3—5 cm. eae pete medium nudis; foliolis 9—11- jugatis plicatis, 4-7 mm. longis ca. 2 latis oblongis vel oblongo- ellipticis supra plus minusve nae subtus villosis apice obtusis vel truncatis saepe emarginatis; stipulis 4-5 mm. longis membranaceis sparse strigosis; racemis 5—10-floris capitato-congestis 0.7—1 cm. diametro saepe in axillis foliorum superiorum subsessilibus (O-5 mm. longe pedunculatis) ; bracteis lanceolatis 3-4 mm. longis; calycibus sessilibus 3-4 mm. longis 390 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII subtomentosis, lobis linearibus 1-1.5 mm. longis; vexillo 6 mm. longo, lamina 2.5 mm. lata apice emarginata; alis 5 mm. longis quam carina 1.5 mm. longioribus; leguminibus ascendentibus 4—5 mm. longis 2 mm. crassis 2—2.5 mm. altis dense albiterque villosis 4-seminatis unilocularibus, intus septum Seer aalrei haud gerentibus, a latere viso margine i la rectis et margine inferiore convexis, apice acutis, basi rotundis, sutura inferiore evidenter 0.7—1 mm. profundeque introflex fea NA. Puna de Atacama, ae de la Puna,” herb. Spegazzini, no. 2111, sub “A. bolivianus?” (La Plata, For this very distinct species, sis I know only from the type, I can suggest no close relative. In habit it somewhat suggests a slender villous form of A. arequipensis, but it is certainly not a close relative of that species. It also suggests some of the coarse forms of A. micranthellus and possibly may be closely related to that species. The subsessile dense axillary clusters of flowers, the villous indument, and the small ovoid villous pods are outstanding characters of A. punae. The fruit is not tough in texture. It completely lacks a false septum. 55. Astragalus micranthellus Weddell, Chlor. Andina 2: 262 (1861).— Original col- ections from near Puno, Peru, Weddell 4503, and from near La Laguna de Potosi, d’Orbigny 1436. Astragalus capitellus Britton, Bull. Torrey Bot. Club 16: 260 (1889).— Type from near La Paz, Rusby 1005. Tragacantha arequibensis var. minima Kuntze, Rev. Gen. 3:73 (1898).— Type from between Uyuni and Machacamarca, Bolivia, Kuntze. Pragacantha plete var, tenuifoliola Kuntze, Biv Gen. 3: 73 (1898).— Type raque, Kuntze. diese: aoacaac Ulbrich, Bot. Jahrb. 37: 417 (1906).— Type from Puna Patanca, Tarija, Fiebrig 2612 RANGE: oS oe on punas, from southernmost Peru, across Bolivia, to northern Argen nee near Juliaca, es Hill 127 and 128 (K); near Puno, Weddell 4503 (Paris, type of A. micranthellus). Arequipa: near Arequipa, 2000-2500 m. alt., 1923, Guenther & Buchtien (Hamb). Moquegua: Ubinas, March 1864, Ramon 11725 (BD). La Paz: La Paz, 3000-4100 m. , Buchtien 2872 (US), epee 626 (US, BD) and Rusby 1005 (NY, type of A. per eaiy Potopoto near La 3790 m. alt., 1861, Mandon (K, BM, Del, Stock) ; Guaqui plateau, 3750 m. alt., a Hill 125 (K); Viardn: ime m. alt., Jan. 29, 1921, Asplund 2110 and 2122 (Unedia) - Guaqui, 3900 m. alt., Jan. 31, 1921, Aabland 2221 (Upsala) ; Corocoro, 4000 m. alt., Feb. 15, 1921, Asplund 2440 Hee Ulloma, 3800 m. alt., Feb. 21, 1921, Asplund 2554 (Upsala); Charana, 4050 m. alt., March 1, 1921, Avian 2670 ating be- tween San Pablo and Achacachi, "1903, Hill 129 (K). Oruro: Hacienda Huanca roma near Eucaliptus, prov. Cercado, 3800 m. alt., 1934, Havin 111 (NY, Stock); Challapata, 3800 m. alt., March 31, 1921, Asplund 3282 (Upsala). Cocha- type of var. tenuifolia). Potosi: near La Laguna de Potosi, d’Orbigny 1436 (Paris, cotype of A. micranthellus) ; Potosi, 4000 m. alt., 1933, clon 366 (US) ; m. alt., March 25, 1921, Asplund 6144 (US) and es (Upsala) ; Atocha, 3700 m. alt., March 20, 1921, Asplund 2999 (Upsala). Sucre: Muyaquiri, prov. Cinti, 1 1934, Hammarlund 386 (Stock). Tarija: Escayadu, 3500 m. alt., 1904, Fiebrig 3498E 1947] JOHNSTON, ASTRAGALUS 391 (BD); Puna Patanca, 3700 m. alt., 1904, Fiebrig 2612 (BD, type of A. patancanus; G, K, Paris). ARGENTINA, Jujuy: Cerro de Chani, 4000 m. alt., 1929, Venturi 8526 (US); Moreno, 3500 m. alt., Fries 809 (Stock); north of Tres Cruces, 1921, Castellanos 53 (G); Santa Catalina, 3650-4100 m. alt., 1901, Claren 11465 in pt., 11476, 11419 and 11499 in pt. (Stock). Salta: Cerro de Cachi, 4000-4500 m. alt., 1927, Venturi 6681 and 6694 (US). Catamarca: Cerro Negro, dept. Andalgala, 3500 m. alt., fhe ie 1628 (G, US), 1629 (G). Tucuman: Lara, Trancas, 1912, Rodriguez 2 AG). The plants associated here are slender stemmed, depressed, and pros- trate or spreading. The small flowers are clustered on peduncles shorter than the subtending leaf. The pods are 5-13 mm. long, sparsely strigose and laterally compressed. The valves of the fruit are flattened and usually about 3 mm. broad. The lower suture is strongly intruded. Under favorable conditions the widely spreading stems may become 1-2 long and loosely branched. Most plants, however, have prostrate stems only 5-10 cm. long. Some of the plants from extreme environ- ments at high altitudes become very reduced and compacted, becoming somewhat caespitose or somewhat pulvinate (e.g. Venturi 6694). Most of the collections treated as A. atacamensis by Fries, Nova Acta R. Soc. Sci. Upsala IV, 11: 134 (1905), actually belong to A. micran- thellus. Among the collections cited by Fries, A. micranthellus appears in several mixtures; with A. Bustillosii nae 11465), with A. minimus (no. 916) and A. pvmnprilas (no. 11499 sub “A. modestus”). Hosseus, Bol. Acad. Nac. Cordoba 26: 157 (1921) a noted some of these Gare and misidentifications. 56. Astragalus triflorus (DC.) Gray, Plant. Wright. 2:45 (1853), as to name-bring- ing synonym only. Phaca triflora DeCandolle, Astrag. 62, t. 1 (1802).— Type given as collected in Peru by Dombey. Astragalus viciiformis Ulbrich, Bot. Jahrb. 37:550 (1906).— Type from near Mollendo, Weberbauer 1504. RANGE: Gani hills, in the Loma Formation, from central Peru to northern ile. ma: Cerro San Augustin, 300-400 m. alt., 1909, Weberbauer 5241 Raimondi 11593 (BD); Mollendo, sandy flat at lower edge of the fertile belt, 1902, WER arO eet. 1504 (BD, type of A. viciiformis) ; Mollendo, Johnston 3569 (G, K) and Hitchcock 22364 (US); Mejia, 30-40 m. alt., 1923, Guenther & Buchtien 273 and 272B (Hamb); Cachendo, 900-1000 m. alt., Guenther & Buchtien 272B and 274 (Hamb) ; Posco, 575 m. alt., 1923, Guenther & Buchtien 272 (Hamb). Moquegua: Torata, 2200-2300 m. alt., Weberbauer 7418 (US, K, BD). Indefinite: Peru, Dombey (Paris, type of P. triflora CHILE. Tarapaca: aoe Werdenana 754 (G, K, BM, BD). Antofa- gasta: Tocopilla, Johnston 3621 (G) and Jaffuel 1008 (G): Antofagasta, Jaffuel 1146 ( This very distinct species, a small slender annual of the coastal hills of Peru and northern Chile, was one of the very first members of its genus to have been described from South America. Though the original 392 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII plate and description are accurate, the species has been discredited and misunderstood and its name consistently misapplied. DeCandolle based his P. triflora upon specimens collected by Dombey in Peru, probably near Lima. The original plate is a careful drawing of a specimen now pre- served at Paris! The species is unquestionably the same as_ that described over a hundred years later as A. viciiformis Ulbr! Much of the confusion enveloping our plant for over a century may be traced back to Kunth, H. B. K. Nov. Gen. 6: 495 (1824). That botanist gave an excellent description and illustration of a plant which Humboldt & Bonpland had collected on the volcano of Jorullo in Mexico and pub- lished them under the newly proposed name, Phaca Candolliana. He associated that name with the following synonymy, ‘“Phaca triflora Decand. Astrag. p. 62. t. 1 (auct. Mu. Paris). Persoon Synops. 2. p. 331.” The Mexican plant is very different from the coastal annual of Peru and Chile. Kunth, however, seems to have thought them to be conspecific and various subsequent authors have inferred that the name he coined, P. Candolliana, was merely a substitute name for the “Inappropriate and misleading” P. triflora DC. Even DeCandolle so considered it, for in his Prodromus, 2: 273 (1825), he gives P. Candolliana HBK as a synonym of P. triflora DC. Most authors have agreed with Jones, Astrag. 105 (1923), that the names P. Candolliana and P. triflora were not only “‘in- tended to be the same” but were originally applied to conspecific plants as well. The prevailing opinion seems to have been that DeCandolle’s description and plate were crude and poor (Jones calls them ‘“fanciful’’) and that his material may have come, not from Peru, but rather from Mexico. ‘This is, of course, thoroughly incorrect and not in accord with the facts. DeCandolle’s plate is a good representation of his species, which, we now know, did come from Peru. There are good reasons for considering that the name P. Candolliana HBK. was intended as a substitute for DeCandolle’s binomial, P. triflora. Personally, however, I feel that Kunth’s binomial can well be restricted to the Mexican plant he figured and described. This has been recently done by Rydberg, No. Amer. Fl. 24: 350 (1929). The accepted name for our plant of Chile and Peru was published by Gray, l.c., as follows: “A. (Phaca) triflorus. Phaca triflora, DC. Astrag. t. 1 & Prodr. 2. 273. P. Candolliana, H. B. K. Nov. Gen. & Sp. 6. p. 495. t. 586.” It is evident that Gray accepted P. triflora and P. Candolliana as conspecific and as applying to the Mexican species as well as to the Texan plant he had in hand. This, of course, was incor- rect, but since his nomenclatorial transfer is clearly and expressly founded upon Phaca triflora DC, even though he misunderstood DeCandolle’s concept, the resulting combination, Astragalus triflorus (DC.) Gray must apply to our plant of ao and Peru. Some authors, however, have insisted that the binomial A. triflorus must be applied to the plant of Texas which Gray was noes when he made the combination, but this is clearly counter to the rules of nomenclature. This confusion of ‘‘species- 1947] JOHNSTON, ASTRAGALUS 393 concept” and ‘“‘nomenclature” reached its extreme with Sheldon, Minnesota Bot. Studies 1: 140 (1894), who cited “A. triflorus A. Gray. Pl. Wr. 2: 45. 1853. excl. syn.” as a synonym of A. Candollianus (HBK.) Sheldon and proposed the combination ‘A. triflorus (DC.) Sheldon” for our Phaca triflora DC. of South America. The two most recent monographers have agreed in treating A. triflorus Gray as based upon Phaca triflora DC. Jones, Astrag. 105 (1923) considered it to be a species of both North and South America, while Rydberg, No. Amer, Fl. 24: 350 (1929), treated it as exclusively South American. 57. Astragalus Famatinae, sp. nov. Planta perennis plus minusve argenteo-strigosa; caulibus gracilibus rigidiusculis prostratis 5-18 cm. longis 0.5—1 mm. crassis foliosis (interno- iis 0.5-5 cm. longis) e caudice oe radice valida profunda coronato erumpentibus; rhachi folii 5-20 mm. longa supra medium vel usque ad basim foliolata supra canaliculata; foliolis 4—5- -jugatis 2-9 mm. longis 1-2 mm. latis ellipticis vel anguste oblongis plus minusve plicatis margine saepe paulo involutis apice rotundis vel obtusis rariter emarginatis subtus_ strigosis supra plus minusve glabris; stipulis 2—3 mm. longis strigosis ocreatis; racemis axillaribus densifloris 8-12 mm. diametro; u floribus saepe 10-15 capitato-congestis; calycibus 2-3 mm. longis, lobis 0.5—1 mm. longis as nigris et albis strigosis; pedicellis 0.5-1.5 mm. longis; bracteis 0.5-1 mm. longis oblongis; vexillo corollae 5—7 mm. longo, lamina 3—4 mm. ee alis 4.5—-6.5 mm. longis quam carina 2 mm. longiori- bus; ovario strigoso; leguminibus oblongis compressis vel prismaticis ca. 1 cm. longis 3—3.5 mm, altis 2—3 mm. crassis sparsissime strigosis 8-10 seminatis, a latere viso utrmque marginibus convexis, sutura inferiore late vel anguste introflexis, intus septum membranaceum ca. . alt gerentibus. ARGENTINA. La Rioja: La Incrucijada, Sierra Famatina, 1879, Hieronymus ¢& Niederlein 482 (type, Berlin); Los Berros, Sierra Famatina, 1879, Hieronymus c& Niederlein 586 (BD). San Juan: Quebrada del Salto, Cord. de Colangiiil, Jan. 15, 1930, Perez Moreau 30/98 (G). Although with a gross aspect more suggestive of A. micranthellus, this well marked species seems to be much more closely related to the habitally dissimilar A. tarijensis. The fruit has a somewhat tough texture and seems to be more or less prismatic with the lower face flattened or shallowly and broadly introflexed. Some of the fruit on the type has the lower suture strongly and deeply intruded, but this may be the result of lateral pressure on the fruit when the specimens were pressed. ‘There is a narrow but evident membranous false septum inside the fruit. The fruit of A. Famatinae, hence, has a different form (prismatic and oblong) and a firmer texture, as well as a false septum, which distinguishes it from the fruit of A. micranthellus. The plant is furthermore clothed with a coarser and somewhat silvery strigosity and hence differs from A. micranthellus in indument. The thicker, densely strigose leaflets with entire or only obscurely emarginate apices, and the firmer texture of the pods, readily distinguish A. Famatinae from all forms of A. tartjensis. 394 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII The type of the species has been cited by Hosseus, Bol. Acad. Nac. Cordoba 36: 152 (1921), as A. modestus. This is certainly incorrect. Hosseus has cited another isco (no. 535) collected by Hieronymus & Niederlein in the Sierra Famatina. This latter, only a young specimen at anthesis, I examined as a sinaie in Rerlin. n my notes this is assigned to “A. micranthellus,’ but now for various reasons I would wish to re-examine the collection before accepting that identification. The material cited from San Juan is questionably referred to 4. Famatinae. It represents a plant evidently related to that of the Sierra Famatina but has stems only 0.5-1 cm. long and has much contracted internodes. The leaves are consequently. somewhat tufted and the in- florescence almost scapose. I believe it is a reduced plant conspecific with those from La Rioja. The material of it, however, is very fragmen- tary and its present disposition must be considered tentative. 58. re tarijensis Weddell, Chlor. Andina 2: 262 (1861). eer from _be- Puna Iscaiachi and Valle de Tarija, Bolivia, Weddell 4 Pcie t Hieronymi Ulbrich, Bot. Jahrb. 37: 418 (1906). eae upon collec tions from Cienega, Tucuman, Lorentz & Hieronymus 580 and Calderillo, Tarija, Fiebrig 2882. RANGE: — of northwestern Argentina and adjacent Bolivia, mostly above 3000 m. altitu BOLIVIA. cae: Calderillo, 3400 m. alt., 1904, Fiebrig 2882 (FM, BD, cotype of A. Hieronymi); along the descent from Puna de Iscaiachi into Valle de Tarija, 2500-3000 m. alt., Jan. 29-31, 1846, Weddell 4021 (type of A. tarijensis, Paris). ARGENTINA. Jujuy: Tilcara, 3000-3200 m. alt., Venturi 6698 and 7421 (US) ; Cerro La Soledad, 3500 m. alt., 1928, Venturi 9029 (G, US). Salta: between Cuesta del Arca and Trancas, Dec. 1896, Spegazzini (LP); Pampa Grande, Jan. 1897, Spegazzini (LP); La Laguna, Sierra del Cajon, 3900 m. alt., Jan. 1914, Rod- riguez 1317 (G). Catamarca: Cerro Negro, Feb. 9, 1916, Joergansen 1887 S); C. de Aconquija, 3800 m. a Marth 15, 1917, dance 1337 (US); Cerro del Campo Grande, Jan. 1874, Schickendantz 256 (BD); Granadillas, Nov. 1877, Schickendantz 88 (BD); Laguna del Tesoro, 4600 m. alt., 1925, Venturi 6629 (US) ; Sierra de Anconquija, 4400 m. alt., 1926, Venturi 6623 (US). Tucuman: Lara, Trancas, Jan. 27, 1912, Rodriguez 280 (G); Cumbres Calchaquies, Tati, 4000-4400 m. alt., Burkart 5402 and 5387 (G) and Parodi 10846 (G); La Quefua, 3000 m. alt., Burkart 5397 (G); La Puerta, 4000 m. alt., Burkart 5395, 5396 and 5388 (G); Penas Azules, 3400-3500 m. alt., Barkart 5398 (G) and Parodi 10911 (G); Cerro de San an Venturi 4681 and 6622 (US); Estan. ae Rik 3800-4400 m. , Venturi 3183 (G, US), 4007 (US) and 4749 (G, er P). La Rioja: aera 6, Camino a La Mejicana, Sierra Famatina, 3300 m me 1927, Parodi 7954 (G) The types of A. tarijensis and ri ae are very similar in aspect and details. They are evidently conspecific and representative of the most common form of this very variable species. This common form has slender well developed ascending stems, green folded narrowly obcordate leaflets, black-strigose calyces, lanceolate bracts and _ strigose ovary. There are, unfortunately, frequent departures from this distinctive typical form. At very high altitudes the leafy stems become very short and the plant matted. The apex of the leaflets, usually broadly retuse, may 1947] JOHNSTON, ASTRAGALUS 395 become rounded and only weakly emarginate, or even entire and actually cute. In outline the leaflets are commonly obcordate or oblong-obcor- date. Some plants, evidently forms of A. tarijensis, have distinctly ob- long, elliptic or even lanceolate leaflets. Most forms of the species have practically glabrous herbage of a characteristic dark green color. But I am also forced to admit in the species some plants having the herbage somewhat cinereous with evident scattered appressed hairs. When the common form of the species varies in only one of the details mentioned, its affinities are not obscured. When, however, several of these aberrant tendencies are exhibited in one alent its general appearance becomes very changed and quite different from that of the typical form. After a con- sideration of these puzzling forms I have finally been forced to place them under A. tarijensis as chance combinations of aberrant characters. They occur within the range of the normal form of the species and are usually represented by single collections. A study of a large series of A. tarijensis makes it clear that they are merely variants of this unstable species. 59. Astragalus Weddellianus (Kuntze), comb. Astragalus modestus Weddell, Chlor. Andina i a6 (1861). Not Boiss. & Hohen, 49).— Type from Cuesta de Periguani east of Camargo, Weddell 3932. Tragacantha Weddelliana Kuntze, Rev. Gen. 2:942 (1891) —Based upon A. modesta Wedd. sos Ric: hag ot Field Mus. Pub. Bot. 11:25 (1931).— Type from Pauc mbo, Cuzco, Herrera 231 Astragalus. sia Mcbride, Field Mus. Pub. Bot. 13 [FI. Peru 3:]: a (1943). m Macusani and Hac. Salcedo, dept. Puna, Soukup 5 RANGE: ie from southern Peru to northwestern Argentina. PERU. Cuzco: Hacienda Chuchapaya, = del Paucartambo, 3800 m. alt., Feb. 1929, Herrera 2315 (FM, type of A. eri). Puno: indefinite baeahty. Soukup 429 (G) and 549 (G, isotype of A. sare 3). La Paz: Rinconada, 4200 m. alt., Dec. 16, 1920, Asplund 1804 (Upsala) ; Desaguadero, 3900 m. alt., Feb. 3, 1921, Asplund 6142 (US); Guaqui to Desaguadero, 3850 m. alt., Feb. 3, 1921, oa ns psala). Potosi: between San Sebastian and Dites, 4060 m. alt Nov. , Cardenas 281 (G, US). Chu- quisaca: Cuesta de Periguani, between eee ard aaa: prov. Cinti, 3500 m. alt., Jan. 11-13, 1846, Weddell 3932 (type of A. modestus, Paris). ARGENTINA. Tucuman: puna between La Puerta and Los Cardones, Cum- bre Calchaquies, ci 4400 m. oS Jan. 30, 1933, Burkart 5384 and 5408 (G). atamarca: Reales Blancos, Feb. 3, 1930, Castellanos 30/488 (G Assembled a under the name A. Weddellianus is a eauiabie group of plants all evidently related to A. tarijensis and differing from it only in the reduced habit of growth, glabrous ovary and fruit, and generally more northern distribution. These distinctions are weak and perhaps the plants can be accommodated under A. ¢arijensis as further variants of that variable species. The type of A. modestus has leafy stems 1-2 cm. long, 2—4-flowered peduncles 5-10 mm. long, practically glabrous stems and leaves, and 6—8-jugate elliptic-oblong leaflets. The type of A. Pilgeri is sparsely and persistently strigose but is otherwise similar to the type of A. modestus. The cited collections of Asplund and Cardenas have more reduced stems 396 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XXVIII and peduncles than in the type of A. Pilgeri but agree with it in other details including pubescence. The collections from Argentina have gla- brous herbage and so more closely resemble the ey of A. modestus. ‘The cited material from Puna, including the type of A. punensis, is glabrous and differs in its larger corollas and deeper calyx from other collections examined. Possibly further collecting may prove it to be a distinguishable entity worthy of recognition. 60. —— Bustillosii Clos in Gay, Fl. Chile 2: 117 (1846); Wedd. Chlor. Andina : 25 — Type from Cordillera de la Patos, Gay 489 ).— Based upon collections from Ag. 1861) ma ae Phil. Fl. Atac. 14 (1860 Veras, and Puquios, Philippi. Piivaeates depauperatus (Phil.) Reiche, bey Univ. . Chile 2: 80 (1898). Not Ledeb. 31). Phaca saxifraga Phil. Cat. Pl. Itin. ial 14 (1891).— Type from Penon, prov. Atacama, Philippi. Astragalus brachycalyx Phil. Cat. Pl. Itin. Tarapaca 15 (1891); Reiche, Anal. Univ. pe 97: 538 (1897), and FI. Chile 2: 81 (1898). Not Fischer (1853).— Type m Near Machuca, prov. Antofagasta, Philippi. uscieeny atacamensis Kuntze, Rev. Gen. 2: 940 (1891).— Based upon P. depauperata Phil. Astragalus atacamensis (Kuntze) Fries Nova Acta R. Soc. Sci. Upsala, ser. IV, 11; 134 (1905). — heshigrigrvins Speg. Anal. Mus. Nac. Buenos Aires II, 4: 264 (1902).— Bas n A. brachycalyx Phil., en Fischer RANGE: a cei gravels at pay altitudes from ‘ntolican and Jujuy south to Tucuman, Mendoza and Coquimbo. CHILE. Antofagasta: ean lat. 22° 36’, ilippi (Santiago, type of A, oo Atacama: efion, lat. 26° 50’, a eri Philippi as P. saxi- fraga (US, BD); Desert of Atacama, Philippi as P. depauperata (K, BD, Boiss) ; Laguna — 3400 m. alt., vig Manni 5940 (G, US, K); Laguna Chica, 3200 m. alt., 1926, Johnston 5970 (G, , Stock); Laguna Valeriano, 4000 m. alt., 1926, Johnston 6041 (G, US, K, Stock). “Coa imbo: Banos del ng _ 1904, Reiche G); Cord. de Dafa Rosa, herb. Ree } (K): Cord de Los Patos, Gay 489 (Paris, type of A. Bustillosii) Polcura, ond de snl ra: 1888, Philippi Chile 97: 537 (1897), and Santa Catalina, 3650 m. alt., 1901, Claren 11465 in (Stock). Salta: Pr Quesera, ca. 3500 m. alt., 1932, Keidel 1 (LP); Cerro . Cachi, 3600 m. alt., 1927, Venturi 6705 (US). Tucuman: Tafi del Valle, fl. violet, 1907, herb. poten ce uan: Vega de la Sepultura, Valle Cura, ca. 3800 m. alt., Jan. 7, 19 rez Moreau 30/95 (G). Mendoza: Puente del Inca, March 1908, herb. ee 2029 (LP); below Paso de la Cruz Piedra, 3000 m. alt., Kuntze (NY). Among the South American species of the genus A. Bustillosii is unique in the possession of slender rhizomes. The slender stems of the plant elongate and spread through the moist sand or gravel in which it grows. A single plant can form a large area of turf. It much resembles A. mini- mus in habit, foliage, and gross aspect and perhaps it is related to that species. It is quickly distinguished from this possible relative, however, by its very different underground parts, its larger flowers, and very different fruit. The pods are strongly compressed laterally and flattened. The lower suture is inflexed only near the base of the pod and the false 1947] JOHNSTON, ASTRAGALUS 397 septum is absent or only very weakly developed. The valves are com- monly 10-13 mm. long and 5—7 mm. broad. The pedicels and peduncles are only slightly more developed than in A. minimus. ‘The species is a rather stable one and has a range which lies to the south and west of A, minimus. With the possible exception of A. brachycalyx Phil., the species I have listed above are all clear synonyms. Through the kindness of Prof. Espinoza I have examined part of the type of A. brachycalyx. It appears to be a depauperate plant of A. Bustillosii from an unfavorable habitat. The material is scanty and shows only a single immature fruit but seems to fall reasonably well into the present species. 61. Astragalus uniflorus DC. Astrag. 243, t. 50 (1802), and Prodr. 2: 306 (1825) ; Weberb. in Engler & Drude, Veg. bail 12: 209, fig. 45 (1911).— Type collected y at Huasa-Huas che eee Hersogi Ulbrich, Mededei Rijks Herb. 27:53 (1915).— Type from e Choquecota, Cord. Quimza Cruz, Bolivia, Herzog 2340. RANGE: pe mountains of central Peru south to the Bolivian Plateau. PERU. Junin: Hacienda Arapa near Yauli, 4400 m. alt., 1902, Weberbauer 290 (BD, basis of Weberbauer’s illustration!) ; Oroya, 4300 m. alt., 1903, Weberbauer 2620 (BD); Alpamina Mines, 1904, es a 5111 (BD); Cerro de Pasco, ca. 4200 m. alt., 1923, Macbride 3066 (FM) ; Huasa Huasi, 1779, Dombey (Paris, type). Apurimac: crest of Socollaccas Pass, 4000 m. alt., open puna among grass, lavender, Nov. 11, 1935, West 3829 (UC). Cuzco: La Raya, ca. 4300 m. alt., 1925, Pennell 13526 (FM.) i La Paz: Huaycho to Molo, 4200 m. alt., 1903, Hill 123 (K); Sorata, 3900 m. alt., Feb. 1886, Rusby 1925 (NY); near Sorata, 3800 m. alt., Nov. 1857, Mandon 713 (G, K, BM, BD, Boiss, Del, Paris) ; hills les Waren ais, 4050 m. alt., 1904, Hill 124 (K); near top of Huallata Pass, 4233 m. alt., Conway 39 (K) ; Corocoro, 4400 m. alt., Feb. 16, 1921, . sae 2487 (Upsala) ; Palca to La Paz, 4200-4700 m. alt., Pflanz 136 and 226 (BD); La Paz to Palca and Illimani, 3600-4800 m. alt., Hauthal 317 and 325 (BD); Sehr la Apacheta, Illimani, 4670 m. alt., Stiibel 24a and 29a (BD); Cord. Quimza Cruz, valley of the Choquecota, 4300 m., alt., Oct. 1911, Herzog 2340 (BD, type of A. Herzogii). Oruro: Cord. de Azana- que, 4400 m. alt., Dec. 26, 1926, Troll 2995 (BD). Chuquisaca: Cuesta de 3934 (Paris). Cochabamba: Cord. de la Llave, prov. Ayopaya, 4000 alt:, Dec. ee uae 4137 and 4138 (Paris). Indefinite: Bang 1792 (G) and 1879 (G, NY, U A low caespitose plant of high altitudes with glabrous herbage and large, pee brightly colored reddish purple flowers. Collectors, attracted only by its conspicuous flowers, have failed to collect the fruit of this ee ae all the collections examined only three (Asplund 2487, Bang 1879 and Rusby 1925) show the fruit. The legume is a remarkable one. It is glabrous, rugulose and completely 2-celled with a coarse septum almost as tough and firm as the valves, and is rigidly inflated and dorsi-ventrally somewhat compressed so that measurements across it about equal those for its total length and exceed its dorsi-ventral measurements. Hence it becomes 4—9 mm. long and thick and only about 3—7 mm. high. Viewed from below the legume appears didymous, being obtuse or rounded at the base, broadly retuse at the apex, and having a narrow deeply in- flexed lower suture. The valves are flattened on the back and sides and 398 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XXVIII hence are somewhat rectangular. The flowers are solitary on peduncles that become 5-10 mm. long in fruit. The pedicels are 1-2 mm. long. The calyx-tube is elongat The species is best om from the department of La Paz in Bolivia. This Bolivian plant, of which a form has been described as A. Herzogii, is clearly conspecific with the original A. uniflorus of the western Andes of Peru. The collection at Paris labeled ‘“‘in altis et frigidis, Huasa-Huasi, 1779, Dombey,” is apparently the basis for the plate, as well as for the description originally given by De Candolle. The species is a very dis- tinct one and is readily recognized. 62. Astragalus Venturii, sp. nov. Planta perennis caespitosa; ramis 2—4 mm. crassis stipulis ocreaceis aureis strictis obtectis e caudice horizontali laxe ramoso orientibus; ramis foliatis perbrevibus 5-10 mm. longis, internodiis stipulis brevioribus; foliis vix numerosis ascendentibus, rhachi 6-20 mm. longa, foliolis saepe adpressos vix abundantes strigosos non rariter subflaventes gerentibus; pedunculo laxe 1-—2-flori 3-12 mm. longo; pedicello 1-2 mm. longo: calycibus 4—5 mm. longis sparse strigosis pilos nigrescentes non rariter gerentibus, lobis subulatis 1-2 mm. longis saepe nigro-strigosis; vexillo ca. 7 mm. longo (lamina 5—6 mm. lata) quam carina 3 mm. et alis 1 mm. longiore; ovario glabro; leguminibus glabris compressis unilocularibus (vix septatis) margine inferiore saepe plus minusve inflexis, valvulis 4—7 mm, longis et 2.5—5 mm. latis. ARGENTINA. Jujuy: Tilcara, 3200 m. alt., fl. blue, Feb. 1927, Venturi 7417 ype Nat. Herb). Tilcara, 3200 m. alt., Feb. 1927, Venturi 7411 (US); Cerro La ae sale Humahuaca, 3500 m. alt., fl. blue Jan. 1927, Venturi 9018 (G, US). Sal ro de Cachi, fl. bluish, March 1927, Venturi 6674 (US); Cerro de Cachi, 4400 m. fl. violet, March 1927, Venturi 6691 (US). Tucuman: Sierra del Cajon, dept. Tafi, 3500 m. alt., fl. violet, May 1926, Venturi 7609 (US). Catamarca Sierra ss Anconquija, dept. Santa Maria, 4600 m. alt., fl. violet, Jan. 1926, Venturi 7605 (U The cans of this plant is uncertain. Among the caespitose species of northwestern Argentina it is quickly distinguished by the yel- lowish green color imparted to the plant by the yellowish stipules that ensheathe the caudex and, partially, by the yellowish hairs on the herbage. The only species readily confused with the species is A. Burkartii. The available material is not fully mature and is inadequate for the proper description of its fruiting structures. The immature legumes are variable, some appearing to have both sutures prominent while others have the lower suture strongly intruded. 63. Astragalus confinis, sp. nov. Planta perennis caespitosa ; ramulis ears paucis gracilibus horizon- talibus 1-2 mm. crassis; ramulis foliatis 5-12 mm. longis paucis dense albo-strigosis, internodiis 2—5 mm. longis; folis pallide — rhachi 2-5 mm. longa, foliolis 2—4-jugatis oblongi-linearibus 3-10 mm. longis 1—2 mm. latis margine involutis, stipulis 2-4 mm. longis strigosis laxe vel vix imbricatis; pedunculis 1-4 cm. longis apice 2—5-floris; pedicellis 1 mm. 1947 ] JOHNSTON, ASTRAGALUS 399 longis; vexillo 6-7 mm. ee (lamina 3-4 mm. lata) quam carina obtusa ca. 2.5 mm. et quam alis 1 mm. longioribus; calyce 3—4 mm. longo pallide strigoso, lobis triangulari-lanceolatis 1-1.5 mm. longis; ovario glabro; leguminibus glabris ca. 8 mm. longis 3-4 mm. latis margine inferiori in- flexis, septo membranaceo angustissimo inconspicue. PERU. Arequipa: near Arequipa, 2000-2500 m. alt., Feb. 13, 1923, Guenther & Buchtien 1082 (Hamb); Arequipa, 2300 m. alt., Feb. 11, 1923, Guenther & Buchtien 1081 (type, Hamburg). CHIL Tarapaca: between Pachica and Aico, 2400 m. alt., April 15, 1927, Troll 3305 (BD). A species evidently related to A. hypsogenus of northwestern Argentina and adjacent Bolivia from which it differs sharply in its fewer leaflets, sparser indument, slightly smaller flowers, and glabrous ovary and fruit. Furthermore, the leafy stems seem to be more elongate than in its southern relative. Except for the rudimentary septum and lack of hairs, the fruit of A. confinis is very similar to that of its relative. 64. Astragalus hypsogenus, sp. nov Planta caespitosa ; caudice eral a laxe longeque ramoso, ramulis 1-2 mm. crassis cum stipulis 3-6 mm. longis imbricatis sparse strigosis vestitis; foliis pallide strigosis saepe sericeis sed non rariter subglabrescen- tibus, rhachi 2—5 cm. longa, foliolis lanceolatis vel oblongis plicatis vel margine involutis 3-17 mm. longis 2—3 mm. latis 4—7-jugatis distantibus apice acutis; pedunculis 2—6 cm. longis apice 2—10-floris; pedicellis 1-2 mm. longis; calyce ca. 5 mm. longe strigoso pilis non rariter nigris vestito, lobis 1-2 mm. longis: vexillo 8-10 mm. longo (lamina 4—6 mm. lata apice emarginata) quam carina 3—4 mm. et quam alis angustis 1 mm. longiore; leguminibus saepe sublunatis strigosis 10-15 mm. longis 3-5 mm. altis margine inferiore plus minusve introflexis margine superiore a latere viso plus minusve leviter concavis, septo nullo. BOLIVIA. Potosi: Atocha, 3700 m. alt., March 19, 1921, Asplund 6135 (US) and 2974 (Upsala). ARGENTINA. Jujuy Esquina Blanca, Feb. 13, 1921, Castellanos 32 (G); Cerro ineatasi dept. Caciitiocs, 450Q m. alt., fl. light blue, 1930, Venturi 10126 (type, Gray Herb.; isotypes, US, BM, Stock); Laguna Tres Cruces, Claren 11695 (Stock) ; Moreno, 3800 m. alt., 1901, Fries 821a (Stock); Abra del Palamar, 3600 alt., 1901, ere 621 (St saa Sierra del Aguila, 3700 m. alt., fl. bluish, 1929, Venturi 8694 (US); Santa Catalina, 3650 m. alt., Claren 11442 (Stock). Los Andes: Chorrillos, March, 1930, Budin 12 (G); San Aneonts de los Cobres, Feb. 25, 1927, 1932, Keidel 4 (LP); Cerro de Cachi, 4000 m. alt., fl. bluish, March 1927, Venturi 6675 (US). Tucuman: between Estan. Santa Rosa and La Cueva, dept. Chicli- gasta, 3600 m. alt., fl. violaceous, Jan. 1926, Venturi 7606 (US). A very distinct species with evident relationships only with A. confinis and A. crymophilus. Among the species of northwestern Argentina it is readily recognized by its tufts of silvery strigose leaves and its pedunculate subumbellate cluster of flowers. In general appearance the fruit of A. hypsogenus is suggestive of a small fruit of A. arequipensis. ‘The upper edge is somewhat concave in lateral outline and the apex is acute. The lower part of the fruit is either flattened or inflexed. There is no false septum developed within the pod. Among the collections cited above 400 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII those by Fries and by Claren have been identified by Fries, Nova Acta R. Soc. Sci. Upsala ser. IV, 11: 135-36 (1904), as A. aida and A. bellus. 65. Astragalus crymophilus, sp. nov. Planta caespitosa; caudice denso breviter ascendenterque congesti-ra- moso; ramulis caudicis 3-10 mm. longis stipulis glabris membranaceis im- bricatis dense vestitis; foliis argenteis dense sericeo-strigosis, rhachi 2—5 . longa infra medium nuda, foliolis 2—3-jugatis linearibus plicatis dis- ‘itibus 8-19 mm. longis 1-2 mm. latis; pedunculo 2-15 mm. longo sericeo-strigoso apice 2—6-flore; pedicellis 0.5-1 mm. longis; calycibus 2.5-3 mm. longis sericeo-strigosis, lobis ca. 1 mm. longis linearibus vel lanceolatis tubo campanulato subaequilongis; vexillo 5-6 mm, longo, lamina 2-3 mm. lata obovata apice emarginata quam alis angustis 1 mm. et quam carina obtusa 2 mm. longiori; ovario albiter strigoso; legumine oblongo compresso strigoso 6-8 mm. longo 2—4 mm. alto 1-2 mm. crasso, margine superiori a latere viso recto vel leviter convexo, margine inferior valde introflexo, rae interiore membranaceo angusto O—O0.5 mm. alto incon- spicuo; ovulis 1 2 BOLIVIA: Uyuni, dept. Potosi, 3700 m. alt., March 1923, Asplund 6134 (type, S. Nat. Herb.) and 3148 (Upsala) ; pur Patanca, dept. Tarija, 3800 m. alt., March 27, 1904, Fiebrig 2912 (G, US, K, , BD). ARGENT INA: Mina Perdida, dept. - Catalina, Jujuy, 4100 m. alt., Jan. 25-26, 1901, Claren 11499 in pt. (Stock). A well marked species most closely related to A. Aypsogenus, from which it differs in its denser more silvery indument, fewer linear leaflets, oblong rather than lunate pods and narrow but definite false septum within the fruit. The collection from Jujuy consists of a single plant mixed with a reduced form of A. micranthellus. Fries determined and reported the collection as A. modestus. 66. Astragalus pusillus Vogel, Verhandl. K. Leop.-Carol. Akad. Naturf. 19: suppl. 19 (1843) ; Weddell, Chlor. Andina 2: 259 (1861).— Type from Alto de Toledo, aa Mevyen Ranc High a titudes from southern hy southward across the Bolivian plateau to the snr te of northwestern Argentina. PERU. Puno: Poto, prov. Sandia, 4500 m. alt., 1902, Weberbauer 993 (BD) ; cordilleras near lake Ayapata, Lechler 1730 (K); Azangaro, June 1854, Lechler 3242 K); Crucero, 1920, Stordy (K); Pucara, 3700 m. alt., Weberbauer 438 and 440a (BD); Moho, 1848, Weddell 4414 (Paris). Arequipa: above Chivay, prov. Caylloma, 4200 m. alt., 1914, Weberbauer 6896 (BD, FM); Alto de Toledo, 4650 m. alt., 1831, Meyen (BD, type); near Arequipa, 2200-2500 m. alt., 1923, Guenther & Buchtien 1114, 1117, 1118 and 1120 (Hamb). sas hk deal Coalaque near Lagunillas, 4400 m. alt., 1923, Guenther & Buchtien 1866 (H H Arica: Cord. de Volcan Tocora, Ancara, prs m. alt., Werdermann 1128 ied BD, US). LIVIA. La Paz: Achachachi, Ancorairnes and Guatata, 4000-4200 m. alt., 1857 , Mandon 712 (G, K, BM, BD, Del, Paris, Stock); La Paz, 4100 m. alt., Buchien 64/624 (K, BD, US); Lancha, between La Paz and Coroica, 5000 m. alt., Mandon 718 (K); La Lancha, Canyon of the Chuquiagillo, 1851, Weddell (Paris): Cerro Quimsachata, 13 km. s. of Tiaguanaco, 4200 m. alt., West 6379 (UC); Challa- velina, Feb. 9, 1927, Troll 3055 (BD); Ulloma, 3900 m. alt., Feb. 2, 1927, Troll 3119 1947] JOHNSTON, ASTRAGALUS 401 (BD) ; Ulloma, 3850 m. alt., Feb. 21, 1921, Asplund 4423 (Upsala) ; Mina La Union, 4400 m. alt., May 4, 1927, Troll 1939 (BD); Charafia, 4000 m. alt., March 1, 1921, Asplund 6143 (US) gud 2671 in pt. (Upsala) ; Corocoro, 4300 m. alt., Feb. 16, 1921, Asplund 2485 (Upsala) ; Calacoto, 3800 m. alt., March 7, 1921, Asplund 4432 (Upsala). Potosi: Totorapalca, prov. Lina 3400 m. alt., 1934, Hammarlund 426 (Stock); Potosi, 4000 m. alt., 1932-33, Cardenas 165 (G) and 466 (US). Indefinite: near Coldo, Andes of Pelectiica: 4800 m t., April 1865, Pearce (BM); prov. de Carangas, ARGENTINA. Salto: is Potreros, south of Nevado del Castillo, March 1873, Hieronymus & Lorentz 56 (BD, K, FM). Bolivian material of this well marked species has been frequently con- fused with A. peruvianus. The latter, however, is readily distinguished by its smaller fruit, its persistent silvery indument, its narrow acute to rounded leaflets, and its usually firm tight yellowish stipular sheaths that commonly form an evident collar about the leaf-rosette. The real rela- tionships of our present species, however, seem to be with the habitally sensu A, arequipensis. Its sparsely villous pods are ca. 5 mm, long and 3 m road. The lower suture is inflexed and there is a false septum. One of th most distinctive developments of A. pusillus is its compacted inflorescence. Terminating the branches of the multicipital caudex and nestling in the center of the rosette of rather coarse leaves, sessile flowers are to be found intermixed with bracts and broad white papery stipular- sheaths. The foliage is villous becoming glabrescent in age. The indu- ment is not lustrous. The leaflets are oblong to obcordate, retuse, and 2-5 mm. broad. The plant is one of the peculiar habit-types character- istic of the Bolivian altiplano. The only Argentine material I have seen comes from southern Salta. Hosseus, Bol. Acad. Nac. Cien. Cordoba 26: 154, fig. 27 (1922), however, reports A. pusillus from the Sierra Famatina, La Rioja, on the basis of six collections which he compared and found conspecific with the material obtained in Salta by Lorentz & Hieronymus (no. 56) 67. rw Pequot vanes Bot. Jahrb. 37: 422 (1906) ; Sasa? a Jahrb. 2 (1910).— Type from Calderillo, southern Bolivia, Fiebri oe ae Paz on 4050 m. alt., March 1, 1921, a 2671 in pt., nae as A. pusilla (Upsala). Potosi: Potosi, 4000 m. alt., March 1932, Carden 5, mixed with A. pusilla (G). Tarija: Calderillo, 3600 m. alt., March 1904, exes 3184a (BD, type; US, FM); southern Bolivia, 1903, Fiebrig 3576 (BD). Very evidently related to A. peruvianus, and perhaps only a luxuriant form of that species. From its relative it differs in having, not 1 or 2, but 4—7 flowers clustered in an exceedingly abbreviated subumbellate axillary raceme. The flowers are intermixed with broadly lanceolate bracts. These bracts permit the plant to be distinguished from those forms of A. peruvianus which produce several flowers on a given se of the caudex, for in the latter cases the flowers are separated, not by bracts, but by leaves and stipules. The leaves in A. Urbanianus are larger than in A. peruvianus, the lanceolate to paEine blade measuring 3-12 mm. and the leaf-rachis 1-5 cm. in lengt 402 JOURNAL OF THE ARNOLD ARBORETUM [voL. XxvlI 68. ers ae ar per operas Pub. Field Mus. Bot. 8:98 (1930), and l.c. 3[Fl. Peru 3:]: 1943).— Type from between Tarma and Morococho, oa, & for leis 1052. ERU. Near Tarma, Junin, 4200 m. alt., June 1922, Macbride & Featherstone 1052 (FM, type) The type has no fruit. The flowers are in a much shortened raceme and form a cluster among the leaves. The leaves are abundantly villous with loosely appressed silky hairs about 2 mm. long. The plant is evi- dently related to A. Urbanianus but differs in the more abundant and longer hairs of its indument. It may not be specifically distinct. 69. Astragalus peruvianus Vogel, Verhandl. K. Leop.-Carol. Akad. shige sain 19: suppl. 18 (1843).— Type collected near Pisacoma in southern Peru, Meyen Rance: High altitudes in southern Peru, Bolivia and northwestern Argentina. PERU. Ayacucho?: Santa Inez silver mines (between Ayacucho and Pisco), May 1910, Weberbauer 5450 (FM, BD). Moquegua: Carumas near Volcan Ticsani, 1925, Weberbauer 7327 (FM, BD). Tacna: near Pisacoma, 4500 m. alt., 1831, Mevyen (BD, type) OLIVIA. La Paz: La Cumbre, trail to the Yungas, ca. 4700 m. alt., Marc 1921, Asplund 6130 (US) and 3846 (Upsala) ; Chacaltaya, 4800 m. alt., 1908, Buckie 1813 (US, BD) ; near La Paz, 3000 m. alt., 1890, Bang 660 (G, NY). Pot Cerro de Potosi, 4800 m. alt., Feb. 1933, Cardenas 406 in pt. (US); Potosi, ieee. m. alt., ARGENTINA. Oe Cumbre de Calchaquies, 4200 m. alt., Spegazzini 2104 (LP); El Pelado, 1912, Rodeieues 442 (G). Typical A. peruvianus usually has the coarse elongate branches of the caudex tightly ensheathed with glabrous tough commonly stramineous stipules. The leaves, with abundant appressed lustrous hairs, are usually small having leaflets mostly 1 mm. or rarely 2 (—2.5) mm. long on rachises 5-10 (—15) mm. long. The solitary (or rarely geminate) flowers are borne among the stipules of the functioning leaves. The species is closely related only to A. Urbanianus and A. Dillinghami, both of which have larger leaves and abbreviated racemes of flowers. Some forms of A. peruvianus suggest A. deminutivus in gross habit but these forms may be distinguished from that species by their very different fruit, lustrous more appressed indument, glabrous stipules, etc. Our present plant has been most frequently confused with A. pusillus, but that plant has non-lustrous indument, broader emarginate leaflets, much broader papery stipules, and clustered flowers. Besides the specimens cited above I have in my notes thirteen other collections assigned to A. peruvianus. They represent collections identified before I had recognized A. Urbanianus as distinct from A. peruvianus, and some of them may possibly represent the former species. Those from Peru are, from near Arequipa, 2000-2500 m. alt., 1923, Guenther & Buchtien 1077, 1353, 1354, and 1355 (Hamb) ; Saracochl near Lagunillas, 4400 m. alt., 1925, Guenther & Buchtien 11646 (Hamb), and those from Bolivia are, from Rinconada, Yungas Pass, northeast of La Paz, 4100 m, alt., 1928, Troll 1891 (BD); Chuyuncayani, 1851, Weddell (Paris) ; Chacaltaya, northeast of La Paz, 1928, Troll 1935 (BD); La Lancha near 1947] JOHNSTON, ASTRAGALUS 403 La Paz, 5000 m. alt., Mandon 718 (K, Del); Huari pampa, between La Paz and Palca, 4200-4400 m. alt., 1910, Pflanz 483 in pt. (BD): Quem- sachata near Tiahuanaco, 4300 m. alt. 1877, Stibel 64 (BD); Copaca- bana, 4050 m. alt. 1903, Hill 122 (K); and Quebrada Honda, Potosi, d’Orbigny 1310 (Paris). 70. Astragalus Werdermanni, sp. nov. Planta dense caespitosa; ramulis caudicis congestis stricte ramosis, vetustioribus stipulis imbricatis membranaceis pallidis dense vestitis: foliis cinereis sparse breviterque strigosis, rhachi 13-18 mm. longa supra medium foliolata, foliolis 5—7-jugatis orbiculato-obcordatis evidenter pli- catis 2—2.5 mm. longis 2-3 mm. latis paulo supra medium latioribus apice conspicue emarginatis basi ioral sparse strigosis pilis a costa oblique abeuntibus; stipulis 3-4 mm. longis strictis sparse strigosis; pedunculis 1—2-floris, fructiferis 3- Tie mm. longis; pedicellis 0.5-1 mm. longis; calyci- bus sparse strigosis ca. 2. mm. longis, tubo membranaceo cupulato, lobis triangularibus ca. 0.5 mm. longis viridibus; corolla ignota; leguminibus subbilocularibus sparse Seen 5-7 mm. longis 2—3 mm. crassis 3—4 mm. altis, margine superiori a latere viso recto, margine inferiore a latere viso valde convexo, sutura inferiore valde inflexa in loculo leguminis septum membranaceum 2-3 mm. altum gerente; seminibus pluribus CHILE: Chislluma, Cord. Volcan Tacora, Arica, depressed, a few plants between rocks, 4500 m. alt., A 1926, Werdermann 1439 (type, Berlin). A well marked species probably most closely related to A. casapaltensis Ball, Jour. Linn. Soc. London 22: 36 (1885), a species known only from high altitudes in central Peru (cordilleras of Lima and Junin), which differs from our present plant in its fewer, more conspicuously emarginate leaflets that bear more persistent and more appressed shorter hairs. The proposed Chilean species is readily recognized by its very broadly obcor- date, strongly folded leaflets which are strigose with numerous scattered closely appressed hairs oriented obliquely to the midrib. The fruit has a very well developed, almost complete, membranous false septum. vs Astragalus minutissimus Weddell, Chlor, Andina 2: 257 (1861).— Type from “dept. Cuzco,” 29, PERU. Cuzco: “dept. Cuzco,” Gay 629 (Paris, type). Puno: Mullapata near Moho, 3125 m. alt., 1919, Shepard 110 (G, NY, US); between Huaicho and Moho, 1903, Hill 126 (K). IVIA: La Paz: Isla del Sol, Lake Titicaca, 3840 m. alt., March 1910, Buchtien (US). Cochabamba: Challa Pass, 4500 m. alt., March 17, 1892, Kuntze (NY). This species is very closely related to 4. alpamarcae Gray, U.S. Explor. Exped. 1: 417 (1854), which is known only from a few collections in the cordillera of central Peru (Cerro de Pasco, Junin, Mathews 924; Alpa- marca, Lima, U. S. Explor. Exped., type; and Casapalca, Lima, Macbride & Featherstone 837). The plants of the northern, older species have 7—9 pairs of leaflets. The leaves in our present southern form bear only 3-5 pairs of leaflets. Otherwise the plants seem very similar. The differ- ences between A. minutissimus and A. alpamarcae are probably not of 404 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII specific value. Certainly the two are very much alike in general appear- ance. In gross aspect A. minutissimus is frequently very similar to A. deminutivus. The fruits of these species, however, are very different, and furthermore the leaflets of A. minutissimus show a very much stronger tendency towards glabrescence than do those of A. deminutivus. 72. serie ere minimus Vogel, Verhandl. K. Leop.-Carol. Akad. Naturfors. 19: suppl. 18 (1843).— Type from ‘Alto de Toledo, east of Arequipa, Meyen. Astragalus ‘ola rd Mem. Torrey Bot. Club 3:19 (1893).— Type from PERU. wet: ee de Toledo, 4600 m. alt., Meyen (BD, type). OLIVIA. La Paz: near La Paz, 3000 m. alt., Bang 660 (NY few . colliculus) ; Charafia, 4000 m. alt., March 1, 1921, ‘tied 6145 (US) and ere (Upsala). CHILE. Arica: Ancara, Cord. Volcan Tacora, 4300 m. alt., April 192¢ Werdermann 1136 (G, BD). ARGENTINA. Jujuy: Moreno, 3500 m. alt., Dec. 14, 1901, Fries 916 (Stock). Los Andes: San Antonio de los Cobres, Feb. 21, 1927, Castellanos 27/813 (G). Tucuman: between La Puerta and Los Cardones, Cumbre Calchaquies, 4400 m alt., Jan. 30, 1933, Burkart 5386 (Burkart). This species is closely related to A. minutissimus and A. alpamarcae, but has a more prostrate habit and an indument of shorter much more appressed hairs. In gross habit it much resembles A. Bustillosii which differs in its rhizomes, larger flowers, and much larger compressed fruit. As in A. Bustillosii, our present species frequently - the fruit elevated on a short but definite 1—2-flowered peduncle 2-6 mm. long. The type specimen of A. colliculus, Bang 660 at New York, represents two forms of the present species. Bang’s collection was evidently a great mixture, for in addition to the two forms of A. minimus at New York, material of A. peruvianus and of a depauperate form of A. sacle has_ been eer under his no. 660 to other herbaria. Rusby description of A. colliculus seems to include details derived from i material of all the oe mentioned. ‘To make matters worse, Mandon 712 at New York, which was cited by Rusby, is also a mixture consisting of plants of A. pusillus and A. minutissimus. If Rusby’s species is not to be discarded as a confused mixture, it must be taken as a synonym of A. minimus since the material of Bang 660 retained at New York by Rusby all belongs to that species. Cc ’ 73. Astragalus pulviniformis Johnston, Physis 9: 309 (1929). ns collected at Vega Cadillo in northwestern San Juan, Argentina, Johnston ARGENTINA. Vega Piri Rio Tagua, San Juan, 3000 m. ae nen Johnston 6128 (G, type; K, Paris, Stock). A very well marked species which has its closest relations in A. Reichei and A. deminutivus. Its dense pulvinate habit, strongly marcescent dusky leaves, scattered hyaline stipules, slender elongating stems, and larger fruit readily distinguish it from the species mentioned. 74. ach Reichei Spegazzini, Anal. Mus. Nac. Buenos Aires ser. 2, 4: 264 (1902). mpacta Phil. Cat. Pl, Itin. Tarapaca 14 (189 1). —Type from Queb. Ma ia prov. Antofagasta, 3200 m. alt., Philippi. 1947] JOHNSTON, ASTRAGALUS 405 Astragalus compactus (Phil.) Reiche, ee Univ. Chile 97: 539 (1897), and FI. Chile 2: 82 (1898).— Not Lam. ( BOLIVIA. Rosario, prov. pace fen La Paz, 3750 m. alt., 1920, Shepard 27 (G); above Animas Mine, west of Chocaya, dept. Potosi, 4250 m. alt., West 6069 aa 6070 (G CHILE. Pefia Blanca, Cord. Arr. Covacagra, prov. Tarapaca, 4000 m. alt., Wer dermann 1117 (G, K, BD); Atacama oe oo Machuca, prov. Aiiehibasta, 3200 m. alt.], ex Philippi as P. compacta (K, This species has close relationship a with the more easterly and southerly ranging A. flavocreatus. ‘The short simple or strictly branched stems are erect or ascending from the crown of the deep root. They are covered with the leaves of seasons past and crowded to form a dense mass commonly intermixed with some dirt. The stipules of A. Reichei are pale and are not so firm nor so closely appressed to the stem as are those of A. flavocreatus. The herbage of A. Reichei is slightly more grayish than in its relative. 75. Astragalus flavocreatus, sp. nov. Planta perennis e radice profunda recta oriens; caudice saepe laxe ramoso plus m minusve horizontali, ramulis funiculiformis gracilibus 1—1.5 mm. crassis stipulis stramineis strictis imbricatis vestitis vestigiis foliorum haud gerentibus; foliis apice paca stipulosorum caudicis aggregatis pilis haud abundantibus (0.2—0.6 mm. longis rectis adpressis) vestitis vetuste deciduis; rhachi Te mm. longa tertiam partem inferiorem efoliolata ; foliolis 0.5-1.5 mm. longis obovatis vel ellipticis saepe sub- perinconspicuis inter stipulis occultis 1-2 mm. longis; floribus solitariis; calyce 2-3 mm. longo haud vel vix pedicellato imam ad basim per bracteam 1-2 mm. longam strictam suffulto, tubo cupulato pilis sparsis adpressis vestito, lobis angustis ca. 0.5 mm _ longis reviter villosis; corolla violacea, vexillo 6-7 mm. longo, lamina suborbiculari ca. 5 mm. diametro tng emarginata; carina obtusa quam vexillo ca. 3 mm. breviore e; alis 4-5 m longis supra medium ut videtur patentibus; leguminibus subglobosis ca. 2 mm. longis villosis unilocularibus biseminatis, sutura inferiore inflexa. BOLIVIA. Cerro de Potosi, 4800 m. alt., 1933, Cardenas 406 (G) and 406 in pt. (US); Potosi, 4000 m. alt., 1932, Cardenas 166 (G ARGENTINA. Salta: Cerro de Cachi, 4000 m. we, 1927, Venturi 6704 (US). Catamarca: Sierra Anconquija, 4200-4600 m. alt., Venturi 6626 and 6627 (US). Tu aaapen Cumbre Calchaquies, dept. Tafi, between La Puerta and Los Cardones, 4400 m. ., Jan. 30, rae ame 5385 (type, Gray Herb.); Est. Santa Rosa, dept. a 3600 m. alt., , Venturi 6625 (US). La Rioja: Alto Blanco, Biorta Famatina, 1928, Lana pi (G) This plant is evidently related to A. Reichei, but its habit of growth, as well as various intangibles of pubescence, stipules, etc., permit it to be distinguished from that species. The caudex of A. Reichii is a compacted mass of strictly ascending stems and persistent old leaves. It is very different from the more spreading superficial caudex of leafless rope-like stems developed by A. flavocreatus. The tightly ensheathing straw- 406 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII colored stipules which clothe its more elongate and loosely branched stems, and the general lack of persistent old leaves, make A. flavocreatus a more neat and attractive plant and permit it to be quickly distinguished from other caespitose species. The flowers of A. flavocreatus have a bract tightly appressed about one side of the calyx. This is the bract borne at the apex of the reduced peduncle. From its axil the solitary flower is produced. The calyx is sessile. The apparent pedicel, accordingly, is really the reduced peduncle. This pedicel-like peduncle in A. flavocreatus is usually twice the length of that found in A. Reichei. It is usually well hidden in the stipular sheathes. 76. Astragalus deminutivus, sp. nov. Planta multiceps cinerea dense caespitosa 2-15 cm. diametro; ramulis caudicis cum stipulis et reliquiis conspicuis congestis foliorum abundanter vestitis; foliis villosulis numerosis deminutivis, rhachi 1-2 cm. r- tiam partem inferiorem efoliolata; foliolis 5—7-jugatis orbicularibus vel anguste ovatis saepe ca. 1 mm, longis margine laxe involutis apice rotundis; stipulis gang — chartaceis albidis villoso-strigosis valde imbricatis ongis; floris axillaribus solitariis subsessilibus; calycibus 3-5 mm. ‘anak villosis imam ad basim bracteam lanceolatam gerentibus, tubo chartaceo profundo, lobis linearibus herbaceis 1-1.5 mm. longis; vexillo corollae caeruleo 7-8 mm. longo quam carina 2—3 mm. longiori, lamina 3—4 mm. lata apice emarginata; alis caeruleis quam carina pallida obtusa 1—1.5 mm. longioribus supra medium ut videtur patentibus; ovario strigoso ; leguminibus oblique subglobosis ca. 2 mm. longis et 1. crassis subtus introflexis; seminibus 1-2 saepe 2 mm. Bane. BOLIVIA. La Paz: General Campero, 3900-4200 m. alt., March 5, 1921, Asplund 6140 (US) and 2780 (Upsala). Potosi: Cerro de Potosi, 4800 m. alt., 1933, Cardenas 406 in pt. (US); Uyuni, 3700 m. alt., March 24, 1921, Asplund 6141 (US) and 3138 (Upsala); Rio Mulalo, prov. Poreo, 1 934, Hammarlund 464 (Stock). Tarija: Puna Patanca, 3700 m. alt., 1904, Fiebrig 2608 (type, Gray Herb; K, BD, Paris). ARGENTINA. Los Andes: Susquis, March 7, 1927, Castellanos 27/818 (G). Jujuy: Laguna Tres. Cruces, 1901, Claren 11657 (Stock). The specimens I have cited above are certainly conspecific and merit description as a new species. They include the plant —— called A. Reichei by Fries, Nova Acta R. Soc. Sci. Upsala ser. IV, 14: 134 (1905). Hosseus, Bol. Acad. Nac. Cordoba 26: 154, fig. 28 (1922), given an illustration of Claren 11657, which is the basis of Fries’ note, and properly expresses a doubt as to the correctness of Fries’ identification. Closely related to A. deminutivus and perhaps only a robust form of this species are two collections made in the dept. of Potosi by Asplund. These collections are, Jyuni, 3700 m. alt., March 24, 1921, Asplund 6139 (US) and 3137 (Upsala), and Atocha, 3700 m. alt., March 20, 1921, Asplund 6138 (US) and 3025 (Upsala). They differ from typical A. deminutivus in having the pods more elongate and larger (oblong and 3-5 mm. long), the seeds smaller (1.3-1.5 mm. broad) and more numerous (3-6), the leaves larger (rhachis ca. 3 cm.), and the leaflets larger (1-2 mm.) and somewhat elongate. On Cerro de Chani, Jujuy, Argentina, 1947] JOHNSTON, ASTRAGALUS 407 Venturi (no. 8527, US) has obtained still another form which is tenta- tively placed under A. deminutivus. This form has glabrous stipules. 77. Astragalus Dielsii Macbride, Field Mus. Pub. Bot. 11: 24 (1931), and Lc. 13 [FI. eru 3:]:396 (1943).—Type from Yanashallos, west of Huallanca, dept. Ancash, Macbride & Featherstone 2479 PERU. Yanashallas, ean 4800 m. alt., Oct. 2, 1922, Macbride & Featherstone , type; BD); mountains meat at Oroya, 4300 m. alt., Nov. 23, 1902, Weberbauer 1704 (FM, BD), This is an outlying northern relative of A. deminutivus. Its stipules are conspicuously and permanently strigose, and though white at first soon become very dusky in color. UNPLACED SPECIES saa aig Se acoing: ener eigen T2221 Sait a Pub. Field Mus. Bot. 13 [Fl. Peru 3:]: 394 (1943).— Peru, Ruiz & P Bases upon a flowering specimen at Madrid said to — been collected somewhere in Peru by Ruiz & Pavon. Among the notable features of the type are the numerous leaflets (9-12 pairs), the black appressed hairs on its calyx, and the rather elongate stems and accordingly laxly caespitose habit. It appears to be a relative of A. minimus Vogel and is, perhaps, even a form of that species. Lacking fruit, however, its identity remains uncertain. Phaca diminutiva shail Cat. Pl. Itin. Tarapaca 14 (1891).—‘“Inter Aguas calientes et Socai . lecta”; not A. deminutiva Johnston A very small Eee tee plant with pallid appressed indument. The single specimen was collected with the type of Phaca cryptantha Phil. Reiche, Fl. Chile 2: 80 (1898), places both P. cryptantha and P. diminu- tiva as synonyms of A. clandestinus (Phil.) Hieron. which is called A cryptobotrys in this paper. This is perhaps correct. In any case the name may be discarded since the trivial epithet is preoccupied under Astragalus. Prof. Espinosa could not locate the type in the Museum at Santiago. see Gilliesi Philippi, Anal, Univ. Chile 31: 393 (1862), and a 33: 48 (1864).— “Portezuelo del Portillo, lado de Mendoza,” 1860-61, This species was described without mature fruit. ae a the description it might be a form of A. Arnottianus (Gillies) Reiche, or perhaps even a form of A. carinatus H. & A. I asked Prof. Espinosa for the privilege of examining a fragment of this type and was loaned material labeled ‘Astragalus Gilliesii, Portillo, lado de Mendoza, 1861-62, W. Diaz.” This consisted of foliage and fruit of A. Berteri and flowers of A. Cruck- shanksti. The material is not covered by the original description. Astragalus gracilis Romero, Bol. Direc. Nac. Estad. y Estud. Geogr., La Paz, sec. Epoc. III, Nos. 31-33, pg. 45 (1920).— Not Nutt. (1817), nor Stev. (1842). This species was described as follows in Romero’s paper on the flora of the vicinity of La Paz, Bolivia: ‘Astragalus gracilis, sp. n. B. D. R. 408 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XxvuI Esta plantita erecta, delgada, con estipulas aladas, tiene foliolos linearos, semi-plegados, agudos y 8 yugados; sus espigas llevan pocas florecillas, la corola es blanca azuleda de 1 cm. de largo. Especie que la hemos recogido en los cerros de esta ciudad |La Paz].” This description, for the transcript of which I am indebted to Prof. Cardenas, is not detailed enough for the recognition of the species. The name is invalid in any case since it has two earlier homonyms. Astragalus infirmus Steudel, Nomencl. ed. 2, 1: 161 (1840), nomen. This name was applied to Bertero 820 from Chile. Among Steudel’s specimens at Paris I could find no Astragalus bearing this number. The name, possibly, may refer to an Adesmia. Astragalus lanuginosus Clo s in Gay, Fl. Chile 2: 121 (1846); Reiche, Anal. Univ. Chile 97: 566 (1897), and Fl. Chile 2: 109 (1898). Not Kar. & Kir. (1841). Tragacantha Gayana Kuntze, Rev. Gen. 2: 941 (1891).— Based upon A. lanu- ginosus Clos. ne er tne aap genie Anal. Mus. Nac. Buenos Aires ser. 4, 4: 265 2 ased u A. lanuginosus Clos. an This ae was ee upon specimens lacking both flowers and fruit. It is represented at Paris by two sheets, evidently parts of a single collec- tion. The label written by Gay reads, “492 Pr. de Coquimbo, Astragalus ? lanuginosus Clos, Fl. Chil. Cl. Gay, nait sur les collines exposé et sablonneuses des basses cordilleres de hurtado—a la Coipa, haut 1639 m., passé fleur au janvier, tres rare.” If this plant is really a species of Astragalus it is one entirely unknown to me. However, I incline towards the opinion that it is an Adesmia or some other genus, rather than an Astragalus. —— eccinee (Kuntze) Hauman, Anal. Soc. Cien. Argentina 86: 281 (1918). Phac moena Philippi, Linnaea 28: 620 (1857).—In cordillera de la Campania, oe nes tiago, Nov. 1833, Germain. Astragalus amoenus (Phil.) Reiche, Anal. Univ. Chile 97: 569 (1897), and FI. Chile 2: 112 (1898).— Not Fenzl (1842). se aie Philippiana Kuntze, Rev. Gen. 2: 941 (1891).— Based upon P. amoena Phil. Privo santiagensis — Anal. Mus. Nac. Buenos Aires ser. 2, 4: 265 2) ased upon P. oena Phil. A spreading silky-strigose eceuai with united stipules. The type has flowers but no fruit. ae l.c., has suggested that it may be a form of A. Germaini. This may be sorneck though the description might also apply to a form of A. Cruckshanksii ee pictus Steudel, Nomencl. ed. 2, 1: 163 (1840).— Based on A. bicolor Des Aural bicolor Desfontaines, Cat. Hort. Paris, ed. 3, additamenta 473 (1832-33), , from “Chili.’ —Not L am, (1783). No i. has ever been associated with the two names above cited. Desfontaines treats the species as follows, “bicolor H. P. Chili, Temp. ¥ ” The name is a nomen nudum and its application wholly obscure. 1947] JOHNSTON, ASTRAGALUS 409 Astragalus quindecimjugas (Phil.) Reiche, Anal. Univ. Chile 97: 540 (1897), and Fl. Chile 2: 83 (1898). Phaca quindecimjuga Philippi, Anal. Univ. Chile 21: 443 (1862), and Linnaea 33:41 (1864).—“Cord. de Dofia Ana, prov. Coquimbo, ca. 4000 ft. alt. Volckmann.” Described as a plant with connate stipules, slender sparsely silky ascend- ing stems, glabrescent leaves, oblong-linear 15-jugate (8 mm. long and 4 mm. wide) leaflets, and elongate peduncles. The corolla is said to have the wings very evidently shorter than the keel. Concerning the fruit Philippi says, “El unico fruto que tiene el ejemplar es todavia verde i casi globoso.” Reiche describes the fruit as, ‘“‘Legumbre (en estado maduro desconocida) casi globosa, pelada con varias semillas.” ‘Fragments of the type loaned me by Prof. Espinosa show the leaves to be folded, glabrous except for a few appressed hairs along the midrib and about the margin near the apex. The ovary is glabrous and the flower in all details similar to that of A. palenae. I believe that A. guindecimjugus is based upon a col- lection from southern Chile and falsely labeled as from the andes of Coquimbo. It probably represents A. palenae var. grandiflorus Speg. Astragalus sphaerocarpus Desfontaines, Cat. Hort. Paris ed. 3, additamenta 473 (1832-33), nomen. This is a garden name which has never been associated with a descrip- tion. Desfontaines published merely the following concerning it: “sphaerocarpus H. P. Chili. Temp. %.” ARNOLD ARBORETUM, Harvarp UNIVERSITY. 410 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII SECOND SUPPLEMENT TO THE SPONTANEOUS FLORA OF THE ARNOLD ARBORETUM ERNEsT J. PALMER THE FIRST CATALOGUE Of the Spontaneous Flora of the Arnold Arbore- tum, published in 1930!', contained the names of 610 flowering plants and ferns, and a supplement in 1935” reported the discovery of 173 additions. The present paper adds 148 names, three of which are substitutes for ones published in previous lists. This brings the number of species, varieties, forms and hybrids that have so far been found growing without cultivation to the rather surprising total of about 930. In the more than ten years that have elapsed since the last publication a diligent search has been carried on but, naturally, with somewhat diminishing results. Approximately sixty percent of the plants listed in this supplement are native to the Boston area and forty percent are in- troduced. But in some cases it is difficult to determine whether plants of native species are really indigenous or have come from seeds of culti- vated plants. This is particularly the case with trees and shrubs, since such a complete collection of them has been established in the Arboretum, and the fruits and seeds are readily distributed by birds and other agencies. In trying to distinguish between native and introduced plants in the list, those native in the area have usually been counted in the first division, although the probability of their having come from cultivated plants is sometimes indicated in the notes Among native plants of considerable interest recorded for the first time in this list may be mentioned Deschampsia flexuosa (L.) Trin., Alopecurus geniculatus L., Spartina pectinata Bosc ex Link, Scirpus planifolius Muhl., Carex virescens Muhl., C. aestivalis M. A. Curtis, C. laxiculmis Schwein., Anychia canadensis (L.) B.S.P., Corydalis sempervirens (L.) Pers., Mi- mulus ringens L., Campanula aparinoides Pursh, and Prenanthes altissima L. Anacharis occidentalis (Pursh) Victorin and Lemna minor L. were apparently recent introductions and both seem to have disappeared, although it is not unlikely bi they grew in the Arboretum area before and that they may reappea The Arboretum affords cad opportunities for the introduction of exotic plants because of the varied and almost world-wide sources from which material is obtained. Seeds of weeds and other plants may be brought in with packing or soil about living material, or in seed, fertilizer or loam, and sometimes in rubbish that is dumped in certain places. The 1 Jour. Arnold Arb. 11: 64-119. 1930. - Ibid. 16: 81-98. 1935. 1947] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 411 common agencies of birds and winds are also active. Many plants appear that are obviously only transient waifs and the source from which they come can in some cases be determined with reasonable certainty. Some waifs have been included in the lists when they seemed to be of sufficient interest, but a number of others of which specimens have been made for a record have been excluded. It is difficult to know where to draw a line in such cases; but in general plants that are almost certainly transient and that are of known origin have been excluded. Most of these appear on rubbish dumps, and they include such garden vegetables as maize, potato, pumpkin, beans, and spinach. Among other less common plants are species known to have been under investigation or to have been used experimentally by members of the staff of the Arnold Arboretum or of the Bussey Institution. A few others have been collected that have per- sisted from old gardens but which show no tendency to spread. Specimens of most of these may be found in the special herbarium of SPONTANEOUS Prants. All of the collections reported in this list and the numbers shown are those of the author, except as otherwise specified. It is the intention of this paper, as was the case with the previous ones, to employ plant names approved by the latest usage. But so many changes have been made in recent publications in the interpretation of names and in the acceptance of earlier names for many species that the nomenclature of the earlier papers is in need of revision. Perhaps some- day it will be possible to bring together and to revise all of the lists. Thanks are due to Dr. L. H. Bailey for identifying most of the species of Rubus and to Dr. S. F. Blake and others for verifying or supplying determinations of a few plants in other difficult groups. ENUMERATION OF THE ADDITIONAL PLANTS COLLECTED! Anacharis occidentalis (Pursh) Victorin. Water-weed. Shallow water and muddy margins of pond, South Street tract. No. 44701. Abundant and in flower in 1938, but not seen since. *Bromus arvensis L. Field Chess. Dry open banks north of Adminis- tration Building. No. 46002. *Festuca elatior L. var. arundinacea (Schreb.) Wimm. Gravelly bed of Bussey Brook, South Street tract. No. 46689. Festuca rubra L. prolifera (Piper) Fernald. Open grassy slopes, Lilac group. No. 4 Poa nemoralis L. an grassy ground near Dawson House. No. 42857. *Poa trivialis L. Waste ground, old quarry along Bussey Street. No. 43161 Eragrostis Frankii (Fisch., Mey. & Lall.) Steud. Meadow near Celtis group. No. 39666. *Fragrostis pilosa (L.) Beauv. Waste and cultivated ground. No. 42973 1 [Introduced plants are marked by an asterisk (*). 412 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII *Eleusine indica Gaertn. Goose Grass. Waste ground, old quarry along Bussey Street and along path near Cercidiphyllum group. Nos. 40221, 46858 Glyceria pallida (Torr.) Trin, var. Fernaldii Hitchc. Border of pond near Forest Hills gate. No. 28155. Deschampsia flexuosa (L.) Trin. Hair Grass. On conglomerate outcrops and in dry gravelly soil, top of Hemlock Hill, Central Woods, and in Malus group, base of Peters Hill. Nos. 42842, 46169. Very local. Agrostis stolonifera L. var. compacta Hartm. Creeping Bent Grass. Muddy margins of brook near Arborway wall southeast of Administration Building. Nos. 45248, 46160, 46674. *Agrostis alba var. vulgaris ives Thurb. Wet margins of small pond above Fraxinus group. No Alopecurus geniculatus L. a ater Saath In shallow water along mar- gin of brook near Arborway wall. Nos. 45427, 46677. *Alopecurus carolinianus Walt. Southern Foxtail. Moist cultivated ground, south of Hemlock Hill. No. 45960 Spartina pectinata Bosc ex Link. Prairie Cordgrass. Moist open ground along brook near Jamaica Plain gate. No. 44694. A large clump, seldom flowering. Panicum agrostoides Spreng. Wet open ground, low meadow near Ad- ministration Building. No. 46535. Uncommon. Panicum clandestinum L. Bank along road near Robinia group. No. 46694 Panicum dichotomiflorum Michx. var. geniculatum (Wood) Fernald. Waste and cultivated ground. Nos. 38229, 43097. *Setaria Faberii aia Rich waste ground, old quarry along Bussey Street. No. 470 Andropogon an Michx. (A. scoparius var. villossisimus Kearney ex Scribn. & Ball). Dry open ground, Peters Hill. No. 44689. Un- common. Cyperus strigosus L. var. capitatus Boeck]. South Street nursery. No. 43408. Eleocharis ovata (Roth.) R. & S.. Muddy margin of pond. No. 35152. Eleocharis capitata (L.) R. Br. var. borealis Svenson. Springy ground, south side of Peters Hill and along ditch in low meadow. Nos. 45974, 46007. Scirpus planifolius Muhl. Dry open woods, top of Hemlock Hill. Nos. 45966a, 46010. Rare and local. Scirpus expansus Fernald. rev drainage ditch, low meadow near Ad- ministration Building. No. 43408. Scirpus atrovirens Muhl. var. georgianus (Harper) Fernald. Margins of brook near Arborway wall. No. 45218 1947] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 413 Carex scoparius Schkuhr var. Mga Tuckerm. Wet springy ground, southeast side of Peters Hill. No. 40293. Carex tribuloides Wahlenb. var. reducta ae Wet open ground, low meadow near Administration Building. No. 46006. Carex cephalantha (Bailey) Bicknell. Wet ground about spring, south- east side of Peters Hill. Nos. 27976, 37723, 42599. Reported in the original list as C. echinata Murr. Carex canescens L. Wet ground along brook north of Administration Building. No. 40038. Carex virescens Muhl. Rocky slopes, south side of Hemlock Hill. No. 42833. Carex aestivalis M. A. Curtis. Rocky slopes, south side of Hemlock Hill. No. 42832. Rare and local. Carex laxiculmis Schwein. Open woods, Oak group. No. 42838. Un- common. Carex hystericina Muhl. Along ditch, low meadow near Administration Building. No. 46645. Symplocarpus foetidus (L.) Nutt. Skunk Cabbage. Wet ground in a number of places throughout the Arboretum. This interesting plant, one of the first to bloom in spring, was referred to in former papers, but through oversight it was omitted from the list. Lemna minor L. Small Duckweed. Stagnant pools along brook near Arborway wall. No. 46381. Transient. Juncus dichotomus Ell. var. platyphyllus Wiegand. Open ground, low meadow near Administration Building, and in cultivated ground, South Street tract. Nos. 45666, 46557. Juncus tenuis Willd. var. anthelatus Wiegand. Open ground, top of Hem- lock Hill. No. 46692. *Muscari botryoides (L.),Mill. Grape Hyacinth. Grassy bank, Populus group, South Street tract. No. 44721. Well established. *Tris Pseudacorus L. var. pallida Hort. ex Bailey. Pale Yellow Iris. This form with very pale flowers is found in the low meadow near the Adminis- tration Building. No. 45894. Salix discolor Muhl. var. eriocephala (Michx.) Anders. Wet open ground, low meadow near Administration Building. No. 46594. Betula papyrifera Marsh. Paper Birch. Open a top of Peters Hill. No. 37755 Celtis occidentalis L. Hackberry. Top of Hemlock Hill. No. 46369. Probably introduced here. *Morus alba L. White Mulberry. Waste ground, South Street tract. No. 43169. *Rumex triangulivalvis (Danser) Rech. f. (R. mexicanus of auth., not Meisn.). Aijong brook and about pond, South Street tract. No. 43173. 414 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Polygonum aviculare L. var. angustissimum Meisn. Waste open ground. No. 44694a. Polygonum aviculare var. arenastrum (Bor.) Rouy. Waste open ground. No. 44693a. *Polygonum caespitosum Blume var. longisetum (DeBruyn) Steward. A weed in waste ground. No. 46382, S. F. Blake & E. J. Palmer. *Polygonum Bistorta L. Bistorte, Snakeweed. A large colony near the bridle path in Conifer group. No. 46620. *Chenopodium Botrys L. Jerusalem Oak. Waste ground and rubbish dumps, about foundation of old Bussey mansion. Nos. 46649, 47073. *Chenopodium glaucum L. Oak-leaved Goosefoot. Along drainage ditch, low meadow near Administration Building. No. 42970. *Chenopodium paganum Reichenb. Lamb’s Quarters. Waste ground, old quarry along Bussey Street. No. 46086 Chenopodium hybridum L. Maple-leaved Goosefoot. Waste ground about foundation of old Bussey mansion. No. 47072. Atriplex patula L. Orach. Waste and cultivated ground. Nos. 43078, 47939. Not common but apparently spreading. *Axyris amaranthoides L. Rocky waste ground, old quarry along Bussey Street. No. 46980. Probably only a transient weed. *Amaranthus hybridus L. Amaranth, Pigweed. Waste ground, South Street tract. No. 43083. *Amaranthus hybridus {. hypochondriacus (L.) Robinson. Prince’s Feather. A weed in waste ground, old quarry along Bussey Street. No. 45249, Anychia canadensis (L.) B. S. P. Forked Chickweed. Open wooded hillside, North Woods, near Robinia group. No. 45214. Rare and local. *Arenaria serpyllifolia L. Thyme-leaved Sandwort. A weed in cultivated ground, Nos. 43081, 44783 *Agrostemma Githago L. Corn Cockle. Waste ground, old quarry along Bussey Street. No. 43007. Probably only a transient waif *Silene Armeria L. Sweet William Catchfly. Grassy ee near Ad- ministration Building. No. 42688. Probably a transient waif. Clematis virginiana L. Virgin’s Bower. In tangle of shrubs along Meadow Road. No. 44912. Probably escaped from cultivation. Menispermum canadense L. Moonseed. Among weeds and shrubs, Jug- lans group. No. 36535a. Probably escaped from cultivation. *Eschscholtzia californica Cham. California Poppy. On open bank near Administration Building. No. 44884. Escaped from cultivation. Corydalis sempervirens (L.) Pers. Pink-flowered Corydalis. At foot of conglomerate ledge, top of Hemlock Hill. No. 46183. Rare and local. *Brassica campestris L. Rutabaga. A weed in waste ground near Ad- ministration Building. No. 44313. 1947] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 415 *Brassica juncea (L.) Cosson. Indian Mustard. Waste ground, old quarry along Bussey Street. No. 47010. *Ribes rubrum L. Red Currant. Moist ground, thickets near Pterocarya group. No. 44767. *Spiraea japonica L. f. Japanese Spiraea. Along margin of Bussey Brook near stone bridge. Nos. 46274, 47746. *Malus micromaius Mak. Rocky ground, old quarry along Bussey Street. No. 42697. Sorbus americana Marsh. American Mountain Ash. Rocky open ground, top of Hemlock Hill. No. 46368. Probably escaped Aronia floribunda Spach. Chokeberry. Open rocky ground, top of Hem- lock Hill. No. 46866. Introduced here. Aronia prunifolia (Marsh.) Rehder. Wet ground about spring, south side of Peters Hill. No. 46977. Perhaps native. *Potentilla anelin L. Dry open ground, South Street tract. No. 42820. *Rosa suffulta Greene. Among weeds and shrubs, Oak group. No. 39664. Rubus idaeus L. var. canadensis Richards. Red Raspberry. Low meadow opposite Aesculus group. No. 45253a. Rubus alius Bailey. Open ground, slopes of Bussey Hill. No. 46612. Rubus semisetosus Blanchard. Open woods near top of Peters Hill. No. 46693. Rubus rhodinsulanus Bailey. Open woods and thickets, Peters Hill. Nos. 43422, 43529. Rubus novanglicus Bailey. Thickets and rocky open ground. Nos. 44356, 44361. Rubus permixtus Blanchard. Rocky ground, top of Hemlock Hill. No. 46016. Rubus frondosus Bigel. ‘Overlook,’ C. E. Faxon. 1919, Rubus multiformis Blanchard. C. E. Faxon. 1913. Rubus philadelphicus Blanchard. Rocky ground, Central Woods. No. 46653 Rubus Rosa Bailey. LLow meadow opposite Aesculus group. Nos. 45248, 46613. *Rubus laciniatus Willd. Persistent from cultivation, near site of old Bussey mansion. No. 46662. Prunus pennsylvanica L. f. Pin Cherry. Rocky ground, top of Hemlock Hill. No. 45962. *Medicago sativa L. Alfalfa. Waste ground, South Street tract. No. 431066. *Astragalus glaucophyllus Bunge. Grassv bank near Bussey greenhouse. No. 44324 416 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XXVIII Desmodium Dillenit Darl. Tick Trefoil, Stick-tight. Border of pond near Forest Hills gate Lespedeza capitata Michx. Dry open ground, edge of woods, south side of Peters Hill. No. 43526. Lespedeza capitata X hirta. Plants that appear to be this hybrid were found growing with the supposed parent species on the south side of Peters Hill. No. 43526a. *Oxalis corniculata L. (O. repens of coh ). On base of rotting stump, near pond and Robinia group. No. 4712 Zanthoxylum americanum Mill. Prickly i, Thickets, rocky ground on top of Hemlock Hill. No. 46419. Probably introduced from culti- vated plants. *Euphorbia maculata L. (E. Preslii Guss.).. Waste ground about foun- dation of old Bussey mansion. No. 46957. A recent introduction, but apparently spreading. The plant reported under this name in the original list is now known as Euphorbia supina Raf. *Kuphorbia virgata Waldst. & Kit. Under trees, Conifer group. Leon Croizat. 1938 Vitis riparia Michx. (V. vulpina of auth., not L.). River-bank Grape. Slopes of Peters Hill, Crataegus group, and in Conifer group. Climbing on trees and shrubs. Nos. 45669, 46562. No doubt introduced from cultivated plants. Vitis aestivalis Michx. Summer Grape. Thickets, bank near Robinia group. No. 46695. Probably introduced. *Viola blanda Willd. Waste ground, old quarry along Bussey Street. No. 44726. x Viola abundans House (V. fimbriatula X< sagittata). Moist meadow, Maple group. No. 42810. x Viola aberrans Greene (V. fimbriatula papilionacea). Nos. 36407, 42682. Viola fimbriatula %_ scabriuscula. Grassy open ground near base of Peters Hill. No. 40023 Viola lanceolata sagittata. Grassy open ground, Crataegus group, near base of Peters Hill. No. 42816a *Elaeagnus umbellata Thunb, Oleaster. Thickets and waste ground. No. 46972. Frequently escaped and becoming abundant. Oenothera laciniata Hill. Waste ground, fertile soil, near Dawson House. No. 42856 Rhododendron viscosum (L.) Torr. White Swamp Azalea. Low ground along Bussey Brook, Conifer group. No. 39603. Plants coming up after persistent mowing, and probably native. *Fraxinus pennsylvanica Marsh. var. lanceolata (Borkh.) Sarg. Green Ash. On rubbish dump, South Street tract. No. 43519. Introduced here and probably not native in the Arboretum. 1947] PALMER, SPONTANEOUS FLORA OF ARNOLD ARBORETUM 417 *ITpomoea purpurea (L.) Roth. Purple Morning Glory. On rubbish dump, South Street tract. No. 43072. Cuscuta pentagona Engelm. (C. arvensis Beyrich). Field Dodder. Along border of Bussey Brook, growing on tall herbs. No. 46400. *Lithospermum arvense L. Corn Gromwell. In cultivated bed near Administration Building. No. 44312. Transient. *Physostegia virginiana L. False Dragon Head. Waste ground, border of pond, South Street tract. No. 44700. ‘Transient. Lycopus virginicus L. Moist fertile ground at foot of Hemlock Hill and along brook. Nos. 28245, 46372 *Mentha piles L. Poppe eink Waste ground near pond, South Street tract. No. 47260. Prunella nes L. var. /anceolata (Bart.) Fernald. Wet ground about spring, southeast side of Peters Hill. No. 44685. *Solanum carolinense L. Horse Nettle. Waste ground, old quarry along Bussey Street. No. 45212. Persistent and spreading for several years. *Lycium halimifolium Mill. Matrimony Vine. Rocky banks and waste ground. No. 38174. Escaped from cultivation. Mimulus ringens L. Monkey Flower. Along drainage ditch, low meadow, and margins of Bussey Brook near Bussey Street. Nos. 43418, 46362. Gratiola neglecta Torr. (G. virginiana of auth., not. L.). Hedge Hys- sop. Muddy margins of pond, South Street tract. No. 42997. *Veronica arvensis L. Corn Speedwell. Cultivated and waste ground. Nos. 42997, 43117. Not common. *Galium Mollugo L. A weed in cultivated beds. Nos. 44716, 46378, 46408a *Galium Aparine L. var. Vaillantii (DC.) K. Koch. Grassy open ground, Conifer group. Nos. 45986, 46167. Galium palustre L. Wet ground along margins of Bussey Brook. Nos. 42836, 45983. Viburnum cassinoides L. Withe-rod, Wild Raisin. Open woods, Central Woods, and top of Hemlock Hill. Nos. 45516, 46867. Viburnum recognitum Fernald. (V. dentatum of auth., not. L.). Arrow- wood. Thickets, rocky ground on top of Hemlock Hill. No. 46602. *Viburnum Opulus L. Cranberry-tree. Wet ground, south side of Pe- ters Hill. No. 46975. Campanula aparinoides Pursh. Marsh. Bluebell. In wet grassy and weedy ground, southeast side of Peters Hill. Nos. 45502, 46363. Very loca *Liatris spicata (L.) Willd. Blazing Star. Rocky waste ground, old quarry along Bussey Street. Nos. 42972, 42975. 418 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Solidago gigantea Ait. var. leiophylla Fernald. (S. serotina Ait.) Smooth Golden-rod. Along margins of Bussey Brook near stone bridge. No. 47421. Rare. Solidago altissima L. Tall Golden-rod. Open ground, south side of Peters Hill. No. 47113. Solidago bicolor L. var. lutea Farwell. Dry open slopes of Peters Hill. No. 47937. Rays bright yellow. Solidago canadensis juncea. Dry open ground, south side of Peters Hill. No. 47937 Solidago canadensis X rugosa. Open ground near Walter Street nursery. No. 47085 Solidago juncea < puberula. Open ground near top of Peters Hill. No. 7. Aster novi-belgii X vimineus. Open springy ground, southeast side of Peters Hill. No. 46546. Growing with supposed parent species. Aster Kier < paniculatus. Border of low meadow, near Rhus group. No. 46967. Erigeron hid Michx. Robin’s Plantain. Moist grassy ground, Peters Hill and near Arborway wall. Nos. 36387, 36415. *Xanthium pennsylvanicum Wallr. Cocklebur. Waste ground about pond, South Street tract. No. 46551. *Helianthus petiolaris L. Sunflower. South Street nursery. No. 44698. *Helianthus doronicoides Lam. Open woods, Oak group. Nos. 42978, 43084 *Coreopsis major Walt. var. stellata (Nutt.) Robinson. Among tall herbs and shrubs, Juglans group, and near Dawson nursery. Nos. 25885, 47239a. Bidens vulgata Greene var. puberula nr Greene. As a weed in cultivated ground, Deutzia beds. No. *Cosmos bipinnatus Cav. Cosmos. tiete ground, old ed along Bussey Street. No. 43521. Probably only a transient w *Helenium autumnale L. Autumn Sneezeweed. Open grassy bank east of Administration Building. No. 45611. * Matricaria sada Cage aa Porter. Pineapple-weed. Waste and cultivated ground. No. *Chrysanthemum ena ee Bernh. Feverfew. Bank near Ad- ministration Building. No. 449 * Doronicum Pardalianches L. cecal s Bane. Open ground along bridle path, near Centre Street gate. Nos. 43115, 47171. *Cichorium Intybus L. f. album Farwell. Dry open ground, Peters Hill. No. 44690. ~~ ARBORETUM, VARD UNIVERSITY. 1947] VAN STEENIS, NOTES ON NEW GUINEAN SPECIES 419 NOTES ON A NUMBER OF NEW GUINEAN SPECIES* C. G. G. J. vAN STEENIS APONOGETONACEAE Aponogeton loriae Martelli in Nuovo Giorn. Bot. Ital. ser. 2, 3: 472, pl. 8. 1897. BritisH New GuINEA: Central Division: Kubuna, L. J. Brass 5567, Nov. 1933, alt. 100 m., common in shallow, stony streams in folest leaves brown, com- pletely submerged, spikes of greenish yellow flowers protruding just above the water. Western Division: Penzara, between Morehead and Wassi Kussa Rivers, L. J. Brass 8671, Dec. 1936, in stream in savanna forest. The New Guinean specimens have sometimes been referred to A. crispus Thunb. by F. v. Mueller (Descr. Not. Pap. Pl. 2: 51. 1886), and to A. monostachyum L. f. by Hemsley (Kew Bull. 113. 1899), but they differ from both species in their greenish yellow flowers and the persistent spathe gradually decaying from the apex to the base. The leaves, for the most part, gradually taper into the petiole. This Papuan species is also known from the southern Celebes, and possibly occurs also in Queensland. JUNCAGINACEAE Triglochin procera R,. Br. var. dubia Benth. Fl. Austral. 7: 168. 1878; Buchenau in Pflanzenr. 16(1V.14): 14. 1903. British New GuINEA: Western Division: Penzara, between Morehead and Wassi mies Rivers, Brass 8447, Dec. 1936, savanna forest, common in streams which The first record of the genus within en the species and a variety also known from Australia and Tasmania PONTEDERIACEAE Monochoria hastata (L.) Solms in DC. Monog. Phan. 4: 523. 1883. Pontederia hastata L. Mant. Pl. 362. 1771. NortHEeAST New GurneEA: Morobe District, vicinity of Kajabit Mission, Mrs. M.S. Clemens 10636, Aug.-Dec. 1939, alt. 240-600 m. DIsTRIBUTION: From India to S$. China and throughout Malaysia; as yet not reported from Queensland. PHILYDRACEAE SS novoguineensis (Krause) spr at in Bot. Jahrb. 65: 260. 1932; Bull. Jar oO penne TT. Ts 12 nei fe sp., Teijsmann in Natit, oe Ned. Ind. 37: 132. 1877. Astelia aid hen Krause in Bot. Jahrb. 59: 559. 1924 * Results of the Richard Archbold Expeditions. 420 JOURNAL OF THE ARNOLD ARBORETUM [vOL. XXVIII NETHERLANDS NEW GurNEA: Idenburg River, 4 km. SW, of Bernhard Camp, L. J. Brass 13431, Mar. 1939, alt. 850 m., rare in rain forest ravines (flowers and fruit white) ; same locality, 6 km, SW. of Bernhard Camp, L. J. Brass 12859, Feb. 1939, alt. 1150 m., occasional in undergrowth of rain forest ravine (flowers and fruit hite). DIstrRIBUTION: New Guinea and Moluccas (Ambon, Buru, Ceram) ; another species in Australia. Philydrum lanuginosum Banks ex Gaertn. Fruct. 1: 62. 1788. British New Guinea: Lower Fly River, Gaima, east bank, savanna areas, L. J. Brass 8353, Nov. 1936, common in sedge stands in shallow swamps (flowers yellow) ; Western Division, Daxw wa, Oriomo River, L. J. Brass 6002, Febr.— March 1934, common on fringes of small savanna swans (flowers yellow). DisTRIBUTION: From SE. and E. Asia to Australia, but within Malaysia hitherto only found in the Malay Peninsula. The record from Java on the basis of a Hillebrand specimen is doubtless due to an error in the locality, and Merrill (Philip. Jour. Sci. 10, Bot.: 88. 1915) showed that the Philippine record also was based on an erroneously localized collection. This is the first record of its occurrence in New Guinea. Both members of this family here recorded were easily identified through Skottsberg’s excellent revision in Bull. Jard. Bot. Buitenz. III. 13: 110-113. 1933. HAEMODORACEAE ee coccineum R. Br. Prodr. 300. 1810; F. v. Mueller Descr. Not. Pap. Pl. 1890. BriTISH ce GuInEA: Tarara, Wassi Kussa River, L. J. Brass 8384, Dec. 1936, savanna forest, common on sour gray clay soil flats (rootstock ved inside; flowers red). The genus was hitherto known only from Australia, but F. von Mueller mentioned it nearly sixty years ago as occurring on the Wassi Kussa River without indicating the collector. DROSERACEAE Drosera indica L. Sp. Pl. 282. 1753. British New Gurnea: Lake Daviumbu, Middle Fly River, L. J. Brass 7827, Sept. 1936, common in thick cover on wet grass-plains Peal purple). Drosera petiolaris'R. Br. ex DC. Prodr. 1: 318. 1824. British New GuINEA: nsanlot Diana. Wassi Kussa River, near Tarara, L. J. Brass 8386, Dec. 1936, on wet sandy flats (one of the first herbs to flower; flowers pink). Drosera ? i casa oe Nov. Holl. Pl. 1: 79, pl. 106, f. 1. 1804; Diels, Pflanzenr. 26(1V.112) : 83. British New pasta Wassi Kussa River, Tarara, L. J. Brass 8752a, Jan. 1937, on wet ground in savanna forest (rare herb; flowers whit te): This specimen was found in a cover with L. J. Brass 8752 and separated at Buitenzorg. The dimensions of the three styles, which are bifid to the base, do not agree with Diels’s description. The “normal” habitat in Malaysia for this species is found at medium or high altitudes; whereas, 1947 | VAN STEENIS, NOTES ON NEW GUINEAN SPECIES 42} D. burmannii Vahl (L. J. Brass 5962) occurs at low altitudes. In Australia, however, e.g. Queensland, D. spathulata Labill. also is found at low altitudes. It may be that this Papuan strain corresponds to the Queensland material. The type was collected at low altitude in Australia. PODOSTEMONACEAE Torrenticola queenslandica Domin in Bibl. Bot. 897: 149, pl. 35, fig. 7-13. 1925; Engler, Nat. Pflanzenfam. II. 18a: 484. 1930. Podostemon queenslandicus Domin l.c. nom. altern. British NEw GuINnEA: Port Moresby area, Roona, C. E. Carr 12415, July 29, 1935, about 700 m., alt., on submerged rocks in the Laloki River (plant dark green; fruit brown). As far as I know this is the second collection of the species, and the first record of the occurrence of the family in New Guinea. The material is in the same advanced stage of growth as the original collection, but flowers are needed for any further remarks on its exact affinity. It exactly matches the figure of the type-collection. The only previous record of this family in New Guinea was made by Zippel (Flora 12: 285. 1829; Alg. Konst- & Letterbode 1: 297. 1829) of Lemnopsis mnioides Zipp. According to Hallier f. (Meded. Rijksherb. Leiden) this is an Utricularia. In 1947 I could not trace the specimen in the Leyden Herbarium. . J. Lam (Blumea 2: 117. 1936) cited a passage from d’Albertis’s “New Guinea: What I did and what I saw” 2: 93. 1880, relating to a peculiar plant which d’Albertis found June 17, 1876. “It grows on the heaps of stones that abound in the river, and seen from a certain distance, its dark color, almost black, and its peculiar shape make it resemble the scales of a serpent. Its branches lie flat, so as to offer as little resistance as possible to the water. It owes to this curious conformation its power of resisting the strength of the current.” According to my wife’s MS. cyclopedia of Malaysian collectors this must have been on the Fly River which d’Albertis ascended from May 22 till June 25, 1876. If the plant was Torrenticola the vegetative phase must have resembled the scales of a serpent. The tiny flowering stems cannot have been meant as one cannot see these ‘‘from a distance” separately. Another argument against identi- fying d’Albertis’s plant as a representative of the Podostemonaceae is the fact that Brass explored the Fly River area, and has, to my knowledge, collected nothing of the kind. CORYNOCARPACEAE Corynocarpus australasica C. T. White, Contr. Arnold Arb. 4: 57. 1933. NETHERLANDS NEW GUINEA: sie New Guinea, Zwaluw bivak, Branderhorst 428, June 10, 1908 (shrub with immature flowers) ; id. Tuinweg near Kp. Kabatiel, Branderhorst 271, Dec. 6, 1907 (tree with fruit). Both numbers were unidentified specimens in Herb. Bog., which un- doubtedly belong to this Papuan-Queensland tree-species. It seems that 422 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Beguin has collected a new species of the genus in Halmaheira (Moluccas). STACKHOUSIACEAE Stackhousia intermedia F. M. ei in Queens]. Agric. Jour. 3(4): 174. 1898; Brouwer in Blumea 3: 174. 1938 British New Guinea: Western Division: Mabaduan, L. J. Brass 6519, April 1936, common grass associate in savanna forests (iiverers and fruit les. Wassi Risse River, Tarara, L. J. Brass 8661, Jan. 1937, common in savanna heresies raked brown-green) ; Wurel, Oriomo River, L. J. Brass 6072, Jan.-March 1934, alt. m.; same locality and date, L. J. Brass 5760, alt. 10-30 m., sporadic and airy uncommon on savanna DISTRIBUTION: From Shite through Malaysia and the Philippines to North Queensland and Micronesia. SAPINDACEAE Sarcopteryx coriacea Radlk. Sapind. Holl.-Ind. 98. 1878. Guinea: Arfak Mountains, Angi Lakes, by @ lake, L. S. Gibbs 5531, alt. 2100 m. (small tree; fruit red). The material consisting of a duplicate from the Brit. Mus. Herb. exactly matches the other collections mentioned by Radlkofer. Neither Gibbs nor Radlkofer cited this number; it was inserted in Herb. Bog. as Euonymus sp. BOMBACACEAE Papuodendron lepidotum C. T. White in Jour. Arnold Arb. 28: 272, pl. 1. 1946. ORTHEAST NEW GUINEA: between Umbili and Wobbe, R. Schlechter 16334, July 30, 1907, alt. 300-400 This specimen, eich was provisionally identified as a new species of Cumingia by Ulbrich in the Berlin Herbarium, is the first known collection of White’s new genus. Camptostemon philippinensis (Vidal) Becc. Malesia 3: 273. 1889; Bakhuizen van 9 Br = - be Jard. Bot. Buitenz. III. 6: 219, 245. 1924; Troll in Flora n. er, 28: Copa Rendon Vidal Phan. Cuming. Philip. 212, pl. 7. 1885. British New Guinea: Western Division: Tarara, Wassi Kussa River, i lier 8517, Dec. 1936, common in mangroves (tree; flowers white); Mabaduan, . J. Brass 6475, April 1936, very common mangrove tree, sometimes forming nearly oe forest (25 m. high, lower part of trunk fluted; bark + scaly, reddish brown when cut; very numerous large flattened brown knobby pneumatophores). Known from the Philippines, Borneo, the Moluccas, and North Aus- tralia; first record from New Guinea. BS STYRACACEAE Bruinsmia styracoides Boerl. & pea in Natuurk. Tijdschr. Ned. Ind. 53: 68. 1893. NETHERLANDS NEW GUINEA: km. SW. of see — mp, Idenburg ui Brass = prseneraes 13506, April es alt. 700 m., frequent on slopes of prim ie (tree 20 m. high). NortHeast NEw GuINEA: sear oe ments . M.S. Clemens 2970, April 1936, alt. 1500 m., forested hills. 1947] VAN STEENIS, NOTES ON NEW GUINEAN SPECIES 423 DISTRIBUTION: From Sumatra to New Guinea, but hitherto in the latter island only reported from the southeast part, the record being based on a Forbes specimen. Styrax agrestis (Lour.) G. Don, Gen. Hist. 4: 5. 1837. NETHERLANDS NEw GutInEA: Idenburg River, Bernhard Camp, L. J. Brass 13953, April 1939, alt. 50 m., common in flooded perhaps permanently swampy rain forest of river plains (straggling tree attaining 12 m.; flowers white) ; same place and date, L. J. Brass 13820, swampy rain forest . + 4m. high). British New GUINEA: Middle Fly River, Lake Daviumbu, L. Brass 7565, Aug. 1936, common in brushy outskirts of lake shore rain forest ie virgate tree 4-5 m. high). NortTHEAst New Guinea: Morobe District, Yoangen to Mongi, Mrs. M. S. Clemens 6613, June 1937, alt. about 1 The species is distributed very widely, from Annam and Hainan through Borneo, Celebes, and the Moluccas to New Guinea, the Bismarcks, the Solomons, and Micronesia. Hitherto it was only reported from the northern parts of New Guinea. It seems to be cee variable in habit, and very tolerant as to habitat. SYMPLOCACEAE Symplocos confusa Brand, Pflanzenr. 6(IV.242): 88. 1901. NETHERLANDS NEW GUINEA: 9 km. NE. of Lake Habbema, Brass & Versteegh 10492, Oct. 1938, alt. 2600 m., rare (tree 20 m. high, 45 cm. diameter, crown not wide-spreading; bark 13 mm. thick, smooth; wood veliew brown). A widely distributed species, described under various names, know from SE. Asia (Ceylon to Luchu Islands) through Malaysia fo New Guinea. BIGNONIACEAE Neosepicaea viticoides Diels in Bot. Jahrb. 57: 500, fig. 7. 1922. British New Guinea: Lake Daviumbu, Middle Fly River, L. J. Brass 7580, within; [leaflets 3-4;] terminal panicle with young foliage); Lower Fly River, east bank, opposite Sturt Island, L. J. Brass 7994, Oct. 1936, in rain forest (high climbing liane; flowers streaked brown; [always 3 leaflets]). NorruHeast New GutINEA: Kaiser blac os eo oe River, R. Schlechter 17557, April 1908, about 250 m. alt., liane in the for Schlechter’s eee is the first ever made of this species. In pre liminary work this plant was identified as a member of the eee and I found Schlechter’s number among the latter family inserted under a provisional name. The species was originally described by Diels as having 5-7 leaflets, but the Brass collections have only 3-4. The petiolule of the terminal leaflet is of the same length as those of the lateral ones and, hence, the leaf is digitately compound. BoTtaNic GARDENS, BuITENzorG, JAVA. 424 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII ENDLICHER’S “GENERA PLANTARUM,” “ICONOGRAPHIA GENERUM PLANTARUM” AND “ATAKTA BOTANIKA” WILLIAM T. STEARN THE MOST IMPORTANT botanical work of the Austrian scholar Stephan Ladislaus Endlicher (1804-49) is probably his Generum Plantarum se- cundum Ordines naturales disposita (quarto, Vienna) which was published in 18 parts between 1836 and 1841. It describes about 7,200 genera, 40 of these being new genera proposed by Endlicher, and was the most detailed and comprehensive survey of the vegetable kingdom available until the publication of Bentham and Hooker’s Genera Plantarum (1862-83). Its arrangement, as pointed out by Jessen, Botanik der Gegenwart und Vorzeit, 428. 1864, owes much to the earlier works of Jussieu, Bartling and Unger, but the descriptions, synonymy etc. were the result of Endlicher’s own immense erudition and industry. Many of the genera are subdivided into sections. It is on account of the sectional names published here and often overlooked by later botanists that the work retains its reference value, and the dates of their publication are accordingly important A very useful account of a copy of the Genera having most of its original wrappers preserved was given by E. Migliorato-Garavini in Annali di Botanica (Roma) 3: 169-175. 1905 and supplemented in op. cit. 22: 44-45. 1941. There is also a short note on its publication in Cata- logue Books British Museum (Nat. Hist.) 2: 529. 1904. Both, however, leave various points in doubt. The Lindley Library of the Royal Horti- cultural Society possesses a copy containing part of the original wrappers. Brockhaus,, Allgemeine Bibliographie fiir Deutschland (Leipzig) listed the contents of the parts as they were received at Leipzig, and other contem- porary references to the work are to be found in the Verzeichnis neuer Biicher of the Leipzig bookseller J. C. Hinrichs and the botanical periodi- cal Flora (Regensburg). By correlating data from these sources the publication of Endlicher’s Genera Plantarum becomes clearer and may be summarized as follows: — Fascicle 1, pages 1-80, published probably Sept. 1836; fasc. 2, pp. 81-160, i-iv, publ. Dec. 1836; fasc. 3, pp. 161-240, publ. June 1837; fase, 4, pp. 241-320, publ. Jan. 1839; fasc. 10, pp. 721-800, publ. March 1839; fasc. 11, pp. 801-880, publ. June 1839; fasc. 12, pp. 881-960, xvii-xx, publ. Nov. 1839; fasc. 13, pp. 961-1040, publ. Feb. 1840; fase. 14, pp. 1041-1120, publ. April 1840; fasc. 15, pp. 1121-1200, publ. June 1840; fasc. 16, pp. 1201-1280, xxi—xxviii, publ. Aug. 1840; fasc. 17, pp. 1281-1360, xxix—xl, publ. Oct. 1840; fasc. 18, pp. 1361-1483, xli-Ix, title, dedication, preface, publ. prob. Feb.-Mar. 1841 1947] STEARN, DATES OF ENDLICHER’S PUBLICATIONS 425 The parts appear to have been received at Regensburg and Leipzig a month or two later than the months stated on the wrappers and accepted above, transport in central Europe being very slow before the building of railways. Thus the first part seems to have been dated “20 Aug. 1836” but the earliest notices of it are in Allgemeine Bibliographie 1836: 752. 4 Nov. 1836, Literarische Zeitung 3: 877. 9 Nov. 1836, Borsenblatt 1836 (no. 45): 45. Nov. 1836; probably it appeared in September 1836. Migliorati was under the impression that the work may have consisted of 19 parts. It is, however, certain from the notices in Allgemeine Biblio- graphie 1841 (17):171. 23 April 1841, Wikstrom, Jahresber. Bot. 1839-42: 2. 1846 and Hinrichs, Verz. Buch. Jan.-Juni 1841: 66. 1841 that the eighteenth part was the last and contained pages 1361-1483; as the preface is dated “Januar. MDCCCXXXXI,” this part was probably published in February 1841. For corrections of the index, which occupies pp. 1429-1483 and contains about 17,500 names, see Botanische Zeitung 3: 301-306. 1845. Five Supplements were published: — (1) Supplementum I occupies pp. 1335-1427 of the main work and was issued in parts 17 and 18 (1840-41) of this. The others were separate publications. (2) Mantissa botanica, sistens Generum Plantarum Supplementum secundum. 114 pages. 1842 (after March). (3) Mantissa botanica altera, sistens Generum Plantarum Supplemen- tum tertium. 11 pages. 1843 (after October 4). (4) Generum Plantarum Supplementum quartum, Pars II. 96 pages. 1847 (according to title-page) or 1848. (5) Generum Plantarum Supplementum quintum. 104 pages. 1850. Pars 1 of Supplementum IV was never published. Supplementum V (or IV Pars 3) lacks a title-page and was completed by Eduard Fenzl after Endlicher’s death. On the wrapper it is entitled ‘““Supplementum Quintum” but the pages bear the heading ‘“‘Supplementum IV Pars III.” The Genera Plantarum has no illustrations but these are provided in a companion work, Endlicher’s /conographia Generum Plantarum (quarto, Vienna) with engraved plates, mostly by Ferdinand Bauer and mostly illustrating Australian species; there are also some plates by Eduard Fenzl, Aloys Putterlinck and Josef Zehner. The number at the right-hand corner of the plate is that under which the genus illustrated is described in the Genera. The plate number is in the lower right-hand corner. The plates may thus be arranged as issued according to the plate numbers or re-arranged systematically by families according to the generic numbers. The Index (pp. xiv-xvi) correlates the two sets of numbers. The Iconographia was likewise published in parts, dated as follows: Fascicle 1, plates 1-12, 1837; fasc. 2, pl. 13-24, 1837; fasc. 3, pl. 25-36, 1837; fasc. 4, pl. 37-48, title, dedication: pp. v—vili, 1838; fasc. 5, pl. 49-60, 1838; fasc. 6, pl. 61-72, 1839; fasc. 7, pl. 73-84, 1839; fasc. 8, pl. 85-96, 1839; fasc. 9, pl. 97-108, 1840; fasc. 10, pl. 109-125, pp. ix—xvi, 1841. 426 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Only four hitherto unpublished names appeared on the plates of the Iconographia, and of these Litsea Baueri Endl. op. cit. t. 44, 1838, p. x. 1841 alone concerned an undescribed species, collected by Bauer in Australia (Nova Hollandia). Coprosma Baueri Endl. op. cit. p. x, t. 111. 1841 was published as a new name for the Norfolk Island plant described in detail by Endlicher, Prodr. Fl. Norfolk 60, n. 117. 1833 as “C. lucida Forst.,” while Anoplanthus uniflorus (L.) Endl. op. cit. t. 72. 1839; p. xii, 1941 and Stirlingia anethifolia (R. Br.) Endl. op. cit. t. 23. 1837; p. xi. 1841 were new combinations, as is made clear by the synonymy etc. in “Conspectus Dispositionis Tabularum” on pp. ix-xiii; this also indicates the material from which the plates were drawn The quarto /conographia was preceded by an illustrated folio work, Atakta botanika. Nova Genera et Species Plantarum descripta et Iconibus illustrata, which, despite the excellence of its large detailed plates, is bibliographically a most unsatisfactory work. The title-page is dated “MDCCCXXXIII” and there is nothing within the work itself to indicate that it was published in 5 parts between 1833 and 1835. Moreover it terminates abruptly in the middle of a word and lacks an index. The first two plates (illustrating Diesingia scandens) are without numbers; the last plate is numbered 40, but the work contains only 35 plates, for tt. 10, 26, 28, 37, 38 were never published. There is no text to tt. 27-40, although several of these figure undescribed species. The names on the plates are, however, validly published, despite the absence of descriptions, because the plates give detailed analyses of floral structure. The name Atakta botanika, chosen to indicate that the work was a betanical mis- cellany without any systematic or geographical order, is from the Greek *draxros (‘undisciplined, not in battle order, irregular’) and proved all too appropriate! Endlicher wrote most of the text, but Fenzl contributed some descriptions. The plates are by Ferdinand Bauer, Endlicher, Fenzl, Franz Xavier, Fieber and Zehner. The publication of Atakta botanika is shown by contemporary references to have been as follows: Contents Publication Date Authority cf. . Hinrichs, Verz. Biicher July-Dec. 1833, p. 55; Fascicle 1 pp. 1-6, tt. 1-5 1833 Wikstrém, Praia Bot. 1833: 50. cf. gg Verz. Biicher Fascicle 2. pp. 7-12, tt. 6-11 1833 or 1834 July-D 1834, p. 56; Fascicle 3. pp. 13-20, tt. 12-24 1834 Wilks, _rorag Bot. 1834, p. cf. igi leeras — Bucher Fascicle 4 pp. 21-26, tt. 25-40 1835 Jan.-Juni 1835, p. 54 Wikstrém, Jahres. Bot. 1834; 53. 18 1947] STEARN, DATES OF ENDLICHER’S PUBLICATIONS 427 The following species are figured in this work: — Alseis floribunda, t. 33; Anamirta Baueriana, n. sp., tt. 39-40; Brachystemma calycinum, 18, t. 16; Ceratotheca sesamoides, 5, t. 5; Dalechampia papposa, n. sp., 22, tt. 20-21; Deppea hedyotidea, 25, t. 24; Diesingia scandens, 1, tt. 1-2; Dadenaes hispidula, n. sp. (= Distichostenon hispidulum), t. 30: Dodonaea humilis, n. sp., t. 31; Exostyles venusta, 26, t. 25; Fenzlia obtusa, n. sp., 19, t. 17; Fenzlia retusa, n. sp., 20, t. 18; Ficinia aphylla, 12, t. 12; Gillia (=Gilia) dianthoides, n. sp., t. 29; Gonotheca Blumei, 23, t. 22; Haloragis ceratophylla, 16, t. 15; Hemispadon pilosus, n. sp., 3, t. 3; Klotzschia brastliensis, 21, t. 19; Kohautia senegalensis, 24, t. 23; Limnanthes Douglasti, t. 27; Lipposa telephioides, He Sis, 13, t.- 135 Polygala Hilariana, 4, t. 4; Portulaca australis, n. sp., 7, t. 6; Pseudanthus pimeleoides, 11, t. 11: ree Sieberi, t. 10; Rubus Zahlbrucknerianus, A, Sp. 35: Schieder ain 15, t. 14; Sipanes radicans, Nn. sp., 7, t. 7: Ungnadia speciosa, n. sp., Morus Brunoniana, Baaliches Atakta Bot. t. 32. 1835 is usually con- sidered conspecific though not identical with M. pendulina Endlicher, Prodr. Fl. Norfolk 40. 1833, the collective species being known as Pseudomorus Brunoniana (Endl.) Bureau. The epithet pendulina (re- ferring to the pendulous male catkins) has, however, priority over Bruno- niana (honouring Robert Brown, 1773-1858) since Endlicher’s Prodromus Florae Norfolkicae sive Catalogus Stirpium quae in Insula Norfolk annis 1804 et 1805 a Ferdinando Bauer collectae was undoubtedly published in 1833, being listed in Hinrichs, Verz. Biicher July-Dec. 1833, p. 55 and reviewed in Allgemeine Bot. Zeitung 16: ii (48): 754. Dec. 1833, while part 4 of the Atakta botanika, containing pls. 25-40, did not appear until 1835. The area of the genus Pseudomorus Bureau (Moraceae) extends from Australia (New South Wales, Queensland) over Norfolk Island and New Caledonia to Hawaii. The four geographically isolated populations which make up this range might be regarded as independent species, but Bureau, Bentham and Skottsberg* treat them as conspecific. Comparison * Accepting Skottsberg’s view of the taxonomy of Pseudomorus as outlined in Acta Horti Gotoburg. 15: 347-350. 1944, the nomenclature of the group is as follows: — Pseudomorus pendulina (Endl.) Stearn, trans. nova. Pseudomorus pendulina var. a pendulina, var. no Morus pendulina Endlicher, Prodr. Fl. Norfolk. 10, 1833 Streblus pendulina (Endl.) F. Mueller, Fragm. Phyt. Austral. 6: 192. 1868. Pseudomorus Brunoniana a pendulina (Endl.) Bureau in Ann. Sci. Nat. Bot. sér. 5, ll: 372. 1869. — andolle, Prodr. 17: 249. 1873, sensu stricto Pseudomorus Brunoniana a pending ere castaneaefolia variatio seabra Bureau in Ann. Sci. Nat. Bot. sér. 5, 11: 373. 1869. — De Candolle, Prodr, 17: 250. 1873. Norfolk Island. Pseudomorus pendulina var. 8 pecan (Bureau), comb. nov. Morus Brunoniana Endlicher, Atakta Streblus Brunoviana (Endl.) F. Mueller, eae Phyt. Austral. 6: 192. 1868. 428 JOURNAL OF THE ARNOLD ARBORETUM [vVOL. XXVIII of abundant material showing adequately the variation within each popu- lation is needed to make clear the extent of morphological overlap; their present classification is necessarily provisional. Stephan Ladislaus Endlicher was born at Bratislava (Pressburg), then part of the Austrian Empire, now in Czechoslovakia, on 24 June 1804 and died at Vienna on 28 March 1849. He studied at the Universities at Budapest and Vienna, obtaining his doctorate of philosophy in Although he intended an ecclesiastical career, the study of natural history and of languages attracted him strongly and in 1826, for family reasons, he resigned holy orders. He then joined the staff of the Imperial Library at Vienna in 1827, being employed on the preparation of the catalogue of manuscripts, and became Keeper (Custos) of the Imperial Natural History Collections (Hofnaturalien-Cabinette) in 1836. On the death of Joseph Franz Jacquin in December 1839 Endlicher succeeded him as Professor of Botany at the Wiener Hochschule and Director of the Botanic Garden. His colleague Eduard Fenzl then became Custos of the Hofnaturalien- Cabinette. During this period Endlicher prepared not only such important botanical works as his Flora Posoniensis (1830), Prodromus Florae Nor- folkicae (1833), Genera Plantarum (1836-41), Enumeratio Plantarum xx Novae Hollandiae (1837) but also a catalogue of the codices in the Imperial Library (1836), and a catalogue of the Chinese and Japanese coins at Vienna (1837). After the completion of the Genera Plantarum Endlicher published a companion volume Enchiridion Botanicum exhibens Classes et Ordines Plantarum (1841), devoted to the definition of the families and merely listing the genera. He also wrote a Synopsis Coniferarum (1847) and collaborated with Eduard Poeppig in writing the text of vols. | and 2 (1835-38) of Nova Genera ac Species Plantarum, quas xxx legit E. Poeppig. This immense literary output is as remarkable for its variety pice Brunoniana (Endl.) Bureau in Ann. Sci. Nat. Bot. sér. 5, 11: — De Candolle, Prodr. 17: 249. 1873, sensu stricto. — Bentham, FI. rer 6: on 1873.—C. Moore et Betche, Handbk. Fl. N. S. Wales, 83. 1893. — M. Bailey, Compr. Cat. Queens]. Pl. 485. 1913. sonido Brunoniana B australiana gris in Ann. Sci. Nat. Bot. sér. 5, 11: 373. 69.— De Candolle, Prodr. 17: 250 i nie New South uae cf. Bentham (1873), Moore et Betche (1893). Pseudomorus pendulina var. Y obtusata Seat comb, sa eau Brunoniana Y obtusata apie n Ann. Sci. Nat. Bot: sér. 5, 11: 374. 1869.— De Candolle, Prodr. 17: 250 New oma cf. Guillaumin, Cat. ate Nouvelle-Calédon. Apét. 26. 1911. Pseudomorus pendulina var. § sandwicensis meres on nov. Pseudomorus sandwicensis ‘Degener, Fl. Haw 21/12. Pseudomorus Brunoniana var. sandwicensis “(Degener) aides in Acta Horti Gotoburg. 15: 347, 1944. Hawaiian Islands (Kauai, Oahu, Hawaii). 1947] STEARN, DATES OF ENDLICHER’S PUBLICATIONS 429 and quality as for its volume; it reveals, as his biographer Wurzbach* truly noted, an exceptional range of knowledge and fertility of spirit, and is distinguished by a depth of understanding, independence and keeness of judgment and acuteness of observation, which place Endlicher among the great taxonomists of the nineteenth century, with Augustin Pyramus de Candolle, George Bentham, Sir Joseph Hooker and Adolf Engler. Botany naturally occupied most of his attention after his appointment as Professor of Botany, but he worked nevertheless at publications on the elements of Chinese grammar (1845) and the laws of Saint Stephan (1849). Popular with his students and a liberal-minded patriot, End- licher took part in the unsuccessful Vienna Revolution of 1848 and was punished by being deprived of his professorship. Life then became so miserable for him that he is believed to have ended his troubles by suicide. Thus, prematurely and tragically, ended a life industriously and fruitfully spent in public service and the advancement of learning. — LIBRAR AL Honmiceurenst SOCIETY, Lo ONDON, ENGLAN * See C. V. Wurzbach, Biographisches Lexikon des Kaiserthums Oesterreich 4: 44-46. 1858 for a biography and bibliography of Endlicher and op. cit. 4: 179-181. 1858 for a biography and bibliography of his co-worker and successor Eduard Fenzl (1808-1879). 430 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII ON THE IDENTITY OF THE GENERA CUPULISSA RAF. AND PLATOLARIA RAF. E. D. Merritt AND N. Y. SANDWITH. Platolaria Raf. Sylva Tellur. 78. 1838, with a single species, P. flaves- cens Raf., was based entirely on Bignonia orbiculata Jacq. —= Anemo- paegma orbiculatum (Jacq.) DC. The record here is clear, but manifestly Platolaria Raf. (1838) has priority over Anemopaegma Mart. (1845). A synonym of Anemopaegma orbiculatum (Jacq.) DC. is Pithecoctenium panamense Benth. Bot. Voy. Sulph. 129, 1844. The case of Cupulissa Raf. Flora Tellur. 2: 57. 1836 [1837], currently placed as a synonym of Bignonia Linn., was not so clear. Because it became desirable to place Cupulissa Raf., in connection with the prepara- tion of a comprehensive Index Rafinesquianus by the senior author, an inordinate amount of time was devoted to this case, with much corre- spondence regarding the matter. The difficulty was that no botanist had made a critical study as to the generic position of Bignonia grandifolia Jacq., which manifestly does not belong in Bignonia as that genus is cur- rently restricted. The species is now shown to belong in Anemopaegma Mart., and thus we have still another generic name, Cupulissa Raf., to account for, earlier than both Platolaria Raf. and Anemopaegma Mart. On the basis of strict priority Cupulissa Raf. takes precedence over the other two proposed names for this particular genus. Several excellent coloured plates of Bignonia grandifolia Jacq. have been published, all apparently based on material derived from the plant first cultivated in Vienna toward the end of the eighteenth century. Probably the chief reason why Jacquin’s striking species was not disposed of earlier was that it remained unrepresented in most European and American herbaria until very recently and again because its fruits were, and still are, unknown. As a matter of fact the species, said to have come from Caracas, and originally described from a living specimen in con- servatory cultivation in Vienna before 1798, was not actually re-collected in Venezuela or elsewhere until 1937, approximately 140 years after it was first described and illustrated. This, in view of the distinctly spectacular nature of the species, with its abundance of large, bright yellow flowers, is all the more surprising. The explanation is that apparently the species is of somewhat local occurrence, and that it grows in regions which have not, until very recently, been intensively explored from a botanical stand- point. Probably the indication of the type locality as “Caracas” has confused the issue, for the species does not grow in the immediate vicinity of the city of Caracas, although it is now known Trom one modern collection in Venezuela. 1947] MERRILL & SANDWITH, CUPULISSA AND PLATOLARIA 431 From an examination of the original description and the several coioured plates it was evident that some group in the vicinity of Anemopaegma Mart. was represented, but from the data available, and with no specimens for study, the senior author could not be certain as to the genus actually represented. Attempts to match Jacquin’s species in the extensive col- lections at the Muséum dvhistoire naturelle, Paris (Dr. Pichon), at the New York Botanical Garden (Dr. Moldenke), at the Missouri Botanical Garden (Dr. Seibert), at the Gray Herbarium and the Arnold Arboretum failed, and the first attempts at the U. S. National Herbarium and at the Chicago Natural History Museum were fruitless. An examination of the material in the herbaria in and near London by the junior author failed to locate any specimen in the Lindley Herbarium (Dr. Gilbert-Carter), at the British Museum (Dr. George Taylor), and at Kew, other than one specimen in the latter herbarium placed under the unidentified specimens of Anemopaegma. This was labelled in Sir William Hooker’s handwriting “Bignonia grandifolia Bot. Mag. Hort. Eblan. |Dublin|.” The specimen was clearly taken from the plant on which the Botanical Magazine plate was based. Some years ago the junior author had written on the sheet “cf. Anemopaegma.” A critical re-examination of the specimen showed clearly that it agrees with Bignonia grandifolia Jacq. and that Anemo- paegma Mart. is the genus represented. An opened calyx shows the characteristic disk and ovary perfectly. The leaflets agree very well with those of Jacquin’s plate, as do the pseudostipules, the tendril, the calyx, the long linear bracts, and the stamens. The curious or “raised points” on the branchlets, petioles, and peduncle are actually raised lenticels, and are fewer and less conspicuous on the Botanical Magazine specimen than as shown in Jacquin’s illustration; in the latter it is sus- pected that the number and eminence was exaggerated. We have seen nothing exactly like them in any other described species of Anemopaegma. The apparent discrepancy in the calyx-margin as shown in the several plates is of no significance; in fact, some of the calyces of the Botanical Magazine material are split and irregularly lacerate exactly as shown in Jacquin’s plate. Neither is there any significance in the fact that the inflorescences as shown in the Botanical Magazine, based on the Dublin specimen, and in the Botanical Register (based on a specimen grown by Mr. Catley at Barnet) exhibit no 3-flowered cymes in the lower parts. The junior author’s first report after having attentively examined three coloured plates (Jacquin’s original, Botanical Register, and Botanical Magazine), and descriptions, and eOnsideUne the several genera suggested as possibilities, includes the statemen “First of all, even without other eee I have no hesitation in re- ferring the plant figured and described by Jacquin as Bignonia grandifolia to the genus Anemopaegma. ‘The three plates represent one and the same spicuously contracted into the disk), as well as the large pulvinate disk. 432 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Other characters such as the simple tendril, the absence of gland-fields at the nodes, the calyx, and the nie bright yellow corolla are all characters of species of Anemopaegm We have been unable to refer ihe. species to any described one now placed in Anemopaegma, or for that matter, in any other allied genus. It is believed that had the fruits been known (they are still unknown and hence undescribed) that the species might long since have been transferred to Anemopaegma. With certain definite data available regarding the generic position of the species requests were again sent to the Chicago Natural History Museum and to the U. S. National Herbarium for a re-examination of the unnamed Venezuelan material in Anemopaegma and in Bignonia. Dr. Steyermark reporting from Chicago, and Mr. Killip, reporting from Wash- ington, are in agreement that Bignonia grandifolia Jacq. is actually repre- sented by one modern collection, this being Pittier 14023 from the Ocumare Valley, Venezuela, May 29, 1937. The senior author has actually ex- amined the Chicago specimen, courteously loaned for the purpose, and agrees with this identification o These new data were transmitted to Dr. Léon Croizat who had recently reported for duty in Caracas. On March 24, 1947, he reported that there are two places in Venezuela known as Ocumare, one Ocumare de la Costa in the State of Aragua, the other, Ocumare del Tuy in the state of Miranda; and that the Pittier collection came from Ocumare de la Costa. The actual place where Dr. Pittier rediscovered the species is about 90 kilometers, air line, or about 140 kilometers by road, from Caracas, on the northern slopes of the coastal cordillera, approximately 600 meters below the pass leading from Maracay to Ocumare; this is the road from Caracas to Puerto Cabello, and, one suspects, the route that Jacquin followed on leaving Caracas on his return to Europe. With the generic identity of Jacquin’s striking species now established, and with the Botani- cal Magazine specimen and one modern collection for examination, the following transfer is made: ANEMOPAEGMA Martius ex de Candolle Prodr. . Pe 1845. Cupulissa Rafinesque Flora Tellur. 2: 57. Platolaria Rafinesque Sylva Tellur. 78. oe Anemopaegma grandifolium (Jacq.) comb. nov Bignonia bape Jacq. Pl. Hort. Schoenbr. 3:19, pl. 287. 1798; Willd. Sp. Pl. 3: 296. ; Poir. in Lam. Encycl. Suppl. 1: 633. 1810; Edwards Bot. Reg. 5: pl. 418, ee Spreng. Syst. Veg. 2: 830 [sphalm. grandiflora]. 1825; Hook. Bot. Mag. 57: pl. 3011. 1830; Van Géel Sert. Bot. Cl. xiv. 1830 (the hinoinial erroneously credited Willdenow; the coloured plate based on Bot. Reg. pl. 418); Reichb. Fl. Exot. 3:48, pl. 213. 1835 (the coloured plate a reproduction of that of Van ca DC. (Prodr. 9: 159. 1846; Knuth Repert. Sp. Nov. Beih. 43: 638. 1927 (Init. Fl. Venezuel.). Cupulissa — Raf. Fl. Tellur. 2:57. 1836 [1837], based on Bignonia grandifol Messrs. W illdenow. Sprengel, De Candolle, and Knuth, saw no speci- 1947] MERRILL & SANDWITH, CUPULISSA AND PLATOLARIA 133 mens representing the species, accepting it on the authority of Jacquin or of later authors. On the other hand Poiret adds at the end of his description “‘v.s.” indicating that he had seen a specimen. ‘There is no evidence that Van Géel or Reichenbach knew the species in cultivation for their coloured plates are manifestly copied from the one in the Botanical Register. Their statement as to the year of introduction into Europe, 1815, is of course erroneous, and was perhaps suggested by the Botanical Reg- ister entry of 1819: ‘We believe the plant is of very recent introduction.” We suggest that the species did not persist in conservatory cultivation in Europe because of its rank growth habit. With this disposition of Bignonia grandifolia Jacq., we now take up the problem of the proper generic name. Cupulissa Raf. (1836) has priority followed by Platolaria Raf. (1837), and Anemopaegma Mart. (1845). With 60 binomials already published in Anemopaegma Mart. and because that name has been consistently used since 1845 as the generic one for this group, we unhesitatingly recommend that Anemopaegma Mart. ex DC. Prodr. 9: 187. 1845, be added to the list of conserved generic names, and that Cupulissa Raf, Flora Tellur. 2: 57. 1836 [1837], and Platolaria Raf. Sylva Tellur. 78. 1838, be rejected. Rafinesque’s original descriptions of his two new genera are character- istically short, yet there is no doubt as to exactly what he had in mind. That of Platolaria Raf. Sylva Tellur. 78. 1838 is: ‘448. Pratovartia Raf. diff. Bignonia Siliqua maxima orbiculata plana. Scandens, fol. digit. fl. racemosis — Very distinct by mere fruit said to be akin to that'of Bign. cerulea, see 458. The flowers are not described.” “449, Platolaria flavescens Raf. Bign. orbiculata Jaq. auct. — fol. 5 natis ovatis acum. cirrhosis, racemis axil, sub 10fl.— Carthagena, fl. yellowish.”’ The description of Cupulissa Raf. Flora Tellur, 2: 57. 1836 [1837], is somewhat more ample. The problem here was a proper interpretation of Bignonia grandifolia Jacq., which was the sole basis of the proposed new “203. Cupuxissa Raf (cup split) diff. from Bignonia, cal. cupularis integro latere fisso, Cor. tubulosa campanul. limbo undulato subeq. Slobo, lob. inf. fisso. filam. basi glandul. hirsutis, didyn. quinto ster. ovar. supra disco glanduloso, stigma bilamel. Scandens, fol. conjug. fl. racemosis. — Type C. grandifolia, foliolis 2 ovatis undul. cirrhosis verrucosis, racemis multifl. pedic. bibract. flexuosis. — Carracas, large yellow flowers. Big- nonia grandif. Jaq. hort. 287. Bot. Reg. 418, Bot. Mag. 3011, auctoris. Probably several Bignonias belong here, the real have cal. dent. cor. bilabiate &c.” The above is the Rafinesque record for the two genera. As no other botanist has ever recognized either of them while Anemopaegma Mart. is universally accepted, and further because of the large number of species involved, we believe that we are justified in recommending the conservation of Martius’ generic name and the rejection of the two earlier Rafinesque 434 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII nes. Of the two Rafinesque names Cupulissa Raf. is earlier than Pla- tolaria Raf. so that if others do not agree with our conclusions, as to the retention of Anemopaegma Mart., Cupulissa Raf. would be the logical choice. ARNOLD ARBORETUM, Harvarp UNIVERSITY. AND RoyaL Botanic GARDENS, Kew, SuRREY, ENGLAND. 1947] KOBUSKI, STUDIES IN THEACEAE, XVI 435 STUDIES IN THE THEACEAE, XVI BIBLIOGRAPHICAL NOTES ON THE GENUS LAPLACEA CLARENCE E. KosuskI WHILE WORKING over the genus Laplacea in connection with my study of Theaceae I have discovered that by some peculiar oversight the type- species of the genus Wikstroemia Schrader has never been officially trans- ferred to Laplacea, the accepted name by which Schrader’s entity is now known. As far as the generic name Laplacea is concerned, there is no involved synonymy if one abides by the ‘‘Rules of Nomenclature,” since the name has been adopted as one of the “‘nomina conservanda.” However, before Laplacea was placed on the conserved list considerable feeling was ex- pressed by various authors in favor of one or another of the generic names now recognized as synonyms of the genus, Blake, in Contrib. Gray Herb. 53: 39. 1918, made a very detailed study of the situation and maintained that the original name Wikstroemia Schrader should be retained for the genus, regardless of the fact that a much larger genus in the Thymelaeaceae possessed the same name and had been placed on the list of conserved names in 1905. At the same time he proposed new combinations for all the then existent species and varieties of Schrader’s genus. However, as Blake pointed out, Wikstroemia Schrader was described in 1821, while Wikstroemia Endlicher, a genus of the Thymelaeaceae, was introduced much later, in 1833. Also the name Wikstroemia Endlicher, borne by a large genus and generally accepted, was long antedated by the name Capura Linnaeus which was introduced in 1771. There was good logic in Blake’s stand, which, however, went for nought, since the name Wikstroemia had already been adopted and conserved for the larger genus in the Thymelaeaceae. Laplacea was eventually conserved for the genus of the Theaceae. It is interesting to note that when the name Laplacea finally appeared as the conserved name in Kew Bull. 1940: 112. 1940, the synonym or rejected names listed were Haemocharis Salisb. and Lindleya Nees. Waikstroemia Schrad. was not listed. This may be accounted for in part, possibly, by the fact that W7kstroemia Endlicher of the Thymelaea- ceae had appeared on the conserved list many years earlier. However, there seems to be no reason for the oversight considering the fact that previous publicity had been given Schrader’s genus by Blake, Rehder and Sprague, and in all three cases in the interpretation of Rules of Nomenclature. It was while reviewing Blake’s paper and tracing his references, some of 436 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII them obscure and difficult to obtain, that I realized that the correct combination for Schrader’s original species had never been made. Because of the obscurity of the publications, I relate below in detail the circumstances concerning the early publications of the names Wik- stroemia fruticosa Schrader and Lindleya semiserrata Nees. In the short span of 33 days in the year 1821 the real story of the genus was unfolded. On May the fifth, 1821, Schrader, in Gottingische gelehr- ten Anzeigen (No. 72, p. 710), a publication which evidently appeared three times weekly, published the new genus Wikstroemia, and on the following page listed a single species W. fruticosa, spelled ‘“fructicosa.” This new binomial was based on a specimen (no. 15), collected by Prinz Maximilian von Neuwied in Brazil. The generic description was in Latin and as complete as any of the descriptions for members of the Theaceae at that time. There could be no questioning of the date since a date appeared on every leaf of the publication, varying, of course, with the time of publication. In the same month, at Regensburg, on May the twenty-first, Nees, in volume 4 of Flora, known also as Botanische Zeitung Regensburg (no. 19, p. 299), published a new genus Lindleya, giving no specific name, however. Nees had received a duplicate set of the Brazilian plants collected by Prinz Maximilian von Neuwied. By an odd coincidence Nees based his new genus Lindleya on the same Neuwied number which Schrader had cited in his publication of approximately two weeks earlier. It may be assumed that Nees saw Schrader’s publication of May 5, 1821, very shortly after it appeared, for on June 7 Nees (Flora vol. 4, p. 328) listed the combination Lindleya semiserrata, but merely as a synonym of Wikstroemia fruticosa. He mentioned that his own work on the collection was in manuscript form and in the hands of the collector (Neuwied) at the time. He further stated that he would rescind his earlier abstract (presumably that of May 21, 1821), since its publication was antedated by that of Schrader. He wrote also that, since it would be very instructive to see just how he agreed or disagreed with Schrader’s treatment, he would offer a bit of amusement for the readers of ‘‘Flora”’ by listing his synonyms along with the original names of Schrader. There seem to have been approximately fifty numbers in the set of specimens worked over by Schrader. Nees’ set was less complete, since he listed twelve numbers as missing from his set. Of the approximate thirty-five numbers which the two workers had in common, Nees offered synonyms for fifteen of Schrader’s new species. Of course these synonyms of Nees were all actually new combinations. His manner of listing is as follows: ‘15. Wickstroemia fruticosa Schr. ist Lindleya semiserrata m.”’ Just what feeling existed between the botanists of that time is difficult to ascertain; also the circumstances regarding the publication of the identifications on the Prinz Maximilian von Neuwied collection. At any rate, in the same year, Sprengel, in Vet. Akad. Handl. Stockholm 1821: 167. 1821, published a second genus, Wikstroemia (Compositae), named after 1947] KOBUSKI, STUDIES IN THEACEAE, XVI 437 the same Dr. J. E. Wikstrém. He ignored Schrader’s genus of the same name, failing to mention its existence. A footnote by the editor drew attention to Schrader’s earlier Wikstroemia but stated that it was under- stood to be merely a synonym of Nees’ Lindleya. The exact month of the last-mentioned publication is not certain. However, the complete action involving this confusion in synonymy took place in less than eight months! In the following vear (1822) Humboldt, Bonpland & Kunth (Nov. Gen. Sp. Pl. 5: 207) introduced the genus Laplacea, the name now conserved by the “International Rules.” The date printed in the front of the volume was 1821, which might have confused the issue even further. However. according to Barnhart in Bull. Torrey Bot. Club 29: 595. 1902. the date of publication has been ascertained as 1822, rather than 1821. The tvpe- species was in no way involved by the creation of the genus Laplacea, since H. B. & K. described L. speciosa from Peru in their work, not the species under discussion. Laplacea speciosa was designated as the type of the genus when the generic name Laplacea was conserved. Four years later (1826), Martius and Zuccarini in Nov. Gen. Sp. 1: 107, t. 66, entered still another name, Haemocharis, and used the binomial H. semiserrata. Lindleya Nees was reported in the synonymy of Haemo- charis, but not the binomial L. semiserrata. The next year (1827), Cambessedes, in St. Hilaire, Fl. Bras. Mer. 1: 300, accepted Laplacea and transferred thereto Haemocharis semiserrata, at- tributing the parenthetical authorship to Martius & Zuccarini. Since that time the species has been recorded under either Laplacea or Haemocharis with the specific name ‘‘semiserrata.”’ A detailed record of the synonymy of this species follows: Laplacea fruticosa (Schrader) Kobuski, comb. roemia gees Schrader in GOtting. Gel oor 1821 (71): 711. May 5, 1821: “fructicosa.”— Pontin [Editor] in Vet. Akad. Handl. Stockholm 1821: 168. 1821, obs. in footnote. — Blake in Contrib. Gray Herb. 53:39. 1918. Lindleya semiserrata Nees in Flora 4(1): 328. June 7, 1821, nom. nud., as syn. Haemocharis semiserrata (Nees) Martius & Zuccarini, Nov. Gen. Sp. 1: 107, t. 66. ge ai in Mém. Soc. Phys. Hist. Nat. Genéve 1: 144 in i 56). — Szyszylowicz in Nat. Pflanzenfam. III. 6: 185, 189. 1893. ae ae (Nees) Sprengel, Syst. Veg. Cur. Post. 4(2): 260, 408. 1827. Paphicen semiserrata (Nees) Cambessedes in St. Hilaire, Fl. Bras. Mer. 1: 300. 32 in Tus. Genéve 16: 407, t. 1, fig. A. 1828.— Spach, Hist. Nat. Veg. 4: 76. 1835.— Hooker in Curtis’s Bot. Mag. 70: t. 4129. 1844.— Wawra in pera Ne Bras. 12(1): 289. 1886.— Melchior in Nat. Pflanzenfam. ed. Be Diss 136. Laplacea aa Schott in Sprengel, Syst. Veg. Cur. Post. 4(2): App. 408. ee prae ees sae eee in Hoeven & De Vries, Tijdschr. 9: 100. 1842; iter. ex Mohl, Bo eit. 1:95. 1843. Laplacea ae ae in Linnaea 22: 549, 1849, Haemocharis camellioides (Sonder) Kuntze, Rev. Gen. Pl. 1:62. 1891, Lindleya fruticosa Hooker & Jackson, Index Kew. 2: 89. 1894, lapsu. This species is probably the largest and most widespread in all the 438 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII genus. To date it has been recorded from Brazil, the Guianas, Colombia, Venezuela, Peru and Bolivia in South America, Panama and even up into Mexico. Of the last I am doubtful. Also several varieties have been described under Laplacea semiserrata. Eventually a more detailed study will show whether these varieties are worthy of recognition. The brief study made so far causes the author to be skeptical of the true varietal lines, and with this in mind he is unwilling to make further new combinations until these entities are studied in relationship with the other described species of the genus. ARNOLD ARBORETUM, Harvarp UNIVERSITY. 1947] WHITE, PLANTS FROM NORTH EASTERN SONORA, MEXICO 439 NEW PLANTS FROM NORTHEASTERN SONORA, MEXICO AND NOTES ON EXTENSIONS OF RANGE! STEPHEN S. WHITE BEGINNING IN 1938 the writer spent four consecutive summers in northeastern Sonora, Mexico, principally in the region of the Rio de Bavispe, collecting plants for the University of Michigan Botanical Gar- dens and the Arnold Arboretum of Harvard University. Although not particularly inaccessible this area has received practically no attention from botanists. The only previous collectors in this region, besides those of the Mexican Boundary Survey, were Hartmann and Lloyd, botanists of the Lumholtz Archeological Expedition, who collected along the route of travel from Bisbee, Arizona through the Sonoran: towns of Fronteras, Oputo, Granados, Bacadéhuachi, Nacori and thence into Chihuahua. This expedition was in Sonora only about three months. More than 4000 numbers were collected in northeastern Sonora by the writer and those who at different times accompanied him (LeRoy H. Harvey, Edwin A. Phillips and José Vera Santos). Among them are the following novelties as well as certain other species whose presence in the area represents a notable range extension. A detailed report on the collec- tions as a whole is in preparation and will be published at a later date. Ceterach Dalhousiae (Hook.) C. Chr. In 1940 Edwin A. Phillips collected this little known fern in the Canon de Bavispe, a few miles west of the village of Bavispe (E. A. Phillips 547). Northeastern Sonora must therefore be included in the peculiar range of this plant, formerly known only from southeastern Arizona, Abyssinia and the Himalayan region of Asia. Mr. Phillips also found in this same loca- tion Asplenium exiguum Bedd. (E. A. Phillips 550), which has a similar distribution but which had been collected previously in northern Mexico. Populus monticola Brandeg. Originally described from Baja California, this tree was later collected by Palmer at Guaymas, Sonora. The writer’s collection (S. S. White 691) from Arroyo del Pulpito, north of Colonia Oaxaca, probably represents the northern limit of the species, since this point is only about 60 miles south of the United States border, where it is unknown. Ostrya virginiana (Mill.) Koch. C. H. and M. T. Muller have collected this species in Nuevo Leon, but 1 Papers from a aac aa of Botany and Botanical Gardens of the University of Michigan, No. 440 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII the writer has seen no record or its presence in western Mexico, except the recent report by Gentry? who collected it in the Sierra Surotato in northern Sinaloa. The writer found it in the Sierra de El Tigre, Sonora, growing with walnut, sycamore, ash, maple and oak (S. S. White 583, 605, 3388, 3504). This wide-spread tree of eastern North America thus has at least two isolated outposts on the Pacific watershed, separated by a distance of 500 miles or more from the nearest eastern station. Undoubtedly Ostrya in Sonora and Sinaloa is a relic of Miocene and Pliocene times when O. oregoniana Cheney (considered to be identical with or hardly distinguishable from O. virginiana) flourished in the north- western United States; at present it is not possible to tell, however, whether O. oregoniana ranged as far south as Sonora and Sinaloa, or whether O. virginiana extended westward to these localities through south- ern United States or northern Mexico. Phacelia heterophylla Pursh var. sonorensis var. nov. Herba perennis, robusta, erecta, 15-30 cm. alta, ubique canescenter hispida et glandulosi-pubescens. Folia breve petiolata, ovata, majora 5 cm, longa et 2.5 cm. lata. Corolla albida. Typus: S. S. White 4641, “Puerto del Cumarito, Sierra de la Cabel- lera, Sonora, Mexico,” in fears Universitatis Michiganens sis. This variety differs from the typical form of the species in being densely, although minutely, glandular in all its parts except the corolla and internal floral organs; the leaves are also smaller and generally lack the one or two small leaflets which are commonly present in P. heterophylla. The type was collected at an elevation of 5400 feet in an oak-grassland association. The Sierra de la Cabellera lies north of the Rio de Bavispe and west of the Sierra del Pulpito. Pentstemon campanulatus (Cav.) Willd. var. subglandulosus var. nov. Herba erecta, perennis, 60 cm. alta. Folia elliptica, acuta, in specimine typico majora 7 cm. longa et 2 cm. lata, inferiora petiolata, superiora sessilia, ab media parte ad apicem grosse serrata vel serrulata, floralia integra. Typus: S. S. White 2645, “Cantén de Huépari, Sonora, Mexico,” in Herbario Universitatis Michiganensis. This variety differs markedly in aspect from P. campanulatus, but except for the size and shape of the leaves the only other significent dif- ference noted is in the distribution of the glandular hairs. In the variety they are confined to the calyx and pedicels, but in the typical form of the species these parts as well as the corolla, peduncles, upper portion of the stem and even the upper leaves are often copiously glandular-pubescent. The type locality lies midway between the towns of Oputo and Hua- chinera at an elevation of 4300 feet. The vegetation is oak-grassland. * Gentry, Howarp Scott. Notes on the vegetation | oo Surotato in northern Sinaloa. "Bull. Torrey Bot. Club 73: 451-462, f. 1-5. 19471 WHITE, PLANTS FROM NORTH EASTERN SONORA, MEXICO 1 Lonicera cerviculata sp. nov. F m. altus, ramis glabris, juvenilibus purpurei-glaucis, deinde flavescentibus et nitidis. Folia oblongi-ovata, supra glabra, subtus pu- berula et perspicue glauca, basi truncata, apice obtusa vel subacuta, 4— 8. 5 cm. longa, latitudine circiter 4 longitudinis. Dua vel tria summa paria foliorum perfoliata, oblonga, alia breviter petiolata, estipulata. Inflores- centiae spicatae, 1-3 terminales et aliae axillares solitariae in axillis paenultimi paris foliorum. Flores verticillati, a bracteis exterioribus 2 mm. lon ngis et interioribus vix 1 mm. longis subtendentes. Calycis pars supra ovarium constricta 1 mm. longa, glauca, lobulis 5 ovatis vix 1 mm. longis. Corolla 12-14 mm. longa, aurantiaca vel rubella, siccitate flava, subventricosa, intus et extus glabra, lobulis 5 aequalibus. Staminum filamenta 1 mm. longa; antherae 2 mm. longae. Ovarium 3-loculatum, stylo glaberrimo in tubo corollae incluso. Fructus globosus, 4-5 mm crassus, in tubulum calycis desinens. Semina 1-6, albida, compressa, ovoidea, foveolata, 3-4 mm. longa et ca. 2 mm. lata. Typus: S.S. White 3422, ‘Las Tierritas de El] Temblor, Sierra de FE] Tigre, Sonora, Mexico,” in Herbario Universitatis Michiganensis. This species belongs to the subgenus Periclymenum L. and to Rehder’s subsection Phenianthi®, although its flowers are much smaller than those of other members of this group. Besides the type the writer also refers to this species his collections 2832, 3934, 4195 and 4277, all of which, with the exception of 3934, are from the Sierra de El Tigre, not far from the type locality. The other specimen was found at Rancho de la Nacha, northeast of Nacozari. These have at times a single terminal spike and no axillary spikes; in some there is only one pair of connate leaves. In none of the specimens cited are any spikes to be found in the axils of the third pair of leaves, even when this pair is connate. The name of this species refers to the neck-like calyx-tube which per- sists on the fruit, preserving the same size and shape that it had in flower. Its habitat is the pine-oak vegetational zone at elevations of 4500 to 6000 feet. Lonicera cerviculata may be distinguished from L. arizonica Rehd., apparently its nearest relative, by its smaller flowers, elongate rather than capitate spikes, and the lack of ciliate hairs on the leaf margin. University OF MICHIGAN, ANN ARBorR, MICHIGAN. 3 REHDER, ALFRED. Synopsis of the genus Lonicera. Rep. Missouri Bot. Gard. 14: 27-232, pl. 1-20. 1903 442 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII NOTES ON THE ASIATIC FLORA Hur-Lin Lr A NEW SPECIES OF SCROPHULARIA FROM UPPER BURMA IN A sTupy of the genus Scrophularia of China, it is noted that in the alpine regions of western China and adjacent areas there occur many endemic species of very limited range. Continuous to western Yunnan in Upper Burma there appears an apparently new species which is herein described. The material is based on the collections of the herbarium of the New York Botanical Garden. Scrophularia birmanica sp. nov. mentis parce pubescentibus; staminodio reniformi; stylo glabro, 5 mm ota. ype, on stony alpine meadows, at altitudes of 3660-4020 meters. on western flank of N’Maika-Salwin divide, Upper Burma, collected in flower, June, 1925, by G. Forrest, no. 26859; holotype in the herbarium of the New York Botanical Garden. UPPER BURMA: Western flank of N’Maika-Salwin divide, G. Forrest 26859, 27308, in the herbarium of the New York Botanical Garden. This species is related to S$. Delavayi Franch. of northwestern Yunnan, differing in being of smaller size, with smaller and shorter petiolate leaves, rounded calyx-lobes, and narrower corolla. THE GENUS DIPHYLLEIA The genus Diphylleia of the Berberidaceae is of particular interest because of its phylogenetic and distributional significance. Like other perennial herbaceous genera of the family, such as Podophyllum, Jeffer- sonia, Caulophyllum, and Achlys, it deviates from ordinary dicotyledons in having irregularly arranged vascular bundles in the stem. These 1947] LI, NOTES ON THE ASIATIC FLORA 443 genera are somewhat similar in habit and habitat. They are small genera with only a few species, and these occur in discontinuous areas in eastern Asia and North America, each occupying a more or less limited range. These plants are geophilous herbs with well-developed rootstocks and a few large palmately-lobed leaves. Most of them grow on mountain slopes at fairly high altitudes as undergrowth in deciduous forests, usually in association with certain genera of the Ranunculaceae like Cimicifuga. Trautvetteria, etc. The range of Diphylleia, like that of Podophyllum, Jeffersonia, and Caulophyllum, is discontinuous in two widely separated regions, namely, eastern Asia and eastern North America. In eastern North America the genus is found only in a very limited area in the Blue Ridge from Georgia to Virginia at altitudes from about 1000 to 1650 meters. In eastern Asia there are two separate areas. One is in the alpine regions from central Japan northward to Sakhalin and in the Amur region on the continent. The other is in central and western China, attaining an altitude of 3700 meters. The taxonomy of the genus as it occurs in these three separate areas is in need of clarification. The American species is D. cymosa Michx. The plant that is found in insular and maritime northeastern Asia from central Japan to Sakhalin and the Amur region is known as D. Grayi F. Schmidt. Diels (Bot. Jahrb. 29: 336-337. 1900) recorded the genus as occurring in central China in Hupeh province and named the plant D. cymosa Michx. He considered the genus to be monotypic, and consequently the Japanese D. Grayi was reduced to varietal standing. Kumazuwa (Jour. Fac. Sci. Univ. Tokyo III. Bot. 2: 346-380, f. 1-20. 1930), in outlining the range of the genus, considered D. Grayi as limited to Japan, and D. cymosa as being found in eastern Siberia and central China, as well as in North America. It is uncertain whether his conclusion was authenticated by actual speci- mens or whether it was based on earlier records like that of Diels. On studying all available specimens of the genus from America, China, and Japan, I am convinced that at least three species are recognizable, one in eastern North America and two in eastern Asia. The insular Asiatic plant found in Japan and Sakhalin is distinct from the plant found in China in Hupeh, Szechuan, and Yunnan, and this fact is substantiated by many specimens from both regions. Numerous specimens from eastern North America have also been observed, and they prove to be of a single species. No specimen has been seen from eastern Siberia. Thus it is impossible to say whether the plant from the Amur region is the same as the insular species or the central-western Chinese species. From the point of geographical proximity it is more likely to be identical with the former than with the latter. There is also the possibility that it is a distinct species or variety. The Chinese species, hitherto unnamed, is described as follows: Diphylleia sinensis sp. nov. Planta perennis, rhizoma crassum, radicibus parvis teretibus, fibrosis; bid JOURNAL OF THE ARNOLD ARBORETUM LVOL. XXVIII caulibus singulis, carnosis, 60-90 cm. altis, leviter puberulis; foliis singulis vel pluribus, longe petiolatis: petiolis ad 45 cm. longis, leviter puberulis; amina palmata, rotundata, apice ad medium lobata, ad 15—20 cm. longa, 20-30 cm. lata, margine distanter et valde inconspicue dentata, dentibus acutis, 10-25; inflorescentiis terminalibus, singulis vel raro_ paribus, elongatis, Cymosis, ramis iterum dichotomis vel raro simplicibus: pedi- cellis 5-25 mm. longis; petalis 5, ovatis vel obovatis, circiter 6 mm. longis et 4 mm. latis, subrotundatis; staminibus inclusis, filamentis crassis, 2 mm. longis, antheris elongatis, 2 mm. longis; ovariis ovoideis, 2 mm. longis; fructu subsphaerico, circiter 8-9 mm, diametro, glabro. — CHINA. Szechuan: western Szechuan, E. H. Wilson 814 (typE), July and August, 1908. Hupeh: A. Henry 6820. Yunnan: Mount Fu-chuan, southwest of Wei-hsi, an — divide, J. F. Rock 16971, May-June 1928, in shade of fir forest, alt. 3700 1 Wilson oy ae both flowers and fruits. The other two collections are fruiting materials only. The specimens are all in the Gray Herbarium. The American species, D. cymosa Michx., differs from both Asiatic species in that the leaves are more deeply lobed, with the lobes pointed, and the inflorescence is glabrous. The two Asiatic species have more shallowly lobed leaves and puberulous inflorescences. ‘The leaves of the insular species, D. Grayi Schmidt, are intermediate between the American D. cymosa and the Chinese D. sinensis. They are less deeply lobed than in the former but more so than in the latter. The Japanese plant further differs from the Chinese plant in its more glabrous leaves, its sessile or subsessile terminal leaf, its more branched inflorescence, its larger petals and larger fruits. In D. Grayi, the Japanese species, the inflorescence rises at the exact base of the blade of the subtending leaf. In the Chinese plant the inflorescence rises on the leaf stalk at a point about 5 to 8 cm. below the leaf-blade. In this respect the American and Chinese species are similar, while the geographically intervening Japanese species stands out as different from both. SoocHow UNIVERSITY, SoocHuow, CHINA. 1947] REHDER, CULTIVATED TREES AND SHRUBS 445 NOTES ON SOME CULTIVATED TREES AND SHRUBS, VI ALFRED REHDER Euonymus Fortunei (Turcz.) Handel-Mazzetti f. kewensis (Bean), comb. nov. Euonymus radicans var. kewensis Hort. ex Bean, Trees Shrubs Brit. Isl. 1: 542 (1914).—[V. Gibbs] in Aldenham House Gard. List Surpl. Pl. 18 (1929), nom. subnud.; exclud. syn. Buonvios. pepe Hort. ex H. A. Hesse in Moller’s Deutsch. Gartn.-Zeit. 47: 21 fig. 93 This an ae usually been referred as a synonym to E. radicans f. minima which it resembles in habit and general character, but its leaves are even smaller, being only 5-8 mm. long, while in f. mmima they are 8-15 mm. long. Aesculus Wilsonii Rehder in Sargent, Pl. Wilson. 1: 498 (1913). Aesculus chinensis sensu Diels in Bot. Jahrb. 29: 450 (1900), non Bunge (1833). Actinotinus sinensis Oliver in Hooker’: Icon. Pl. 18: t. 1740 (1887); 19: corrective note inserted before t. 1851 (1889, June).— Hemsley in Gard. For. 2: 124 (1889, March).— ‘Pro parte, quoad folia; quoad flores = Viburnum plicatum f. tomentosum (Thunb.) Rehd. In the corrective note cited above, it is stated that Actinotinus sinensis is based upon a trick played by one of Dr. Henry’s collectors, and is made up of the inflorescence of a Viburnum inserted into the terminal bud of Aesculus chinensis. The specific name of the Viburnum is not given and that of Aesculus is not quite correct. There can be no doubt, however. that the inflorescence represents V. plicatum Thunb. f. tomentosum (Thunb.) Rehder in a Arnold Arb. 26: 77 (1945), better known as V. tomentosum Thunb., a name unfortunately invalidated by the earlier homonym of Lamarck (1778). The leaves agree perfectly with 4A. Wilsoni Rehd., the only representative of the genus Aesculus in Western China. In the same region V. plicatum f. tomentosum is found, so that material of both species could have been collected at the same time and put together to create this fake specimen. Phillyrea latifolia var. media f. pendula (Ait.), comb. Phillyrea media Y pendula Aiton, Hort. Kew. 1:11 ea): —Rehder in Bailey, Stand. Cycl. Hort. 5: 2583 (1 916) “var.” Phillyrea pendula Willdenow, Enum. Pl. Hort. Berol. 12 (1809).— Link in Jahrb. Gewachsk. 1, 1: 159 [1818]. nev ander: Arb. Brit. 2: 1204 (1838) “P. (m.) It is doubtful if this form is still in cultivation; I have seen neither a living plant nor a recent reference to one, nor an herbarium specimen of it. Viburnum lantanoides f. praecox (Kache), grad. Viburnum lantanoides praecox Kache in Gatieneere® i 496, fig. (1912, Sept.). 446 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Viburnum alnifolium ({.) praecox Hesse in Mitt. Deutsche Dendr. Ges. 1912(21): 371, fig. [1913]. This form differs chiefly in its flowering time which is about three weeks earlier than in the typical form. The specific name V. lantanoides Michx. had been universally used as the valid name for this species until in 1898 Britton in Britton & Brown (Hl. Fl. N. U.S. 3: 229) took up the name V. alnifolium Marshall of 1785 and was followed by a number of botanists. Marshall’s plant, however, represents only partly a Viburnum; it is based on Miller’s Viburnum americanum (Gard. Dict. ed. 8, V. no. 8. 1768) which is mainly referable to Hydrangea arborescens L. This is further confirmed by a specimen of Miller’s V. americanum in the British Museum which according to S. F. Blake in Rhodora, 20: 14 (1918) is Hydrangea arborescens L. The description of V. alnifolium Marshall is an almost verbatim copy of Miller’s description without any additional new character which would indicate that Marshall did not know much of the plant he was describing, and as in almost all other cases, he does not give any reference to the author of the name. See also Mackenzie in Torreya, 27: 81-83 (1927). ARNOLD ARBORETUM, HArRvaARD UNIVERSITY. 1947 | THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 447 THE ARNOLD ARBORETUM DURING THE FISCAL YEAR ENDED JUNE 30, 1947 DURING THE PAST YEAR many changes have occurred at the Arnold Arboretum. Dr. Merrill retired as Director on reaching the age of 70, but remains as Arnold Professor of Botany. Mr. E. J. Palmer retired at the age of 73, but will remain another year for a special part-time project. Dr. Leon Croizat and Mr. Vladimir Asmous resigned during the year. Dr. Charlotte G. Nast, Curator of the Wood Collections, left to accept a teaching position. None of the resulting vacancies has been filled. We have, however, appointed Mr. Richard Fillmore to succeed the late Mr. Judd as Propagator. Mrs. Beatrix Farrand continues to serve as Consulting Landscape Gardener. Horticultural Activities. During the past year nearly 4,000 man days have been spent on the grounds. The major project has been the removal of duplicate and overcrowded plants —a total of more than a thousand. Some of the extra shrubs were moved to nursery rows at the Case estate in Weston. The larger trees were cut up for fire-wood which we hope will pay for some of the costs. Some new plantings have been made around the Administration Building where we plan to feature the introductions and new hybrids introduced by the Arnold Arboretum. A planting of ground-covers along the road will be extended along the meadow as funds permit. The opening of vistas and opening up of overcrowded areas, as well as the new plantings has been done in accord with the suggestions of our consulting landscape gardener, Mrs. Beatrix Farrand. During the current year 155 species and varieties of plants were prop- agated. We received 267 species or varieties from other institutions and distributed 283. Photographs in natural color are made for records and for display. There is a need also for black and white prints for use in publications. We have few good pictures for such purposes The National Shade Tree Conference, leaded by over 500 delegates, held its field demonstration in the Arnold Arboretum last summer. The Case Estate in Weston has proved to be a valuable adjunct to the Arboretum. Miss Louisa Case died in the fall of 1946 and left additional land and buildings to the Arboretum, together with farm equipment and a substantial endowment. Approximately 46 acres of land were sold to the town of Weston for school purposes, but 145 acres remain in the Case Estates. Most of this land is suitable for experimental work. Semi-permanent plantings of Ribes and Berberis have been started at the Case Estate together with other species which should not be given 448 JOURNAL OF THE ARNOLD ARBORETUM [voL. Xxvill valuable display space at the Arnold Arboretum proper. Various experi- mental work has been started on cultural practices and fertilizers. Some of the Bussey Institution and Cabot Foundation work is being done at Weston, The Department of Landscape Architecture of Harvard Uni- versity has been given the use of an area for demonstration work. Both at the Arnold Arboretum and at the Case Estates in Weston we are using more mechanized equipment. Power mowers, cultivators, and saws permit more work with fewer men. Weed-killing sprays, flame throwers, and a mulch of spent hops are used to control weeds with a minimum of hand labor. The plant breeding program is beginning to pay dividends in new varieties of ornamental plants. Last spring two new forsythias, Arnold Giant and Arnold Dwarf were distributed to cooperating nurserymen. Arnold Giant is a tetraploid form of Forsythia intermedia produced by colchicine treatment. Arnold Dwarf is a delicate dwarf forsythia produced by species hybridization. Next spring we shall distribute a new dwarf flowering cherry of exceptional merit. Cytogenetic studies of Sorbus-Aronia hybrids have aided in determining the relationships of these genera. Grafting experiments in the Pomoideae and in Prunus have proved to be of theoretical interest and of possible economic value. The common lilacs budded on tree lilacs continue to make good growth. This should result in our being able to grow the common lilac free of root suckers. Comparative Morphology. — Professor Bailey has continued his in- vestigations of the comparative morphology of various dicotyledonous families, with special emphasis upon salient trends of structural specializa- tion in carpels. Dr. B. G. L. Swamy, a Government of India Fellow, is making a comprehensive study of the Saxifragaceae and related families, Mr. James E. Canright of the Magnoliaceae, and Mr. R. W. Vander Wyk of the Annonaceae. Mr. William Spackman, Jr., is continuing his inves- tigations of dicotyledonous woods, with special reference to problems of identifying Mesozoic and Tertiary fossils. The work in comparative morphology has been seriously handicapped during the year by the loss of Dr. Charlotte G. Nast, Curator of the Wood Collections, who has accepted a position at the Nebraska State Teachers College at a much higher salary. Taxonomy. — Dr. Merrill has continued his taxonomic and_biblio- graphic work and completed his Index Rafinesquianus. His extensive Botanical Bibliography of the islands of the Pacific was published by the Smithsonian Institution early in 1947. During the year he was honored by election to foreign membership in the Royal Society of Edinburgh, the Royal Swedish Academy of Sciences, and was appointed as honorary collaborator of the Botanical Garden, Buitenzorg, Java. He also received the George Robert White medal of honor from the Massachusetts Horti- cultural Society in appreciation of his services to horticulture. 1947 | THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 449 During the past year Professor Rehder completed his bibliography of cultivated trees and shrubs, and part of this work is now in proof. The catalogue begun by Professor Rehder, containing all available references in the periodical literature useful in the classification and identification of woody plants, now contains 144,551 cards — an increase of 3,526 during the past year. Professor Johnston has worked on the flora of Panama, and has com- pleted a study of the plants of San José Island. Professor Raup’s “Bot- any of the southwestern Mackenzie’ has been published in Sargentia. Dr. Smith completed a study of //licium, Schisandra and Kadsura before he left, in late February, for a year of field work in Fiji. Dr. Kobuski has been busy with editorial and curatorial duties, but is continuing his work on the Theaceae. Mr. Palmer is working on Crataegus and Quercus. Dr. Perry is studying the Papuan collections made during the Richard Archbold Expeditions, while Dr. Allen is continuing work on South American Lauraceae. Dr. Frans Verdoorn is working on his Index Botanicorum, to which more than 120,000 references were added during the year. The herbarium now contains 626,999 specimens. A total of 35,898 specimens were received by exchange, purchase or for identification. The Arboretum sent 13,046 specimens to other institutions. The Library. — At the end of the fiscal year the library contained 46,520 bound volumes, 200 unbound volumes, and about 14,000 pamph- lets. Through the Interlibrary loan system we loaned 137 volumes in addition to 435 volumes to other Harvard libraries. We borrowed 188 volumes from other libraries, including those of Harvard. Several exhibits of our rare books were on display for the Visiting Com- mittee, various Garden Clubs and the Massachusetts Horticultural Society visitors. With the end of the war the flow of visitors and students has increased greatly. Among outstanding gifts were a Gazetteer of Chinese Place Names from the Army Map Service, and six volumes of Backer’s Flora of Java, a gift from Dr. Verdoorn. Financial Report.—- Endowment funds received last year included $55,464.91 from the Marion Roby Case estate, $10,350.00 from the J. B. Case Fund, by sale of property to the town of Weston, $150,000.00 from the Katherine T. Balch bequest and contributions to the Judd Memorial Fund, amounting to $2,862.50 — a total of $218,577.41 for the fiscal year. In addition funds for current use totaled $4,970.25, including $2,320.00 for cultural purposes, a grant of $500.00 from the Bache Fund of the National Academy for Dr. Merrill’s work, and a grant of $600.00 from the same fund to Dr. A. C. Smith. Harrison W. Smith contributed $1,500.00 for Chinese exploration work, and the publication fund was increased by $50.25. 450 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVill The total income for the year was $152,409.43, while expenses amounted to $170,914.89, leaving an apparent deficit of $18,505.46. The deficit was covered by drawing on a credit balance built up during the war vears, and by drawing on contributions for cultural purposes which had also accumulated during the war vears. e endowment of the Arnold Arboretum has increased more than three-fold since Sargent’s administration, yet the labor and scientific personnel has not been increased materially. Costs of labor, equipment, and materials have increased almost as rapidly as has the income from endowment. The income from endowment is adequate for routine main- tenance, and present projects, but new major projects must be financed by annual or special gifts. Bibliography of the Published Writings of the Staff and Students July 1, 1946 — June 30, 1947 Croizat, L. Euphorbia maculata L. Bull. Torrey Bot. Club 74: 153-155. 1947. ————— Nomenclatural transfers and corrections in the Euphorbiaceae. Trop. Woods 88: 30-32. 1947. Novelties in American Euphorbiaceae. Jour. Arnold Arb. 27: 289-291. 46. Trochodendron, Tetracentron, and their meaning in phylogeny. Bull. Tor- rey Bot. Club 74: 60-76, 15 fig. 1947. —_—, Cg anor landscape changes at the Arnold Arboretum. Arnoldia 6: 8, pl. 7. ea e E. ieee in the Theaceae, IX. In: Kanehira, R. & Hatusima, A. The Kanehira-Hatusima 1940 Collection of New Guinea ie V. Bot. Mag. (Tokyo) 55: 504-510, f. 1-5. 1941 Studies in the a Xv. A review of the genus Adinandra. Jour. Arnold Arb. 28: 1-98. ’ MANG eae P. C. The origin and evolution of maize. Advances in Genetics 1: 161-207, f. 1-11. 1947. Merritt, E. D. An vere bryologist’s observations on the Philippine moss flora. Bryologist 50: 4-13, 1947. mos seen herbarium. Rhodora 48: 201-205. 1946. — The pear Arboretum the fiscal year ended June 30, 1946. Jour. Arnold Arb. 27: 486-509. The Arnold oS ae University, Jamaica Plain, Massachusetts. Bull. Gard. Club Am. X, 7: 9-13. 1946. ———— A cldamar es hbliography of the Islands of the ‘Pacific. Contrib. U. S. Nat. Herb, 30: 1-32 Botanical eee Botanic Garden, Bussey Institution, Arnold Arbore- tum, Maria Moors Cabot Foundation for Botanical Research (Annual Reports 1945-1946). Off. Rep. Harvard Univ. Reprint pages 1-28 Notes on the dates of issue of Desvaux’s Journal de Botanique. Jour. Arnold Arb. 28: 247-250. 1947. On the identity of Brandisia Souliei Bonati. Jour. Arnold Arb, 28: 251, 252, pl. 1. 1947. 1947] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 451 Thomas Barbour. Jour. Wash. Acad. 36: 283, 284. 1946, Unlisted ria a binomials in Eaton’s “Manual of Botany.” Am. Fern Jour. 37: 6—1 947. ——— & PErrR oo C4 Some additional records for the Guam flora. Jour. Arnold Arb)i2: ee 328. 1946. R, J. R. New plant names published by Amos Eaton between the years ‘si7 a 1840. Bartonia 24: 26-7 PatMeER, E. J. Crataegus in rag and central United States and adjacent Canada. Brittonia 5: 471-490. 1946. Raup, H. M. The botany : 5 ane Mackenzie. Sargentia 6: 1-275, pl. 1-37. fig. 1-16. 1947. REEDER, J. R. Notes on Papuasian Saxifragaceae. Jour. Arnold Arb. 27: 275-288, 1 fig. 1946. Renper, A. Notes on some cultivated trees and shrubs, IV. Jour. Arnold Arb. 7: 316-318. 1946; V. 28: 253-258. Sax, H. J. & Sax, K. The A aaa of generic hybrids of Sorbus. Jour. Arnold Arb. ve 137-140, pl. 1. 1947, Sax, K. The Bussey an Arnoldia 7: 13-16. 1947. —_——— new plants are made. Horticulture 25: 127, 128, illus. 1947. re ; Plant breeding at the Arnold Arboretum. Arnoldia 7: 9-12, pl. 2. 1947. Te ergs effects on X-Ray induced chromosome aberrations. Genetics 32: 75-78. 1947, ScuwarTeN, L. Index to American botanical literature. Bull. Torrey Bot. Club 3: 381-396; 481-494; 596-605. 1946. 74: 89-98; 188-196; 267-278. 1947. Illustrissimi auctoris bibliographia. In: Merrilleana: A selection from the general writings of Elmer Drew Merrill, Sc.D., L1.D. Chron. Bot. 11: 144-157. 46 Situ, A. C. Studies of Pacific Island plants, V. New and noteworthy flowering plants from Fiji. Jour. Arnold Arb. 27: 319-322. 1946. & Bearp S. The genus Eschweilera in Trinidad and Tobago. Jour. Arnold Arb. 27: 305-313. 1946. VeRDooRN, F. De los fines y eae ti de oe oP Biologica y la Biografia. Rev. Fac. Nac. de Agronomia 6: International sees ac in emit ture. United Hortic. II 2:12, 13. 1947. The need for International Visitors’ Research Stations in certain areas of the tropics. Austr. Jour. Sci. 9: 51-54. 6. Notas para los colaboradores os pias Botanicorum. Rev. Fac. Nac. de Agronomia (Medellin) 6: 393-413. On the need for a series of generic oe particularly for the cae ae plants of the = and New World tropics. (Abstract). Am. Jour. Bot. 33 Suppl: 27a, 28a. 1946. Publis toae pronlens and international cooperation in biology (Abstract). Am. ats Bot. uppl: lla. 6: Sobre la nseecidad de existencia de Estaciones Internacionales de Investi- gacion para Visitantes en algunas zonas de las regiones tropicales. Ciencia 7: 156-160. 1946 ————— Visitors’ Research Stations in the tropics. Nature 158: 243, 244. 1946. Summary in Gard. Chron. III 120: 109. 1946. (as editor) eas i and pate aeons Biological Newsletter. Waltham, Mass. vol. 1 Jan. 1947 e Wyman, D. Hedges for sie Peele. es aS rt. Mag. 25: 207-226, illus. 1946. ———— Now is the time to judge broad-leaved evergreens, Horticulture 25: 126, illus. 1947. 452 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII On color in evergreens. Horticulture 25: 94. 1947, The past year at the Arnold Arboretum. Arnoldia 7: 1-8. 1947. Spring in the Arboretum. Arnoldia 7: 25-28, pl. 4. 1947. Take an interest in making your plantings new! Parks & Recreation “29: 302-305. 1946. Use several varieties to lengthen lilac time. Horticulture 25: 278, illus. 4 Watch the elms. Arnoldia 7: 17-23, pl. 3. 1947. KARL SAx, Director 1947] THE ARNOLD ARBORETUM DURING THE FISCAL YEAR 153 Staff of the Arnold Arboretum 1946 — 47 Kart Sax, S.D., Professor of Cytology and Acting Director. Joun GEoRGE JACK, Assistant Professor of Dendrology, Emeritus. AtFreD Reuper, A.M., Associate Professor of Dendrology and Curator of the Herbarium, Emerit us. JoserH Horace Fautt, Ph.D., Professor of Forest Pathology, Emeritus. ELMER Drew MerriLt, $.D., LL.D., Arnold Professor of Botany. IrnvING WIDMER BaIrLey, S.D., Professor of Plant Anatomy. Paut CuristoPpH MANGELSDORF, S.D., Professor of Botany. ALBERT CHARLES SMITH, Ph.D., Curator of the Herbarium. Ivan Murray JounsTon, Ph.D., Associate Professor of Botany. Hucu Miter Ravp, Ph.D., Associate Professor of Plant Geography. CLARENCE EMMEREN KosuskI, Ph.D., Assistant Curator of the Herbarium. DonaLp Wyman, Ph.D., Horticulturist. ERNEST JESSE PALMER, Collector and Research Assistant. CAROLINE KATHRYN ALLEN, Ph.D., Assistant in the Herbarium. Lity May Perry, Ph.D., Technical Assistant. Leon Croizat, J.D., Technical Assistant. FrANS VERDOORN, Ph.D., Bibliographer. CHARLOTTE GeorciA Nast, Ph.D., Curator of Wood Collections. LAZELLA SCHWARTEN, Librarian. VLADIMIR CONSTANTIN Asmous, A.B., Assistant Librarian. BEATRIX FARRAND, L. H. D., Consulting Landscape Gardener. Susan DeLano McKeE Lvey, A.B., Research Assistant. CoNnsTANCE MANSFIELD GILMAN, Business Secretary. RoBert GEROoW WILLIAMS, B.S., Superintendent. Ricuarp Harotp FiritMore, M.S., Propagator. ee ee 1947] INDEX INDEX Synonyms are printed in ztalics; new names in bold-face type. Acmopyle, A Fijian, 141 Acmopyle Pancheri, 141, 142 — Sahniana, 142 Acranthera Arn. ex Meisn. (Rubiaceae), Monograph of the Genus, 261 Acranthera, —ser. Ciliolatae, 297, 29 —ser. Iteophyllae, 298, 301 —ser. Monopodiales, 287, 291 — ser. tanec aes — 300 — ser. Sympodiales, — ser. i a oie 302 — subgen. Ab toplianidesma: 280, 295 —subgen. Amphoterosanthus, 279, 283 —subgen. Androtropis, 279, 281 —subgen. Athroophleps, 280, 304 —subgen. Cleomocarpus, 279, 282 —subgen. Dichroanthes, 280, 286 —subgen. Eu-acranthera, 279, 280 —subgen. Mitracme, 280, 297 —subgen. Pha i a 280, 285 —abbreviata, 287, —anamallica, 280 —athroophlebia, 305 —atropella, 298, 304 —aurantiaca, 287, 291 — axilliflora, 287, 294 — Hallierii, 298, 303 — hirtistipula, 287, 295 — involucrata — ophiorrhizoides, 283 —parviflora, 296 — philippinensis, 298, 301 Acranthera Ruttennii, 305, 307 7, 29 ee 297, 298 — Yate 4 a ee , 281 Actinotinns sinensis, 445 Adena Aduenen audieain: 310 — Capillus-veneris, 313 — latifolium 3 subcodes: 13 pete Sean in oe XV. oo Gen Prine — sect. Peas y —sect. Eleutherostyla, 2 — —tonkinensis, 18 — bracteata — caudata, 1, 76 cea 49, 59 — cordat Be ty on 65 456 JOURNAL OF THE ARNOLD ARBORETUM Adinandra cordifolia strigosa, 50, 66 a hag ae 11, 20 49 — Howii, Ai, 28 — Mannii, 94 , 19 Adinandra maquilingensis, 35, 39 8 —megaphylla, 10, — 3 — Millettii, 18, 19 — Millettii, 10, 11 60 -- therein 49, 53 — ri — rostra "40 — ea ae 22 —— glabra, 55 —— parvifolia, 55 — stylosa, Agnus castus incisa, 258 ——— multifida, 258 Agrostemma Githago, 414 Alopecurus carolinianus, 412 —geniculatus, 412 Amaranthus hybridus, 414 — — hypochondriacus, 414 Amelanchier asiatica, 139 [VOL. XXVIII 1947] Amelanchier florida, 138 —sp., Amelasorbus Hoseri, 139 — Jackii, — Raciborskiana, 139 Anacharis occidentalis, 411 Andropogon ra ce, — scoparius, Androtro pis neon 281 Anemia cheilanthoides, 310 — fulva, — hirsuta, 313 — Phyllitidis, 310 10 5 43 Anogramma chaerophylla, 313 Antrophyum lanceolatum, 310 fo ees in New Guinea and a Proposed ew Section of the Genus, Notes on Two Species of, 25 Araucaria sect. Colymbea, 260 —sect. Eutacta, —sect. Intermedia, 260 rea 259 — Klinkii, Archbodiodendron, 2 — calosericea, pen per 414 Arnold Arboretum during the Fiscal Year Ended June 30, 1947, The, 447 Arnold Arboretum, Second Supplement to the Spontaneous Flora of the, 410 Arnold Arboretum, 1946-47, Staff of the, 4 5: Aronia arbutifolia, 138 florib 415 Asiatic Flora, Notes on the, 442 Asplenium malayo-alpinum, 313 —monanthes, 313 —— Galeotti, 313 Astelia novoguineensis, 419 -belgii xX vimineus, 418 , 418 Bolivia and Astragalus in entina, Chile, 33 Actes aconcaguensis, 351 —affinis, 348 —alberjilla, 381 — alfalfalis, 383 —amatus, 341, 376 — Ameghinoi, 356 INDEX Astragalus amoenus, 408 munategui, 340, 362 — Amunateguianus, 362 — araucanus, 3 — chrysanthus —chubutensis, 340, 367 — clandestinus, — — Flossdorfii, 365 — crypticus, 3 —cryptobotrys, 340, 365 8: — Dielsii, 345, 407 458 JOURNAL OF THE ARNOLD ARBORETUM (VOL, XXVIII Astragalus dilectus, 383 ec atiig minimus, 345, 404 — Dillinghami, 344, 402 — mino — distans, 368, 369 met ee 345, 403 — distinens, 341, 368 — modestus, 395 — Dodti, 342, 380 — monospermus, 346 — dolichostachys, 351 — Monteroi, 341, 375 — Domeykoanus, 342, 388 —monticola, 339, 356 Bi hit as 380 — Moyanoi, 362 — Dus 359 —— villosula, 360 erie 354 —nanus, 358 — Edmonstonei, 338, 347 —nubigenus, 356 —elatus, 3 —— Arnottianus, 356 —elongatus, 381 —nudus, 342, 380 Famatinae, 343, 393 — oblongifolius, 350 ferrugineus, 376 — ochroleucus, 345 — filifolius, 381 — oliganthus, 387 — flavocreatus, 345, 405 — Orbignyanus, 386 — flavus, 347 — niger 358 — Garbancillo, 342, 384 — orite: 8 Mandoni, 384 —orthoeaeps, 342, 387 — Germaini, 342, 383 — ovallensi. 46 — — azureus, 383 aiaaees oo 358 — Gilliesi, 407 —-— Duseni, 359 — glaucophyllus, 415 — prandifiora, 359 — gracilis, 407 ape 358 — grandis, 353 — paposanus, 339, 351 — gratus, 381 — Parodii, 341, 371 — Herzogti, 397 —patagonicus, 340, 363 — Hieronymi, 394 — patancanus, 390 — Hohenacheri, 348 — pauranthus, 341, 373 — hurtadensis, 408 —pehuenches, 339, 353 sogenus, 344, 399 — pencanus, — Illinii, 3 eruvianus, 344, 402 — infirmus, — Philip pi, — inflatus, ¢ — Philippianus, 408 ay eT) 341, 372 —pictus, 408 — laetevirens, — Pilgeri, 39 — Landbecki, 361 — Pissisi, 338, 347 -— lanuginosus, 408 — placens, — laxiflorus, 351 — — oblongifolius, 350 — — dissitiflora, 351 — 376 — leucomallus, 376 — prostra ~ — limariensis, 338, 348 — pulche . — litoreus, 376 ie Seon 345, 404 Looserii, 338, 349 — punae, 343, 38 — macrocarpus, 353 — punensis, 395 — — petiolatus, 353 — pusillus, ,4 — macrophysus, 355 qui ndecimjugas, 409 —— robustus, 355 — Rahmeri 358 — Mandont, 384 — Reedii, 35 — maulensis, 34 — Reichei, 345, 404 — megalocarpus, 353 — Rengifoi, 381 — melanogonatus, 381 — —lejocarpa, 370 — Meyenianus, 3 — Rhudolphi, 382 — — paucifoliolata, 356 — Ruiz-Lealii, 341, 373 — micranthellus, 343, 390 —rupestris, 356 1947 | Astragalus Sanctae-Crucis, 342, 387 — San Romani, 35 Boer eee: 408 — Sege eo —— Echegarayi, 354 — subandinus, 376 —tehuelches, 340, 367 — tricolor, 376 sas tpi lonas 343, 391 — trifoliatus, 379 —uniflorus, 343, 397 —unifultus, 384 — Urbanianus, 344, 401 — vagus, 339, —valerianensis, 339, 354 — valparadisiensis, 341, 379 352 — Werdermanni, 344, 403 ee patula, Axyr amarante, 414 serie papyrifera, 4 Bibliography of ity Published any of the Staff ae Students, July 1, 1946 — June 30, Bidens vulgata oe 418 eae ue og 432 Bia : he he gan ahe ee in + Nev oe by L. J. Brass, I, I, 207 Blechnum blechnoides, 310 Brandisia, ee pli ee of the Genu 7 Brandisia oe Bonati, on the Identity {251 INDEX 459 Brandisia, 130 subgen. Coccineabotrys, 131 —subgen. Eubrandisia, 131 —subgen. Rhobrandisia, 131 Brass, I, II, The Cyperaceae Collected in New Guinea by L. J., 99, 207 Brassica campestris, 414 — juncea BREMEKAMP, Monograph of the Genus Acranthera Arn. ex Meisn. (Rubiaceae), 261 —A Monograph of the Genus Pomazota Ridley (Rubiaceae), 186 h of the Genus Streblosa 5 HOLZ, JOHN & aon E. Gray. Fijian Fee 141 Bulbostylis barbata, a Buxus flaviramea Calyptranthes lane, Camellia axillar — japonica Campanula aparinoides, 41 Camptostemon philippinensis, 422 — canescens, 413 — capillacea, 101 —cephalantha, 413 460 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Carex constricta, 112 Chun, Woon-Young & Ying Tsiang. Spe- —con er 104 cies Novae Fagacearum Sinicarum, 320 — curta, Cichorium Intybus album, 418 — “cine ena 102 Clematis virginiana, 414 Doniana, 115 Cleyera japonica, 33 — eremostachya, 99 ipingensis, 33 —euphlebia, 108 —-— montana, 4 — fascicularis, 116 —lasiopetala, 54 — Gaudic engery 107 — Millett 1 — hystericina Coniogramme fraxinea, 310, 314 eater clad 104 Coptophyllum, 195 —laxiculmis, 413 — bracteatum, 201 —melanophora, 106 — capitatum, 198, 199, 201 — montivaga, 109 — pilosum, — neoguineensis, 105 Coreopsis major stellata, 418 — perilei Corydalis sempervirens, —— Corynocarpus australasica, 421 afflesiana continua, 104 Cosmos bipinnatus, —rara capillacea, Cryptogramme japonica, 311 — rhizomatosa, Cumingia philippinensis, 422 —scoparius moniliformis, 413 Cupressus pisifera Sanderi, 253 — subtransversa, 115 Cupulissa Raf. and Platolaria Raf., On —tribuloides reducta, 413 the Identity of the Genera, 430 —tricholoma, 110 Cupulissa, 432 — tricuspidata, 113 — grandifolia, 4 viresce 13 Cuscuta arvensis, 417 Carya sect. Apocarya, 254 pentagona, — sec acania, 25 Cyathea fulva, 313 Cassia heii te. 126 — Harrisii, 313 —insularis, 126 — mexicana, 313 Celtis occidentalis Cyclopeltis semicordata, 314 Cephaloscirpus a aioes 212 Cyperaceae Collected in New Guinea by Ceterach Dalhousiae, 439 L. J. Brass, I, II, The, 99, 207 Chamaecrista clarensis, 126 Cyperus, 213 Chamaecy paris obtusa ericoides, 253 —subgen. Eucyperus, 213 — — Sanderi, —subgen. Juncellus, 219 Chrmitahsiuls insularis, 126 —subgen. Kyllinga, 225 Cheilanthes membranacea, 314 —subgen. Mariscus, 221 —microphylla, 314 —subgen. Pycreus, 22 ME heey: 310, 314 —subgen. Torulinium, 226 —tenuifolia, 314 —aquatilis, 219 Chelonopsis ‘aldifiora, 252 —auricomus, 214 — Souliei, 252 — — microstachyus, 214 Chenopodium Botrys, 414 — brevifolius, 22 —glaucum, 414 — — stellulatus, 225 —hybridum, 414 — canescens, — paga 414 —Carrii, 213 Chinese Fagaceae, New and Noteworthy, —cinereobrunneus, 216 230 —compactus, 22 Chrysanthemum Parthenium, 418 —-—decolorans, 221 Cuun, Woon-Younc. New and Note- —cylindrostachys, 224 srotbey Chinese Fagaceae, 230 —cyperoides, 224 —& Hsranc-Hao Hsur. Notes on Chi- — diffusus, 2 nese Sterculiaceae, 328 — —celebicus, 215 —& Yinc Tsranc. A New Lactiterous ening ing 215 Sauk in South China, 245 — princeps, 2 1947 | Cyperus diffusus pubisquama, 215 | — dilutus, 221 — — decolorans, 221 —distans, 214 — ferax, —— novae-hannoverae, 226 —globosus, 220 — Haspan, 219 — Holoschoenus, — — fusci-squamatus, 223 — Iria, — javanicus, — Kyllingia, 225 — — humilis, 225 - melanospeniaue) — monocephalus, 22 — — floribundus, 218 — pennatus, 222 a — pumilus nervulosus, 221 — Sip ib 219 —yvacem 4 -- ae ee 221 ——ornans, 221 — stoloniferus, 213 — strigosus capitatus, 412 Cystopteris ie 310, 314 Cytogenetics of Generic Hybrids of Sor- bus, The Dacrydium aby 141, 142 — taxoides, 42 he. oo 314 — cicutaria, 3 ee een flexuosa, 412 INDEX Desmella Aneimiae, 310 Desmodium Dillenii, 416 Desvaux’s Journal on the Dates of Issue of, 24 Dioscorea linearis, — nipensis Diphylleia sinensis, 443 Diplazium crenulatum, 314 —expansum, 31 Doronicum Pardalianches, 418 — Clarkei, 310 — diplazioides, 315 — effusa, —equestris, 315 — firma, 315 —heteroclita, 315 = melanochlaena, 315 10 ——gu Sameera 311 Elaeagnus umbellata, 416 Elaphoglossum aaah 311 — lingua, ie — Pringlei Eleocharis ‘pitas, 412 — 20 —- ee 226 — laxiflora, 227 — ovata, — pellucida, 227 — plantaginea, 227 — spiralis, 227 — variegata laxiflora, 227 de Botanique, Notes 462 JOURNAL OF THE ARNOLD ARBORETUM Eleusine indica, 412 Endlicher’s “Genera oo 7 “Icono- graphia Generum _ rum” an “Atakta Botanika,” Eragrostis on ih pilosa, pe ke lus, 418 Eriolaena spectabilis, 330 —sp., 330 Eschscholtzia a 414 Eugenia moaens — Sturrockii, a Euonymus Fortunei kewensis, 445 —radicans kewensis, 445 Eu phorbia maculata, 416 , 416 sd lusiopetata Evonymus Sey pelt _ H. ede Fern Hosts of ust Fungi, 309 Festuca elatior arundinacea, 411 411 ennsylvanica lanceolata, 416 Fuirena umbellata, 228 Fungi, Tropical Fern Hosts of Rust, 309 Galium Aparine Vaillantii, 417 — Mollugo, 417 — palustre, 417 GLeEaAson, H. A, MIT Two New Riowering ‘Plants iron St. Lucia, 3 Glyceria ~— Fernaldii, 412 Gonyanera, Gordonia oe 73 75 E. & Joun T. Bucuuotz. A Fijian Acmopyle, 141 Gumsia chloroxylon, 330 — tiliacea, 330 Haemocharis camellioides, 437 rsa, 43 ei eipa deat 437 modorum coccineum, 420 actin autumnale, eae ‘dronicoies 418 — petiolar Helicteres = hiris 330 Helmholtzia novoguineensis, 419 [VOL. XXVIII Hemionitis palmata, 316 Hicoria sect. A pocarya, 254 Hsr1anc-Hao ON- YOUNG Notes on Chinese Sterculia- ceae, Hyalopsora Cheilanthis, 311 —obovata, 311 — Polytaenii, 310 Hypaelyptum microcephalum, 228 207 — eulerhanen 209 — pa , 209 — uadr isicaeiia. 209 oe a or inact 209 —scabrum, 207 Ipomaea purpure Iris Pseudacorus pli, 413 Isolepis barbata, Jounston, Iva av ena in Ar- gentina, Bolivia and Chile, Juniperus sabina Sanderi, 253 ngeri, 253 — Sanderi, 253 Kajewskiella, a New Rubiaceous Genus from the Solomon Islands, 331 wa ty — trichan Kieseria pene 125 KosuskI, CLARENCE E, Studies in the Theaceae, XV. A Review of the Genus Adinandra, 1 — Studies in the Theaceae, XVI. Biblio- graphical Notes on the Genus Laplacea, 435 saa pi bingoes: 225 225 — — subtriceps, 22 — Plant in South China, A New, , Studies in the Theaceae, XVI. on the Genus, Pat cea Bibliographical Notes 3 Laplacea camellioides, 437 1947 ] Laplacea fruticosa, 437 — inaequalilatera i Lasiagathis humilis, 200 Lemna minor, Lepironia articulata, 213 — bancana, Lespedera en 416 6 ie Hor Lin. “Wetes on the Asiatic Flora, 442 —Relationship and Taxonomy of the Genus Brandisia, 127 Liatris spicata, 41 ia oligantha, 333 m ee 27 — californica aureu m a 256 m macrocarpum, 25 — elegantissimum, 256 — Ibota aure eum Vecast. 256 —— Vicaryi, 256 Ean erit cs tricolor, 256 —lodense, 255 a= TOOUS —nanum compactum, 255 —ovalifolium aureo- iar anniiine. 256 a ne: aureo-marginatum, 256 —— — variegatum, 256 —— variegatum, 256 — Tschonoskii macrocarpum, 257 — Vacaryi, 256 oe nanum, 255 20 Lindle sa ratios, 437 — semis Lipocarpha argentea, 229 , 228 Se enenalenen Lithocarpus seis ee 230 — Chifui, 320 —chrysocoma, 321 — Garrettiana, 234 INDEX 463 wae de rca 233 — sap byiloidla 236 Lycopus virginicus, 417 Lygodium heterodoxum, 313 —micans, 311 —polymorphum, 311 —volubile, 311 Malus micromalus, 415 — purpurea pendula 4 Mapania subgen. Cephaloscirpus, 212 1 Mariscus Sieberianus, 224 8 matrices 418 Medicago sativ: 15 Menispermum pee 414 Mentha piperita, 417 MErrRILL, E s on the Dates of Issue of Desvaux’s Journal de Botani- que —On the Identity of Brandisia Souliei Bonati, 25 —& L. M. Perry. suena Genus Islands, 3 & N. Kajewskiella, a New from the Solomon a On the Identity of the Genera oa Raf. and Pla- 310 i eaepatlen ber iat 311 — columbiensis, 464 Milesia Coniogrammes, 310 311 a comer Pa 311 — tenuis, 311 Milesina’ arisanense, 311 iogram 3 Mimulus ringens, 417 Monochoria hastata, 419 Monograph of the Genus Acranthera Arn, ex Meisn. (Rubiaceae), A, 261 Monograph of the Genus Pomazota Rid- ley cae gery A, 186 Monograph of the Gems —— Kor- thals (Rubiaceae), A, Morus alba, 413 11 New and Ac ewouhe Chinese Fagaceae, 230 New Guinea and a Proposed New Sec- tion of the Genus, Notes on Two Species of Araucaria in, 2 Un Oo New Guinea by Brass, I, The Cyperaceae Collected in, 99 Guinea by Brass, II, The Cyperaceae Collected in, 207 New Guinean Species, Notes on a Num- ber of, 419 New Lactiferous Plant in South China, A, 245 New Pants from Northeastern Sonora, Mexico and Notes on Extension of ca 439 Notes on a oo of New Guinean Species Notes on See Sterculiaceae, 3 Notes on Some Cultivated Slaid and Notes on the Asiatic Flora, 442 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Notes on the Dates of Issue of Desvaux’s Journal de Botanique, 24 Notes on Two Species of Araucaria in New Guinea and a Proposed New Sec- tion of = Genus, 259 soria Jenmanii, 316 Cenuthane pierevey 416 O : ti of Brandisia Souliei Bonati, 251 On the Baan of the — — lissa Raf. and Platolaria Raf., PALMER, ERNEST 3 Second Supplement to the Spontaneous Flora of the Arnold Arboretum, 410 Panicum agrostoides, 412 41 Papuodendron lepidotum, 422 Parabarium Huaitingii, 245 Paramapania amboinensis, 209 0 — attenuata, 2 — montana, — parvibractea, 209 — si x, 20 Pasania cyrtocarpa, 231 — Garrettiana — Paviei, — vesti Pellaea | auth 311, 316 — hastata, 311 —ternifolia, 311 — viridis, 311 Pentstemon campanulatus subglandulo- sus, Perry, L. M. & E. D. Merritt. Kajew- skiella, a New Rubiaceous Genus from the Solomon Islands, 331 Phaca acutidens, 3 —amoena, —andicola, 370 —araucana, 346 — Bustillosi, 356 — canescens, 350 —carinata, 370 — carrizalensis, 352 — Chamissonis, 346 —chilensis, 346 — chrysantha, 347 1947] INDEX 465 Phaca clandestina, 365 Genera Cupulissa re and, 430 , 404 Platygyne volubili —concinna, 381 a nemor — coquimbensis, 352 —trivialis, 4 — Cruckshanksii, 361 Podostemon qu mao sarap 421 —cryptantha, 365 Polybotrya osmundace 16 — curvicaulis, 352 Polygonum aviculare hich 414 — depauperata, 396 — — arenastrum, — diminutiva, — Bistorta, — dissitiflora, 351 —caespitosum longisetum, 414 — dolichostachya, 351 Polypodium arisanense, 311 — Domeykoana, 388 —aureum, 316 — Edmonstonei, 348 —ellipsoideum, 316 —el — fissidens, 316 — elongata, 381 — loriceum, 316 — flava, — Martensii, 316 — grata, 381 —plesiosorum, 316 — inflata, 353 — Veitchii ere 316 — Landbecki, 361 Polyspora axillaris, 73 — laxiflora, 351 Polystichum amabile, 311 — macrocarpa, 353 —rachichlaena, 316 ee 355 Pomazota Ridley eee A Mono- nan graph of the Genus, boone 356 Pomazota, 195 — ochroleuca, 345 —ser. Monopodiales, 201 — oligantha, 387 —ser. Sympodiales. 196 — oreophila, 358 —assimilis, 194, 198 — orites, 38 — batuensis, 195, 203 — pa —capitata, 194, 199 atagonica, 363 —involucrata, 201 — Pissisi, —pilosa, 194, — podocarpa, 348 —reptans, 194, 196 — pulchella, 366 —rivularis, 194, 201 — quindecimjuga, 409 —scabiosiflora, 195, 202 — Rahmeri, 3 —simalurensis, 194, 197 — Reedi, 357 — sylvestris, 194, 199 — robusta, — Vanleeuweni, 194, 197 — San Romani, 354 Pontederia hastata, 419 — saxifraga, 39 Populus montic 39 striata, Potentilla intermedia, 415 — tricolor, 376 Privet Lodens —triflora, 391 Prunella vulgaris lanceolata, 417 — uspallatensis, 357 Prunus pennsylvanica, 415 — ee Pseudomorus Brunoniana, 428 — verticillata, 349 — — australiana, 428 Phacelia heterophylla sonorensis, 440 — — obtusata, 428 Phillyrea latifolia media pendula, 445 — — pendulina, 427 — media pendula, 445 — — —castaneaefolia scabra, 427 — pendula, 445 — — sandwicensis, 428 Philydrum lanuginosum, 420 — pendulina, 427 Phyllanthus chryseus, —— australiana, 427 Physostegia virginiana, 417 — — obtusata, 428 Pityrogramma calomelanos, 311 —— pendulina, 427 —sulphurea, 31 —-— sandwicensis, 428 Platolaria, 432 — sandwicensis, 428 Platolaria Raf., On the Identity of the Psidium confertum, 123 466 hihi 278 — siam 281 shat sip singalensis, 176 — spec., — tortilis, 162, 164 Pteridium aquilinum, 312 Pteris longifolia, 316 — quadriaurita, 316 Pterygiella nigrescens, 136 Puccinia Lygodii, Pyrus communis, 139 — purpurea ert 254 — ee 324 — dispa —- Garrettiana, 234 — patelliformis, 241 eer — Sini 243 Rajania linearis, 118 — Serkan 329 on Some Culti- V, VI, 253, E Notes o d Trees ory Shrubs, ee and Taxonomy of the Genus Brandisi 7 Restio his 213 Retinis po i Sanderi, 2 — Set Genus oo Studies the Theaceae, XV, Rhododendron any ald alt album, 254 416 — philadelphicus, 415 — rhodinsulanus, 415 — Rosa, 415 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XXVIII Rubus semisetosus, 415 Salix discolor eriocephala, 413 SanpwitH, N. Y. & E. D. Merrity. On the Identity of the Genera Cupulissa Raf. and Platolaria eo Sarco nee den coriacea, Sarosanther — excelsa, Me —lasiopetala, 54 Sax, Harty J. & Karri Sax. The Cyto- senintics of Generic Hybride of Sorbus, ra , Kart & Hatrty J. Sax. The Cyto- panne of nacre Hybrids of Sorbus, 137 Schima Wallichii, 8 Scirpus atrovirens aes 412 4 Scrophularia birmanica, 442 Second Supplement to the Spontaneous Flora of *the Arnold Arboretum, 410 Siderobombyx Brem. Nov. Gen. Rubia- earum Hedyotidearum, 204 Siderobombyx, 204 = kinabaluensts 205 Sreinid t Lago a 33 Silene Arm SMITH, ee . "' H. A, GLeAson. Two New Flowering Plants from St. Lucia, 33 Solanum carolinense, 417 Solidago altissima, 418 —bicolor lutea, 418 —canadensis x juncea, 418 ugosa, 418 — gigantea leiophylla, 418 erula, 418 Solomon Islands, Kajewskiella, a New Rubiaceous Genus from the Sorbaronia alpina, 137, 138 — Arsenii, 137 — Dippelii, 137, 138 — fallax, — heterophylla, 137 —hybrida, 137 1947] INDEX 467 Sorbaronia Jackii, 137 Streblosa multiglandulosa, 160, 168 Sorbopyrus sri, 139 —myriocarpa, 160, 167 — — bulbiformis, 1 — palawanensis, 161, 182 Sorbus, are Cxtogenetics of Generic Hy- ——Elmeri, 162, 18 brids of, —— Merrillii, 162, 183 Sorbus ee 137,138 — platyphylla, 180 — americana, 137, 415 —polyantha, 161, 176 — Aria, 138, 13 — puberula, 164 —hybrida, 139 — scabridula, 161, 176 —scopulina, 138 — tortilis, 160, 162, —sp., 139 — undulata, 160, 170 Spartina pectinata, 412 —urticina, 161, 179 Species Novae Fagacearum Sinicarum, ae 161, 180 320 — — Stapfii, ; Spiraea japonica, 415 — urticina, 181 Spontaneous Flora of the Arnold rae — Wallichii, tum, Second Seeger to the, Streblus Brinn. 427 Stackhousia intermedia, 4 — pendulina, 427 STEARN, WILLIAM T. a tice “Gen- Studies in the Theaceae, ee A Review era Plantarum,” “Iconographia Gene- of the Genus Adinandra rum Plantarum” and “Atakta Botani- Studies in the Theaceae, Seyi Biblio- ka,” 42 graphical Notes on the Genus Laplacea, Sterculia Henryi cuneata, 329 435 — micrantha, 328 Styrax agrestis, 423 — subracemosa, 328 Supplement to the Spontaneous Flora of Sterculiaceae, Notes on Chinese, 328 the / Arnold Arboretum, Secon Streblosa Korthals (Hahacacy. A Mon- Sutherlandia Darumbium, 354 ograph of the Genus, 145 Symplocarpus foetidus, 413 Stcblos, 1 150 Symplocos confusa, 423 r, Costatae, 161, 177 Tectaria irregularis, 316 — ser. Ecostatae, 161, 182 Ternstroemia bancana, 93 —subgen. Eu-streblosa, 159, 162 — coriacea, —subgen. Para-streblosa, 161, 177 —crenulata, 92 — anambasica, 160, —epiphytica, 46 ssimilis, 161, 181 — insignis, 33 —axilliflora, 162, 183 —integerrima, 69, 75 —— angustifolia, 162, 184 — reticulata, 7 — — latifolia, 162, 184 ae 89 — — laxiflora, 162, 184 — anetensis, 46 — bracteata, 160, 166 Thoracostachyum oon 208 — bracteolata, 160, 172 — hypolytroides, 2 —bullata, 160, 171 — montanum, fe —chlamydantha, 161, 177 — pandanophyllum, 208 — deliensis, 160, 173 — parvibractea, 209 — glabra, 161, 180 — simplex 9 — glabra, 183 — subcapitatum, 208 — hirta, 174, 176 Torrenticola queenslandica, 421 —hypomalaca, 161, 174 Torulinium ferax, — Johannis-Winkleri, 160, 169 Tragacantha andina, 35 lampongensis, 160, —arequibensis minima, 390 — lanceolata, 161, 179 — — tenuifoliola, —leiophylla, 161, 175 — atacamensis, —longiscapa, 161, 178 — bella, 367 — maxima, 160, 168 — bisinflata, 353 —microcarpa, 160, 164 — brevialata i — microcarpa, 175 — Cruckshanksii glabrescens, 362 463 JOURNAL OF THE ARNOLD ARBORETUM Tragacantha firma, 355 a, 40 Se 408 — procumbens glabrescens, 388 95 Tropical he saat of Rust Fungi, 309 TSIANG, YIN ON-YOUNG CHUN. A New Tater Plant in South China, 2 oe Novae Fagacearum Sini- carum, 320 Flowering Plants from St. Uredo blechnicola, 310 VAN STEENIS, C. G. Notes on a oe of New Guinean Species, 419 ronica arvensis Viburnum ee praecox, 446 — cassinoides, 41 — dentatum — lantanoides praecox, 445 — — praecox, 445 — Opulus, 417 —plicatum tomentosum, 445 [VOL. XXVAIL Viburnum recognitum, 417 Viola aberrans —abundans, 416 — blanda, ee x papilionacea, 416 416 eolata xX sagittata, 416 Vitex shape — incisa, 2 258 —— multifida, 258 — laciniatus, si — Negundo, ae heerophyl 258 mult em 258 Vitis sestvaliy 416 — ripar a ne — vulpi Wallichia snob 330 Notes on Two ce of w Guinea and a Pro- d New Section of the Genus, 259 iEN S. } ew Plants from Xanthium ee 418 NXerotidae sp., Xyris Jupicai, oe Zanthoxylum americanum, 416