JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY C. E. WOOD, JR. EDITOR FRANCES M. JARRETT LAZELLA SCHWARTEN ASST, EDITOR CIRCULATION VOLUME XL CAMBRIDGE, MASS. 1959 Reprinted with the permission of the Arnold Arboretum of Harvard University KRAUS REPRINT CORPORATION rk 1968 SETS - / ? 1969 ? Yr Ve : Fa 4 g DATES OF ISSUE No. 1 (pp. 1-112) issued January 6, 1959. No. 2 (pp. 118-204) issued April 10, 1959. No.3 (pp. 205-326 issued July 15, 1959. No. 4 (pp. 327-447) issued October 15, 1959. Printed in U.S.A. TABLE OF CONTENTS STUDIES IN ARTOCARPUS AND ALLIED GENERA, I. GENERAL Con- SIDERATIONS. By Frances M. Jarret STUDIES IN ARTOCARPUS AND ALLIED GENERA, II. A REvISION OF PRAINEA. By Frances M. Jarrett A Discussion OF THE Paciric RAILROAD Reports AS ISSUED IN THE Quarto EpitTion. By Susan Delano McKelvey ............ STUDIES IN THE GENUS CoccoLoBa, VI. THE SPECIES FROM THE Lesser ANTILLES, TRINIDAD AND Topaco. By Richard A. Howard THE GENERA OF THE NYMPHAEACEAE AND CERATOPHYLLACEAE IN THE SOUTHEASTERN UNitTep States. By Carroll E. Wood, r STUDIES IN ARTOCARPUS AND ALLIED GENERA, III. A REVISION OF ARTOCARPUS SUBGENUS ArTocarPus. By Frances M. Jarrett A YELLOW-FLOWERED FORM OF RHODODENDRON CAROLINIANUM. By Leonard F. Frisbie ANDRODIOECISM IN THE FLOWERS OF TROCHODENDRON ARALIOIDES. By Hsuan Keng THE EMPETRACEAE AND DIAPENSIACEAE OF THE SOUTHEASTERN Unitep States. By C. FE. Wood, Jr. and R. B. Channell ........ THE EFFECT OF JUVENILITY ON ROOTING OF CUTTINGS FROM APPLE SEEDLINGS. By Donald L. Smith STUDIES IN THE GENUS CoccoLosBa, VII. A Synopsis AND Key To THE SPECIES IN MEXICO AND CENTRAL AMERICA. By Richard A. Howard STUDIES IN THE mab CoccoLosa, VII. A Synopsis AND Kry TO THE SPECIES IN EXICO AND CENTRAL AMERICA oe cluded). By Richard A. Howard THE POLLEN oF EpHepRA. By Margaret Wolfe Steeves and Elso S. Barghoorn ANATOMICAL STUDIES OF BARK REGENERATION FOLLOWING SCORING. By Kyaw Soe an THE GENERA OF THE PRIMULALES OF THE SOUTHEASTERN UNITED States. By R. B. Channell and C. E. Wood, Jr. ............ THE CYTOGENETICS OF FACULTATIVE APOMIXIS IN MALUS SPECIES. By Karl Sax STUDIES IN ARTOCARPUS AND ALLIED GENERA, III. A Revision O TOCARPUS SUBGENUS ARTOCARPUS (Continued). By pide M. Jarrett 38 68 94 161 172 176 205 298 STUDIES IN ARTOCARPUS AND ALLIED GENERA, III. A REVISION OF ARTOCARPUS SUBGENUS ArTocARPUS (Concluded). By Fran- ces M. Jarrett THE GENERA OF OLEACEAE IN THE SOUTHEASTERN UNITED STATES. By Kenneth A. Wilson and Carroll E. Wood, Jr. .............. A ReviseD KryY TO THE CHINESE SPECIES OF JASMINUM. By Clar- ence EF. Kobuski THE GENERA OF PLUMBAGINACEAE OF THE SOUTHEASTERN UNITED SraTes. By R. B. Channell and C. E. Wood, Jr. ................ STUDIES ON THE STIPULES OF SIX SPECIES OF VITACEAE. By J. J. % | Shah THE GENERA OF THEACEAE OF THE SOUTHEASTERN UNITED STATES. By Carroll E. Wood, Jr. THE Dtrectror’s REPORT BIBLIOGRAPHY OF THE PUBLISHED WRITINGS OF THE STAFF AND STupEntTs, JuLy 1, 1958—JuNE 30, 1959 STAFF OF THE ARNOLD ARBORETUM, 1958-1959 InpEx To Vou. XL TITLE-PAGE AND TABLE OF CONTENTS 327 369 385 391 398 413 420 434 437 438 i-lv JOURNAL OF THE ARNOLD ARBORETUM VoL. XL JANUARY 1959 NUMBER 1 STUDIES IN ARTOCARPUS AND ALLIED GENERA,' I. GENERAL CONSIDERATIONS FRANCES M. JARRETT MEMBERS OF THE GENUS Artocarpus are the most commonly encountered representatives of the Moraceae in the lowland forest of the Old World tropics (Africa excepted) apart from the ubiquitous Ficus. The genus also includes two important food-plants, the Breadfruit (A. inmcisus) and the Jack (A. heterophyllus), which now have a circumtropical distribution as cultivated plants, owing to the efforts of man, which include, of course, the well-known voyage of Captain Bligh ending in the mutiny on the Bounty. The fruits of a number of other species are edible, some being planted for the sake of these, and many species provide valuable timber, although they do so in general as scattered members of the forest flora, for the only example of gregarious behaviour is that of Artocarpus incisus (probably little, if at all, modified from the original form) acting as a dominant member of the forests of the river-swamps of New Guinea. The genus Artocarpus is singled out by its compound fruit or syncar- pium, which attains a large size in some species, notably in the two cauliflorous species, the Jack and the Chempedak (A. integer, frequently cultivated in Malaysia). The structure of this fruit has been the subject of special attention in this study with the object of determining the differ- ences between it and the syncarpous fruits of two other small genera of the same area and alliance, Parartocarpus, which has frequently been confused with Artocarpus, and Hullettia, which has proved, somewhat unexpectedly, to show a relationship with Parartocarpus, having previously been placed in a different subfamily of Moraceae. A third small genus, Prainea, which is more closely related to Artocarpus than either of these, although the female inflorescence is not a syncarpium, is also included. It provided another immediate problem, since details of the leaf anatomy had led an earlier author (Renner, 19077) to unite Prainea with Artocarpus, necessitating a re-evaluation of his work and its extension to cover all the species here ‘ This series of papers is based on a thesis presented to the University of Cambridge, D. England, in July 1956 for the degree o * The dates refer to the bibliography at the end of the paper. 2 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL recognized, from which it has been concluded that his action was not The four genera to be included in this study are thus Artocarpus J. R. & G. Forster, which is here recognized as consisting of about fifty species, Prainea King (four species), Parartocarpus Baillon (three species) and Hullettia King (two species). They are lactiferous trees or, less commonly, shrubs restricted in general to the lowland tropical evergreen forest of the Asiatic-Malaysian region although some species may extend to slightly higher altitudes or into areas of deciduous forest. A few of the latter are restricted to regions with a distinct dry season and are themselves either deciduous or evergreen, but the rest are merely tolerant of a short dry season, often adopting the deciduous habit when they grow in such a climate. With an African genus, Treculia Decaisne ex Trécul, which has perhaps six species, they form a group within the Old World Moraceae character- ized by condensed, capitate, unisexual inflorescences which is sufficiently distinct to justify independent study, although it cannot be entirely certain that it represents a natural assemblage of genera. These genera have been assigned to the subfamily Artocarpoideae tribe Artocarpeae and, as will be shown later, constitute the Old World members of this tribe. Although the three subfamilies of the Moraceae — Moroideae, Artocarpoideae and Conocephaloideae — which are distinguished by the position of the ovule and of the stamens in bud, may be valid groups if some exceptions are allowed for, it is not felt that the present classification of the Moraceae offers a sufficiently satisfactory basis for a detailed discussion of the posi- tion within the family of the genera being studied, and there will be given only a brief review of Bentham and Hooker’s classification of the Artocar- peae. Full revisions of Prainea, Artocarpus, Parartocarpus and Hullettia will be given, in this order, in subsequent papers, while in this introductory paper the more general problems of the generic distinctions will be con- sidered. The genus Artocarpus was described by J. R. & G. Forster in 1776 and has been monographed as a whole once before in 1847 by Trécul, who established two subgenera, Jaca and Pseudojaca, which are here main- tained (the former as subg. Artocarpus), on the basis of “alternate” (i.e.. spiral) vs. distichous arrangement of the leaves and amplexicaul vs. non- amplexicaul stipules, as well as on a character of the male flowers which later proved untenable. Parartocarpus was described by Baillon in 1875, but was not very clearly distinguished by him from Artocarpus, and when King monographed the species of Artocarpus in India and the Malay Pen- insula in 1889, he included unawares two species of Parartocarpus, the discordant characters of which led him to reject Trécul’s subgenera. King had also described Prainea and Hullettia in 1888 but, mistaking the posi- tion of the ovule, he assigned them to the Conocephaloideae. In 1902 Beccari published an account entitled ‘“‘Nuove specie di Artocarpeae Malesi e Papuane” in which he pointed out the affinity between Prainea and Arto- carpus and briefly commented on the differences between the latter and 1959 | JARRETT, ARTOCARPUS AND ALLIED GENERA, I 3 Parartocarpus. The final paper of major importance in the classification of the genera is Renner’s study of the leaf anatomy of the Moraceae: Artocarpoideae and Conocephaloideae, published in 1907. Renner re- established Trécul’s subgenera but treated them, together with Prainea, as three sections under Artocarpus, a proceeding which might be justifiable on the basis of leaf anatomy alone, but not if the inflorescence structure is also taken into account. Various observations had thus been made on the differences between Artocarpus, Prainea and Parartocarpus, but the status and distinguishing characters of the genera had not been fully elucidated, nor had the true affinities of Hullettia previously been noted. Although the inflorescence structure of these genera is complex and its morphological derivation is not immediately apparent, no study of the development had been made, apart from a few superficial observations published by Baillon in 1863. In this introductory paper the results of studies carried out in the course of this revision on the morphology and development of the inflorescences are described and their significance is discussed. Other characters that have been found to be of importance at the generic level, namely, seed structure and germination, shoot morphology and leaf anatomy are then considered. Reference is made as necessary to the work of earlier authors, but fuller historical accounts will be found under each genus. Before enter- ing on the more detailed discussions a summary follows of the characters of these four genera, with their resemblances and differences as established in this survey, and a brief review of their position within the Moraceae. The genera included in this study have condensed, unisexual, usually many-flowered, axillary inflorescences. (In Parartocarpus occasional ab- normal inflorescences are bisexual.) The flowers are sessile upon or sunken into a pulvinate, globose, or cylindric, fleshy receptacle bearing interfloral (Prainea and Artocarpus) or involucral (Parartocarpus and Hullettia) bracts, the (upper) surface of the receptacle keing completely covered. They have no rudiments of the organs of the opposite sex. The stamens or ovaries are enclosed in tubular or 2—4-lobed or -partite perianths (Prainea and Artocarpus) which may fuse partially or completely to form a syncar- pium (Artocarpus), or they are sunken in cavities in the receptacle (Parar- tocarpus and Hullettia). If the latter, the anthers or styles are exserted either between indurated processes covering the surface and (?) represent- ing perianths (Parartocarpus), or through simple perforations in the naked receptacle (//ullettia). The male flower has one (Prainea and Artocarpus) , two (Hullettia) or up to three (Parartocarpus) stamens, which are erect in bud, and the female flower has a unilocular ovary with a pendulous ovule and a simple or bifid style. The seeds and embryos are large with little or no endosperm and no period of dormancy. The embryo has either equal or unequal, appressed cotyledons, and is straight (Praimea, Arto- carpus and Hullettia) or curved, the cotyledons being folded just above the base and incumbent (Parartocarpus). Germination (otherwise un- known) is hypogeal in Artocarpus and epigeal in Parartocarpus. The leaves are penninerved (except in Artocarpus altissimus) with microscopic epi- 4 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL dermal gland-hairs providing characters of taxonomic value and (in Prainea and Artocarpus subg. Artocarpus) resin-containing cells in the mesophyll. They are spirally (Artocarpus subg. Artocarpus, Parartocarpus and Hullettia) or alternately and distichously (Prainea and Artocarpus subg. Pseudojaca) arranged with paired or (in Parartocarpus) single stip- ules enclosing the terminal bud (amplexicaul in Artocarpus subg. Arto- carpus, acicular in Hullettia). From this brief account the more important distinguishing characters of the genera may be derived. In Prainea and Artocarpus the stamens and ovaries are enclosed by perianths, usually mixed with interfloral bracts, and a clearly differentiated involucre is lacking. The embryo is straight, or nearly so, and in Artocarpus germination is hypogeal, with no elongation of the hypocotyl. The primary difference between the genera is that in Prainea the perianths in the female inflorescence remain free, but in Arto- carpus they are fused with each other to form a syncarpium, though the fusion usually occurs only in the superficial layer, leaving the perianths free at the level of the ovary. The subgenera of Artocarpus can be distin- guished from each other and from Prainea by the characters of the shoot. In Artocarpus subg. Artocarpus the leaves are spirally arranged on the shoot with large amplexicaul stipules leaving annulate scars, while in subg. Pseudojaca and Prainea they are alternate and distichous with small lateral stipules; but the epidermal gland-hairs of Prainea and Artocarpus subg. Artocarpus have multicellular heads and the mesophyll contains resin-cells, while in Artocarpus subg. Pseudojaca the heads of the gland-hairs are usually unicellular and the mesophyll lacks resin-cells. In Parartocarpus and Hullettia the stamens and ovaries are enclosed in cavities of the receptacle, the walls of which have been shown in the former to be intercalary in origin, and there is a well developed involucre of bracts which are basal in Parartocarpus and vary to an equatorial posi- tion in Hullettia. In Parartocarpus the embryo is curved, having the appressed cotyledons folded just above the base and incumbent, and ger- mination is epigeal, the hypocotyl elongating markedly and carrying the cotyledons well above the ground, where they separate and become green. In this genus the surface of the inflorescence is armoured from indu- rated processes some of which are fused basally around the openings to the cavities in the receptacle. This suggests that they represent perianth segments, but their similarity to the interfloral bracts of the allied but less reduced African genus Treculia indicates that they may be derived from the latter. In Hudllettia the fleshy receptacle is naked and both perianths and interfloral bracts appear to be completely absent. The shoot has spirally arranged leaves in both genera, but in Parartocarpus each leaf has a single triangular, intrapetiolar, non-amplexicaul stipule (distinguishing the genus from species of Artocarpus subg. Artocarpus in which the syn- carpium is superficially similar), whereas in Hullettia the stipules are paired, lateral and acicular. Within the subfamily Artocarpoideae of the Moraceae, which is charac- terized by the erect position of the stamens in bud and the pendulous 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 5 ovule (compared with incurved stamens in the Moroideae and the erect ovule of the Conocephaloideae), these four genera, together with Treculia, represent the Old World members of the tribe Euartocarpeae of Bentham and Hooker (now to be written as Artocarpeae).* These authors also in- cluded Cudrania, which should have been assigned to the Moroideae and will not be considered further, and Balanostreblus, which has been shown (Jarrett, 1958) to be based on a species of the South American genus Sorocea introduced into the Botanic Garden, Calcutta. The classification of the Moraceae is still fundamentally that of Bentham and Hooker (1880) and it may be in some need of revision, since the New World genera of the Artocarpeae do not appear to be closely related to those of the Old World. However, the latter form a fairly well-defined group, characterized by capitate, many-flowered inflorescences in which the flowers completely cover the surface of the receptacle (i.e., without the naked strips found in some American genera), and in which there were probably originally both interfloral bracts and a basal involucre. Treculia, which is the least reduced genus, is the only one to have both interfloral and involucral bracts well developed, as well as perianths enclosing stamens and ovaries. There is also a rudiment of the ovary in the male flower. It is allied with Parartocarpus by the presence of an involucre and the orientation of the embryo. Nevertheless, it must be realized that there is nothing except general similarity to justify the classification of Prainea and Artocarpus with Treculia, Parartocarpus and Hullettia; the superficial re- semblance of the syncarp in Artocarpus and Parartocarpus is due to parallel evolution. The study of this group was suggested to me by Mr. E. 4. Corner, and was carried out under his guidance at the Botany School, Cambridge, England. I should like to record my grateful appreciation of Mr. Corner’s continued and stimulating interest in the research and I am further in- debted to him for the use of his notes made in Malaya on field characters and details of seed structure and germination. In addition, numerous speci- mens of inflorescences preserved in alcohol, which had been collected by him, were loaned for study from the Botanic Gardens, Singapore. These, together with preserved material sent from Kebun Raya, Bogor, Indonesia, and by Mr. T. B. Worthington, Peradeniya, Ceylon, and Mr. G. R. Proctor, Science Museum, the Institute of Jamaica, Kingston, Jamaica, were of great value in studying the structure of the inflorescences. Mr. Chew Wee Lek, now at the Botany School, Cambridge, also generously made avail- able to me inflorescences and young seedlings of Treculia africana Decaisne ex Trécul preserved in alcohol which he had collected in Singapore. Seeds of various species of Artocarpus were kindly sent for germination by the director of the Botanic Gardens, Singapore. * Actually subtribe Euartocarpeae of tribe Artocarpeae of Bentham and Hooker, but the ranks and suffixes of these subdivisions have been altered to accord with the modern practice of treating their “Urticaceae” as consisting of the Ulmaceae, Moraceae, Cannabinaceae and Urticaceae (s.s.). 6 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL The material which has been examined in the course of this revision is that in the following herbaria: Arnold Arboretum, Harvard University, Cambridge, Mass. (A) British Museum (Natural History), London, England. (sm) Herbarium Bogoriense, Bogor, Java, Indonesia. (Bo) Botanical Museum and Herbarium, Copenhagen, Denmark. (c) Indian Botanic Garden, Calcutta, India. (cAL) Botany School, University of Cambridge, England. (ccE) Forest Research Institute and Colleges, Dehra Dun, India. (pp) Herbarium Universitatis Florentinae, Firenze, Italy. (Fr) Gray Herbarium, Harvard University, Cambridge, Mass. (GH) The Herbarium, Royal Botanic Gardens, Kew, England. (x) Rijksherbarium, Leiden, Netherlands. (L) New York Botanical Garden, New York. (ny) Muséum National d'Histoire Naturelle, Laboratoire de Phanérogamie, Paris. ) P Philippine haa Herbarium, Manila, Philippines. (pny) Sarawak Museum, Kuching, Sarawak. (sAR Herbarium of rh Botanic Gardens, Singapore. Botanical Museum and Herbarium, Utrecht, Nethatbandy (vu) U. S. National Museum (Department of Botany), Smithsonian Institution, Washington, D. C. (us) Thanks are due to the directors and staffs of these institutions for the loan of material and also for hospitality at the British Museum (Natural History), The Herbarium, Kew, the Rijksherbarium, Leiden, the Muséum d'Histoire Naturelle, Paris, and the Botanical Museum and Herbarium, Utrecht. The work on which this revision is based was carried out during the tenure of a Maintenance Grant from the Department of Scientific and Industrial Research, the Frank Smart Studentship for Research in Botany in the University of Cambridge, and a Caroline Turle Scholarship of Newn- ham College, Cambridge. A grant was also made by Newnham College toward the expenses of a visit to Leiden and Paris during March and April, 1955. Grateful acknowledgment is made for this support. MORPHOLOGY AND DEVELOPMENT OF THE INFLORESCENCES The morphology of the inflorescences in the genera under review has been studied by dissection and sectioning at anthesis and maturity and by examining different stages in the development where suitable material has been available. Material preserved in alcohol has been used as far as possible and has been supplemented by dried inflorescences, expanded by boiling in water, which can be quite suitable for study if they have been dried quickly (being sliced if they are at all large) and protected from insect attack. The development has not been studied previously except for brief observations on cultivated species of Artocarpus published by cane (1863) and Subba Rao (1940), which will be commented on below 1959 | JARRETT, ARTOCARPUS AND ALLIED GENERA, I 7 The arrangement of the inflorescence in the Urticaceae and Moraceae has been established as basically cymose by Bernbeck (1932) who showed that the dorsiventral inflorescences found in this alliance originate in the suppression of apical growth in a mono- or dichasium, the appearance of the primordia of the flowers and their subtending bracts being delayed, so that they arise on a variously shaped cushion of tissue representing the compressed inflorescence axis, or receptacle. He was able to do this both by making comparative studies in one genus (e.g., Urtica) showing pro- gressive reduction in the inflorescence, and by analysing the arrangement and order of origin of the primordia in genera exhibiting a dorsiventral inflorescence only. The moraceous genera which Bernbeck studied were Dorstenia and Ficus: in the former the dichasial arrangement of the ear- liest primordia on the flattened receptacle was clearly distinguishable, al- though later this became obscured by the simultaneous origin of many primordia, but in the latter, though the first bracts were arranged cymosely, the flower primordia all appeared almost simultaneously. It has been observed in the present study that the inflorescence of Artocar pus incisus shows a similar degree of reduction, the flower primordia in both male and female heads arising almost simultaneously and in no discernible order. The same phenomenon had earlier been noted for A. heterophyllus (as A. integrifolia) by Subba Rao (1940) in a paper on the cytology and embryology of this species. In both these species interfloral bracts are lacking, but in others where they are present (e-g., A, hispidus and A. gomezianus ssp. zeylanicus) they arise and mature a little earlier than the flower primordia, though they bear no definite relation to the latter. In the less reduced inflorescence of Treculia africana, however, the sympodial arrangement of the first few interfloral bracts is evident. These bracts, like those of most species of Artocarpus, terminate in a peltate scale and at the apex of the inflorescence there is always one conspicuously large scale, with a few somewhat smaller ones arranged meridionally and the rest of the surface covered by scales of irregularly decreasing size. The inflorescences of Prainea and Artocarpus, on the one hand, and of Parartocarpus and Hullettia, on the other, have, as has already been indi- cated, a fundamentally different structure and hence they will be consid- ered independently in the following account. The structure of the inflo- rescences is illustrated in figs. 14,4 the species having been chosen both because they were typical and because, in Artocarpus hispidus and Parar- tocarpus forbesii, inflorescences preserved in alcohol were available. In both male and female inflorescences of Prainea the perianths are tubular and free from each other, and are mixed with peltate to spathulate. long-stalked interfloral bracts. The male perianth is perforate or bilobed above and the apex of the female perianth is thick-walled and clavate, with a contracted mouth which is either perforate or very shortly 2—4-lobed. The female inflorescence of P. papuana is illustrated in fig. 1, a—e and the male inflorescence in fig. 3, a-c. As the female inflorescence ripens only * For convenience of reference the figures will be numbered consecutively throughout this series of papers. 8 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL those perianths in which seeds are formed enlarge markedly. These become fleshy and appear as conical to globose projections from the general sur- face of the inflorescence, which is formed by the apices of the unexpanded flowers and the heads of the bracts, giving the typical appearance shown in fig. 1, c. Beccari (1902), in recognizing the affinities of Prainea with Artocarpus, stated that the only difference between the genera lay in the structure of the female inflorescence, but this is only correct with respect to the reproductive organs. Fic. 1. The female inflorescence in Prainea and Artocarpus. papuana: a, b, longitudinal section and entire head at anthesis (« 5 d (X %); d, fruiting perianth in longitudinal section (x ne } (X 1). f-j, Artocarpus hispidus: {-h, longitudinal section, Mere view and tangential section in plane u-v at anthesis (x 10); i, longitudinal section at maturity (X 12); j, part of the same (x 1). k, A. incisus, embrvo (X 1). Conventions used in figs. 1-3: ovaries and stamens are shown in black and receptacular tissue is crosshatched when cut in section. Abbreviations: 1, inter- floral bract; 2, perianth; ee perianth; 4, ovary; 5, testa; 6. unexpanded perianth; 7, involucral bra In the inflorescences of Artocarpus the flowers are also usually mixed with interfloral bracts, but these are frequently shed from the syncarp at or before anthesis and in some species they are altogether lacking. The nerianths in the male inflorescence are likewise free from each other, vary- ing from tubular to 2—4-partite, but in the female inflorescence the peri- 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 9 anths, which are tubular with a contracted and perforate apex, are fused with each other to form the characteristic syncarp from an early stage in development. This fusion is, however, only partial in most species and is of an unusual nature. Each perianth consists of a proximal region which is thin-walled and encloses the ovary, and a distal region with thickened walls and a narrow lumen through which the style passes to the exterior. Proximally the perianths remain free from each other, but distally they are connate, forming a continuous layer evidently capable of withstanding a considerable degree of tension as the syncarp matures. The perianths may _ 2. The female inflorescence in Parartocarpus and Hullettia. a—f, Parar- tocarpus forbesii: a-c, longitudinal section, surface view and tangential section in plane w-x at anthesis (X 5); d, longitudinal section at maturity (x 1%); e, part of the same (x 1); f, embryo (X 1). g-k, Hullettia dumosa: gH, longitudinal section, surface view and tangential section in plane y—z at anthesis 5); j, mature head (X '%); k, the same in longitudinal section (X 1, approx. ). 10 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL be completely fused distally to form a smooth covering layer to the syn- carp with perforations through which the styles are exserted at anthesis, or their apices may remain free and project as cylindric, conical or obtuse processes on the surface. In Artocarpus (subg. Artocarpus) hispidus, which is illustrated in fig. 7, f-j, the distal parts of the perianths are not com- pletely fused to each other and the syncarp is echinate from their rigid, free apices. The longitudinal section of the female inflorescence at anthesis in fig. 1, f shows one perianth cut sagittally, while those on either side are cut in a tangential plane, and in those on the extreme right and left the proximal region and the free apices are seen in face view, the plane of the section having passed between neighbouring perianths. In contrast to Prainea the entire syncarp usually enlarges more or less uniformly although relatively few of the flowers set seed. In the proximal layer, however, where the perianths are free, only those containing seeds enlarge, and the others appear as narrow ribbons in the mature syncarp, as shown in fig. 7, j. In this and several other species of subgenus Artocarpus, including the Jack (A. heterophyllus), the proximal region of the fruiting perianths be- comes fleshy and provides the edible portion of the syncarp, in addition to the seeds, which are often roasted, though in the seedless form of the Breadfruit (A. incisus) the entire head is eaten. The syncarp has been examined in nearly all the species of Artocarpus subg. Artocarpus recognized in this study and the perianths have been found to be consistently free from each other at the level of the ovaries. The same structure is found in many species of subg. Pseudojaca, but in others the fusion between perianths is complete, although each perianth may still be distinguishable on the surface as an areole or papilla. In this subgenus the entire syncarp is fleshy and is eaten in a number of species. It has been possible to examine very young inflorescences of Artocarpus rigidus, a species closely allied to A. hispidus, and the longitudinal section illustrated in fig. 4, a@ shows that the perianths are at first entirely free from each other. The later fusion must be due to the pressure exerted on each other by the thick-walled distal portions of the very closely set pri- mordia. The thinner-walled proximal portions, having less resistance to pressure, would remain free. The partial fusion of the perianths was described and illustrated for Artocarpus heterophyllus (under the name Sitodium cauliflorum) by Gaert- ner in 1788 and again in 1939 by Corner, writing on this species and on Artocarpus integer, but other authors have assumed that the ovaries were embedded in a solid mass of tissue. Nearly all have regarded this tissue as representing the perianths, but Baillon, stated (1863) that the tissue surrounding the ovaries must be axial in origin, on the grounds that the belated fusion of neighbouring parts is very rare, as is indeed the case. He was able to study material of Artocarpus incisus preserved in alcohol, but, nevertheless, observing the very young pistils each in the centre of a ring-like emergence, he described them as becoming gradually more deeply sunken in pits of the receptacle and failed to notice the more com- ple, adult structure. 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 11 In Prainea the male and female inflorescences are more or less globose and are not always easily distinguishable before anthesis, but in Artocarpus there is a considerable range of variation especially in the male head, and the two sexes usually differ markedly in size and shape. In neither genus is there a clearly differentiated involucre, although there may be few sessile bracts similar to the interfloral bracts at the base of the head. In Arto- carpus there may be a slight enlargement of the top of the peduncle and in A. heterophyllus this forms a distinct annulus. Fic. 3. The male inflorescence. a—c, Prainea papuana: a, entire head (X 1); surface view ope ee section (X 5). d-f, Artocarpus hispidus: d, entire head (X* 1); e, f, ee view and longitudinal section (X 20). gH, Parartocarpus forbesit: eS e head (x 1); h, i, surface view and longitudinal section (X 5). j-l, Hullettio pay ne k, entire head (x 2); 1, j, surface view and aa nel section ( 5) In Parartocarpus the inflorescences of the two sexes are similar to each other in structure, with both stamens and ovaries contained in cavities in the apparently continuous tissue of the capitula. Careful examination has revealed no sign of fusion having taken place between neighbouring peri- anths in the walls separating the cavities and developmental studies show that they are intercalary in origin. Details of the structure in Parartocar- pus forbesii are shown for os female inflorescence in fig. 2, a—f and for the male inflorescence in fig. 3, g-4; comparison of the tangential section of the female inflorescence at the level of the ovaries (fig. 2, c) with that of Artocarpus hispidus (fig. 1, ) will show the difference between the 12 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL syncarps of the genera. The surface of the receptacle in Parartocarpus is closely covered by indurated processes, which are acutely conical or spinous in P. forbesii. Some are arranged in groups of two to four around the openings of the cavities containing the stamens or ovaries, and are fused basally to each other around the opening. The rest, which are identical in appearance, are free from each other and cover the intervening surface, bearing no relation to a cavity. The processes surrounding the openings are fused to each other only at the base at anthesis, but, since they grow from below, the fusion becomes more marked during the ripening of the syncarp. In spite of its different origin, this syncarp is superficially very similar to that of some species of Artocarpus in which the perianth apices are indurated. Wah Ey Ie Ga) eal <£3\(¢>\ i ||| Fic. 4. Details of inflorescence structure. a, b, Longitudinal sections of young female inflorescences in A. rigidus (X 12.5) and A. styracifolius (X 2.5). c¢. d, Transverse sections through bases of perianths in female inflorescence of Parar- tocarpus venenosus at anthesis and in male inflorescence of P. forbesii before anthesis ( 5). e-h, Development of male head in Parartocarpus bracteatus: e, f, longitudinal and tangential sections from head 7 mm. in diameter (* 10) g, h, the same from head 25 mm. in diameter (x 5). 4\ €\ ¢ ay The development of the inflorescences is illustrated by the sections, both longitudinal and tangential at the level of the stamens, through male heads of Parartocarpus bracteatus in fig. 4, e-h. In this species the processes are spinous and the considerable enlargement of the cavities relative to the processes during their growth is evident on comparison of the figures. In the younger inflorescence (diameter of the head 7 mm.) the stamens are merely finger-like projections, while in the older one (diameter 25 mm.) they are nearly at anthesis. It has not been possible to examine any younger stage but it is probably justifiable to assume that the stamens arise between the processes on the surface of the receptacle and are only later enclosed within cavities. The intercalary origin of the walls is con- firmed by the vertical lines of cells that are seen in longitudinal section. The processes are indurated from an early stage in development and can 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 13 therefore grow only at the base, while they must also enlarge laterally to cover the increasing surface area of the receptacle. It has not been possible to determine whether the processes surrounding each cavity are originally free, or whether they arise from a ring of tissue surrounding the androe- cium or gynoecium. Varying opinions have been expressed as to the identity of the processes. In view of the similarity of the “fertile” and “sterile” processes it may be assumed that they are all equivalent; thus they presumably represent either perianths or interfloral bracts. Baillon, who was working with limited material of the male inflorescence, inaccurately stated in the orig- inal description of Parartocarpus beccarianus (1875) (= P. venenosus (Zoll. & Mor.) Becc.) that the stamens were intermingled with stalked bracts (the processes on the surface representing their heads), and hence supposed that the difference between this and the male inflorescence of Artocarpus lay in the absence of perianths. Boerlage (1897), who de- scribed the same species independently under the name Gymnartocarpus venenosa (based on Artocarpus venenosa Zollinger & Moritzi) likewise distinguished it from Artocarpus by the absence of perianths, but he de- scribed the stamens and ovaries correctly as sunken in cavities of the receptacle, regarding the processes alone as representing bracteoles. He failed to observe the fusion of the processes grouped around each cavity, which is not conspicuous in this species, since they are short and often truncate. The fusion is shown in the tangential section through the bases of the processes in fig. 4, c. Beccari, in discussing the inflorescence struc- ture of Artocarpus and Parartocertus in 1902, was the first to note the fusion of some of the processes in the latter. He considered that they represented the apical lobes of tubular perianths, which were completely fused to each other below, and that the free processes represented the solid apices of sterile perianths. While he realised that the processes thus differed from the tubular perforate perianth apices of Artocarpus he did not observe any difference in the internal structure. Although the fusion of the processes around each cavity indicates that Beccari’s interpretation is correct, apart from the fact that the tissue en- closing the stamens and ovaries is intercalary in origin and only the proc- esses themselves can be regarded as representing the perianths (and even in these the base is also intercalary), the allied African genus Treculia must here be considered, since its less reduced condition may throw light on the structure of Parartocarpus. In Treculia africana, well differentiated perianths, interfloral bracts and basal involucre are all present (illustrated in Engler, 1898, ¢. 12-14). The stalks of the abundant interfloral bracts are fused for about half their length, so that the flowers are enclosed in cavities between them. That this tissue represents the stalks of the bracts is shown by its continuity with the free distal portions of the stalks and the abrupt transition to receptacular tissue below the level of the flowers. In the male inflores- cences the flowers consist of (2—)3—4(—5) stamens enclosed in a mem- branous tubular perianth. In the female inflorescence each ovary may be 14 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL accompanied by up to four linear perianth segments, and numerous abor- tive male flowers are also present. In this species each bract is terminated by a peltate scale, but in other species of the same genus the shape of the heads of the bracts is various. In Treculia zenkeri Engler (l.c., ¢t. 15) the bracts have some peltate and some capitate heads, while in the male head of Treculia obovoidea N.E.Br. (as examined in Zenker 2526) the bracts terminate in minute spinous processes surrounding the cavities containing the flowers and bluntly conical processes covering the intervening surface. Hutchinson (1917) considered that Acanthotreculia winkleri Engler (1908) probably represented the female inflorescence of the latter species and in this the free spinous processes surrounding each emerging style were de- scribed as conspicuous and strongly differentiated from the other short and obtuse processes. The processes in both Treculia zenkeri and T. obovoidea have a general similarity to those on the surface of the inflorescences of Parartocarpus and the possibility must be considered that in the latter the bracts have secondarily assumed the function of perianth segments, those surrounding the openings of the receptacular cavities becoming par- tially fused to give added protection; that perianths containing the stamens and ovaries have been lost; and that walls of intercalary origin enclosing the latter have replaced the stalks of the now sessile bracts. The inflorescence structure of Hullettia is fundamentally similar to that of Parartocarpus, although interfloral bracts and perianths are apparently entirely lacking and the fleshy, pubescent receptacle is naked, having per- forations in the surface leading to cavities containing either stamens or ovaries. The female inflorescence of Hullettia dumosa is illustrated in fig. 2, g-k and the male inflorescence of H. griffithiana in fig. 3, j-l. No material has been available for studies in development, and the homology of the structure with that of Parartocarpus is inferred from examination of the inflorescences at anthesis and maturity and from other characters which the genera have in common. The most immediately obvious of these similarities is that in Hudlettia, as in Parartocarpus and Treculia, there is a clearly differentiated involucre. In Parartocarpus this consists of three to eight basal bracts which are tri- angular to ovate and indurated or thickly coriaceous. In Hullettia there are three to six fleshy flattened triangular projections which at anthesis are marginal on the obturbinate or pulvinate receptacle and on the sub- globose mature syncarp vary ‘from a basal to an equatorial position. In both genera the bracts may be somewhat obscured, especially on the mature syncarp of Hullettia. In Parartocarpus the inflorescences are globose or nearly so, like those of Prainea, and the male and female inflorescences can only be identified before anthesis by dissection, but in Hudllettia it is only the very young inflorescences that are indistinguishable externally. Details of the flowers have not been included in this discussion of the generic characters of the inflorescences, but those of the ovary will be considered in the following section. In Prainea and Artocarpus there is little variation in the male flowers and in distinguishing the species the entire inflorescence is the unit of variation. However, the male perianth 1959 | JARRETT, ARTOCARPUS AND ALLIED GENERA, I 15 in Artocarpus subg. Artocarpus is nearly always tubular and bilobed above, whereas in subg. Pseudojaca it is frequently 2—4-partite, and, in the former, slight differences in the size of the anthers are of some assistance in estab- lishing the subdivisions. There are also small differences in the stamens of the various species of Parartocarpus and Hullettia. In conclusion, some aspects of the variability of the inflorescence within Artocar pus, which i is in marked contrast to its relative uniformity within the other genera, will be commented on. In the syncarp, as in the male inflorescence, there is some overlapping in the ner of the two sub- genera, so that they are most readily defined by the characters of the shoot. However, in subgenus Pseudojaca the syncarp is always more or less globose, with a surface varying from smooth to tuberculate, but never (except in A. styracifolius, discussed below) bearing distinctly elongate processes. In subgenus Artocarpus, on the other hand, the syncarp is al- ways covered (at least at anthesis) by distinct processes, except in A. sepi- canus and A. kemando, and in these, as in a majority of the other species, the syncarp is ellipsoid to cylindric. The variety in the shape of the processes on the syncarp in Artocarpus subg. Artocarpus is considerable and will be illustrated in a plate accom- panying the key to the subgenus. In a few species solid sterile perianths are present with elongate apices projecting markedly beyond those of the fertile perianths; in A. multifidus, A. elasticus, A. sericicarpus and A. tamaran the dimorphism is complete, but in A. teysmannii intermediates occur, all except the longest apices being perforate. This feature has either been unnoticed or its significance has been unrecognized by earlier au- thors, Elongate sterile perianths also occur in the male inflorescences of the first and last of these species. That these elongate processes represent sterile perianths and not enlarged interfloral bracts is shown by the find- ing of the latter in the same inflorescences, and by the close similarity of the perforate and solid processes in the syncarp. In subgenus Pseudojaca, on the other hand, the flexuous, solid, terete processes that cover the surface of the syncarp of A. styracifolius, and be- tween which the styles emerge through perforations in the surface, appear to be enlarged interfloral bracts. They have a single longitudinal vascular bundle, and in the young syncarp they look very similar to the clavate interfloral bracts found in the male inflorescence. A longitudinal section at this stage is shown in Fig. 4, bd. STRUCTURE OF THE SEED AND ITS GERMINATION These genera are characterized by large seeds which have little or no en- dosperm and no period of dormancy or ability to withstand desiccation. Differences in the shape and orientation of the embryo and in the mode of germination confirm alliances indicated by the structure of the inflores- cences. The ovule is pendulous at anthesis, being attached at or a little below the apex of the ovary, and the style is terminal or nearly so, but in the mature ovary the position of the style and the attachment of the seed 16 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL vary from apical to sub-basal, owing to differential growth in the walls of the ovary. In Prainea papuana, the only species of the genus of which material was available for detailed study, the style is strictly terminal throughout, but, while at anthesis the ovule is attached laterally slightly below the apex of the ovary, at maturity it is sub-basal. The mature ovary is mem- branous and the testa is totally absorbed except for a thickened basal saucer-shaped region, at one side of which is the hilum. The embryo, which is illustrated in fig. 1, e, is symmetrical, with two equal cotyledons having their long axes parallel to the median plane of the ovary but their appressed faces at an angle of about 45° to it. The radicle and plumule are small and apical. In Artocarpus there is great variation in the ovary and embryo and also in the consistency and relative thickness of the fruiting perianth, ovary and testa. In subgenus Artocarpus the terminal or lateral position of the style in the mature ovary delimits two natural groups of species. The seed is attached just below the base of the style and, since the radicle is usually directed approximately at the hilum, in the first group the embryo is longitudinally aligned and the radicle is apical, whereas in the second the long axis of the embryo is more or less oblique and the radicle is ventral. In the first group the appressed faces of the cotyledons are at an angle of 30-90° to the median plane of the ovary and the radicle and plumule are apical and minute so that the appearance of the embryo and its orien- tation relative to the ovary are similar to Prainea, although the hilum is apical instead of sub-basal. In these species the inner wall of the ovary forms a horny endocarp, while the exocarp decays leaving the “seed” free within the proximal region of the perianth which may be thickened and fleshy. In species in which the style is lateral or sub-basal in the mature ovary and the long axis of the embryo is hence more or less oblique, the appressed faces of the cotyledons may be in the median plane of the ovary (in which case the embryo is usually symmetrical), or they may also be oblique or at right angles to the median plane (in which case one cotyledon lies above the other in the seed and is frequently reduced in size). Such an embryo is illustrated in fig. 1, k. The variation in this second group will be discussed in more detail in considering the subdivi- sion of subgenus Artocarpus. In this group the cotyledonary petioles and the radicle lying between them are relatively well developed and, if the latter is large enough, it is curved downward. The chalaza is always basal, with a conspicuous ventral raphe, so that in the first group the embryo is anatropous, whereas in the second it is campylotropous. In Artocarpus subg. Pseudojaca the style is terminal or nearly so and the embryo is longitudinally aligned, with the appressed faces of the equal cotyledons in the median plane of the ovary or at varying angles to it. In some species of this subgenus, also, the seed is enclosed by an indurated endocarp which lies free in the cavity of the perianth at maturity The primary difference between the embryo in these genera and in Parartocarpus is that in the latter it is folded. The ovary has a terminal 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 17 style at anthesis, but at maturity it is lateral and the hilum lies just above it. The ovary wall differentiates into a bony endocarp and an exocarp which finally decays. The testa is attached to an indurated hilar plug and is thin-walled, with a relatively small apical vascular cap. The embryo fills it almost completely but, unlike Artocarpus, there are distinct remains of the endosperm in the region of the hilum. The markedly unequal cotyledons are appressed to each other and are incumbent, being folded transversely a little above the base so that the large, straight radicle lies along the median line of the inner and smaller cotyledon. The plane of the appressed faces of the cotyledons is longitudinal and at right angles to the median plane of the ovary. The tips of the cotyledons are enclosed in the vascular cap of the testa, which makes a distinct indentation in them, and the radicle lies on the side of the seed facing the style, its tip being directed upward to the hilum and enclosed in a pocket of endosperm. The embryo of P. forbesii is shown in fig. 2, f. The same type of embryo occurs in Treculia (africana) and has been illustrated by Trécul (1847) and Engler (1898) (but in Acanthotreculia Engler (1908) the embryo is shown with very unequal, straight cotyledons and a small, apical radicle). The embryo of Hullettia, on the other hand, is straight, with equal or slightly unequal cotyledons arranged longitudinally; the small radicle, however, is basal. The ovary is pergamaceous and the testa has a shallow, thickened, apical cap (not shown in fig. 2, k) attached centrally, imme- diately below the subterminal style, and making an indentation in the tips of the cotyledons, but is otherwise thinly membranous. Thus, whereas in Prainea and Artocarpus the chalazal region is basal, in Parartocarpus and Hullettia it is apical (relative to the ovary), and since the tips of the cotyledons are in contact with this region, except in some species of Arto- carpus, the orientation of the embryo with respect to the ovary is also reversed. It has been possible to study germination in the following species of Artocarpus: (subg. Artocarpus) A. rigidus, A. integer, and A. heterophyl- lus, and (subg. Pseudojaca) A. borneensis. In addition I have been able to use Mr. Corner’s notes on germination in Artocarpus (subg. Artocarpus) anisophyllus, A. lanceifolius and A. elasticus and Parartocarpus bracteatus and P. forbesii, and to examine herbarium specimens of seedlings of Arto- carpus incisus and A. nobilis. Germination in Artocarpus and Parartocarpus differs chiefly in whether or not there is elongation of the hypocotyl. In Artocarpus there is no elongation, and germination is hence hypogeal. The cotyledons usually remain enclosed within the testa and the plumule is enabled to emerge from between them by the elongation of their petioles (illustrated for A, hirsutus and A. chaplasha in Troup, 1921, figs. 323, 325). In A. hetero- phAyllus (l.c., fig. 324, and Gaertner, 1788) and A. integer, however, the cotyledons separate to allow the emergence of the plumule. A further variation occurs in subgenus Artocarpus in the nature and arrangement of the first leaves borne by the seedling. The first pair subsequent to the seed leaves may be opposite and fully developed and be followed by spirally 18 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL arranged leaves (A. anisophyllus, A. lanceifolius, A. rigidus, A. chaplasha, A, hirsutus and A. nobilis) or the leaves may be spirally arranged from the first, the youngest being scale leaves (A. incisus, A. heterophyllus, A. in- teger and A. elasticus). The two types correspond, but only approximately, to the species groups distinguished above by the terminal and lateral posi- tion, respectively, of the style, and will likewise be discussed in more detail under subgenus Artocarpus. In Artocarpus subg. Pseudojaca the seedling bears scale leaves before the fully developed leaves, as in the second group above, but in Artocarpus borneensis (ssp. griffithiz) these are strictly alternate and distichous from the first. It is not clear from Troup’s illustration (l.c., fig. 326) of A. lakoocha whether the leaves are spirally or distichously arranged. Photographs of seedlings of A. rigidus and A. borneensis are reproduced in fig. 5 In Parartocarpus germination is epigeal and the hypocotyl elongates markedly, carrying the folded cotyledons well above the ground. As they straighten and open out, becoming green and photosynthetic, the endocarp and testa are split off. The plumule grows out from between them and the first pair of leaves is opposite, with paired stipules, but the later leaves are spirally arranged, each with a single stipule. The seedlings of Treculia africana show an exactly similar mode of germinaton. MORPHOLOGY OF THE SHOOT The morphological characters of the shoot which are of importance in distinguishing the genera under review are principally those provided by the arrangement of the leaves and by the stipules, which enclose the ter- minal bud and are either intrapetiolar or lateral to the base of the petiole, being paired except in Parartocarpus. They are illustrated in fig. 6. It is in these characters that the most easily discernible distinctions are to be found between the two subgenera of Artocarpus. In subg. Artocarpus the leaves are spirally arranged and the stipules are intrapetiolar and amplexi- caul, leaving annulate scars, whereas in subg. Pseudojaca the leaves are strictly alternate and distichous and the stipules are non-amplexicaul, leav- ing lateral or partially intrapetiolar scars. The latter arrangement is also found in Prainea but the leaves of this genus can usually be recognized by the abruptly and narrowly acuminate apex and the conspicuous and regular inarching of the lateral veins (in addition to details of the an- atomy). In Parartocarpus the leaves are spirally arranged and in the axil of each is a single non-amplexicaul stipule which forms a scar extending slightly to either side of the petiolar scar. J. J. Smith (1922) was the first author to realize that these stipules were single and not paired. In the seedling, as already mentioned, they may be paired or have a bifid apex, indicating that each has originated from two stipules fused along their intrapetiolar margins. The leaf arrangement, together with the non- amplexicaul stipular scars, distinguishes the genus from both the preceding ones. The leaves of Hullettia are also spirally arranged but the stipules . 5. Seedlings of Artocarpus; left, Artocarpus (subg. Artocarpus) rigidus; neensis; insets, details of stipules. Pie “4 right, Artocarpus (subg. Pseudojaca) bor- [6S6I i. “LLAWUVE a 4 SNdAVIOLAYN 7 VYANAD GHITIV UNV I 20 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL are distinctive, being linear, acute and frequently persistent. They are lateral to the leaf-axils and the scars are small and round. Differences in the arrangement of the foliage are associated with the two types of phyllotaxy. In the groups with spirally arranged leaves the ultimate shoots and usually also the main branches are ascending. In Parartocarpus and Hullettia the leaves are markedly clustered at the tips of the shoots and there is greater elongation of the internodes at the base of each lateral branch, so that one leaf cluster is separated from the next. Fic. 6. Details of stipules; a, Artocarpus sericicarpus (X 1); b, c, A. gome- zianus ssp. seylanicus (X 1); d, terminal bud of the same ( 2); e, Pararto- carpus bracteatus (X 1), f, terminal bud of the same ( 2); g, Hullettic dumosa (X 1), In the groups with alternate leaves, on the other hand, at least the ulti- mate shoots are more or less applanate. As already mentioned, study of the seedlings of species belonging to both subgenera of Artocarpus has shown that the difference in leaf arrangement may exist from the earliest stage Leaf form in Artocarpus is variable, ranging from simple to pinnatifid. In most species the adult leaf is simple, but in one group within subg. 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 21 Artocarpus it: is frequently pinnatifid, while in another species, A. aniso- phyllus, it is pinnate. .In many other species of this subgenus the juvenile leaves are more or less pinnatifid and sometimes very large — to six feet in length in A. elasticus (Corner, 1940). Artocarpus tamaran is remark- able in having a nearly pinnate juvenile leaf, distinguished from that of A, anisophyllus by a continuous wing of lamina along the rachis. In some species of subg. Pseudojaca the juvenile leaves are irregularly pinnatifid or the lamina may be reduced to a sinuous wing along the midrib, but in the rest, as in the other genera, the leaf is always entire and the juvenile leaf is merely elongate. Juvenile leaf-forms are shown in fig. 7. As noted by Corner (1940) the pinnate leaves of Artocarpus anisophyllus are un- usual in falling as a whole instead of disarticulating at the base of each pinna. Fic. 7. Juvenile leaves in Artocarpus; a, A. odoratissimus; b, A. tamaran; c, d, A. dadah (X 1/7). In the field Artocarpus and Parartocarpus are easily distinguished by characters of the bark. In Artocarpus the bark on the roots is red and that on the trunk peels off in flakes. In Parartocarpus the bark of the roots is yellow, while the trunk is covered by numerous large and prominent lenti- cels (conspicuous in the photograph in Browne, 1955). LEAF ANATOMY A detailed account of the leaf anatomy of the Artocarpoideae and Conocephaloideae was published by Renner in 1907. From this it may be concluded that the most important anatomical characters distinguishing the genera under consideration are those of the microscopic glandular epi- dermal hairs and of the spongy mesophyll. Renner considered that the fact that leaf anatomy of Prainea was in some ways intermediate to that of the two subgenera of Artocarpus necessitated the unification of the two genera, but this view is here rejected as a misinterpretation of the vegeta- tive characters. 22 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Renner’s observations have been repeated in the course of this revision and expanded to cover all the species here recognized. They have in general been confirmed, but in the following discussion a few points of dis- agreement are indicated. The leaves were examined by means of surface and transverse sections cut from fragments of herbarium specimens boiled up in water, or from material preserved in alcohol when available. The sections were mounted in a solution of polyvinyl alcohol and lactic acid (Metcalfe and Richardson, 1950). In addition to normal epidermal hairs the Artocarpoideae have gland- hairs on or sunken into the surface of the leaf, with a unicellular stalk and a variously shaped head. Renner stated that he was unable to find any anatomical characters distinguishing the tribes of the Artocarpoideae, but this may have been due in part to the fact that some genera were misplaced or studied from misidentified specimens. However, the Old World genera of Artocarpeae do show a wide range of variation in the gland- hairs, especially within the genus Artocarpus. In Prainea and Artocarpus subg. Artocarpus the heads of the gland- hairs are 4—16-celled with a thick-walled stalk. In the former they are always and in the latter they are frequently sunken in pits in the epider- mis. In Prainea the heads are depressed-globose and divided by vertical and horizontal walls into four to eight cells. In Artocarpus subg. Arto- carpus the heads are either globose or depressed-globose with one or two tiers (rarely four) of four to six cells, or they are flattened and peltate with eight or rarely sixteen cells. An account of the gland-hairs is included in each specific description for this subgenus. In several species there is a hypodermis of cells which are either isodiametric or elongate in surface view. Both the gland-hairs and the hypodermis provide characters of taxonomic value and will be discussed further in considering the classifica- tion of the species of the subgenus. In most species of Artocarpus subg. Pseudojaca, on the other hand, the gland-hairs have a unicellular globose head and a thin-walled stalk. Renner found the head to be 2—4-celled in Artocarpus lakoocha and in this study it also has been found to be 1—6-celled in A. tonkinensis. The gland-hairs of Parartocarpus likewise have a thin-walled stalk and a unicellular head, but the latter is markedly elongate and thus cylindric. In Hullettia, which was omitted by Renner as being of doubtful affinities, the gland-hairs have a thick-walled stalk and a globose unicellular head. According to Renner the gland-hairs of Treculia have an elongate head often divided by a longitudinal wall. In none of these are the gland-hairs sunken, except for a few species of Artocarpus subg. Pseudojaca in which they are only slightly so, and in none is there a hypodermis. In the characters of the spongy mesophyll the two subgenera of Arto- carpus again differ widely, while Prainea occupies a somewhat intermediate position, All the genera being studied have a bifacial leaf with one or two rows of palisade cells, and a well developed spongy mesophyll. This is more or less compact in Artocarpus subg. Pseudojaca, Parartocarpus and Hul- lettia, but in Artocarpus subg. Artocarpus it is extremely loose, consisting 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 23 of a mesh of long-armed (‘‘hyphenartig’’) cells and including some con- spicuous ellipsoid or globose cells with resinous contents. The resin cells are absent in A. heterophyllus and A. integer, but not, as stated by Renner, in A. odoratissimus (as A. mutabilis Becc.). In Prainea the spongy meso- phyll is rather loose with relatively large cells nearly all having slightly resinous contents. 0 WY, iS ae Eo fixe (inal — e\9) ae g IBY (SQ KS 2) Crea ey NO Fic. 8. Details of leaf anatomy. a-i, Epidermal glands: a—c, Prainea limpato, Artocarpus odoratissimus and A. lancetfolius i aa section (the last also showing hypodermis); d, e, A. treculianus in longitudinal section and sur- face view; 4, imteger in longitudinal section; g, A. . in surface view; h. i, Parartocarpus venenosus and Hullettia dumosa in longitudinal section. j. k. Enlarged cape cells in Hullettia dumosa in surface view and longitudi- nal section. 1, m, Spongy mesophyll in Prainea frutescens and Artocarpus ri- gidus in longitudinal section (the latter showing resin cell in contact with palisade tissue) (> 200, approx.). = * On the grounds that, firstly, Prainea showed more similarity in ana- tomical structure to Artocarpus subg. Artocarpus than the latter did to subg. Pseudojaca, and that, secondly, according to Beccari (1902) the only difference between the genera lay in the free or fused female peri- anths, Renner treated all three groups as sections under Artocarpus. How- ever, the mesophyll of Prainea could well be regarded as a type from which the two subgenera of Artocarpus have diverged, while subg. Arto- carpus and subg. Pseudojaca may show an advance and a reduction re- spectively in the development of the gland-hairs. In leaf and _ stipular arrangement Prainea resembles Artocarpus subg. Pseudojaca. The inter- relations between the three groups thus appear reticulate, indicating that parallel evolution has occurred. This is least likely to have occurred in the unique sycarpium of Artocarpus and most probable in the develop- ment of distichous leaves and non-amplexicaul stipules. It is therefore concluded that Prainea and Artocarpus should be retained as distinct genera and that the anatomical characters of the latter have probably arisen 24 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL by divergence from a common ancestor with Prainea. The anatomical characters, in conjunction with the difference in leaf and stipular arrange- ment, provide the best evidence that the subgenera of Artocarpus represent natural groups. In their indumentum (apart from the gland-hairs) all the genera are variable, except for Prainea, in which the leaves are glabrous. The hairs frequently have enlarged bases and papillate walls, though, as was noted by Renner, there may be a few hairs with hooked tips and these are always smooth walled. In Prainea there are likewise enlarged cells in the epidermis which may be produced into a very short point. In Artocarpus the hairs are usually restricted to the veins, but in some species of subg. Pseudo jaca nearly all the cells of the areolae grow out to form thin-walled, often crisped hairs, so that the lower surface appears greyish-glaucous from the minute tomentum. In Hullettia there are numerous enlarged cells with strongly thickened and narrowly pitted walls in the lower epidermis, which bear hairs of varying length and cause the lower surface to appear minutely punctate under a lens. GENERAL CONCLUSIONS The conclusions reached in the foregoing discussion are summarized below in the order in which the characters were taken up. The inflorescences of Prainea and Artocarpus are covered by well de- veloped perianths enclosing the stamens and ovaries and mixed with inter- floral bracts, but they lack a clearly differentiated involucre. In the female inflorescence of Artocarpus the perianths are fused to form a syncarpium. Dissection shows that the proximal portions of the perianths, enclosing the ovaries, usually remain free, while the distal portions fuse, either completely or leaving the apices free, to form an external wall. The perianths have been found to be completely free in the very young in- florescence. In some species (in subg. Pseudojaca) the fusion between the perianths is complete, but it is assumed that the structure of the syncarp is homologous. In Parartocarpus the stamens and ovaries are sunken in cavities which are described as receptacular, and have been shown to be separated by walls that are largely if not entirely of intercalary origin. The surface of the receptacle is covered by indurated, spinous, conical or truncate processes that may represent either perianth segments fused basally around the opening of each cavity and solid sterile perianths between these, or second- arily modified interfloral bracts, taking into account the less reduced structure of the related African genus Treculia. In Treculia perianths are present (tubular in the male, and of linear segments in the female) and are sunken with the stamens of ovaries between the fused stalks of interfloral bracts, the heads of which somewhat resemble the processes in Pararto- carpus. In both these genera there is a well developed basal involucre. The ovule has been found to be pendulous rather than basal in Hudllettia and the genus is therefore removed from the Conocephaloideae to the Arto- 1959 | JARRETT, ARTOCARPUS AND ALLIED GENERA, I 25 carpoideae. The stamens and ovaries are apparently enclosed in cavities as in Parartocarpus and there is an involucre of fleshy bracts. It is assumed that the structure of the receptacle is homologous, and that the surface is naked, perianths and interfloral bracts being absent. The orientation of the embryo and the mode of germination confirm the above alliances. In Prainea and Artocarpus the embryo is straight (or nearly so) and germination in the latter is hypogeal. In Parartocarpus and Treculia the embryo is curved, with folded incumbent cotyledons, and germination is epigeal. The embryo of Hullettia resembles that of Pararto- carpus in its orientation but is straight. Fic. 9. sn of the genera. 1, Prainea; 2, Artocarpus; 3, Pararto- carpus; 4, Hullet The arrangement of the leaves and stipules serves to distinguish Parar- tocarpus and Hullettia from each other, and also from Prainea and Arto- carpus. In the last genus this character readily defines two natural sub- genera, Artocarpus and Pseudojaca, further differentiated by details of the glandular epidermal hairs and the spongy mesophyll which also sepa- rate them from Prainea. Renner’s observations (1907) on leaf anatomy have been expanded and confirmed but his reduction of Prainea to the status of a section under Artocarpus is not upheld. In vegetative characters the interrelations between the three groups are reticulate, but these indicate the occurrence of parallel evolution and the differences in the female in- florescences override them. Prainea, Artocarpus, Parartocarpus and Hullettia, together with Tre- culia, form a group of genera united by their condensed capitate inflo- rescences and large embryos. Prainea and Artocarpus are closely related to each other, while Treculia, Parartocarpus and Hullettia form a less closely allied series, both groups showing increasing reduction of the in- florescences. Whether these genera together form a natural group within 26 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL the Moraceae cannot be certain until an up-to-date revision of the entire family has been carried out, but, from a review of the other Old World genera, it seems probable. The eee of Artocarpus and Parartocarpus are striking examples of parallel evolution. These two genera and Hullettia each represent a highly specialized evolutionary end-point, though reflecting a general tendency for the female inflorescence to become a syncarpium having the ovaries enclosed in either floral or axial tissue. KEYS TO THE OLD WORLD GENERA OF THE ARTOCARPEAE NATURAL KEY Inflorescences without an involucre; embryo straight. 2. Female perianths free, only fruiting perianths enlarging and projecting beyond surface of mature head; leaves alternate and distichous; stipules and scars lateral or partially oe gland-hairs with hag 8-celled head; mesophyll loose, cell-contents resinous. . . Prainea. 2. Female perianths fused to form a ae entire aa clr at maturity. . Artocar pus. ‘ Leaves spirally arranged: stipules plea. scars i gland- hairs with 4-16-celled head; mesophyll long-armed with globose or ellipsoid resin-cells. Subgenus Artocarpus. Leaves alternate and distichous: stipules anid scars lateral or partially intrapetiolar; gland-hairs with 1(—6)-celled head; mesophyll compact, lacking resin-cells. Subgenus Pseudojaca. Inflorescences with an involucre of triangular, ¢ OV ate | or orbicular bracts; em- bryo curved (except Hullettia). 4. Involucre multiseriate; stamens enclosed in tubular perianths, these and ovaries sunken between partially fused interfloral bracts; stipules am- plexicaul, scars annulate. (Treculia). 4. Involucre uniseriate; stamens and ovaries sumkensi In ‘cavities in receptacle; stipules non-amplexicaul, scars not annulate 5. Surface of receptacle armoured from indurated, spinous, conical or truncate processes; stipules fused, intrapetiolar, ae as Parartocarpus. 5; Surface of receptacle not armoured, smooth: stipules pte and lateral, MWB AL.. 6 ou once na we nae eb eex eee eee eee take , . Hullettia. — Ww — ARTIFICIAL KEY BASED ON VEGETATIVE CHARACTERS 1. Stipules amplexicaul, leaving annulate scars. 2. Leaves spirally arrange 2. Artocarpus subg. Artocarpus. 2. Leaves, at least on leimate branches, alternate and distichous. (7Treculia). 1. Stipules not amplexicaul, scars not annulate. 3. Leaves spirally arrange 4. she paired, linear, scars lateral, round; leaves minutely punctate beneath. 4. Hullettia. Stipules fused, intrapetiolar: scars Honeate: leaves not as above. 3. Parartocarpus. 3. Leaves ; alternate and distichous, 5. Leaves abruptly and distinctly acuminate, glabrous; gland- _— sunken, head 4—8-celled; mesophyll loose, cell contents resinous. . Prainea. 1959 | JARRETT, ARTOCARPUS AND ALLIED GENERA, I 27 5. Leaves usually not abruptly acuminate, often pubescent; gland-hairs not sunken, head 1(-6)-celled; mis os compact, without resin. EE aan ee, weg rtocarpus subg. Pseudojaca. NOTES ON THE TAXONOMIC TREATMENTS Keys. These are intended to indicate, as far as is practicable, the natural affinities of the species and to enable the identification of most material bearing either male or female inflorescences and, where possible, of sterile specimens. Distinguishing characters for sterile material of species with overlapping areas of distribution have been given in the notes under the species. Types. These have been listed, where known, at the end of the ref- erences for each accepted name and synonym, All these specimens have been examined unless otherwise stated. Detailed localities have been given here only when the type or a duplicate has not been seen, since they otherwise appear in the lists of specimens. Lectotypes have been selected for most species where the original description gave syntypes and for those species where the holotype has been destroyed, but isotypes exist. Neotypes have only been chosen, where necessary, for accepted names. SPECIFIC DESCRIPTIONS. These have been prepared from herbarium material, with the addition of details from field notes. The inflorescences have been described from material preserved in alcohol when this has been available. Otherwise the male flowers have been described from mate- rial boiled up in water. Where measurements in published descriptions differ greatly from those found in the specimens seen they have been given, together with the source. Measurements of the female inflorescences are probably frequently too small, since fully ripe fruits are not often collected, and many of the species will certainly have larger sapling leaves than indi- cated here. Descriptions of new species have been prepared from the types and supplemented by details from other collections enclosed in brackets. Details of the stipules have only been given for Artocarpus subg. Arto- carpus, Parartocarpus and Hullettia, since in the other groups they do not show significant variation between the species. For the same reason, details of the glands on the leaves and of the presence or absence of a hypodermis have only been given for Artocarpus subg. Artocarpus. In de- scribing the leaves the term “main veins” includes the midrib and the lateral veins, while the term “reticulum” includes the intercostals unless these have been mentioned separately. In counting the lateral veins all those at the base have been included as well as all those clearly distinguish- able at the apex. The inflorescences have been described as immature before anthesis and the female inflorescence has been described as submature between anthesis and maturity. In Prainea and Artocarpus the interfloral bracts (as opposed to the involucral bracts of Parartocarpus and Hullettia) are referred to as bracts in the descriptions. In describing the syncarp of Artocarpus and Parartocarpus the variously shaped projections of the 28 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL surface, excluding peltate interfloral bracts. are referred to as processes or papillae. The morphological nature of these, which has already been dis- cussed, differs both within Artocarpus and between Artocarpus and Parar- tocarpus. The terms “core” and “wall” refer respectively to the receptacle and the surface layer formed by the fused perianths. CITATION OF SPECIMENS. These have been listed by areas, arranged from west to east in continental Asia (excluding Malaya) and according to the enumeration of geographical units by van Steenis (1950) in Malaysia, except that in the latter the order of the first four units has been changed to Malaya, Sumatra, Borneo, Java, in order to reflect the geographical continuity of Malaya with Asia and the close connection of Borneo with Malaya and Sumatra. In the Philippine Islands, Luzon and Mindanao have been further subdivided into provinces, and specimens from eastern New Guinea have also been listed according to recent political subdivisions. The herbaria in which the collections have been seen are indicated by the abbreviations of the Index Herbariorum, Part I. ed. 3. 1956. These have been given above in conjunction with the acknowledgements. In addi- tion, the following abbreviations may be noted which have been used preceding the collection numbers. bb — Bossen buitengewesten; i.e., collections 0 by the Netherlands Indies Forest Service from areas — Jav BS — Bureau of Science, Manila, Philippine CF — Conservator of Forests series, Forest Institute, Kepong., Malaya. FB — Forestry Bureau, Manila, Philippines. NIFS — Netherlands Indies Forest Service; this has been prefixed to those collections made by the service which do not have bb numbers. (They are numbered in several series, distinguished by various abbreviations indicating where they were collected. ) PB — Piante Bornensi; collected by a PP — Piante Papuane; collected by Becc SB — _ Blancoanae; collected by Merrill to illustrate his book of that itle. SFN — cine field number. Localities have been cited as given on the original label, apart from some minor corrections. Where important alterations have been necessary, these have been added in brackets. The Dutch oe has been transliterated through- out as w. The presence of male and female inflorescences has been indi- cated, but in Prainea, Parartocarpus and Hullettia the sex of immature inflorescences can often be determined only by dissection, and these have been recorded as ° VERNACULAR NAMES. Only those names have been cited which are of fairly common occurrence for a given species, or which are of particular interest. The following names are applied to any species of Artocarpus subg. Pseudojaca: Chay or Cay chay (Annamite), Kwai muk or Hung kwat muk (China), Tampang (Malay), and Cubi and Anubing (Philippines). The name Tampang is also applied to Prainea limpato. In Artocarpus subg. Artocarpus, Terap is the Malay name for A. odoratissimus, A. scorte- 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, I 29 chinu, A. elasticus, A. sericicarpus and A. tamaran. Many other species of this subgenus are iain by their own names, which are usually accurately applie BIBLIOGRAPHY BaILLon, H. Sur l’organisation du fruit de l’arbre a pain. Adansonia 4: 79-83. t. 5. 1863 s exoticae novae (cont.). Adansonia 11: 292-312. 1875. | Parar- tocarpus beccarianus, 294-295. | Beccari, O. Nelle Foreste di Borneo 625-636. Appendix. Nuove specie di Artocarpeae Malesi e Papuane. 1902. BENTHAM, G., and J. D. Hooker. Genera Plantarum 3: 341-395. 1880. [Urti- caceae, oi BERNBECK, F. Vergleichende Morphologie a Urticaceen- und Moraceen-inflo- reszenzen. Bot. Abh. ed. Goebel 19: 1-10 32. moe J. G. Gymnartocarpus venenosa ae Icones Bogorienses 1: 73-76. 1897. t. 24-25, BRownNE, F. G. Forest Trees of Sarawak and Brunei 357. ¢. 47. 1955. Corner, E. J. H. Notes on the systematy and aeibution of Malayan Phanero- gams, II. The Jack and the Chempedak. Gard. Bull. Singapore 10: 56-81. bal ee L959: Wayside Trees of Malaya 1: 652 [Artocarpus anisophyllus], 653 [A. elasticus |. 1940. ENGLER, A. Monographieen afrikanischer Pflanzen-familien und -gattungen. I. Moraceae (excl. Ficus) i-iv, 1-50. ¢t. 1-18. 1898. [Treculia, 31-35, t. 12-15.] Moraceae africanae IV. a Jahrb. 40: 543-549. figs. 1, 2. 1908. [Aeanthatrecli 546-549. fig. Forster, J. R. and G. ee Generum Plantarum 101-102. ¢. 51, 51a. GAERTNER, J. De Fructibus et Seminibus Plantarum 1: 344-346. ¢. 71, 72. 1788. HvutTcuHInson, J., in D. Pratn. Flora of Tropical Africa 6(2): 228. 1917. JARRETT, F. M. A note on the identity of rr genus Balanostreblus (Moraceae). Jour. pol Arb. 39: 107-110. plate. 1 KING, G., . D. Hooker. Flora of acu India 5: 546 [Prainea], 547 ee 1888. Kinc, G. The species of hoy tas indigenous to British India. Ann. Bot. Gard. Calcutta 2: 1-16. ¢. 1-14. 1889. MeETcALFE, C. R., and F. a RicHARDSON. The use of polyvinyl alcohol and related compounds as a mounting medium for microscope slides. Kew Bull. 1949: 569-571. 1950. RENNER, O. Beitrage zur Anatomie und Systematik der Artocarpeen und Cono- cephaleen, insbesondere der Gattung Ficus. Bot. Jahrb. 39: 319-448. 1907. ae J. J. Plantae novae vel criticae ex herbario et horto Bogoriensi. Bull. Jard. Bot. pe: III. 4: 230-240. t. 4-10. 1922. [Gymnartocarpus triandra, 233-235, STEENIS, C. G. G. J. VAN. Flora Malesiana I. 1: Ixx-cxvi. 1950. SUBBA Rao, A. M. Cytology and embryogeny in iat integrifolia Linn. Half-yearly Jour. Mysore Univ. Sect. B. Sci. 1: 6 1940. TreEcuL, A. Mémoire sur la famille des Artocarpées. . Sci. Nat. Bot. III. 8: 38- 157. t. 1-6. 1847. Troup, R. S. The Silviculture of Indian Trees 3: 876-885. figs. 323-326. 1921. 30 JOURNAL OF THE ARNOLD ARBORETUM VOL, XL STUDIES IN ARTOCARPUS AND ALLIED GENERA, II. A REVISION OF PRAINEA FRANCES M. JARRETT Prainea King in Hook. f. Fl. Brit. Ind. 5: 546. 1888; King, Ann. Bot. Gard. Calcutta 5(2): 162. t. 196. 1896: Engler & Prantl, Nat. Pflan- zenfam. Nachtrage II-IV. 122. 1897; Boerl. Handl. Fl. Ned. Ind. 3: 337, 372. 1900; Becc. For. Borneo 635, 1902, Webbia 5: 563. 1923; Ridley, Fl. Malay Penin. 3: 358. 1924; Van Steenis, Bull. Jard. Bot. jugs: IIT. 12: 259. 1932; Jarrett, Jour. Arnold Arb. 40: 8, 11. fig. 1, a—e, fig. 3, a-c. 1959. Typr Species: Prainea scandens King. Artocarpus section Prainea Renner, Bot. Jahrb. 39: 366. 1907. Trees (? or climbers). Leaves alternate and distichous, simple, entire, penninerved, glabrous; epidermal gland-hairs sunken, heads globose, 4—8- celled; spongy mesophyll loose, cell contents somewhat resinous; juvenile leaves elongate. Stipules paired, small (to 5 mm. long), lateral or intra- petiolar, scars non-amplexicaul. Inflorescences unisexual, capitate, heads globose, or the male short- obovoid, pedunculate, solitary or paired in leaf-axils; flowers mixed with numerous narrowly peltate, clavate or spathulate interfloral bracts; peri- anths free, enclosing a single stamen or ovary; involucre absent. At anthesis perianths and bracts compactly covering surface; male head with numerous flowers; perianths tubular, bilobed or perforate above; stamens shortly exserted; female head with fewer flowers (c. 20-100); perianths tubular, clavate, fleshy above, the mouth contracted, very shortly 2—4-lobed or merely perforate; ovary unilocular, style apical, bifid, the branches ex- serted, ovule subapical. Mature female head with 1-20 flowers forming fruit, perianths greatly enlarged and projecting from surface, fleshy, sub- globose to ellipsoid; remaining perianths elongating somewhat or not at all and surface hence loose or compact; mature ovary membranous, style apical, seed large, attached laterally near base, testa absorbed except the thickened saucer-shaped basal portion, endosperm none, embryo straight. longitudinal, cotvledons equal, appressed faces oblique to median plane of ovary, radicle and plumule minute, apical; interfloral bracts persistent. DisTRIBUTION: Malaya. Sumatra, Borneo, Moluccas, New Guinea. The genus Prainea, described by King in 1888, was based on a single species from Malaya, P. scandens, which he illustrated in 1896 in “A Century of New and Rare Indian Plants.’ King described the ovule as basal and hence placed the genus incorrectly in the Conocephaloideae. However, although the mature seed is attached sub-basally (in P. papuana, the only species of which adequate material has been available for study ) 1959} JARRETT, ARTOCARPUS AND ALLIED GENERA, II 31 if the ovary is examined at anthesis the position of the ovule is found to be lateral just below the apex of the loculus In 1902 Beccari, in the appendix to “Nelle Foreste di Borneo” entitled ‘“Nuove Specie di Artocarpeae Malesi e Papuane,” briefly discussed the inflorescences of Prainea, stating that he had found the ovule to be pendu- lous [at anthesis] in spirit material of his P. cuspidata (= P. limpato) and that the genus was allied to Artocarpus, differing only in the structure of the female inflorescence. He also questioned King’s description of Prainea scandens as a climber. He described as new the three other species of Prainea here recognized and a fourth which was an identification of Rum- phius’ Metrosideros spuria (= Artocarpus fretissii Teysmann & Binnen- dijk Fic. 10. Distribution of the species of Prainea, with inset showing styles; 1, P. scandens; 2, P. frutescens; 3, P. limpato; 4, P. papuana. Renner’s reduction of the genus in 1907 to the status of a section under Artocarpus has been discussed fully in the introductory paper of this series and rejected on the grounds that the resemblance in leaf structure to subg. Artocarpus and in leaf and stipular arrangement to subg. Pseudojaca is of less significance than the difference between the free perianths of the female inflorescence of Prainea and the syncarpium of Artocarpus. In uniting Prainea with Artocarpus, Renner correctly identified Beccari’s Prainea cuspidata with Miquel’s Artocarpus limpato (1861) and listed four species which are equivalent to those recognized here. He found the ovule to be pendulous in Prainea frutescens and thus the only species in which this character has not been established is P. scandens. e species are readily separable and the characters used may be seen from the key. No marked variation occurs within the genus, although P. scandens and P. frutescens, on the one hand, and P. limpato and P. papuana, on the other, are somewhat more closely allied to each other by leaf characters and the length of the peduncle. 32 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL The genus Prainea is entirely restricted to Malaysia, with three species in western Malaysia, of which one is endemic to Malaya, and another to Borneo, and a single species in the Moluccas and New Guinea. Van Steenis, in discussing the distribution of the Styracaceae in 1932, cited Prainea as an example of a genus which, like that family, occurs in both the western and the eastern parts of Malaysia, but not in the Philippines. The absence of Prainea from the Philippines has been confirmed thus far and, moreover, the genus is apparently lacking from nearly all of the geologically unstable area between the Sunda and Sahul continental shelves. Species of Artocarpus showing a similar pattern of distribution will be noted in a discussion of the geographical distribution of this group of genera at the conclusion of this series of papers. KEY TO THE SPECIES OF PRAINEA _ Male peduncles 5-15 mm.; female ae 10-25 mm.; twigs c. 1.5 mm. thick; leaves with 6-12 pairs lateral v 2. Male heads to c. 15 mm. across; ae heads to c. 35 mm. across with fruiting perianths projecting up to 25 mm.; stylar branches 2 mm. long, ACULES. soricoc.e Cu. 25-4 3h 65 te PoE oR Rae Ree ee eee Ys 1. condens. 2. Male heads 4—6 mm. across; female heads to c. 10 mm. across with fruit- ing perianths projecting less than 10 mm.; stylar branches rs 5 mm. long, t Oe ae chee oe ser ae ee doe Se oe eee 2. P. frutescens Male peduncles 15-60 mm.; female i 35-110 mm. twigs 2-5 mm thick; leaves with 9-24 pairs lateral v 3. Male heads to c. 15 mm. across; pate heads to c. 35 mm. across with fruiting perianths projecting up to 15 mm.; stylar branches forked; ma- ture leaves with lateral a distinctly prominent beneath, areolae bullate above only, or not at all. 6.202... 3. P. limpato. 3. Male heads 3-8 mm. across; ok —— to 20 mm. across with fruit- ing perianths projecting up to 10 mm.; stylar branches obtuse; mature leaves with lateral veins not or shallowly prominent sesaieaag areola often bullate on both surfaces. ......... 20.0.0... 0.0. eee ee P. papuana. =" Prainea scandens King in Hook. f. Fl. Brit. Ind. 5: 547. 1888; King, Ann. Bot. Gard. Calcutta 5(2): 162. t. 196. 1896; Ridley, FI. Malay Penin. 3: 358. 1924. Syntypes, Malaya, King 3549, 3693 (CAL, not seen); duplicates examined (BM, kK). Artocarpus scandens Renner, Bot. Jahrb. 39: 367. 1907. —" Climber(?), height to 20 m. Twigs c. 1.5 mm. thick, finely rugose, minutely appressed pubescent, soon glabrescent. Leaves 9-17 * 3.5-7 cm., obovate-elliptic, varying elliptic or oblong-elliptic, abruptly contracted into a narrow acumen to 1.5 cm. long, base cuneate, margin entire; midrib prominent beneath, lateral veins slightly so; lateral veins 9-12 pairs, curved, inarching at margin, often with secondary loops; intercostals not parallel; glabrous, dark green, drying greenish above, pale to reddish brown beneath, petiole 5—15 mm. long 1959 | JARRETT, ARTOCARPUS AND ALLIED GENERA, II Ri) Inflorescences solitary in leaf-axils. Male head (not at anthesis) to 15 mm. across, short-obovoid; perianths tubular, bilobed above, 1 mm. long; stamen immature, filament cylindric, anther oblong, 0.3 mm. long; bracts slenderly stalked, heads peltate, 0.4 mm. across; peduncle to 15 x 1.5 mm.. velutinous; female head at anthesis with bifid styles exserted, branches 2 * 0.2 mm., acute. Mature female head to 35 mm. across (excluding fruiting perianths), globose, loosely covered by numerous flow- ers and bracts, yellow, drying brown; fruiting perianths 1-7, projecting to 25 mm., ellipsoid, 15 mm. across, broadly and obtusely attenuate; re- maining perianths 12-15 mm. long, apices clavate, perforate, minutely pubescent, 0.8 mm. across; bracts numerous, slenderly stalked, heads clavate. 0.8 mm. across, pubescent; peduncle 20-25 2.5 mm., puberu- ent. DIstRIBUTION: in lowland evergreen forest to 2500 ft., endemic to Malava. Malaya. PERAK: Larut, King 3549, Nov. 1882 (bm, kK, 46), King 3693 Dec. 1882 (BM, K. 2). SELANGOR: bridge where Gap road crosses Sungei Semangko, Strugnell Kepong FN 49703 (sinc, ¢, detached inflorescence with leaves of Artocarpus rigidus). The material of this species is inadequate, but King’s collections agree well on vegetative characters and the mature female head on Strugnell Kepong FN 49703 matches that of King 3693. This species appears to differ from all others in Prainea and Artocarpus in being a climber instead of a tree. The labels on both King’s collections describe the plants as “creepers.” King 3693 originally read “a tree with slender hanging branches 50-60 ft. high . . . clinging to another tree” and was altered later. 2. Prainea frutescens Becc. For. Borneo 635. 1902. Holotype, Borneo, Beccari PB 667 (¥1); isotypes (FI, K, P). Parartocarpus sp. Benth. & Hook. f. Gen. Pl. 3: 375. 1880. Artocarpus frutescens Renner, Bot. Jahrb. 39: 367. 1907. Trees. height to 30 m. Twigs c. 1.5 mm. thick, finely and acutely rugose, appressed puberulent, soon glabrescent. Leaves 5-16 & 3-7 cm., obovate- elliptic to obovate- or elliptic-oblong or elliptic, abruptly contracted into a narrow acumen to 2 cm. long, base cuneate, rarely oblique with one side rounded, margin entire; midrib prominent beneath, lateral veins slightly so; lateral veins 6-10 pairs, curved, inarching at margin; intercostals not parallel; glabrous, drying greenish to pale or reddish brown; petiole 7-12 mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 4-6 mm. across, subglobose to short-obovoid; perianths tubular, 0.8 mm. long, minutely pubescent; stamen 1.2 mm. long, filament cylindric, anther ob- long, 0.15 mm. long; bracts slenderly stalked, heads peltate, 0.3 mm. across. ciliate; peduncle 5-12 x 1 mm., velutinous; female head with 34 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL bifid styles exserted, branches 0.6 0.2 mm., truncate. Mature female head to 10 mm. across (excluding fruiting perianths), globose, compactly covered by 20-30 flowers and numerous bracts; fruiting perianths 2 (in only mature head seen) projecting to 7 mm., ovoid, 6 mm. across, obtuse; remaining perianths c. 2.5 mm. long, apices obtuse, minutely pubescent, 5 mm. across, mouth shortly bilobed; bracts coherent in groups, slen- derly stalked, heads narrowly peltate, 0.25 mm. across, pubescent; pedun- cle 13-16 & 1.5 mm., short-pubescent. DISTRIBUTION: in lowland evergreen forest to 200 ft., endemic to Borneo. neo. SARAWAK: near Kuching, Beccari PB 667, Nov. 1865 (FI, K, P. 6, Haviland 3111 (BM, Bo, 9). EAST AND NORTHEAST Borneo. W. Kutei: bb 16183 (A, Bo); Longbleh, Sungei Pekan, bb 29606 (a, Bo, L, SING, 6); Mujup, bb 16785, bb 16911 (A, Bo, L). 3. Prainea limpato (Miq.) Beumée ex Heyne, Nutt. Pl. Ned.-Ind. ed. 2. 1: 579. 1927 sat tg limpato Miq. Fl. Ind. Bat. Suppl. 421. 1861; Renner, Bot. 39: 7. Syntypes, Sumatra, Diepenhorst HB 2092, HB 2500 ae one Diepenhorst HB 2092 (wv). pio diepenhorstii Miq. Fl. Ind. Bat. Suppl. 439. 1861. Holotype, Sumatra, Teysmann HB 716 (v); isotype (Bo). Ficus pire att (Miq.) King, Ann. Bot. Gard. Calcutta 1: 181, 1888. _ Ree nay Becc. For. Borneo 636. 1902. Holotype, Borneo, Beccari PB 2825 (¥1); isotypes (FI, K, P). Prune a, Merr. Philip. Jour. Sci. 29: 364. 1926, Pl. Elmer. Born. 1 olotype, Borneo, Castro and Melegrito 1610 (vc, not seen); a (A, BO, K). Trees, height to 30 m. (60 m. fide Gusdorf 115), buttressed, bark dark red-brown, peeling off in flakes. Twigs 2-5 mm. thick, finely rugose, ap- pressed puberulent, soon glabrescent. Leaves 10-33 X 4-13 cm., oblong, varying oblong- or obovate-elliptic, or elliptic, apex abruptly contracted or often subtruncate below a narrow acumen to 1.5 cm. long, base cuneate. or often oblique with one side subcordate, margin entire: midrib and lateral veins prominent beneath, intercostals slightly so, areolae often slightly bullate above; lateral veins 11—24 pairs, straight or slightly curved, inarching at margin conspicuous; intercostals forming rectangular net: glabrous, drying dark to pale reddish brown; petiole 8-22 mm. long. Inflorescences solitary or paired in leaf-axils. At anthesis: male head c. 15 mm. across, globose; perianths tubular, perforate, 1.4 mm. long, minutely pubescent; stamen 1.8 mm. long, filament cylindric. anther oblong, 0.4 mm. long; bracts slenderly stalked, heads peltate. 0.3 mm. across, ciliate; peduncle 30-40 & 2 mm., glabrous to short-pubescent: female head with bifid styles exserted, branches 0.5 0.2 mm., forked at the tips. Mature female head to 35 mm. across (to 55 mm. including fruiting perianths), globose, loosely covered by numerous flowers and bracts, yellow-green, drying brown; fruiting perianths 8—20, projecting to 1959 | JARRETT, ARTOCARPUS AND ALLIED GENERA, II 35 15 mm., ellipsoid, 10 mm. across, obtuse or broadly and obtusely attenuate; remaining perianths c. 4 mm. long, apices clavate, perforate, minutely pubescent, 1.5 mm. across; bracts slenderly stalked, heads spathulate, 3 mm. across, ciliate; peduncle 75-110 & 2-3 mm., puberulent. VERNACULAR NAMES: Tampang (Malay), Sumatra, Borneo; Limpato, Sumatra. DISTRIBUTION: in lowland evergreen forest to 1700 ft., Malaya, Sumatra, Borneo. Malaya. Kepau: Weng road, near Baling, Best SFN 21266 (K, SING, infl.). Sumatra. ATJEH: Bivak Aer Putih, waterfall near Pendeng, Gajolanden. Van Steenis 9270 (Bo, L, 2). West Coast: Fort Elout, Teysmann HB 716 (Bo, pee Priaman, Diepenhorst HB 2092 (po, L, vu, 2), HB 2500 (Bo, L, Uv, 3), 2914 (Bo). INDRAGIRI: Batang Peranap, bb 30095 ee L); Muara Serangge, ; 30066 (Bo, L); Peranap, bb 30105 (aA, Bo, L). Dyampr: Danau Lama, bb 13639 (BO). PALEMBANG: Lematang Ilir, Gunong Megang, NJFS T 340 (Bo, L ¢), NIFS T 521 (po, K, L, U, 2); Lematang Ilir, Semangus, 6b 32063 (A. Bo, L, SING, 6), bb 32085 (A, BO, L, SING); Mulak Ulu, Grashoff 323 (Bo, L). LAMPONGS: Manggala, Gusdorf 115 (Bo, 2). ENGGANO: forest near Bua bua, Lutjeharms 4558 (A, BO, K, L, P, SING). Borneo. ane Gunong Braam, near Sabungo, Beccari PB 2825, Nov. 1866 (FI, K, P, 6. 2). SOUTH AND SOUTHEAST ea Tanah Bumbu, Kampong Baru, bd pai (Bo). EAST AND NORTHEAST BORN Berouw, Domaring, bb 18817 (A, Bo, L); Bulungan, Mara, bb 10809 (Bo) ; Une Mahakam. D. Parei, bb 20743 (A, Bo, L). British NorTH Borneo: Liman Wood 1790 (Bo, 6); Tawao, Elphinstone Prov., Elmer 21799 (A, BM, BO, GH, K. L. P, SING. U, 2). BANGUEY IsLAND: Castro & Melegrito 1610, July—Sep. 1923 (Bo, K, 4). The appearance of the mature female head is rather variable, owing to differences in the number of fruiting perianths and in the extent to which the other perianths elongate after anthesis. The leaf form in which the apex is subtruncate below the acumen is distinctive and is found only in this species of the genus. 4. Prainea papuana Becc. For. Borneo 635. 1902. Holotype. New Guinea, Beccari PP 675 (¥1); isotypes (FI). Artocarpus papuanus Renner, Bot. Jahrb. 39: 367. 1907, non Diels. 1936. Prainea ial ei hes J. J. Smith, Bull. Jard. Bot. Buitenzorg III. 6: 80. yntypes. Ternate, Beguin 691, Halmaheira, Beguin 1919, 1980, 2242 (BO); Shek Beguin 1980 (Bo) ss dl papuanus S. Moore, Jour. Bot. 61 Suppl. 52. 1923. non Becc. Diels. Bot. Jahrb. 67: 175. 1936. Syntypes, New Guinea. Forbes 337, BM); lectotype, Forbes 411 (BM). ~~ rae Trees, height to 35 m., bark brown, peeling off in flakes. Twigs 2-3 mm. in diameter, finely rugose, appressed puberulent, soon glabrescent. Leaves 8-26 4-12 cm., oblong, varying elliptic, abruptly narrowed into an acumen to 1.5 cm. long, base cuneate or rounded, often oblique with one side subcordate, margin entire; midrib prominent beneath, lateral veins 36 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL slightly so, areolae often bullate on both surfaces; lateral veins 9-20 pairs, straight or slightly curved, inarching at margin conspicuous; intercostals not parallel; glabrous, drying red-brown, grey or green above, dark to pale brown beneath; petiole 8-18 mm. lon Inflorescences solitary or paired in leaf-axils. At anthesis: male head 3-8 mm. across, globose; perianths tubular, clavate, shortly bilobed above, 1.8 mm. long, minutely pubescent; stamen 2.5 mm. long, filament slender, cylindric, anther oblong, 0.4 mm. long; bracts slenderly stalked, heads clavate, 0.4 mm. across; peduncle 15-60 & 2 mm., velutinous; female head olive green, with bifid styles exserted, branches 0.7 < 0.5 mm., obtuse. Mature female head 10-20 mm. across (excluding fruiting perianths), glo- bose, loosely to compactly covered by c. 25-100 flowers and numerous bracts, orange, drying brown; fruiting perianths 2—12(—25), projecting to 10 mm., subglobose to conical, 10 mm. across, mucronate; remaining perianths c, 2.5 X 1 mm., clavate, mouth contracted, obtusely 4-lobed: bracts slenderly stalked, heads clavate, 0.5 mm. across, often coherent in groups; peduncle 30-90 2 mm., velutinous. DIsTRIBUTION: in lowland evergreen forest to 3500 ft., Moluccas (Hal- maheira group), New Guinea. Moluccas. Mororar: Gunong Moku, Anang 263 (L); Mt. Permatang, along the Sangowo River, Kostermans 959 (1, SING); Tjaw, Kostermans 836, 837 (1). HALMAHEIRA: Galela, Soa Tobaru, Beguin 1919, Dec. 1921 (Bo, k, 1926 (Bo, L, 2), 1980, May 1922 (Bo, L, P, SING, U, 2); Kampong Toliwang, Idjan/Moehtar 364 (kx); W. Pitu, Beguin 2242, Nov. 1922 (BO, K, L, P, U). TERNATE: Laguna, Beguin 691, May 1920 (Bo, L, 2). BATJAN: Masurung. bb 23133 (BO, L). New Guinea. VoGeLkop: Andai, Beccari PP 675, 1872 (1, 6, 2); Mano- kwari, Pomi, Kolonisatie, bb 15897 (Bo, 2); n. of Ransiki, 80 km. s. of Manokwari, Kostermans 4746 (a, K, L, SING, 2); Sorong, Beccari s.n. (FI). DutcH North New GUuINEAa: Bernhard Camp, Idenburg River, Brass & Versteegh 13167, 14122 (a, 2). Papua. Central Division: Boridi, Carr 14976 ‘“G, 2); Koitaki, Carr 12213 (BM, x, L, SING, infl.), 12586 6 (BM, K, SING, infl.), 12886 (BM, K, L, SING, 2); Sogeri, Forbes 331, 411, 1885-6 (BM, K, L, 2). Milne Bay District, Cape V ogel Peninsula: near Medino village on n. coast, Hoogland 4666 (a, 2); Menapi, Brass 22000 (a, 2). MANDATED TERRITORY OF NEw GvuINEA. Morobe District: Womersley NGF 2994 (A, BO K, L, 2): Quem- bung, Sattelberg, Clemens 8099 (a, 2). D’ENTRECASTFAUX Is. Normanby I.: Lebudowa River, Brass 25517 (kK, L, 2); Waikaiuna Bay, Womersley & Brass NGF 8674 (A. K. 2); near Waikaiuna Plantation, Jackson NGF 4109 (A, K, L, SING, 2). JAPEN: Serui, bb 30434 (Bo, L), bb 30655 (L). LovursiaDE ARCHIP.: Misima I.. Mt. Sisa, Brass 27523 (k, L, 9). In his description of Parartocarpus papuana, S. Moore cited a third specimen, Forbes 503, but this has not been found either at the British Museum (Natural History) or elsewhere. The female head in Prainea papuana shows a variability similar to that of P. limpato in the number of seeds formed and in the degree of elonga- ton of the perianths after anthesis. The latter is partially correlated with 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, II 37 a geographical variation in the diameter of the head at anthesis. In east- ern New Guinea the female head then has a diameter of 10-15 mm. and the surface generally remains compact at maturity, but in the Vogelkop and the Moluccas the diameter is only about 5 mm. at anthesis and later all the flowers usually elongate somewhat so that at maturity the surface becomes loose. A parallel variation occurs in the size of the male head. However, the collections from Dutch North New Guinea are intermediate, with a diameter of 8-10 mm. at anthesis and with the surface remaining fairly compact. The creation of subspecies is therefore not regarded as justifiable. SPECIES EXCLUDENDA Prainea rumphiana Becc. For. Borneo 636. 1902 = Artocarpus fretis- sii Teysm. & Binnend. Abh. Naturf. Ges. Halle 9: 189. 1866. 38 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL A DISCUSSION OF THE PACIFIC RAILROAD REPORTS AS ISSUED IN THE QUARTO EDITION SUSAN DELANO MCKELVEY THE VOLUMES CONSIDERED in these pages contain the findings of those who participated, both in field and in laboratory, in the first systematic examination by the United States government of the immense area extend- ing from the Mississippi River to the Pacific Ocean and from Canada on the north to Mexico on the south, and which, in the half century from 1803 to 1853 had become part of our national domain. The only official expeditions yet to have crossed the region, those of Lewis and Clark and of Frémont, had acquired, everything considered, some interesting scientific facts, but these were inadequate when it became necessary to determine ‘‘the most practicable and economical route” for a railroad. For the factor of practicability, concerned with such matters as grades distances, and weather conditions, demanded trained topographers, astron- omers and meteorologists; while the factor of economy, interlocked to be sure with the practicable, but even more intimately related to a railroad’s construction and maintenance as well as to the subsistence of prospective settlers along its line of march, necessitated knowledge, in terms of iden- tification, of rocks, minerals, soils, timber, plants and all forms of animal. bird, fish and reptilian life, as well as estimates of their abundance and availability. Such facts could only be supplied by geologists, mineralogists, botanists, zoologists and others proficient in the natural sciences, at that time a covering term for all that related to the systematic study of nature as a whole, whether animal, vegetable or mineral. On earlier official expeditions the government had demonstrated little enthusiasm when asked to include men interested in such matters. Now the expert knowledge which only they could supply had become vital and the Smithsonian Institution and its members were called upon to provide instructions in the technicalities of field collecting —the making of the specimens and the keeping of the records upon which the reports of the scientists must of necessity be based — and to enumerate the equipment necessary for its accomplishment. This they did, wholeheartedly, even specifying red pencils and stressing the importance of adding tartar emetic to all alcohol — “besides adding to its preservative powers, [this] will remove any temptation to drinking it on the part of unscrupulous wee ‘ Because the expeditions were to cross the trans-Mississippi West at, near, four widely separated parallels of north latitude, a siete coverage was assured as far as natural history collections were concerned and the scientific reports, some including fine plates, which eventually were appended to those descriptive of the routes and which were written by such authorities as S. F. Baird, W. P. Blake, J. S. Newberry, J. Torrey. 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 39 Asa Gray and many others, constituted and still must be regarded as in- valuable records. Although publications dealing with special aspects of the surveys never fail to mention that important scientific work was accomplished, they do not elaborate upon the statement. Long interested in collectors — often ignored to the point of anonymity —I determined, with the optimism of ignorance, to “run through” the series of reports, twelve volumes in thir- teen, and see what could be learned of the men upon whose accomplish- ments in the field the scientific edifice was erected. The great quarto edi- tion, occupying three to four feet of standing room, has been characterized s “monumental,” “sumptuous” — it has been stated! that its publication cost the government over one million dollars, the surveys themselves four hundred and forty-five thousand — and, although its pristine magnificence is no longer in evidence, its very dimensions may have contributed to my confidence that the content would be well organized, clearly presented and easily digested. This did not prove to be the case and any “running through” was reduced to a walk and eventually a full stop, for, with a mental preference to understand one thing before proceeding to another, I was soon “bogged down” in the fine points of military terms and distinc- tions, as well as in far deeper swam It finally became clear that, before attempting any study of the reports, three preliminary tasks must be faced: first, the government documents in which they were issued must be understood; second, the content of all volumes of the series must be analysed in order to be able to cite accurately from their pages; and, third, a satisfactory means of locating the data contained in the analyses must be devised. These tasks were completed for my own enlightenment and convenience. But, convinced that anyone approaching the reports for the first time must be faced with similar needs to mine, it is possible that my findings may have a wider usefulness. I shall comment upon what these three tasks involved before turning to their solution which constitutes my paper proper. (1) The issues of the quarto edition have been enumerated before now. But because enumerations vary greatly in form of presentation, it was necessary to compare and to check, one against the other, a number of authoritative bibliographies in order to feel certain that an accurate and complete picture had been obtained —a slower task than the brevity of my summary might indicate. Especially helpful to this end were the presentations contained in Mr. George Leslie Albright’s ‘Official explorations for Pacific railroads 1853— 1855.” issued in 1921 (University of California Publications in History, Volume X), in a Catalog of books represented by Library of Congress printed cards, published in 1948 (Ann Arbor, Michigan, Volume 154, 634— 635). and in Professor Howard Taft’s Artists and illustrators of the Old West 1850-1900, published in 1953 (Charles Scribner’s Sons, New York). Another volume, the Checklist of United States public documents (edition ‘See Tatt, H., op. cit. this page, (5 & fn. 13). 40 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL 3. 1: 1274-1275), published in 1911, which distinguishes all numbered Congressional Documents by Serial Set and Serial Number, served as a final check. These authorities are occasionally referred to in my paper. (2) Examining the quarto edition for the first time, problems inherent in its makeup became apparent immediately. I mention some of the more baffling: There is no index covering the content of the entire series, and under- standably so. For, with the exceptions of Volumes VIII, IX and XI, none has pages numbered consecutively throughout — Volumes II and NX. for examples, have seven sets of page numbers each — and it is therefore necessary, and I believe always will be necessary, to ‘leaf through” a vol- ume to locate an included report. Although page numbers in individual reports are for the most part clear, seemingly endless blank or only occa- sionally numbered pages, in front or back matter especially, are. at best, an annoyance if one needs to cite therefrom. Because it had been impossible, obviously, to assemble on time for a waiting printer, the great amount of material due from widely scattered sources, reports are not always present where a volume’s table of contents indicates and subsidiary papers, usually scientific, are not always found in proximity to the report to which their content relates. A note on an inserted page or an easily overlooked footnote states that such and such a paper will appear later, although when or where is not predicted. Only by time-consuming search through later volumes — involving in all instances the “leafing through” procedure already mentioned — can matter delayed in publication be located; it should be added that, when found, it is usually referred to its appropriate place in an earlier volume! And what is true of belatedly published papers is also true of matter elucidating the content of reports, whether this be in the form of plates of all sorts, of maps, of profiles, etc. Such material is usually enumerated in a report’s list of illustrations, but checking often proves that certain items are lacking, or missing at the point indicated. To enumerate the discrepancies in illustration-content existing in the many issues of any one volume — they appear to have been dictated by the convenience. or perhaps necessity of a volume’s compiler — was clearly impossible. One could not hope to do more than supply the total number included in the report under scrutiny and, when such existed, their numerical captions. Attempts to record the whereabouts of a report's scattered but affiliated content, textual and illustrative, resulted in the analyses contained in this paper at they were essential became evident when cross-references necessitated accurate citation. Furthermore, although analyses could not do away with the broken pagination of the volumes, an irremediable prob- lem, they could lessen some of the other difficulties inherent in the makeup of the series; for to be able to ascertain, within the compass of a few pages, the content of its volumes, would certainly be far more convenient than to search repeatedly through the ponderous tomes themselves. Two other matters related to the compilation of the volumes had best 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 41 be mentioned here — others will be referred to at appropriate places in my analyses. The first concerns dates of publication. The printers were obliged to proceed with their tasks; consequently the year supplied on the main title page of a given volume may be earlier than one recorded on an included letter, report, etc. When, in my analyses, later years than those which appear on the title pages are recorded, they are based on this internal evi- dence and merely represent the year before which, it may be assumed, a given volume could not have been issued. For one example see under Volume II. To scientists dates of publication can be immensely important. Dr. I. M. Johnston’s paper, “Publication dates for botanical parts of the Pacific Railroad Reports’ (Journal of the Arnold Arboretum, Volume XXIV, 237-242, 1943), assigns, on the strength of records found in old letters, contemporary publications and the like, “Reasonably exact dates to the various botanical reports, accurate in most cases to within a month or two.” Dr. Johnston also takes into consideration the first, or octavo edition of the reports, an issue not discussed in this paper. (See my fn. 3.) His conclusions are summarized in tabulated form at the end of his paper. Only in rare instances do the authors of scientific reports included in the quarto series mention the publication date of their papers. When they do so the fact is noted in my analyses. The second matter relates to the manner in which certain illustrations are incorporated in the volumes. Because of the great number contained in Volume XI the unfortunate results are especially apparent and my com- ments are confined to that volume, although to a lesser degree the criticism is applicable wherever similar illustrations are included. Volume XI is, in text, perhaps the most generally useful in the entire series of Pacific Railroad Reports. It is regrettable, therefore, that its associated and very fine maps, profiles, panoramic views, etc., represent- ing the graphic results of the topographers’ arduous task, should have been included in a manner which makes their manipulation impossible and is destructive of the illustrations themselves. Most are extremely large and many-times-folded. Bound tightly in with and outbalancing the text which constitutes the first half of the volume, to unfold them, to keep any single one spread out for study or any two for comparison, and to execute the even more difficult refolding without tearing, are all impossibilities. Yet, because none bears any exterior notation, they must be opened to full size for identification — the placement of the beautifully lettered and explicit captions seems to have been selected to that very end! Even in copies of Volume XI which have obviously stood unused, many of the illustrations have already disintegrated along their exposed, uneven folds; lack of dusting or the reverse may have been responsible, or possibly the paper was not suitable in the first place. Had they been mounted on linen (see fn. 4), visibly captioned to conform to the enumeration found in the volume’s “Contents,” and supplied loose in a container, their usefulness would have been assured and their life expectancy prolonged; as issued, 42 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XL they symbolize endless labor and much money spent to little purpose. It is possible that in some instances these valuable records might still be mounted although in others it could not be done now. (3) The Ann Arbor Catalog, cited on p. 39, supplies two synopses which serve as guides to the content of the quarto edition. The first, ““Con- tents (Outline), condenses into some 15 lines the routes described in its volumes, usually according to the parallels which they followed. The second, “Contents (By authors),” supplies an alphabetical list of the men who contributed reports, and then records, for the most part by a descrip- tive term, the subject matter of their papers, followed by the volume or volumes in which they are to be found. Presumably intended to supple- ment each other, these synopses presuppose a very great familiarity with the content of the series and, in my opinion, are not sufficiently explicit to. offer much assistance to anyone approaching the reports for the first time. Moreover, necessitating frequent reference to the volumes themselves, they present many of the already mentioned difficulties inherent in the edition’s makeup. The “Alphabetical index of authors” which concludes this paper supplies, I believe, a more complete roster of contributors than does the somewhat similar list of the Catalog. And, when used in conjunction with my anal- yses, it provides, without reference to the volumes themselves, the precise association of every report as well as an outline of that report’s entire subject matter and the pagination thereof. In possession of these facts one should be sufficiently informed to be able to turn immediately to the appro- priate volume as well as to the information sought in any given report. REPORTS OF EXPLORATIONS AND SURVEYS TO ASCERTAIN THE MOST PRACTICABLE AND ECONOMICAL ROUTE FOR A RAILROAD FROM THE MISSISSIPPI RIVER TO THE PACIFIC OCEAN 2 An enumeration of the Congressional Documents containing the reports, An analysis of the content of the twelve volumes, in thirteen, of the quarto edition in which the reports were issued in final form. An alphabetical index of authors to be used in conjunction with the analyses. CONGRESSIONAL DOCUMENTS CONTAINING THE REPORTS Although a few of the Pacific Railroad Reports enumerated below were issued in an octavo edition,* the revised, complete and authoritative “Present on the title-page of every volume in the series as well as on the title- -pages of included reports, this is inconveniently long for repeated citi ation and is usually short- ened to Pacific Railroad Reports, an raga which G. K. Warren used as early as 1859 in his table of contents of Volu , oS]. “The Checklist of United phates are eens states, Esra of the . reports first appeared in 8° in House Document 129, 33d Congress, Ist session, in asl nos. 736. 737, and 739. Serial no. 738, the projected v. 3 of this edition, was never printed.” See p. 39. 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 43 reports, the ones usually cited and the ones most often found in libraries, were prepared under the direction of the War Department and were issued for both Senate and House of Representatives in a quarto edition of twelve volumes, in thirteen, under the long descriptive title cited above. The quarto edition is found as follows: Volumes I through XI: 33d Cong., 2d Sess., Senate Ex. Doc. No. 78 |Serial Nos. 758 through 7684]; 33d Cong., 2d Sess., House Ex. Doc. No. 91 {Serial Nos. 791 through 801]. Volume XII: 35th Cong., 3d Sess., Senate Ex. Doc. No. 46 | Serial No. 992 °]; 36th Cong., 1st Sess., House Ex. Doc. No. 56 [Serial Nos. 1054, 1055 ®]. All the issues cited above are numbered Executive Documents and are included in the Serial Sets. There are, in addition, two printings, one of Volume XI, one of Vol- ume XII. which lack document numbers and are not included in the Serial Sets. These are: Volume XI: 36th Cong., 2d Sess., Senate Ex. Doc.* Volume NII: 36th Cong., 1st Sess., Senate Ex, Doc.® There may be variations in these unnumbered issues, related primarily to such matters as the date of publication on a volume’s title-page, their illustration-content or illustration-placement and so on, but as far as text is concerned, all issues, whether numbered or unnumbered documents, appear to be accepted as the same. AN ANALYSIS OF THE CONTENT OF THE THIRTEEN VOLUMES OF THE QUARTO EDITION Volume I — 1855 33d Cong., 2d Sess., Senate Ex. Doc. No. 78 |Serial No. 758 °| The title-page of Pacific erie Reports, Volume I, is dated 1855. No included matter bears a later dat me parts, certainly, of the reports must have been issued for the Senate in this octavo, unillustrated, and ae edition; for example: 33d Cong., 1st Sess., Senate Ex. Doc. No. 52 [Serial No. 698]. eres (45, fn. 1; 160), and Taft (255, fn. 13). ‘Taft (255, fn. 2) records: ‘“. . . The ice in the: Senate Serial Set (768' and 768°) of volume 11 are mounted on Tinen ae: uiring two books (parts 1 and 2) to contain them.” It has not been possible to ome a copy of Volume XI a mounted maps in the libraries of Boston or of Cambridge and the issue must be ra ° Tait (255, fn. 4) states: “This volume (Serial No. 992) duplicated in a Gale book the een in Serial Nos. 1054, 1055 [see below] but is designated on the title page as ‘Supplement to Volume 1’”. Card 2 of ee Ann Arbor Catalog notes that this “Supplement to vol. I,” was issued in 1859, W. A. Harvie, printer. “The two volumes of this issue are piles on the Sg eae title-pages of Pacific Railroad Reports as Volume XII. Book I, Volume XII. “The title- aoe records that this was printed in Washington = ee. W. Bowman and is dated 1 . The maps are bound in with the text. See Taft (255, kes 2) title-page records that this was orbed in Washington by Thomas H. Ford and is dated 1860. As in House Ex. Doc. No. 56 its two parts — I ay Book II) are bound separately, making two volumes. See Taft (255, fn * Because my analyses have been based upon a mixed set of i ‘qusite edition, the 44 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL “Contents of Volume I,” [iii], lists five papers. The subject matter of the first four, distinguished here as (1) through (4), is enumerated in the “Index to Report of the Secretary of War and Office Reports,” [v]—viii, and that of the fifth, (5) below, in the ‘Alphabetical Index to Governor Stevens’s Report,” [637 |-651. Page numbers run consecutively in (1) through (4); in (5) they are independently numbered. The title-pages of (1) through (5) read respectively: (1) Report of the Secretary of War on the several railroad explorations. [1]-35. For reprintings of tes of this report see “Explanatory note’ quoted under Volume III, fn. 18. (2) An examination by direction of the Hon. Jefferson Davis, Secretary of War, of the reports of explorations for railroad routes from the pair to the Pacific, made under the orders of the War Depart- nt in 1853-54, and of the explorations made previous to that oe which have a bearing upon the subject: by Capt. A. A. Hum- phreys & Lieut. G. K. Warren, Corps of Topographical Engineers. [35]-111. ConTENTs: Table of contents, [35]; An examination of the reports of explorations for railroad routes from the Mississippi to the Pacific, [37]-38; Humphreys & Warren report (6 chapters), [39]—111. r reprintings of portions of this report see “Explanatory note” quoted under Volume III, fn. 18. (3) Memoranda on railways, prepared Office of Pacific Railroad Surveys; by Capt. Geo. B. McClellan, Corps of Engineers, under instructions from Hon. Jefferson Davis, Secretary of War. [113]-13 4) Report upon the cost of transporting troops and supplies to California, Oregon, New Mexico, etc., etc. By Major General Thomas S. Jesup, Quartermaster General, US. Army. [131]—134 -— — mn ) Report of explorations for a route for the Pacific railroad, near the forty-seventh and forty-ninth parallels of north latitude, from St. Paul to Puget Sound. By I. I. Stevens, Governor of Washington Territory. [i]—vii; [1]-651. ConTENTS: Table of contents, [iii]—vii; Part I. Reports from the field, [1]—72; Part II]. Report (18 chapters), [73]—159: Miscel- laneous documents related to special aspects of the survey,'” 160-635; Note, [636]; Alphabetical index to Governor Stevens’s report. Pre- pared by E. B. Hunt, U.S.A., |637]-651. For maps and profiles see under Volume XI. A “Note,” [636], states: ““The Narrative of the Explorations, accompanied by Views, illustrating the Features of the country, the Natural History, Botanical and other Scientific Reports, with Illustrations, will be found in a __ Subsequent Volume.” See under Volume XII, Books I and II. one in the Arnold Arboretum Library, the printing is cited in every instance to facilitate checking should such be desired. * For authors of certain of these documents see fn. 56. 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 45 Volume II — 1857 33d Cong., 2d Sess., House Ex. Doc. No. 91 |Serial No. 792] The title-page of Faces jc Railroad Reports, Volume II, is dated 1855, but Pope's ‘Explanatory note” to Blake’s geological report, inserted before the title- page of (4A) below, is dated February 18, 1857. “Contents of Volume II,” [iii], lists six papers, distinguished here as (1) through (6); a seventh, (4A) below, is not listed. Each of these seven papers has its own set of page numbers. They are cited on their respective title-pages as follows: (1) Report of explorations for a route for the Pacific railroad, by Capt. J. W. Gunnison, Topographical Engineers, near the 38th and 39th parallels of font latitude, from the mouth of the Kansas River, Mo., to the Sevier Lake, in the Great Basin. Report by Lieut. E. G. Beck- with, Third Artillery. [1]—128. 13 Plates (captioned but not num- bered). CONTENTS: Introductory letter, [3]; Contents, [5]—-8; Report (10 chapters), [9]—118; Appendix A. Letters relating to the progress of the survey of the route near the 38th and 39th parallels, in charge of Captain Gunnison, [119]—124; Appendix B. Explanations of the map and illustrations, [125 |- For associated botanical report see under (2) below. For associated zoo- logical reports see under Volume X. For maps and profiles see under Volume XI, (2) Report of explorations for a route for the Pacific railroad, on the line of the forty-first parallel of north latitude. By Lieut. E. G. Beckwith, Third Artillery. [1]—132. ConTENTS: Introductory letter, [3]; Contents,'' [5|]—8; Report (10 chapters),1* [9]—112; Appendix. Instructions from the Secretary of War, [113]—114; Botanical report, [115 ]—132. For zoological reports see under Volume X. For plates, maps, profiles, see under Volume XI. The title-page of the included Botanical report reads: Report on the botany of the expedition: by John Torrey and Asa Gray. [115]-132. 10 plates (Nos. I-X).'% CONTENTS: Contents, [117]; Botanical report, Part I, [119]- 125; Part II,’* 125-131; Explanation of the plates, 131-132. (3) Synopsis of a report of the reconnaissance of a railroad route from “For “Errata” see p. 8, and for “Additional errata” see sheet inserted between pp. 8 and [9]. * Of these Chapter X, “Geology,” includes 4 plates of fossils (Nos. 1-4) and 3 woodcuts. * Of these, No. I is associated with (2), Nos. II-X with “Part I describes collections made under (2) and Part Il collections made under (1). For the wording of the long descriptive headings of Parts I and II, see Pacific Railroad Reports, Volume II, where the heading of Part I appears on pp. 117, 119, and that of Part II on pp. 117, 119, 125. 46 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Puget Sound via South Pass to the Mississippi River: by Fred W. Lander, Civil Engineer. [1|—45. Contents: Contents, [3]; Legislative and executive action in reference to this exploration and report, [5]—6; Introduction, [7]—28; Report, |29|—45 (4) Report of explorations of a route for the Pacific railroad, near the thirty-second parallel of north latitude, from the Red River to the Rio Grande, by Brevet — John Pope, Corps of Topographical Engineers. [i]—iv; [1]—1 CoNnTENTs: Pope letter Semi report, [ii]; Contents, [iii ]—iv; Report (8 chapters), [1]—50; 12 Appendices (A through L),'° [51]- 156; Botanical report, [157|-178; Index, |[179|—185. For map and profile see under Volum The title-page of the included Botanical report reads: Report on the botany of the expedition: by John Torrey and Asa Gray. [157]-178. Index,'® [179 |-185 aripcanrete with the route described in (4) but with its own set of page number (4A) Report on the geology of the route, near the thirty-second parallel: prepared from the collection and notes of Capt. Pope, by William P. Blake, Geologist of the Office of the United States Pacific Railroad Surveys. [1]—50. Geological map. Geological section. Contents: Explanatory note to geological aca [inserted be- fore title-page]; Blake letter prsiprnas his report, [3]; Contents, [5]-6; Report (6 chapters), |[7]—40; Catalogue of the geological collection made by Brevet Captain John en 1853, [41]-42; Note in explanation of the map and section, |43|]—44; Index to the report on the geology of the route near the thirty-second parallel, [45 ]—50. (5) Report of explorations for that portion of railroad route, near the thirty-second parallel of north latitude, lying between Dona Ana, on ® Most of these are highly technical. Four, however, have a general interes “Appendix A. Diary of the expedition, by J. H. Byrne, Assistant Computer” [51]-95, and “Appendix E. Table of camping places along the direct line of survey “Appendix B. Preliminary report on the anbural history,” [94], signed by pares F. Baird, refers to the scope ~ importance of Pope’s zoological collection, and t anticipated publications thereo Appendix D. Note upon the geological report,’ [98], concerns the delayed pub- S sci upon the geological collection, and the preliminary report of the octa ° This covers nara contained in Pope’s report proper as well as in the botanical ™ Pope wrote: “The mineralogical collections made by me were placed in the hands of M. Jules Marcou for examination, and carried by him to France. They were sub- sequently returned in a atin condition, and with many of the labels displaced. This fact will account for y errors in the report, a and section prepared by Mr. Blake.” This is dated Washington, February 18, 185 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 47 the Rio Grande, and Pimas Villages, on the Gila. By Lieut. John G. Parke, Corps of Topographical Engineers. [1|—28. ConTENTS: Report, We ~24; Appendix (in two parts, A and B), [25]-26; Index, [27]-28 For Parke’s later report see under Volume VII. ( ON Na Extract from report of a military reconnaissance, made in 1846 and 1847, by Lieut. Col. W. H. Emory. [1]-22. ConrENts: Humphreys letter explaining the inclusion of this earlier report, [3]; Extract from Lt. Col. Emory’s report, [5]—20; Index, [21]-22. Volume III — 1856 33d Cong., 2d Sess., House Ex. Doc. No, 91 [Serial No. 793] The title-page of Pacific Railroad Reports, Volume III, is dated 1856. No included matter bears a later date. Volume III has five sets of page numbers “Contents of Volume III,” [1], divides the subject matter into two parts, distinguished here as (1) and (2). The first bears the heading (1) Extracts from the (preliminary) report of explorations for a railway route, near the thirty-fifth parallel of north latitude, from the Missis- sippi River to the Pacific Ocean, by Lieutenant A. W. Whipple, Corps of Topographical Engineers. [3]|—36. ConTENTS: Explanatory note,!® [3]; Extracts, [3]-32; [Portion of| Appendix B. Table containing the distances, altitudes, * * * of each camp from Fort Smith to the Pacific Ocean, 33-36. The title-page of the second reads: (2) Report of explorations for a railway route, near the thirty-fifth parallel of north latitude, from the Mississippi River to the Pacific Ocean: by Lieutenant A. W. Whipple, Corps of Topographical Engineers; assisted by Lieutenant J. C. Ives, Corps of Topographical Engineers. ConTEntTs: ' Letter to the Secretary of War, [vii|—viii; General ** Signed by Humphreys, the ‘‘Explanatory note” reads: “The reports of the Secre- tary of War and the revising officer, which appear in volume one, were founded, so far as they relate to the route near the 35th parallel, upon the anaes report o Lieutenant Whipple. The following chapters, being those which are principally pas to in those revisory reports, are, therefore, republished. They are chapters two, three, four, aby and eleven, with a portion of appendix B. The preliminary profiles are also reprin preliminary report of Lieutenant Whipple,” see octavo edition of Pacific Railroad hee (House Ex. Doc. No. 129). For “reports of the Secretary of War and the revising officer which appear in volume one” see under that volume. Reprintings of the “preliminary profiles” have not been located in the printings of the quarto series examined. ** Page numbers in front matter preceding Part I appear to be confused but are cited as given. 48 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL table of contents,2° [ix]—x; Part I, [i]-viii; [1]-136; Part II, [1]-77; Part III, [1]-127; Part a ‘fii sai, [1]-175. The title-pages of Parts I, II, III and IV read respectively: Part I. Itinerary. [i]—viii, [1]—136. 10 plates (captioned but not numbered). 10 woodcuts. ConTENTs: Errata, [iv]; Contents, [v]—vii; Illustrations, [viii] ; Introduction, [1]—4; Itinerary (16 chapters), [5]—136 Part II. Report of the topographical features and character of the country. [3]-77. 2 plates (captioned but not numbered). 2 wood- cuts. 8 plates (diagrams, Nos. 1—8 ConTENTS: Contents, [5]; Illustrations, [6]; General sketch of the route, |7|; Sections I through VII (of the route), [8]—45; General remarks upon various subjects, |45|—77. For Maps and profiles see under Volume XI. Part III. Report upon the Indian tribes, by Lieut. A. W. wera Thomas Ewbank, Esq., and Prof. Wm. W. Turner, [1]—12 plates (captioned but not numbered). 30 woodcuts. 3 maps ae 1, 2, 3; inserted in text). ContTENTsS: Contents, [3]; Illustrations, [5]; Report (7 chap- ters), [7]—-127. Part IV. Report on the geology of the route, [i]—[xili]; [1]—175. 3 plates (1 in text, unnumbered; for Nos. I, II, fossils, see fn. 21); 18 woodcuts. Geological map. 2 Geological sections. 9 small Geo- logical sections (included in No. 2 below). ConTEeNTs: Contents, [v]|—vii; Illustrations, [viii]; Errata and addenda, [ix]; Additional errata, [xi]; Introductory letter (to No. 1), [xiii]; No. 1. General report upon the geological collections. By William P. Blake, Geologist of the Office of the United States Pacific Railroad Explorations and Surveys (10 chapters),?? [xiii]; [1]- 116; Catalogue of the geological collection, [117]—119; No. 2. Resumé and field notes, by Jules Marcou,”? Geologist and Mining Engineer to the expedition; with a translation by William P. Blake, [121]—-164; Resumé of a geological reconnaissance, extending from Napoleon, at the junction of the Arkansas with the Mississippi, to the Pueblo de los Angeles, in California. By Jules Marcou, Geol- ogist and Mining Engineer,?? [165]—175. * This covers the subject matter of the entire Whipple report, Parts I through VI and its 13 Appendices (A through I, K through N), but only Parts I natn IV are included in Volume III. For Parts V, VI ae Appendices see under Volume IV. *1 Chapter IX, [99]-105, is headed: ‘Descriptions and notices of the oss collected upon the route. By Professor James Hall.” It contains 2 plates (Nos. I, II). “A “Note,” [121], reads: “This paper is a copy of Mr. Marcou’s field book, and is an exact transcript of the original rough notes as they were taken while on the road 3 Fn. 1, [165], states: “This resumé . is reprinted from the preliminary or first report of Lieutenant Whipple, in 8vo., Chap. VI, p. 40. (Reports of Pacific Railroad Surveys, House Doc. 129: Washington. 1855.)” 1959] McKELVEY, PACIFIC RAILROAD REPORTS 49 Volume IV — 1857 33d Cong., 2d Sess., House Ex. Doc. No. 91 [Serial No. 794] The title-page of Pacific Railroad Reports, Volume IV, is dated 1856, but the “Introduction” to Part V, No. 4, [59], is dated January 12, 1857.24 Volume IV has three sets of page numbers. It is a continuation of the Whipple report begun in Volume III and includes Part V, Part VI, No. 1, and “Appendices to report.” These sunyeons are distinguished here as (1), (2) and (3) and their title-pages rea (1) Part V. Report on the botany of the expedition. fi] vii: [1]-195. CONTENTS: Contents, |[v|—vii; Nos. 1, 2, 3, 4, [1]—167; Index to botany, [169]—182; No. a, [185 ]- 193. For included betaateal profile and for plates see under Nos. 2, 3, 4, 5, elow. Part V includes five papers, headed respectively: No. 1. General description of the botanical character of the coun- try. By J. M. Bigelow, M.D. [1]-16. No. 2. Description of forest trees. By J. M. Bigelow, M.D. [17]- 26. Botanical profiles. No. 3. Description of the Cactacee. By George Engelmann, M.D., of St. Louis, and John M. Bigelow, M.D. [27]—58. 24 plates (Nos. I-XXIV). ConTENTS: No. 3, |27|—53; Explanation of the plates of Cacta- cee, [54]—58 No. 4. Description of the general botanical collections. By John Torrey. [59]—182. 25 plates (Nos. I-XXV). NTENTS: Introduction, [59]; No. 4, [61]—161; Explanation of the plates, [163]—167; Index to botany, [169 ]—182. No. 5. Description of the mosses and liverworts. By W. S. Sullivant. [185]-193. 10 plates (Nos. I-X). ConTENTsS: No. 5, [185]-191; Explanation of the plates, 192- 193. (2) Part VI. Report on the zoology of the expedition. [1]-—17. This includes one paper, headed: No. 1. Field notes and explanations. By C. B. R. katie M.D., Physican and Naturalist to the expedition. [5]—-17 An inserted sheet, following p. 17, bears a “Note.” reading: “The remainder of the Zoological hae will appear in a subsequent volume, it being impossible to prepare it in time for publication in connexion with the other portions of this report.” For Part VI, Nos. 2, 3, 4, 5, see under Volume X. ce ** See under Whipple report, Part VI. Zoological 4 ee X, [7], fn.), sige i be stated — it would seem incorrectly —- that . . appeared in Vol. IV, 50 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL (3) Appendices to report. [1]—288. These 13 appendices are distinguished by the we A through I, Kk through N, and consist, according to a “Note” , of ‘ the original unreduced records of the astronomical, magnetic, a sjetearoloaieal observa- tions. . Volume V — 1857 33d Cong., 2d Sess., Senate Ex. Doc. No. 78 |Serial No. 762 | The title-page of Pacific Railroad Reports, Volume V, is dated 1836, but the title-page of Part II. Geological heat is dated 1857 and Blake's letter trans- mitting that report is dated Apr 57 Volume V has four sets of page numbers The title-page, covering the content of Volume V, reads: Report of explorations in California for railroad routes, to connect with the routes near the 35th and 32d parallels of north latitude, by Lieutenant R. S. Williamson, Corps of Topographical Engineers. Front matter preceding and following this title-page and related to William- son’s report in its entirety appears on 12 unnumbered pages and includes the following: Contents of Volume V; Letter to the Secretary of War; General table of contents; etter, Instructions from the War Department The ‘General table of contents” divides the report into Parts I, II, II and IV, and three Appendices (A, B, C). These subdivisions, distinguished here as (1), (2), (3), (4) and (5), are cited on their respective title-pages as follows: (1) Part . Report. |1]—43. 12 plates (Nos. I-XI1). 12 wood engravings. CONTENTS: a [5]; Illustrations, [5]; Report, |7]—43 For maps and profiles see under Volume XI. (2) Part II. Geological report, by William P. Blake, Geologist and Min- eralogist to the expedition. [i]—xvi; |1]-370; [i]—xiii. ConTENTS: Blake letter submitting the report, [v]; Contents, [| vii]—xii; [Contents of] Appendix, [xiii]; Illustrations, xiv—xvi; I. Itinerary, |1]—130; II. Geology of portions of the route, | 131]—310; Appendix (Articles I through VII), [311]-370; Index, [i]—xin. For maps, sections, plates, wood engravings, see under I. Itinerary; II. Geology of portions of the route; and Appendix (Articles I. II. III, V and VII). The three subdivisions of the Geological report are cited on their respective title-pages as follows: I. Itinerary, or notes and general observations upon the geology, mineralogy, and agricultural capabilities of the route (10 chapters, I-X), [1]-130. 9 plates (Nos. I-IX). 53 wood engravings. II. Geology of portions of the route (10 chapters, XI-XX). [131 ]- 310. 8 plates (3 maps, 4 views, section). 34 wood engravings. Geological sections. Geological map Appendix. [311]-370. For plates see under Articles I, I, III, V, VII, below. 1959] McKELVEY, PACIFIC RAILROAD REPORTS 51 The Appendix includes seven Articles (of a contemplated eight) which are headed respectively : Article I. Notice of fossil fishes. By Professor Louis Agassiz. [313]-316. 1 plate (No. 1) Article II. Description of the fossil shells.2? By T. A. Conrad. [317]-329. 8 plates (Nos. II-IX Article III. Catalogue of the recent shells, with descriptions of the new species. By Augustus A. Gould, M.D. [330]-336. 1 plate (No. XI). Article IV. Letter from Professor J. W. Bailey, describing the structure of the fossil plant from Posuncula River. [337]. For plate (No. XII, figs. 1, 2) see under Article V. Article V. Description of the fossil wood from the Colorado Desert. By Prof. Geo. E. Schaeffer. [338]-339. 1 plate (No. XII, figs. 1-4). ase VI. Chemical examination of soils and incrustations. D. Easter, Ph.D. [340]-358. sie eaten Agte le VI, 340-343; Catalogue of the geological collection with descriptions of several of the specimens, [343 |-— Article VII. Description of plants collected along the route, by W. P. Blake, and at the mouth of the Gila.*° By John Torrey. {359|-370. 10 plates (Nos. I-X). Contents: Article VII, [359]-367; Description of the plates, | 368 |—370. Article VIII." (3) Part III. Botanical report: by E. Durand and T. G. Hilgard, M.D. [1]-15. 18 plates (Nos. I-XVIIT). (4) Part IV. Zoological report.?* ates: ‘‘These descriptions were published = 1855. See Appendix to the Preliminary Geological Report, 8vo; Washington, 1855.’ * An introductory paragraph, [359], signed by Torrey, mentions id oe Sean ollecte Tuma, by included, in addition to those of Blake, plants “. . . co d near t ¥ Major Thomas and Lieutenant Du Barry, of the gee States army ” * A page inserted at this point bears a ‘‘Note” reading: “Atl VII was not received in time for publication with the other eatin mf this rep Article VIII has been cited in the contents of the Appendix Par II. iscsi report, [xiii] ), as: Ratan of fossil microscopic organism, from Monterey: Professor J. W. Bailey.” It has not been found in the series of Pacific Railroad Reports examined. “SA “Note” following the above title-page states: ‘“‘The Zoological Report will r in a subsequent volume, it being foun sii cena = prepare it in time for por included the followin papers: “No. 1. Mammals, by oie S. EF, Baird” NO w.2: Birds, by Dr. A. L. Heermann, Physician and Naturalist to the expedition’; “No 3. 52 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XL (5) Appendices. [1]—14. ContTENTs: Appendix A. Distances and altitudes, [3 |—4; Appendix B. Latitudes and longitudes, [5]; Appendix C. Data for profiles, [6|—14. Volume VI — 1857 33d Cong., 2d Sess., Senate Ex. Doc. No. 78 |Serial No. 763 | The title-page of Pacific Railroad Reports, Volume VI, is dated 1857. This date is corroborated in a footnote to “Contents” of Abbot’s zoological report, (Volume X, [7]), which reads: “Numbers 1, 2, and 3 [of Part IV. Zoological report, my (4) below | Hi 7 found i in Vol. VI of the Pacific Railroad Surveys, printed and published in 1857 Volume VI has five vi of page ‘numbers. The title-page covering the content of Volume VI reads: Report of Lieut. Henry L. Abbot, Corps of Topographical Engineers, upon explorations for a railroad route, from the Sacramento Valley to the ‘Columbia River, made by Lieut. R. S. Williamson, Corps of Topographi- cal Engineers, assisted by Lieut. Henry L. Abbot, Corps of Topographi- cal Engineers. CoNTENTS (front matter): Letter to the Secretary of War,”® [3]|-4; General table of evens, [5]; Introduction. Instructions from the War sesame [7]|-1 Abbot report five main subdivisions, distinguished here as (1), (2), (3), (4) and (5). Their respective title-pages read: (1) Part I. General report. [17]—134. 12 plates (Nos. I-XI, XIII). 3 woodcuts ConTENTS: Contents, [21]—23; List of illustrations, [24]; General report (7 chapters), [25]- 129; Index to General report, |130|—134. For maps and profiles see nas Volume XI. (2) Part II. Geological report, |1|-85. CoNTENTS: Contents, [5|—7; List of illustrations, [8]; Nos. 1, 2, 3, and 4 of Geological report, [9]-85. Part II contains four papers, headed respectively: No. 1. Report upon the geology of the route. By J. S. Newberry, M.D., Geologist of the expedition (8 chapters), [9]—68. 1 plate (No. I). 11 woodcuts. No. 2. Description of the tertiary fossils collected on the survey. By T. A. Conrad. [69|—73. 4 plates (Nos. II-V) No. 3. Report upon an analytical examination of water and Reptiles, by Dr. Edward Hallowell”; ‘““No. 4 Fishes, by Dr. Charles Girard.” For these a see under Volume X. ” The Hon. Jefferson Davis held the post of Secretary of War from March, 1853, to March, 1857. The letter is addressed to the Hon. John B. Floyd who had been appointed to that office by President Buchanan in March, 1857. 1959] McKELVEY, PACIFIC RAILROAD REPORTS 53 minerals from the Hot Springs in Des Chutes Valley. Conducted under the direction of Prof. E. N. Horsford.*° [74]-78. No. 4. Catalogue of the minerals and fossils collected on the survey. [79]-85 (3) Part III. pai Bigs {| 1]—10 ConTENTS: Contents, [5]; List illustrations, [7]-8; Nos. 1 and 2 of Botanical a. 19] —94; Index to Botanical report, [97 ]-102. For plates and woodcuts see under Nos. 1 and 2 below. Nos. 1 and 2 are headed respectively : No. 1. Report upon the botany of the route. By John S. Newberry, M.D., Botanist of the expedition (2 chapters), [9 |—64. Ghapter I. Geographical botany. |[9]|-—19. Chapter IT. ee of the forest trees of northern California and Oregon. [20]-64. 10 plates (Nos. I-X). 28 woodcuts. No. 2. General catalogue of the plants collected on the expedition. By J. S. Newberry, assisted by Asa Gray and John Torrey, as specified in the proper places, [65 ]—94. No. 2 has four subdivisions: Exogenous plants.*! [65]-90. 6 plates (Nos. XI-XVI). II. Endogenous plants. By John Torrey, |90]—92. III. Mosses and liverworts. By W. S. Sullivant, 93-94. IV. Lichens. By Edward Tuckerman, 94. — (4) Part we Zoological aa ie [1]-114. CONTENTS: Cont [5]; List of illustrations, [6]; Prefatory note to "Part IV, 7]: Se 1, 2, and 3 *? of Zoological report, [9 ]—114; Index to Zoological report, fi jaiv. The subdivisions included in the Zoological report are the following: No. 1. Report upon fishes collected on the survey. By Charles Girard, M.D. [9|-34. 11 plates. (Nos. XXIIa, XXIIb, XXVa, XXVb, XLa, XLVI, LXII, LXVI, LXVIII, LXX, LXXIV). No. 2. Report upon the se of the route. By J. S. Newberry, M.D.** (2 chapters), [35]—110. For allocation of included plates see under Chapters I, II, below. The chapters are headed: A peer letter; [74], refers to this as ‘ . the report of my assistant, Mr. L. M. Dor ; in Canin ” [5], of Botanical report, the authors of “Exogenous plants” are named as Asa Gray, John Torrey, and J. S. Newberr * For omission of ah en No. 4, which had been ainda in “Contents,” [5], ee following No. 3 A fn., [38], states: ae species para ae e in this report have been determined, and their See prepared, by Prof. S. F. Baird, Assistant Secretary of the Smith- sonian Institutio 54 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Chapter I. Report upon the mammals: [35]—72. 3 plates (Nos. I, III, XXII). Chapter II. Report upon the birds.*# [73]-110. 2 plates (Nos. XXVI, XXXIV. No. 3. Report upon the land shells collected on the survey. By W. G. Binney, Member of the Academy of Natural Sciences of saa “age [111]—-114. Jote” on an inserted page, following p. 114, states: “No. 4 has not been sonpiied in time for publication with the rest of this Report. It will appear in a or. volume.” No. 4 had been cited in the Contents, ” [5], of Part IV. Zoological report as: “Report upon reptiles collected on the survey. By S. F. Baird, Assistant Secretary of the Smithsonian Institution.” For this see under Volume X. — Ww 4 Appendices. [1]—64. Distinguished by letters A through F, these 13 Appendices relate to astronomical observations and the like. They are enumerated in “General table of contents,” [5], of the Abbot report. An inserted page following the Appendices lists “Errata for Volume VI.” Volume VII — 1857 33d Cong., 2d Sess., Senate Ex. Doc No. 76 {Serial No. 764] The title-page of Pacific Railroad Reports, Volume VII, is dated 1857. No included matter bears a later date. Volume VII has five sets of page number “Contents of Volume VII,” [3], records —- subdivisions, distinguished here as (1) and (2). The title-page of the first reads: (1) Report of explorations for railroad routes from San Francisco Bay to Los Angeles, California, west of the Coast Range, and from the Pimas Villages on the Gila to the Rio Grande, near the 32d parallel of north latitude, by Lieutenant John G. Parke, Corps of Topo- graphical Engineers, assisted by Albert H. Campbell, Civil Engineer. ConTENTS: Letter to the Secretary of War, (7); General table of contents, [9]; Introduction. Instructions from the War Department, f11]—-15; Part. I. General report, [17]—[23],*° [1]-12; Part I. Geological report, [1]-204; Part III. Botanical report, [1]—28: Appendices (A through E), [1]—118. % A fn., [73], reads: “As the final determination of the species of birds collected by the expedition has not yet been completed by Prof. Baird, the names here given are to be considered as temporary. In his general Rosie upon the birds of the Pacific Railroad Surveys, hereafter to appear, the names and pages of the species in this article will be carefully quoted, and any errors “oh determination thus rectified.” For Baird’s “general report upon the birds” see under \ e IX. *® The 23 pages (8 blank) sae precede the text proper of Part I. General report, are, with the exceptions of pages 14, 15, 22, unn mbered. The content of some relate to Parke’s report as a whole, Hn others ead front matter to Part I. I have inserted the missing page numbers and have allocated the subject matter to accord with my understanding of its proper association. 1959] McKELVEY, PACIFIC RAILROAD REPORTS 55 For maps and profiles see under Volume XI. Parts I, II, III, and Appendices are cited on their respective title-pages as follows: Part I. General report. [17]—[23]; [1]-42. 8 plates. For allocation of the plates see under Nos. 1 and 2 below. For maps and profiles see under Volume XI. ConTENTS: Contents, [21]—[23]; [1]—23; List of illustrations, [23]; Nos. 1 and 2, [1]|-42. The two subdivisions of the General report are headed: No. 1. Report of explorations from San Francisco Bay to Los Angeles. [1]-18. 3 plates (Nos. I, II; the third not numbered). No. 2. Report of explorations from Pimas Villages to Rio Grande. [19]-42. 5 plates (Nos. VI [= IV], V, VI, VII, VIII). Part II. Geological report, by Thomas Antisell, M.D. ao of the expedition. [1]—204. 2 maps (geological). 24 plate For allocation of plates see under Chapters XXIX and XXX below. CoNTENTS: Contents, [5]-10; List of illustrations, [11]—13; Geological report (30 chapters), [15]-197; Description of the plates, [19]-2 Of the 30 chapters of the Geological report, Chapter XXIX (“Report upon the palaeontology of the survey. By T. A. Conrad”), [189]-196, contains 10 plates (Nos. I-X); and Chapter XXX (‘Order of position of Californian strata”), [197], contains 14 plates (Nos. I-XIV). Part III. Botanical report: by John Torrey. [1]—28. 8 plates. For allocation of plates see under Chapter I below. CoNnTENTS: Contents,?° [5]; Chapters I, II, II, [7]-28. The subdivisions of the Botanical report are headed respectively: Torrey, M.D. [7]—22. 8 plates (Nos. I-VIII Chapter II. Synoptical tables of botanical localities. By Thomas Antisell, M.D. [23 ]-26 Chapter I. List and descriptions of the plants collected. By John ). Chapter ITT. Deseuition of the plates [27 ]-28. Appendices. [1]—-116. 11 plates. For allocation of plates see under Appendix A. The five Appendices are headed respectively: . By John Torrey, M.D. Appendix A. Remarks on meteorology and barometic results, with meteorological plates. By Albert C. Campbell, A.M. Civil Engineer to the expedition. [3]—14. 11 plates (Nos. I-XI). Appendix B, Report upon the route from San Diego to Fort * This contains a list of plates. 56 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Yuma via San Diego River, Warner’s Pass, and San Felipe Canon. By Charles M. Poole, Chief Engineer. [15]-28. Appendix C. Translation of an archive from Tucson. [29]—30. Appendix D. ao of maps and profiles, with table of latitudes. [31]-3 Appendix E. ae of distances, and barometric and meteorologi- cal observations and results. By Albert H. Campbell, A.M., Civil Engineer to the expedition. 35-116. The title-page of the second subdivision of Volume VII reads: (2) Conclusion of the official review of the reports upon the explorations and surveys for railroad routes from the Mississippi River to the Pacific Ocean. [1]-37. For General map and General profiles see under Volume XI. ContENTS: Contents, [3]; Prefatory note, [5]; Nos. 1, 2, 3, 4, 5, [7 |-37. The five subdivisions of the “official review” are headed respectively: No. 1. Extract from the annual report of the Secretary of War. December, 1855. [7]-—10. No. 2. Report upon the progress of the Pacific Railroad explorations and surveys. November, 1855. By Captain A. A. Humphreys, Corps of Topographical Engineers. [11]—18. No. 3. Extract from the annual report of the Secretary of War, December, 1856. [19]—22. No. 4. Report upon the progress of the Pacific Railroad explorations and surveys. November, 1856. By Captain A. A. Humphreys, United States Topographical Engineers. [23 |—34. No. 5. Table exhibiting the comparative lengths, cost, etc., of the different routes, with explanatory remarks. By Captain A. A. Humphreys, Corps Topographical Engineers. [35]—37 Volume VII ends with an inserted page: “Errata for Volume VII.” Volume VIII — 1857 33d Cong., 2d Sess., Senate Ex. Doc. No. 79 (Serial No. 765) The title- -page of Pacific Railroad Reports, Volume VIII, is dated 1857. This date is corroborated in the “Preface” to Part II. Birds (Volume IX, [xiii]) which refers to the account of the mammals (Part I. below) as “. . having been published in 1857. The pages of Volume VIII are ‘numbered consecutively throughout. Following “Contents of Volume VIII,” [iii], is a title-page reading: General report upon the zoology of the several Pacific Railroad routes. CoNTENTS: Letter to Captain Humphreys, Topographical Engineers, 1959] McKELVEY, PACIFIC RAILROAD REPORTS 57 [vii]; General table of contents,®* [ix]; Introduction. General sketch of lines explored,** [| xi]—xvii. The title-pages covering the content of Volume VIII read: Part I. Mammals: by Spencer F. Baird, aa Rasetca of the Smithsonian Institution. [xix |—xlviii; [1 |—75 woodcuts. 43 plates. (Nos. XVII-XXVIII, XXX-LX). CoNTENTS: Contents, [xxiii]; Preface, |xxv]—xxix; Table of higher groups, xxx—xxxi; List of species, |xxxii|—xlviii; Introductory remarks, [1]; [Orders of mammals, I through VI], [3]-684; Appendix . List of authorities referred to in the preceding report, [685]—703; Appendix B. Alphabetical list of localities referred to in the preceding report, [704]—715; Alphabetical index. I. Systematic index of common names, |717|—720; Alphabetical index. II. Systematic index of scien- tific names, [721]—735; Explanation of the plates,®® [737]—750; List of wood-cut figures, 750-752; Systematic list of illustrations, 755-757. Volume IX — 1858 33d Cong., 2d Sess., Senate Ex, Doc. No. 78 [Serial No. 766] The title-page of Pacific Railroad Reports, Volume IX, is dated 1858. No included matter bears a later The pages of Volume IX are numbered consecutively throughout. It is a continuation of the “General report upon the zoology of the several Pacific Railroad routes” begun in Volume I. Following “Contents of Volume IX,” [iii], and “Letter to Captain Humphreys, Topographical Engineers,” [v], are title- -pages readings: Part II. Birds: by Spencer F. Baird. Assistant Secretary Smithsonian Institution. With the co-operation of John Cassin and George N. Law- rence. [vii]—lvi; [1]- CoNTENTS: Contents, [xi]; Preface,4° [xiii]-xvi; I. Table of the * This cites the content of Volume VIII (Part I. Mammals) ; of Volume IX (Part II. Birds) ; and of Volume X (as to Part III. Reptiles and Part IV. Fishes). along the m nected” therewith. Included also are men who, working independently, made collections contributing to the report. For a somewhat similar paper see “Preface” (Volume IX, [xiii]—xvi). * A fn., [747] explains to which reports the plates of mammals enumerated in the seca reports of the different railroad parties. . .”, and tells how plates waepciated with special reports should be distinguished. A list of officers in command of the different lines of survey and the zoologists associated with them, is also supplied. Despite the explicit directions, attempts to check the plates were not successful. “ This useful enumeration of collectors is somewhat similar to the one supplied in “Introduction. General sketch of lines explored” (Volume VIII, [xi]-xvii) for the reason that some of the men collected in more than one branch of natural history. But it contains additional names and covers some different territory. 58 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL higher groups, [xvii]-xxiv; II. List of species, |xxx|—Ivii; Introductory remarks, [1]—2; [Orders of birds, I through VI], [3]—920; Appendix A. Additional remarks on North American birds, [921]-925; Appendix B. Birds found at Fort Bridger, Utah, [926]—937; Appendix C. List of authorities referred to in the preceding report, [928]—954; Alphabeti- cal index. I. Systematic index of common names, [955 ]—963; Alphabeti- cal index. II. Systematic index of scientific names, [965 |—1005 Volume XK — 1859 33d Cong., 2d Sess., Senate Ex. Doc. No. 78 |Serial No. 767] The title-page of Pacific Railroad Reports, Volume X, is dated 1859. This date is corroborated in a footnote to the “Contents” of Abbot’s report (Part IV. Zoological report, Volume X, [7]), which states: “No. 4 Patt on reptiles col- lected on the survey. By S. F. Baird] dates 1859.” See fn. 49. Volume X has seven sets of page numbers “Contents of Volume X,” [3], names two subdivisions: first, Parts III and IV of the “General report upon the zoology of the several Pacific Railroad routes” (continued from Volumes VIII and IX); and, second, five groups of zoological papers which had not been ready when the reports with which they are affiliated has been published; they are distinguished here as (1), (2), (3), (4), and (5). The title-pages of Parts III and IV read respectively: Part III. Reptiles: by Spencer F. Baird, Assistant Secretary of the Smithsonian Institution, [7]-18. 13 plates (Nos. XXIV-XXXVI). Contents: Letter of Humphreys,*! [11]; Explanation of the plates, [13 |—16. Part IV. Fishes: by Charles Girard, M.D. [i]—xiv; [1]-400. 21 plates (Nos. VII, VIII, XIII, XIV, XVII, XVIII, XXIIc, XXVI, XXIX, XXX, XXXIV, XXXVH, XL, XLI, XLVII, LI, LIX, LXI, LXIV, LXV, LXXI). ContTENTS: Contents, |v|—xiv; Introductory remarks,** [1]-2; {Orders of fishes, I through X], tele ~284: List of the plates,*® [385 |- 388; Alphabetical index, 389-400 The title-pages of the five — of reports forming the second subdivision of Volume X read respective (1) Report of Lieut. E. G. Beckwith, Third Artillery, upon explorations for a railroad route, near the 38th and 39th parallels of north latitude, *. This reads: ‘The General Natural History Reports having been extended so much beyond the limits originally contemplated, the War Department has considered it advisable to omit the publication of the Report on Reptiles. The plates to accompany this report having been prepared and printed, they, with a brief explanation of the hgures a is them, are herewith given.’ “These appear in two forms in the printing under analysis. One was presumably a aaa draft and should have been omitted. ‘3A jn. states: “Plates XXIIa, XXIIb, XXVa, XLa, XLVI, XLVII, LXVI, LXVIII, LXX and LXXIV of this list will be found in vol. VI, part IV, of the present series. Other plates missing from this volume will be found in vol. X.” 3 5 1959] McKELVEY, PACIFIC RAILROAD REPORTS 59 by Captain J. W. Gunnison, Corps of Topographical Engineers, and near the forty first parallel of north latitude, by Lieut. E. G. Beck- with, Third Artillery. Zoological report.44 20 plates. For allocation of the plates see under Nos. 1, 2, 3, and 4 below. ConTENTS: Contents,’ [5], List of illustrations, [6]; Nos. 1, 2, 3, and 4, [7]—27; Alphabetical index, [28 The four included zoological reports are headed respectively: No. 1. Report upon mammals collected on the survey. By S. F. Baird. [7|—9. 3 plates (Nos. IV, VI, X). No. 2. Report upon birds collected on the survey. By S. F. Baird. [13]-16. 7 plates (Nos. XII, XIII, XIV, XV, XVII, XXXII, XXXV). No. 3. Report on reptiles collected on the survey. By S. F. Baird. [17]-20. 4 plates (Nos. XVII, XVIII, XXIII, XXIV). No. 4. Report on fishes collected on the survey. By Charles Girard, M.D. [21]-27. 6 plates (Nos. XXIII, XLIX, LIV, LVI, LXXIII, LXXV). ya Bo —m Report of explorations for a railway route (near the thirty-fifth paral- lel of north latitude), from the Mississippi River to the Pacific Ocean. By Lieutenant A. W. Whipple, Corps of Topographical Engineers: assisted by Lieutenant J. C. Ives, Corps of Topographical Engineers. Part VI.*° Zoological report. [1]-34. 34 plates. For allocation of plates see under Nos. 2, 3, 4 and 5 below. CONTENTs: Contents,*” [7]; List of illustrations, [9]—10; Nos. 2, 3,4, and 5, [11]-59; Alphabetical index, [61]-64. The four included posi reports are headed respectively: No. 2. Report on mammals collected on the survey. By C. B. R. Kennerley, M.D. [11]-18. 6 plates (Nos. VIII, XI, XII, XIII, XIV, XVI). No. 3. Report on birds collected on the route. By C. B. R. Ronee. M.D. [19]-35. 11 plates (Nos. XVIII, XIX, XX, MXIT, XXVIII, XXIX, XXX, XXXI, XXXII, XXXVI, XXXVII). “A fn, states: “The report to which this article belongs will be found in Vol. II of this series.” “Of five eae as listed, only four are included. The missing report, named in ‘‘Contents” as “No. 5. Report on insects ene on the survey. By John 1 ocaues Seay ve not been found in the quarto seri 4° For Parts I, II, III and IV, see Volume III; for ee V, and Part VI, No. 1, see Volume IY. “Of five zoological ee cited, only four are included. Of the missing report (No. 1. Field notes and explanations. By C. B. R. Kennerley, M.D.) it is stated in a fm.: “No. 1 has already appeared in Vol. IV, 1856.” According to my analysis Volume IV was published in 1857, certainly not earlier. See Volume IV, Introduction to Part V, No. 4, [59], which is dated January 12, 1857. 60 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL No. 4. Report upon the reptiles of the route. By S. F. Baird. (37|-45. 3 plates (Nos. XXV, XXVI, XXVII). . 5. Report upon fishes collected on the survey. By C. Girard, [47|-59. 14 plates (Nos. III, IV, V, VI, EX, X. XXI, oan XXV, XXXV, XLb, LII, LVII, LVIII). (3) Report of explorations for a railroad route near the 32d parallel of north latitude, lying between Dona Ana, on the Rio Grande, and Pimas Villages, on the Gila, by Lieutenant John G. Parke, Corps of Topographical Engineers. Zoological report. |1]—24. 4 plates. For allocation of plates see under Nos. 1 and 2 below. ConTENTs: Contents, [5]; List of illustrations, [6]; Introductory letter, [7]; Nos. 1 and 2, [9|-24 The two included zoological reports are headed respectively : No. 1. Report upon the birds collected on the survey. By A. L. Heermann, M.D. [9|-—21. 3 plates (Nos. I, IV, VI). No. 2. Report upon reptiles collected on the survey. By Edward Hallowell, M.D. [23]-24. 1 plate (No. IT). (4) Report of explorations in California for railroad routes to connect with the routes near the 35th and 32d parallels of north latitude. By Lieutenant R. S. Williamson, Corps of Topographical Engineers. Part IV. Zoological report, |i]—|viii]; [1]—91. 26 plates. For allocation of plates see under Nos. 1, 2, and 4 below. Contents: List of illustrations, [vi]; Contents,** [vii]; Intro- ductory letter. By A. L. Heermann, M.D. |viii]; Nos. 1, 2, 3 and 4 [1]—91; Alphabetical index, [93 ]—97. The four included zoological reports are headed: No. 1. Report upon the reptiles collected on the survey. By Dr. Edward Hallowell. |1|—23; List of reptiles collected. By Dr. A. L. Heermann, Naturalist to the expedition, |24|—25; Explanation of the plates, [26|-27. 9 plates (Nos. I, III, IV, V, VII, VII [sic], VIII, IX, X). No. 2. Report upon birds collected on the survey. By A. L. Heer- mann, M.D. |29|-—77; List of birds collected between San Francisco and Fort Yuma, California, during the survey of railroad routes from the Mississippi to the Pacific Ocean, under the command of Lieutenant R. S. Williamson, Top. Engs. [79]—80. 7 plates (Nos. II, III, V, VII, VIII, 1X, X). No. 3. Report on mammals collected on the survey. By S. F. Baird. fer) e2. No. 4. Report on fishes collected on the survey. By Charles Girard, — aed 24 ag states: “The report to which the present part belongs will be found in volume V 1959] McKELVEY, PACIFIC RAILROAD REPORTS 61 M.D. [83]-91. 10 plates (Nos. I, XII, XXII, XXVII, XXVIII, XXXI, XXXVI, XXXVIII, XXXIX, XLVII). (5) Report of Lieut. Henry L. Abbot, Corps of Topographical Engineers, upon explorations for a railroad route, from the Sacramento Valley to the Columbia River, made by Lieut. R. S. Williamson, Corps of Topographical Engineers, assisted by Lieut. Henry L. Abbot, Corps of Topographical Engineers. Part IV. Zoological report. [1]—13. 4 plates. For plates see under No. 4 below. ConTENTs: Contents,!” [7]; List of illustrations, [8]; No. 4, [9]-15. The included zoological report is headed: No. 4. Report on reptiles collected on the survey. By S. F. Baird. [9]-13. 4 plates (Nos. XI, XXVIII, XXX, XLIV). Volume XI — 1859 33d Cong., 2d Sess., House Ex. Doc. No. 91 [Serial No. 801] The title-page of Pacific Railroad Reports, Volume XI, is dated 1855, but the title-page of Warren’s Memoir as well as the title-page preceding the maps, profiles, plates, etc., which form the last half of the volume, are dated 1859. Following the “Contents of Volume XI,” [3], and “Letter to the Secretary of War,” [5], is a title-page reading: Memoir to accompany the map of the territory of the United States from the Mississippi River to the Pacific Ocean, giving a brief account of each of the exploring expeditions since A.D. 1800, with a detailed description of the method used in compiling the general map. By Lieut. Gouverneur K. Warren, Corps of Topographical Engineers, U.S.A. [3]—120. 4 plates (Nos 1. 41 Tir IV." ConTENTS: Letter to Captain A. A. Humphreys, Topographical Engi- neers, [9]; Contents, [11]—-12; List of illustrations, [13]; Introductory remarks, [13]—16; Memoir (6 chapters), [17]—115; Alphabetical index, [117 ]-120. The remainder of Volume XI is preceded by a title-page reading: Topographical maps, profiles, and sketches, to illustrate the various reports of survevs for railroad routes from the Mississippi River to the Pacific Ocean. This is followed by “Contents,” [iii]-iv, in which the illustrations, segregated under ten headings, are meticulously enumerated and described. These headings and the total number and type of illustration allotted to each in the “Contents” follow: “© A fn. states: ““Numbers 1, 2, and 3 will be found in Vol. VI of the Pacific Railroad Surv eys, printed and published i in 1857. No. 4 dates 1859.” ” These are reduced copies of old maps, intended for insertion in the text. 62 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Sketch of route of the 41st parallel — Beckwith’s Report, Vol. II. 8 plates (views). 5 panoramas (views) Official review of the reports — Vols. I and VII. General map. General profile. Route of the 47th ee 49th parallels — Stevens’s Report, Vol. I. 3 maps (Nos. 1, 2, 3). Profi Routes in Oregon a California — Abbot’s Report, Vol. VI. 2 maps (Nos. 1, 2). 2 profiles (Nos. 1, 2). Route near the 41st parallel — Beckwith’s Report, Vol. II. 4 maps (Nos. 1, 2,3, 4). Profiles of the route Route near the 38th and 39th parallels — Beckwith’s Report, Vol. II. Pro- files of the route. 4 maps (Nos. 1, 2, 3, 4) Route near the 35th parallel — Whipple’s Report, Vol. III. 2 maps (Nos. 1, 2). Profiles Routes in southern California — Williamson’s Report, Vol. V. 4 maps. 2 profiles (sheets Nos. 1, 2). Routes in California on the 32d parallel — Parke’s Reports, Vol. VII. 2 maps (Nos. 1, 2). Profiles Route of the 32d il Pope’s Report, Vol. IIT. Map and profile. Volume XII. Book I °!' — 1860 36th Cong., 1st Sess., House Ex. Doc. No. 56 [Serial No. 1054| The title-page of Pacific Railroad Reports, Volume XII, Book I, is dated 1860. No included matter bears a later The volume has two sets of page numbers. “Contents - ounne XII. Book I,” [3], records two subdivisions: Part I and Appendice Following “Contents of Volume XII. Book I” is a title-page reading: Narrative and final report of explorations for a route for a Pacific railroad, near the forty-seventh and forty-eighth parallels of north latitude, from St. Paul to Puget Sound. By Isaac I. Stevens, Governor of Washington Territory. This is followed by “Letter to the Secretary of War,” by the “General table of contents” and by an “Introduction. " The first Stevens report appeared in Volume I. His second and final report and the scientific papers associated with it necessitated two volumes which are distinguished on the title- “pasts of Pacific Railroad Reports as “Volume XII. Book I” and ‘‘Volume XII. ok II.” “ The subdivisions are in accord with the contents of Book I as issued. For the reason that, with one exception, none of the pages in Volume I preceding page 20 is numbered and the exception, page “18,” bears no numerical relationship to those which precede or follow it, and for the reason that the “General table of contents, it seems probable that there were “last minute” changes in the compilation of Books I vial II which were not incorporated in the introductory pages of Book I. See also . 34, 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 63 The title-page of the first subdivision of Volume XII, Book I reads: Part I. General report. [15]—358. 70 plates (Nos. I-LXX). 2 maps. 1 a a Sean ts: Contents, [19]—27; List of illustrations, [29]—30; Gen- eral pot (18 chapters), [31] =456. oa The second subdivision of Volume XII. Book I is preceded by a title-page reading: Appendices. [1 ]—25.°4 CONTENTS: ndix A. Heights and distances. [3]—12; Appendix B. Meteorological register, [13 ]—25. Volume XII. Book II — 1860 36th Cong., 1st Sess., House Ex. Doc. No. 56 |Serial No. 1055] The title-page of Pacific Railroad Reports, Volume XII. Book I, is dated 860. No included matter bears a later date. The volume has two sets of page numbers. It is a continuation of the Stevens report begun in Volume XII. Book I. “Contents of Volume XII. Book II,” [3], divides the subject matter into two Parts which are cited on their respective title-pages as follows: Part II. Botanical report. |5]—76. 6 plates. For allocation of plates see under Nos. 2 and 3 below. ConTENTS: Contents, [9]; List of illustrations, [11]; Nos. 1, 2 and 3 of Botanical report, [13]—71; Index to Botanical report, [73 |—76. The three included papers are headed respectively: No. 1. Report on the botany of the route. By J. G. Cooper, M.D. [13 ]-39. No. 2. Catalogue of plants collected east of the Rocky Mountains. By Professor Asa Gray. [40]-49. 5 plates (Nos. I, I, III, IV, V). No. 3. Catalogue of plants collected in Washington Territory. By J. G. Cooper, M.D. [50|]-71. 1 plate (No. VI). Part III. Zoological report. [i]—[ix]; [1]-399. 46 plates. For allocation of plates see under Nos. 1, 2, 3, 4, 5 below. ContTENTS: Contents, [v|—-vi; List of illustrations,” [vii]—vil. °° The “General report” is subdivided into Narrative of 1853 (Chapters I-X, [31]- 195); Narrative of 1855 (Chapters XI-XII, [196]—225) ; Geographical memoir (Chap- ters XIII-XVI, [226]-331); Railroad feb art and estimate (Chapter XVII, [332]- 351); Computation of altitudes from barometrical observ ations. Table of heights and distances. Meteorological register (Chapter XVIII, [352]-358). Following the Appendices and with pages in numerical sequence thereto is an Alphabetical index, [27]-41. Its content does not cover the Appendices and its existence = is enumerates between 80 and 90 plates, more than half of which are referred to zoological reports contained in Volume XII. Book IJ, and the remainder to zoological reports published in earlier volumes of the series. Attempts to check the enumeration 64 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XL Prefatory note to Part III, [ix]; Nos. 1 through 7 of Zoological report, |1]-389; Index to Zoological report, [391 ]—399. Part III contains 7 papers, headed respectively: No. 1. Report upon insects collected on the survey. By John T. Leconte, M.D. [1]—77. 2 plates (Nos. I, II). No. 2. Report upon the mammals collected on the survey. [73]—138. 5 plates. For allocation of plates see under Chapters I and II below. No. 2 contains 3 reports headed respectively Chapter I. Report by J. G. Cooper, M.D. [73]—88. 4 plates (Nos. V, VII, IX, XV). reseed II. Report by Dr. Geo. Suckley, U.S.A. [89]-106. 1 plate IT). (No. Chapter ITT. al of Dr. Geo. Suckley, U.S.A., and Geo. Gibbs, Esq. |107]|-1 No. 3. Report upon the birds collected on the survey. [140]—291. 8 plates. For allocation of plates see under Chapters I, II, below. No. 3 contains 2 subdivisions, headed respectively: Chapter I. Land birds, by J. G. Cooper, M.D. pee ce 7 plates (Nos. XI, XVI, XXVIII, XXI, XXIII, XXIV, Chapter IT. Water birds, by Dr. G. Suckley, Toe [227]-291. 1 plate (No. XXXVIII). No. 4. Report upon the reptiles collected on the survey. By J. G Cooper, M.D. [292]-306. 11 plates. (Nos. XII, XIII, XIV, XV, XVI, XIX, XX, XXI, XXII, XXIX, XXXI). No. 5. Report upon the fishes collected on the survey. By Dr. G. Suckley, U.S.A. [307]-368. 21 plates (Nos. I, XI, XV, XVI, XIX, XX, XXXII, XXXIII, XLII, XLIII, XLIV, XLV, L, LI, LV, LX, LXIII, LXVII, LXIX, LXXII, LXXV). No, 6. Report upon the Mollusca collected on the survey. By William Cooper. [369 |]—386. No. 7. Report on the Crustacea collected on the survey. By J. G. Cooper, M.D. [387]—389 ALPHABETICAL INDEX OF AUTHORS TO BE USED IN CONJUNCTION WITH THE ANALYSES 56 Aspot, H. L. (Sacramento Valley to Columbia River): Report, v. VI. Acassiz. L. (Williamson route): Fossil fishes, v. VI. of this List with plates in the given reports have not been satisfactory, either as to the total number of plates or as to their allocation. See fn. 39, “To the first Stevens report published in Volume I are appended (pp. 160-635), 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 65 ANTISELL, T. (Parke route): Geological report; Synoptical table botanical local- ities. v. VII. *ARNOLD. R. 160-177, BaILey. J. W. (W aie eon route): Fossil plant Posuncula River, v. V. Batrp, S. F. (Pope route): Preliminary report natural history, v. I. (General report on zoology): Mammals, v. VIII. Birds, v. IX. Reptiles, v. X. (Beckwith and Gunnison routes): Mammals; Birds; Reptiles, v. X. (Whipple route): Reptiles, v. X. (Williamson route): Mammals, v. X. asi route): Reptiles, v. X. BeckwitH, E. G. (41st aarti route): Report, v. II. (Gunnison route, 38th and 39th parallels): Report, v BiceLow, J. M. (Whipple pout: Botanical character country; Forest trees, eee a BIcELow, J. M. See ENGLEMANN, G. Binney, W. G. (Abbot route): Land shells, v. VI BLakE, W. P. (Pope route): Geology, v. II. (Whipple route): peners report geological collections, v. 1. (Williamson route): Geological report, v *BiopceT, L. 566-571, v. I. *Bur. F. H. 586-597, v. 1. Byrne, T. H. (Pope route): Diary, v. II. CAMPBELL, A. H. (Parke route): Table distances, barometric observations, v. a Conrap. T. S. (Williamson route): Fossil mine. v. V. (Abbot route): Tertiary fossils, v. VI. (Parke route): ae .V ee J. G. (Stevens route): see plants Washington Terri- ; Mammals; ne birds; ane Crustacea, v. XII. Book II. See also: ie 180: 219-221 Cooper. W. (Stevens aan Mollusca, v. XII. Book II. AvIS, J. Report on es several railroad explorations, v. I. *DonELson, A. J. ~247: 269-279; 358-359; 360-363, v. I. DornBacu. L. M. te route): Examination water, minerals, v. VI. *Doty, J. 441-442; 442-445; 445-446; 543- 553; 553-565; 572-584, v. I. *DUNCAN a); K: 7203-2195 v Duranp. E. and Hitcarp, T. G. (Williamson route): Botanical report, v. V. FASTER, J. D. (Williamson route): Chemical examination soils, incrustations; Eso W.H. Extract from report of a military reconnaissance made 1846, 1847, ie eee G. and BiceLow, J. M. (Whipple route): Cactaceae, v. IV. Ewprank. J. See WHIPPLE, A. W. together with letters, etc., etc., a number of signed documents describing work done in connection with particular sections and aspects of his survey; some of these (pp. 160-449) were appended to the report when submitted; others (pp. 449-635) were added at a later date. In my analysis of the content of the report these papers are merely noted as “Miscellaneous documents related to special aspects of the survey, 160-635”, and neither the authors nor the long descriptive titles of their papers are appear in the Stevens report 2 Volume I are supplied. In instances where authors contributed papers which are cited in the analyses, and app — ae also, the first take precedence and the last are preceded by the notation See al. 66 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL *GipBs, G. 402-434; ie 473-486, v. I. Gipss, G. See SUCKL Grrarp, C. (Abbot set “Fishes, v. VI. (General report on zoology): Fishes, v. X. (Beckwith and Gunnison routes): Fishes, v. X. (Whipple route): Fishes, v. X. (Williamson route): Fishes, v. X. GouLp, A. A. ene: route): Recent shells, v. \ ~~ A. (Stevens route): Catalogue plants from east of Rocky Mountains, . XII. Book II. Grav. A., ae J.. and Newserry, J. S. (Abbot route): Exogenous plants, VI. Gray, A. See Torrey, J. *GROVER, C. 222-223; 247-248; 396-398; 486-488; 488-498; 498-515. v. I. GUNNISON, J. W. For route of, see BECKWITH, G. Hatowent, E. (Parke route): Reptiles, v. X. (Williamson route): Reptiles, eon: A. L. (Parke route): Birds, v. X. (Williamson route): Birds. v. X. Hixtcarp, T. C. See DuRAND Humpuereys, A. A. Official review reports and explorations, 1855, 1856, v. VII. Humpnreys, A. A. and Warren, G. K. Examination of reports of explorations made 1853, 1854, and earlier, v. I Hunt, E. B. Index to Stevens pate v. I. Jesup, T. S. Cost transporting troops, supplies, v. I. mesic C. B. R. (Whipple route) : Zoological field notes, v. IV. (Whipple >: Mammals; — v. X. co J. 160-177, *LANDER, F. W. Reconnaissance Puget Sound to Mississippi River, v. IT. See also: 186-187; 224-225, Leconte, J. T. (Stevens ee Insects, v. XII. Book II. *MACFEELEY, R. a 290, cir eet G. enioet on railways, v. I. See also: 180-183: 188-202, JI. ee J. (Whipple route): Resumé, field notes; Geological reconnaissance Napoleon to Pueblo de los Angeles, v. III. *MINTER, J. F. 377-389, v. I. *Mowry, S. 389-395; 602-613, v. I. *MULLAN, J. 301-319; 319-349; 349-352; 437-441; 516-527; 527-537; 538- 540, v. I. NEWBERRY, J. S. (Abbot route): Geology; Botany (Geographical botany; Forest trees northern California ‘ad Oregon); Zoology (Mammals; Birds). v. VI NeEwserry, J. S. See Gray, A. Parke, J. GC. (32d parallel, Dona Ana to Pimas Villages): Report, v. II. (San Francisco Bay to Los Angeles and Pimas Villages to Rio Grande near 32d parallel): Report, v. VII. PooLe, C. M. (San Diego to Fort Yuma, Parke route): Report, v. VII. Pore, J. (32d parallel route, Red River to Rio Grande): Report, v. IT. *SAxTON, H. 249-250; 251-269, v. I. SCHAEFFER, G. E. (Williamson route): Fossil wood. v. \ *STANLEY, J. M. 447-449, v. STEVENS, I. I. “— and 49th parallels): Report, v. I. Narrative and final report, v. XII. Book I. 1959 | McKELVEY, PACIFIC RAILROAD REPORTS 67 *SUCKLEY, G. (Stevens route): Mammals; Water birds; Fishes, v. XII. Book II. See also: 177-179; 291-301, v SucKLEY, G. and Giss, G. (Stevens route): Mammals, v. XII. Book II. ee Ww S. (Whipple route) : Mosses and liverworts, v. IV. (Abbot oute): Mosses and liverworts, v. sris A. W. 184-186; 226- 230: 276-281; 352-358; 399-400, v. I. TORREY, (Whipple fouls: Description general botanical collections, v. IV. pene route): Plants collected by W. Blake, v. V. (Abbot route) : Endogenous plants, v. VI. (Parke route): Sees report (Description of plants collected; Description of plates), v. VII. Torrey, J. See GRAY, Torrey, J. and Gray, A. (Beckwith and Gunnison routes): Botany, v. II. (Pope route) : ar report, v TUCKERMAN, (Abbot route): Lichens, v. VI. TURNER, W. a See WHIPPLE, A. W. Warren, G. K. Memoir of explorations since A.D. 1800, v. XI. Warren, G. K. See HuMPHREYS, A. Wurprte, A. W. Extracts from preliminary report, v. III. (35th parallel route) : me 8h Wuippte, A. W., Ewxpank, T., and Turner, W. W. (Whipple route): Indian oe Pe 18 W ILLIAMSON, R. S. (Route to connect with those near 35th parallel): Report, Vs 68 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL STUDIES IN THE GENUS COCCOLOBA, VI. THE SPECIES FROM THE LESSER ANTILLES, TRINIDAD AND TOBAGO * RicHarp A. Howarp THE PRESENT STUDY Continues a series of papers on the genus Coccoloba as it occurs in the West Indies. A single study of the genus as a whole has been considered impossible for a number of reasons. Not only is the genus a large one, but reliable data based on field observations, particu- larly in relation to hybridity and possible apomixis, have been lacking. In addition, the application of many names is difficult, for too many species, including some only recently described, have been based on anomalous specimens (adventitious shoots, for example). Many names either have been misapplied or have passed unnoticed and many species are represented only by the type collection. A further difficulty is the lack of strong and easily defined characteristics which can be used in dividing the genus taxo- nomically. Some of these obstacles have been overcome in part, as has been pointed out in earlier papers of this series. However, other difficulties, such as our inadequate knowledge of the range of variation in these plants, still remain and much more field work will be needed. The artificial geographical division which was the basis for my previous studies of the genus as it occurs in Cuba, Jamaica, Puerto Rico and the Bahamas, Hispaniola, and now the Lesser Antilles and Trinidad, has al- lowed progressive steps which will be the foundation for further work on the genus in South America, Central America and Mexico. In 1950 I completed a field study of Coccoloba in the Lesser Antilles, Trinidad and Tobago which supplied much of the material and informa- tion used in this paper. I am grateful to the directors of the American Philosophical Society and the directors of the Milton Fund of Harvard University for grants which made the field work possible. Specimens used in this study have been borrowed from a number of herbaria and botanical institutions represented by the standard abbreviations given. I wish to express my grateful appreciation to the directors and the curators of these institutions for the use of these specimens. The same difficulties encountered in earlier work apply to this study. Although field work in the Caribbean area has resolved many problems, many others remain and further studies will be necessary for a complete understanding of many of the species. The variation in leaf-form and -size when correlated with the growth habit of any particular plant should be incorporated into a general description of the species. In doing so, the *The preceding papers in this series will be found in Jour. Arnold Arb. 30: 388-424. 1649; 37: 317-339. 1956; 38: 81-106. 1957; 38: 211-242. 1957; 39: 1-48. 1958. 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 69 construction of a key to species to cover all variations seen in the field and represented in the herbarium collections becomes more difficult. In some species only one sex is known; in some, the fruiting material is un- known or is inaccurately associated with the flowering material. Fortunately, I have been able to see the type material of all of the more recently described species and all but a few of the older species. Many of Lindau’s species from Trinidad were based on specimens in the herbarium of the Imperial College of Tropical Agriculture, Port of Spain, Trinidad. Fragments of these specimens were placed by Lindau in the herbarium of Krug and Urban in the Botanisches Museum in Berlin. The fragments re- main in Berlin, but the more complete material in Trinidad has largely been destroyed by the tropical environment and by periods of neglect. The geographic area considered in this paper is currently regarded as an unnatural one. The Lesser Antilles (the Leeward and Windward Islands from Antigua to Grenada) form a group of relatively young islands, mostly volcanic in origin. The natural vegetation of this area is presumed to have been derived by migrations southward from the Greater Antilles, by mi- grations northward from Trinidad and South America and by the in situ development of some endemic or localized species. Trinidad and Tobago, in contrast, appear floristically as well as geologically as a part of the South American mainland. Of the seventeen taxa considered in this paper, two (Coccoloba uvifera and C. venosa) are widespread, appearing along the seacoasts of the Antilles, Central America and South America. Four species and one form, C. diversifolia, C. krugii, C. pubescens, C. swartzit and C. swartzii f. pubescens, are found in the Greater Antilles and extend to varying degrees into the Lesser Antillean islands. Only one taxon, C. < boxii, representative of a hybrid population, occurs only in the Lesser Antilles. Coccoloba dussii occurs only in the Lesser Antilles and Trinidad, but is very similar to a species from South America. Coccoloba ascendens and C. novogranatensis occur in the Lesser Antilles and Trinidad and in South America, with the range of C. novogranatensis extending west and then north through Central America, while the range of C. ascendens con- tinues south and east in Brazil. Five species (C. cruegeri, C. fallax, C. latifolia, C. nitida and C. striata) occur in Trinidad and South America but have not been found in the Lesser Antilles. One species, C. nigrescens, is currently known only from the Bocas Islands near Trinidad, but addi- tional material will probably reveal its relationship or identity with other southern species. Thus, in contrast to the Greater Antilles where the genus Coccoloba has evolved a great many endemic species, neither Trinidad nor the Lesser Antilles has experienced a parallel elaboration of the genus. The Lesser Antillean species of Coccoloba are predominantly Greater Antillean species which have migrated southward. The species found in Trinidad are bas- ically South American and only a few species have migrated northward into the Lesser Antilles. Following the key, the species are described and are listed in alpha- betical order. 70 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL KEY TO THE SPECIES A. Inflorescence paniculate, stems thick, striate- ia hollow; ray um- bonate between the conspicuously reticulate veins. .......... C. latifolia, AA. Inflorescence racemose, solitary or fascicled. B. sea generally fasciculate; ocreae large and flaring, conspic- TV Sai 6 of a a a cr ee a ae C. fallax. BB. Tallore cance solitary, rarely with a smaller one » at the base. s or woody plants with scrambling branches. “D. Branches scandent or entire plant with flattened stems through bilateral cambial activity; flowers and fruit on pedicels ae than ocreolae. .................... _ C. duss DD. Branches uniformly thickened; pedicel scarcely, if at all, ex- ceeding ocreolae in flower; fruit o conspicuous pedicels E. Leaves of normal shoots oa oblong-elliptic, coriaceous, shining when dry, flat, coarsely and conspicuously reticulate veined, the apex rounded to emarginate; fruit large, 1.8-2 cm. long, 1-1.2 cm. thick; fruiting ‘aca is rugose when dry, thick and brittle. ....000. C. ascendens, EE. Leaves of normal shoots not oblong-elliptic, thinner in texture, usually acute or acuminate at the apex; fruit smaller; hypanthium thin and vascular F. Leaves ovate, chartaceous, flat, the ‘apex acuminate, the ultimate venation finely reticulate: stems hollow; in- florescence slender. ....................... striat ata. FF. Leaves ovate to ohovate- oblong 0 on normal branches, the Secs usually umbonate between ae veins, the apex cute; branches solid; inflorescence stou C. nitida. CC. Trees or shrubs, icie not noticeably scandent. G. oo conspicuously longer than the ocreolae in flower and fru H. - es basically orbicular, as broad as or ag than long, one or both basal lobes overlapping the petio . Leaves er aey aid rugose and pubescent fruit globose or ovoid, 0.5—-0.6 cm. long. . pubescens. II. Leaves esa ay fleshy, not rugose, eee fruit obpyriform, 1.2—2 cm. long. . uvifera. HH. Leaves basically nae to elliptic, ‘longer than broad, nar- d or rounded at the base C. diversifolia. GG. Pedicels shorter pe or ees exceeding the ocreolae j in flower and generally in frui J. Plants odio ocreae membranaceous, ee cen and usually deciduous for all of its length (Trinidad species K. Leaves of normal shoots usually broadest above the middle. narrowed toward the base, the midrib not sharply keeled below, the blades turning black on drying C. nigrescens. Kk. Leaves of normal shoots usually eer below the mid- dle, round to cordate at the base, the midrib sharply keeled, “the blades at most wma: slightly on drying. ee C. cruegeri. JJ. Plants not noticeably deciduous: ocreae ‘not membranaceous, the base at least coriaceous and persisting. 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 71 L. Achene surrounded by the perianth lobes, the ocreolar sheath commonly elongating with the pedicels, the bracts USUGIY lab. «ou cake ne eee .yetnvs.s G, Venosa. LL. Achene surrounded by the hypanthium, the perianth lobes appressed and imbricate or coronate on the achene, the bracts not conspicuously dark in color. M. Perianth lobes appressed and imbricate in fruit, the fruit triangular in cross section; leaves ovate to sub- orbicular, 2-5 cm. long, cordate at the base. C. X boxii), the fruit round in section; leaves larger. N. Leaves cordiform-ovate, broadest below the middle, 11 X 6.5 to 27 X 17.5 cm. long and broad, the base obliquely auriculate-cordate or rounded and evenly cordate. ....... sk a Saree a Bowe NN. Leaves ovate to ovate-elliptic, smaller, generally narrowed below the middle. O. Petioles arising above the base of the ocreae, the base of the blade auriculate at the junction with the petiole, the ultimate venation not con- SOICUOUS: A Ste Ae we nc Oe Ms novogranatensis. OO. Petioles arising from the base of the ocreae, the base of the blade slightly decurrent on the petiole, the ultimate venation reticulate and slightly elevated. ..... _. C. swartzit. Coccoloba ascendens Duss ex Lindau, Engl. Bot. Jahrb. 13: 156. 1890. Coccoloba ascendens Duss mss. in herb. Krug & Urban; Ann. Inst. Col. Marseille 3: 168. 1897, in part. Coccolobis quadrifida Britton, Bull. Torrey Bot. Club 53: 467. 926. Coccoloba urbaniana Lindau, Engl. Bot. Jahrb. 13: 155. 1890 (as to Trinidad specimens cited but not description). Climbing shrub or liana, much branched, the branches to 50 feet long, glabrous, occasionally hollow; ocreae membranaceous, deciduous, cut to the base, 17-20 mm. long, acuminate at the apex, glabrous to lightly puber- ulent; petiole inserted below the ocrea, stout to slender, 1—2.5 cm. long, flattened adaxially with a narrow compressed groove, glabrous; leaf blades elliptic, obovate, oblong, rarely ovate, 11 x 6, 13 x 8 to 17 & 14.5 cm. long and broad, coriaceous, glabrous on both surfaces, the apex rounded, acute or emarginate, the base rounded to cordate, the midrib and primary veins prominent below, distinct but not prominent above, the primary veins 5-7 pairs, arcuate anastomosing near the margin, the ultimate venation reticulate and conspicuous on both surfaces; juvenile leaves of the current year’s lateral shoots elliptic-lanceolate, 14 x 7 to 20 8.5 cm. long and broad, the apex acute to acuminate, the base acute; leaves of adventitious shoots on petioles 3-5 cm. long, the blades ovate to elliptic-ovate, Sh) 20 to 45 & 30 cm. long and broad, the apex acute, the base rounded to slightly cordate: inflorescence axillary or terminal, 13-17 cm. long, the basal ocrea 72 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL 1 cm. long, densely ferrugineous puberulent, the rachis puberulent, the bracts broadly ovate, 2 mm. long and broad; ocreolae 2 mm. long, shal- lowly lobed; flowering pedicels to 2 mm. long; hypanthium 1.5 mm. long, the perianth lobes 1.5—-2 mm. long and broad; stamens 8, the filaments to 2 mm. long, united at the base; pistil rudimentary, to 1.5 mm. long; stam- inate flowers 2—3 per nodule, the pistillate flower borne singly; hypanthium and perianth similar but stamens rudimentary, functional pistil to 2 mm. long; fruiting inflorescence to 30 cm. long, the rachis to 4 mm. diameter at the base, the mature pedicels 3-4 mm. long, the fruit conspicuously articulate; fruit 1.8—2.4 cm. long and 1-1.2 cm. in diameter; fruiting hy- panthium thick, slightly woody, with 11 bundles, the achene chestnut brown, smooth and shiny when dry. Locat NAMES: liana baur (Dominica), liane cacao (St. Lucia). cuchape (Trinidad). DIsTRIBUTION: Guadeloupe to Trinidad (and possibly South America). Guadeloupe: Ravine Chaud, Questel 752 (us), Duss 2180 in part (c, Us); Baines Jaunes, Stehlé 388 (us), 1036 (NY), 1501 (us), 1989 (us). Howard 11796 (GH), 11812 (A, GH); Fonfarabre, Stehlé 3023 (Ny); without location, Duchassaing (8B). Dominica: Carib reserve on Castle Bruce trail, WH. & e 3334 (GH); Between Pointe Ronde and Milton Estate, W.H. & B.T. Hodge 2669 (Gu); Between Riverdale and Deux Branches, Howard 11758 (GH), 11771 (GH). Martinique: Fonds St. Denis, Morne Rouge, Duss 36 (x, Ny, US); Morne Juin, Hahn 1005 (B-LEcTOTYPE, GH, US). St. Lucia: Quilesse, P. Beard 1089 (GH, MO, Ss), Howard 11676 (cu); Barre d’Isle, Howard 11388 (GH); Castries-Dennery Rd., Howard 11336 (cH): Patience near Micoud, Howard 11602 (GH). Grenada: Grand Etang, Howard 10660 (GH): without location, Broadway 1206 (rm). Trinidad: Arima valley near Simla. A.C. Smith 10103 (A); Aripo Savannah at Waller Field, Howard 10316 (cu), 10341 (cH), 10452 (GH), 10507 (A); Aripo road near 34 mile post, Broadway 5992 (Kk); Blan- chisseuse road N. of Arima, Broadway 7467 (Bm), Howard 10371 (a, GH): Britton, Britton & Freeman 2139 (GH, K, NY); Caparo, Broadway 2760 (FM): Cleaver wood west of Arima, Simmonds 189 (TRIN); Cumaca road, Simmonds 322 (TRIN); Between La Brea and Irois, Crueger 2694 (TRIN), 2695 (TRIN), 2696 - TRIN); Long Stretch, 22—23-mile post, Broadway 6942 (BM, FM, GH, K, MO, ; Maravel, collector unknown (TRIN 5953); Mt. Tucuche, Baker 14329 (TRIN), yee (TRIN), Williams 11012 (Ny-type of C. quadrifida, TRIN): Southern Guyapo road, Broadway 6142 (A, BM, K, $); Toco road near Sangre Grande, Howard 10367 (cH). Coccoloba ascendens in the Antilles is a clearly distinct species charac- terized by being a liana with shiny, coriaceous and heavily veined leaves and having a large fruit with a thick, vascular hypanthium. In the orig- inal description, Lindau cites material from Brazil which I have not seen and which is not represented in the material from the Berlin Herbarium. It would be unusual if this species were not found in the Guianas or Brazil; there may even be an older name for the species. Lindau cited a number of specimers in the original description without indicating a holotype. The specimens in the Berlin Herbarium carry no 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 73 designation to indicate a type and therefore the collection Hahn 1005 from Martinique is selected as the type collection, with the lectotype being in the Berlin Herbarium. In 1897 Duss redescribed Coccoloba ascendens, in- cluding with his description the citation of specimens which Lindau had previously described as C. dussii. It is unfortunate that one large collec- tion, Duss 2180, selected by Lindau as the type of C. dussii, is a mixe collection and was included by Duss in his description. While the specimen of Duss 2180 in the Berlin Herbarium represents the taxon described by Lindau as C. dussii, a great number of the other collections of this number represents C. ascendens. In habit Coccoloba ascendens, as I have seen it in the field, is a scram- bling shrub or a liana. The plants were generally climbing to the tops of the tallest trees. If these were small, the plants of C. ascendens formed a heavy tangle on and through the low shrubs. Leaf variation on a single plant was tremendous. Different types of leaves seem to be produced by the mature wood, the young shoots and the adventitious branches. The greatest variation in leaf size and type was seen in specimens in St. Lucia. I made one collection (Howard 11336) where trees had recently been cut along the Castries-Dennery road. The plants on the edge of this clearing had some branches sprawling out into the cleared area while other branches climbed into the undisturbed woods. On one such plant some stems were hollow while others retained the solid pith. Some stems had lanceolate- elliptic leaves, acuminate to acute at the apex and the base, while other branches of the same plant had leaves broadly ovate-oblong, acute at the apex and truncate to cordate at the base. Variation in texture was evi- dent in the thin, young shade leaves in contrast to the larger coriaceous leaves in full exposure to the sun. Similar variation was found on plants in the forest reserve in the center of the island. A plant of a coastal river valley had thick basal stems with leaves to 45 cm. long and 30 cm. wide, while the scrambling branches produced oblong leaves averaging 13 9 cm. long and broad. Coccoloba quadrifida described by Britton is clearly a restricted growth form of C. ascendens. Britton reports his plant to be a small tree instead of the usually liana type, but the cited material, as well as subsequent col- lections from the same area, show clearly the variation that makes C. quad- rifida the same as C. ascendens. Many of the specimens cited above bear an unpublished specific name referring to the island of Guadeloupe. These are not distinguishable from C. ascendens in other parts of its range. Coccoloba urbaniana was described by Lindau in his monograph of the genus in 1890. He cited specimens from Trinidad (Crueger 2694, 2696) and from Puerto Rico (Sintenis 1527, 1585); however, the description published fits only the Puerto Rican material. In a subsequent treatment of the West Indian species (Symb. Ant. 1: 225. 1899), Lindau again cites the Puerto Rican specimens but fails to list the Crueger specimens any- where in the treatment. Britton later redescribed the Puerto Rican ma- terial as C. borinquensis, stating that Lindau had erroneously included this 74 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL in C. urbaniana of Trinidad. As no type was selected for C. urbaniana and the original author chose to limit the species to Puerto Rico, I feel that the species name has no validity in the Trinidad flora. Coccoloba urbaniana is best considered as a form of C. swartzi (Jour. Arnold Arb. 37: 328. 1956). A most striking variation was found in the plants of this species grow- ing in the Aripo Savannah. A number of specimens were found with the staminate and pistillate flowers on different branches of the same plant. Most of the species of Coccoloba which I have studied in the field have been completely or functionally dioecious. This was also true of the ma- jority of plants found in the savannah area. However, there were enough monoecious plants seen in this area to suggest the need for careful field study in the South American species of this genus. One collection seen but not cited above is Duss 72, represented by one specimen in the New York Botanical Garden herbarium. The handwritten label reports the plant to be from Antigua. During my visit to Antigua I failed to find a single ecological location which might possibly have sup- ported the species. Since other recent collectors have not reported the species, I suspect that the label of this specimen is in error. Coccoloba * boxii Sandwith (pro sp.); Howard, Jour. Arnold Arb. 37: 329. 1956 Coccolobis boxi Sandwith, Jour. Bot. 78: 97-98. 1940. Tree to 8 m. tall; current year’s branches cinereous, striate, sulcate, pubescent; ocreae to 1.5 cm. long, densely pubescent, the base persistent, sub-coriaceous, the apex membranaceous-evanescent; leaves cordiform- ovate, 11 & 6.5 cm. long, 27 « 17.5 cm. broad, coriaceous, light, midrib pubescent; lateral veins 8-12, slightly prominent; petiole densely pubes- cent, 0.7—1.3 cm. long; apex obtuse, the base oblique, auriculate-cordate or rounded to nearly evenly cordate; inflorescence solitary, simple, 17—22 cm. long; rachis 1.5-2 mm. wide, sulcate, densely minutely pubescent, the nodules 1—4-flowered, the bracts broadly triangular-ovate, rounded-obtuse, 1.3 mm. long, 2 mm. wide, conspicuously pubescent, the ocreolae 2 mm. long, membranaceous, flaring, almost bilobed, glabrous at the base; pedicels to 0.75 mm. long, the hypanthium 1.3 mm. long, the perianth lobes ovate- obtuse, 2 mm. long, 1.75 mm. broad, the interior smaller, the filaments (in bud) 0.6 mm. long; ovary ovoid-ellipsoidal 2.3 mm. long, 1.3 mm. in diameter, glabrous; styles shorter than the ovary, about 0.75 mm. long. Antigua: Pelican Bay, Box 539 (BM, US), 1497 (BM-HOLOTYPE, US). The suggested hybrid nature of this species and the possible parents, C. uvifera and C. swartzii or C. swartzii forma pubescens, have been dis- cussed in a previous paper. Coccoloba cruegeri Lindau, Engl. Bot. Jahrb. 13: 209. 1890. Coccoloba ernstii Johnston, Proc. Amer. Acad. 40: 685. 1905. Tree to 12 m. tall, d.b.h. 30 cm., the wood extremely hard and tough. 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 75 the bark gray; young stems glabrous, ash-colored, striate, solid; ocreae 4-5 cm. long, membranaceous and translucent when dry, deeply split, de- ciduous: leaves deciduous, the petioles 2-3 cm. long, deeply grooved above, glabrous. inserted at the base of the ocrea, the blades ovate-elliptic to sub- orbicular, 14 12 to 22 * 16 cm. long and broad, coriaceous when ma- ture, glabrous, the apex acute to rounded, the base rounded, the margin flat, the midrib prominent above, sharply keeled below, the primary veins 6-8 pairs, slightly prominent above, sharply keeled and prominent below, slightly decurrent on the midrib, straight becoming arcuate and anastomos- ing near the margin, the secondary veins and nerves reticulate; leaves of ad- ventitious shoots with ocreae to 7 cm. long, on petioles 5 cm. long, blades mostly elliptical to 54 cm. long and 32 cm. broad, with 11 pairs of con- spicuous primary veins; inflorescence racemose, laxly flowered, 7 cm. long, the rachis angled when dry, puberulent, the bracts ovate, 0.5 mm. long, the ocreolae membranaceous 0.75 mm. long, the pedicels tenuous, as long as the ocreolae, the hypanthium glabrous, | mm. long, the perianth lobes ovate to oblong, 1 mm. long, only staminate flowers known, these borne singly with functional stamens exserted on filaments 1.25 mm. long, the pistil rudimentary; immature fruit (according to Lindau) subglobose, the lobes of the fruiting perianth conspicuous to the base. DistrRIBUTION: Trinidad and Margarita. Trinidad: Without specific location, Crueger 113 [TRIN 778] (GOET-TYPE, NY, TRIN); Monos Island, Howard 10415 (cH); Aripo Savannah, Howard 10505 (cH); Toco Road between Arima and Sangre Grande, Howard 10365 (GH). Lindau assigned Goccoloba cruegeri to his section Campderia, believing that the perianth lobes surrounding the achene remained free in fruit. The associated characters of this section of black-colored bracts and normally large. conspicuous ocreolae were not described by him and cannot be seen on the type material. The specimens in Berlin and Goettingen do not have fruits, in spite of Lindau’s description- of them, and the material in the Trinidad Herbarium has only fragments of a flowering inflorescence axis. Therefore it seems unwise to assign C. cruegeri to the section Campderia. Lindau cites specimens for this species as “Crueger 113, 778 (Engl. Bot. Jahrb. 13: 210. 1890). Examination of the original material which is preserved in the Trinidad Herbarium shows that Crueger assigned his own collection numbers to some specimens and failed to do so to others. However, nearly all sheets were assigned numbers for the Trinidad Herba- rium and in the present case the 113” is Crueger’s number while the “778” is the serial number of the collection in the Trinidad Herbarium. The orig- inal collection of this material was assigned by Grisebach to Coccoloba plumieri, according to the label on the specimen, although this informa- tion was never published. The label also indicates that the collection was made in the vicinity of Caroni. There is a strong similarity between Coccoloba nigrescens and C. crue- geri, and perhaps future collections of more adequate material will indicate either that they are the same or that material called C. cruegeri is in reality 76 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL a hybrid population. Both species are unique in the Trinidad area in hav- ing deciduous leaves. Coccoloba nigrescens appears to develop flowers when the leaves are very small, while C. cruegeri has nearly mature leaves before any indication of the inflorescence is produced. I found only one plant in flowering condition and on this the flowers were just beginning to open. These were borne singly at the nodules and were staminate. The several plants seen in the field all possessed multiple trunks but were without indication of injury. One trunk possessed leaves much larger than normal as described above. In both normal and adventitious shoots the leaves were characterized by the remarkably sharply keeled midribs. In this characteristic the plant could be distinguished easily from neigh- boring plants of C. nigrescens. Several of the local residents were familiar with plants assigned to these two species. They recognized these as “grapes” similar to Coccoloba uvifera but could not recall having seen fruits on the trees. Coccoloba ernstii Johnston described from Margarita Island is to be re- ferred here. The type specimen (Johnston 250) in the Gray Herbarium indicates that this tree, also, is deciduous and that the young leaves and inflorescences are comparable to the material cited from Monos. Johnston indicated the alliance to C. cruegeri in his original description. Coccoloba diversifolia Jacq. Enum. Pl. 19. 1760, Hist. Stirp. Amer. 114. pl. 76. 1763 Coccoloba cubensis Meisner, DC. Prodr. 14: 162. 1857. Coccoloba hcl Lindau, Engl. Bot. Jahrb. 13: 158. 1891, and all recent authors, not Jacq Coccoloba longifolia Schmidt, Rep. Spec. Nov. 24: 73. 1927, not Fischer ex Lind. Shrub or tree to 8 m. tall; branches terete, often geniculate by limited growth, glabrous, the nodes rarely slightly swollen; ocreae coriaceous in the persistent lower portion, membranaceous and deciduous above, 3—5 mm. long; leaves of normal shoots with petioles 7-10 mm. long, glabrous, arising from the base of the ocreae; blades ovate, oval, oblong, ie lanceolate or obovate, variable on one branch, 4 3.5, 7 x 54:5 % 4 12 & 8 cm. long and broad, coriaceous, often shining above, dull beneath, glabrous, the apex rounded, obtuse, acute or acuminate, the base cuneate, rounded or subcordate, the margin entire; midrib and primary veins slightly prominent above, the secondary venation reticulate on both surfaces, the primary veins 3—7 pairs, arcuate, anastomosing before reaching the margin; leaves of adventitious shoots on petioles 1—2.5 cm. long, with blades of varying shapes 17 & 8, 24 & 13, to32 & 12.5 cm. long and broad; leaves of windswept specimens often much smaller than those of normal shoots with blades 2 & 1.3 to 3 x 2 cm. long and broad: inflorescence terminal 4.5-18 cm. long, the rachis glabrous, the flowers on pedicels 2-4 mm. long, the staminate flowers 2-4 at each locus, the pistillate flowers borne singly at each locus, the bracts ovate, less than 0.5 mm. long, the ocreolae membranaceous, less than 0.5 mm. long, the hypanthium 1 mm. long. the 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 77 perianth lobes ovate to oblong, 2-3 mm. long, 1-2 mm. broad, the func- tional stamens 1 mm. long, the sterile stamens rudimentary; fruiting pedicels 3—4.5 mm. long, the fruit globose to obpyriform, 10 & 7, 12 X 8, 13 < 8 mm. long and in diameter, the apex rounded, the perianth lobes imbricate and appressed. DISTRIBUTION: Florida, the Bahamas, the Greater Antilles and Antigua. Antigua: Pelican Bay, Howard 11990 (cH). Coccoloba dussii Lindau, Notizblatt Berlin 1: oe 1896, Symb. Antil. 1: 226. 1899; Duss, Ann. Inst. Col. Marseille 3: 167. 1897. Coccoloba scandens Benth. in obs. Lond. Jour. Bot. 4: 625. 1845, nomen. Coccoloba excelsa Smith (?), Kew Bull. 1893: 272. 1893. Woody shrub with scrambling branches or a vine; young stems terete, minutely short pilose becoming glabrous, the older stems flattened through bilateral development; ocreae to 1 cm. long, membranaceous, deciduous; petiole inserted below the ocrea, 2.5—4 cm. long, minutely pilose; blades ovate to oblong, 6.5 & 4.5, 12 * 7.5 to 17 x 10.5 cm. long and broad, thin-coriaceous, glabrous, the apex acute to short-acuminate, the base rounded to slightly cordate, the margin entire, the midrib prominent below, slightly keeled above, the secondary veins 7-8 pairs, arcuate, anastomos- ing near the margin, the ultimate venation reticulate and conspicuous on both surfaces; leaves of adventitious shoots on petioles 3.5 cm. long, the blades broadly ovate to elliptic-ovate, 25 « 16 to 30 & 19 cm. long and broad, the apex rounded to acute, the base broadly cordate; leaves of younger shoots elliptic-lanceolate, 19 % 10 to 24 x 10.5 cm. long and broad, the apex acuminate, the base rounded or acute; inflorescence ter- minal, 9-15 cm. long, the basal ocrea 1.3—4.3 cm. long, the rachis minutely pilose or puberulent, the bracts less than 1 mm. long, the ocreolae 0.5 mm. long, the flowering pedicels to 3 mm. long, the hypanthium 2 mm. long, the perianth lobes ovate, 2 mm. long, the staminate flowers borne 2 or 3 at a node, the pistillate flowers not known; fruiting pedicels 4-6 mm. long, usually slightly reflexed, slender, the mature fruit short stipitate, spherical, dark blue-purple in color, extremely astringent, the surface of dried fruit rugose, the fruiting perianth tightly adherent to the achene, the achene dull brown, rugose Loca NAMES: Raisinier marron, Raisin marron (Guadeloupe). DIsTRIBUTION: Guadeloupe, St. Lucia, St. Vincent, Grenada, Trinidad and Tobago. Guadeloupe: Ravine Chaude (Lamentin), Duss 2180 (B-TYPE, C, FM, NY); Between Pointe Noire and Ravine Chaude, Duss 3711 (B, FM, NY, US); Between Lamentin and Pris d’Eau, Howard 11852 (a, GH); Grande Savane, Gourbeyre, Duss 3430 (8, NY); Goyave, Pont Moreaux, Stehlé 931 (us); Basse Terre near Duclos, Petit Bourg, A.C. Smith 10347 (A). St. Lucia: Without location, An- derson s.n. (K). St. Vincent: Silver Spoon district above Three Rivers, Howard 11161 (cH), 11185 (GH), 11189 (GH). Grenada: Grand Etang, Howard 10656 78 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL (cH). Tobago: Easterfield, Broadway 4371 (Ny); The Widow, Broadway 4653 (kK, Mo). Trinidad: Between Arima and Sangre Grande, Howard 10362 (cH); Blanchisseuse Road, Broadway 6456 (A); Morne Bleu, Britton, Freeman & Bailey 2271 (GH). Similar to many other lianas, Coccoloba dussii varies according to the attitude of the stem and its branches. Over fifty feet of liana was pulled from trees or untangled from prostrate growth in several locations to de- termine the range of leaf variation in single plants. Basal leaves which persisted on the older stems were usually larger than the others on the plant. Leaves on apical portions of the stem were usually longer and nar- rower than the other leaves, while the smallest leaves occurred on the lateral flowering or fruiting branches. Such flowering branches are repre- sented by most of the herbarium specimens cited. Flowering branches which were twisted around limbs of the supporting trees as an aid in climbing had the smallest leaves and the shortest inflorescences. The bilateral development of the older stems was conspicuous. Such stems were often 5—6 times as wide as thick. Young stems were cylindrical but slightly fluted, with a conspicuously large pith. Coccoloba dussii is similar to C. excelsa and in sterile condition it is often difficult to distinguish between them. In most South American material of C. excelsa the leaves, young stems and inflorescences are conspicuously pubescent. In some South American material the pubescence has been lost, leaving relatively large hair bases with evidence of hairs only along the midrib, on the lower surface of the lamina and in the axils of the lateral veins. Such hair bases have often been mistaken for glands or secretions. Both functionally staminate and pistillate flowers of C. excelsa are known and in each the ocreolae surrounding the flowering pedicels are large, membranaceous and more or less persistent, almost equalling the pedicels in length in flowering condition. Only staminate flowers are known for Coccoloba dussii and in these the ocreolae are small and less than 1/5 the length of the pedicels. The inflorescence rachis and the ocreolae are only slightly puberulent. I have not observed C. excelsa in the field, but on the basis of herbarium material alone the character of the ocreolae in flowering condition is sufficient to distinguish the species. The fruits likewise -are similar, as is the growth habit, and the appearance of the leaves in dried condition. In general the leaves of C. dussii tend to be more umbonate between the veins and to have thinner and slightly longer petioles. The flowers are commonly reflexed on the inflorescence, although this characteristic may be associated with the growth habit of the plant. Further field study and additional collections may indicate that C. dussit should be considered a variety of, or even synonymous with, C. excelsa. Lindau described Coccoloba dussii in 1896, citing only one specimen, Duss 2180, in the herbarium of Krug and Urban. The type in the Berlin Herbarium consists of one branch and is as Lindau described. Other specimens bearing this collection number are mixtures of C. dussii and C. ascendens. In describing Coccoloba excelsa, Bentham notes that “In Forsyth’s 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 79 herbarium, I found, under the name of C. scandens, an imperfect specimen of a plant gathered by Anderson in Saint Lucia, very much like the above, but with the leaves perfectly smooth, the racemes much longer and the bracts very small. These are the only two species as yet known to be climbers.” I have seen the plant from Bentham’s herbarium which is now in the collection at the Royal Botanic Garden at Kew. It is a fragment consisting of two leaves and one inflorescence; the latter is staminate, but very few flowers remain. The specimen is better referred to C. dussi# than to C. excelsa. It is to be expected that more thorough exploration of St. Lucia will reveal C. dussii, which has not been reported in recent collections from that island. Coccoloba dussii is known from Guadeloupe, but has not been found on Martinique or Dominica to the present. Coccoloba fallax Lindau, Engl. Bot. Jahrb. 13: 172. 1890. Coccoloba crescentiifolia Griseb. Fl. Br. W. I. 163. 1859, not Cham. Tree to 10 m. tall (Swabey), d.b.-h. 6 cm.; branches thick, strongly striate; ocreae coriaceous, 2—4 cm. long, flaring, strongly bilobed, conspicu- ously striate, glabrous; petioles arising from the base of the ocreae, 2.5— 4.5 cm. long, stout, deeply grooved, striate, glabrous; leaf blaces oblong- ovate to oblong-obovate, 14 & 5, 23 & 10, 25 & 11, 36 X 13 cm. long and broad, thick-coriaceous, glabrous, the apex acute to acuminate, the base rounded or narrowed, the ultimate base slightly cordate, the midrib conspicuous above, keeled, sulcate or grooved at both sides, strongly keeled and conspicuous below, the veins 7-10 pairs, ascending arcuate, anastomos- ing near the margin, slightly impressed above, conspicuous below, the ulti- mate venation reticulate but inconspicuous; inflorescence lateral, fascicu- late, racemose, laxly flowered, the basal ocreae conspicuous, coriaceous 3 cm. long, mostly persisting, commonly strongly keeled or striate, glabrous or puberulent, the rachises 7-23 cm. long, puberulent, the flowers borne singly at each nodule, the bracts ovate, minute to 0.5 mm. long, puberulent, the ocreolae puberulent, scarcely exceeding the bracts, the staminate flow- ers with pedicels 2-3 mm. long, puberulent, the hypanthium 1 mm. long, the perianth ovate, the lobes 1.5 mm. long, the functional stamens with filaments 2—2.5 mm. long, the pistillate flowers not seen; fruiting pedicels 2-3 mm. long. Loca NAMES: Black grape (Trinidad). DISTRIBUTION: Trinidad and Venezuela. Trinidad: Without locality, Crueger 114 (GOET-TYPE), 779 (TRIN); S. W. Reserve, Swabey [TRIN 12292] (ny); Cat’s Hill Reserve, Swabey [TRIN 12543, 12536]; Black River, Freeman 8102 (NY, TRIN); Windbelt Reserve, Brooks [TRIN 12502] (Ny). Chacachacare: Britton, Freeman & Watts 2694 (GH, NY, TRIN). Venezuela: Cristobal Colon, Broadway 685 (GH); Cariaquita, Bond, Gillin & Brown 244 (cu); Fundo la Argentina, Delgado 209 Coccoloba fallax is very similar to C. densifrons and C. padiformis. When more adequate material is available, especially when there is a good 80 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL series of fruits and more information on the growth habits, these species may be considered to be merely varieties, or even identica I was unable to find Coccoloba fallax in Trinidad or on Chacachacare. Data on the specimens cited are inadequate for a complete understanding of the form of the plants. No female flowers are known and the few fruits associated with specimens collected by Swabey (TRIN 12536) and Freeman (TRIN 8102) are crushed and in poor condition. The fruits were probably round or spherical with imbricated perianth lobes. The achenes were a pale tan color in the material seen, apparently also spherical and very thin-walled. Swabey reports that one specimen which he collected grew as a clump tree of 2 to 6 stems. Some of these reached 30 feet in height. Neither these nor other specimens studied appeared to represent adventitious shoot material. At present Coccoloba fallax seems distinct on the basis of the fascicled inflorescence of simple racemes, the conspicuously large ocreae, particularly those of the inflorescence, and the strongly keeled midribs of the leaves. Coccoloba krugii Lindau, Engl. Bot. Jahrb. 13: 145. 1890; Symb. Antil. 1: 222. 1899; Howard, Jour. Arnold Arb. 37: 337. 1956 Coccoloba borgesenti Schmidt, Rep. Spec. Nov. 24: 75. 1927. Coccoloba bergesenii forma ovato-lanceolata Schmidt, Rep. Spec. Nov. 24: 76. 1927. Shrub or small tree to 6 m. tall; branches terete, anion saan ge- niculate and nodose; ocreae membranaceous, persistent, 3—5 mm. long; leaves of normal shoots with petioles 5—6 mm. long, Ba at the ean aris- ing from the base of the ocreae; blades ovate to suborbicular, 2 « 1.8, 4 X 3.5, 5 & 4 cm. long and broad, thin-coriaceous, glabrous or rarely with a few hairs near the attachment of the petiole, the apex obtuse or rounded, the base cordate or rounded, the margin flat or recurved; midrib flat above, slightly prominent below, the primary veins 4—6 pairs, straight bifurcating and anastomosing near the margin, flat on both surfaces, the secondary venation minutely reticulate below, smooth above; leaves of ere arias shoots with petioles 1 cm. long, the blades cordate or elliptic 7 X 6 cm. long and broad; inflorescence terminal 5—8 cm. long, the oak glabrous, the staminate flowers 1-3 per node, the pistillate flowers borne singly, the bracts broadly ovate, membranaceous, 1 mm. long; ocreolae membranaceous, flaring to 1 mm. long; sedicels wanting or shorter than the ocreolae, the hypanthium 1 mm. long, the perianth lobes ovate, to 2 mm. long, the filaments of fertile stamens 1.5 mm. long; fruit ovoid or angularly fusiform, strongly triangular in outline, 4-5 mm. long, 3—3.5 mm. in diameter, the perianth lobes appressed, about half the length of the fruit. DIsTRIBUTION: The Bahamas, Hispaniola, Jamaica, Puerto Rico, the Virgin and the Leeward Islands. Antigua: Goble creek near Gaynors, Box 1388 (A, GH, US). 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 81 Coccoloba latifolia Lam. Encycl. 6: 61. t. 316, f. 4. 1804; Lindau, Engl. Bot. Jahrb. 13: 133. 1890. Small tree to 15 m. tall, d.b.h. to 45 cm.; branches thick, angular and vertically ridged at least when dry, hollow but septate at the nodes, gla- brous; ocreae large, conspicuous, coriaceous, tubular or sub-campanulate, 3-7 cm. long, glabrous, acute to strongly bilobed; leaves of flowering branches variable in size, the petioles generally 2-3 cm. long, glabrous or puberulent or occasionally with long pilose hairs; blades generally or- bicular to obovate, 13 *& 10, 20 * 15 to 28 & 28 cm. long and broad, coriaceous, glabrous or with scattered pilose hairs on the midrib and veins on the lower surface, the apex rounded or emarginate to short acuminate, the base rounded to deeply cordate, the margin flat or strongly recurved, entire, often undulate; midrib and veins impressed above, conspicuous be- low, the primary veins 9-12 pairs, arcuate and anastomosing near the margin. the secondary venation conspicuous, the blade strongly bullate be- tween the primary veins or between the secondary veins as well as the pri- mary or scarcely bullate and nearly flat; leaves of adventitious shoots larger with petioles to 5 cm. long and blades to 48 & 50 cm. long and broad or larger. similar in other characteristics; inflorescence terminal, panicled to 30 cm. long and broad, with solitary or 2—3-fascicled undivided pseudo- racemose branches or rarely the lowest branch paniculate, the rachis puberulent, the staminate flowers 1-3 per node, the pistillate flowers usually borne singly but some apparently 3—4 per node, the bracts minute, to 0.5 mm. long, the ocreolae membranaceous to 0.75 mm. long, the pedicels very thin 1-2 mm. long, the hypanthium tapered at the base, 1-1.6 mm. long, the perianth lobes 1.5 mm. long, the functional stamens with filaments to 2 mm. long; fruit ovoid, 7-10 mm. long, 5-6 mm. in diameter, the base rounded to slightly attenuate, the apex rounded with imbricate perianth lobes, the immature fruits showing 9 conspicuous basal nerves when dry, the mature fruit black, fleshy when fresh, the outer pericarp commonly loose. the achene dark brown shining, smooth. LocaL NAMES: Grape, cuchape, stave wood, big leaf. DisTRIBUTION: Trinidad, Tobago, Venezuela, British Guiana, Brazil. Trinidad: San Fernando, La Retraite, Crueger 2690 (B, TRIN); Tamana, Marshall | TRIN 12313] (Ny); St. Joseph, Dannouse, s.n. (TRIN); Aripo Savanna, Britton, Broadway & Hazen 316 (GH, Ny), Howard 10355 (cH); Talparo, Simmonds |TRIN 14652]; Arima, Eggers 1377 (c); Piarco Savanna, Britton, Britton & Hazen 113 (ny); Chancellor Road, St. Ann’s, Broadway 8989 (TRIN); Caron. Eggers 912 (mM); cultivated, Broadway 4332 (TRIN). Monos Island: Howard 10416 (GH). A well-defined and easily recognized species. The trees are common in savanna areas and probably for these reasons have not been collected and so the species is poorly represented in herbaria. The wood tends to be soft and the branches are relatively brittle and fragile. An equal hazard to the collector is the occurrence of colonies of stinging and biting ants in the hollow pith and in the ocreal sheath. 82 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XL Coccoloba nigrescens Lindau, Engl. Bot. Jahrb. 13: 192. 1890. Tree to 10 m. tall; youngest stems pubescent with long strigose hairs, becoming glabrate; ocreae lax, membranaceous to extremely thin and trans- lucent even when fresh, to 3.5 cm. long, divided 2/3 its length, acute at the apex, pubescent with long scattered hairs, deciduous to the base; petioles attached at the base of the ocreae, stout, 5-7 mm. long, densely pubescent on the adaxial surface when young, becoming glabrate; leaf blades obovate-oblong to ovate-oblong, 12 x 7 to 16 X 9 cm. long and broad, thin-coriaceous when mature, glabrous, the apex acute to rounded, rarely short-acuminate, the base cuneate or rounded, the margin slightly recurved, the midrib prominent on both surfaces, not keeled, the primary veins 6-9, mostly straight becoming arcuate and anastomosing near the margin, the lesser nerves inconspicuous; inflorescence terminal or lateral, to 10 cm. long, the peduncle 6 mm. long, strigose pubescent becoming gla- brous, the bracts triangular, acute, puberulent, 0.75 mm. long, the ocreolae erect, truncate, not exceeding the bracts; staminate flowers 2 (rarely 1) per nodule, the pedicels shorter than the ocreolae, the hypanthium 0.75 mm. long, the perianth lobes ovate to 0.75 mm. long, the stamens exserted on filaments 1 mm. long; pistil rudimentary, the pistillate flowers not known; old fruiting inflorescence with stout pedicels to 1 mm. long; fruit not DistTRIBUTION: Islands near Trinidad. Chacachacare Island: Crueger 116 (B, GOET-TYPE, TRIN), Howard 10440 (GH). Monos Island: Crueger [TRIN 3244], Britton & Hazen 1742 (FM, GH, NY, TRIN), Broadway 7457 (TRIN), Howard 10422 (GH). Coccoloba nigrescens is apparently one of the few species of the genus which loses all of its leaves at certain seasons. However, there is no indica- tion of this characteristic in the shape or texture of the mature leaf. Lindau described the species from specimens collected in May while the leaves and inflorescences of the plant were very young and immature. Subse- quently, Britton and Hazen collected the species again on an adjacent island in the month of April and obtained specimens comparable in quality. Lindau commented on the fragility of the leaves which turn black in drying. The. Broadway collection cited above was collected in November and con- sists of mature leaves from a sterile shrub. The collections which I made were obtained in the month of February in the type locality and in the same area where Britton visited. These also possess fully mature leaves. The leaves turn dark on drying and all parts of the plant are glabrous, lacking the hairs apparently characteristic of the younger parts. My com- panion in the field, a local resident and employee at the leper colony, as- sured me that this plant does indeed lose its leaves. At present this species is poorly represented and not well known. Further collections are needed, especially to determine whether this species is truly distinct from C. crue- geri. At present C. cruegeri may be distinguished in the juvenile leaf con- dition by the lack of the strigose pubescence and in mature leaf form by 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 83 the longer petioles, larger adventitious shoot leaves and the sharply keeled midrib. Both of the Crueger specimens which represent the type collection are in very poor condition. Coccoloba nitida HBK. Nov. Gen. 2: 176. 1818. Polygonum arborescens Vell. Flor. Flum. 163. 1825, 4: pl. 43. 1827. Coccoloba marginata Hooker, Jour. Bot. 4: 626. 1845. Coccoloba guianensis Meisner, Linnaea 21: 264. 1848. Coccoloba martii Meisner, Fl. Bras. 5: 37. 1855. : Coccoloba martii Meisner var. major Meisner, Fl. Bras. 5: 38. 1855. Coccoloba nitida HBK. var. rotundata Meisner, Fl. Bras. 5: 88. 1855. Coccoloba nitida HBK. var. ee Meisner, Fl. Bras. 5: 38. 1855. Coccoloba trinitatis Lindau, Engl. Bot. Jahrb. 13: 182. 1890. Plants of variable habits most commonly tree-like at the base with scrambling, liana-like branches, the nodes extremely long and the leaves much reduced at the ends of these or liana-like from the base, rarely shrub- like; branchlets terete, smooth or striate, minutely puberulent, the lenti- cels elliptical, conspicuous, whitish; ocreae obliquely truncate, 1—1.5 cm. long on normal as well as scrambling shoots, ferrugineous puberulent when young becoming glabrate, the base coriaceous and persistent usually tightly investing the stem; petioles inserted at the base of the ocreae, usually 1.5—2 cm. long, minutely puberulent, canaliculate above, striate; leaf blades of flowering branches ovate, elliptic, oval-oblong or obovate-oblong, 10 & 5, 13 & 6 to 20 & 8 cm. long and broad, coriaceous, shining or dull, usually drying darker above than below, glabrous on both surfaces, the apex acute, rarely acuminate or almost rounded, the base rounded to cordate, the margin flat rarely slightly recurved, the midrib and veins slightly impressed above, conspicuous below, the primary veins 8-10 pairs, arcuate, anastomosing near the margin, the ultimate venation finely reticu- late; leaves of conspicuously scrambling shoots either much smaller or even larger (to 24 * 10 cm.) than those of normal flowering branches, generally of shinier texture; leaves of adventitious shoots on petioles 1.2-3.5 cm. long, with blades generally oblong or elliptic-oblong, 17 « 10 to 21 & 12 cm. long and broad, the apex usually rounded, the blade often strikingly umbonate between the primary veins; inflorescence generally terminal but often lateral, a solitary raceme, 7, 10, 22 to 36 cm. long, occa- sionally fascicled with one or two small racemes at the base or with 1 or 2 basal racemes developing from the main axis; rachis puberulent; bracts ovate 0.5—-1 mm. long, the ocreolae strongly bilobed, membranaceous, puberulent, 1.5-2 mm. long, the staminate flowers 2 or 3 at each node, on pedicels shorter than the ocreolae, the hypanthium conical 1—-1.5 mm. long, the perianth lobes ovate, 1-1.5 mm. long, the stamens 1.5—2 mm. long, the pistil abortive; pistillate flowers bea singly on pedicels about equalling the ocreolae, the stamens rudimentary and shorter than the perianth lobes, the pistil exserted, the fruiting pedicels 2—2.5 mm. long; fruit ovoid recall 12 mm. long, 9 mm. diameter), the perianth lobes 84 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL slightly coronate, the hypanthium in fruit smooth to strongly striate when dry, the achene dark brown, smooth. Locat NAMES: Black grape (Trinidad), masari (British Guiana), pipoca (Brazil) DistriBUTION: Trinidad, the Guianas, Brazil. Trinidad: Arima, Howard 10359 (GH), 10368 (GH), 10506 (GH); Aripo Savannah, Crueger 2693 (B-TYPE of C. trinitatis, TRIN), age 10507 (GH), Britton, Britton & Hazen 272 (GH, NY), Baker sm. (HT 15122), Broadway (TRIN 10480); Caroni, Crueger 671 (TRIN); Long Stretch, eons 329 (TRIN 14255), sm. (TRIN 15378); Santa Cruz, Dannouse sn. (NY, TRIN); Valencia, Broadway 5584 (A, BM, MO, S), 6652 (A, B), 8988 (TRIN), Britton, Britton & Hazen 1023 (GH, NY); Britton & Britton 2103 (GH, NY); without locality, Fendler 1010 (8M). British Guiana: Amakura River, De La Cruz 3552 (GH); Assakatta, De La Cruz 4270 (GH); Barima River, De La Cruz 3363 (GH); Bartica River, De La Cruz 1962 (GH); Kaieteur Plateau, Maguire & Fanshawe 23316 (A); Kuyuwini River, Smith 3030 (A); Mazaruni Station, For. Dept. Brit. Guiana 4269 (NY); Upper Rupununi River, De La Cruz 1442 (GH). Dutch Guiana: without location, Samuels 419 (GH), Wullschlagel By (B). French Guiana: Acaronany, Sagot 887 (Ny); Paramaribo, tig 1 492 (ny); without location, Broadway 307 (GH), Leblond s.n. (B), Jelski (8), saan 84 (NY), Poiteau (8). Brazil: Maranhao, Island of Sao Luiz, Froes 11698 (A), 11716 (A); Minas Geraés, river San Francisco near Salgado, Martius (M-TYPE of C. marti); Bahia, Joazeiro, Martius (M-type of C. nitida). In the field there is considerable variation among plants of this species. I was fortunate to have one full week in the Arima area of Trinidad, where attention was given to the growth habits and the variation seen in plants then referred to Coccoloba trinitatis Lindau. The three collections made at that time (Howard 10359, 10368 and 10506), comprise the equivalent of 47 herbarium sheets. However, each number was made from only one plant and each number was collected on a separate wooded island in the Aripo Savannah. By carefully following the branches of each tree I was able to determine that a tenuous “‘liana” of one part of the forest canopy was actually the same individual as the “different” plants of the forest floor which were, in reality, either flowering branches or adventitious shoots. As a result of this experience in the field, it seems clear to me that the specimens Lindau saw and referred to Coccoloba nitida, C. guianensis, C. marginata and C. trinitatis could all have come from one plant. They obviously did not come from one plant, but nevertheless I conclude that they represent only one species. Many of the morphological differences by which Lindau distinguished species can be found on a single plant in the field. Moreover, he failed to understand the sexual differences between staminate and pistillate plants. Meisner, Lindau’s predecessor, treated as varieties the different leaf sizes on a series of specimens. These varieties “major” and ‘‘minor” are easily found on one branch in the living plant. Coccoloba nitida will be recognized in future collections as a liana, for the flowering branches are usually short lateral shoots on a scrambling branch system. The flowers are a clearer white than most of the Coccoloba 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 85 species I have seen in the field and the staminate inflorescence in full flower is conspicuous. The long inflorescence, the membranaceous, bilobed ocreolae and the ovate fruits aid in the determination of this species. Coccoloba novogranatensis Lindau, Engl. Bot. Jahrb. 13: 192. 1890. Coccoloba caribaea Urban, Symb. Antil. 5: 337. 1907. Shrub 2 m. to tree 27 m., trunk commonly muscular, the bark red; branchlets terete, glabrous, the nodes slightly swollen; ocreae cylindrical, the upper portion membranaceous and deciduous, the lower portion coria- ceous and persistent; leaves of normal shoots with petioles 8-11 mm. long, puberulent when young, with age glabrous except in the groove, attached above the base of the ocreae; blades ovate to ovate-elliptic, 6 « 4,9 * 5, 10.5 & 5.5 cm. long and bread thin-coriaceous, glabrous above, glabrate below except for long hairs along the midrib at the base of the blade, the apex attenuate, the base narrowly cordate, the margin entire; midrib and primary veins inconspicuous above, prominent below, the primary veins 5 or 6 pairs, the ultimate venation finely reticulate; leaves of adventitious shoots generally similar but occasionally obovate and narrowed at the base, ovate-lanceolate or lanceolate-elliptic, the blades 11 & 7 to 18 & 15 cm. long and broad, the basal lobes of leaves on adventitious shoots often over- lapping; inflorescence terminal on short lateral shoots, 3-9 cm. long, the basal ocreae 5-10 mm. long, bilobed at the apex, the rachis glabrous, the flowers commonly borne singly at each nodule (occasionally 2 in staminate plants), the bracts oblong; 2 mm. long, 0.5—0.75 mm. broad, the ocreolae 3 mm. long, bilobed at the apex, membranaceous and persistent, the pedicels 2.5—4 mm. long, the hypanthium 0.75 mm. long, the perianth lobes ovate, 1.5 mm. long and broad, the functional stamens 2.5 mm. long, the func- tional pistil 2 mm. long; fruit generally spherical with coronate perianth lobes, occasionally narrowed at the base, 1 crn. long, 5—9 mm. in diameter, the achene brown, shining. DiIsTRIBUTION: St. Vincent, Grenada, Trinidad, Tobago, Margarita, Colombia, Venezuela and Panama. St. Vincent: H.H. & GW. Smith 1790 (8, Ny). Grenadines: Bequia, How- ard 11231 (GH), 11241 (GH); Cannuoan, Howard 11091 (cH); Carriacou, How- ard 10815 (GH), 10818 (GH), 10890 (cH); Large Island, Howard 10965 (cx); Mayero, Broadway 5748 (TRIN), Howard 11038 (GH); Mustique, Howard 11223 (GH); Petit Martinique, Howard 10935 (cH); Petit St. Vincent, Howard 10896 (GH); Tobago Cays, Howard 11024 (GH); Union, vias 10988 (GH). Grenada: St. Georges, Broadway 1760 (B, FM, GH, NY), 1795 (FM. GH, NY), Howard 10692 (GH); Grand Anse, Broadway 2503 (FM, se 3361 (B, M); Sauteurs, Howard 10705 (GH). Trinidad: Caledonia Island, Hart 5834 (TRIN). Tobago: The Lodge, Broadway 4019 (B, FM), Howard 10453 (GH), 10454 (cH), 10458 (GH). Carrera Island: Broadway s.n. (NY). Margarita: Johnston 274 (GH). In the original description of Coccoloba novogranatensis, Lindau cited a Karsten collection from Sabanilla, Colombia, and a Triana specimen 86 JOURNAL OF THE ARNOLD ARBORETUM [| VOL, XL without specific locality. In the Berlin Herbarium there are two sheets representing this species, but in contrast to his usual practice, Lindau did not indicate by abbreviation, or otherwise, his choice of a type. The Triana specimen is a fragment of a leaf in a packet with a pencil sketch of a fruit on the packet. The Berlin sheet attributed to Karsten consists of two flowering fragments of different ages. It would appear that the Karsten specimen in Berlin could be considered the holotype; however, when botanists from the Field Museum photographed this specimen, tem- porary annotations were placed on the “Karsten” sheet to indicate that one of the two flowering fragments was a Karsten collection and the other a Triana collection. The labels have since been removed. Thus I have been unable to verify whether this sheet is truly a mixed collection. Both flowering fragments apparently represent the same species, but I am unable to state positively which is the type. It is likely that additional Karsten material from Colombia exists in other herbaria and that a lectotype should be selected at a future date. I have seen in the herbarium of the Muséum National d’Histoire Na- turelle in Paris two full sheets of Triana 978, a number cited by Lindau. The collection was made between Anapoima and Apulo, Province of Bogota in Colombia. Each of the two sheets has a branch with young flowers and a second branch in fruiting condition. No fruits remain with the specimens. I have not seen additional material collected by Karsten. Lindau indi- cated that his specimens of this species came from Berlin, Leningrad and Vienna. The first two collections are intact and I have seen all the material. The Vienna collection of this genus was destroyed during World War IT. While it is possible that Karsten material exists in some other herbarium, it seems practical at this time to designate Triana 978 as the type collec- tion and one of the two sheets in the Paris herbarium is selected as the lectotype. In the herbarium material from the Muséum d’Histoire Naturelle in Paris there are three additional sheets to be referred to Coccoloba novo- granatensis. All were collected by Pleé. Two sheets, one with a printed and one with a script label, were collected at Maracaibo (Venezuela). The third carries a printed label “Martinique.” I cannot distinguish the speci- mens as distinct and different collections and since C. novogranatensis has not as yet been collected in Martinique, I suspect that this sheet too should bear a Maracaibo label. Urban did not indicate a holotype in his publication of Coccoloba cari- baea, but the specimen Broadway 1760 in the Berlin Herbarium is indi- cated as the type in Urban’s handwriting. This collection is in flower and is comparable in all characteristics to the fragments on which Lindau based C. novogranatensis, There have been no recent collections assigned to C. novogranatensis; in fact, the name has been overlooked in recent years. A number of recently described species from Colombia and Peru and from Central America will be assigned in synonymy to C. novogranatensis in a later paper. A large number of Coccoloba novogranatensis populations were studied 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 87 in the Lesser Antilles and a considerable morphological variation was rec- ognized as being of an ecological nature. The size, color, texture and atti- tude of the leaves on the shoots seemed to depend upon whether the popula- tions are growing in full sunlight, in shade, in an area subject to salt spray or in moist situations. A similar set of variations has also been recognized in C. venosa and collections of these two species from the Lesser Antilles have been confused by various authors. For example, the plants on Mt. Royal in Cannuoan in the Grenadines which John Beard refers to C. venosa are more properly referred to this species. Coccoloba venosa and C. novo- granatensis may be readily distinguished by referring to the point of at- tachment of the petiole to the ocrea. Coccoloba pubescens L. Syst. ed. 10. 1007. 1759; Hooker, Bot. Mag. t. 3166. 1832; Fawcett & Rendle, Jour..Bot. 51: 123. 1913; Howard, Jour. Arnold Arb. 38: 227. 1957. Scortea arbor americana, amplissimis foliis, aversaparte nervis extantibus hirsutie ferruginea refertis; Pluk. Phytographia ¢. 222, f. 8. 1691 Coccoloba grandifolia Jacq. Enum. 19. 1760. Coccoloba rubescens L. Sp. Pl. ed. 2. 523. 1762. Coccolobis pubescens Sandwith, Jour. Bot. 78: 98. 1940. Coccolobis antiguensis Sandwith, Jour. Bot. 78: 98. 1940. Mature tree to 13 m. tall, d.b.h. 5 cm., much branched above a well-de- fined trunk: branches terete, swollen at the nodes, the lenticels not con- spicuous, tomentose to pilose; ocreae to 1 cm. long, generally completely deciduous. pubescent; leaves of completely mature plants vary consid- erably in size and shape, the petioles 3-6 mm. long, inserted below the ocreae, densely de -pubescent, the blades broadly orbicular to orbicular- ovate, 4 & 6, 7.5 & 10 cm. long and broad grading into size of leaves of adventitious ne rugose or bullate, the apex rounded, the base cordate, the basal lobes rounded and only rarely approximate, sparsely pubescent above to glabrate, densely to sparsely pubescent below or glabrate, the margin undulate, the venation of 5 pairs of primary veins, arcuate to the margin, strongly anastomosing, slightly impressed above, conspicuous and reticulate below; adventitious shoots generally strict and sparsely branched, to 10 m. tall; branches stout, terete, slightly swollen at nodes, strongly grooved or striate; ocreae 2 cm. long, membranaceous and evanescent above, coriaceous and persistent below, the petioles stout 1—2 cm. long, densely tawny pubescent; blades large, generally orbicular except for the terminal leaf, frequently broader than long, 30 * 40, 50 x 80 cm. long and broad, coriaceous, rugose or bullate when mature, thin and plane when young, the apex rounded, the base rounded to cordate, the basal lobes commonly encircling the stem, the terminal leaflet commonly rhombic, longer than broad when mature, densely tomentose, the veins slightly im- pressed above, all venation conspicuous and reticulate below; midrib and secondary veins persistent pubescent above, the others sparsely pubescent when young, becoming glabrate above, the veins and leaf surface pubes- 88 JOURNAL OF THE ARNOLD ARBORETUM [VoOL. XL cent or becoming glabrate below, the margin irregular, commonly undu- late; inflorescence terminal, often stout, the basal ocreae to 7 mm. long, membranaceous, the peduncle to 1.5 cm. long, the rachis minutely and often densely puberulent, 10-18 cm. long on mature shoots, to 45 cm. long on adventitious shoots, the bracts broadly ovate, about 1 mm. long, puberulent, the ocreolae membranaceous spreading, 1 mm. long, minutely puberulent or glabrate; staminate flowers 2—4 at each locus, the pistillate flowers 1-3 at each locus, the pedicels 2-3 mm. long, the hypanthium 0.6—-1 mm. long, the perianth lobes broadly orbicular, 1.5 mm. long. 2 mm. wide, puberulent, the fertile stamens 2.5 mm. long, the sterile stamens rudimentary 0.5—1 mm. long, the fertile pistil glabrous or rarely slightly puberulent, on the ovary, the sterile pistils glabrous, rudimentary, 0.5— 1.5 mm. long; fruit globose to ovoid, 5—6 mm. long and 4.5 mm. in diameter, the fruiting perianth imbricate at the apex, not coronate, the fruiting hypanthium with conspicuous vascular bundles; achene subglobose, dark brown, shining, ated triradiate at the apex, the fruiting pedicels puberu- lent, 3-4 mm. DistTRIBUTION: Dominican Republic, Puerto Rico, Virgin Islands and the Lesser Antilles south to St. Lucia and Barbados. Barbuda: Ponthieu sn. (rm). Antigua: Carrs Gut, Box 301 (us), 1496 (A, US, ISOTYPES of C. antiguensis); Macarthy Hills, Box 1495 (pm, us); Wallings Area, J. Beard 283 (a, Mo), Howard 11984 (GH); Blubber Valley. pie 11863 (GH), without location, W ullschlaegel 486 (M). Montserrat: Cudjo Head, Shafer 457 (FM, Ny, US); Harris’ Lookout, Howard 11873 (GH). Siete Ward’s Mountain, Howard 11933 (GH). Guadeloupe: Pigeon, Duss 3379 (us); Désirade, Stehlé 195 (GH, NY); Deshaies, Stehlé 2018 (us); Jarry, Questel 602 (us); Vieux Fort, Howard 11847 (GH), 11849 (GH); without location, Duss 1743 (NY), 2193 (FM, GH, MO, NY). Dominica: Calibishie, W.H. & B.T. Hodge 3142 (GH); Pointe Ronde to Milton Estate, Aug H. & B.T. Hodge 2668 ate Pointe Baptiste near Calibishie, W.H. & B.T. Hodge 3499 (GH). Marti Presqu’ile de la Caravelle, Egler 39-252 (ny), Duss 1743 (Ny), Howard 117 128 (cH), 11729 (ny). St. Lucia: Between Le Toc and Cul de Sac Bay. Howard 11371 (cH); Vieux Fort, Howard 11475 (GH). Barbados: Turner’s Hall Woods, Eggers 7158 (A) The growth forms of Coccoloba pubescens have been discussed in earlier papers (Jour. Arnold Arb. 38: 229-231. 1957, 39: 37-39, 44-46. 1958). Immature forms with wand-like branches and very large leaves are the most common habit found in the Lesser Antilles. Beard illustrates such growth forms in his publication “Vegetation of the Windward and Lee- ward Islands” (Oxford Forestry Memoir 21: f. 32. 1949). A few isolated mature plants with the smaller leaf type were seen and collected on St. Lucia (Howard 11475), Guadeloupe (Howard 11849) and Martinique (Howard 11729). The variation in the texture, size and pubescence of the leaves within the populations studied in the field will include that described by Sandwith for his species C. antiguensis. I have previously placed that species in the synonymy of C. pubescens (Jour. Arnold Arb. 38: 231. 1957). 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 89 Coccoloba striata Benth. Hooker Lond. Jour. Bot. 4: 626. 1845. Coccoloba guianensis Griseb. Fl. Brit. W. I. 163. 1859, not Meisner. Coccoloba grisebachiana Lindau, Engl. Bot. Jahrb. 13: 195. 1890. Coccoloba pittieri R. Knuth ex Pittier, Man. Pl. Usuales Venez. 355. 1926. Tree with scrambling branches; branches hollow, striate, glabrous; ocreae membranaceous above, coriaceous and persistent below, strongly ribbed: petioles attached above the base, strongly grooved above, 1.5—2 cm. long, glabrous; blades ovate to ovate-elliptic, 9 & 5.5 to 14 & 8 cm. long and broad, chartaceous, usually turning dark on drying, the apex shortly and obtusely acuminate, the base rounded to cordate or slightly cordate-auricu- late, the midrib and primary veins almost immersed above and only slightly prominent below, the primary nerves 6-8 pairs, arcuate and anastomosing near the margin, the ultimate venation densely reticulate; leaves of scram- bling shoots broadly ovate to 23 cm. long and 18 cm. broad, on petioles 3 cm. long; inflorescence to 27 cm. long, the rachis thin and glabrous or slightly puberulent, the flowering nodules widely separated in the pistillate plants and less conspicuously so in the staminate plants, the peduncle about | cm. long, the bracts ovate to triangular generally 1 mm. long, the ocreolae flaring, scarcely as long as the bracts, the pedicels equalling the ocreolae in the staminate flower; staminate flowers generally 2-4 per nodule, the hypanthium 0.75 mm. long, the perianth lobes ovate to oblong, 0.5 mm. long, the functional stamens with filaments 1 mm. long; pistillate flowers not seen; fruiting pedicels 2-2.5 mm. long, the ripe achenes red, ovoid, about 9 mm. long and 6 mm. in diameter, the apex acuminate, the lobes inconspicuous, the fruit slightly stalked at the base, obscurely 3-angled. British Guiana: Without specific location, Richard Schomburgk 1265 (ps). Trinidad: Mount Tocuche, Crueger (TRIN 776-IsoTYPE of C. grisebachiana), Baker (tR1N 14804), Freeman (TRIN 9045), Britton, Hazen and Mendelson 1262 (GH, NY). Venezuela: Carababo, Guaremales, Pittier 8880 (A. GH. ISOTYPES of C. pittiert); Carruao, Pittier 11847 (a). This species is distinctive among the other species of Coccoloba reported from Trinidad on the basis of the hollow scrambling branches, the at- tachment of the petioles above the distinctive base of the ocrea and by the widely separated flowering nodules on the inflorescence rachis of the pistillate plants. The extremely fine reticulate pattern of the ultimate venation of the leaves is also ea ara although this expression is not always seen in the herbarium collectio One sheet of Pittier 8880 in the Gray Herbarium collections cited above probably represents an adventitious shoot. A short piece of the stem 9 cm. long contains one node. The stem is hollow and the ocrea extends 5.5 cm. above the apex of the petiole scar and 1.5 cm. from this point down the stem. The ocrea is coriaceous, appressed and deeply bifid. The leaf on the same herbarium sheet is not attached but is oblong in shape and 40 cm. long and 17 cm. broad. The margin is obviously undulate when fresh. Unless this is a mixed collection, which does not seem likely, the dissim- 90 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL ilarity of the leaves of fertile and adventitious shoots is as large in this species as in other species of Coccoloba. The material cited above from Trinidad is all from pistillate plants in fruiting condition or with old inflorescence axes. The types of Coccoloba striata and C, pittieri are staminate plants. There is no question in my mind that these are all conspecific, although I have not seen this species in the field. Coccoloba paraensis Meisner, based on Spruce and Martius’ material from Para and Amazonas, and C. glaziovi Lindau, based on a Glaziou specimen collected from Rio de Janeiro, probably represent the same species and should be assigned to synonymy here. Regrettably, I have insufficient material either to include or exclude these species with certainty at the present time. Coccoloba swartzii Meisner, DC. Prodr. 14: 159. 1856; Lindau, Engl. Bot. Jahrb. 13: 157. 1890; Howard, Jour. Arnold Arb. 30: 420. 1949, 37: 317-339. 1956. Coccoloba swartzii var. (2) portoricensis Meisner, DC. Prodr. 14: 160. 1856. Coccoloba barbadensis Lindau, Engl. Bot. Jahrb. 13: 148. 1890; Duss, Ann. Inst. Col. Marseille 3: 166. 1897, not Jacquin. Coccoloba diversifolia Lindau, Symb. Antil. 1: 223. 1899; Griseb. FI. Brit. W.I. 163. 1859, and most recent authors, not Jacquin. Trees 8 to 20 m. tall: branches terete, the youngest puberulent, becoming glabrate, the nodes slightly tumid; ocreae 10-12 mm. long, the basal por- tion 3-5 mm. long, coriaceous, persistent, the upper portion membranaceous and deciduous, puberulent to glabrate; leaves of normal shoots with peti- oles 10-18 mm. long, puberulent to glabrate, the blades ovate to elliptic, 2.2 1.3.7 5,11 & 9, 15 & 7.5 cm. long and broad, coriaceous, usually turning black on drying, glabrous, having pit-like depressions on the upper surface and small glands on the lower surface, the apex acute, often rounded, the base narrowed, rounded or slightly cordate and usually oblique, the margin entire; midrib and veins inconspicuous or flat above, prominent below, the primary veins 6 or 7 pairs, arcuate, anastomosing, the secondary venation conspicuous, reticulate; leaves of adventitious shoots with petioles 1.5—2.5 cm. long, the blades generally ovate to lanceo- late 23 & 8.5, 45 & 18.5, to 70 & 25 cm. long and broad, the apex acute to acuminate. the base rounded: inflorescence terminal 10-15 cm. long, the rachis glabrous or with glandular exudate, rarely papillose; staminate flowers in clusters of 3-5 flowers at each node, tightly surrounded by membranaceous ocreolae which form a truncate cylinder after the flowers have fallen, the pistillate flowers borne singly at each node, the bracts ovate 5 mm. long, the ocreolae 1.1.5 mm. long, membranaceous, the pedicels, shorter than the ocreolae; hypanthium 0.5 mm. long, the perianth lobes 1-1.5 mm. long, the fertile stamens 1—-1.5 mm. long, the sterile stamens rudimentary, 0.5 mm. long; fruit ovoid 8-10 mm. long, 6 mm. diameter, the perianth lobes 1-1.5 mm. long and coronate in fruit; achene dark brown. 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 91 DIsTRIBUTION: Jamaica, the Bahamas, Dominican Republic, Puerto Rico, St. Croix, St. Jan, Virgin Gorda, St. Thomas: Saba, St. Kitts, Mont- serrat, Nevis, Antigua, Guadeloupe, Dominica, Martinique, St. Lucia and Barbados: Antigua: Pearl es Box 975 (us). St. Kitts: Mt. Misery, R.A. & E.S. Howard 11938 (GH). Nev : R.A. Howard 11921 (GH). Montserrat: Cental Hills, R.A. & E.S. Howard ees (GH), 11866 (GH), 11871 (GH), 11872 (cH); aes Upper Soufriere, R.A. & E.S. Howard 11882 (GH); Gardu Gut, Shafer 323 (F, NY, US), 615 (F, Ny, US). Guadeloupe: Bois de Gombeyre, Duss 3251 (F, GH, NY, US); Bouillante to Pointe Noire, R.A. & E.S. Howard 11843 (cH). Dominica: Antilles near Marigot, R.A. & E.S. Howard 11754 (GH); Salybia, W.H. Hodge 3404 (GH); South Chiltren Estate, W.H. & B.T. Hodge 1583 (cu). Martinique: Casa Pilote, Hahn 1187 (F, GH, us), R.A. & E.S. Howard 11712 (GH); Diamant, Duss 37, 248 (ny); Mt. Pelee, Duss s.n. (Ny); Montagnes des Trois Flotz, Hahn 629 (¥, GH); Presqu’ile de la Caravalle, R.A. & E.S. Howard 11727 (cH), Egler 39-230 (Nv); Trois Islets, R:A. & E.S. Howard 11731 (cu). St. Lucia: Castries-Dennery Road, R.A. & E.S. Howard 11335 (GH), 11329 (GH), 11355 (GH); Gros Piton, R.A. & E.S. Howard 11506 (cu); Le Toc to Cul de Sac Bay, R.A. & E.S. Howard 11377 (cH); Vieux Fort, R.A. & E.S. Howard 11404 (GH). Barbados: Turners Hall Wood, Eggers 7161 (us) Coccoloba swartzii forma pubescens Howard, Jour. Arnold Arb. 30: 420. 1949 Similar to the species, but the young shoots, petioles, the lower end of the upper leaf surface, especially the midrib, the ocrea and the inflorescence rachis, at least at the base, puberulent to pilose pubescent. Antigua: Blubber Valley, Box 1411 (us-HoLoTYPE), R.A. & E.S. Howard 11860 (GH), 11985 (GH), 11986 (GH); Carr’s Ghaut, R.A. & E.S. Howard 11994 (GH); Orange Valley, Box 1184 (GH, Us); Pelican Bay area, R.A. & E.S. Howard 11990 (GH); Sugar Loaf Mt., Box 1543, 1544 (us), A.C. Smith 10489 (A). Barbuda: Codrington Village, Fairchild 3830 (A, us); Martello Tower, J.D. Beard 372 (A, MO); without location, Box 602 (us). Guadeloupe: De Pouce 86 (FM) Coccoloba uvifera L. Syst. Nat. ed. 10. 1007. 1759. Polygonum uvifera L. Sp. Pl. 365. Guaiabara uvifera House, Amer. a a 8: 64. 1922. Tree of strand areas, 2-17 m. tall, branches terete, stout, papillose to pilose, the nodes not tumid; ocreae rigid, coriaceous at the base, mem- branaceous at the apex, 3-8 mm. long, papillose to pilose; leaves of normal shoots with petioles stout, 7-10 mm. long, papillose to pilose, the blades orbicular to reniform, 6 & 8,11 & 13,13 X& 18 cm. long and broad, thick and fleshy when fresh, coriaceous when dry, glabrous and minutely punc- tate on both surfaces, the apex rounded truncate or emarginate, the base rounded to broadly cordate one lobe often extending around the petiole; midrib and primary veins prominent on both surfaces, frequently brightly colored when fresh, the primary veins 3—5 pairs, usually straight, bifurcate 92 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL and weakly anastomosing near the margin, commonly barbate in the axils of the basal veins, the secondary venation minutely reticulate or obscure; leaves of adventitious shoots usually variable in size and shape, commonly obovate; inflorescence stout, 15-30 cm. long, the rachis puberulent; stam- inate flowers in clusters of 1—7, the pistillate flowers solitary at each locus, the bracts ovate, 1-1.5 mm. long, 2 mm. broad, puberulent, the ocreolae membranaceous, 1 mm. long, puberulent, the flowering pedicels 1-2 mm. long, the perianth yellow-white or greenish, the hypanthium 2—3 mm. long, the perianth lobes 4 mm. long, 3—4 mm. wide, the fertile stamens to 4 mm. long; fruiting pedicels 3-4 mm. long; fruit obpyriform, 1.2—2 cm. long, 8— 10 mm. diameter, narrowed at the base, rounded-truncate at the apex, the perianth lobes appressed against the apex of the achene, the perianth rose-purple when mature, the achene black. DIsTRIBUTION: Throughout tropical America along the coasts. Antigua: Carr’s Ghaut, R.A. & E.S. Howard 11992 (GH); without location, Box 1434 (A, US). St. Barts: Euphrasen s.n. (s), Forsstrom s.n. (s), Questel 82 (Ny). Montserrat. Harris’ Lookout, Howard 11874 (GH). St. Kitts: Britton & Cowell 247 (ny), Sargent sm. (A). Guadeloupe: St. Francois, Howard 11787 (GH); without location Duss 2183 (us). Dominica: Carib Reserve, W.H. & B.T. Hodge 3316 (GH); Cabrites, Howard 11750 (GH); Hatton Garden, W.H. & B.T. Hodge 2057 (GH); Pointe Ronde, W.H. & B.T. Hodge 2686 (cH); Scot’s Head, W.H. & B.T. Hodge 1605 (cH). Martinique: Casa Pilota, Howard 11720 (cH); Presquile de la Caravalle, Howard 11726 (cH); Salinas Pie Beach, Egler 39-19 (ny); without location Belanger s.n. (B), Duss 1742 (us), Sieber 103 (Mo, US). St. Vincent: Kingshill, Howard 11125 (GH). Grenadines: Bequia, Dalton for H.H. Smith B-128 (cH); Mayero, Howard 11041 (cu), 11042 (cH); Petit St. Vin- cent, Howard 10912 (cH); Kick ‘em Jenny, Howard 10782 (GH). Grenada: Grande Anse, Broadway s.n. (rM, GH), J.S. Beard 203 (A). Barbados: Christ Church, Dash 129 (rm); St. Philip, Bovell 129 (us); without page Curran. (rm, Mo), Eggers 7347 (A). Trinidad: Galera Point, Broadway s.n. (A); Monos island, Williams (TRIN). Tobago: Buccoo Bay, Elmore V-4 (FM); eas Lands, Palo Saco, Russell (TRIN 13071); Farm Road, Williams (TRIN 11143); Lower Scarborough, Broadway 3617 (rm, s); Mt. St. Georges, Howard 10480 (GH); Speyside, Turley (TRIN 15244). Coccoloba venosa L. ee Nat. ed. 10, 1007. 1759; Fawcett & Rendle. Jour. Bot. 51: 123. 1913. Uvifera arbor americana, fructu aromatico punctatus, Pluk. Alm. 394. ¢. 237, 4. as to leaf only. Coccoloba punctata L. Sp. Pl. ed. 2, 523. 1762. Coccoloba nivea Jacq. Hist. Stirp. Amer. 115, pl. 78. 1763; Enum. Pl. 19. 1762. Coccoloba excoriata Duss, Ann. Inst. Col. Marseille 3: 168. 1897, not L. Guaiabara venosa House, Am. Midl. Nat. 8: 64. 1922 as Guaibara. Trees to 15 m. tall; branches terete, glabrous, the nodes not tumid; ocreae membranaceous, deeply cleft, acuminate on one side, or truncate, to 2 cm. long; glabrous or with flattened glands; leaves of normal shoots with petioles 5-10 mm. long, glabrous, the blades oblong-lanceolate to elliptic, 8 & 4, 10 & 4.5, 16.5 & 6.5, 21 & 9, 27 & 10.5 cm. long and 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VI 93 broad, membranaceous, glabrous except for clusters of hairs in the axils of the veins, sparsely glandular below, the apex short-acuminate, the base narrowed and slightly cordate or cuneate or obtuse; midrib and primary veins slightly prominent on both surfaces, the primary veins 8—13 pairs, straight or arcuate, bifurcate and anastomosing at the margins; leaves of the adventitious shoots about the same size, the internodes much elongate and the ocreae to 4 cm. long; inflorescence terminal or terminal on short lateral branches, the rachis puberulent, angular; staminate flowers in clus- ters of 2—5, the pistillate flowers solitary, the bracts lanceolate-ovate, to 1.5 mm. long, black, puberulent to pilose or commonly with a fringe of hairs at the apex; ocreolae membranaceous, enlarging with the expanding bud, each flower with an ocreola, to 2 mm. long, the flowering pedicels 1-2 mm. long, glabrous; hypanthium less than 0.5 mm. long, the perianth lobes broadly ovate, 1.5-2 mm. long and broad, slightly unequal, the fertile stamens to 1 mm. long; fruiting pedicels 1.5—2.5 mm. long, the perianth lobes fleshy, white or pink, enclosing the black achene, the hypanthium scarcely evident in the fruit, the fruit 3-4 mm. long and broad, the fruit broadly ovoid. DistrRIBUTION: Cuba (introduced), Hispaniola, Puerto Rico, Jamaica (?), Virgin Islands, Lesser Antilles and Trinidad. Antigua: Galley Bay, Box 1081 (us); Morris, J.S. Beard 351 (A, MO); Wal- lings, Howard 11989 (GH), without location, Fairchild sn. (us). St. Martin: Forsstrom sn. (s), Boldingh 2993B (ny). St. Barts: Forsstrom s.n. (NY, S), Le Gallo 2371 (A). Montserrat: Harris’ Lookout, Howard 11875 (cu). Tortola: Schafer 1125 (us). Rogen Bouillante to Pointe Noire, Howard 11842 (A); Gourbeyre, Duss 3250 (FM, NY, US); Riviere, Stehlé, Quentin & Bena 5313 (us); Les Saintes, ee 1763 (us). Dominica: Roseau to Canefield, Hodge 443 (GH. NY, US). Martinique: Carbet, Duss 1744 (ny); Casa Pilota, Hahn 967 (G. US), 1182 (s), Howard 11714 (GH). St. Lucia: Soufriere, Howard 11580 (GH). Grenadines: Kick ‘em Jenny, Howard 10792 (GH); < a Ronde, Howard 10713 (GH), 10717 (GH). Grenada: Broadway s.n. (Mo, , US). Barbados: Dodd's Botanical Station, Waby 131 (rm); Turners Hall Woods Gooding 594 (ny), Dash 262 (Ny). Trinidad: Castara, Sandwith 1831 (Ny); ene s Hill Re- serve. Swabey 12535 (TRIN); Galera Point, Broadway 2788 (BR, NY, TRIN); Guanapo, Dannouse s.n. (TRIN); Quinam Reserve, W illiams 12015 (xy, TRIN): Southwest Reserve, Marshall 12404 (TRIN), Swabey 12271 (TRIN); Southern Watershed Reserve, Russell 12261 (TRIN); Toco, Freeman a seca 11768 (NY. TRIN). Tobago: Mason Hall, Broadway 4160 (A, FM, GH TRIN), Howard 10470 (GH); Point opposite Melville Island, Freeman & v illiams 11422 (TRIN). Little Tobago: Swabey 12959 (TRIN). 94 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL THE GENERA OF THE NYMPHAEACEAE AND CERATOPHYLLACEAE IN THE SOUTHEASTERN UNITED STATES CARROLL E. Woop, Jr. THE TREATMENTS Of the ranalian families Nymphaeaceae and Cerato- phyllaceae presented here continue a series of studies toward a biologically oriented generic flora of the southeastern United States made possible by the interest and support of George R. Cooley and a grant from the National Science Foundation. The first paper in this series of generic treatments (The genera of the woody Ranales in the southeastern United States. Jour. Arnold Arb. 39: 296-346. 1958) includes a general explanation of the aims and scheme of this work. As noted there, the area covered in these studies is bounded by and includes North Carolina and Tennessee, Arkansas and Louisiana. The pattern of the descriptions, notes and references is the same as outlined and followed in the introductory paper. It may be called to attention again, however, that the descriptions are based upon the species which occur within this area, items in brackets being supplementary and not applying to our plants. The abbreviations of journals follow those of Schwarten and Rickett (Abbreviations of titles of serials cited by botanists. Bull. Torrey Bot. Club 76: 277-300. 1958). References which have not been checked are marked by an asterisk. The illustration of Nelumbo by Dorothy H. Marsh is from living mate- rials obtained through the enthusiastic assistance of Richard A. Eaton and Kenneth A. Wilson. NYMPHAEACEAE (WaATER-LILY FAMILY) Aquatic perennial {or annual] rhizomatous herbs, with alternate cordate to peltate floating or emersed leaves with involute vernation, and solitary, axillary, perfect, at least partially cyclic flowers. Plants usually with air- spaces, latex, vascular bundles without cambium or vessels, the leaf-tissues (especially) often with sclereids. Sepals 3-6|—12], green to petaloid. free or slightly united, hypogynous [in ours]. Petals 3-many, showy and colored, to stamen-like, free [in ours], hypogynous, or, with the stamens, inserted on the surface of the ovary, sometimes transitional to stamens. Stamens 3-200, extrorse or introrse; pollen more or less **monocotvl- edonoid,”” (monocolpate or derived types) or (in Velumbo) 3-colpate. Gynoecium of |1]|2—many carpels. free or united, superior to inferior. Ovules anatropous, many—2—1, pendulous from the top, the walls. or the abaxial suture of the carpels. Fruit a many-seeded berry, a nut, or 1-3 seeded, small and indehiscent. Seeds operculate (except Nelumbo). with or without an aril; embryo with thick and fleshy cotyledons, small and — 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 95 with abundant perisperm and little endosperm or (in Nelumbo) large and lacking both. (Including Cabombaceae, Nelumbonaceae. A family of eight well-marked genera (five of which occur in our area), of very wide distribution in quiet fresh waters. The genera are linked together by the aquatic rhizomatous habit, the usually long-petioled simple leaves (with a strong tendency towards peltation), the solitary, long-peduncled flowers with at least partially cyclic arrangement of parts, the anatropous, pendulous parietal ovules, the presence of air spaces, the latex-producing habit, and the monocotyledonoid vascular bundles. In respect to other characteristics (e.g., pollen morphology, embryology, oper- culum and other features of seeds, seedlings, anatomy, floral structure, etc.) the relationships are reticulate. Although a large body of literature exists, many of these latter items have been investigated only partially and it is difficult to draw complete comparisons throughout the family. Various divergences occur between the genera, especially in respect to the structure of flower and fruit, these generally being associated with special mechanisms which function in connection with pollination and with the protection, dispersal, and survival of seeds. The structural features of all these plants need further study and interpretation in connection with their biology. The treatment of Caspary, who regarded the family as composed of three well-marked subfamilies, is followed here. Some authors, however, split the group into three or even five families (in as many as three orders). These latter treatments seem unjustified, particularly in respect to the Cabomboideae (Cabomba and Brasenia) and to Euryale, Victoria and Barclaya, all of which clearly seem to be linked to Nuphar and Nymphaea. Nelumbo, the single genus of the Nelumboideae, is perhaps the most di- vergent member of the family, but even this genus may be regarded merely s being specialized in regard to leaves, dispersal mechanism, embryo and pollen. (See also Brasenia and Nelumbo below.) REFERENCES: BaILton, H. Monographie des Nymphéacées. Histoire des Plantes 3: 77-104. 1871 Fhe Gi BLeNK, P. Uber die durchsichtigen Punkte in dem Blattern. Flora 67: 49-57, 97-112, 136-144, 204-210, 223-225, 275-283, 291-299, 339-349, 355-370, 371-386. 1884. [Transparent dots on leaves of Nymphaeaceae, 100-102. | BruyNeE, C. pe. Idioplastes et diaphragmes des Nymphéacées. Compt. Rend. Acad. Sci. Paris 175: 452-455. 1922. Caspary, R. Les Nymphéacées fossiles. Ann. Sci. Nat. Bot. IV. 6: 199-222. 1856. ——. Nymphaeaceae. Nat. Pflanzenfam. III. 2: 4-10. 1888. Coox, M. T. The embryology of some Cuban Nymphaeaceae. Bot. Gaz. 42: 376— 392. 1906. [Includes Brasenia, Cabomba piauhyensis, Nymphaea ampla, N. pubescens, Nuphar advena var. cubana. n the embryology of Nymphaeaceae. Bot. Gaz. 48: 56-60. 1909. [Reviews similarities and disagreements; further notes on Nymphaea odo- rata, Nuphar advena, Nelumbo luteum. | Cutter, E. G. Studies of morphogenesis in the Nymphaeaceae. I — Introduc- tion: Some aspects of the morphology of Nuphar lutea (L.) Sm. and Nym- 96 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL phaea alba L. 11 — Floral development in Nuphar and Nymphaea: bracts and calyx. Phytomorphology 7: 45-56, 57-73. 1957 Gray, A. Cabombacaceae, Nelumbiaceae, Nymphaeaceae. Genera Pl. U. S. 1: 91-104. 1848. Gruss, J. Die Haustoren der Nymphaeaceen. Ber. Deutsch. Bot. Ges. 45: 459-— 466. 1927. | Mucilage-producing hairs in Nuphar, Nymphaea, Victoria. | sada et acre D. T. On some points in the morphology and anatomy of e Nymphaeaceae. Trans. Linn. Soc. London II. 5: 287-299. 1897. ae , F., F. REHNELT and L. DiIrTMANN. Das Buch der Nymphaeaceen oder Seerosengewaclise. 158 pp. Darmstadt, 1907. [ Monograph, keys, cultiva- tion, etc LANGLET, O., and E. Soperserc. Uber die Chromosomenzahlen einiger Nym- phaeaceen. Acta Horti Berg. 9: 85-104. 1927. Lawson, G. On the Nymphaeaceae. Trans. Roy. Soc. Canada 6: 97-125. 1889.* Li, H. L. Classification and phylogeny of Nymphaeaceae and allied families. Am. Midl. Nat. 54: 33-41. 1955. | Nymphaeaceae split into 5 families, 3 ae PLANCHON, J. E. Etudes sur les Nymphéacées. Ann. Sci. Nat. Bot. » 19s 17-63. 1853. [Systematic treatment. | Enumeration succincte des espéces de la famille des Nymphéacées. Revue Hort. Paris IV. 2: 63-68. a Raciporski, M. Die Morphologie der Cabombeen und Nymphaeaceen. Flora 78: 244-279. 1894. [Flower and vegetative shoots, especially. | . Beitrage zur Kenntniss der Cabombeen und Nymphaeaceen. Flora 79: 92-108. 1894. Supplementary and in reply to Schumann. | SCHAFFNER, J. H. Some morphological peculiarities of the Nymphaeaceae and Helobiae. Ohio Nat. 4: 83-92. 1904. [Attempts to show that Nymphaeaceae are monocotyledons. | SCHILLING, A. J. Anatomisch-biologische Untersuchungen iiber die Schleimbil- ung der Wasserpflanzen. Flora. 78: 280-360. 1894. [Mucilage hairs; in- cludes Brasenia, Cabomba, Nuphar, Nymphaea, Victoria, Nelumbo SCHUMANN, K. Die Untersuchungen des Herrn Raciborski iiber die Nymphae- aceae und meine Beobachtungen iiber diese Familie. Ber. Deutsch. Bot. Ges. 12: 173-178. 1894. [Criticism of Raciborski. ] Ursprunc, A. Zur Kenntnis der Gasdiffusion in Pflanzen. Flora 104: 129-156. 1912. | Bubbling of gas from leaves of Nymphaeaceae; critical review of literature. | WEBER, F., and G. Kenpa. Notizen iiber Nymphaeaceen-stomata. Protoplasma 40: 158-164. 1951. WEBERBAUER, A. Beitrage zur Samenanatomie der Nymphaeaceen. Bot. Jahrb. 18: 213-258. 1894. [All genera considered. | KEY TO THE GENERA OF NYMPHAEACEAE A. Carpels united, either along their sides or along the outer margins by adna- tion to a cup-like “‘receptacle;” stigmas radiate; ovules numerous in each locule; stamens numerous, introrse; fruit an irregularly dehiscent berry, ripening in the water; leaves with a basal sinus. Subfam. NyMPHAEOIDEAE. B. Perianth wide-spreading, of 4 sepals and 12-32 showy white, pink, blue or yellow petals; carpels sunken in a cup-shaped fleshy receptacle or hypanthium on the outer surface of which petals and stamens are in- 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 97 serted; carpels prolonged alae into slender incurved projections (car- Pellary styles) SCEGS AMA. © fo iste hues nce wy eda ee 1. Nymphaea. B. Perianth subglobose, of 6 concave yellow (green- or red-tinged) sepals and numerous scale-like or stamen-like “petals” inserted with the numer- ous stamens on the receptacle beneath the ovary; carpels completely united, the stigmas radiate and sessile on a disc; seeds not arillate MN oi 5 las aint Oye A ap AE MRD STS oles, noe eae Nuphar. A. Carpels free (although in Nelumbo embedded in the sand sal ovules soli- tary or 1-3 in each carpel; stamens hypogynous, few to many (3-36) and extrorse to slightly introrse, or very numerous and & xtrorse; gore leathery or hard, indehiscent; leaves (at least the floating or pene ones) peltate, lacking a basal sinus. C. Perianth of 6-8 segments, the flowers small; receptacle small, with 4-18 free, superior carpels; fruit small, 1-3 seeded; leaves all floating or sub- mersed; plants more or less coated with mucilage. Subfam. CABOMBOIDEAE. D. Plants with dissected, opposite submersed leaves and small, a floating leaves; perianth petaloid, white or purple; stamens SR res. = Da ae a ecaE Stag: 5 eis abo omba. D. Plants with only undivided, alternate, peltate floating leaves; sepals persistent, the petals dull purple; stamens 18-36; plants heavily CUALOC WIC) TIMGIAVE Com yARee RRO n gsc. yeaa ied a geen Brasenia. C. Perianth of numerous segments, the flowers large and showy receptacle large, top-shaped, with the many uniovulate carpels sunken separately in cavities in the upper side, only the stigmas protruding; receptacle en- larging greatly in fruit, the carpels maturing into nuts; stamens very numerous, extrorse, hypogynous; all leaves floating or emergent on strong petioles, centrally peltate, large, glaucous; plants lacking mucilage. Sub- Sv. MINED UMBONOMDEAR Merete Oe sree oe cae Seed as 5. Nelumbo. Subfam. NYMPHAEOIDEAE Casp. 1. Nymphaea L. Sp. Pl. 1: 510. 1753, partim, emend. J. E. Smith in Sibth. & Smith, Fl. Graec. Prodr, 1: 360. 1808-9, nom. cons. Perennial rhizomatous aquatic herbs with floating ovate to orbicular leaves, cleft at the base, the submerged stipulate petiole inserted at the base of the deep sinus; aerial leaves exceptional; submersed leaves often present, filmy and delicate. Flowers solitary, showy, white (to pink), blue or yellow, borne either at the surface of the water or raised above it. Sepals 4, nearly free, one completely outside the others; petals 12—ca. 32, in whorls of 4, 8, 8 (in subg. BRACHYCERAS), or 4, 8, with the arrangement of petals within these not clear (in subg. NymprHagA), in this last the petals often transitional to stamens. Stamens numerous, 50-150, in- trorse, with or without an appendaged connective. Gynoecium of 7-25 carpels, apocarpous or syncarpous, in a circle abutting and adnate to a central prolongation of the floral axis and sunk in the cup-shaped fleshy receptacle (or hypanthium?), upon the outer surface of which petals and stamens are inserted, and produced above it as “carpellary styles;” upper surfaces of carpels stigmatic to form a broad, concave radiate stigma. Ovules numerous, anatropous, pendulous from the sides of the ovary 98 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL locules. Fruit an irregularly dehiscent spongy berry, ripening under water; seeds 1—many per carpel, hard, operculate, embedded in mucilage and each surrounded by a bell-shaped floating aril, embryo small, with 2 equal fleshy hemispherical cotyledons, embedded in a thin layer of endosperm and abundant perisperm. (Castalia Salisb., nom, rejic.) TYPre species: JV. alba L. (From the Greek and Latin, Nymphaea, a water-lily, from Nympha, goddess of waters, meadows and forests.) — WATER-LILIES, POND-LILIES, WATER-NYMPHS. A genus of about 35 species of very wide occurrence (lacking in New Zealand and the Pacific slope of North America) in quiet, fresh (rarely brackish) waters. Represented with us by three species in two of the five subgenera. Subgenus BrAcHycerAs Casp., characterized by carpels free at the sides, thick, fleshy carpellary styles, and appendaged stamens, and _ in- cluding about 12 species of the tropics and subtropics, occurs in our area as NV. elegans Hook., a pale-violet-flowered species of southern Florida, the West Indies and Mexico. The sepals and 4 outer petals persist, be- coming turgid in fruit. Subgenus NyMPHAEA (Subg. Castalia DC.) characterized by carpels united at the sides, ligulate carpellary styles, obscurely veined sepals, and outer stamens with petaloid filaments, includes about 7 species, all of the northern hemisphere. In our area the subgenus occurs as the yellow- flowered N. mexicana Zuccarini (.V. flava Leitner), 2n = 56, of Florida, southern Louisiana, Texas, and Mexico, one of the most a in range, and the white-flowered NV. odorata Ait., a quite variable species. This latter occurs both as the typical form and the much larger var. gigantea Tricker (Castalia lekophylla Small). Some authors also recognize var. stenopetala Fern. and var. minor Sims. In view of this diversity and the two chromo- some levels recorded for this species (2u= 56, 84), further critical studies of NV. odorata are needed. A perhaps parallel case is the complex V. alba L., of Europe, in which 84, 105, 112, and 160 chromosomes have been re- ported. Flowers of all species are proterogynous. Some are self-fertile, others self-sterile. Our species are all day-flowering; at the latitude of Philadel- phia Conard records the flowers of N. mexicana as open for each of 2 days from 10-11 a.m. to 3-4 p.m.; those of NV. odorata open for 3 or 4 days from 6—7 A.M. to 12-1 p.m.; and those of V. elegans open from 7-8 A.M. to 12-1 p.m. When receptive, the stigmatic cup is filled with a sweetish, watery liquid. No nectar is secreted (although some species are very fragrant); insects visit the flowers for the abundant pollen. Numerous hybridizations (particularly within subg. BRAcCHYCERAS and subg. Loros) since about 1850 have produced many showy horticultural plants. Crosses between members of subg. NYMPHAEA have resulted in hardy forms (all sterile) with white, yellow, pink, orange or red flowers. Natural hybrids occur within this subgenus where two or more species grow in the same pond. All attempts to cross species of different subgenera 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 99 have failed: within a subgenus interspecific hybrids may be either sterile or fertile. Recorded chromosome numbers range from 28 to 224. The chromosomes of plants used in hybridization work apparently have not been determined. One of the smallest-flowered species (V. tetragona Georgi) and one of the largest-flowered (V. gigantea Hook.) are both high polyploids (2n = 120 and 224, respectively)! Nymphaea mexicana spreads rapidly by runners and produces on geo- tour shoots hibernating bodies or “brood bodies” which consist of a cluster of fleshy roots more or less resembling a “hand” of bananas, Rich in starch, these brood bodies are an important duck food. Nymphoides aqua- tica (Walt.) Kuntze (Gentianaceae) produces similar clusters of fleshy roots at the base of the inflorescence; the two are often confused. The nomenclatural confusion centering around the application of the name Nymphaea to this genus or to that now known as V uphar 72 Benth has resulted in the conservation of both of these names in the sense in which they are used here. REFERENCES: See also Nuphar: BIEBL, BRAND, CONARD, COOK, ROBINSON, SCHUSTER, and SPEGG. Curtiss, A. H. The yellow water lily of Florida. Pi. World 5: 106-109. 1902. [V. mexicana. | Conarp, H. S. The waterlilies: a monograph of the genus Nymphaea. Carnegie Inst. Wash. Publ. 4: i-xiii. 1-279. pl. 1-30, figs. 1-82. Washington, 1905. | The definitive monograph including history, structure, development, physi- ology, taxonomy, distribution, hybrids and culture. | Water-lilies: monocots or dicots? Am. Bot. 42: 104-107. 1936. | Dicots! N.B. tri-lacunar nodes, etc. | . The banana floatingheart ( Nymphoides aquaticum). Proc. lowa Acad. Sci. 44: 61-64. 1937. [A comparison of the hibernating bodies of %. mexicana and Nymphoides aquatica. | and H. Hus. _Water-lilies and how to grow them. i-xili. 1-228. pl. 1-31. New York. Dau, A. O. Cytological studies in Nymphaea L. (Abs.) Biol. Bull. 97: 257. 1949. |ma — 28 in N. odorata from Hyannis, tas ss. | GERM, H. Scheinbare Traumabewegungen von Nymphaea-Blattstielen. Verh. Zool.-Bot. Ges. Wein 92: 245-259. 1951. |[Temperature. geotropic effects in det “a leaves of N. alba. | e Kompensation der Wassertiefe durch das Wachstum der Blatt- und aia der Seerosen. Loc. cit. 69: 505-508. 1957. | Nymphaea alba. | Gessner, F.. and F. WeINFuRTER. Die geotrope Kriimmung der Nymphaea-Blatt- stiele. Ber. Deutsch. Bot. Ges. 65: 46-50. 1952. IrvINE. F. R.. and R. S. Trickett. Waterlilies as food. Kew Bull. 1953: 363- 370. 1953. Mackenzie. K. Kk. Proper use of the name Nymphaea. Rhodora 29: 234-237. 1927. | kor Nuphar.| Masters, M. T. The TS of the air-canals in the stalks of nvymphaeas. Jour. Roy. Hort. Soc. 26: 840-843. 1902. MELLINK, J. F. A. Zur ‘Thyllenteaee Bot. Zeit. 44: 745-753. 1886. | Hairs near wounds in petiole of N. alba. | 100 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL Nasu, G. V. Water lilies and other aquatics: their relation to horticulture. Jour. N. Y. Bot. Gard. 8: 202-220. 1907. P[r1nG], G. H. Hybrid nymphaeas. Missouri Bot. Gard. Bull. 22: 47-90, 93-108. 1934. |Tropical spp.; many data on fertility of hybrids in subg. Bracuy- CERAS and Loros. | . VAN P. Witson. ah to propagate tropical waterlilies. Home Gard. (N. Y.) 2: 71. Nov. 1943 Otis, C. H. The transpiration of as water plants: its measurement and its relationships. Bot. Gaz. 58: 457-494. 1914. [Includes Nymphaea. ] Ranps, R. L. The water lily in Maya art: a complex of alleged Asiatic origin. Smithsonian Inst. Bur. Am. Ethnol. Bull, 151: 75-153. RosHarpt, P, A. Schwimm- und Wasserblitter von Nymphaea alba L. Ber. sag Bot. Ges. 33: 499-507. 1915. Zahl und Verteilung der Spaltéffnungen in ihrer Abhangigkeit vom Licht beobachtet am Blatt von Nymphaea alba L. Ber. Schweiz. Bot. Ges. 1922: 22-25. ; SMALL, J. K. The water-lilies of the United States. Jour. N. Y. Bot. Gard. 32: 117-121. 1931. [Includes photographs of N. elegans, N. mexicana, N. odorata and var. gigantea. | SPRAGUE, T. A. The standard-species of Nymphaea L. Rhodora 30: 53-59, 1928. |[N. alba.] WacuTer, W. Beitrage zur Kenntniss einiger Wasserpflanzen, 3. Flora 84: 343-348. 1897. [Experiments on heterophylly in Nymphaea. | 2. Nuphar J. E. Smith in Sibth. & Smith, Fl. Graec. Prodr, 1: 361. 1808-1809, nom. cons. Perennial aquatic herbs with stout, creeping rootstocks, emersed, floating or submersed narrowly lanceolate to orbicular entire leaves with a deep sinus at the base; submersed leaves thin and delicate. Sepals usually 6[S-12], the 3 outer greenish to yellowish, the inner yellow, tinged with red or green. “Petals” numerous, linear to oblong, thick, stamen-like or scale-like, bearing a nectary on the outer surface and inserted with the very numerous introrse stamens on the receptacle under the ovary. Gynoe- cium of 5S—23[-30] carpels, resembling that of Papaver, with a stigmatic disc, each carpel with a stigmatic ray; ovary multilocular, with numerous ovules on the inner walls. Fruit an ovoid to columnar berry, dehiscing irregularly by the swelling of a gelatinous covering of the seeds; seeds usually ovoid, yellow to brown, opening by a small operculum, lacking an aril. Embryo with 2 fleshy cotyledons. Pollen monocolpate, oblate-sphae- roidal, spiny. 2n=34. (Nymphaea L. partim, emend. Salisb. and Nympho- zanthus L. C. Rich., nom. rejic.) Type species: N. luteum (L.) Sibth. & Smith. (The name from Greek nouphar, the flowers of a medicinal plant mentioned by Dioscorides, perhaps N. luteum.) — SparrERDocK, Cow- LILY, YELLOW POND-LILY. Perhaps 19 species in North America, two in Europe and several poorly known in eastern Asia; probably about 14 in our area, but some of these not well understood. T he group is taxonomically difficult, especially from nerbarium materials, in which the features of the plants are not well re- 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 101 tained. Ecological and ontogenetic variations add to the difficulty. Entire populations need study; specimens should be preserved in liquid and care- fully pressed flowers and adequate notes made in the field. Taxonomic characters used in the group include habit; shape and pubescence of leaf- blades; cross-sectional shape of petioles; presence or absence of sub- mersed leaves and their shape; number of sepals and color, shape and size of “petals;”’ shape of pistil; lobing, rays and color of stigmatic disc; shape and coloration of fruit; and color, size and indument of seeds. Special efforts should be made to press open flowers carefully and to record vari- ation in numbers of stigmatic rays; at least some discs should be pressed separately and longitudinal sections of pistils and fruits are valuable. All of the chromosome counts made thus far indicate a genus with a uniform chromosome number of 2n= 34. Largely sterile hybrids are known between the European V. pumilum and N. luteum, between NV. advena and N. sagittifolium, and between N. variegatum and N. microphyllum (of the northern U. S. and Canada). (This last also exists in a fertile form, N. x rubrodiscum Morong.) Where N. advena and N. variegatum over- lap in range (e.g., southern Michigan) intermediate plants are known, but these have not been studied in detail. The existence of hybrids of this kind has led to the treatment of the European and American representa- tives of the genus as a single variable species with 9 subspecies. Most entities appear to be definite units which maintain themselves both ecologi- cally and geographically, however. The flowers are proterogynous and are visited by pollen-collecting insects (sweat bees, flies and beetles) which work over the introrse anthers exposed by the recurving of the filaments as anthesis proceeds. The fruit ripens generally below the surface of the water, splitting irregularly, the carpels tending to separate and float as decay of the berry occurs. As in Nymphaea, long known for the occurrence of sclereids in blades and petioles, bizarre stellate sclereids occur at the intersections of vertical diaphragms in the petioles. Both genera have been the subject of numerous anatomical and physiological investigations, a great many of which have dealt with the European Nuphar luteum and Nymphaea alba. REFERENCES: Beat, E. O. Taxonomic revision of the genus Nuphar Sm. of North America and Europe. Jour. Elisha Mitchell Sci. Soc. 72: 317-346. 1956. [ Keys, descriptions, maps; 1 sp. (N. luteum) with 9 subsp. recognized. | Brest, R. Freilandbeobachtungen an Spaltoffnungen von Hydrocharis, Nym- phaea und Nuphar. Verhandl. Zool.-Bot. Ges. Wien 92: 249-253. 1951. {Outdoor observations on stomata of Nymphaea alba and Nuphar luteum. | Branp, F. Uber die drei Blattarten unserer Nymphaeaceen. Bot. Centralbl. 57: 168-171. 1894. [Submerged, floating and air leaves of N. luteum and Nymphaea alba. | Conarp, H. S. Nymphaea and Nuphar again. Rhodora 18: 161-163. 1916. Coox, M. T. Development of the embryo-sac and embryo of Castalia odorata and Nymphaea advena. Bull. Torrey Bot. Club 29: 211-220. 1902. [Early stages from the first, later from the second. | 102 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XL FERNALD. M. L. Nymphozanthus the correct name for the cow lilies. Rhodora 21: 183-188. Submersed leaves of Nuphar advena; Nuphar sagittaefolium in Virginia. Jn The seventh century of additions to the flora of Virginia. Rhodora 44: 394-396. 1942. | Includes NV. x es | A small ite of novelti II. A southeastern Nuphar. Rhodora 52: 65-67. 1950. . puteorum “Fern | FRITEL, P. H. Sur lattributi tion au genre sta de quelques espéces fossiles de la nore arctique. Bull. Soc. Géol. France 13: -297. GREEN, M. H. and W. S. Buck. The stereids in the petioles of Nymphaea advena. Proc. Pa. Acad. Sci. 11: 20-22. 1937. [Nuphar Richie HESLOP- Harrison, ¥. nig See intermedia Lede b., a sumed ad ale de Britain. Watsonia 3: 1953. and numerous eee eae N. luteum and N. ilies as dis- tinct species. | Lainc, H. E. Respiration of rhizomes of fied react and other water ~ plants Am. Jour. Bot. 27: 574-581. 1940. [N. advena. | — . Respiration of the leaves of ate Aa and Typha latifolia. ‘key, Jour. Bot. 27: 583-585. 1940. The mae eng of the internal atmosphere of Nuphar advenum and other water plants. Am. Jour, Bot. 27: 861-868. 1940. [1941]. ——. Effect of ea of oxygen and ppl of water upon growth of rhizomes of semi-submerged water plants. Bot. Gaz. 102: 712-724. 1941, | Maximum growth in NV. advena with 1-0 per cent oxygen; pressure equiva- lent to 4-6 ft. of water. | Love... J. H. Three fluvial pee and their visitors. Asa Gray Bull. 6: 60-65. 1898. | Includes * ‘N. adver = N. variegatum. | MILLER, G. S. and P. C. ek The North American species of Nymphaea. Contrib. U.S. Nat. Herb. 16: 63-108. 1912. [= Nuphar; based on served and herbarium materials; includes photographs of fruits and seeds. maps, keys; 19 sp. recognized in N. Am. | Moronc, T. Revision of the North American species of Nuphar. Bot. Gaz. 11: 164-169. 1886. PONCE DE LEON y CARILLO, P. Nuphar advena cubana var. nova. Revista Soc. Cuba. Bot. 4: 9-16. 194 Rogrinson, C. Flowers and ade I. Bot. Gaz. 14: 120-126. 1889. | Pollina- tion of NV. advena. | SCHUSTER, J. Zur Systematik von Castalia und Nymphaea. Bull. Herb. Boissier IT. 7: 853-868, 901-916. 1907. | Nymphaea and Nuph ar. | SEATON, 5S. The development of the embryo-sac of Nymphaea advena. Bull, Torrey Bot. Club 35: 283-290. 1908. |All stages. | Specc, H. Die Feuerwurzel. Deutsch. Apoth.-Zeitg. 96: 317-319. 1956.* | In- cludes morphology of N. luteum and Nym phae a alba. | WarpLaw, C. S. The effect of isolating the apical meristem in Echinopsis, Nu- phar, Gunnera and Phaseolus. Phytomorphology 2: 240-242, 1952 Subfam. CABOMBOIDEAE Casp. 3. Cabomba Aubl. Hist. Pl. Guiane Frang. 1: 321. ¢. 124. 1775. Strictly aquatic herbs with opposite submersed leaves and sometimes a few alternate floating leaves. Roots adventitious from the nodes. Sub- 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 103 mersed leaves petiolate, the blade divided into 5-7 parts at the summit of the petiole, each part divided dichotomously and trichotomously several times into many long, narrow segments. Floating leaves (when present ) peltate, small, narrow and tapered to both ends, or narrow and forked [or resembling miniature leaves of Brasenia|. Flowers borne singly on long axillary peduncles, the perianth white or purple [or yellow], sepals 3 and petals 3, slightly united near the base. Petals clawed, the blade often auriculate, Stamens 3—6, extrorse, the pollen 1-sulcate. Carpels [1]2-4, free, elongate, with small terminal stigmas; ovules usually 3, one on the adaxial. one on the abaxial suture and one on the wall near either. Fruit indehiscent: seeds 3, pendulous, operculate, covered with elongated processes; embryo as in Nymphaea, small. TYPE SPECIES: C. aquatica Aubl. (The name presumably a barbaric one.) —— FANWoRT, WATER- SHIELD. Seven species of warm and temperate regions of the New World. Two species, C. caroliniana Gray (2n = 24) and C. pulcherrima (Harper) Fas- sett. in our area. Cabomba caroliniana, with yellow-based white petals rounded at the summit, ranges from south-central Texas to southern Flor- ida. northward to Kentucky, southern Michigan and along the Atlantic coast to New York, Connecticut, and southeastern Massachusetts. Often used as an aquarium plant, this species appears to be adventive north of Virginia. Cabomba pulcherrima, with purple, emarginate perianth seg- ments. occurs at low altitudes from southwestern South Carolina to south- western Georgia and adjacent Florida. The terminal divisions of the leaves of both species have spathulate tips. As an aquarium plant Cabomba is most likely to be confused with Limnophila (Ambulia) heterophylla Benth. (Scrophulariaceae), which has sessile ternate leaves with the segments pinnately divided. REFERENCES: BEAL. W. J. Notes on Cabomba caroliniana A. Gray, Bull. Torrey Bot. Club 27: 86. 1900. Cook. M. T. The embryology of some Cuban Nympheaceae. Bot. Gaz. 42: 376-392. 1906. [Includes C. piauhyensis. | Fassett. N. C. A monograph of Cabomba. Castanea 18: 116-128. Dec. 1953. J. D. H[ooxer|. Cabomba aquatica. Bot. Mag. ¢. 7090. 1889. Ramuorst, J. D. VAN, an _ A. Frorscutitz. A new variety of Cabomba caroliniana Gray. Acta Bot. Néerl. 5: 342-343. 1956. [Var. paucipartita, a cultivated aquarium plant, origin unknown, with fewer and broader terminal leaf-segments. | 4. Brasenia Schreb. Gen. Pl. ed. 8. 1: 372. 1789. Aquatic herbs with floating, alternate, long-petioled, elliptic, centrally peltate leaves and small, solitary, axillary dull-purplish flowers. Rhizome small. bearing leaves and axillary runners which creep on the surface of mud, root at the nodes, and produce leafy branches and other runners. 104 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Submersed leaves present, at least in seedlings. Submersed parts of plant heavily covered with a mucilaginous jelly. Sepals 3-4, colored within, persistent. Petals 3—4, linear, sessile. Stamens 18—36, the filaments fili- form, and anthers slightly introrse; pollen 1-sulcate. Carpels 4-18, free, oblong-ovate, the ovules 2 in each, anatropous, pendulous from the abaxial suture. Fruit clavate, indehiscent, with 1 or 2 small, dull, grayish white, operculate seeds. Typr species: B. Schreberi Gmel. (Origin of name not explained by Schreber.) — WATER-SHIELD, WATER-TARGET, PURPLE BON- NET, PURPLE WEN-DOCK. A single species widely but sporadically distributed in ponds and slow streams in the West Indies, Mexico and Central America, Florida to Texas, north to Prince Edward Island, southern Quebec, southern Ontario, and Minnesota; also Idaho, California to British Columbia and Alaska: eastern Asia, Australia and Africa. This curious plant with centrally peltate leaves is clearly linked to Cabomba through leaves, floral structures, and pollen morphology. Both plants are sometimes treated as a separate family, the Cabombaceae. Seeds, seedling ontogeny, pollen morphology and embryology, as well as struc- tural features, do not bear out such a segregation, however. It may be noted that in Brasenia the mode of seed germination and the early seedling stages are essentially the same as in Nymphaea and Nuphar. The first leaf is awl-like, the second narrowly oblong with the petiole at the margin at the lower end. Subsequent leaves are eccentrically peltate, then cen- trally peltate, a total of 6-9 submersed leaves being produced before the first characteristic floating leaves. The submersed parts of the plant are notable for the extremely heavy covering of mucilaginous jelly, the secretion of numerous ephemeral glandular hairs (found also, but to a lesser degree, in the other genera, with the exception of Nelumbo). Specimens should be prepared with waxed paper, for the leaves become quite brittle in drying and adhere firmly to newsprint. Dianthesis has been reported in Brasenia in Japan by Tokura (cf. H. L. Mason, Fl. Marshes Calif. 491. 1957). The proterogynous flowers are raised above the water and are open for the first time from about 6 to 9 A.M., are then drawn into the water to reappear the following morning when the pollen is shed, and finally are withdrawn again. (Cf. dianthesis in Persea | Lauraceae] ). In the southern part of our range the plant is evergreen, but at least in the northern United States it produces in autumn winter buds which absciss from the parent plant and overwinter at the bottom of the pond. These reddish, translucent bodies consist of the thickened stem-tip with dwarf leaves with thickened petioles and the characteristic gelatinous cover- ing. Dispersal is effected not only by the winter buds but by the fruiting peduncles which become detached and float. 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 105 REFERENCES: Curyster, M. A. The winter buds of Brasenia. Bull. Torrey Bot. Club 65: 277-283. 1938. [Observed in Maine. |] Hixt, E. J. An observation on the water-shield (Brasenia peltata). Pl. World 3: 153. 1900. [Detached fruiting peduncles Keuer, I. A. The glandular hairs of Brasenia peltata Pursh. Proc. Acad. Nat. Sci. Phila. 1893: 188-193. 1894. NriewLanp, J. A. Habits of waterlily seedlings. Am. Midl. Nat. 4: 291-297. 1916. [Brasenia, Nuphar and Nymphaea. RayMonp, M. and P. DANSEREAU. The Sas aeons distribution of the bipolar Nymphaeaceae, Nymphaea tetragona and Brasenia Schreberi. Proc. 7th Pacific Sci. Congr. 7(5): 122-131. 1949. (1983), ScHRENK, J. On the histology of the vegetative organs of Brasenia peltata, Pursh. Bull. Torrey Bot. Club 15: 29-47. 1888. [Notes esp. mucilage, internal hairs and submerged leaves. | Toxura, A. On the blooming of Brasenia Schreberi J. F. Gmel. (1). eae Jap. Bot. 13: 829-839. 1937. (In Japanese.) [Dianthesis; cf. H. L. Mason, Fl. Marshes Calif. 491, 493. 1957. Subfam. NELUMBONOIDEAE Casp. 5. Nelumbo Adans. Fam. Pl. 2: 76. 1763. Aquatic herbs with fibrous roots borne at the nodes of slender horizontal vegetative rhizomes and thickened storage rhizomes, and with large, di- morphic (floating and emersed) peltate leaves and large yellow or pink flowers borne on stout peduncles well above water. Perianth of ca. 14-26 petaloid tepals, the outer 2—5 persistent (and thus calyx-like), the inner caducous in 2—4 days. Stamens ca. 200, with a conspicuous fleshy terminal appendage, extrorse, spirally inserted; pollen tricolpate. Carpels usually 12-28 (9-39), bore singly in cavities in the pithy obconical receptacle, basally attached, flask-shaped, the ovary oblong-cylindric, completely im- mersed, the style short, neck-like, the stigma circular, centrally depressed ; ovule solitary in each carpel, anatropous, pendulous from the top of the locule. Carpels becoming in fruit acorn-like very hard-walled nuts in the greatly enlarged receptacle. Embryo with only a rudimentary primary root, filling the fruit with two large fleshy cotyledons which inclose in a delicate stipule-like sheath a green plumule with two peltate involutely folded leaves and two rudimentary leaves. Seedling floating upon germi- nation; all roots adventitious. 2n=16. (Nelumbium Juss.) TYPE SPECIES: N. nucifera Gaertn. (Name derived from the Ceylonese name for N. nucifera.) — YELLOW LOTUS, YELLOW NELUMBO, GREAT YELLOW LILY, PoND-NUTS, WATER-NUT, WATER-CHINQUAPIN, WONKAPIN, YONKAPIN, YOCKERNUT, Duck ACORNS, RATTLE-BOX, ALLIGATOR BUTTONS. Two species, V. lutea (Willd.) Pers. (West Indies, Central America, Florida to Texas, and locally northward to southern New England, New York, southern Ontario, Minnesota and Iowa) and N. nucifera Gaertn. (China to Australia, India and Persia and introduced and established 106 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL locally through cultivation in the United States). The former species has sulfur-yellow to white flowers and nearly spherical fruits, the latter pink to white flowers and somewhat larger ellipsoidal fruits. Nelumbo lutea is one of our showiest and most interesting native plants. The greatly enlarged top-shaped receptacle with numerous separate carpels is a unique dispersal mechanism. The petals fall after 2-4 days leaving the 1. Nelumbo. a-p, NV. lutea: a, habit, « 1/20: ver-bud, 4; c, flower with petals scene sepals A etd ent, Ys d, stamens — note extrorse dehiscence, appendages, * 1; ceptacle at anthesis, the stigmas of carpels protruding, & '%; f, ede ae carpels, vertical section, X ™%; g, carpel at anthesis, & 2; h, carpel, vertical section, with pendulous anatropous ovule, X 2; i, receptacle with nearly mature green fruit, vertical section, X 4 j, dry r receptacle with mature fruit, X 14; k, green fruit at maximum size, * 1: 1, mature fruit, the ovary wall sclerified, & 1: m, embryo poe mature but green fruit, one cotyledon removed — note absence of hypocotyl and radicle, ? n, seedling with two leaves — note eccentric peltation, involute vernation, ad- ventitious roots, X 1%; 0, seedling with four leaves (only bases of petioles numbered in seque nce Re the fruit coat removed —note distribution of roots, beginning of r me, X 4%; p, tuber (a single swollen internode) with terminal bud and leaf- bud to right. & ™%; f, i, semi-diagrammatic 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 107 outermost 4—5 tepals which persist into fruit. Within a few days the peduncle bends just below the receptacle, tilting it first to about 45 degrees and later to an almost horizontal position. As the fruits near maturity the receptacle is returned to an upright position, but later is bent down- ward through 180 degrees. The dry receptacle breaks off, falls into the water and floats with the cavities downward. The individual fruits shrink greatly (14-24 their volume) in ripening and the ovary wall becomes very hard and impermeable. Germination of fully ripened fruit may thus be greatly delayed. (Nelumbo seeds are the longest lived of any known; roughly 1000-year-old seeds of V. nucifera from peat beds in the Pulantien basin of southern Manchuria have germinated consistently.) The fruits strongly resemble those of the chinquapin, Castanea pumila, in appearance and the very edible cotyledons (the green plumule is bitter) taste much like those of the same plant. The plant grows best in organic soil in water up to 6(8) feet deep and spreads as much as 50 feet radially each year by slender runners with internodes up to 5 feet long, the longest of any plant in our flora. It may thus under favorable conditions cover acres crowding out other aquatic vegetation and becoming a pest. The roots are entirely adventitious in 6 groups at the nodes, where lateral branching also occurs. In autumn one or two nodes at the growing tip develop into enlarged banana-like tubers 8-28 cm. long, the overwintering structures from which growth proceeds the following spring. The tubers of both species are highly edible: NV. lutea was planted by the American Indians for both seeds and tubers and N. nucifera is widely cultivated in Asia for the same items. Although a unique plant, Nelumbo appears to be related to the other Nymphaeaceae, differing from them primarily in the seed-dispersal mecha- nism, the more fully developed embryo (an enlarged and somewhat more specialized version of that of Nymphaea and Nuphar), the elimination of all submersed leaves, the more completely peltate leaves, and the tricolpate pollen. It is notable that in the leaves there are usually about 25 main veins, all but one of which branch dichotomously, this last being simple. Seedling leaves are apiculate at the end of the simple vein and are eccentri- cally peltate in the opposite direction, indications of the derived and specialized nature of the centrally peltate ale which have proceeded to an evolutionary level beyond Brasenia (q.v.). Leaf vernation is involute as in the other genera. Nelumbo is placed by some in a separate family (Nelumbonaceae Dumort., 1829; Nelumbiaceae Lindl., 1836), or even ina separate order, but it seems better for the present to retain the genus in the Nvmphaeaceae in the status accorded it by Caspary. REFERENCES: Anonymous. Longevity of Nelumbo praia Nature London 168: 905. 1951. |See OHGA. CHANEY, Lipsy, below Berry, E. W. Geologic history indicated by the fossiliferous deposits of the Wilcox Group (Eocene) at Meridian, Mississippi. U. S. Geol. Surv. Prof. Paper 108E: 61-66. 1917. |Nelumbo protolutea; map with past and present distribution of Nelumbo. | 108 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XL Cuaney, R. W. How old are the Manchurian lotus (Nelumbium nelumbo) seeds? Garden Jour. 1: 137-139. 1951. [See Lippy; drawings of modern and fossil seeds. | Conarp, H. S. Phyllody in Nelumbo. Trans. Proc. Bot. Soc be -350...331. 1904. [Dwarf N. nucifera ‘Cihwan’ or ‘pygmaea alba Prick with petals transitional to stamens and le a carpels in a single cavity in receptacle, suggesting apocarpous Nymphae passer lege D. E. A naturalized her of ray pe Ss bslate Mgt n New Jersey. Bull. Torrey Bot. Club 85: . [Mon ] eeu C. H. The meiotic cytokinesis of rear a — ae Bot. 4 296-306. 1922. [N. lutea, n = 8.} FARWELL, O. A. The color of the flowers of Nelumbo pentapetala. Rhodora 38: 272. 1936. FERNALD, M. L. The name of the American lotus. Rhodora 36: 23-24. 1934. GLeAson, H. A. The preservation of well-known binomials. Phytologia 2: 201- 212. 1947. [N. lutea vs. N. pentapetala. | Hatt, T. F. and W. T. PENrounp. The biology of the American lotus, Velumbo lutea (Willd.) Pers. Am. Midl. Nat. 31: 744-758. 1944. [The best general account of our plant; includes distribution, morphology, germination, ecol- ogy, control Heritace, B. Preliminary notes on Nelumbo lutea. Bull. Torrey Bot. Club 22: 265-271. 1895. | Careful study of rhizome-system and tubers Jones, J. A. Overcoming delayed germination of Nelumbo lutea. Bot. Gaz. 85: 341-343. 1928. [5 hours in concentrated sulfuric acid. Lipsy, W. F. Radiocarbon dates, II. Science 114: 291-296. 1951. [No. 629, seeds of N. nucifera from Pilantien Basin, s. Manchuria: 1040 + 210 yrs. ] Lyon, H. L. eee on ia embryogeny of Nelumbo. Minn. Bot. Stud. 2: 643-655. 1901. LW. Meyer, W. C. Dormancy yee growth studies of the American lotus, Velumbo lutea. Pl. Physiol. 5: 225-234. 1930. [Germination, growth of seed: effect of temperature, pH and soil. | Ouca, I. On the longevity of fruits of apes nucifera. Bot. Mag. Tokyo 37: 87-95. 1923. [From southern Manchur On the structure of some ancient, a still viable fruits of Indian lotus, with special reference to their prolonged dormancy. Jap. Jour. Bot. 3 1-20. 1926. The germination of century-old and recently harvested Indian lotus fruits, with special reference to the effect of oxygen supply. Am. Jour. Bot. 13: 754-759. 1926. [1927.] A comparison of the life activity of century- old and recently harvested Indian lotus fruits. Loc. cit. 760-765. A double maximum in the rate of absorption of water by Indian lotus seeds. Loc. cit. 766-772. . On the fertilization of Nelumbo nucifera. Cytologia, Fujii Jubl. Vol. 1033-1035. 1937. [Embryo as in Nuphar and Nymphaea; fertilization 6-8 hr. after pollination; embryo sac Polygonum type. ] Ouno, N. Uber lebhafte Gasausscheidung aus den Blattern von Nelumbo nuci- fera Gaertn. Zeitschr. Bot. 2: 641-664. 1910. [Physiology of streaming of gas bubbles from leaves. ] PEARL, R. Variation in the number of seeds of the lotus. Am. Nat. 40: 757-768. 1901. [9-15-35-39, normal distribution; statistical treatment. | Ropertson, V. Flowers and insects. III. Bot. Gaz. 14: 297-304. 1889. [In- cludes N. lutea. ] 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 109 SHAw, M. F. A microchemical study of the fruit coat of Nelumbo lutea. Am. Jour. Bot. 16: 259-276. 1929. [Includes histology. | STURTEVANT, E. D. The oriental Nelumbium naturalized in America. Gard. & Forest 2: 172, 173. 1889. [Near Bordentown, N. J.; tuber formation at 18 inches. | Taytor, Mrs. H. J. The history and distribution of yellow Nelumbo, Water Chinquapin, or American lotus. Proc. Iowa Acad. Sci. 34: 119-124. 1927. Wanc, H.C. A preliminary report on the morphological studies of the vegeta- tive body and yaaa of ane bud of Nelumbo nucifera Gaertn. Acta Bot. Sinica 5: eee WETTSTEIN, R. von. poe uber den Bau und die Keimung des Samens von Nelumbo nucifera Gartn. Verh. Zool.-Bot. Ges. Wien. 38: 41-48. 1888. [Structure and germination of seed. WIGAND, A. Nelumbium speciosum W. Bot. Zeit. Leipzig 29: 813-826. 1871. a German.) [{Development, morphology, anatomy and starch distribu- WU, v. " Des lotus sacres issus de graines vieilles de plusieurs siecles. Nature Paris 1957: 47. Ste York, H. H. The embryo-sac and embryo of Nelumbo. Ohio Nat. 4: 167-176. 1904. [Confirms Lyon; ee relationship to monocotyledons. | CERATOPHYLLACEAE (Hornwort FAmMILy) The family includes only a single small genus, Ceratophyllum L., of nearly cosmopolitan distribution, the plants growing entirely submerged in quiet, fresh waters. The group is easily recognized by the rather brittle and rough, whorled, dichotomously dissected leaves, the minute imperfect flowers, each solitary in the axil of one leaf of a whorl, the equally solitary branching, and the complete lack of roots. Ceratophyllum was first placed near the Nymphaeaceae in 1837 by Asa Gray, largely on the basis of comparisons drawn with Nelumbo [itself highly specialized but mostly along other lines], the chief points being the simple, one-seeded ovaries, the pendulous orthotropous, exalbuminous seeds, the large and fleshy cotyledons, and the unusually developed plumule. Unfortunately, the most strongly emphasized evidence, the ovule, is not orthotropous in Nelumbo, Brasenia and Cabomba, but anatropous. How- ever, Strasburger and others have maintained the same position for the Ceratophyllaceae, citing the convex receptacle, the numerous extrorse stamens, the superior ovary with the single pendulous parietal ovule, and the details of embryogeny (which seem to agree well with Nymphaeaceae). It may be noted that the stamens of Ceratophyllum seem to be spirally (not cyclicly) arranged, that the orthotropous ovule has but a single integument (instead of the two usual in Ranales) and that the pollen has lost all features which might be used as an indication of the relationships of the genus 1. Ceratophyllum L. Sp. Pl. 2: 922. 1753; Gen. Pl. ed. 5. 428. 1754. Aquatic herbs with whorled dissected leaves; monoecious, with minute sessile, axillary flowers lacking a perianth but with an 8—12-cleft involucre 110 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL in place of a calyx. Plants entirely submerged, branching, a single branch produced at a node. Roots lacking (even in the embryo) but leafy branches sometimes modified as “rhizoids.”’ Leaves 6-10 at a node, rather rigid, 1—4-dichotomously dissected, with two rows of minute teeth along the upper side of the ultimate segments, tipped by two bristles, usually with a purplish hair of tannin-filled cells between. Flowers usually solitary in the axil of one leaf of a whorl. Staminate flowers with an involucre of 8-12 segments, slightly united at the base, each tipped with a purplish hair, with a convex receptacle and 10-20 extrorse, 4-locular stamens with large, sessile, often 2-cuspidate anthers; pollen large, nearly smooth, thin-walled, acolpate. Pistillate flower of a solitary pistil with involucre, the style fili- form, oblique, the stigma a lateral pocket, the ovary ovate, 1-locular, with a single orthotropous 1-integumented suspended ovule; embryo sac development normal (Polygonum type). Fruit an achene tipped by the indurated persistent style [or its base] and often with additional basal or marginal spines. Seeds filling the fruit, nucellar tissue forming the only seed coat, with a large embryo with 2 ovate cotyledons, a well-devel- oped plumule of several nodes with simple or forked leaves, and lacking a hypocotyl and radicle; endosperm scant or lacking. Type SPECIES: C. demersum L. (The name from Greek, keras, a horn, and pfvyillon, leaf, in reference to the stiff, divided leaves.) — HorNwort, COoNnTAIL, Six or more species, three of which occur in our area: C. demersum L., of very wide distribution, from Canada to Argentina and also in Europe, Asia, Africa, and perhaps Australia; C. echinatum Gray, distributed from Quebec to Michigan and south to Florida and eastern Mexico; and C, floridanum Fassett, known only from the Florida Keys. The genus is almost completely adapted to an aquatic life. The plants lack cuticle, stomata and roots (even in the embryo, the anchoring function being taken over by leafy branches which grow into the mud); both xylem and phloem are reduced, mechanical cells are lacking, and about a third of the plant is occupied by air spaces. The greatly reduced flowers are borne abundantly and pollination is entirely subaquatic. Fruits develop only if water temperatures are maintained above 80°F (cf. Guppy), a phenomenon which may account for the scarcity of fruiting collections in herbaria. The plants may vegetate indefinitely, however, even under ice in winter; vegetative reproduction occurs by fragmentation Species of Ceratophyllum are highly variable and are difficult taxonomi- cally. The principal taxonomic characters used are those of fruits and leaves, the fruits being unarmed, or with either long or short spines at the base and the apex or along the margin, the leaves varying in the amount of forking, width of segments, degree of tapering at base, coarseness and distribution of the teeth. The female flowers are so reduced as to be use- less in classification, but stamens need further examination, particularly in connection with their terminal cusps and hairs. Chromosome numbers reported from European plants include 24 in C. demersum and 40 and 72 in C. submersum L. These, combined with the wide range of morphological 1959] WOOD, NYMPHAEACEAE AND CERATOPHYLLACEAE 1a characteristics, suggest that much more needs to be done toward an under- standing of the biosystematics of the genus. REFERENCES: Barton, H. Monographie des Pipéracées et des Urticacées. Hist. Pl. 3: 465- 537. 1871. [Ceratophyllum, 479-481, 483, 495. | Darwin, C. and F. The Power of Movement in Plants. x + 592 p. London, . [Movements of Ceratophyllum, 211. |] ENGLER, A. Ceratophyllaceae. Nat. Pflanzenfam. III. 2: 10-12. 1891. Fassett, N. C. North American Ceratophylium. Com. Inst. Trop. Invest. Ci. Salvador 2: 25-45. 1953. [4 sp., descriptions, maps, keys, variation. | Gtuck, H. Biologische und morphologische Untersuchungen iiber Wasser and Sumpfgewachse. II. Untersuchungen uber die mitteleuropaischen Utricularia- Arten, tiber die Turionenbildung bei Wasserpflanzen, sowie ber Ceratophyl- lum. xviil + 356 pp. Jena. 1906. [Biology of Ceratophyllum; formation of ““rhizoids. GopperT, H. H. Uber den rothen Farbestoff in den Ceratophylleen. Bot. Zeit. 6: 147, 148, 1848 Gray, A. Remarks on he structure and affinities of the order Ceratophyllaceae. Ann. Lyceum Nat. Hist. N. Y. 4: 41-50. Grier, N. M. Notes on comparative regeneration in Elodea and Ceratophyllum. Am. Bot. 26: 80-84. 1920. Guppy, H. B. Water-Plants and their ways. Their dispersal and its observation. Science Gossip II. 1: 145-147. 1894.* Their thermal conditions. Loc. cit. 178-180.* Ceratophyllum demersum. Loc. cit. 195-199.* [Important orig- inal observations. HESSLAND, I. On the occurrence of subfossil Ceratophyllum submersum L. Sv. Bot. Tidskr. 40: 235-256. 1946. [ Bibliography; fossil finds in Sweden and Denmark; some consideration of ecology. | Irmiscu, T. Kurze botanische Mittheilungen. 8. Dauer der Ceratophyllum- Arten. Flora 36: 527-528. 1853. [Shoots of C. may vegetate through the winter. | Jones, E. N. The morphology and biology of Ceratophyllum demersum. Univ. Iowa Stud. Nat. Hist. 13(3): 11-55. 1931. [Includes vegetative mor- phology, eg and eae biology in an Iowa lake KIRCHNER, O. , E. Loew C. SCHROTER. eae Again Lebensge- schichte i plat eee Noueesag 2(3): 51-73. 1917. [Compre- hensive account of structure and biology; extensive bibliography. | ee J.-E.-F. pe. Sur l’anatomie et le développement de Ceratophyllum. . Vet.-Akad. Handl. 9(10): 1-22. 1885. [Includes embryology. hase F. Uber die map inicrinace germ einiger Suswasserpflanzen und ihre Anpassungen an das Wasser und gewisse wasserbewohnende Insekten. Kosmos Stuttgart 10: 7-12. 1881. [Obeer tations on pollination of Cera- ee | McAteer, W. L. Eleven important wild-duck foods. Bull. U. S. Dep. Agr. 205: 1- . 1915. [C. demersum, 24, 25.] Masnven, A. Contribution a l’anatomie vasculaire de Ceratophyllum demersum L. Proc. Internat. Congr. Bot. 7: 381. 1950. MouraviEFF, I. Recherche sur la microspore du g. Ceratophyllum L. Le vacuome. Bull. Mens. Soc. Linn. Lyon 17: 37-42, 51-56, 70-77. 1948.* 112 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL . Sur les membranes externes de l’epiderme du g. Ceratophyllum L.— la cuticule. Bull. Mens. Soc. Linn. Lyon 21: 236-243. 1952 Sur la structure inframicroscopique des cuticules bagh Ceratophyllum demersum L. Ann. Univ. Lyon III. Sec. C. 7: 5-23. 19 MUENSCHER, W. C. Fruits and seedlings of Fale eae sie ‘te, Jour. Bot. 27: 231-233. 1940. [Comparisons of C. demersum & C. echinatum. | PEARL, R. Variation and differentiation in Ceratophyllum. Carnegie Inst. Wash. Publ. 58: 1-136. 1907. [‘Biometrical eam of intra-individuality and differentiation in Ceratophyllum |demers Pai | Ropier, E. The movements of a submerged wae plant. Nature London 16: 554-555. 1877. [Translation from Compt. Rend. Acad. Sci. Paris 84: 961-963. 1877. Ceratophyllum demersum. | SANDWITH, Mrs. C. The hornworts and their occurrence in Britain. Proc. Bristol Nat. Soc. 6: 303-311. 1926. SASTRI, R. L. N. The embryology of Ceratophyllum demersum L. (abs.) Proc. Indian Sci. Congr. Assoc. 42: 226 . SCHILLING, A. J. Anatomisch-biologische Untersuchungen iiber die Schleimbil- dung der Wasserpflanzen. Flora 78: 280-360. 1894. [Includes C. demersum. | SCHLEIDEN, M. J. Beitrage zur Kenntniss der Ceratophylleen. Linnaea 11: 513-542. 1837. |The first thorough account.| See also Linnaea 12: 344-346. STRASBURGER, E. Ein Beitrag zur Kenntniss von Ceratophyllum submersum und phylogenetische Erorterungen. Jahrb. Wiss. Bot. 37: 477-526. 1902. JOURNAL OF THE ARNOLD ARBORETUM VOLE APRIL 1959 NUMBER 2 STUDIES IN ARTOCARPUS AND ALLIED GENERA, III. A REVISION OF ARTOCARPUS SUBGENUS ARTOCARPUS FRANCES M. JARRETT Artocarpus J. R. & G. Forster, Char. Gen. 101. ¢. 51, 51a. 1776, nomen conservandum; Linn. f. Suppl. Pl. 61, 411. 1781; Lamarck, Encycl. Méth. 3: 207. 1789; Willd. Sp. Pl. ed. 4. 4: 188. 1805; Persoon, Syn. Pl. 2: 531. 1807; Blume, Bijdr. 479. 1825; Roxb. Fl. Ind. 3: 521. 1832; Blanco, Fl. Filip. 666. 1837, “Arctocarpus”’; Trécul, Ann. Sci. Nat. Bot. III. 8: 109. 1847; Miq. Fl. Ind. Bat. 1(2): 284. 1859, Suppl. 417. 1861, Ann. Mus. Lugd.-Bat. 3: 211. 1867; King in Hook. f. Fl. Brit. Ind. 5: 539. 1888; King, Ann. Bot. Gard. Calcutta 2 1. 1889; Becc. For. Borneo, 625. 1902; Renner, Bot. Jahrb. 39: 363. 1907; Elmer, Leafl. Philip. Bot. 2: 609. 1909; Gagnep. Bull. Soc. Bot. Fr. 73: 86. 1926; Gagnep. in Lecomte, Fl. Gén. Indoch. 5: 731. 1928; Corner, Wayside Trees of Malaya, 649. 1940; Jarrett, Jour. Arnold Arb. 40: 8, 11. fig. 1, f-k, fig. 3, d—-f. 1959. TyPE SPECIES: Artocarpus communis J. R. & G. Forster. Saccus or Soccus Rumph. Herb. Amb. 1: 104-115. #. 30-34. 1741. Sitodium [Banks & Solander ex] Parkinson, Jour. Voy. Endeavour, 45. 1773, nomen subnudum. Rademachia Thunb. Vet. Akad. Handl. Stockholm 37: 251. 1776; Houttuyn, Nat. Hist. II. Pl. 11: 446. 1779. Type spectes: Rademachia incisa Thunb. (= Artocarpus communis J. R. & G. Forster). Sitodium [Banks & Solander fia Thunb. Philos. Trans. Roy. Soc. London 69: 465. 1779, nomen illegitim Sitodium te [& Solander] e. ex Gaertn. Fruct. 1: 344. ¢. 71, 72. 1788, nomen illegitim Polyphema Teh Fl. Cochinch. 546. 1790. Type species: Polyphema jaca r. (= Artocarpus Ce ee Lamarck). ae 0. Kuntze, Rev. Gen. . 1891, nomen illegitimum. Small to large trees. Leaves spirally arranged (subg. Artocarpus) or alternate and distichous (subg. Pseudojaca), simple and entire to pinnat- ifid, or pinnate (A. anisophyllus), penninerved (basally trinerved in A. altissimus), thinly to thickly coriaceous, glabrous to pubescent or hispid; epidermal gland-hairs superficial to deeply immersed, heads globose to 114 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XL flattened and peltate, 1—16-celled; spongy mesophyll long-armed and usu- ally very loose (subg. Artocarpus), with ellipsoid to globose resin-cells (except in A. integer and A. heterophyllus), or compact, lacking resin-cells (subg. Pseudojaca); juvenile leaves entire to pinnatifid, or pinnate (A. tamaran), or with the lamina reduced to a sinuous wing along the midrib. Stipules paired, large, intrapetiolar, amplexicaul, scars annulate (subg. Artocarpus) or small, non-amplexicaul, scars lateral or intrapetiolar (subg. Beton Inflorescences unisexual, capitate, head cylindric to clavate, obovoid, ellipsoid or globose, lobed (in syncarp) or not, pedunculate (rarely sub- sessile), solitary or paired in leaf-axils, rarely borne on short shoots on older wood (subg. Pseudojaca) or rami- or cauliflorous (A. integer, A. heterophyllus) ; numerous flowers compactly covering surface, the perianths enclosing a single ovary or stamen, mixed with abundant stalked, peltate, varying to clavate or spathulate interfloral bracts, these often caducous from syncarp, or in some species entirely lacking, or in one (A. styractfolius) enlarging to form flexuous processes on syncarp; involucre absent (top of peduncle forming an annulus in A. heterophyllus). At anthesis: male head with perianths tubular and bilobed or perforate above, to 2—4-partite, stamens short- to long-exserted, anther-cells globose to oblong; female head with perianths tubular, thin-walled below and enclosing ovary, thick-walled above with a narrow lumen containing the style, partially or completely (in some species of subg. Pseudojaca) fused with one another to form syncarp; partial fusion occurring only between distal thick-walled region of perianths to form an external wall, with proximal thin-walled region remaining free; perianths completely fused above to form a smooth or areolate surface with the styles exserted through perforations, or the apices free, forming cylindric, conical, umbonate or truncate, rigid or flexuous processes on surface, all with styles exserted from their tips, or some proc- esses sterile (in some species of subg. Artocarpus), solid, elongate and projecting beyond the rest (some of these species also with sterile, solid, elongate flowers in the male head); ovary unilocular, the style apical to lateral, simple or bifid (branches sometimes very unequal), the ovule sub- apical to lateral. Mature syncarp formed by the enlargement of the entire female head, with one to many flowers forming fruit; proximal portions of fruiting perianths, if free, either becoming fleshy or not, the remaining perianths elongated and ribbon-like; mature ovary thin-walled to fleshy or horny, or developing an indurated endocarp freed by decay of rest of wall, the style apical to sub-basal, the seed large, attached subapically to sub-basally, testa membranous to pergamaceous, endosperm none, the embryo straight, orientation longitudinal or oblique, cotyledons equal or not, appressed faces at an angle of 0-90° to median plane of ovary, radicle and plumule small to minute, apical or lateral; germination hypogeal. DIsTRIBUTION: Ceylon, India, Pakistan, Burma, Siam, Indochina, south- ern China, Malaysia, Solomon Islands: two species (4. communis, A. heter- ophyllus) cultivated throughout the tropics. 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 115 HISTORY OF THE GENUS AND NOMENCLATURE OF BREADFRUIT, CHEMPEDAK AND JACK The genus Artocarpus is represented by indigenous species in an area extending from India and Ceylon to southern China, and through Malaysia to the Solomon Islands, while two species — the Breadfruit and the Jack — are now cultivated throughout the tropics. That it was, nevertheless, un- known to Linnaeus is somewhat surprising in view of the distinctiveness and economic importance of some of the species and the excellent descrip- tions which they had received from earlier authors. The Jack was probably briefly mentioned by Theophrastus in about 300 B.C. as occurring in India and it figured more extensively in many later travellers’ tales. The Breadfruit was first mentioned at the end of the sixteenth century by voyagers in the Pacific, where it was already wide- spread and of great importance to the Polynesians; it was recorded as far west as Java in 1642 and had become circumtropical in cultivation at least by the close of the eighteenth century. PrE-LINNAEAN BOTANICAL DESCRIPTIONS. Among early botanical writ- ers, Rheede described in 1682 both the Jack, which was already cultivated through most of tropical Asia and Malaysia, and a species of the Western Ghats, Artocarpus hirsutus, under the vernacular names Tsjaka-maram and Ansjeli (Hort. Ind. Malab. 3: 17, 25. t. 26-28, 32). At about the same time Rumphius prepared very full descriptions of the Jack and the Chempedak (then, as now, almost restricted to and less commonly cultivated in Malaysia) as Saccus arboreus major and S. a. minor: of the seedless and seeded forms of the Breadfruit and an allied wild species of the Moluccas as Soccus [sic] lanosus, S. granosus and S. silvestris, respectively; and, finally, of another Malaysian species, Artocarpus fretissii, as Metrosideros spuria (Herb. Amb, 1: 104-115. t. 30-34. 1741, 3: 26. t. 13. 1743). These are the most important pre-Linnaean botanical references to members of the genus, but several species from the Philippines were listed under their vernacular names, with brief descriptions, by J. G. Kamel (in Ray, Hist. Pl. 3, App. 51, 52. 1704). No attempt was made to identify Rumphius’ plants by Linnaeus in 1754 (Herb. Amb. [resp. O. Stickman] ), except for Metrosideros spuria, which he wrongly equated with Ochna jabotapita L. Nor did Burman provide any further identifications in his indices to the “Herbarium Amboinense” published in 1755 and 1769. In 1891 Kuntze held that the generic name Saccus had been validly pub- lished by Rumphius and, accordingly, he made many new combinations under it. These are illegitimate, however, and are omitted from the syn- onymy of the various species. Post-LINNAEAN BOTANICAL DESCRIPTIONS. Once the genus had become known to post-Linnaean botanists, it was described, within about twenty years, under four different generic names, Sitodium (1773. 1779 and 1788). Artocarpus (1776), Rademachia (1776) and Polyphema (1790), but the 116 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL name Artocarpus has been in continuous and undisputed use until doubt recently arose as to its priority, resulting in its conservation over Sitodium (1773). A review of the facts concerning these various publications follows, in which an attempt is made to determine their status and hence to estab- lish whether the conservation of Artocarpus was necessary. This review is extended to cover the nomenclature of the Breadfruit, Chempedak and Jack — the only species in which any serious problems have arisen — since this will involve the addition of only two items to the material literature and will avoid later repetition, as well as provide a detailed picture of the early botanical history of the genus. Except for the earliest description of Sitodium and Houttuyn’s publication of Rademachia the discussion will be largely based on the two papers published by Corner in 1939, although all the references have been carefully re-examined. For each name the Rumphian and post-Linnaean synonyms cited will be given and the factors taken into consideration in deciding its status will be clearly stated. SITODIUM, ARTOCARPUS AND RADEMACHIA. The Breadfruit was the first species of the genus of which a description was published under a Linnaean binomial. This was done, under the name Sitodium-altile [sic], by Stan- field Parkinson in 1773 in his edition of “A Journal of a Voyage to the South Seas, in His Majesty’s Ship, The Endeavour,” which was written by his brother Sydney, who accompanied Sir Joseph Banks and Dr. Solander as an artist on Cook’s first voyage around the world (1768-1771) but died on the way home from Batavia. In this was included a series of notes on plants seen in Tahiti, to most of which were appended, without acknowledgement, hyphenated ‘‘Latin names” corresponding to the bino- mials in Solander’s manuscript “Primitiae Florae Insularum Oceani Paci- ci, . . .” now in the British Museum (Natural History), London. This journal had been generally overlooked until Fosberg drew attention to it in 1939 (Am. Jour. Bot. 26: 229-231) and stated that Parkinson’s descrip- tion would necessitate the conservation of Artocarpus over Sitodium, which was finally achieved on these grounds in 1955 (Taxon 4: 162. 1955). In anticipation of this action the new combination Artocarpus altilis (Park- inson) Fosberg was published by him in 1941 (Jour. Wash. Acad. Sci. 31: 95) as the correct name for the Breadfruit. However, Sydney Parkinson’s notes are totally unscientific and mostly very brief; he was apparently quite unaware of the significance of his action in using those of the binomials that were new, as was his brother in publishing them, and few of them have any claim at all to valid publi- cation. The account of the Breadfruit happens to be rather longer than the others but is merely a general description of which the ‘‘botanical” part is given below. Sitodium-altile. . This tree grows to between thirty and forty feet high, has large palmated leaves, of a deep grass-green on the upper-side, but paler on the under; and bears male and female flowers, which come out single at the bottom or joint of each leaf. The male flower fades and drops off; the female, or cluster of females, 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 117 swell and yield the fruit, which often weighs three or four pounds, and is as big as a person’s head when full grown. It is of a green colour; the rind is divided into a number of polygonal sections; the general shape a little longer than round, and white on the inside, with a pretty large core. The fruit, as well as the whole plant, is full of a white clammy juice, which issues plentifully from any part that is cut: . In order for this description to be validly published it would have to be adequate as a ‘‘generico-specific” description, as was pointed out by Dandy and Exell in litt., cited by Merrill (The Botany of Cook’s Voyages. Chron. Bot. 14: 330. 1954) in the course of a detailed discussion of Parkinson’s journal. Merrill’s view that the description is not valid seems fully justi- fied, the more so if the circumstances of publication are taken into con- sideration. This question must, however, always remain a matter of opin- ion and, since it has now been settled, ae any who feel that they must recognize the validity of Parkinson’s description, by conservation, its chief importance is in determining the specific epithet to be applied to the Bread- fruit. The source of the generic name Sitodium is nowhere directly indi- cated in the literature of the eighteenth century, but it is now generally recognized that the authors should be cited as [Banks & Solander ex] (or [ Solander ex] ) Parkinson, etc. A botanical description of the Breadfruit was published in 1776 under the name Artocarpus communis, with Soccus [lanosus and granosus]| in synonymy, by J. R. & G. Forster, who were the botanists on Cook’s second voyage around the world (1772-1775) (the generic name being derived from the Greek artos, bread, and karpos, fruit). In the same year another description of the Breadfruit, together with one of the Chempedak, was published under the generic name Rademachia, with the specific epithets incisa (synonyms: Soccus lanosus, S. granosus and S. silvestris) and integra (synonyms: Saccus arboreus major and S. a. minor), respectively, by C. P. Thunberg in a paper dated Batavia, d.15 Junii 1775. The name was given in honour of Herr Radermacher, then ‘“Baillou” of Batavia, from whom Thunberg received much assistance, especially later, when he returned from Japan. It has not been possible to determine which of these two names was published first. The preface to the Forsters’ “Characteres Generum Plantarum” was dated November 1775, and there is no reason to doubt that the title-page is correct in giving the date as 1776, but no more precise evidence is available. Thunberg’s description was published in the July— September issue of the Kongl. Vetenskaps Acadamiens Handlingar for 1776, but Merrill (1.c., 359) stated that Hultén had been unable to find any record of the date of publication in Stockholm. Although it seems likely that Artocarpus was actually published before Rademachia, unless this can be established, the first person to choose between the two names must be followed. It has always been assumed that this choice was made by Linnaeus filivs in 1781 in his “Supplementum Plantarum” when he used the name Artocarpus, combining both generic descriptions, though without giving Rademachia as a synonym. He recognized two species, A. incisus and A. integrifolia [sic], basing his descriptions on those of Thun- 118 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL berg and citing his names, and the same Rumphian names, but not A. communis, in synonym It has, however, been ake that in 1779 Houttuyn had republished, in an account entitled “Beschryving van den Oostindischen Broodboom” which he included in his *‘Natuurlyke Historie,’ Thunberg’s description of Rademachia (from a manuscript in Thunberg’s hand sent to him by Rader- macher) and had chosen his name to replace that of the Forsters on the grounds that the description was more complete, in that the formation of seeds was mentioned, whereas the description of Artocarpus was based solely on the seedless form of the Breadfruit. Houttuyn cited “Artocarpus” in synonymy under Rademachia incisa, in addition to the Rumphian names, and this would certainly have necessitated the conservation of Artocarpus over Rademachia, but for the fact that he did not know of the earlier publication of Thunberg’s name in Sweden and hence was, giving his state of knowledge and intentions priority over the fortuitous circumstances, acting in contravention of present-day rules of nomenclature. Houttuyn’s specific descriptions were based, in part at least, on material sent to him by Radermacher and, from his account of Rademachia integra, it seems that he had received the foliage of the Chempedak, but the fruit of the Jack. He also described very briefly a third species, R. rotunda (Artocar- pus rotunda (Houtt.) Panzer, Pflanzensyst. 10: 380. 1783; in a translated and amplified version of Houttuyn’s paper), which it has not been pos- sible to identify, although Merrill (Jour. Arnold Arb. 19: 331. 1938) re- duced Artocarpus rigidus to synonymy under Panzer’s name. Rademachia rotunda will therefore be discussed further under that species. ( Merrill did not comment on Houttuyn’s reduction of Artocarpus.) THE BreApFRuIT. In determining the correct name for the Breadfruit, if the specific epithet altilis is rejected as invalidly published and Hout- tuyn’s action is ruled illegitimate, it remains to consider whether the choice made by Linnaeus filivs in forming the name Artocarpus incisus for this plant is legitimate. The use of the name Artocarpus communis in its place was first formally recommended by Richter in 1894 (Bot. Centralbl. 60: 169) on the grounds of its priority, and he was followed by various authors, including Merrill. The question was next fully discussed by Corner in 1939 (Gard. Bull. Singapore 10: 280) who concluded that, as the Rules then stood, Linnaeus filius was perfectly free, in making the choice between two simultaneously published names, to take the generic name from one genus and the specific epithet from the other (although he had chosen incisus over communis only by implication, this is all that is now required). Ac- cording to Article 57 of the International Code of Botanical Nomenclature, 1956, this is still true; there is nothing to indicate that the choice of generic name determines the specific epithet, where both are simultaneously published. Therefore, if the generic names and specific epithets are consid- ered independently, the argument is valid and it is only after much delibera- tion that the decision has been made to reject it in the interests of stability of nomenclature and in order to follow the general principles of the Rules. 1959} JARRETT, ARTOCARPUS AND ALLIED GENERA, III 119 In the present example, Article 57 is in conflict with Principles III and IV, especially the latter: ““Each taxonomic group can bear only one correct name [which for a species is a binary combination; see Art. 23], the earliest that is in accordance with the Rules, ... .” Artocarpus became the correct generic name through the choice made by Linnaeus filius; thus the earliest combination based on the Breadfruit that is in accordance with the Rules is Artocarpus communis (1776). The name Artocarpus incisus (1781) is also legitimate at present (apparently it is not in conflict with Article 11), but, if any change were considered necessary in Article 57 in order to deal with this unusual problem, the Article would undoubtedly be brought into line with the Principles. St. John (Pacific Sci. 2: 109. 1948), in the only other detailed considera- tion of the nomenclature of the Breadfruit, concluded that the name Arto- carpus incisus should be used, following both Article 57 (then no. 56) and established custom. However, there has been uncertainty as to the correct name of the Breadfruit for over fifty years. Artocarpus communis has been gaining gradually in acceptance over A. incisus and is now perhaps more widely used. With the addition of Artocarpus altilis as a third possibility, the situation has become more confused during the last two decades. It is evident that “established custom” cannot provide a stable solution to the problem. Moreover, in view of the long-standing uncertainty, any argu- ment based on custom has lost much of its force in this particular instance, although the general desirability of following custom, where there is genuine doubt as to the consequences of the Rules, is not questioned. It is regretted that the name Artocarpus incisus was used in the first paper of this series. THE CHEMPEDAK. Turning now to the second species described by Thun- berg and Linnaeus filius, this was until recently erroneously identified as the Jack, to which the name Artocarpus integrifolia was universally applied from the beginning of the nineteenth century until it was partially replaced in the literature by the legitimate combination 4. integer. This was made by Merrill in 1917 (as A. integra, Interpr. Rumph. Herb. Amb. 190). to replace Linnaeus fidius’ superfluous name, and was used by him in his consideration of Rumphian names in the same sense, i.e., as the correct name for Saccus arboreus major. In 1939, however, Corner demonstrated conclusively, in a paper on the distinguishing characters and nomenclature of the Jack and the Chem- pedak (Gard. Bull. Singapore 10: 56-81. ¢. 7, 2), that Thunberg’s de- scription had been based solely on the Chempedak. Although the latter had cited both Saccus arboreus major and S. a. minor as synonyms, and had given the Malay name of the tree as Tjampeda, but that of the fruit as Nanca (the Malay name for the Jack) and thus had, apparently, attributed the fruit of the Jack to the foliage of the Chempedak (as is also indicated by the material received by Houttuyn), only the vegetative parts were included in his description (with a note in Swedish that the roots, stems, branches and flowers were as in the preceding species [ Rademachia incisa | ). From the shape and colour of the leaf and the indumentum of rigid, patent 120 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL hairs on the lower surface, twigs and peduncles, Corner showed that the de- scription could refer only to the Chempedak (the Jack being glabrous), so that, in spite of the confusion in the Malayan and Rumphian names, the application of Thunberg’s name is unambiguous. The distribution was given by Thunberg as Java, around Batavia, and (from Rumphius) Am- boina and other places. Corner stated that Professor Svedelius had exam- ined the type specimen at Uppsala and had found that it consisted of ma- terial of the Chempedak agreeing with Thunberg’s description and includ- ing male inflorescences without the annulus at the base of the head that is found only in the Jack. Mixed with it was material of the Jack (see photographs in Corner, l.c.) which, from the note on the back of the sheet, “e Ceylona Thunberg,” had probably been collected in Ceylon, where the Chempedak does not occur (presumably on Thunberg’s visit there in 1777— 78 on the way back from Batavia). Professor Svedelius thought that by an error the two plants of different origin had been put on the same sheet when Thunberg’s collections were mounted some years later. This was con- firmed by the discovery, in the course of the present revision, of a sheet in the Rijksherbarium, Leiden, bearing only the Chempedak and labelled by Thunberg “RADEMACHIA integra Thunb.,” which was presumably part of the material sent to Houttuyn. THE Jack. Thunberg showed that he had completely failed to perceive any difference between the Jack and the Chempedak in a: paper published in 1779 with the title “Sitodium incisum et macrocarpon, ususque fructuum qui exinde nascuntur, descripta a Carolo Petro Thunberg, M.D.,” in which he altered the generic name Rademachia (now spelling it more correctly as Radermachia) to Sitodium without giving either his reasons or any source for the name, although he had presumably obtained it from Banks, who communicated the paper to the Royal Society of London. The generic de- scription (with Soccus, Saccus, Radermachia, and Artocarpus as syn- onyms) had only minor changes in wording and Rademachia incisa and R. integra were cited (in addition to the previous Rumphian names) under the two species, respectively, in synonymy. Therefore, although they are validly published here, the generic name Sitodium and the specific epithet macrocarpon must be rejected as nomenclaturally superfluous and thus illegitimate. Under Sitodium macrocarpon, Thunberg added to his previous Latin description of the Chempedak an account of the juvenile leaves and the fruit of the Jack (with a popular account of its uses in Ceylon, in- cluding recipes! ). His paper renders illegitimate as a later homonyn Gaert- ner’s independent publication in 1788 of the generic name Sitodium, at- tributed by him to Banks. This was based on another description of the fruit of the Jack under the name Sitodium cauliflorum, but, as stated by Corner, the latter must also be regarded as a superfluous name for the Chempedak, since the author’s intention, as indicated by the indirect cita- tion of Artocarpus integrifolia (‘‘Artocarpus foliis integris. LINN. [ed. Mur- ray,| syst. veg. 838.” [1784]) together with Soccus arboreus [sic] in syn- onymy, was to replace the earlier name, and this intention must be given 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 121 priority over the description and plates. Gaertner was the only author previ- ous to Corner (l.c.) to describe and draw the female perianths (called by him “baccae partiales”) as free below, though fused above. Lamarck, in 1789, was the first post-Linnaean author to make a distinc- tion between the Jack and the Chempedak, when he described in his ‘“‘En- cyclopédie Méthodique” three new species based on the Jack: Artocarpus heterophyllus (as A. heterophylla), with Saccus arboreus major as a syn- onym; A. jaca, which he recognized as close to the former, with Rheede’s Tsjaka-maram among its synonyms; and A. philippensis. Under Artocar- pus jaca he recognized a variety “8. Eadem ramulis petiolis foliisque sub- hirsutis,” giving Saccus arboreus minor, Artocarpus integrifolia and Rade- machia integra as synonyms. Unfortunately Lamarck’s “illegitimate” ac- tion, while making this perfectly correct distinction, in thus subordinating the variety based on the earlier names and descriptions applicable to the Chempedak (of which he had seen no specimens) under his new species must, following Corner, make Artocarpus jaca yet another superfluous name for the Chempedak. The latter therefore, since no previous effective choice had been made between the two other simultaneously published names (for both of which Lamarck had mentioned the characteristic an- nulus at the base of the inflorescence head), took A. heterophyllus as the more appropriate, so that this becomes the correct name for the Jack. Examination in 1955 of the type material of the three species in the herbaria of Lamarck and de Jussieu (the latter including that of Commerson) at the Muséum National d’Histoire Naturelle, Paris, has confirmed that it consists solely of specimens of the Jack and that the species were based on somewhat marked differences in the leaves. Lamarck also included Arto- carpus incisus (synonyms: Soccus lanosus, S. granosus, Rademachia incisa, “Artocarpus Forst.’’) in his account of the genus and described a fourth new species, A. hirsutus (as A. hirsuta), which he based entirely on Rheede’s Ansjeli. In spite of this distinction between the Chempedak and the Jack, the transference of the name Artocarpus integrifolia (with Rademachia integra) from the former to the latter was effectively achieved by Willdenow in 1805 in his edition of the “Species Plantarum” (the next work in which the genus was treated at all fully). He then “corrected” Lamarck’s nomen- clature, changing the name of his second species based on the Jack from Artocarpus jaca to A. integrifolia. He cited A. jaca, together with Radema- chia integra and Tsjaca-maram, in synonymy, and recognized two varieties, “8. ramulis petiolis, foliisque subtus hirsutis” with Saccus arboreus minor as the only synonym, and “‘y. foliis indivisis trilobisve’’ with Artocarpus heterophylla and Saccus arboreus major as synonyms. PoLyPHEMA. The description by Loureiro in 1790 of a new genus Poly- phema to take the place of Linnaeus filius’ Artocarpus integrifolia was due to a complete misunderstanding of the latter’s generic description, which led Loureiro to suppose that it could only correspond to the first species, Artocarpus incisus, not seen by him, and to make several distinc- 122 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL tions between it and the plants he was studying which he felt justified their segregation as a new genus. Under this he described the Jack as Polyphema jaca, with Saccus arboreus major as a synonym, and the Chempedak as P. champeden, with Saccus arboreus minor as a synonym, and thus correctly separated the elements confused in A. integrifolia (though only in the syn- onymy). His description of the Jack appears to be entirely original (the type is in the British Museum (Natural History), London), but that of the Chempedak is largely based on Rumphius’ account and it is doubtful whether he had ever seen a specimen. The plant is not found in Indochina in spite of his statement that it occurs in high woods there, and he may have been referring to a native species which he also used in making the description; the precise derivation of his account will be discussed further under Artocarpus integer. Willdenow did not mention Loureiro’s generic name, but in 1807 Spren- gel made it the basis of an illegitimate name for the Chempedak, Artocarpus polyphema, under which this plant has been most widely known in the litera- ture, although it was corrected to A. champeden in 1812 (by Stokes, Bot. Mat. Med. 4: 330; but universally ascribed to Sprengel, Syst. Veg. 3: 804. 1826). HISTORY OF THE GENUS FROM 1807. A few other species were described in Artocarpus, but the first major contribution to the taxonomy of the genus was made by Trécul in 1847 in his paper on the Moraceae: Arto- carpoideae and Conocephaloideae entitled ‘Mémoire sur la Famille des Artocarpées.”” He based his account mainly on the material then in the Muséum National d’Histoire Naturelle in Paris, and distinguished fifteen species, of which seven were new. In view of his limited material his work was very accurate and it is the only complete revision of the genus previously undertaken. He divided Artocarpus into two subgenera, Jaca (= subg. Artocarpus) |‘‘Perigonium masculum diphyllum, foliolis plus minusve inter se cohaerentibus; stipulae duae oppositae, amplexicaules, altera marginibus alteram. tegens; folia alterna” | and Pseudojaca |“ Peri- gonium masculum tetra- rarissime triphyllum; stipulae 2 minimae, axillares vel sublaterales, non oppositae, nec amplexicaules; folia disticha” |. These subdivisions have been found to be well marked, although, as was pointed out by King (1889), the distinction based on the male perianth is not valid, since this is more variable within subgenus Pseudojaca than Trécul realized, even within the species studied by him. His use of the term oppo- site with respect to the stipules was a source of confusion to King, as shown by Renner (1907), since he restricted it to stipules having their midlines 180° apart. In the years 1861 and 1867, Miquel published altogether 16 new species and combinations in Artocarpus (in addition to four that had appeared earlier), but many of the descriptions were based on poor material and it has only been possible to identify them by studying the types. More than one species was described under several different names and, although Miquel recognized Trécul’s subgenera, he included in subg. Artocarpus a 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III iZS species of Sloetia (Moroideae) and in subg. Pseudojaca one species (A. limpato) later transferred to Prainea and three which have been reduced to Parartocarpus venenosus (Zoll. et Mor.) Becc. (itself originally de- scribed as a species of Artocarpus in 1845). The next important paper on the genus was King’s monograph of 1889, “The species of Artocarpus indigenous to British India,” in which he de- scribed seventeen species (seven of them and two varieties as new) known to him as native in India, Ceylon, Burma and Malaya. They were usually correctly defined (except for the two varieties), but, owing to his inability to study the types of Miquel and other authors, several of King’s names must be replaced by earlier ones. He rejected Trécul’s subgenera and arranged the species in two groups, in the first of which [the distal por- tions of| the ‘“anthocarps”! were only partially united with each other so that the syncarp was spiny or tuberculate, while in the second they were completely united and the surface of the syncarp was smooth. Except for two new species included at the beginning of his first group (Artocarpus forbesii and A. bracteata) the subdivision does, however, correspond to Trécul’s subgenera; this distinction between them happens to be fairly satisfactory in the species which King was considering, although, as was pointed out in the introductory paper, it does not hold for the entire genus. These two species should have been referred to Parartocarpus (in which the leaves are spirally arranged with non-amplexicaul stipules and the inflorescences are involucrate) and they represent the two other species which will be recognized in the genus. King did not mention Parartocarpus, and, although the description had been published in 1875, it may not have been available to him. The syncarps of these two species closely resemble those of some species of Artocarpus and King mistook the solid spinous process on the surface (the nature of which was discussed in the introduc- tory paper) for tubular, perforate perianth apices. He did not consider leaf arrangement (explicitly at least), although his two groups differed in this respect also, and he rejected Trécul’s perianth and stipular characters. The rejection of the stipular character was due to a double confusion: as already noted, King failed to comprehend Trécul’s peculiar use of the term opposite; it has further been found in this study that he assigned a sterile specimen of Artocarpus rigidus to his A. bracteata, which caused him to describe the shoot as annulate, as it was in all the other species of his first group except A. forbesii, so that the stipular character would not give a ‘natural’ classification, whereas the syncarp character apparently would. He considered that whether or not the stipules were amplexicaul depended merely on their size, and evidently attached no special signif- icance to the occurrence of a basal involucre, which he observed only in A. bracteata. In 1902 Beccari described nine new species and one new variety of Arto- carpus from Borneo and New Guinea in an appendix to his ‘‘Nelle Foreste 'Te., the female ae this term can, strictly, be applied only to those perianths enclosing mature seeds. It has not been used, since it would merely complicate the specific descriptions. 124 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL ) di Borneo,” in which he also considered the characters in which Prainea and Parartocarpus differ from Artocarpus, though without fully clarifying them (see under Prainea above and Parartocarpus to follow), and removed King’s two wrongly assigned species to Parartocar pus. Finally, Renner, in a paper entitled ‘“Beitrage zur Anatomie und Sys- tematik der Artocarpeen und Conocephaleen, insbesondere der Gattung Ficus” published in 1907, discussed the systematy of the genera being studied, chiefly on the basis of their leaf anatomy, and, reinstating Trécul’s subgenera, concluded that they, together with Prainea, should be treated as sections under Artocarpus. Renner’s action in uniting Prainea with Arto- carpus has been fully discussed in the introductory paper to this series and has been rejected, since the differences in leaf anatomy (and leaf arrange- ment) between the two subgenera (here maintained at this rank) could well have arisen by divergence from a common ancestor with the somewhat intermediate Prainea, whereas the unique syncarpium of Artocarpus indi- cates a close taxonomic relationship between the subgenera and contrasts markedly with the free female perianths of Prainea. Later authors have ignored both Renner’s reduction of Prainea and Trécul’s subgenera. Ren- ner listed the anatomical characters for each of the species of Artocarpus available to him, citing the specimen(s) he had examined, but suggested no further taxonomic innovations, although his arrangement of the species follows a fairly “natural” order. The examination of leaf anatomy has been repeated for each of the species here recognized and details of the microscopic epidermal gland-hairs and of the hypodermis, if present, are given for each of the species in subg. Artocarpus, since their characters are of value in establishing the series. Two purely taxonomic treatments should perhaps be mentioned, namely, Elmer’s “Synopsis of Artocarpus” treating 16 species of the Philippines (1909), and Gagnepain’s account of the genus in the ‘Flore Générale de Indochine” (1928) for which he described eight new species in 1926, several of which must be reduced. There should also be added here a note on a paper which has only recently come to my notice, on the structure of the jackfruit, by D. D. Sundararaj and K. Ramanujam, “Studies on the floral morphology of Jack, Artocarpus heterophyllus Lamk. (Syn.: A. integrifolius, L.f.),” S. Indian Hort. 1: 56-61. figs. 1-12. 1953. These authors have described and illus- trated the structure of the jackfruit at anthesis and maturity, and have emphasized, as was done in the introductory paper to this series, that the perianths are free below but fused above, in this species leaving the apices free, so that they have recognized three regions in each perianth. They have also pointed out again that, as had been made clear by Corner (Gard. Bull. Singapore 10: 56-81. 1939), the true fruits are enclosed by the free fleshy lower portions of the fruiting perianths. They describe the syncarp as a false fruit. However, they do not seem to have examined any very young female heads, in which the perianths would presumably be entirely free (as in 4. rigidus, described in the introductory paper). They also state positively, but without bringing forward any evidence, that the an- 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III io nulus represents the involucral bract of the inflorescence, whereas Corner had noted (l.c., 62, 76) that he had been unable to discover its morpholog- ical nature, although it was probably comparable to the bracts in Pararto- carpus. This may well be the origin of the irregular row of bracts and the slight swelling of the top of the peduncle seen in some other species of Artocarpus but it is possible that the annulus represents a secondary enlargement of the latter. GENDER OF ArTOCARPUS. Although the International Code of Botanical Nomenclature now states that Artocarpus must be treated as a masculine noun, since the Greek masculine noun karpos has been further altered to the masculine form in Latinization (Recommendation 75A), the genus has been considered as feminine by almost all previous authors except Linnaeus filius (1781), Renner (1907) and Corner (1939, 1940). The original specific epithet in Artocarpus — communis —was ambiguous in this re- spect, while Linnaeus filius treated his first species as masculine — A. incisus — but the second as feminine — A. integrifolia. Renner pointed out that this error probably arose in type-setting, since both the other species on the same page had feminine terminations, and he supposed that it had been perpetuated through the custom in Latin of treating all names of trees as feminine. In this revision only the original form in which a given name was published is indicated, if different from that now accepted (in addition to significant typographical errors). CLASSIFICATION OF THE GENUS The two subgenera of Artocarpus, established by Trécul in 1847, are here recognized on grounds which have already been fully discussed in the introductory paper in which they were also keyed out with the genera (p. 26, above). The characters distinguishing them, of which the vegeta- tive are of more practical importance than the reproductive, may be sum- marized as follows: in subg. Artocarpus the leaves are spirally arranged with large, amplexicaul stipules leaving annulate scars and they have a long-armed, usually very loose spongy mesophyll with ellipsoid to globose resin-cells, and epidermal gland-hairs with a 4-16-celled head, whereas in subg. Pseudojaca the leaves are alternate and distichous with small non- amplexicaul stipules leaving lateral or partially intrapetiolar scars, the mesophyll is compact, and the heads of the gland-hairs are usually uni- cellular. The syncarp in subg. Pseudojaca is globose or somewhat lobed with a fleshy, smooth or papillate surface, while in subg. Artocarpus both shape and surface are variable, though the syncarp usually bears distinct indurated or fleshy processes, or, where it is smooth or nearly so, it is ellipsoid to cylindric. While the species in this genus are, on the whole, distinct and readily separable, this is not always the case with the subdivisions between the level of the subgenus and the species. There is also a marked contrast 126 JOURNAL OF THE ARNOLD ARBORETUM, [VOL. XL in the range of variation exhibited within the two subgenera and this is re- flected in the classification here suggested for the species within them. In subgenus Artocarpus there is a wide range of variation — within the fairly strict limits imposed by the structure—in the syncarps (especially in the shape, length, consistency and indumentum of the processes) and the species can be recognized on the differences in these alone, although they are supported by more or less marked differences in the male inflorescences (in shape, size and surface, and in peduncle indumentum and length) and in the shoot (in indumentum and leaf size and shape). The pattern of variation warrants the creation of two sections with two and four series, respectively, the basis for which will be discussed below. In subgenus Pseudojaca, on the other hand, while two sections will also be created, one is for an anomalous species, Artocarpus altissimus, and the other presents little variation between the species. The inflorescences are so reduced that the entire male and female heads provide the unit of varia- tion in establishing the species. The heads of different species vary in size and shape. in the length and indumentum of the peduncle, and, in the female head. in the appearance of the surface at anthesis and maturity, its indumentum, the degree of exsertion of the styles, and whether or not the interfloral bracts persist or lobing develops. There are also vegetative differences between the species —in shape, size, prominence of venation, petiole length and indumentum of the leaf — but, while most species can, with practice, be identified when sterile, the differences are often not read- ily definable. As might be expected where there are limited possibilities for variation. parallel evolution has occurred, especially in the vegetative characters, and the species may only be definable by using a combination of characters. The only discontinuous ‘‘morphological” variation occur- ring in this section is in the shape of the interfloral bracts, which is used to separate three species as one series of limited geographical distribution, leaving the remainder as a large group which is probably best treated as a single series. Although subsidiary groups can be recognized within the latter, the definitions necessary for these are somewhat complex, and there are also several species showing reticulate relationships, and a few of which the alliances are doubtful. In the keys to both subgenera proportions have been used — length to breadth of inflorescence head in subg. Artocarpus, and length of head to length of peduncle in subg. Pseudojaca. These recognize and utilize characteristic differences between species groups in the shape of the head or in the relative length of head and peduncle. These differences hold, with a few exceptions. for both male and female inflorescences, thus making possible the construction of a ‘natural’ key which can be used (at this dichotomy at least) at any stage in the development of the inflorescence. since the relationships remain fairly constant throughout its growth. The use of these ratios also allows for the considerable variation in size, rather than shape. which may occur in the inflorescences, depending largely on the part of the plant on which they are borne. There are some species, especially in subg. Pseudojaca, where these relationships do not 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 127 always hold, but they are relatively few in number and have been allowed for as far as possible. ECOLOGICAL AND BIOLOGICAL NOTES The species of Artocarpus, except for a few belonging to subg. Pseudo- jaca, are large trees of the high forest and are mainly restricted to ever- green forest in the tropical everwet zone or in areas with a relatively mild monsoon climate. They are usually found below an altitude of 3000 ft.. though several species may occur up to 5000 ft. One species, Artocarpus communis, is abundant in swamp forest in New Guinea, and others are re- corded from swampy areas or from the banks of streams. The majority of the species in everwet forest are themselves evergreen, although they may, especially in subg. Pseudojaca, show marked flushes of new growth, but Artocarpus dadah is deciduous. Of the species restricted to regions with monsoon climate some, including A. lakoocha, are deciduous, but others, such as A. gomezianus, are evergreen. The latter (as ssp. gomezianus) is the only entity within the genus to show a marked climatic discontinuity — between Lower Burma and Siam, and central and eastern Java. Artocarpus lakoocha, which is the most wide-ranging species on the continent of Asia, is also the most tolerant of cool temperatures and low rainfall, although it is damaged by frost. In considering biological aspects of the variation in growth habit and in the syncarp in this genus several more or less parallel trends may be distinguished. The first tendency is shown in the arrangement of the leaves, which in subg. Artocarpus are borne spirally on usually ascending twigs. whereas the ultimate shoots in subg. Psewdojaca, with their alternate and distichous leaves form more or less applanate sprays of foliage. Associated with the latter habit is a general reduction in massiveness, but within each subgenus there are also tendencies to reduction in the size of parts, cul- minating in subg. Artocarpus in A. kemando, and in subg. Pseudojaca in A, styracifolius. There is in subgenus Artocarpus considerable variation in the form of the adult leaves, which are pinnate in A. anisophyllus (though unusual in not disarticulating when they fall) and pinnatifid in A. communis and its allies. all these species having very stout twigs. In the saplings of several species the leaves are very large and highly dissected (twice or thrice pinnatifid and to six feet in length in A. elasticus). Members of this subgenus were mentioned by Corner in his exposition of the “Durian Theory” (Ann. Bot. IT. 13: 367-414. 1949) as showing in their sapling and adult stages the transition from a “pachycaul” habit, with a sparingly branched trunk and large, compound leaves, to the modern “leptocaul” tree, with slender twigs and horizontal sprays of simple leaves. They were also noted as exhibiting the principle of ‘“‘axial conformity” or correspondence between the massive- ness of the axis and the size and complexity of its appendages, which may be extended through A. anisophyllus and A. communis to the most slenderly constructed species of both subgenera. Except in two cauliflorous species 128 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL of subg. Artocarpus (A. integer and A. heterophyllus) there is a general correspondence throughout the genus between the size of the twigs and leaves and that of the syncarps, which are borne with the male inflores- cences among the foliage (although before they mature the subtending leaves may have fallen). A different trend was pointed out by Corner in A, heterophyllus, in which the twigs were slender with simple leaves, tend- ing to form an applanate spray (cf. subg. Pseudojaca), while the enormous syncarp was borne on the trunk or large branches, a necessary development if the massive fruit was to be retained. The syncarps within the genus also vary between two types of organiza- tion, namely, those in which each seed is surrounded by a pulpy perianth and the rest of the syncarp is more or less fibrous or indurated, and those in which the entire syncarp is soft and fleshy but the fruiting perianths are not markedly hypertrophied. The first type is best developed in the armoured syncarps of the first section of subgenus Artocarpus (which in- cludes A. rigidus), and the second is found in the small, nearly smooth syncarps of A. kemando and its allies, and of subg. Pseudojaca. The re- maining species in subg. Artocarpus have small to fairly large, moderately fleshy syncarps, usually with well developed firm, or fleshy and often flex- uous processes on the surface. They lack (so far as evidence is available) conspicuously enlarged fruiting perianths, except in the very large syncarps of the two cauliflorous species. However, the perianths in A. elasticus and A, sericicarpus are apparently somewhat pulpy. In the species with an armoured syncarp and also in a number of species in subg. Pseudojaca the seeds are enclosed in more or less well differentiated horny endocarps which may become free by the decay of the rest of the pericarp, while in other species of subg. Artocarpus the entire pericarp is indurated. This protec- tion is lacking in the cauliflorous species, but in A. heterophyllus there is a horny outer testa. In spite of the varying nature of the attractive element in the syncarp of Artocarpus the mode of dispersal is much the same throughout the genus, which is primarily adapted for distribution by arboreal mammals, although the smaller fruits of subg. Pseudojaca, many of which are yellow, red or purple, may also be eaten by birds. (Troup (Silvicult. Ind. Trees 3: 884. 1921) records that the fruit of A. lakoocha is eaten by parrots and minahs, and also by monkeys.) The larger fruits in the genus, according to Ridley (The Dispersal of Plants. 1930), are bitten to pieces on the tree by monkeys, squirrels and civet-cats. He describes how the orange syn- carps of Artocarpus rigidus (for which the English name is the Monkey Jack) are eaten in Malaya by a macaque monkey, which readily tears apart the spiny covering to reach the edible sweet orange pulp surrounding the seeds, and how as much fruit may be thrown away as is eaten. It is pre- sumably by means of this wastage of portions of the fruit, which are let fall by the animals eating them, that the seeds are more widely dispersed, but they will, in any case, germinate in syncarps which fall to the ground from the tree. Ridley suggests that the enormous syncarps borne on the trunk in A. integer (and presumably also in A. heterophyllus) are eaten 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 129 by forest ungulates, such as wild cattle, pigs and elephants, which can easily reach them. It should be noted that the seeds in this genus have no period of dormancy, but germinate immediately, and that they are unable to withstand dessication. The mode of pollination in Artocarpus is variable. Corner (Wayside Trees, 650. 1940) recorded that male heads in A. heterophyllus, A. integer and A. dadah had a sweet scent of honey and burnt sugar, attracting small flies and beetles, which were the pollinating agents, but that A. rigidus, A. communis and A. elasticus were apparently wind-pollinated, since the male heads had no scent, but gave off clouds of pollen when they were moved. Van der Pijl (Ann. Bogor. 1:79-82. 1953) studied pollination in Artocarpus heterophylius and found that the anthers emerged over a period of some days and produced a sticky pollen. The flies bred on the decaying flower- heads which had fallen from the tree and thus established a brief symbiotic relationship with the plant during the period of anthesis. Subgenus Artocarpus Artocarpus subgenus Jaca Trécul, Ann. Sci. Nat. Bot. III. 8: 110. 1847. Artocarpus section Jaca Renner, Bot. Jahrb. 39: 363. 1907. Leaves spirally arranged, simple and entire to pinnatifid, or pinnate (A. anisophyllus) ; gland-hairs superficial to deeply sunken, heads 4—16-celled; spongy mesophyll long-armed and usually very loose, with ellipsoid to globose resin-cells (except in A. integer and A. heterophyllus). Stipules large, intra-petiolar, amplexicaul, scars annulate. IJnflorescences with or without sterile, solid, elongate perianths. Male head, perianths tubular, shallowly (rarely deeply) 2 (—3)-lobed above, or perforate. Syncarp globose or subglobose, covered by indurated processes, or ellipsoid to cylindric, varying subglobose, covered by firm or flexuous processes or less commonly areolate. This subgenus is divided into two sections, Duricarpus and Artocarpus, based on the shape of the inflorescence heads and the presence or size of the interfloral bracts, on various characters of the syncarps, and, less cer- tainly, on the mode of germination; but there are also three anomalous species which will be discussed below. The first section is a fairly homo- geneous group of seven species, characterized primarily by the syncarp which has the surface armoured (whence its name) by the indurated free apices of the perianths. The free proximal region of those perianths that form seeds becomes fleshy, and encloses an ovary having a terminal style and developing a more or less clearly differentiated horny endocarp (be- coming free by the decay of the rest of the pericarp in A. anisophyllus and A. lanceifolius). The embryo is symmetrical with a minute apical radicle and the appressed faces of the cotyledons at an angle of 30—90° to the median plane. The syncarp is globose or nearly so, and the male head varies from globose to cylindric or clavate, but is never more than three and one half (rarely four) times as long as broad. In both there 130 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL are conspicuous peltate interfloral bracts, but these are mostly lost from the syncarp at or before anthesis, although a few frequently persist. The mode of germination is known in several species and the first pair of leaves subsequent to the cotyledons is consistently opposite, though later leaves are spirally arranged. For comparison with the series of the next section, it may be noted that a hypodermis is present in the leaves of only two species and that it consists of cells which are elongate in surface view. while the gland-hairs have a globose, 4—16-celled head. On the basis of the leaf characters, together with a difference in the shape of the male heads, two species are separated as series Laevifolii, leaving the rest in series Asperi- The second section, Artocarpus, is larger and more heterogeneous, but the species are united and are distinguished from sect. Duricarpus by their more or less elongate inflorescences, the absence or small size of the inter- floral bracts and the fleshy syncarp with the fruiting perianths becoming markedly fleshy in only two species, and the mature ovaries with a lateral or sub-basal style and no clearly differentiated endocarp (so far as the in- ternal structure is known; exceptions presented by the anomalous species will be noted below). The contrast in the shape of the inflorescences is more marked in the male head which is usually at least four times as long as broad (though as little as one and one half times in 4. heterophyllus). The syncarp is, however, usually ellipsoid to cylindric and only rarely sub- globose, and the surface is either covered by firm, or more or less fleshy and often flexuous processes of varying shape or, less commonly. merely areolate. Interfloral bracts are frequently entirely lacking and. when present, they are scattered and inconspicuous, with small heads to 0.2 mm. across. The position of the stvle, though variable, is always at least one third of the way down the ventral face of the ovary at maturity. and the radicle is likewise ventral, so that the long axis of the embryo is oblique. The appressed faces of the cotvledons lie either in the median plane of the ovary, in which case the embrvo is usually svmmetrical, or they are also more or less oblique, with the uppermost cotyledon frequently reduced in size. Details of germination are known for only four species (4A. communis, A, integer, A. heterophyilus and A. elasticus) but the first leaves appear to be always spirally arranged, except, perhaps, in 4A. communis, and to be preceded by scale leaves. The classification of the species within section rtocarpus has presented some problems, in addition to those originating in the three species that have been separated as anomalous. It has been concluded that four series should be recognized and these are based, as in sect. Duricarpus. primarily on characters of the leaves, namely the shape of the gland-hairs. and the presence or absence of a hypodermis and of resin-cells. However. with one exception, each is further distinguished by characters that are unique to it. Series Rugosi, with seven species. is readily distinguished by the anatomical characters of the leaves, which have a continuous hypodermis of isodiametric cells and gland-hairs with a depressed-globose. 4-celled head, and by the variously ridged surface of the male head, from which it receives its name. 1959] JARRETT. ARTOCARPUS AND ALLIED GENERA, Il 131 In the other species of the section the surface of the male inflorescence is smooth (except occasionally in A. integer) and an irregular hypodermis is found only in some specimens of A. blancoi. Series Incisifolii, with six species, is also well distinguished by the flattened, peltate, 8(—16)-celled heads of the gland-hairs, the frequently pinnatifid adult leaves, the inflated hairs usually covering the syncarp processes and male perianth lobes, and the relatively large size of the anthers, which range from 0.3-1.5 mm. in length (in no other species of the subgenus do they exceed 0.5 mm.). The remaining series. Cauliflori and Angusticarpi, each with two species, have entire adult leaves with gland-hairs having a depressed-globose, 6—10-celled head and consistently slender twigs, as compared with series /ncisifolii. Series Cauliflori is distinguished from series Angusticarpi (and all other species of the subgenus) by the very large syncarps borne on the branches and trunk. with markedly hypertrophied fleshy fruiting perianths, the absence of resin-cells from the leaves, and possibly also by the germination of the seed, in the course of which the two cotyledons separate to allow the emergence of the plumule. Although the cauliflorous syncarp of these species gives them a very distinctive appearance, the characters just listed do not appear to be of sufficient importance to justify separating this series from the rest in the section. Series Angusticarpi thus remains as an un- specialized. though generally reduced, group which does not show clear alliances with any other series. The appearance of the leaves is quite distinct from that of series Cauliflori, as may be seen from the key to the subgenus. The value of the svncarp in classifying this section, although considerable in distinguishing the species, is limited at the level of the series, since there has been parallel evolution in the development of sterile perianths forming elongate, solid processes on the surface in one or more species of all series except Cauliflori: in series Incisifolii and Angusticarpi there may also be similar perianths in the male inflorescence. While the embryos may well provide characters distinguishing the series — those of series Incisifolii have relatively well developed cotyledonary stalks and radicles and the entire pericarp wall is indurated — they are very variable from species to species and evidence is unfortunately insufficient to determine their taxo- nomic value. Details are given for each species in the section when avail- able. The validity of these groups is supported by their geographical distribu- tion. Section Duricarpus and series Rugosi are both centred in western Malaysia (Malaya, Sumatra, Borneo, Java) and the former occurs north- ward to Sikkim and Indochina, and is perhaps represented by indigenous species in the Philippines, while the latter has one species reaching southern Tenasserim and Palawan, and another extending through the Philippines, to Celebes and the Moluccas. Series /ncisifolii, on the other hand, has species in the Philippines, the Moluccas and New Guinea (and perhaps Melanesia if rtocarpus communis is indigenous there). The most wide- spread groups are series Cauliflori and Angusticar pi. The first has one species (4. integer) extending from western New Guinea to Malaya and reaching peninsular Burma and Siam, although it is absent from the Philippines, 132 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL and another (A. heterophyllus) perhaps indigenous in peninsular India. The second has one species in Malaya and Sumatra, and one witha distribu- tion similar to that of A. integer but extending to the Nicobar Islands instead of north of the Malayan border. On the basis, primarily, of the shape of the inflorescence heads, three species, which are otherwise anomalous in that each is, in various charac- ters, intermediate between the sections, are placed at the end of sect. Arto- carpus. Owing to this intermediacy and the fact that the structure of the mature syncarp is known in only one of them, they are not assigned to any of the series here established. Although a separate series should, perhaps, be created for each one of them, this is deferred, pending a more complete knowledge of the syncarps. Of these species, Artocarpus hirsutus, from peninsular India, has an ex- ceptionally long and slender male inflorescence indicating an affinity with sect. Artocarpus, but a subglobose to shortly cylindric syncarp covered by indurated, narrowly cylindric processes, which is somewhat similar to the syncarps of sect. Duricarpus. From the account given in Troup (Silvicult. Indian Trees 3: 876. fig. 323. 1921) it is evident that the mode of germina- tion is also the same as that found in sect. Duricarpus. The characters of the perianth and ovary do not, however, appear to be in agreement. The internal structure of the mature syncarp is not known, but the submature fruiting perianths are thin and the ovary is membranous with a subapical style. In view of these differences, together with the shape of the male head and the complete absence of bracts from the inflorescences (although this last character may have arisen by parallel evolution), this species is assigned to sect. Artocarpus. The induration of the syncarp processes is assumed to be of less taxonomic importance, although it provides, together with the hispid twigs and the appearance of the leaves (the gland-hairs have a depressed-globose, c. 6-celled head), a superficial resemblance to A, rigidus, The two other species, Artocarpus nobilis and A. sepicanus, are remark- able in that, although one comes from Ceylon and the other from New Guinea, they share the characters of narrowly cylindric inflorescences (cf. sect. Artocarpus) and well developed, peltate interfloral bracts (cf. sect. Duricarpus) which are persistent and conspicuous on the syncarps. In other respects, however, the appearance of the syncarps is very different: in A. nobilis the surface is covered by short, umbonate, indurated processes, while in A. sepicanus the syncarp is fleshy, with a pubescent, apparently completely smooth surface (cf. species of subg. Pseudojaca) and the styles are exserted through perforations in this. The internal structure is known only in A. nobilis, in which the fruiting perianths are thin, the Ovary is pergamentaceous with a subapical style, and the symmetrical embryo lies obliquely in the median plane of the ovary with the relatively large radicle ventral (cf. series /ncisifolii), although in its germination, as in A. hirsutus, the first pair of leaves is opposite. A. nobilis also shows an alliance with series Incisifolii in the shape of the gland-hairs, which have a peltate, flattened, 8-celled head, but in A. sepicanus the gland-hairs. which are 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 133 unusual in being deciduous at a very early stage, have a globose, 6—8-celled head. These three species and, more especially, the curious resemblances between A. nobilis and A. sepicanus are a reminder of the long evolutionary history that can be assumed for this genus. It would appear that they are “relicts”, and the persistence of a few such species that do not readily fit into a plascneation intended to reflect the affinities of the groups dominant today is scarcely surprising. The series recognized within section Artocarpus are not, however, neces- sarily closely allied and it is not suggested that any one of them is directly ancestral to the others. Section Duricarpus may represent either an offshoot from sect. Artocarpus or an independent line of evolution within the genus. The anomalous species, as has already been implied, are probably the remnants of a wider variation exhibited by the genus in the past, and this may also be true of series Cauliflori. KEY TO THE SPECIES OF ARTOCARPUS SUBGENUS ARTOCARPUS Male head globose to clavate or short-cylindric, length/breadth = 1-3.5(-4), the surface largely covered by the peltate heads of numerous bracts; syncarp subglobose, length/breadth = 1-1.5, the surface covered by indurated proc- esses; adult leaves pinnate or entire (rarely trilobed above the middle), the latter usually with fewer than 20 intercostals. Syncarp processes narrowly cylindric, 5-8 mm., hispid, peduncle 4.5-6.5 cm. (male head narrowly cylindric, 7— i. > 0) a ee South- PERNA etl nly cased De Re EE acs: 8 7-8 . A, hirsutus. ne sence not as above. Sect. DURICARPUS. . Twigs and leaves sue! male head ellipsoid to short-cylindric; syn- carp processes glabrous or pubescent. Series LAEVIFOLII. eaves pinnate; syncarp processes elongate, pedals cylindric, 6-8 & 1-1.5 mm., glabrous. Malaya, Sumatra, Borneo a eA Lk or) aed eee 1. A. aaah Leaves simple; syncarp processes short, broadly eerie trunca c. 3.5 & 3 mm., and surface are tesselate, or slightly tapering, obtuse, c. 1.5 x 1.5 mm., and somewhat separated, appresse pubescent. Malaya, Sumatra, Bader Wee ees 65 2. A. lanc caleles . Twigs and leaves (at least on the main veins beneath) hispid or hispid- pubescent; male head globose to obovoid or clavate; syncarp processes hispid. Series ASPERIFOLII. 5. Syncarp processes broadly cylindric, 1.5-3.5 mm. long. 6. Leaves smooth above, hispid-pubescent on the main veins only beneath; male peduncle 1-5 X 0.3 cm.; syncarp processes hispid from patent, aac) slightly crisped hairs. Indochina, British _ a Ww INOrtn BOIMmeCO. Macnee ee ne tr5 3. A. melinoxylus. 6. Leaves had above, hispid-pubescent throughout beneath; male peduncle 6-7.5 X c. 0.15 cm.; syncarp processes hispid from appressed hairs. Northeastern India, rare Andaman and. NICODAL, $5. ngs su eases 23. A. kemando. . Leaves obovate- elliptic varying elliptic, shortly and obtusely acuminate to retuse, young leaves ap- pressed-puberulent above; peduncles eo onaNs, in male c. 0.5 mm., in female 0.3-0.8 cm.; syncarp velutinous, tesselate a very aes truncate proc- esses. Malaya, Sumatra. . . A, maingayi. ie) “I Artificial keys for the identification of specimens bearing either male or female inflorescences are provided at the end of the treatment of this subgenus, and the second of these is illustrated. Section Duricarpus Jarrett, sect. nov. Folia adulta pinnata (A. anisophyllus) vel simplicia, integra, juvenilia pinnatifida; hypodermis absens, vel imperfectum, cellis elongatis composi- tum. /nflorescentiae bracteis interfloralibus peltatis, syncarpio subdeciduis. Capitula mascula subglobosa, obovoidea, clavata, ellipsoidea vel breviter cylindrica. Syncarpia globosa, vel subglobosa, processibus induratis ob- tecta; ovaria stylis terminalibus; semina testis pergamentaceis, rubris, peri- carpiis induratis, perianthiis carnosisque inclusa; embryum in longitudi- nem positum, cotyledonibus aequalibus, radicula supera. TYPE SPECIES: Artocarpus rigidus Blume. Although the distinctive feature of this section is the syncarp, there is considerable variation in its appearance due to the differing shapes of the indurated processes (free perianth apices) covering it, which may be either narrowly cylindric (A. anisophyllus and A. odoratissimus, with clavate tips in the latter), or tapering so that the syncarp is echinate (A. hispidus, A. rigidus), or short and obtuse or truncate so that the surface is more or less clearly tessellated (A. lanceifolius, A. melinoxylus, A. chaplasha). Other characters, however, suggest that these resemblances do not neces- 138 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL sarily indicate the closest relationships of the species. Artocarpus anisophyl- lus stands out by reason of its pinnate leaves, but in the presence of a hypo- dermis, the deeply sunken gland-hairs, and the glabrous shoot, a close affinity is shown with A. lanceifolius, which is confirmed by the ellipsoid to cylindric male inflorescences, and the strongly differentiated endocarp. Artocar pus odoratissimus, A. hispidus and A. rigidus are likewise united by their obovoid, subglobose or clavate male inflorescences, the elongate, hispid syncarp processes, and the hispid shoots. Between them may be placed A. melinoxylus and A. chaplasha, which resemble the latter group in the characters of the shoot and the male inflorescence, but A. lanceifolius in the surface of the syncarp. These are the only two species in the section of which the distinctness is in any doubt, and the characters separating them are listed under A. melinoxylus. As noted above, two series are dis- tinguished on the basis of leaf anatomy, indumentum of the shoot and shape of the male head, of which the first, series Laevifolii, includes Arto- carpus anisophyllus and A. lanceifolius, while the rest of the species are placed in series As perifolit. Series Laevifolii Jarrett, ser. nov. Ramuli et folia subglabri; folia hypoderme imperfecto, cellis elongatis composito, glandulis profunde immersis, capitibus globosis, 4—6-cellis; inflorescentiae masculae breviter ellipsoideae vel subcylindricae. Type species: Artocarpus anisophyllus Miq. 1. Artocarpus anisophyllus Miq. Fl. Ind. Bat. Suppl. 422. 1861, “anisophylla;”’ Renner, Bot. Jahrb. 39: 366. 1907; Merr. Pl. Elmer. Born, 45. 1929; Corner, Wayside Trees, 652. ¢. 191, 192. 1940. Holo- type, Sumatra, Teysmann HB 3698 (wu); isotypes (BO, kK, L). Artocarpus klidang Boerl. Handl. Fl. Ned. Ind. 3: 333, 371. 1900, in clavis. ee superba Becc. For. Borneo, 625, 1902; Merr. Philip. Jour. Sci. 18: 1921, Enum. Philip. Pl. 2: 43. 1923; Ridley, Fl. Malay Penin. 3: 356. oN Holotype. Sarawak, Beccari PB 2007 (FI); isotypes (FI, K). Evergreen trees, height to 45 m. Twigs 10-20 mm. thick, rugose, ap- pressed-puberulent and minutely punctate; annulate stipular scars c. 1.5 mm. broad, prominent; lenticels scattered. Stipules 4-17 cm. long, broadly ovate-lanceolate, acute, exterior rugose, subappressed-pubescent, hairs yellow to brown. Leaves imparipinnate; rachis 40-80 cm. (fide Elmer in Merrill, 1929, to 150 cm.) long, base to 7 mm. in diameter, terete, finely rugose; pinnae 5—12 on each side, usually in subopposite pairs with longer and shorter pairs alternating and the latter raised above the rachis, terminal pinna often bifid; juvenile leaves with distal pinnae decurrent on rachis ind the pinnae pinnatifid; pinnae 3.5-40 2-13 cm., oblong to oblong- or ovate-lanceolate, with an acumen to 2 cm. long, the base rounded, varying cuneate, often inequilateral, glabrous, the margin entire or shallowly and distantly crenate; main veins prominent beneath, intercostals slightly so; 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 139 lateral veins 7—20 pairs, straight; intercostals parallel only towards margin; dark green, drying red-brown to pale greenish or yellowish brown; hypoder- mis of one cell-layer present, interrupted over the areolae, cells elongate in surface view; gland-hairs immersed, heads globose, 4—6-celled; petiolules to 40 mm. long. Inflorescences, male and female heads usually paired in the same leaf- axil. At anthesis: male head 30-75 & 15-20 mm., ellipsoid-oblong, smooth, covered by flowers and bracts; perianths tubular, 1.8 mm. long, shortly bilobed above, minutely pubescent; stamen 2.0-2.3 mm. long, filament slender, cylindric, anther-cells subglobose, 0.2 mm. long; bracts slenderly stalked, heads infundibuliform, to 0.6 mm. across, pubescent; peduncle 50— 65 X 4 mm., surface as in the twigs; female head with pubescent peltate bracts being shed and simple styles exserted to c. 1 mm. Syncarp to 11 * 8 cm. (fide Elmer, l.c., to 20 cm. in length), subglobose, brown, drying pale brown, covered by closely set, cylindric, rigid, shallowly fluted, obtuse, minutely punctate processes, 6-8 % 1-1.5 mm.; wall c. 6 mm. thick: fruiting perianths numerous, proximal free region orange, pulpy, “seeds” (endocarps) ellipsoid, 17 & 10 mm.; core c. 20 mm. across; peduncle 60-130 & 10 mm., surface as in the twigs. VERNACULAR NAMES: Bakil, Sumatra, Borneo; Mentawa or Mentaba, Banka. Borneo; Puan or Pupuan, Borneo. DIsTRIBUTION: in evergreen forest to 4000 ft., Malaya, Sumatra, Banka, Billiton, Lingga, Borneo Malaya. Matacca: Maingay 1483 (x, . 2). SINGAPORE: Garden Jungle, Ridley 4113 (CAL, SING, 3), 6429 (BM, CAL, K, SING, 2), 6678 (BM, CAL, K, SING, a 58 } - MacRitchie Reservoir, s. side, Sinclair SFN 40658 (kK, 2); Seleter Reser- VOIr, s. ‘side. Sinclair SFN 39249 (1, SING). (Fide Corner, 1940, frequent in low- land forest.) Sumatra. INDRAGIRI: Kwala Belilas, bb 27580 (a, Bo, L); Muara Serangge, bb 29991 (A, BO, L). BENKULEN: Lebong, Bosch Pengadang, Olivier 29 (BO, L). PALEMBANG: Banjuasin, Bajunglintjir, NJ/FS E 1028 (Bo, 2), 1163 (Bo, L, u, 2); Batu-radja, Teysmann HB 3698 (Bo, K, L, U); Lematang Ilir, Gunong Megan. NIFS E 952 (Bo, K, L, SING, U, 2), 959 (80, L, P, U, $, 2), 1029 (Bo, U, 2); Lematang Ilir. Semangus, bb 31984 (A, BO, SING), 32042 (Bo, K); Lematang Ulu, Grashoff 277 (Bo); Muara Dua, Grashoff 454 (Bo, L); Rawas, Grashoff 1052 ‘BO, L). LAMponcs: Semangka, Kuta-Agung, Gusdorf 289 (Bo, L). BANKA Lobok Besar, Kostermans & Anta 1230 (a, L) ; Muntok, Batu- balai, Peysmann 7245 (Bo, C, K, L); Sungei Liat, Teysmann 7247 (BO, C, K, L). N: Rossum 48 (Bo, L, 6, 2), Teysmann s.n., (BO); Tandjungpandang, Teysmann s.n. (Bo); Tandjungpandang, Bantan, bb 8677, (Bo, L). Lincca: Teysmann 7282 (BO, L. P). Borneo. SARAWAK. Beccari PB 2997 (1, K, 2); Lundu, Foxworthy 344 (tL). West Borneo. De Vriese s.n., 1857-61 (Bo [labelled Riedel s.n., Menado, Cel- ebes], L); Sanggau, Nek Swale bb 14307 (Bo). SOUTH AND SOUTHEAST BORNEO. Pleihari, Kintap, bb 8145 (xo, 1): Puruktjahu, Muara Djaan, bb 10506 (Bo, L): Tanah Bumbu, Kampong Rare, bb 13053 (Bo, L, 6). EAstT AND NORTHEAST Borneo. Balikpapan: S. Tunan, bb 25588 (Bo, L). Berouw: Domaring, bb 18861 (so). Bulungan: Salim Botu, S. Kumoh, bb 11275 (Bo, L). E. Kutei: Bengalon 140 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XL Rapak, bb 15, Ai 15319 (Bo, L); Peridan, bb 9862 (Bo, ro Sangkulirang, Pala- wan, bb 11859 (Bo, L); Sangkulirang, Ronggang, bb 797 1 (Bo, L . Leban, bb 14685 (Bo, a W. Kutei: Blu-u, Jaheri 1460 (80); Ibo Antislong; bb 16484 (Bo, P); Kahala, bb 28365 (Bo); Kelumpang, bb 16928 (A, Bo, L); Longbleh, bb 16137 (A, BO, L), 16141 (Bo), 16149 (Bo, SING), 16152 (BO, P), 16488 (BO, sING); Mujup, bb 16752 (A, Bo, L). BritisH NortH Borneo. Elphinstone Prov., awao, Elmer 21600 (A, BM, BO, C, GH, K, L, P, SING, U, ?); Kinabatangan Besar, Kori Timber Camp, Cuadra A 2133 (x, SING) ; Mt. Kinabalu, Penibukan ridge, Clemens 40424a, 50210 (pm); Sepilok For. Res., 15 miles w. of Sandakan, Wood SAN 16549 (A, K, L, 3, 2). This species is the only one in Artocarpus to have pinnate adult leaves. It was noted by Corner (1940) that these are unusual both in that they do not disarticulate when falling, as do most compound leaves, and in the arrangement of the pinnae, with longer and shorter pairs alternating and the latter lying in a plane above the former (well shown in the photograph, l.c.). The only other occurrence of pinnate leaves in the genus appears to be in saplings of A. tamaran; the latter can readily be distinguished (apart from the indumentum of the shoot) by the continuous narrow wing of lamina on both sides of the rachis. A sheet at Bogor identified under an unpublished name purports to have been collected by Riedel in Menado, Celebes, but this is presumably an error of labelling, since De Vriese s.n., West Borneo, at Leiden, bears the same determination. Merrill (1921, 1923) recorded this species (as A. superba) from Basilan (Hutchinson FB 3989), Mindanao (Zamboanga, Merrill 8280; Port Banga, Whitford & Hutchinson FB 9164; both sterile) and Palawan (Cenabre et al. FB 28009), but no duplicates of these collections have been traced. 2. Artocarpus lanceifolius Roxb. FI. Ind. 3: 527. 1832, “Janceaefolia;” Wight, Ic. Ind. Or. 2: ¢. 679. 1843; Tréc. Ann. Sci. Nat. Bot. III. 8: 122. 1847; King in Hook. f. Fl. Brit. Ind. 5: 543. 1888; King, Ann. Bot. Gard. Calcutta 2: 13. ¢. 11. 1889; Renner, Bot. Jahrb. 39: 366. 1907; Ridley, Fl. Malay Penin. 3: 354. 1924; Foxworthy, Malayan For. Rec. 3: 128. 2 plates. 1927; Burkill, Dict. 256. 1935; Corner, Wayside Trees, 656. f. 197. 1940. Artocarpus lanceofolia Roxb. Hort. Beng. 103. 1814, nomen nudum. ssp. lanceifolius Evergreen trees, height to 35 m., with small buttresses, bark dark grey, peeling off in flakes. Twigs 6-8 mm. thick, rugose, appressed-puberulent and minutely punctate; annulate stipular ‘scars c. 1 mm. broad, inconspic- uous; lenticels scattered. Stipules 1.5—4.5 cm. long, ovate-lanceolate, acute, exterior rugose, shortly appressed-pubescent. Leaves 10-33 X S-I7 cm., elliptic, varying to ovate, obovate or oblong, rarely narrowly lanceolate- elliptic, short-acuminate, acumen rounded varying acute, base cuneate, decurrent, rarely inequilateral, thickly coriaceous, glabrous, the margin entire or shallowly and distantly crenate; juvenile leaves pinnatifid; main 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 141 veins prominent beneath, reticulum not or scarcely prominent; lateral veins 9-14 pairs, straight; intercostals parallel; deep green, drying pale or reddish brown with straw-coloured reticulum; hypodermis of one cell- layer present, interrupted over the areolae, cells elongate in surface view; glands deeply sunken in narrow pits, heads globose, 4—6-celled; petiole 10-30 mm. long. Inflorescences solitary in leaf-axils or the male ones paired. At anthesis: male head 30-60 X 12-18 mm., ellipsoid to cylindric, smooth, covered by flowers and bracts; perianths tubular, 2 mm. long, shortly bilobed above, puberulent; stamen 3.5 mm. long, filament slender, cylindric, anther-cells oblong, 0.4 mm. long; bracts slenderly stalked, heads infundibuliform, to 0.5 mm. across, sparsely ciliate; peduncle 25-70 X 2.5-5 mm., shortly appressed-pubescent; female head with pubescent peltate bracts mostly shed and bifid styles exserted to 0.5 mm. Syncarp to 8 X 7 em. (to 12 cm. across, fide Foxworthy, 1927), subglobose, olive- or chestnut-brown, drying brown, tesselated from closely set, indurated, cylindric, truncate, appressed-pubescent processes, c. 3.5 3 mm.; wall c. 8 mm. thick; fruit- ing perianths numerous, proximal free region whitish, pulpy (fide Burkill, 1935), “seeds” (endocarps) ellipsoid, 20 % 10 mm.; core Cc. 15 mm. across; peduncle 50-100 & 8 mm., surface as in the twigs, or shortly appressed- pubescent. VERNACULAR NAMES: Keledang or Klidang (Malay), Malaya, Riouw- Lingga Archip. Users: the timber is valuable and the pulp surrounding the seeds is eaten by jungle tribes. DisTRIBUTION: in evergreen forest to 2000 ft., Malaya, Sumatra (West and East Coast), Banka, Riouw-Lingga Archip. Malaya. Perak. Larut: King 3452 (BM, CAL, K, 6), 3826 (CAL, K, ? ); Gunong Boobo Range, King 7631 (CAL, K, P, SING, ?). PAHANG: Kuala Lipis, Phillips 670 (stNG); Temerloh, Hamid CF 5737 (sinc). NEGRI SEMBILAN: Gunong BM, CAL, SING, @). JoHORE: Pulau Tinggi, above Kampong Tandjong Balang, 8); Penang Hill, Nanen 35845 (sinc, 2); Puram (?) Bukit, Curtis s.n., July 1893 (stnc). SINGAPORE: Bukit Timah, Corner s.n., Feb. 1931 (sinc, 2); Bukit Timah For. Res., Jalan Tiup Tiup, Sinclair SFN 40249 (x, L); Gardens Jungle, Ridley 6541 (BM, CAL, K, SING, 6). (Fide Foxworthy, 1927, also in Kedah, Kelantan and Trengganu. ) Sumatra. West Coast: de Vriese & Teysmann 20 (L); Padang, Lubuk peraku, bb 6117 (Bo); Painan, Barung Belantae, NJFS SWK/I-50 (Bo); near Pajakumbuh, Mt. Sago, Meijer 4708 (ccE, ?). East Coast: Simelungun, Masikat, bb 5349 (Bo, L); Sibolangit, Lérzing 5445, (Bo, L, 2). BANKA: Teys- mann 7248 (Bo, C, K, L, P); Muntok, Aer Limau, 6b 7613 (Bo); Sungei-elan, Teysmann 7250 (Bo, C, K, L, P, 2); Sungei-elan, Kampong permis, Teysmann 7251 (Bo. P, 2). Rrouw Arcuip.: Bintang, Teysmann 7283 (BO, CAL, K, L, P, ¢ ). Lixcca ArcuiP. P. Sinkep: Bakong, bb 2045, 13670, Kassim 6 (Bo, L); Bakong, 142 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL Santel, bb 17400 (Bo, sinc) ; Djago, bb 3945 (Bo); Mapur, Marok Tuwa. bb 3846 (Bo); Tadjur Tandjong, bb 8386 (Bo); Tandjung Batang, bb 4014 (Bo, L). Roxburgh’s original description was very brief and no type specimen has been found, so that the identification must rest on his drawing of a leaf and a syncarp which was published by Wight in 1843 and can be recognized with certainty as representing this species. ssp. Clementis (Merrill) Jarrett, stat. nov. gee Xe clementis Merr. Jour. Str. Br. Asiat. Soc. 85: 164. 1922. Holotype, British North Borneo, Clemens 10770 (pNu, not seen, photograph in A); isotypes (A, BO, K). Differs from ssp. lanceifolius as follows: leaves varying to ovate- -elliptic or ovate-lanceolate; syncarp covered by somewhat widely spaced, indurated, slightly tapering, obtuse, appressed-pubescent processes, 1-1.5 < 1 mm., with scattered peltate bracts persistent between them: iio c.3 mm. thick: “seeds” (endocarps) numerous, ellipsoid, 12-15 & 8 mm.; core c. 15 mm. across; peduncle 40 *& 6 mm., minutely punctate. VERNACULAR NAME: Keledang (Malay). DIsTRIBUTION: in evergreen forests to 3600 ft., endemic to northeastern Borneo. Borneo. EAST AND NORTHEAST BorNEO. Balikpapan: Pemaluan. bb 24763 (Bo, L); S. Karnain, bb 26374 (Bo, L, SING). Berouw: Betemaran Tidi. bb 19062 (Bo); Domaring, bb 18831 (A, BO), bb 18835 (A, BO, 1), 18851 (Bo): Inaran, bb 12078 (Bo); Long Lanuk, bb 18445 (Bo), 18461 (Bo, L, SING). Bulungan: Kabiran, S. Bengalun, bb 11686 (Bo); Mara, bb 10750 (po). E. Kutei: Muara Karangan, Remendai, bb 7962 (Bo); Sangkulirang, Palawan, bb 11867 (Bo); Sangkulirang, Pingadan, bb 12992 (Bo); Sangkulirang, Rantau Banan, 00 15236, 15238, 15247 (Bo); Loa Djanan, w. of Samarinda, Kostermans 6627 Tandjong Bangko, nr. Mahakam River estuary, Kostermans 7096 (Bo. K. L, 2). 7, Kutei: Djembajan, Sungei Djambu, bb 24672 (a, Bo, L); Kendisi. bb 16681 (BO, L), 16695 (A, Bo, L); Mo-Antjalong, bb 16482, 16486 (A, Bo, L): near Mt. Kemul, Endert 3683 (L). Tidung: Loban, bb 17882 (po, L). BritisH NortH Borneo. Mt. ae in riled Puasa Si (kK); Mt. Kinabalu, Gurulau Spur, Clemens 10770, Nov. 1915 (A, Bo, K, 2); 12 miles w. of Ranau, Bondu Tahan, Wood & Charrington pe 16381 (A, 9), pais bb 26174 (Bo). The description of the syncarp is based on that of the type of Artocarpus clementis, which is slightly decayed, but almost apes had the processes rather widely spaced at maturity. In the other, younger, syncarps seen from Borneo the processes are also small but they ‘are closely set and ap- parently separate only when the syncarp nears maturity. The three collec- tions cited from British North Borneo have rather distinctive ovate-elliptic to ovate-lanceolate leaves, but they merely represent an extreme of the range of variation which is exhibited by the rest of the material seen from Borneo, nearly all sterile, and which is not separable from that of the species as a whole. In view of the discontinuity in the distribution of the species between a western area including Malaya and Sumatra and a smaller area 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, II 143 to the east in northeastern Borneo, it is likely that only ssp. clementis is represented in the latter. The sterile collections from Borneo are tentatively listed under this subspecies, but only specimens bearing syncarps or showing the extreme leaf-shape can be assigned here with certainty. Series Asperifolii Jarrett, ser. nov. Ramuli et folia hispidi vel hispidi-pubescentes; folia sine hypoderme, glandulis subimmersis, capitibus globosis, 4—16-cellis; inflorescentiae mas- culae globosae ad clavatae. Type species: Artocarpus rigidus Blume. 3. Artocarpus melinoxylus Gagnep. Bull. Soc. Bot. Fr. 73: 87. 1926, “melinoxyla;”’ Gagnep. in Lecomte, Fl. Gén. Indoch. 5: 736. 1928. Syntypes, Indochina, Chevalier 37007, Poilane 1218, 4588, 6873, 7079, 7621 (Pv); lectotype Poilane 7079 (P). ssp. melinoxylus Trees, height to 15 m. Twigs 3-7 mm. thick, rugose, hispid, hairs yellow, appressed (longer and patent on juvenile shoots); annulate stipular scars 0.5 mm. broad, not or scarcely prominent; lenticels scattered. Stipules c. 2.5 cm. long, ovate, acute, hispid-pubescent, hairs yellow, subappressed. Leaves 7-30 & 4-18 cm. (-35 X 25 cm.) elliptic to obovate- or ovate- elliptic, short-acuminate, base rounded, varying cuneate, margin entire; main veins only prominent beneath or intercostals slightly so; glabrous above or with scattered hairs on the main veins, appressed subhispid- pubescent on the main veins beneath; lateral veins 6-13 pairs, straight, intercostals parallel; drying pale to dark red-brown, lighter beneath; hypodermis absent; glands half immersed, heads globose, 6—8-celled; petiole 8—28 mm. long. Inflorescences solitary in leaf-axils, At anthesis: male head (one only seen at maturity) 50 * 45 mm., subglobose, smooth, covered by flowers and bracts; perianths deeply 2(or 3)-lobed, 1 mm. long; stamen 1.5 mm. long, filament slender, cylindric, anther-cells ellipsoid, 0.2 mm. long; bracts stoutly stalked, heads peltate, to 0.4 mm. across, these and perianths sparse- ly ciliate; peduncle 50 3 mm., appressed hispid-pubescent; female head with pubescent peltate bracts mostly shed and simple styles exserted to 1.0 mm. Syncarp to 6 cm. across, globose, drying red-brown, covered by closely set, indurated, cylindric, obtuse, hispid processes, c. 3 > 2 mm., hairs patent and slightly crisped; wall c. 5 mm. thick; fruiting perianths numerous, proximal free region fleshy, “seeds” (pericarps) ellipsoid, 10 X 6 mm.; core 20-25 mm. across; peduncle 70-135 > 6 mm., appressed- hispid. VERNACULAR NAME: Cay mit nai, Annam. Uses: the wood is stated in field notes to be of good quality. DIsTRIBUTION: in forests to 5000 ft., endemic to Indochina. 144 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL Indochina. ANNAM. Near Hué, Nui Bah Ma, Poilane 29977 (p, 2). Near Nhatrang: Massif de C6 Ihn, Poilane 4588, Sept. 1922 (p, 2); Massif de la Mere et |’Enfant, Poilane 6873, May 1923 (p, 2). Quang Tri prov., Mai Lanh, Poilane 1218, Mar. 1920 (xk, Pp, 6). Near Tourane: Clemens 3431 (p, é); Balete, Clemens 4021 (kK, P, 2); Ba na, Poilane 7079, Aug. 1923 (A, P, 2); Mt. Bani [Ba na], Clemens 3510 (a, K, P, 9); Lien Chien, Poilane 7621, Aug. 1923 (A, Pp, 2); Thua-Luu div., Lang-co For. Res., Chevalier 37007, Apr. 1918 (p, 6). Laos. Luang Prabang prov., n.e. of M. Ngai [probably Muong Ngoi], Poilane 20689 (Pp, 4 ssp. brevipedunculatus Jarrett, ssp. nov. Holotype, British North Borneo, Wood SAN A 1733 (A); isotype (kK). Differt ab typo inflorescentiis ad anthesin capitulis masculis subglobosis, c. 15 mm. diametro, pedunculis, 8-17 2.5 mm. suffultis, femineis pedun- culis 13-18 X 3 mm. suffultis [in syncarpio submaturo ad 25 mm. longis]. VERNACULAR NAME: Temponek. DISTRIBUTION: in lowland forest, endemic to British North Borneo. Borneo. BritisH NortH Borneo: Beaufort, Wood SAN A 1733, May 1955 (A, K, 6, 2); Jesselton, Daw Kepong FN 71664 (A, K, 2). This new subspecies agrees closely with the type from Indochina, except in the smaller size of the male head and in the length of the peduncles. The elongation of the latter, throughout the genus, occurs largely before anthe- sis and they can be assumed to be considerably shorter than in ssp. meli- noxylus, even in the mature syncarp. In spite of the difference in the size of the male inflorescence in the two subspecies the dimensions of the flowers are the same; the perianths are unusual for this subgenus in that they are very deeply lobed. Although taxonomically Artocarpus melinoxylus is very near A. chap- lasha, it differs in a number of characters which appear to justify main- taining it as a distinct species. The male inflorescence in A. chaplasha has a longer, more slender peduncle and the flowers are larger in all their parts, while the perianth is only shortly bilobed. The indumentum of the syn- carps is also different, being appressed on the syncarp processes in A. chaplasha instead of patent and slightly crisped. In A. chaplasha the leaves have the entire upper surface and the venation beneath subappressed hispid- pubescent, but in A. melinoxylus the leaves have a smooth upper surface and only the main veins are appressed hispid-pubescent beneath. 4. Artocarpus chaplasha Roxb. Hort. Beng. 66. 1814, nomen nudum, Fl. Ind. 3: 525. 1832; Wight, Ic. Ind. Or. 2: ¢. 682. 1843; Tréc. Ann. Sci. Nat. Bot. III. 8: 112. 1847; Kurz, For. Fl. Burma 2: 432. 1877; King in Hook. f. Fl. Brit. Ind. 5: 543. 1888; King, Ann. Bot. Gard. Calcutta 2: 13. ¢, 12. 1889; Renner, Bot. Jahrb. 39: 365. 1907; Troup, Silvicult. Ind. Trees 3: 880. ¢. 325. 1921; Parkinson, For. Fl. Anda- man Is. 254. 1923; Champion, Indian For. 60: 524. ¢t. 50. 1934. Holo- tvpe, “East India,” Roxburgh s.n. (xk). 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 145 Artocarpus chama Buch.-Ham. ex Wall. Cat. no. 4657C. 1831, ‘nomen nudum. Urostigma chrysophthalmum Miq. Hook. Lond. Jour. Bot. 6: 575. 1847. Ficus chrysophthalma Miq. Ann. Mus. Lugd.-Bat. 3: 285. 1867. Deciduous trees, height to 40 m., bark grey, becoming brown, peeling off in flakes. Twigs 4-10 mm. thick, feos, densely hispid- pubescent: hairs yellow, appressed (patent on juvenile shoots); annulate stipular scars c. 0.5 mm. broad, not or scarcely prominent; lenticels scattered. Stipules 1.5-2.5 cm. long, ovate-lanceolate, acute, hispid-pubescent, hairs yellow, appressed. Leaves 14-23 & 9-14 cm., obovate-oblong, varying obovate or elliptic, acute, base rounded or shallowly cordate, margin entire; juvenile leaves pinnatifid; main veins prominent beneath, intercostals slightly so; very shortly subappressed-hispid above, venation beneath shortly subap- pressed hispid-pubescent, both surfaces varying subglabrous; lateral veins 8-12 pairs, straight; intercostals parallel; green, drying pale or reddish brown, lighter beneath; hypodermis absent; glands half immersed, heads depressed-globose, c. 8-celled; petiole 8-30 (—40) mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 15-30 X 10-30 mm., ellipsoid to short-obovoid or subglobose, smooth, covered by flowers and bracts; perianths tubular, 2 mm. long, bilobed above, minutely pubescent; stamen 2.5 mm. long, filament slender, cylindric, anther-cells ellipsoid, 0.3 mm. long; bracts slenderly stalked, the heads peltate, to 0.5 mm. across, pubescent; peduncle 60-75 1.5 mm., shortly appressed hispid-pubescent; female head with pubescent peltate bracts mostly shed and simple styles exserted to 0.5 mm. Syncarp to 6 cm. [to 10 cm., fide King, 1889] across, subglobose, yellow, drying brown, covered by closely set, indurated, cylindric, obtuse, appressed-hispid processes, c. 1.5 X 2 mm.; wall c. 5 mm. thick; fruiting perianths numerous, proximal free region . ‘seeds” (pericarps) ellipsoid, 7 X 4 mm. [fide King, 1889, to 20 mm. long|; core c. 20 mm. across; peduncle 55-80 3 mm., appressed hispid- pubescent. VERNACULAR NAMES: Chaplash (Bengali), India; * Taung-peing, Burma. Uses: an important timber tree and often planted, although Champion (1934) stated that elephants found the seedlings very palatable; Parkinson (1923) noted that the leaves were used for elephant fodder and the fruits were edible. DisTRIBUTION: in evergreen, semi-evergreen and moist deciduous forest to 5000 ft., in regions with a monsoon climate (rainfall at least 80 inches), northeastern India (Sikkim to Assam and Chittagong), Lower Burma (to northern Tenasserim), Andaman and Nicobar Is. India. Not localized: East India, Roxburgh sn. (K); Dulka Thae, Gamble L Terai, Tin sing tung, sine nom. et num. (K). WEST AND aver BENGAL: Duars, ‘Throughout this series of papers the word India will be used in a geographical sense, to denote the Indian subcontinent, and will include the modern political subdivi- sions of India and Pakistan. 146 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XL Talpaigwa, Haines 529 (kK); towards Ilafoo, Griffith 4668 (kK); Tipperah (Hill), near Ranir Bazar, Noagaon, Debbarman 1096 (cat, 2). Assam: Lister s.n, (cAL), Mann s.n. (cat, 6), Masters s.n. (80, L, P); Abor, Pilung, Burkill 38127 (caL); Gowhatly, Clarke 371654 (pm); [? =] Guwalatty, near Aber [? Abor], Jenkins s.n., Apr. 1835 (caL); Cachar, Bundookmora, Keenan s.n. (K); Jatookia, Watt 11308 (pm, 2); Kamrup, sine nom. 942 (caL, 6); Khasia Hills. Hooker f. & Thomson s.n, (A, BM, C, by p, U); Khasia Hills, Doodputli, Hooker f. & Thomson s. n., Nov. 1856 (K); Mangeldai to foot of Bhutan Himalaya, Schlagint- weit 13480 (BM); Naga Hills, Vad 214 (L); Nowgong, Hooker f. & Thom- SON S.N., Tuly 1850 (kK); Sibsagar, Ligri Pukri, Watt 10405 (cat, 2). Cuitta- GONG: Kagi Ke Hath, Hooker f. & Thomson s.n., Jan. 1851 (K); Kodala. aye 26008 (cAL), King 392, 487 (BM, CAL, 2); Rancamet. Lister s.n, (CAL); gamati, Magban. Hooper s.n., Nov. 1898 (CAL); Rangamati and Damara, I} “allich 4657C (BM); Seetakoone, Hooker f. & Thomson s.n., Jan. 1851 (xk). 3a 4 CI ' 7b o oY ) 0 "0 \ : Ta 5 G7 ee i) | vi A : St YQ ~ ie) 8 ee 28 ) ° s Y Qe Pog e 6 oe Distribution of the species of sect. Duricarpus. 1, Artocarpus osha _ en in the Philippines); 2, A. lanceifolius, a, ssp. lanceifolius, b, s A, eS a, ssp. ‘melinoxylus, b, ssp. eaten re "4 we Wir as A. odoratissimus (probably introduced in a enclosed by broken ines: 6, A. hispidus; 7, A. rigidus, a, ssp. as 7 He asperulus. 1959} JARRETT, ARTOCARPUS AND ALLIED GENERA, III 147 Lower Burma. (?) Abel s.n. [“East Indies” | (K, 6); Pegu, Tonkyeghat, Kurz 1498 (CAL pp. L, 9), 1499 (caL); Rangoon, Dickason 5528, 6948 (a), McClel- land sn. (K, 6). TENASSERIM. Kyauktalon [not located], Meebold 14612 (CAL, 4). Thaton: Martaban, Kurz 1498 (cax p.p., @). Andaman Islands. Aurz s.n., Nov. 1881 (CAL). Parkinson 167 (pp, 6), Rogers s.n.. 1904 (kK, ¢, 2); Manglu- tan. Parkinson 463 (K, 6 9%); South Andaman, King s.n., Apr. 1890 (caL), Kurz sn. (Kk. 2); South Andaman, n. coast, Aurs 5.1. (x, 2); S. Andaman, Corbyn's Cove. Kurz s.n. (CAL). Nicobar Islands. Jelinek 165 (CAL, 38); Karnicobar. Kakena, Kamphovener 2672 (c); Karnicobar and Chowry, Kamp- hovener 2394 (c); Teressa, Kamphovener 2741, 2742 (Cc). Cultivated. Ixpra: Calcutta, Hort. Bot., Lane 7935 (cat. 6), Wallich 4657 (xk, 6). 4657D (BM, CGE, P, 3), s.m., Apr. 1815 (c); Dehra Dun, sayy Research Institute. Raizada s.n., May 1947 (pp, 6, 2), 5.2, June 1950 (pp, @). Artocarpus chaplasha is somewhat similar vegetatively to A. rigidus ssp. asperulus (in addition to A, melinoxylus, the distinguishing characters of which have been given above), but it may be identified by the larger, typi- cally obovate-oblong leaf, since in Tenasserim where the two entities may overlap. the leaf shape of A. rigidus ssp. asperulus tends to ovate. ace however. of interest to note that the distributional areas of these three entities. and hence, presumably, their ecological requirements are comple- mentary. .1. chaplasha is restricted to areas with a moderate to well- marked monsoon climate (but a rainfall of at least 80 inches) and is recorded definitely as far south as the Thaton district of Tenasserim. A. rigidus ssp. asperulus is recorded only as far north as the adjacent Amherst district (there is one unlocalized collection by Meebold of each species) and occurs, primarily in coastal areas, in the moister peninsular regions of Burma and Siam, extending to Cambodia and southern Annam:; it is replaced by 4. melinoxylus in the mountains of Annam and northeastern Laos, which have a rather more uniform, though not more abundant, rainfall. 5. Artocarpus odoratissimus Blanco, FI. Filip. 671. 1837, “Arctocarpus odoratissima.” Fern.-Villar, Noviss. App. 203. 1880; Elmer, Leafl. Philip. Bot. 2. 618. 1909; Wester, Philip. Agr. Rev. 8: 108. ¢. 7d, 9b. 1915. 17: 24. 1924(a), Bull. Bur. Agr. Philip. 39: 129, t. 156, 326. 1924(b): Merr. Jour. Str. Br. Asiat. Soc. 76: 80. 1917, Sp. Blan- coanae. 124. 1918, Enum. Philip. Pl. 2: 42. 1923. Neotype, Mindoro, Merrill SB 1019 (BM). Artocarpus tarap Becc. For. Borneo, 626. 1902; Renner. Bot. Jahrb. 39: 365. 1907. Syntypes. Sarawak, sae a PB 9, PB 2528. PB 2588, PB 2697 (¥1); lectotype. Beccari PB 2697 Artocarpus mutabilis Becc. ee pes 627. 1902; Renner. Bot. Jahrb. 29: 365. 1907. Holotype, Sarawak, Beccari PB 758 (¥1); isotypes (FI, K, P). Evergreen trees. height to 25 m. Twigs 4-10 mm. thick, rugose, hispid- pubescent. hairs yellow to rufous, patent; annulate stipular scars c. 0.5 mm. wide. not raised: lenticels scattered. Stipules 1-8 cm. long, ovate, acute, hispid-pubescent, hairs yellow to rufous, appressed or patent. Leaves 16-50 & 11-28 cm., broadly elliptic to obovate, obtuse to shortly acumi- 148 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL nate, base cuneate, often slightly decurrent, margin entire or distantly and shallowly crenate, often bearing tufts of hairs towards and at apex, varying lobed in upper half, lobes one (rarely two) on each side; juvenile leaves pinnatifid; main veins and intercostals prominent beneath, reticulum slightly so; both surfaces subappressed hispid-pubescent, scabrescent above, with the main veins softly and densely yellowish pubescent; lateral veins 13-15 pairs, straight; intercostals parallel; dark green, drying reddish to blueish brown above, red-brown beneath; hypodermis absent; glands slightly sunken, heads globose (8—)16-celled, cells in two tiers; petiole, 20-30 mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 40-90 25-35 mm., ellipsoid to clavate, smooth, covered by flowers and bracts; perianths tubular, 1.8 mm. long, shortly bilobed above, minutely pubescent; men 2.0 mm. long, filament slender, cylindric, anther-cells subglobose, 0.15 mm. long; bracts slenderly stalked, heads peltate, to 0.4 mm. across, pubescent; peduncle 25-70 x 3-5 mm., indumentum as shoot; female head with pubescent peltate bracts mostly shed and simple styles exserted toc. 1.5 mm. Syncarp to 16 X 13 cm., subglobose, greenish yellow, drying pale red-brown, covered by closely set, rigid, cylindric, hispid processes, 8-13 & 1 mm., fluted below, the tips clavate; wall c. 8 mm. thick: fruiting perianths numerous, proximal free region white, juicy (fide Wester, 1915), “seeds” (pericarps) ellipsoid, 12 8 mm.; core c. 15 mm. across; peduncle 55-140 & 8 mm., indumentum as shoot. VERNACULAR NAMES: Oloi or Loloi, Mindoro; Marang or Madang, Min- danao, Basilan, Sulu; Terap (Malay), Borneo. Uses: the syncarp is esteemed, particularly in the Philippines, for the sweet, juicy, aromatic perianths surrounding the seeds, which may also be roasted; according to Wester (1915) the flavour is better than that of the Jack. DIsTRIBUTION: in evergreen forest to 3000 ft., Borneo; in regions with abundant and equally distributed rainfall (? introduced), Philippines (Mindoro, Mindanao, Basilan, Sulu Archip.). Borneo. Korthals s.n, (L, U). SARAWAK. Haviland 3116C (cat, é, 2); Upper Rejang River, Gat, Clemens s.n., July 1922 (ny); Gunong Mattan, Beccari PB 2528, Nov. 1866 (FI, K), 2588, Nov. 1866 (FI, K, P, ¢, 2); Kuching, Beccari PB 2697, Nov. 1866 (FI, K, 2), Haviland 2190 (cat, kK, SING, &, 2 ), Haviland & Hose 3316 (kK, 8, 2), 3316B (BM, L, 3, 2); base of Santubong, Sinclair SFN 38372 (k, sinc, 9); Siul, nr. Kuching, Beccari PB 758, Oct. 1865 (FI, K P, 6, 2). West Borneo. De Vriese sn. (K); Landak, s. nom. et num. (BO) ; Liang gagang, Hallier 2577 (Bo); Melawie, Tjatil B. Tengkujung, bb 26328 (BO, L). SOUTH AND SOUTHEAST BorNEO, Muara Teweh, Lué Katjang, bb 10926 (so); Puruktjahu, Muara Djaan, bb 10514 (Bo). EAST AND NORTHEAST BoRNEO. Berouw: Dedemuaer, bb 19103 (A, Bo, L); Tandjong-redeb, Labanan, bb 11520 (Bo). Tidung: bb 17732 (Bo); G. Muara Tagel, Amdjah 136 (Bo, L, 9). W. Kutei: Blu-u, Jaheri 1488 (Bo). BritisH NortH BorNEo. Burbridge s.n., 1877- 78 (BM, K); Kabili-Sepilok For. Res., Puasa Herb. For. Dept. B.N.B. 7037 (stnc); Mt. Kinabalu, Dallas, Clemens 26228 (A, BM, BO, L, NY, SING, 2); Tenom, Pangie mile 80, Cuadra A 3282 (1, sinc). SreatiK. St. Lucia, Cuadra 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 149 A 2403 (kK, L). NuNuKAN. bb 26196 (Bo, L). TarAKAN. N. E. Kalimantan, Passanggrahan, Meijer 1851 (K, L, 2). Philippine Islands. M1nporo. Calapan, Merrill 2583 (us, 2), SB 1019, May 1916 (A, BM, BO, GH, K, L, P, 2); Pinamalayan, Ramos BS 41090 (a, 2); Upper Sinariri River (“near a Conklin PNH 17522 (A, L, PNH, 2). MINDANAO. 2). BASILAN. Hallier 4504 (L), Merrill SB 1018 (a, BM, BO, GH, K, L, P), Miranda FB 17880 (x, P, 6), Reillo FB 15416 (1), Wester BS 38345 (a); San Rafael, Moro, Hutchinson s.n., May 1906 (Ny). SuLu ArcuiP. Jolo, Vidal 3846 (kK). Cultivated. Borneo. Sunes Beccari PB 9, May 1865 (FI, K, P, 3); Sandakan, Villamil Herb. For. Dept. B.N.B. 200 (1). Mauritius. Chapelier S.N., Commerson S.n. (P), Apart from the very characteristic inflorescences, this species can be recognized by the length of the indumentum on both surfaces of the leaves. In Artocarpus elasticus, which may be confused with A. odoratissimus when sterile, the leaves are also markedly scabrid above, but from very short appressed hairs. Although this species was described from and is better known in the Philippines, its distribution there is limited and it was probably introduced to the islands from Borneo, where it was described under the names A. mutabilis and A. tarap by Beccari. The second of these was reduced to A. odoratissimus by Merrill (1917) and later he stated (1924) that the species apparently was always planted in the Philippines, which is con- firmed by such field notes as are available (quoted above). Wester (1924a) also noted a tradition of the Moros in the southwestern Philippines that the Marang was introduced by them from the wes 6. Artocarpus hispidus Jarrett, sp. nov. Holotype, Singapore, Corner SFN 37035 (SING); isotype (Kk). Differt ab A. rigido capitulis masculis longiore pedunculatis, ramulis juvenilibus pedunculisque dense hispidi-pubescentibus, pilis patentibus. Arbores [ad 20 m. altae]. Ramuli juniores 4-7 mm. crassi, plus minusve rugosi, dense hispidi-pubescentes, pilis rufis patentibus; cicatrices stipu- larum annulatae, 0.5 mm. latae, non prominentes, nec conspicuae; lenti- cellae rarae. Stipulae 0.5—1[- 2] cm. longae, ovatae, acutae, dense sub- appresseque hispidae, pilis rufis. Folia 15-19 « 7-8 [12-23 x 5-10] cm., obovati-elliptica [vel elliptica], acuta [vel breviter acuminata] basi cuneata, margine versus apicem denticulata [vel integra], supra scabrida, pilis brevissimis appressis praefractisve, basibus inflatis tuberculatis, costa nervis lateralibusque dense appresseque pubescentibus, subtus _hispidi- pubescentia, saturata virida, in sicco rubri-brunnea, vel supra canescentia; folia juvenilia pinnatifida; costa nervi lateralesque subtus prominentes; nervi transversales venulaeque subtus prominuli; nervi laterales utrinque [11-]14-15, recti; nervi transversales paralleli; hypodermis absens; glandulae subimmersae, capitibus globosis, 4-8-cellis; petiolus 10-15 [—20] mm. longus. 150 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XL Inflorescentiae axillis foliorum solitariae [vel geminae]. Ad anthesin: capitula mascula [15- ]30 [15-]20 mm., obovoidea, plana. floribus bracteisque numerosissimis obtecta; perianthia tubulosa, 0.6 mm. longa, supra bilobata, minute pubescentia; stamina 0.7 mm. longa. filamentis tenuis, cylindricis, cellis antherum globosis, 0.1 mm. longis: bracteae tenuiter stipitatae, capitibus peltatis, ad 0.3 mm. latis, pubescentibus; pedunculus 25[12-30] x 2 mm., indumento ut ramulis; capitula feminea bracteis peltatis pubescentibus subdeciduis, stvlis simplicibus 1.5 mm longis exsertis. Syncarpia ad 5.5 cm. diametro, globosa, in sicco fulvi-brun- nea, echinata, processibus pny rigidis, teretibus, obtusis. hispidis, 5—6 mm.: stratum externum syncarpii c. 2 mm. crassum; “‘semina” (peri- carpia indurata) numerosa, ellipsoidea, 13 X 10 mm., perianthiis liberis carnosisque inclusa; axis syncarpii c. 15 mm. diametro; pedunculus 25-35 x 5mm., indumento ut ramulis, (Inflorescentiae typi spiritu vini conservae descriptae. ) DISTRIBUTION: in evergreen forest to 1000 ft., endemic to Malaya. Malayas PERAK: Goping. Scortechini 1979 (BM, CAL, K, SING, 6, 2). TRENG- GANU: Kg. Binjai, on road to K. Brang, Corner s.n., Apr. 1937 (SING). SELANGOR: Siinzel Buloh For. Res.. Ja’amat & Jackson SFN 39416 (a, L. 6). MALAcca: Cantley sn. (sinc, 6). PENANG: Government Hill, Curtis s.n.. Mar. 1894 (sinc). SINGAPORE: Bukit Timah, Corner SFN 37035, June 1940 (k, SING, ¢, 2). Corner s.n., Mar. 1941 (stnc, 6); Changi, Ridley 3357 (BM, CAL. SING, 2); 11% miles Mandai road, Sinclair SFN 40051 (1); Reservoir jungle. Thompson Reach end, Corner s.n., Jan. 1937 (sInc); Sangli, Ridley 6756 (stxc. 2). Cul- tivated. SINGAPORE: Hort. Bot.. Ahmad s.n., Mar. 1926 (SING, ¢ ). The syncarps of Artocarpus hispidus and A. rigidus are very similar but the two species are quite distinct in the length of the male peduncles and the indumentum of the shoot and peduncles, and the differences were noted by Corner on the type. Artocarpus hispidus also differs from A. rigidus ssp. rigidus in the scabrid upper surface of the leaf, and from ssp. asperulus in the leaf shape. King’s description of 4. rigidus, under which he cited Scortechini 1979, also included this species. ~I Artocarpus rigidus Blume, Bijdr. 482. 1825, “rigida:” Trec. Ann. Sci. Nat. Bot. III. 8: 114. 1847; Mig. in Zoll. Syst. Verz. Ind. Archip. 2: 89,95. 1854; Miq. Fl. Ind. Bat. 1(2): 286. 1859, Suppl. 418. 1861, Ann. Mus. Lugd.-Bat. 3: 211. 1867, Kurz, For. Fl. Burma 2: 431. 1877; King in Hook. f. Fl. Brit. Ind. 5: 540. 1888; King. Ann. Bot. Gard, Calcutta 2: 8. ¢. 3. 1889; Koord. & Val. Bijdr. Boomsoort. Java 11: 17. 1906; Renner, Bot. Jahrb. 39: 366. 1907; Ridley. Fl. Malay Penin. 3: 352. 1924: Corner, Wayside Trees, 657. ¢. 198, 199. 1940: Backer, Beknopte Fl. hegeee 6: 14. 1948. Holotype, Java. Blume 1364 (L); isotype (CAL) Artocarpus runcinata Reinw. ex Blume, Cat. Bog. 101. 1823, nomen nudum. Artocarpus echinata Roxb. Hort. Beng. 66. 1814, nomen nudum. Fl. Ind. 3: 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III ileal 527. 1832; Wight, Ic. Ind. Or. 2: ¢. 680. 1843; Tréc. Ann. Sci. Nat. Bot. III. 8: 113. 1847. Holotype, Roxburgh s.n. (BM). Artocarpus Le sala Griffith, Not. Pl. Asiat. 4: 400. 1854. Holotype, Malac- ca, Griffith 4664 Artocarpus kertau Zoll. ex Miq. in Zoll. Syst. Verz. Ind. Archip. 2: 89, 95. 1854; Miq. Fl. Ind. Bat. 1(2): 287. 1859; Koord. Exkursionsfl. Java 2: 95. 1912. Holotype, Java, Zollinger 1009 (Pp); isotypes (Let U), Artocarpus A sere Miq. FI. Ind. Bat. Suppl. 417. 1861, Ann. Mus. 67. Holotype, Sumatra, Laas HB 3369 (U); isotype (Bo). Artocarpus varians Miq. ll.cc. Holotype, Sumatra, Teysmann HB 4358 (wv); 0, L) isotypes eas muricata Hunter ex Ridley, Jour. Str. Br. Asiat. Soc. 53: 114. 1909, pro syn. ssp. rigidus Evergreen trees, height to 35 m., buttressed, bark grey, peeling off in flakes. Twigs 2-6 mm. thick, smooth or rugose, densely to sparsely hispid, hairs rufous, appressed; annulate stipular scars c. 0.5 mm. broad, incon- spicuous; lenticels few, scattered, or none. Stipules 0.5—3 cm. long, ovate- lanceolate, acute, hispid-pubescent, hairs vellow or rufous, appressed. Leaves 9-32 & 5-15 cm., elliptic to ovate-, obovate- or oblong-elliptic, apex acute or short-acuminate, varying obtuse, base cuneate, varying narrowly rounded, margin entire or distantly and shallowly crenate; ju- venile leaves pinnatifid: main veins and intercostals prominent beneath, re- ticulum slightly so; glabrous and smooth above, or rarely slightly scabrid from sparse, very short, appressed hairs, except the appressed-pubescent main veins, venation appressed-hispid beneath; lateral veins (9—)12—20 pairs, straight; intercostals parallel; dark green, drying blue-grey to red- brown above, red-brown beneath; hypodermis absent; glands slightly sunken. heads globose, 8-celled, cells in two tiers: petiole 10-25 mm. long. Inflorescences solitary in leaf-axils. At anthesis: male head 13-20 mm. across. short-obovoid to globose, smooth, covered by flowers and bracts; perianths tubular, 1.2 mm. long, shortly bilobed above, minutely pubescent: stamen 1.3 mm. long, filament slender, cylindric, anther-cells globose, 0.1 mm. long; bracts slenderly stalked, heads peltate, to 0.3 mm. across, shortly ciliate: peduncle 2-6 < 2.5 mm., densely appressed hispid; female head with pubescent peltate bracts being shed, and simple stvles exserted to c. 5mm. Syacarp to 7 cm. across (to 13 cm. fide Corner. 1940) globose, dull orange. drying pale brown, echinate from closely set, rigid. tapering, fluted, acute. hispidulous processes, 7-9 & 1.5 mm.; wall c. 10 mm. thick; fruiting perianths numerous, proximal free region orange, fleshy. “seeds’’ (pericarps) ellipsoid. 12 & 7 mm.; core c. 20 mm. across; peduncle 8—25(—40) « 8 mm.. appressed hispid. VERNACULAR NAMES: Perian, Malaya, Purian or Surian, Sumatra, Pujan, Borneo (Malay); Tempunai or Tempunih, Malaya. Tempunit, Sumatra (Malay): Pussar, Java (Sundanese). Uses: the tree is cultivated for its 152 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL fruit in Malaya and Java, the edible portion being the sweet, pulpy, waxy perianths surrounding the seeds; it also provides timber DISTRIBUTION: in evergreen forest to 1$00/-3000) ft., noted several times as occurring near streams, Tenasserim, Malaya, Sumatra, Simalur, Banka, Billiton, Riouw-Lingga Archip., Borneo, Java (except the eastern province). Lower Burma. TENASSERIM: Helfer 4669 (cAL, K, 6). Malaya. KEDAH: Bukit Rombang, Dmat Kepong FN 27377 (x, 6); Cherok Perah, Meh CF a (SING) ; Katumbah, Meh CF 17882 (stnc). Perak: King 6921 (CAL, K, ¢, Batang Padang district, King 7755 (CAL, K, L, P, 2); Bikum Sungei, eee 376 (pm); Larut, King 6727 (cal, K, 2); Larut, Turu, King 6751 (CAL, L, P, é); Larut, Turu, Gunong Bubu range, King 7612 (caL, SING, 2), 7679 (cat, K, 3, 2); Matang Jambu, Wray gs — SING, @); Teluk Anson, Allen 37244 (A, BO, K, SING). TRENGGANU: Ulu g, Moysey & Kiah 33745 (sinc). Pa- HANG: Gat, near Raub, Burkill & ets 16929 (sinc); Kuala Lipis, Nong CF 1258 (sinc): Temerloh, Awang CF 2398 (sinc), Kassim CF 0731 (snc, é). SELANGOR: Bukit Badat res., Rawang, Bahsin CF 32403 (stnc); Kuala Lumpur, ‘Hose CF 4599 (kK, SING, 2); Sungei Buloh For. Res., Hashim 37 (x, 6); We Hills Res., Hamid CF 4575 (kK, 2). NEGRI SEMBILAN: Franck 1197 (c). MALac- ca: Alvins 590, 1070, 1234, 1317 (stnNc), 1624 (sinc, 2), Maingay 1474 (cat, K, L, ¢, 2), 1475 (K), 1476 (CAL, GH, K, L, , 2); Batu Tiga, Holmberg 766 (stnc); between Roombiya and Aloor Gafah, Griffith 4664, Sept. 1842 (xk); 14-14% miles Sungei Udan For. Res., Sinclair SFN 40597 (k, SING). JOHORE: Gunong Pantai, Corner s.n., Jan. 1937 (stnc); Sungei Pelepal Kiri, Corner s.n., June 1937 (sinc). PENANG: Government Hill, Curtis s.n., Mar. 1894 (sING, 2); Penaea Bukit, Curtis 1984 (sinc); Penang Hill, Nanen s.n., June 1940 (SING); ile King s.n., Aug. 1879 (caL); Waterfall Quarry, Curtis 3603 (CAL, K, SrncaporeE: Changi, Ridley 3357 (K), 4437 (CAL, SING, 6, 2); Garden Tangle, Ridley 6542 (BM, CAL, K, 6, ; . De Vriese 23 (L), De Vriese & Teysmann 18 (L). TAPANULI: Barus, namie bb 31402 (A, BO, L); Padang Lawas, Gunong Tua Djulu, bb 6443 (po). West Coast: Ayer Waringun, Burck s.n., Aug. 1883 (Bo, L); Priaman, L, SING, 2); Karolanden, Lao Pengulu, bb 12503 (so); Langkat, Sungei Sedapan, bb 9132 (Bo); Tasik, Koorders 10456 (Bo). Inpractri: Keritang, bb 28660 (Bo, L, SING); Kuantan, Djake, bb 26487, 26491 (Bo, L); Kuantan, Sungei Besar, bb 24028 (Bo, L); Peranap, bb 30102, 30120 (aA, Bo, L). DJAMBI: Muara Pidjuan, bb 12273 (Bo, L, 6). BENKULEN: Seis Penandjung pandang, bb 2735 (xo, L). PALEMBANG: Praetorius s.n. (L); Banjuasin and Kubustreken, Grashoff 809 (Bo, L); re presee as Kubustreken, Bajunglintjir, NJFS T 27 (Bo, L, 6, @), 195 (BO, K, L, SIN 2); Komering Ulu, Grashoff 575 (Bo); Lematang ore Gunong Magang, me T 284 (BO, L, P, SING, U, 6, 2), 300 (Bo, L, 2), 5 (po, 6); Lematang Ilir, Semangus, bb 32245 (Bo, L); Muara Dua, Grashof a (Bo, L); Muara hee Kisau, bb 9630 (Bo); Muara Mengkulem, Forbes 3041 (A, BM, CAL, L, P, SING, 2); Musi Ilir, Ipil, N7FS T 1071 (Bo, L); Rawas, Grashoff 1032 (BO, L): Sete Djangkar, bb 8106 (Bo). LAmMponcs: Mangala, Gusdorf 47 (Bo, 4, rie Teysmann HB 4358 (Bo, L, uU), 4369 (Pp), 4393 (BO, L, vu), 4419 (Bo, U); Seputik, Suwikis, bb 2844 (Bo 1): Seputik, Tulangbawang, Gunong Sugit, Hae 140 (Bo, 2). SIMALUR: Achmad 814 (Bo); Landschap 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 153 Tapah, Defajan, Achmad 1429 (Bo, K, L, U). BANKA: Teysmann HB 3296 (Bo), 6844 (BO, K, L, P); Blinju, Grashof 14 (Bo, L, 6, 2); Blinju, foot of Gunong Rengkuk, Berkhout 149a (Bo); Djebus, Teysmann HB 3369 (Bo, U), 7243 (L), 7252 (c, L, P); Lobok-besar, Kostermans & Anta 824 (A, K, 6); Muntok, Majang, bb 7596 (Bo); Toboali, Teysmann HB 7265 (Bo, P); Tohrin, sine nom. HB 296 (vu). BILLiTon: Rossum 20 (Bo, K, L, ¢ ), 76 (Bo), Teysmann HB 17580, 17583 (Bo); Tandjong Pandjang, By 6778 (Bo). Rrouw Arcuip.: Tandjong Pinang, Bintan, Teysmann HB 7284 (Bo, c, L, P). Lincca ArcHip. NIFS Ri./I-128 (Bo). P. Singkep: ce Tuwa, near Sungei ee Amat 12 (1); Ulu Sungei Marok Tuwa, Amat 34 (Bo, L). Borneo. SARAWAK. Beicon PB 2478 (xk, P, 6), 2998 (K, P, 2); near Kuching, Haviland 1888 (cat, K, 9); Mt. Poi, Clemens 20305 (k, NY); Sibu Sungei Assan, For, Dept. Sarawak S 0502 (SAR). WEST BoRNEO. Melawie, Tjatit B. Tangkujung, bb 26439 (Bo, L); Sadakan, Pait, bb 8048 (Bo, L). SOUTH AND SOUTHEAST BorNEO. Martapura, Djungur, bb 10384 (Bo); ean Sungei Sangga, bb 9950 (Bo); Tanah Bumbu, Kampong Baru, bb 13356 (Bo, L, ¢, EAST AND NORTHEAST BorNEO. E. Kutei: Sangkulirang, Ranggang, bb 7968 (Bo); Tandjong oe region, mouth of Mahakam River, Kostermans 7186 iP : Djembajan, Sungei Kelasan, bb 25122 (Bo, L); Djemba- jan, Sungei Glen. ‘bb 12766 (Bo); Tandjong Tsue, Endert 1953 (A, K, L). BritisH NortH Borneo. Elphinstone prov., Tawao, Elmer 21514 (A, BM, BO, C, GH, K, L, P, SING, U, 2). P. Laut. Sungei Paring, bb 12897 (Bo, uv, 3, 2), 13258 (BO). Java. Blume 1364 (CAL, L, 2), Reinwardt s.n. (L),-De Vriese 1655 = WEST Java. Batavia: Depok, Beumée 6021, Hallier s.m., Aug. 1896 (Bo), Koorders 31077 (Bo, L), 41048 (Bo, 2), 42776 (BO), 42792 (Bo, 4), rer: '(Bo), Van Steenis 12750 (L); between Kota Bambu and Djembatan duren, Backer s.n., 1902 (Bo, 6); Leeuwiliang, Pasir Angsana, Bakh. van den Brink 6974 (xo, kK, L, 6, 2); Leeuwiliang, Pasir Tjihideung, Bakh. van den Brink 6386 (Bo, L); Leeuwiliang, Tjibata, Bakh. van den Brink 6796 (Bo, L, 9); Tjiampea, Koorders 30364, 30365 (B0), 30366 (A, BO, L, 6). Buitenzorg: Handjere, Janglappa, N/F'S Ja 6206 (L). Cheribon: Kuningan, Houter 67, 138 (Bo). Preanger: Palabuan- ratu, Koorders 8738 (Bo, L), 12561, 12562 (Bo, L, 6), 12570 (A, BO, L, P, 2, 2), 15677 (Bo, L, 6), 33049 (Bo); Palabuanratu, Buniwangi, sine nom. et num., Mar. 1873 (Bo); Palabuanratu, Tjibareno, Winckel 1858 (Bo, L, u, 2); Sang- grawa, Koorders 8739 (Bo). CENTRAL JAVA. Pekalongan: Loutresten, E. Tegal, Beumée 1889 (Bo); Pemalong, Bruscomps 8 (Bo). NusA KAMBANGAN: Tijilatjap. Koorders 27032 (Bo, 2). Lesser Sunda Islands. Bali: [? cult.] sine nom. et num, (L). Cultivated. INpra: Pras Hort. Bot., Wallich 2142 (P), eae (BM, CGE, K), 5.m., 1819 (BM, 9). JAvA: Bantam, Zollinger 1009 (A, L, P. U); Bogor, Hort. Bot., Fe ‘2982 (BM, P, U, 2). Unlocalized ee Abel s.n. (East Indies) (k, 2); Kurz ooee (caL); Roxburgh s.n. (BM). The synonyms given above all refer exclusively to Artocarpus rigidus ssp. rigidus, as do the references, except for King’s descriptions (1888, 1889) which include A. hispidus. Merrill (Jour. Arnold Arb. 19: 331. 1938) reduced A. rigidus to A. rotunda (Houtt.) Panzer, Pflanzensyst. 10: 380. 1783, based on Rademachia rotunda Houtt. Nat. Hist. II. Pl. 11: 455. 1779, which was very briefly described from Javan material as having leaves of the same shape as Artocarpus integer but without “roughness” 154 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL except on the fruit, which was round and grew, so Houttuyn was told, to the size of a child’s head. This could be a crude description of A. rigidus but, in fact the leaves in this species are more or less scabrid beneath, the twigs and peduncles are appressed-hispid and the syncarp is smaller. Since no type has been found at Leiden, the identification would have to be based primarily on the vernacular name Mandelique given by Houttuyn, which is cited by Teysmann and Binnendijk (Cat. Bog. 85. 1866) and by Koorders and Valeton (1906) as a Javan name for A. rigidus. The de- scription does not fit any other species occurring in Java and it seems preferable to treat 4. rotunda as a nomen dubium. The specimen listed above from Bali is probably from a cultivated tree, since there are no other records of A. rigidus from eastern Java or the Lesser Sunda Islands as an indigenous plant. One of the collections from Penang, Curtis s.n., May 1894, has the syncarp processes softly pubescent, but otherwise agrees with A. rigidus ssp. rigidus. ssp. asperulus (Gagnep.) Jarrett, stat. nov. Artocarpus calophylla Kurz, Prelim. Rep. For. Pegu App. A, 124, App. B, 82. lavis, For. Fl. Burma 2: 431. 1877, non Teysm. & Binnend., 1866, quae est nomen nudum; King in Hook: f. Fl. Brit. Ind. 5: 540. 1888: King, Ann. Bot. Gard. Calcutta 2: 8. ¢. 2. 1889. Holotype, Burma, Kurz 5.1. cent Pyare asperula ees Bull. Bot. Soc. Fr. 73: 86. 1926; Gagnep. ecomte, Gén. Indoch. 5: 734. fig. 90. 1928. Syntypes, Indochina, Chevalier 30083, Herb. For. “Cambodge 36930, Pierre 15, 1851, Poilane 6644, Thorel 1067 (»); lectotype, Poilane 6644 (p). Artocar pus ie, ila var. hirta Gagnep. Bull. Soc. Bot. Fr. 73: 87. 1926; n Lecomte, Fl. Gén. Indoch. 5: 735. 1928. Syntypes, Indochina, Pierre 337 7, 3777, Thorel s.n., 1862-66 (P); lectotype, Pierre 3777 (e). Artocarpus chaplasha auct. non Roxb., Gagnep. in Lecomte, Fl. Gén. Indoch. 8 Differs from ssp. rigidus as follows: twigs rather softly and densely sub- appressed hispid-pubescent, juvenile shoots with long patent hairs; leaves obovate-oblong to ovate, base rounded or shallowly cordate, rarely cuneate; very shortly appressed-hispid above, rarely nearly smooth, indumentum beneath as on shoot; lateral veins 9-12 pairs; syncarp with processes hispid from spreading hairs c. 0.5 mm. long. VERNACULAR NAMES: Taung peing, Burma; Kanun pan, Siam; Knol prey or Knor prey, Cambodia; Cay mit nai or Mit nai, Annam. UsEs: as in ssp. rigidus. DISTRIBUTION: in evergreen forest to 3000 ft., Burma, Siam, Indochina. Burma. Kyauktwin [not located], Meebold 15599 (CAL). TENASSERIM. Amherst: Falconer 1015 (cat, @); Mekhrein chaung, Parkinson 5205 (pp, 2); Moulmein, Kurz s.n. (cAL). Tavoy: Kalemaung Res., Ba-Pe 864 (CAL, DD, 2); Kadwe For. Circle, Manson 762 (cat, 9). Siam. Chanburi |Chantaburi ?], Makham, Khao sabap, Put 2368 (ccE); Rayawng, Ban Pe, Kerr 2734 (BM, 2); Kaw Chang (island off se. coast), Klawng Mayom, Kerr 6923 (BM, p, 2), Mar- can 1329 (BM, 2). PENINSULAR SIAM. Surat, Sman 2365 (cGE, 6, 2); Surat, 1959] JARRETT, ARTOCARPUS AND ALLIED GENERA, III 155 Panom, Kerr 12375 (pm, 2). Island off w. coast: P. Terutao, Kerr 14198 (BM, ?). Islands off e. coast: Kaw Pa-ngan, Kerr 1235 (pM, 2); Kaw Tao, Kerr 12799 (BM, kK, ¢). Indochina. CAMBODIA: Hahn 61 (ep); Kamput, Pierre 15 (Pp); Phom-ba, Nem- Tram-Kok, Miller 350 (Pp); Pursat Prov., Pnom Barong, ass Herb. For. ee 36930 (P). COCHINCHINA. Thorel 1067 (A, BM, kK, 2), Thorel 1862-6 (P). Baria prov.: area 36637 (P), Pees aie . la Gutta 4 e S Xuon moc, Chevalier 36606 (P, 2). Bten Hoa prov.: Commission de la Gutta s.n., 1866 (Pe); Bao Chiang, Pierre 1851, p.p. (A, BM, K, P, 2); Song lu, Pierre 3777 (A, P); Trang-bom, Chevalier 30083 (2, 2), Chevalier 36761 (P), Ha 39323 (Pp, 2). Budot, Miller 1728 (Pp). Saigon i? cult.], Pierre 3377 (p, 2). Tay ninh prov., Cay Cing, Pierre ae p.p. (L). ANNAM: Massif de la Mére et l’Enfant, Poilane 6644 (kK, P, 2). _ ConporE: alee 4699 (P) This subspecies is consistently distinguished from the type by the hispid and not hispidulous syncarp processes, and usually also by the rounded or shallowly cordate base and rough upper surface of the leaves. In in- dumentum and leaf shape it is rather variable, though the hairs are usually denser and softer on both shoot and leaf than in ssp. vigidus. The latter feature is most marked in Burma, where, in addition, the leaves are fre- quently ovate. This form was described by Kurz as A. calophylla and is shown in King’s plate (1889), drawn from Falconer 1015 and not Kurz’s own collection, as stated by King, who reversed these two collections in his discussion of the species. In Siam and Indochina the leaf shape is usually obovate-oblong and the form was described by Gagnepain as A. asperula. However, since the variation between the two forms is con- tinuous, they are here united. Both species were distinguished by their authors from A. rigidus by the indumentum of the leaf, but, since the dif- ferences in this and in the leaf shape are not constant and those in the syncarp are so slight, it is preferable to treat the continental entity as a geographical subspecies of A. rigidus. The northern boundary of Malaya appears to separate the two subspecies, with the exception of Helfer’s col- lection of ssp. rigidus from Tenasserim, but more material from a area 1s desirable to show whether there is any intergrading of characte Gagnepain’s A. asperula var. hirta was described from sapling Piccaans with long patent hairs on the shoot. Two Indochinese collections of A. rigidus ssp. asperulus at Paris, Comm. de la Gutta s.n., 1866, and Pierre 15, were labelled A. polyphema Pers., but the native species which may have been included by Loureiro in his confused description of Polyphema cham- peden, on which this name is based, is not identifiable, although it will be further discussed under Artocarpus integer. The identification was not mentioned by Gagnepain in 1928. (To be continued) 156 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL A YELLOW-FLOWERED FORM OF RHODODENDRON CAROLINIANUM LEONARD F. FRISBIE Rhododendron carolinianum Rehder f. luteum, forma nova. A low, evergreen shrub of medium-compact growth. Branchlets brown with scattered scales. Leaves lanceolate, acuminate to acute at the apex, cuneate at the base, the blade to 6.3 cm. long, 2.5 cm. wide, yellowish green, glabrous above, reticulate, densely scaly below with scales separated by one-half their width, the petioles to 1.2 cm. long, scaly. Inflorescences terminal, the rachis 2 cm. long, about 6-flowered, the pedicels about 1 cm long, scaly. Calyx-lobes 5, ca. 2.7 mm. long, scaly; corolla rotate-funnel- shaped, 4 cm. across, mimosa yellow [Horticultural Color Chart 602/1], nearly self-colored, scaly on the outside, the lobes obtuse, the tube 1.0 cm. long; stamens 10, ca. 1.5 cm. long, slightly exserted, hairy at the base; ovary oblong, ca. 1 cm. long, scaly; style glabrous. C apsule narrowly oblong, 1.3 cm. long, s Frutex humilis sempervirens, compacto-mediocriter incremens. Ramuli fusci squamis sparsis. Folia lanceolata, acuminata vel acuta, basi cuneata, lamina 6.3 cm. longa, 2.5 cm. oe flavo-viridia, reticulata, dense squamosa. Inflorescentia terminalis, rachis 2 cm. longus, sex- floribus intextus; pedi- celli ca. 1 cm. longi; corolla ane infundibuliformis, 4 cm. lata, flava,” prope uniformiter colorata, extus squamosa, lobis obtusis, tubulo 1.0 cm. longo; stamina 10, ca. 1.5 cm. longa, parum exserta, basi pilosa; Ovarium squamosum, stilus glaber. Capsula 1.3 cm. longa, squamosa. Type: Plant in cultivation, 2728 Lemons Beach Road, Tacoma 66, Washing- ton; known only to have come from “the mountains of western North Caro- lina,” L. F. Frisbie, May 15, 1958 (Herb. Arnold Arboretum). This very handsome and distinctly different color-form of Rhododendron carolinianum was one of a group of native plants obtained about 1945 by Halfdan Lem of Seattle, Washington. The plants were offered to Mr. Lem as being ‘‘yellow-flowered carolinianum.” He accepted the small plants, but had such small faith in the possibility of the claim that he promptly forgot about the plants in the press of other matters. In the spring of 1957 this writer on a visit to Mr. Lem’s garden was attracted by one of the plants in flower. Due to my keen interest Mr. Lem let me have the form and it was brought to Tacoma. Subsequent efforts to gain information as to the exact natural location of the plants have met with no success. The possible collectors whom we have contacted have all given negative answers, and seemed to know nothing of such a form of the species, so it has proved impossible to obtain more exact information than that given above. At the present writing the plant is thriving, is making excellent growth, and has adapted itself very well to this climate. It will make an excellent 1959] FRISBIE, YELLOW-FLOWERED RHODODENDRON 157 addition to the series of evergreen rhododendrons, and will be a valuable horticultural subject wherever rhododendrons are grown. Here in western Washington, where a wide range of types of the genus is grown, we are very much pleased with this new color form. The hybridizing potential of the /uteum form is intriguing, especially so in the light of such acceptable hybrids of the typical form as ‘Conewago’ and ‘Conestoga.’ This yellow-flowered form of Rhododendron carolinianum is being propagated by ground layering and every effort will be made to see that it is generally available through the Washington Rhododendron Society, Inc. In addition to this color variation the typical form of the species with clear pink flowers is highly valued for decorative garden purposes, and a special place is reserved for the white-flowered form, R. carolinianum var. album Rehder. Good foliage, compact habit, and free flowering with chaste trusses of multiple flowers make this form a favorite which sometimes wins in rhododendron shows over all other species exhibited. The three color forms of Rhododendron carolinianum make a delightful combination in the garden, a grouping of true distinction. But we have found that it pays to be selective, especially with the pink and white forms of the species. Collected plants are not difficult to obtain, and persistent culling will turn up a very few plants of outstanding quality. Here in the West where the genus grows so well, some shaping and pruning are essential if one is to have compact plants. These distinctive forms of native American rhododendron species, both evergreen and deciduous solidly substantiate the importance of the work of the selective collector, a field in which the Washington Rhododendron Society has made a notable contribution. 2728 LEMoNsS BEACH RoaD TACOMA 66, WASHINGTON 158 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL ANDRODIOECISM IN THE FLOWERS OF TROCHODENDRON ARALIOIDES Hsuan KEncG * EXISTING DESCRIPTIONS of Trochodendron aralioides Sieb. et Zucc., the sole living representative of the vesselless angiospermous family Trocho- dendraceae (Smith 1945, Lawrence 1951), are based primarily on her- barium specimens. In all the botanical works consulted the flowers are described as hermaphrodite. The literature is reviewed in detail by Smith (1945). In the spring of 1957, from January to June, the author visited Bamboo Lake, forty miles northwest of Taipeh, Formosa, every two weeks to collect flowertmg buds, flowers, and fruits of Trochodendron. A number of trees growing naturally in the valley and on the hill slopes were ex- amined. About half of them were found to be protandrous: during the maturation of the stamens, the styles are clasped together (Figs. 1, 2). The other half were found to be protogynous, with the styles recurved dis- tally, exposing the ventral stigmatic surfaces, at a time when the stamens had not yet dehisced (Fig. 3). All the flowers of any one tree without exception presented the same condition, i.c., they were either all protan- drous or all protogynous. In mid-April the trees were in full bloom and were being frequented by honey bees (4 pis mellifica, A. indica) and snake-eyed butterflies (V pthima motschulski). The carpels and filaments of open flowers are pale yellowish green, while the anthers are bright yellow in color. A shining, viscid fluid coats the external surface of the gynoecia in a great majority of both pro- tandrous and protogynous flowers, although no nectar has been found. The sweet, pleasant fragrance is detectable from some distance away, especially on a sunny morning. Occasionally a few flowers appear to lack a coating of viscid fluid and remain dr In protogynous flowers the stamens possess well developed anther-sacs which dehisce subsequently to release normal pollen. In addition, the gynoecia are also well developed. On the other hand, in protandrous flow- ers, with an equally normal androecium, the styles are at first tightly clasped, only opening slightly later. The trees which bear protogynous flowers produce normal follicles and fertile seeds. However, the author failed to find a single normal follicle on those trees which bore protandrous flowers. Thus, the species appears to be androdioecious, a condition which Darwin (1896, p. 13) pointed out as being exceedingly rare among flower- ing plants. It would be most interesting to have reports on the floral biol- ogy of Trochodendron from other parts of its range, particularly from the islands of Japan. * The author wishes to express his thanks to Dr. Lincoln Constance and Dr. Herbert G. Baker for their helpful suggestions. PLATE I Jour. ARNOLD Ars. VoL, XL TROCHODENDRON ARALIOIDES Sieb. & Zucc. 160 JOURNAL OF THE ARNOLD ARBORETUM [VoL. XL LITERATURE CITED Darwin, C. The different forms of flowers on plants of the same species. Apple- 896. ton, New York. Lawrence, G. H. M. Taxonomy of vascular plants. Macmillan, New York. 1951 SmitH, A. C. A taxonomic review of Trochodendron and Tetracentron. Jour. Arnold Arb. 26: 123-142, 1945. DEPARTMENT OF BOTANY UNIVERSITY OF CALIFORNIA BERKELEY, CALIFORNIA EXPLANATION OF THE PLATE Fic. 1. Stages in the unfolding of the inflorescence of Trochodendron. Fic, 2. Development of protandrous flowers. Fic. 3. Development of proto- gynous flowers. (AIl natural size.) 1959] WOOD & CHANNELL, EMPETRACEAE & DIAPENSIACEAE 161 THE EMPETRACEAE AND DIAPENSIACEAE OF THE SOUTHEASTERN UNITED STATES C. E. Woop, Jr. AND R. B. CHANNELL THE STUDIES OF EMPETRACEAE AND DIAPENSIACEAE, two small families of both biosystematic and a a ee interest sonutiered below, are part of a generic flora which is in preparation for the southeastern United States.! The genera of these families oude examples on a different level of complexity from those of nine families of Ranalian affinities, - es ments of which have been published previously (Jour. Arnold A 9: 296-346. 1958; 40: 94-112. 1959). Further examples of the eee of the plants of this area and of the problems which they present will be published as space permits in order that some of this material may be ' Preliminary to a manual of the species of seed plants of this area, generic treat- ments of this type are intended to call attention to problems of various kinds which s é of an understanding of the plants of the area will be reached more quickly through this approach than through the traditional o While one may speak of “alpha” and Toa taxonomy and insist that one pre- i t the taxonomist have the knowledge and viewpoint to do this. It is not cient to say “Collect now — study later.” The very items the monographe careful student of evolution may need to know are the one may easily be overlooked (and S often are) by the well-meaning but uninformed collector. Provided with a guide to s if it is to be both analytic and synthetic, and if a modern ee of any area is to have a firm basis in an understanding of the plants involved, the approach must be one of understanding plants as living organisms and the info ioe considered must any of these levels. Certainly a number of taxonomic and phylogenetic, as well as morphological and biological, problems will be apparent ons the thoughtful reader of the descriptions, notes and references which follow. —C. E. W. 162 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL made more immediately available to those interested in the plants of east- ern North America. The general scheme of these studies is outlined in the first paper of this series (see above). It should be pointed out again, however, that the south- eastern United States as defined here is bounded by and includes North Carolina, Tennessee, Arkansas and Louisiana; that the descriptions are based primarily upon the species occurring within our area, any supple- mentary material added for clarity being included in brackets: that the abbreviations used for periodicals are the very useful ones of Lazella Schwarten and H. W. Rickett (Abbreviations of titles of serials cited by botanists. Bull. Torrey Bot. Club 76: 277-300. 1958); and that refer- ences which we have not seen are marked by an asterisk. All of this work, which is being conducted as a joint project of the Gray Herbarium and the Arnold Arboretum, has been made possible through the kind support of George R. Cooley and through a grant from the National Science Foundation. The detailed drawings of the four genera are the careful work of Dorothy H. Marsh. We are indebted to H. L. Blomquist, of Duke University, and to R. K. Godfrey, of Florida State University, for their respective kind- ness in sending for study and use in illustration a large series of specimens of Pyxidanthera brevifolia and excellent fresh specimens of Ceratiola in fruit. EMPETRACEAE (Crowserry FAMILY) 2 A small family of evergreen shrubs of ericoid habit with numerous pul- vinate leaves, inconspicuous apetalous flowers of few stamens and drupa- ceous fruits. Three genera of disjunct distribution and about eight species: Empetrum L., bicentric, with about four boreal species, primarily of arctic to subalpine distribution, and a single species in the subantarctic; Corema D. Don, with C. Conradii (Torrey) Torrey ex Loud., in widely separated cites from the Magdalen and Prince Edward islands and Nova Scotia to New Jersey, and C. alba (L.) D. Don in Portugal and the Azores: and the sienoreuin Ceratiola Michx. confined to our area. The systematic position of the Empetraceae has been the subject of considerable controversy. The family has been variously allied with the Celastrales, Ericales and Sapindales. The erect ovule and ventral raphe have been considered to indicate affinities with the Sapindales. A natural group, it is perhaps best regarded as a reduced apetalous and polygamous or dioecious derivative of the Ericaceae. This view is supported by the embryological data of Samuelsson and is confirmed by the morphological studies of Hagerup. A whole series of well-marked embryological features characterizes the Ericales. These represent standard stages in Ericalean embryology and constitute a combination unknown in any other order. The Empetraceae show close correspondence in all respects. The embrvol- By R. B. Channell. 1959] WOOD & CHANNELL, EMPETRACEAE & DIAPENSIACEAE = 163 ogy of the Celastrales and Sapindales differs in so many ways as to render any link with the Empetraceae obscure indeed. The 3-colporate pollen grains are united in tetrahedral tetrads as in Ericaceae. REFERENCES: FERNALD, M. L.. and K. M. WiEcANpD. The genus Empetrum in North America. Rhodora 15: 211-217. 1913. Goop, R. C. R. The genus Empetrum. Jour. Linn. Soc. Bot. 47: 489-523. 1927. [Analysis of variation; 10 units ae ie Gray. A. Chloris boreali-americana. Mem. Am. Acad. Arts Sci. II. 3: 1-56. 1848. [Corema and other N. Am. genera, 1-14. | Hacerup. O. Empetrum ge eters (Lge) Hagerup. A new tetraploid, seek species. Dansk Bot. Ark. 5(2): 1-17. 1927. [E. nigrum L., 26: E. hermaphroditum, 2n = 52.] a eiiaie: on the Empetraceae. Danske Vid. Selsk. Biol. Meddel. 205: 1-50. 1946. [ Morphological; discussion of relationships. | McEwen. M. The comparative anatomy of Corema alba and Corema Conradit, Bull. Torrey Bot. Club 21: 277-285. 1894. Pax. F. Empetraceae. Nat. Pflanzenfam. III. 5: 123-127. 1895. SAMUELSSON. G. Studien iiber Entwicklungsgeschichte der Bliiten einiger Bicor- nes-Typen. Ein Beitrag zur Kenntnis der systematischen Stellung der Dia- pensiaceen und Empetraceen. Sy. Bot. Tidskr. 7: 97-188. 1913. | Em- brvological; discussion of systematic position. See also MAHESHWARI, Introduction to the embryology of angiosperms, 359-361. | — Ceratiola Michx. Fl. Bor.-Am. 2: 221. 1803. Dioecious shrubs of dry sand strands and sand hills, emitting a charac- teristic odor, Leaves linear-acicular, in ‘“‘decussate” whorls, with a longi- tudinal groove on the lower surface, jointed at the base of the short peti- ole. Old stems roughened by petiolar sterigmata. Flowers axillary, ses- sile, hvpogynous, the perianth consisting of 5—6 bracts and sepals. Male flowers of 2 exserted stamens, the 2-loculed anthers longitudinally dehis- cent; pollen in tetrads. Female flowers with a single pistil, the 3-4 exserted stigmas divergent, flabellate-incised-pinnatifid, united into the single style arising from the summit of the ovoid, 2-locular ovary. Fruit small, greenish-yellow, drupaceous with 2 pyrenes, the seeds with a straight embryo. TYPE AND SOLE SPECIES: Ceratiola ericoides Michx. (C. falcatula Gandoger, Empetrum aciculare Bertol.) (Name from Greek ceras, horn, apparently alluding to the flabellate style-branches.) — ROSEMARY. Often covering extensive areas, Ceratiola is associated with pines and oaks throughout its range from Florida to South Carolina and Mississippi, and is a more or less characteristic shrub of the Pinus clausa-scrub in Florida. It does not occur on the Florida Keys. In some areas, especially on coastal sand dunes, it is associated with a shrub of similar habit and stature, Solidago (Chrysoma) pauciflosculosa Michx. Inhabiting extremely dry situations in pinelands, in sand hammocks and on inland sand dunes, often in almost pure sand with Selaginella arenicola Underw. and Cladonia, 164 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Ceratiola is often ravaged by fire. In southern Florida it may be seri- ously parasitized by the lauraceous woevine, Cassytha filiformis L. Fic. 1. Ceratiola. a—m, C. ericoides: a, habit b, @ flower, x 6; c, pistil, X 6; d, young ovary, semi-diagrammatic vertical section, showing two locules each with a single ascending ovule, X 12; e, mature fruit inclosed in weathered bracts and sepals, & 6; f, mature fruit free of bracts and sepals, xX 6; g, pyrene inclosed in ovary wall, * 6; h, pyrene, X 6; i, portion of vegetative shoot showing position and arrangement of leaves — note articulate, pulvinate, appressed petioles, X 6; j. portion of ¢ shoot in flower, * 1: k, 3 flower, X 6; 1, stamens of a single 3 flower, X 6; m, pollen tetrad, ca. 30 » diameter, X ca. 300. Usually a well-formed shrub, Ceratiola is very exacting in habitat and, like Empetrum and Corema, is difficult to cultivate, presumably because of a mycorrhizal relationship. The common name rosemary refers to the superficial resemblance to Rosmarinus officinalis L. Little is known concerning the agent of pollination (wind?), germina- tion and other biological features of the species. REFERENCES: HECHENBLEICHER, H. Ceratiola ericoides Michx. in Lexington County, South Carolina. (Abs.) ASB Bull. 3: 10. 1956. [Previously reported from the same county by R. M. Harper; successful cultivation reported. | Hooker, W. J. Ceratiola ericoides. Bot. Mag. 54: t. 2758. 1827. HuspeLl, T. H., and F. W. WaLKer. A new shrub-inhabiting species of Schis- tocerca from central Florida. Occas. Pap. Mus. Zool. Univ. Mich. 197: 1-10. 1928. [S. ceratiola, a Ceratiola-inhabiting species of grasshopper. | DIAPENSIACEAE (Dr1ApensraA FAMILY) ® Low, evergreen herbs or subshrubs of acid soils, tufted or creeping, with simple exstipulate, alternate leaves. Flowers regular, the sepals 5, free, imbricate, the petals 5, separate or united, the corolla campanulate or funnel-shaped. Androecium typically of two whorls: an outer of 5 con- nivent fertile stamens, and an inner (when present) of 5 staminodia oppo- * By C. E. Wood, Jr. 1959] WOOD & CHANNELL, EMPETRACEAE & DIAPENSIACEAE 165 site the petals; filaments adherent to the petals (and sometimes monadel- phous, as well); pollen grains single, 3-colporate. Style simple, the stigma 3-lobed, the ovary 3-loculed, superior, lacking a disc at the base; ovules 2-integumented, the placentation axile. Fruit a loculicidal capsule; seeds as in Ericaceae. (Including Galacaceae of Small’s Manual.) A small family of five genera, primarily of eastern America and eastern Asia (with the exception of the circumpolar Diapensia lapponica L.), form- ing a natural group with numerous reticulate relationships. Two tribes, DIAPENSIEAE Gray and GALACINEAE Gray, generally are recognized, al- though variously delimited (see Gray, Drude, Diels). The group is usually agreed to be related to the Ericales but differs from most in the simple pollen grains, the epipetalous stamens and the absence of a disc. In view of the importance accorded embryological evidence in allying the Empetraceae with the Ericales it should be noted that the em- bryological features of Diapensia lapponica do not fit with those of other members of the Ericales. Further embryological studies of other genera are highly desirable. The family is currently placed either with the Ericales or in a separate order Diapensiales. Chromosome structure and basic num- ber are similar throughout those members of the family thus far examined. The stamens and staminodia present interesting problems in function and morphology. REFERENCES: BALDwIn, J. T., Jk. Chromosomes of the Diapensiaceae. Jour. Hered. 30: 169- 171. 1939. [Representatives of all groups except Berneuxia. | Diets, L. Diapensiaceen-Studien. Bot. Jahrb. 50(Suppl.): 304-330. 1914. [ In- cludes morphology, ovules, geography, notes on relationships, etc. | Drupe, O. Diapensiaceae. Nat. Pflanzenfam. IV. 1: 80-84. 1889. Frye, E. M. American Diapensiaceae. New Flora & Silva 6: 192, 193. 1934. Gray, A. Note sur le Shortia galacifolia et révision des Diapensiacées. Ann. Sci. Nat. VI. 7: 171-179. pl. 15, 1878 [1879]. GREVEL, W. ‘Anatomische Untersuchungen iiber die Familie der Diapensiaceae. Bot. Centralbl. 69: 257-267, 309-315, 342-347, 369-377, 401-411. 1897. {Includes our Leas a. | SAMUELSSON, G. Studien iiber die Entwicklungsgeschichte der Bliten einiger Bicornes-Ty = Ein Beitrag zur Kenntnis der systematischen Stellung der Diapensiaceen und Empetraceen. Sv. Bot. Tidskr. 7: 97-188. 1913. [Em- bryological features of Diapensia lapponica. | KEY TO THE GENERA OF DIAPENSIACEAE A. Plant with small awl-like leaves; prostrate, creeping; staminodia lacking; anthers 2-locular, apparently transversely dehiscent, each locule awned on the lower side; capsule with a persistent but easily broken columella. .... Ce epe sete be Rae BS Pyxidanthera. A. Plant rhizomatous with shining oval to orbicular leaves; 5 staminodia present. B. Flowers large, solitary; petals connate, crenate-undulate-toothed; anthers large, oval, 2-locular; staminodia distinct, near base of corolla; capsule with a persistent columella; seeds ovoid or spherical. ........ . Shortia. 166 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL B. Flowers small, numerous, in wand-like racemes on naked scapes; petals entire, distinct; stamens and staminodia monadelphous, forming a tub adnate to the petals: anthers 1-locular; capsule without a columella; seeds angular 3. Galax 1. Pyxidanthera Michx. Fl. Bor.-Am. 1: 152. pl. 17. 1803. Creeping, prostrate, evergreen subshrub from a woody root, the small leaves alternate, oblanceolate, awl-pointed, sessile and hairy on the upper side (at least near the base). Flowers solitary and sessile on short, densely leafy branches. Sepals concave, oblong, reddish. Corolla white, about 5-10 mm. broad, the petals united by the broad stamen-filaments to form a_ tube, the lobes broadly spatulate, cuneate or obovate-cuneate. Stamens alternating with the petals, the filaments white, almost petal-like, the anthers bent inward; anther locules 2, each apparently dehiscing trans- versely and awned at the base; staminodia lacking. Style as long as the corolla tube, increasing somewhat in length in age; ovules 4—6 in each locule of ovary. Fruit with a persistent but brittle and easily broken columella; seeds globular, regularly pitted (seldom collected). 2n = 1 Type species: P. barbulata Michx. (Name from the Greek pyxis, a small box, and New Latin anthera, for anther, from the appearance and dehis- cence of the anthers.) — Prxir, FLOWERING-MOSS. One or two species: Pyxidanthera barbulata, of sandy pine barrens, usu- ally seasonally wet, on the coastal plain from New Jersey to Virginia, and North and South Carolina, and P. brevifolia Wells, of the inner coastal plain, Harnett Co., N. C. to Darlington Co., S. C. The status of this second species, which is associated with Quercus laevis Walt. and Pinus australis Michx, f. on the rolling sand hills, needs careful study, for it ap- pears to intergrade in moister habitats with the more widespread plant. Although P. brevifolia when well developed is characteristic in aspect, its only distinctive features are the smaller size of all parts and the greater hairiness of the leaves and stems, characteristics which may well be attributable to the more xeric habitat and which are matched or approached by various smaller specimens from New Jersey and North Carolina. The mechanical operation of the awned, connivent anthers, which are bee cealy transversely dehiscent, is probably significant in the biology of the small, white flowers which are borne abundantly in March and April. Pollen is not discharged from the anther unless the awned tip is pushed downward. A comparison with the stamens of Diapensia suggests that the orientation of the anther-locules has changed from vertical to hori- zontal in the inwardly bent anthers of this plant so that the dehiscence is not truly “transverse.” REFERENCES: BEAL, W. J. Insects needed to fertilize Utricularia and Pyxidanthera. Am. Nat. 12: 552-554. 1878. MFreHAN, T. Pyxidanthera barbulata. Meehan’s Monthly 8: 33. 34. pl. 3. 1898. | Illustration with details.] See also Bot. Mag. 77: pl. 4592. 1851. 1959] WOOD & CHANNELL, EMPETRACEAE & DIAPENSIACEAE = 167 WeELLs. B. W. A new pyxie from North Carolina. Jour. Elisha Mitchell Sci. Soc. 44: 238, 239. 1929. 2. Shortia Torrey & Gray, Am. Jour. Sci. Arts I. 42: 48. 1842; If. 45: 402. 1868, nom. cons. Low, evergreen, perennial herb, forming dense clumps or carpets, the foliage combining the aspect of Galax and Pyrola. Plants spreading by horizontal rhizomes bearing clustered long-petioled, broadly elliptic to sub-orbicular truncate-emarginate, cordate, lustrous leaves, toothed along the margin, and much smaller leaves or scales. Flowers 2-3 cm. high, solitary on bracted peduncles, nodding. Sepals ovate, imbricate. Corolla open-campanulate, white, pale pink or pale blue, the petals united, the lobes undulate-crenate notched. Fertile stamens with conspicuous, 2-loculed anthers bent sharply inward and connivent in the tube; staminodia borne near the base of the corolla and incurving over the ovary. Style elongate; capsule globular, 3-valved and with a persistent columella bearing the globular or ovoid seeds. 2” = 12. (Sherwoodia House; not Shortia Raf., 1840, nom. rejic.) Type species: Shortia galacifolia Torrey & Gray. (Dedi- cated to Charles Wilkens Short, 1794-1863, of Kentucky, ‘“‘whose attain- ments and eminent services in North American botany are well known and appreciated both at home and abroad.”) — SHORTIA, OcONEE-BELLS, LITTLE COLTSFOOT. In North America a single species and in Japan, Formosa and China perhaps 9-11 species, depending upon specific and generic concepts. Shortia galacifolia, S. uniflora Maxim., of Japan, and S. sinensis Hemsl. form a closely related group of species with solitary flowers and staminodia. Other species with several-flowered inflorescences and partially developed, although sterile, stamens have been treated as the genus Schizocodon Sieb. & Zucc., while four species described from Formosa with several-flowered inflorescences and no staminodia would be placed in Shortiopsis Hayata. All of these species appear to be closely related, however, and are best treated as a single genus, Shortia. The nearest ally is the Sino-Himalayan Berneuxia Decne. Shortia galacifolia, a handsome plant with an intriguing history shares in our area the special fame of Franklinia and “Bartram’s Ixia” (Spheno- stigma coelestinum (Bartr.) R. C. Foster). Collected by André Michaux in the Carolina mountains in 1787, the specimen was seen by Asa Gray in Paris in 1839 and was later described as a new genus. All searches for the plant were futile, but a second species described from Japan was recog- nized by Gray as belonging to the genus. Shortia galacifolia finally was rediscovered by George M. Hyams, in McDowell Co., N. C., in 1877, and, in 1886. was found by Sargent and Boynton near the junction of the Horsepasture and Toxaway Rivers in Oconee Co., C. The plant is now known from two limited areas some 60 miles apart: var. galacifolia, primarily in the drainage of the Keowee River in Oconee and Pickens counties, S. C., Rabun Co., Ga., and Transylvania Co., N. C., and var. brevistyla P. A. Davies, in McDowell and Burke counties, N. C. 168 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL ic, X 4; h, ovary, c i-q, Pyxidanthera. i-p, P. barbulata: i, habit, branchlet from above, < 1; j, flower, < 4; k, tip of flowering branchlet with flower after fall of corolla. « 4: 1, opened corolla, < 4; m, stamens, - — cheer: fruit _ mat (2 lobes removed), X 4; 0, ovary, cross section, x 10: x 2 P. bre evifolia: vpened capsule from above, ar Benecar tn gee ‘islodeed columella in 1959] WOOD & CHANNELL, EMPETRACEAE & DIAPENSIACEAE = 169 The two varieties differ in petal-length, -notching and -venation, length of staminodial hairs and length of styles. The ratio of length of style to length of mature ovary is 1:2.4 in the Keowee area and 1:1.3 in McDowell County. On this and other bases it appears that Michaux’ specimen came from near the spot where Sargent and Boynton first found it. Shortia is still abundant in a part of the Keowee area but should be protected from vandalism and excessive commercialism. It occurs in moderately acid soils with good aeration and usually with abundant moisture, growing best under Rhododendron maximum or Tsuga, but often in association with Pinus Strobus, P. rigida, Liriodendron, Liquidam- bar and Kalmia latifolia and various other Ericaceae. The altitudinal range is from about 600 feet (180 m.) to well above 1600 feet (500 m.) but the plant is hardy far to the north of its restricted range. The pro- terogynous flowers are borne in March and edtly April and the seeds mature from late April to May (or June). It has been suggested that the restricted distribution is related to the lack of a dispersal mechanism and to limited reproduction by seeds, for the seeds often germinate even within the capsule. However, natural reproduction by seeds is reported in Am- herst Co., Virginia, and in the Keowee region. REFERENCES: eer’ D. L. Shortia galacifolia in Gray’s Manual range. Rhodora 58: 38-40. 956. [Amherst Co., Va.; introduced? pices P. A. Distribution ait abundance of Shortia galacifolia. Rhodora 57: 189-201. July 1955. _Geographical variations in Shortia galacifolia. Rhodora 54: 121-124. 1952. [Var. brevistyla. —. Type location of Shortia galacifolia. Castanea 21: 107-113. 1956. [See also Wild Flower 33: 24-31 Duncan, W. H., H. VENARD Per G. W. ee Shortia galacifolia from Georgia. Rhodora 52: 229-232. 1950. [Rabun H[ooker], J. D. Shortia galacifolia. Bot. Mag. 115: e: 7082. 1889. House, H. D. The genus Shortia. Torreya 7: 233-235. ; Jenkins, C. F. Asa Gray and his quest for Shortia galacifolia. Arnoldia 2: 13- 28. 1942. [A detailed historical account. | Li. H. L. On the Sino-Himalayan species of Shortia and Berneuxia. Rhodora 45: 333-337. 1943. [Unites Shortia, Schizocodon and Shortiopsis. | Prince, A. E. Shortia galacifolia in its type locality. Rhodora 49: 159-161. 1947. Ross, M. N. Seed production of Shortia galacifolia. Jour. N. Y. Bot. Gard. 37: 208-211. 1936. [Germination of seeds within the ovary. | SarGENT, C. S. The story of Shortia. Garden & Forest 1: 506, 507. 1888. [History; includes an excellent illustration and an account of Sargent’s visit to Toxaway and Horsepasture rivers. center, X 4. r-x, Galax. G. aphylla: r, habit, & 14; s, flower, from above, « 4; t, opened corolla and staminal tube, X 4; u, stamen and oi . X 10; v, ovary vith 3 sepals and bracteole, X 5; w, fruit, < 4; x, seed, X 2 170 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XL 3. Galax L. Sp. Pl. 1: 200. 1753, partim, emend. Nutt. Gen. 1: 145. 1818, nom. cons, Low herbs with scaly-bracted rhizomes forming a thick, matted tuft bearing long-petioled, round-cordate, crenate-toothed, lustrous evergreen leaves, the plant spreading by slender cord-like rhizomes which later be- come branched, thickened, and leaf-bearing at the tips. Flowers small (4-5 mm.), white, each with minute bracts at the base, borne in a wand- like raceme on a slender, naked scape in May (or July at higher altitudes). Sepals 1/3—1/2 the length of the oblong-spatulate petals. Petals distinct to the base, nearly erect or spreading. Filaments of 5 fertile stamens and 5 staminodia united, forming a tube adnate to the petals at the base, fall- ing with the petals; free tips of the staminodia fleshy, nearly erect, the fertile anthers bent at right angles to the tube, connivent, each anther apparently 1-loculed and opening horizontally (actually across the top), the lower (inner) half tapering to an obtuse point. Pistil about as long as the calyx, the style short, the stigma 3-lobed. Capsule obpyriform, about 3 mm. long, without a columella, the seeds numerous, small, brown, angular, tapering, with a cellular seed coat. 2m = 12, 24. (Not Galax L., 1754, nom. rejic.) TYPE AND SOLE sPECIES: G. aphylla L. (Name from Greek, gala, milk, presumably from the white flowers.) — GALAX, GALAXY, WAND-FLOWER, COLTSFOOT, BEETLEWEED, BEETLE-PLANT, SKUNK-CAB- BAGE. A handsome and distinctive monotypic genus ranging from northern West Virginia and northwestern Maryland to the coastal plain of eastern Virginia and eastern North Carolina, to central Georgia, central Alabama and cen- tral Tennessee, in acid soils, generally in mesophytic associations with Kalmia or Rhododendron and other Ericaceae. The plant reaches its best development in rich, acid, humus-covered soils in the piedmont and Blue Ridge where, from North Carolina to Georgia, great quantities of the leaves, which become bronzed in winter, are gathered and sold to florists. Galax provides one of the clearest, simplest and best studied cases of autoploidy in wild populations. Diploids occur throughout the range of the species, except in the Virginia coastal plain, where only the tetraploids are known, Tetraploid plants grow in the same habitats and areas as diploids, but are especially concentrated in the region from southern Vir- ginia through the North Carolina mountains into South Carolina and Georgia. They appear to be absent from the northern and western parts of the range. Under the most favorable conditions the leaves of tetraploids may reach 15 cm. wide, while those of diploids attain only 10 cm. The plants are usually indistinguishable in the field, however, and no morpho- logical distinctions other than those of size are known. No triploids and no meiotic irregularities have been reported in either diploids or tetra- ploids. The apparently one-loculed anthers are strongly reminiscent of those of Pyxidanthera and suggest a similar mode of development. The flowers are most like those of Pyxidanthera (except for the presence of staminodia) 1959] WOOD & CHANNELL, EMPETRACEAE & DIAPENSIACEAE 171 but in vegetative features Galax suggests Shortia. The genus has been treated either as the sole member of the tribe GALACINEAE or has been placed there with Shortia and Berneuxia. REFERENCES: ANprREwS, E. F. The galax odor. Torreya 15: 16-18. 1915. [Leaves in large masses emit an elusive “faint, sickly carrion scent.” BALDWIN, : T., Jr. Distribution of Galax aphylla in Virginia. Va. Jour. Sci. 2: 68, 69. 1941. ee The genus and its chromosomes. Jour. Hered. 32: 249-254. 1941, ge and tetraploids, their characteristics and distributions. | CLAUSEN, J., D. D. Keck and W. M. Hresey. Experimental studies on the nature oe species. II. Plant evolution through amphiploidy and autoploidy, with examples from the Madiinae. Chap. 8. Ecologic characteristics of natural autoploids. Carnegie Inst. Wash. Publ. 564: 130-148. 1945. [In- cludes Galax. | GraEF, P. E. The ovule and embryo sac of Galax aphylla. Va. Jour. Sci. Il. 9: 319-322. 1958. [Ovule and embryo sac development closely parallels that of Diapensia. Cf. SAMUELSSON. | Pike, A. V. Galax aphylla. Gare Chron. III. 119: 148. 1946. [Notes on the plant and its culture. See also Bot. Mag. 20: pl. 754. 1804.| ARNOLD ARBORETUM VANDERBILT UNIVERSITY 172 JOURNAL OF THE ARNOLD ARBORETUM ‘VOL, XL THE EFFECT OF JUVENILITY ON ROOTING OF CUTTINGS FROM APPLE SEEDLINGS DONALD L. SMITH It HAS LONG BEEN KNOWN that cuttings from young seedlings can be rooted much more readily than can cuttings from mature trees. Extensive tests conducted by Gardner (1929) with seedlings of apple, pear, cherry, elm, locust, pine and spruce showed that cuttings from one-year-old seed- lings rooted easily. He obtained some rooting from cuttings from two- year-old seedlings, but as the seedlings grew older it was difficult or impos- sible to root the cuttings taken from the upper branches. It also has long been known that the basal part of the tree remains in the juvenile stage. More than 150 years ago Thomas Andrew Knight ob- served that shoots from the base of a seedling pear tree possessed juvenile characters and that scions from such shoots, when grafted on pear root- stocks, were much slower to come into fruit than scions from the bearing branches. It is also known that cuttings taken from the lower branches of seedling trees will root more easily than cuttings taken from the tops of the mature trees (Grace, 1939; O’Rourke, 1951). The basal part of the tree appears to remain in the juvenile stage and a seedling tree can be kept in the aromas stage indefinitely by cutting it back to the ground each year (Blair, 1955). These observations on the relation of juvenility to ease of rooting of cuttings are of considerable significance in the clonal propagation of trees and shrubs. Clonal propagation of certain species by cuttings is essential to maintain a uniform root system. It is also essential in species such as the lilac, which suckers from the roots, if the clone is to be maintained true to type by the amateur horticulturist. Seedling trees can be grown to maturity so that the flowers and fruits can be evaluated, but can still be propagated readily by cuttings by forc- ing sucker shoots from the base of the trunk. This can be done by girdling the bark near the base of the trunk or cutting the tree back nearly to the level of the ground. Less drastic methods are bark inversion or “scoring” the bark of the trunk near the base of the tree. Tests were made on four-year-old seedlings of an ornamental apple, Malus ‘Henry F. duPont,’ grown in the Bussey Institution nursery during the summer of 1958. In June 1957 every other tree in the nursery row had a short ring of bark inverted to induce earlier flowering. The bark-inverted trees did not fruit earlier than the controls, but they did produce profuse suckering from the base of the tree. In some cases the sucker shoots bore leaves which were tri-lobed, even though the mature branches bore only entire leaves. In other cases there was little morphological difference be- 1959 | SMITH, JUVENILITY AND ROOTING OF CUTTINGS 173 tween the juvenile and adult leaves. All of the trees tested fruited for the first time in 1958. Cuttings from sucker shoots and from fruiting branches were taken in June. The cuttings, which were about eight inches long, were cut in half to provide basal and tip cuttings. Half of the cuttings were treated with Hormodin No. 2, while the other half were planted with no hormone treat- ment. The cuttings were set in wet sand in greenhouse flats, a wire frame was placed over them, and the entire flat and wire frame were enclosed in polyethylene film. The enclosed flats were set in the shade under the greenhouse bench and given no further attention until examined for root- ing. The results are shown in TABLE I The juvenile cuttings of tree number 16155—4 rooted rather well with or without hormone treatment, but the adult cuttings rooted poorly or only moderately well, even with hormone, and very poorly, or not at all, without hormone. The juvenile cuttings of tree 16155—14 also rooted much better and earlier than did the adult cuttings. With tree 16155—21 the adult basal cuttings rooted comparatively well with hormone. In general the basal cut- TaBLe I. Rooting of Cuttings from Juvenile and Mature Branches TREE TYPE OF NuMBER OF PER CENT RooTED AFTER WEEKS NUMBER CUTTINGS HORMONE CUTTINGS 2 3 4 5 6 16155-4 a" + 10 40 40 50 uy J.T — 10 40 50 tt J.B + 10 80 90 90 a J.B - 10 80 80 ae A,T + 10 20 20 20 os A,T _ 10 0 ) - A,B + 10 @) 50 50 - A,B _ 10 0 10 16155-14 dat + 8 75 87 100 : hie a 10 20 30 J.B + 8 87 87 100 : J.B = 10 100 100 yh A,T + 8 0) 0 ) ¥: Aol — 10 0) 10 A,B +}. 8 ) 12 62 ue A,B _ 10 20 40 16155-21 iY + 9 0 33 1) . J,B + 9 11 77 77 se A,T + 9 0) 0 Z2 44 ‘i A,B ar 9 ) 44 77 tings of either juvenile or adult branches rooted better than the terminal cuttings and the juvenile cuttings from the basal suckers rooted much better than the adult cuttings from fruiting branches. The cuttings which 174 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL rooted in two weeks were more likely to thrive when transplanted than those which required a longer time. Cuttings from root suckers were also found to root more easily than cuttings from mature fruiting branches, even when the root suckers had developed six or seven feet from the base of the tree. Suckers from the roots were numerous on a Malus sargentii {. rosea hybrid (3340) which was 18 years old and growing in sod. Cuttings were made in the same manner as previously described. The results are shown in TABLE II, No rooting was obtained from the cuttings from the fruiting branches, but moderate rooting, with hormone, was produced in four weeks by cuttings from the root suckers. Evidently the roots retain their juvenility for a si distance from the base of the tree than do the branches. the summer of 1957 Dr. Karl Sax found a seedling of Hopa Crab (Malus baccata < pumila atchnilalibiiagice dvds men appsanee to as prom- ising as a dw of the extremely dwarfing rootstock varieties used in ee Cuttings from the original three-year-old seedling gave 91 per cent rooting in two sage and cuttings from lateral branches of the Hopa Crab seedling budded M. sargentii {. rosea gave 100 per cent rooting in two weeks, using conone (Taste II). Even without hormone 85 per cent rooted in four weeks. These rooted cuttings were transplanted to soil, with no loss, for testing as a dwarfing rootstock and to see if they can be kept as permanent juveniles (as a source of cuttings) by keeping them cut back to the ground level each year. TABLE II. fe of ee from sabaail and Adult cenanes TREE “Typ Num UMBER OF i CENT # Roore, Arren Ww EEKS NUMBER C an HORMONE © mo: UTTINGS 2 4 33340 A.T.* + “25 0) 0 AT. -- 25 ) 0 me RT: + 15 13 53 Hopa Sdlg. T + 12 91 . S + 12 100 S — 20 40 85 * A.T. = Adult Terminal; R Root sucker; T - Terminal: Ss = Exiondiey: SUMMARY Juvenile shoots from the base of fruiting seedling apple trees, induced by bark inversion, were found to root more readily than cuttings taken from the fruiting branches. Root suckers of a mature apple tree rooted well, although cuttings from the fruiting branches produced no roots, even with hormone treatment. A three-year-old seedling of Hopa Crab rooted very readily from cuttings to be tested as a dwarfing rootstock clone. The production and retention of juvenility in clonal varieties of trees and shrubs should be of value in clonal propagation. 1959 | SMITH, JUVENILITY AND ROOTING OF CUTTINGS 3 This work was supported, in part, by a grant from the Massachusetts Society for Promoting Agriculture to Dr. Karl Sax. LITERATURE CITED Buatr, D. S. The propagation of Malus species. Proc. Plant Propagators Soc. 1955: 84-91. 1955. GARDNER, F. E. The deena between one age and the rooting of cuttings. Proc. Am. Soc. Hort. Sci. 26: 101-104. Grace, N. H. Vegetative Rea of ae Canadian Jour. Res. 17: 178- 180. 1939. O’RourkeE, F. L. The effect of juvenility on plant propagation. Proc. Plant Propagators Soc. 1951: 33-37. 1951. 176 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL STUDIES IN THE GENUS COCCOLOBA, VII. A SYNOPSIS AND KEY TO THE SPECIES IN MEXICO AND CENTRAL AMERICA RICHARD A. HowArpD THE ONLY COMPLETE MONOGRAPH of the genus Coccoloba was prepared by Gustav Lindau and published in Engler’s Botanische Jahrbiicher 13: 106-229. 1890. Since that time many additional species have been de- scribed in the genus as the result of extensive collecting by botanists in Mexico and in the several countries of Central America. Paul Standley revised Coccoloba as it occurs in Mexico in his treatment of the trees and shrubs of Mexico published in 1922. More recently Standley with Julian Steyermark treated the species of Guatemala as part of their series of papers on the flora of Guatemala. In addition, lists of species, some with critical notes, will be found in the following papers. Other papers and citations are included in the text. ALLEN, P. Preliminary index to the trees of the Golfito-Palmar area, Costa Rica. 37. 1952. SMITH, J. D. Enumeratio plantarum Guatemalensium 6: 36. 1903. STANDLEY, P. C. Flora of the Panama Canal Zone. Contr. U.S. Nat. Herb. 27: 170. 1928. . Flora of Yucatan. Publ. Field Mus. Bot. 3: 252-253. 1930. ———. Flora of the Lancetilla Valley, Honduras. Publ. Field Mus. Bot. 10: 180-181. 1931. . Flora of Costa Rica. Publ. Field Mus. Bot. 182: 413-414. 1937. and S. CALDERON. Lista preliminar de las plantas de El Salvador. 98. eral —) co S. J. REcorp. ay forest and flora of British Honduras. Publ. Field Mus. ae 12: 125-127. 6. YUNCKER, T.G. A nia to the flora of Honduras. Publ. Field Mus. Bot. 17: 361. 1938. . Flora of the Aguan Valley and nes as regions near La Ceiba, Honduras. Publ. Field Mus. Bot. 9: 286. In general, the genus Coccoloba has been overdescribed in Mexico and Central America. Of the eighty names which have been applied in pub- lished form, only thirty-three species are recognized in this paper. In large part this tendency to excessive description is due to floristic treatments which deal individually with the many small countries of Central America. Many species which were considered endemic in adjacent countries have proved to be identical instead. Several West Indian species are now rec- ognized in the Central American flora and the corresponding Central American name has been reduced to synonymy. Seven species had been de- scribed on the basis of sterile branches from adventitious shoots. All of these have now been associated with fertile material and these names placed 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 177 in synonymy. Fortunately only one such name had to be retained with a sterile holotype specimen. The failure to recognize the unisexual con- dition of the plants led to the duplication of specific names and several such species have now been properly associated. Two new species are described and one species has been removed from synonymy and re-estab- lished at the species rank. Lindau in his monograph treated the species from Mexico and Central America in three sections of the genus. While these sections are useful in a preliminary assignment of species within the genus, the sections can not be maintained. When sufficient material is examined, intermediates are found and the sections grade indistinguishably into one another. In the previously published studies | of the genus Coccoloba in the West Indies I had the advantage of extensive field work and an intimate knowl- edge of the variation found in the living plant. Much of the information gained in studying the West Indian species has been applied to the current study. Many of the experienced field botanists who have described species of Coccoloba in Mexico and Central America have regarded the flowers as being perfect. The extensive group of herbarium specimens I have seen, as well as the living plants I have studied in Mexico and Honduras, clearly indicates, however, that the flowers are unisexual or functionally so. The pistillate flowers are normally borne singly at each nodule of the inflores- cence rachis or rarely in clusters of two or three. The stamens of such flowers are rudimentary and included. The staminate flowers with pollen- producing, exserted stamens are borne in clusters of three or more. Only a few species of Mexico and Central America have heavily cori- aceous leaves or leaves much reduced in size. The leaves tend to be more pubescent than those of West Indian species. The variation in size be- tween leaves of normal shoots and/or adventitious shoots is similar in the West Indies and in Central America. The leaves of adventitious shoots are larger. thinner and on longer petioles than those of normal shoots. Short-shoots, or at least shoots of limited growth, occurring laterally on the branches may also produce leaves smaller in size than those of shoots of normal growth. Pubescence varies with age of the leaf. An adequate understanding of the species is best obtained from a study of staminate and pistillate flowers, fruits and leaves from both normal and adventitious shoots. I am grateful to the several collectors indicated in the text who made special efforts to get complete material for me. I have also been fortunate to be able to study nearly all the types of the species considered in this paper. I am particularly grateful to the directors and curators of the Botanisches Museum, Berlin; the Jardin Botanique de lEtat, Brussels: the Royal Botanic Gardens, Kew; the Chicago Natural History Museum, Chicago; the Botanische Staatssammlung, Munich; and the University Herbarium, University of Michigan, Ann Arbor, for the bg materials sent for this study. Equally valuable but less authen- our. Arnold Arb. 30: 388-424. 1949; oe oe 1956; 38: 81-106. 1957; 38: 211-242. 1957; 39: 1-48. 1958; 40: 68-93. 178 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL tic specimens were received on loan from the other herbaria designated by the standard abbreviations given in Index Herbariorum. Distribution of recognized species of Coccoloba in Mexico and Central America GREATER ANTILLES EL SALVADOR NICARAGUA Costa RICA PANAMA Mexico | SouTH AMERICA | LESSER ANTILLES * || Bk. HonDURAS x x & || HONDURAS acapulcensis x xX 3 = nv n = ~ = a J iy g, Bs oa x X x xX M XX X X X || GUATEMALA x x = = = S ss) = ww 5 a aes 4 x x humboldtii x X asserl lehmannit x liebmannit lindeniana lindaviana x x lundellii x manzanillensis x matudat montana x &X x &X nicaraguensts x novogranatensis x > = ay = Ss Ba 3 = “=. eS 5 ae ae a4 « th _ 5 Q a x & tuerckheimii uvifera x x venosa x x be oe ae x xX The following list contains the names used for species of Coccoloba in Mexico and Central America. These names either were used in floras, lists or monographic treatments or apply to taxa based on specimens collected 1959] HOWARD. STUDIES IN THE GENUS COCCOLOBA, VII 179 in this area. The generic name Coccoloba has been conserved. Coccolobis is regarded as an orthographic variant and all binomials published under this spelling are also known as Coccoloba. The many transfers made by Otto Kuntze to the genus Uvifera have been omitted. All of the names in this list are discussed in the text under the recognized specific name. Varie- ties which are treated as distinct from the species are listed below. Varieties not listed are transferred with the species. Campderia floribunda = Coccoloba venosa C. lindeniana = Coccoloba lindeniana C. mexicana = Coccoloba venosa C. nematostachya = Coccoloba man- zanillensis Coccoloba acapulcensis Standley acuminata HBK. isididie major = C. venosa allenti = C. Page te = C. chiapensis barbadensis Jaton belizensis Standley bracteolosa = C. parimensis browniana = C. acapulcensis cardiophylla = C. acapulcensis aracasana Meisner changuinolana = C. lehmannii chiapensis Standley colonensis = C. oe coronata = C. spic corozalensis = C. nel diversifolia Jacq emarginata = Neomillspaughia emarginata (Gross) Blake escuintlensis = C. montana excelsa ie = C. parimensis excoriata = C. venosa floribunda = C. venosa fluviatilis = C. barbadensis gentler = C. swartzii mannii Standley a9 a a guatemalensis = C. tuerckheimii hirsuta = C. belizensis hondurensis Lundell humboldtii Meisner jurgenseni = C. barbadensis lancifolia = C. diversifolia lapathifolia = C. liebmannii laurifolia = C. diversifolia ea ehcp tee a ak GOP PAI OO OG ere oO grandifolia Standley = C. liebmannii Coccoloba lasseri Lundell Lh ae oe eee ae ar accor es = . latifolia = C. tuerckheimii lehmannii Lindau leptostachya = C. barbadensis liebmanni = — c. liebmannii lindeniana (Bentham) Lindau lundellii = C. X lundellii x lundellii Standley macrophylla = C. rugosa Desfontain Pesailents Beurling ie ela C. manzanillensis marginata = C. hondurensis masoni = ee barbadensis matudai Lundell mayana = C, barbadensis molinae = C. venosa montana Standley nematostachya = C. manzanillensis nicaraguensis Standley & L. Williams ivea = C. venosa novogranatensis Lindau oaxacensis = C. barbadensis obovata HBK. orizabae = C. humboldtii padiformis ane parimensis Bentham petrophila = C. humboldtii pubescens = C. liebmannii roseiflora = C. padiformis reflexiflora aes ey riparia = C. obov sessiliflora = C. babadenits schiedeana = C. barbadensis schippit = C. montana spicata Lundell steyermarki = C. montana standleyana Allen strobilulifera = C. acuminata suborbicularis = C. < lundellii swartzii Meisner 180 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Coccoloba tuerckheimii Donnell Smith Coccoloba waittii = C. oe C. umbilicata = C. pyrifolia C. werckl Desfontaines C. yucatana = C LoTR C. uvifera Linnaeus Uvifera oo = Coccoloba C. venosa Linnaeus lehman The following is a synoptic key to the species of Mexico and Central America. To be fully applicable, the key requires both flowers and fruit. It does not apply to sterile material, particularly that of adventitious shoots. Following the key, the species are listed in alphabetical order with a citation of specimens seen and a discussion of the nomenclature adopted. The countries are considered in order from north to south, with the states, provinces, districts, etc., in each country being listed alphabetically. KEY TO THE SPECIES 1. Petioles arising from above the base of the ocrea, the diameter of the leaf scar facto than the distance from the base of the leaf scar to the base of the (Species inadequately known: Leaves lanceolate-ovate, deciduous, the leaves turning black on drying. C 2. fap aban paniculate. 3. Inflorescence branches leaf base rounded or cordate; fruit spher- ical, rounded at the base........... «2 Xi Inflorescence many- ey leaf blade acute at the apex, the base decurrent on the petiole; fruit oval, stalked at the base. edb ace eA ene AROS PA aoe hy cok beck ata us C. tuerckheimii. 2. Inflorescence racemose or spicate. Inflorescence racemose 5. Leaves narrowly oblong to narrowly elliptic. .. . C. lindaviana. 5. Leaves cordate, elliptic or obovate-elliptic. 6 aves cordate, the base strongly cordate, occasionally peltate; flowering and fruiting pedicels 5-15 mm. long; fruit globose, cm. diameter, or abnormally obovoid and 2.5 cm. long; ocreae splitting longitudinally, flaring, the petiole appearing winged at the base; the inflorescence ocrea often appearing spathe-like. Che ea AanAa dep Adee EERE AD SBOE GA 4 C. acapulcensis, Leaves elliptic to obovate-elliptic, flowering and fruiting ped- icels shorter; ocrea not flaring if s 7. Leaves usually narrowed to a cordate-auriculate base; ocre- olae conspicuous, papery, flaring; flowers never reflexed; fruit globose, conspicuously coronate. .. C. novogranatensis. Leaves rounded at the base or narrowed to an obtuse base; ocreolae small and inconspicuous, the flowers commonly reflexed; fruit speak at the base and _ at most slightly coronate, at the apex. ..... - C. reflexiflora. 4. Inflorescence spicate, the fruits sseentiatly sessile. tan-colored when dry, the petioles usually grayish green; leaf blades acute to obtuse - the apex; fruit globose, ee t the ODER. co area ga a Sad Hoag lte es kd eta Be 4 . Spicata. Liseye ea ealxGaceseee ec . nicara pana ° > ~sI 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 181 8. Leaves and petioles turning dark brown or black on drying; leaf blades acuminate at the apex; fruit elongate, the achene stoutly POrOnate AC Ie ANON. baci Sk v's hs ae adeeb an eee obovata. 1. Petioles arising from the base of the ocrea, the base of the leaf scar at or close to the base of the ocrea. (Species inadequately known: ......................... C, standleyana.) 9. Inflorescence paniculate. 10. o 9. Inflo 11 Flowering pedicels stout, not exceeding the ocreolae; young stems, ocreae and inflorescence rachis conspicuously puberulent or tomen- tose to pilose on adventitious shoots, the pubescence golden or chest- nut-colored; the terminal buds large, conspicuously club-shaped, swollen and rounded at the apex, the ocrea of the terminal bud com- monly subcalyptrate; fruit globose, not narrowed to a stalk at the base, obtuse or rounded, not coronate, at the apex. .... C. belizensis. Flowering pedicels tenuous, exceeding or up to twice the length of the ocreolae; young stems, ocreae and inflorescence rachis puberulent, the pubescence commonly noticeable only with a lens; the terminal bud slender, acute at the apex, the ocrea splitting laterally ; mature PPUICS Ot ENOWN? (cee reste ios + 25 be ban ee vee Beaes C. matudai. rescence racemose or spicate. Inflorescence racemose; the flowers and fruits borne on pedicels which exceed the bracts and the ocreolae. 12. Lianas; leaf blade umbonate between the veins. .. C. parimensis. Trees or shrubs, at most the ends of the branches rarely scram- bling. 13. Leaves as broad as or broader than oo orbicular to broadly oblong, apex rounded to emargina 14. Fruits 1-2.5 cm. long, obtuse or rounded at the apex, at most Nas slightly coronate, usually slightly stalked at the bas 1S; oes usually much broader than long, the base strongly cordate with one lobe of the blade usually bo bo 12, slightly overlapping the petiole. C. uvifera. 15. Leaves usually orbicular, a base of the blade essentially rounded or obtuse. _C. X lundelliz. 14. Fruits less than 1 cm. long. 16. Fruit strongly coronate at the apex, the base rounded; leaf blades persistently pubescent only along the midrib and primary veins or glabrate. Mee a ae C. barbadensis hybrid. 16. Fruit Fane the perianth lobes distinct on the up 1/3 or 1/2 of the fruit; leaf blades per- sistently pubescent on the lower surface. 1h Ce a ea C. goldmannii. 13. Leaves noticeably longer than broad. 17. Leaves elliptic-lanceolate in outline, acuminate to long- acuminate at the apex, cuneate or rounded to an obtuse asymmetrical base; petioles 2-6 mm. long. ne ee ee ene C. chiapensis. ive Eeaves not of this type. 18. Foliar ocreae 3-5 cm. long. membranaceous or 182 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL chartaceous, silky pubescent, fruit coronate at the apex, the crown surrounded by the distinct perianth lobes, these fused below. ...... C. manzanillensis. ae Foliar ocreae 1-2 cm. long, coriaceous, or if char- taceous, not silky pubescent. 19. Achene surrounded by the expanded perianth lobes, these free nearly to the base. 20. Leaves elliptic to elliptic-oblong. the flow- er clusters distinct on the inflorescence rachis. ~ 9° C. lehmannit. Leaves narrowly oblong or obovate-ob- long, the inflorescence of closely asso- ciated flower clusters and these not clearly distinct on the rachis and not appearing 24; eaves obovate-oblong. broadest above the middle; midrib and veins usually pubescent below, blades per- sistently barbate in the axils of the primary veins. ........ C. venosa. Leaves narrowly oblong, broadest at middle; leaves glabrous below. C. darienensis. 19. Achene surrounded by the expanded hypan- thium, the perianth lobes coronate or sur- rounding the upper third of the achene. 22. Fruit conspicuously coron eav 23. ate. es lanceolate-ovate in outline, ocreae glabrous; pedicels in fruit sev- eral times as long as the thickness of the inflorescence rachis. gah ee ese 4a eek C. diversifolia. Leaves broader, ovate-oblong to obo- vate-elliptic in outline, ocreae pu- berulent; fruiting pedicels not longer than the thickness of the inflores- cence rachis 24. Fruiting pedicels shorter than the ocreolae; leaves usually oie or with a see bluntly acute apex, concolorou do de hades pb ated C. Set aren Fruiting pedicels slightly longer than the ocreolae; leaves usu- ally with a short but sharply acuminate apex, dark or black on the upper surface and lighter below. ....... C. hondurensis. ht as 22. Fruit obtuse ‘i the apex, the perianth lobes imbricate over the obtuse apex o the achene 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 183 11. 25. Leaves persistently pubescent below, argin entire but usually con- spicuously crispate to undulate. =} 32. Leaves of fertile branches generally with blades 12-20 cm. long and 8-12 cm. broad; gia densely flowered; fruit obtuse at the apex and only slightly coronate........... C. hon ts . Leaves of fertile branches usually with blades 7-11 cm. long and 5-6.5 cm. broad; inflorescence axis sparsely flowered, the pe clusters distinct and separated; fruit rounded at the apex and distinct- LY COPOMALG.. 6 o4as cere db eee Eopees C. swartzit. Ww bho Coccoloba acapulcensis Standley, Proc. Biol. Soc. Wash. 33: 66-67. 1920. Coccoloba cardiophylla Standley, Publ. Field Mus. ra . 8: 1930. Coccoloba browniana Standley, Trop. Woods 10: 927 Coccoloba wercklei Standley, Publ. Field Mus. ay 4: 304. 1929. It is unfortunate that the oldest name applicable to this species is based on anomalous material. To the present, Coccoloba acapulcensis has been distinguished by the peltate leaves and the large fruits. The collections cited below, made since the type collection of the species, show clearly that leaves with peltate, cordate or rounded bases may be found on the same shoot. Miranda (Anal. Inst. Biol. México 14: 29. 1943) reports that peltate leaves appear abundantly on all young specimens of C. acapulcensis but that the cordate base is more frequent on older plants. The fruits of this species are larger than is usual in the genus Coccoloba; however, the fruits of the holotype appear to be abnormal, the probable result of an insect attack. Similar abnormal pedicels also occur on the type specimens of the other species which I have examined and which I now consider to be synonymous with C. acapulcensis. Lundell (Lloydia 2: 83. 1939) placed Coccoloba cardiophylla in the synonymy of C. browniana and reported the affinities of the species to be with C. wercklei, of Costa Rica, and C. acapulcensis, of Guerrero, Mexico. In the original description of C. wercklei Standley suggested that his new species is a relative of the Honduran C. browniana but distinct from it in that C. browniana bears pedicels only 5 mm. long. The series of specimens cited, however, shows the full range of variation in the length of the pedicel between flowering and mature fruiting conditions, as well as the differences 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 185 in the pedicel length of the male and female flowers. It is clear to me that all four species must be grouped together under the oldest accepted name, which is C. acapulcensis. México. GUERRERO: Acapulco, MacDaniels 249 (¥), Haenke 1120 (F), 1125 (F), Palmer 399 (us-type of C. acapulcensis; GH, NY); Tecpan de Galiana, Hinton 14118 (¥, GH). YucaTAN: Chichén Itza, Bequaert 28 (A, F), C.L. & A. A. Lundell 7471 (ps, F, MICH), 7513 (A, DS, F, MICH) ; ha span Gaumer & sons 23865 (DS, F, ae 23905 (F); Quintana Roo, C.L. & A. A, Lundell 7644 (micH); Ebtun, C. L.& A. A. Lundell 7534 (A, Ds, F, MICH): without specific location, Gaumer 24013 (F-type of C. cardiophyila.: A, GH). Guatemala. EL PRrocREsso: Barranquilla, Steyermark 46428 (F). HUEHUETENANGO: Between Democracia and canyon of Chamusu, Steyermark 51229 (F); Paso del Boqueron, Steyermark 51152 (F). JuTIAPA: Quebrada above Ovejero, Standley 77647 (F). British Honduras. ORANGE WALK District: Honey Camp, Lundell 514 (a, ps, F, GH). Honduras. ComayaGua: Comayagua, Standley & Chacén 5126 (¥), 5383 (F), 5484 (F), 6012 (F); El Banco, Rodriguez 2374 (¥, GH); Rio Selguapa, Rodriguez 2616 (¥, GH); Siguatepeque, Yuncker, Dawson & Youse 6138 (¥, GH, MICH). Yoro: Coyoles, Yuncker, Koepper & Wagner 8071 (¥F, GH, MICH); Olanchito, Record & Kuylen H-54 (us-type of C. browniana; GH, y). Costa Rica. El Coyolar, Wercklé s.n. (us 865109-type of C. wercklei); San Pedro de San Ramon, Brenes 21872 (F). Coccoloba acuminata HBK. Nov. Gen. 2: 176. 1817. Coccoloba acuminata var. pubescens Lindau, a Jahrb. 13: 193. 1890. Coccoloba acuminata var. glabra Lindau, |.c. Coccoloba strobilulifera Meisner, FI. Beas: Te 25. 1855. Lindau was correct in concluding that Coccoloba strobilulifera Meisner is the same as C. acuminata. I have examined the specimens cited by Meis- ner and Lindau and have also seen this species in the field. On the basis of these observations, I find it impossible to recognize the two varieties that Lindau established on a characteristic of the pubescence. The varia- tion in the amount of pubescence seems to be related to the age and vigor of the plant. Pubescence is present on the young leaves and shoots but is less conspicuous, or the hairs are broken off so that only the clear hair bases remain, on older specimens. Certainly all intermediate stages can be found between the material cited by Lindau as C. acuminata var. pubes- cens and that cited as var. glabra. In addition to the material cited below, I have seen collections from Colombia, Venezuela, British Guiana, Brazil, Ecuador and Peru. The type is Humboldt 1479, collected along the Rio Magdalena, near Mompox, Colombia. Lindau did not select a type for his var. glabra, but the type of C. strobilulifera Meisner is Moritz, without number and without definite locality in Colombia. Lindau places this species in synonymy under his var. glabra. Guatemala. IzABAL: Quebradas, Pittier 8592 (GH); between Bananera and Sioux Station, Steyermark 38986 (F, MICH). Honduras. ATLANTIDA: Near Tela, Mitchell 82 (¥, GH); near Tela, Standley 54710 (a, F); La Fragua, Standley 186 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL 55720 (F); vicinity of Tela, Standley 56867 (F); spear of San Alejo, Standley 7771 (®); between Tela and Lancetilla, Yuncker 4645 (A, F, GH, MICH). CorTEs: La Lima, Johansen 45 (A, F); Rio Piedras near San Pe drs Sula, Molino 3417 (F); La Lima, Standley & Chacén ei (F), Williams & Molina 12485 (¥). Yoro: Near Progresso, W. Dz 1 (&), Record & Kuylen H-44 (cu), Standley 55013 (A, F); Coyoles, Cala Koepper & Wagner 8034 (¥, MICH), 8061 (f, GH, MICH). Department unspecified: Coyol, Carleton 495 (a, GH); Highland Creek, Pto. Sierra, Wilson 84 (F). Nicaragua. ZELAYA: La Esperanza, Rio Grande, Molina 2109 (F), 2136 (¥, GH). Department unspecified: Region of Braggman’s Bluff, Englesing 183 (F, y), 229 (F, Y). Costa Rica. ALAJUELA: Vicinity of Capulin, Rio Grande de Tarcoles, Standley 40189 (¥). diag Canton de Osa, Palmar Norte to Canablancal, Allen 5226 (GH, MICH). Pana CANAL ZONE: Barro Colorado Island, Aviles 74, 976 (F), L.H. & E.Z. Bailey 616 (GH), Bangham 378 (a, F), Starry 223 (F), Wilson 22, 140 (F). Woodworth & Vestal 397 (A, F); Juan Mina, Chagres River, Bartlett & Lasser 16321 (micH); Gamboa Reach, Maggs IJ 39 (F); Frijoles, Maxon 4707 (¥, cH); Gamboa, Pittier 2608 (¥, GH, MO); Upper Chilibre River, Seibert 1510 (mo); Quebrada Bonita, Steyermark & Allen 17194 (micH, MO). DARIEN: Tucuti, Chepigana, MW. E. & R.A. Terry 1381 (A, F, MO); Yape, Allen 850 (¥, GH, MO); location unspecified, MacBride 2676 (F). Province unspecified: Maume & Gor- gone, Wagner s.n. (Mt) ; Marraganti, Williams s.n. (Ny). Location unspecified: Weddell sn. (Gu, P) Coccoloba barbadensis Jacquin, Enum. 36. 1760, Obs. Bot. 1: 18. pl. 8. 1764; Howard, Jour. Arnold Arb. 37: 317-339. 1956. Coccoloba barbadensis var. mexicana Meisner, DC. Prodr. 14: 153. 1856. Coccoloba schiedeana Lindau, Bot. Jahrb. 13: 187. 1890. Coccoloba jurgenseni Lindau, Bot. Jahrb. 13: 188. 1890. Coccoloba leptostachya Bentham, Bot. Sulph. 159. 1856. Coccoloba oaxacensis Gross, Repert. Sp. Nov. 12: 219. 1913. Coccoloba mayana Lundell, Bull. Torrey Club 64: 547. 1937. Coccoloba masoni Lundell, Lloydia 2: 8. 1939. Coccoloba fluviatilis Lundell, Contr. Univ. Mich. Herb. 7: 8. 1942. In the second paper of this series (Jour. Arnold Arb. 37: 317-339, 1956), I considered the correct identification and application of the names Coccoloba swartzii Meisner and C. barbadensis Jacq. The former name had been overlooked by recent workers on the floras of the Antillean Islands and the latter name had been misapplied. I concluded that the material which is properly called C. swartzii was first identified by Lindau as C. barbadensis and that later he and all recent authors on the West Indian flora used the name C. diversifolia Jacq. for the same material. Coccoloba barbadensis Jacq. was described twice by Jacquin, first in his Enumeratio, and later, supported by an illustration, in his Observationum Botanicarum. A definite location was not given for the species and recent authors have assumed from the name that the island of Barbados was im- plied. However, in the British Museum (Natural History) there are two sheets which can be referred to the Jacquin description and illustration and these are significant in the correct application of the name. One of 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 187 these, from the Miller herbarium, is obviously the material illustrated by Jacquin. Moreover, it bears the Houston catalogue reference which Jacquin cites in the synonymy of his Coccoloba barbadensis. Houston collected in peche (Yucatan) and his specimen is accurately matched in many of the collections cited below In his monograph of the genus Coccoloba for De Candolle’s Prodromus, Meisner assumed that C. barbadensis came from the West Indies and so created a variety, C. barbadensis var. mexicana, for a Schiede specimen from Mexico. Lindau (Bot. Jahrb. 13: 187. 1890) recognized the variety as a species which he named Coccoloba schiedeana, citing in synonymy C. barbadensis var. mexicana Meisner. Lindau cited additional material, all of which, with the exception of one collection, I have now seen. These are all clearly comparable to the original Jacquin material; thus, C. bar- badensis Jacq. must henceforth be considered in the flora of Mexico and Central America. Of the two old specimens in the British Museum (Natural History), one represents a sterile adventitious shoot and the other a vigorous shoot with an old inflorescence from which the fruits have fallen. In all probability, therefore, it will be impossible to describe the fruit which accompanied this original sheet. However, with a knowledge of the variation found in other species of Coccoloba in such characteristics as leaf shape, size of petiole. presence or absence of pubescence and length of fruiting pedicels, it is possible to make an accurate comparison with more recent collections. It is clear that C. barbadensis has puberulent to tomentose ocreae and inflorescence rachises, although in age these become glabrous. Also, leaves are coriaceous and the leaf shape, particularly at the base, is variable. As in other species of Coccoloba, the plants are dioecious and the pistillate Howers, and later the fruits, are borne on short pedicels which in length barely exceed the ocreolae. The fruit, coronate at the apex, is rounded at the base and not attenuated into a stipe. The variations with age and environment in these characteristics have caused Coccoloba barbadensis to be described many times. Seven species are assigned to the synonymy of C. barbadensis in this paper alone. Three additional apparently plage eat names have been applied to the speci- mens in various herbaria. The original variety which Meisner described is also placed in synonymy with is species. Three of these newly recognized synonyms represent only sexual variations, i.e., staminate versus pistillate plants, in the comparisons given by the original authors. Two species were originally distinguished through the failure of the author to recognize the loss of, or to see the residual, pubescence. One species was based on an abnormal fascination of the inflorescence and two others were based on size and texture differences of the leaves. When considered alone or with only the material cited by the original authors, several of the species which I have reduced to synonymy might well represent clones or local popula- tions. When considered with the wealth of material cited below, the differ- ences become of little taxonomic value. It appears desirable to treat Cocco- loba barbadensis as a widespread species of central and southern Mexico 188 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL and northern Central America, recognizing the considerable variation in leaf size, shape and texture. Two excellent series of collections by Dr. Robert Dressler made in the states of Veracruz and Guerrero, Mexico, show the variation in leaf-size, -shape and -pubescence to be found on single plants. Coccoloba leptostachya Bentham has been a troublesome name. Bentham described this plant, without citing a collection, in reporting on the Botany of the Sulphur Expedition. The locality was given as Libertad in Colombia, and recent collections from Colombia and Panama have been assigned to this species. However, examination of the type material in the Herbarium at the Royal Botanical Gardens at Kew shows clearly that the specimen on which Bentham based the species is a Barclay collection which repre- sents C. barbadensis. Other species from the Sulphur voyage reportedly collected in ‘Libertad in Colombia” have more recently been recognized as Central American and, in several instances, the collection is believed to have come from Libertad in El Salvador. Coccoloba leptostachya appears to me to be a similar example of the incorrect citation of a locality. The recent collection Carlson 563 (¥) from “Finca Santa Emilia west of La Libertad at or near sea level” in El Salvador is a perfect match for the Barclay specimen. No material comparable to the type of C. leptostachya has been seen from Colombia. In his monograph Lindau distinguishes Coccoloba jurgenseni from C. schiedeana on the basis of a glabrous inflorescence rachis in the former and a pubescent one in the latter. The distinction is inadequate, especially since the holotype of C. jurgenseni, Jurgensen 157 (G) shows a slight puberulence within the range of variation expected in this genus. Coccoloba oaxacensis Gross was reduced to the synynomy of C. schie- deana by Standley in his treatment of the trees and shrubs of Mexico. He reported at the time that he had “seen no material of C. oaxacensis, and it may be a distinct species.” I have examined the holotype in the Berlin herbarium but find no reason to maintain C. oaxacensis as a species. Coccoloba mayana Lundell was based on a staminate flowering specimen collected by Lundell on the bank of the Rio San Pedro de Martir above El Paso in the Department of Petén, Guatemala. Lundell reported that the species was found only in wet soils on river banks and around “aguadas.” Coccoloba mayana, he states, “is related to Coccoloba schie- deana Lindau, from which it differs in having longer racemes, nodes 1- or 2-flowered, larger flowers, and suborbicular perianth lobes.” None of these characteristics is of specific value. At best, C. mayana as conceived by Lundell is an ecological variation, distinct only in having smaller and thinner leaves, but it does not seem to be worth a subspecific category. Coccoloba barbadensis is a variable species and, judging from the collec- tions seen, occurs primarily in drier areas. Plants from other departments in Guatemala show larger leaves of thinner texture and are thus interme- diate between the type collection of C. mayana and C. barbadensis. Coccoloba masoni Lundell was based on a collection by H. L. Mason from Maria Magdalena Island of the Tres Marias Islands off the coast of 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 189 Nayarit, Mexico. Lundell recognized the dioecious condition of the flow- ers, the obovate or oval leaf shape and the pilose pubescence in the axils and on the veins of the leaves as distinguishing characteristics of this species. Two other collections from the same island group, one made earlier and one made later than the Mason specimen, show more clearly the range of variation in leaf-shape and -pubescence and indicate that C. masoni and the Mason collection on which it is based are the same as C. barbadensis. Only a few recent floras have recognized the dioecious nature of Coccoloba. At present it appears that all species are functionally unisexual and the dioecious character is not of specific value in the case of C. masoni. In describing Coccoloba fluviatilis, Lundell placed his new species in a group which included C. jurgenseni and C. mayana, both of which are here included in C. barbadensis. Lundell felt that C. fluviatilis was distinct be- cause of “large glabrous leaves, subcordate at the base, the stout petioles up to 1.7 cm. long, the glabrous rachis of the inflorescences, and the sub- sessile fruits.” However, all of the size and shape characteristics mentioned by Lundell can be found in many of the single collections cited. The “gla- brous” characteristic, however, is an error, for pubescence can be found on the terminal buds and in the axils of the leaves on the holotype, although the plants do become glabrate in the clonal material which Lundell studied. Seler 1642 from Arroyo de San Carlos, in Mexico, was designated by Gross as the type of an apparently unpublished species named for the large, fasciated inflorescence axis. Many of Gross’ new species have been published in obscure papers, so it is possible that this name may be in print. The fasciated inflorescence rachis is relatively common in other species of the genus and has been seen in three additional collections cited below. The Seler collection is clearly identical with the material assigned to this species. The collection by Edward Palmer which was distributed with the letter ‘““G” in place of a collector’s number bears an unpublished name attributed to Rose. This name refers to the sessile flowers, but the collection is not distinct from C. barbadensis. The specimen attributed to Galeotti and numbered 7218 has been cited and annotated by Lindau as C. schiedeana. The material, por is a mixed collection containing in part C. barbadensis and C. veno Mexico. CAMPECHE: Champoton, Flores 1934 (F); Mundo Nuevo, Karwinsky 733b (LE). Cutapas: Escuintla, Matuda 2028 (A, DS, F, MICH); Javalinero, (micH-type of C. fluviatilis; a, F); Chicomuselo, Matuda 15627 (F); Monte- cristo, Matuda 15948 (Fr); Nandolopez, Acapetahua, Matuda 16642 (Br, F); Mojarra, Tonala, Matuda 17139 (F); a Matuda 17517 (¥); Mon- serrate, Purpus 275 (LE, us). Cota: Paso del Rio, Emrick 168 (F); Man- zanillo, Ferris 6209 (os), Palmer “G” (B, GH, NY); Colima, Palmer 90 (a, c H). GUERRERO: La Mina near Atoyac, Dressler 1798a, 1798b, 1798c on Vallecitos, Montes de Oca, Hinton 10209 (ps, MICH), 10221 (ps, LE, MICH); Vallecitos, Llano de Oca, Hinton 11781 (pS, LE, MICH); Sierrita, El Limon, 190 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Galeana, Hinton 14137 (¥, GH); ae Kenoyer C-277 (micH); La Copradilla, Langlassé 134 (B, GH); El Trienta, C.L. & A.A. Lundell 12588 (micH); Dos Arroyos, Nelson 7029 (GH, pee Palmer 344 (A, GH), 602 (A, F, GH, MICH). JaLisco: Navidad, McVaugh 11903 (micH); Tuxpan, Mexia 1051 (A, F, GH). MicnHoacAn: San Naranjillo, Coalcoman, Hinton 13943 (cH); La Placita, Turner 2028 (micH). Nayarit: Acaponeta, Rose 1414 (xy), Rose, Standley & Russell 14399 (ny); Magdalena Island, Tres Marias Islands, Elmore 1B3 (¥. micH), Nelson 4315 (GH), Mason 1806 (F-holotype of C. masoni; A, DS, GH, MICH), Maltby 167 (Ny); San Blas, Ferris 5331 (A, Ds); Ixitlan, Viereck 1193 (us). Oaxaca: Rincon San Antonis: Endlich (s-holotype of C. oaxacensis): Dunes del’Oc. Pacifique, Galeotti 7218, in part (BR, K, P); Pinotepa Nacional, Jurgensen 157 (c-type C. jurgenseni; ¥-photo & fragment; B, K); Between Llano Grande and Pinotepa, Nelson 2334 (GH); Puerto Angel, Reiche 574 (m); Arrovo de San Carlos, Seler 1642 (8, GH); Almoloya, L. Williams 9906 (¥, w). StNatoa: Labradas, Ferris & Mexia 5291 (a, DS). VERACRUZ: Laguna Encantada, east of San Andrés Tuxtla, Dressler 1804a, 1804b, 1804c (cH), Nelson 452 (A); Tantoyuca, Ervendberg 364 (GH, K, Savana de Mata de Don Juan, Karwinsky 734, 735 (LE); Tajin, Papantla. Kelley 65a (GH); Laguna Tamiahua, south of Tampico, LeSueur 100 (£); location un- specified, Liebmann s.n. (B); Mirador, Ross 802 (m); Tierra Blanca. Ross 883 (m); Zacualpan, Purpus 12022 (Fr, MIcH); Rancho Camaroén, Purpus 13068 (A, F, GH, MICH); Fortin, near Zacualpan, Purpus 2428 (A, F, GH): Papantla, Schiede SM. ie type of C. barbadensis var. mexicana and C. schiedeana by im- plication; B). State unspecified: location unspecified, Houston s.n. (BM-type of oH tarbadensi): Sessé & Mocino 950, 5434 (F). temala. EscUINTLA: Naranjo, J. Donnell Smith 2496 (F, GH); San José, ise 63998 (fF); Rio Guacalate, northwest of Escuintla, Standley 89345, 89350 (¥, MICH); Rio Michatoya, southeast of Escuintla, paongrses 89054, 89060 F, MICH). HUEHUETENANGO: Ciénaga de Lagartero below Miramar, ma mark 51488 (A, ¥). PeTEN: La Libertad, Aquilar 275 (F, MICH. W), 313 (A, MICH), Lundell 3230 (¥, GH, MICH), 3212, 3340 (F, MICH); El Paso, pees: 1498 (micu-holotvpe of C. mayana; DS, GH). QUEZALTENANGO: Rio Ocosito, J. Donnell Smith 1481 (¥, GH). RETALHULEU: Champerico, Standley 66593, 66602, 66631 (F); between Nueva Linda and Champerico, Standley 87699, 87721 (¥); Retalhuleu.Standley 88818 (F). SAN Marcos: Ocos, Steyermark 37780 (Ff). SANTA Rosa: Guazacapan, Standley 78603 (¥); Chiquimulilla, Stand- ley 78785, 79175, 79243 (F): La Sepultura, west of ae Standle y 79378 (F). SUCHITEPEQU Ez: South of Tiquisate, Stevermark 47807 Salvador, AHUACHAPAN: Ahuachapan, Standley & Padilla sa _ as 20313 (GH, MICH). LA Lipertap: Libertad, Barclay s.n. (K-ty C. lepto- stachya), Carlson 563 (F), SANTA ANA: Santa Ana. Standley ate (GH). SONSONATE: Acajutla. Standley 21895, 21974 (cH). —" ’ The following collections are also assigned to Coccoloba barbadensis, but probably represent a hybrid complex. The specimens are all staminate, with sterile, insect-infested fruits on tenuous pedicels twice as long as the ocreolae. The leaves appear to be more coriaceous and the margins are revolute. The leaves are shiny on the upper surface, although this may be an artifact of preservation. Guatemala. ALTA VERAPAZ: Laguna Sapala, Stevermark 44899 (Ff). PETEN: 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 191 Laguna Petexbatum, Steyermark 46224 (a, F); Along Rio Santa Monica between Cedral and Ceibal, Steyermark 46040 (¥, MICH), 46160 (F). The series of collections cited below appears to represent a different hybrid population limited to the state of Sinaloa and found only in the coastal areas. The probable parents are Coccoloba barbadensis Jacq. and C. uvifera L., although in general the leaves are smaller than in either of these. Leaf texture, as well as the size and shape of the fruit, indicate a relationship to C. barbadensis. The leaf shape, especially at the base and apex of the blade, and the venation show similarities to C. uvifera. How- ever, in contrast with most of the recognized C. uvifera hybrids from the West Indies and with C. & lundellii of Central America, C. uvifera seems to be the less dominant parent in this hybrid. Fruits of two types have been found, one essentially globose but strongly coronate and the other more or less oblong and obtusely to slightly coronate at the apex. The first tvpe closely approaches the fruit of C. barbadensis and all fruits opened have fully developed embryos and endosperm. The second type shows the influence of C. wvifera and all fruits examined had either rudimentary seed development or were sterile. Mexico. SINALOA: Altata, Rose 1359 (Ny); Culiacan, J. Gonzalez Ortega 6583 (ps, GH, M), Palmer 1518 (GH, NY); Elota, J. Gonzales Ortega 5879 (DS, GH, M); La Concha. Gentry 6804 (¥, GH); Mazatlan, Howell 10559 (a), Rose, "Standley & Russell 14035 (xy); Villa Union, Rose, Standley & Russell 13953 (ny); locality unspecified, J. Gonzalez Ortega 7203 (F). Most of these specimens were identified as Coccoloba goldmanii or C. masoni. The former is a distinct species, but is known from the type and one other collection. It may, in fact, be related here as part of this hybrid population or be one of the parents in place of C. barbadensis. Coccoloba needs a special study in Sinaloa and southern Chihuahua. Coc- coloba masoni Lundell is clearly referable to typical C. barbadensis Jacq. Coccoloba belizensis Standley, Trop. Woods 16: 38. 1928. Coccoloba hirsuta Standley, Publ. Field Mus. Bot. 4: 303. 1929. Coccoloba belizensis Standley is a clearly defined species and one which is easily recognized. The relatively large, globular terminal buds with tawny to chestnut-brown pubescence are distinctive even on sterile or ad- ventitious shoots. The pubescent inflorescence is branched and consists of racemes arranged as a panicle of nearly equal branches. The leaf size varies considerably in fertile material as well as on sterile shoots which are pre- sumably adventitious. Both staminate and pistillate plants are represented in the collections cited. The mature achene is surrounded by a fleshy perianth which seems to be formed by the equal enlargement of the perianth lobes and the hypanthium. In the Berlin Herbarium there is a specimen studied by Gross which bears an unpublished herbarium name honoring the collector, Campbell. Much of Gross’ work on Coccoloba appeared as small notes in miscellaneous 192 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL papers. Thus it is possible that this specific name has been published and, if so, it antedates C. belizensis. Coccoloba hirsuta Standley was based on sterile material. The type, Standley 54802, as well as the cotype 52823, was collected in the Lance- tilla Valley, near Tela, Honduras. Standley noted the undesirability of basing species on sterile material, but felt that this species was easily rec- ognizable by the “copious long pubescence of the leaves.” In their treat- ment of the genus Coccoloba for the Flora of Guatemala (Fieldiana Bot. 24: 114. 1946) Standley and Steyermark continued to recognize C. hirsuta as a distinct species, referring to, but not citing additional specimens from, Guatemala and thus extending the range. I have seen two collections from the Department of Izabal, Stevermark 38185 and Standley 72945, which are among those which must have been seen by Standley and Steyermark. These authors suggest that C. hirsuta “will be found to have panicled racemes, and to be closely related to C. Tuerckheimii.” I am unable to dis- tinguish between material annotated “C. hirsuta” by Standley and Steyer- mark and cited by Standley and that of occasional sterile specimens which the same authors assigned to C. belizensis. Such collections as C. & W. von Hagen 1344 from the Department of Col6n, Honduras, and Lundell 2756 from the Department of Petén, Guatemala, approach C. hirsuta in the amount and type of pubescence. The Von Hagen collection is sterile and was taken from a 75—100-foot tree. This specimen may be an adventi- tious shoot comparable to the type of C. hirsuta, for the terminal bud is the same in size and shape, but there is less copious pubescence on the stem and much less on the leaves. The Lundell collection grades easily into fertile material which can be clearly defined as C. belizensis and which is found in the same area. Guatemala. ALTA VERAPAZ: Chirreacté, Standley 91620, 91672 (F); between Sachaj and Sacacac, Steyermark 45157a (F). IZABAL: between Bananera and La Presa in Montana del Mico, Steyermark 38185, 38186 (F); Escoba, Standley 72945 (F). PeTEN: La Libertad, Lundell 3463 (Fr, MIcH); Monte Santa Teresa, Lundell 2736 (¥), 2756 (micH). British Honduras. CorozaAL District: San Roque, Gentle 554 (Ff); Corozal, Orange Walk Rd., Gentle 4967 (F, MICH); Corozal, Pachacan Rd., Lundell 4784 (a, F, MIcH). Ext Cayo District: San Agustin, Lundell 6812 (¥, GH, MICH). ORANGE WALK District: Honey Camp, Lundell 637 (ps, F, GH, MICH, US). STANN CREEK District: Carib Reserve, Gentle 3088 (miIcH); Mullins River, Gentle 3365, 3369 (A, MICH); Silk Grass Creek Reserve, Gentle 2982 (micH); Stann Creek, Gentle 2964 (A, MICH); Stann Creek Valley, Stevenson 7 (us-holotype; A, F, Y). ToLepo District Jenkins Creek, Gentle 4080 (A, MICH). District unspecified: Belize, Campbell 117 (8); Crique Negra, Stevenson 104 (¥, y¥); All Pines, Schipp 794 (a, F, GH, MICH); Tower Hill Estate, Karling 14 (¥, GH). Honduras. ATLANTIDA: Lance- tilla Valley near Tela, Howard, Briggs, et al. 451 (a), Standley 52823 (Ff), 53555 (A, F), 54802 (F-type of C. hirsuta; A); San Alejo, Standley 7971 (¥); San Juan near Tela, Yuncker 4817 (A, F, MICH); Tela, Standley 53402 (a, F). COLON: Guarunta, C. & W. von Hagen 1344 (Ff). Nicaragua. Eastern Nicaragua, Shank 93 (yx). 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 193 Coccoloba caracasana Meisner, DC. Prod. 14: 157. 1856. Coccoloba caracasana forma glabra Lindau, Bot. Jahrb. 13: 211. 1890. This species was formerly considered to have orbicular leaves, but the large number of specimens examined indicates that a broadly oblong leaf is more characteristic than a strictly orbicular one. Considerable variation in leaf size and shape is found in the species. Leaves of adventitious shoots may have nearly ovate blades up to 34 cm. long and 28 cm. wide on petioles to 4 cm. long. The straw-colored ocreolae are membranaceous, with several occurring at each nodule because of the presence of numerous flowers, several of which are frequently aborted. In fruit the lobes of the perianth expand to cover more than the upper half of the achene. The forma glabra which Lindau described is scarcely worthy of recogni- tion since the density and persistence of the pubescence varies in single collections and on single plants. In addition to the specimens cited below, the species is known from Venezuela and Colombia and possibly farther south along the Andes. The type has been cited both as Humboldt 732 and as Bonpland 732, One col- lection cites only Caracas, Venezuela, as the locality, while another bears a label referring to the ‘‘valley of Araguensibus. México. Curapas: Las Garzas, Acapetagua, Matuda 2677 (A, DS, F, MICH), 2806 (A, MICH); Mapastepec, Matuda 17516 (F ); Tapachula, Matuda 17704 (F). Guatemala. EscurntLa: Iztapa, J. R. Johnston 1171 (F); San José, Stand- ley 64232 (F, MICH), 64019, 64190 (¥). JuTIAPA: Between Trapiche Vargas & Asuncién Mita, Steyermark 31792 (¥). RETALHULEU: Between Nueva Linda and Champerico, Standley 87673 (¥). SANTA Rosa: Santa Rosa, @rsted 668 (m); Capulin, Standley 79592 (F); Los Cerritos region, Capulin, Standley 79607, 79563 (F, MICH). SUCHITEPEQUEZ: Tiquisate, Steyermark 47700 (F). El Salva- dor. La Paz: Zacatecoluca, Calderon 305 (GH). La UNION: Laguna de Maquigue, aden 20977 (GH). San MIGUEL: San Miguel, Standley 21139 (GH); Laguna Jaguay, San Miguel, Fassett 28813 (a); Laguna de Olomega, Standley 20990 (GH). SAN SaLvapor: San Salvador, Renson 139 (Ny), Standley 23611 (¥, GH). SAN VICENTE: San Vicente, Standley 21261 (¥, GH), Standley & Padilla 3702 (fF). Santa Ana: Hacienda la Barra, Curlson 1031 (fF). SONSONATE: San Antonio del Monte, Standley 22171 (GH). Nicaragua. CHINANDEGA: Chichigalpa, Stand- ley 11187, 11334 (F). CHONTALES: Juigalpa, Standley 9302, 9305 (Ff). GRANADA: Granada, Baker 2435 (A, MICH). LEON: La Paz, Baker 2272 (cH). meee unspecified: C. Wright s.n. (GH); Orsted 670, 671 (c). Costa Rica. GUANACASTE: between Bebedero and Taboga, Brenes 12562 (F); Filadelfia, cen 295 (F); Nicoya, Tonduz 13799 (¥, G, GH), 13974 (Gc, GH); Isla de Chira, Valerio 1467 (Ff); Santa Ana de Nicoya, Leén 989 (¥F). Province unspecified: Los Loros, Brenes 22056 (F). Locality unspecified: Hoffmann 292 (8), Warscewicz sn, (B). Panama. CANAL ZoNE: Ancon, Pittier 2730 (us); Gorgas Memorial Laboratory grounds, White 106 (F, GH, Mo); Rio Agua Salud near Frijoles, Piper 5848 (GH); Victoria Fill near Miraflores Locks, Allen 1701 (¥, GH, Mo, Ny). CHRIQUI: Progreso, Cooper & Slater 270 (F). Los Santos: La Jagua, Bartlett & Lasser 16382 (micH). PANAMA: Sabanas Road, Gillespie P-34, P-34a (ps); Chepo, Kluge 25 (F, Ny). 194 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL Coccoloba chiapensis Standley, Proc. Biol. Soc. Wash. 33: 67. 1920. Coccoloba anisophylla Standley, Publ. Field Mus. Bot. 4: 303. 1929, Coccoloba anisophylla was based on sterile material, but Standley re- ported that ‘“‘the form of the leaves is quite different from that of any other Central American species.” On studying the material cited by Standley and being unable to distinguish C. chiapensis from C. anisophylla, I have re- ferred the latter to the synonymy of C. chiapensis. This species is distinct in having leaves which are oblong to elliptic, broadest near the middle, tapering to an acuminate apex and narrowing to an abruptly obtuse and asymmetrical base. Only staminate flowers are known for this species and additional material is needed. Sterile specimens may resemble C. novo- granatensis but can be distinguished by the basal origin of the petiole and by the pubescence. México. site Finca Irlanda, Purpus 7699 (us-holoty Pe of C. chiapensis; B, GH, NY); location unspecified, Purpus 7599 (GH, NY). Honduras. ATLANTIDA: Lancetilla ce near Tela, Standley 53260 (¥- holoty pe of : anisophylla; A), 53364, 54840 (A, F). Coccoloba cozumelensis Hemsley, Biol. Am. Centr. 4: 108. 1887. Coccoloba yucatana Lindau, Bot. Jahrb. 13: 190. 1890, Coccoloba cozumelensis is readily recognized but difficult to define and separate in a key. The species seems amply distinct on the basis of the pale tan color of the leaves which often dry a lighter yellow-brown on the lower surface, and the sessile fruits which are obtuse at the apex with the perianth lobes not at all coronate. Also, the petioles are puberulent and along the midrib and veins on the lower leaf surface there is a characteris- tic crispate pubescence. Some specimens of Coccoloba cozumelensis grade into the Central Amer- ican phase of C. swartzii and it is possible that a hybrid complex exists in coastal areas where these species occur together. Since I have insufficient material to resolve this problem, the key characteristics used here apply to plants represented by the distinctive holotypes of each. Two recent collections from the state of Chiapas are referred here. One collection, Margery C. Carlson 2071, made at Tuxtla Gutiérrez. is anno- tated as a new species by Standley and Williams. The name used refers to the tenuous spike. This collection, in flowering condition, is interme- diate between typical material of C. cozumelensis and specimens referred to C. spicata, but other than the larger leaves of Carlson 2071.1 find no reason for recognizing the collection as a new species. The second collec- tion, Dressler 1408, from El Real, Chiapas, is intermediate between typical C. cozumelensis and the Carlson collection. Standley placed Coccoloba yucatana in the synonymy of C. cozumelensis. I have re-examined authentic material of this and confirm his decision. México. CAMPECHE: Tuxpena, Lundell 853 (A, DS, F, GH, MICH). CHIAPAS: 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 195 Tuxtla Gutiérrez, Carlson 2071 (r, NY); El Real, east of Ocosingo, Dressler 1408 (GH). QuINTANA Roo: Cozumel, Gaumer 18 (B-holotype of C. yucatana; GH, K); Coba, C.L. & A.A. Lundell 7804 (A, DS, F, M, MICH), 7830 (A, DS, F, micu). Tasasco: Reforma, Balancan, Matuda 3173 (a, MicH). British Hon- duras. Betize District: Maskall, Gentle 1068 (A, F, GH, MICH), 1349 (A, F, MICH). CorozAL District: San Antonio, Bartlett 13031 (A, F, MICH), Lundell 4815 (A, F. MICH); Corozal, Gentle 524 (Ff, MIcH). Et Cayo District: Little Mountain Pine Ridge, Bartlett 13060a (micH); Mountain Pine Ridge, San Agustin, Lundell 6648 (pS, F, GH, MICH). ORANGE WALK District: Honey Camp, Mever 50 (F). Guatemala. IzapaL: Rio Dulce. Wilson 387 (fF). PETEN: Carmelita. Egler 42-249 (F); La Libertad, Lundell 3570 (¥, GH, MICH). Coccoloba darienensis, sp. nov. Arbor, 12 m., ramulis teretibus, glabris; ocreis chartaceis, oblique trun- catis. 1 cm. longis, glandulis (?) resinosis, adpressis; petiolis ad basem ocreis gerentibus, teretibus, supra canaliculatis, glabris, 7-11 mm. longis; laminis anguste oblongis vel lanceolati-oblongis, apice acutis vel breviter acuminatis, basi acutis, 10.5 & 3.5 vel 14.5 & 5 cm. longislatisque glabris, tenuiter coriaceis, nerviis primariis 9 vel 10, arcuato-adscendentibus; in- florescentibus racemosis, 8-14 cm. longis, floribus crebris, ncdulis confluen- tibus, pedunculis 1 cm. longis, rachis striatis, puberulentibus; floribus ignotis, bracteis ovatis vel oblongis, apice obtusis, 1 mm. longis, puberulen- tibus vel breviter pilosis, ocreolis membranaceis, raro bracteolisque pu- berulentibus vel apice ciliatis; pedicellis fructiferis 1 mm. longis, glabris; fructu ovoideo basi rotundato, 5-6 mm. longo, 5 mm. diametro, obtuse trigono, hypanthio non manifesto, lobis perianthii ad basem distinctibus, late ovatis. imbricatis, acheniis nitidis, nigris. Panama. DarIEN: Pinogana, P. H. Allen 934 (GH-type: F. MO). This collection had been identified as Coccoloba acuminata, which is cer- tainly a related species. The strongly divided fruiting perianth would allow this species to be placed in Lindau’s section Campderia. Coccoloba darien- ensis is distinct from C. acuminata on the many-flowered inflorescence in which the flowers are confluent, not clustered, and the axis is terete but striate or grooved and not swollen below each nodule. Furthermore, C. darienensis differs in the evident fruiting pedicels. while C. acuminata can be regarded as having only spicate inflorescences without evident pedicels in fruit. Coccoloba diversifolia Jacquin, Enum. Pl. 19. 1760. Coccoloba laurifolia Lundell, Bull. Torrey Club 66: 594. 1939, and authors, yossibly Jacquin, Hort. Schoenbr. 3: 9, ¢. 267. 1798. Coccoloba lancifolia Lundell, Bull. Torrey Club 66: 593. 1939. I have previously discussed the proper application of the epithet Cocco- loba diversifolia Jacq. (Jour. Arnold Arb, 30: 421-424. 1949). All recent publications, however. have used C, laurifolia Jacq. for material more prop- 196 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL erly called C. diversifolia. Coccoloba laurifolia, based on material from Caracas, Venezuela, was described and illustrated by Jacquin, but I have been unable to find any collections from northern South America which exactly duplicate the characteristics which he illustrated. It is probable that C. /aurifolia is the same as C. diversifolia, described by Jacquin twenty- eight years earlier. Oddly enough, no specimens of this which could be attributed to Jacquin have been seen in European herbaria, although a few specimens collected from cultivated plants and labelled “C. laurifolia Jacq.” are present in the herbaria at Berlin, Leningrad and Geneva. How- ever, all of these are true C. diversifolia and fail to compare favorably with the original illustration of C. laurifolia Jacq. Lundell compares Coccoloba lancifolia with the original description of C. laurifolia, reporting that his new species differs in its small, lanceolate and pointed leaves. Coccoloba diversifolia, as its name implies, is a variable species in respect to leaf shape, but in all the characteristics which I can observe, it is the same as C. lancifolia. In their treatment of the genus for Guatemala (Fieldiana Bot. 24: 114. 1946), Standley and Steyermark placed C. lancifolia Lundell in synonymy of C. laurifolia, but without com- ment. I concur with their decision, but call the species C. diversifolia Jacq. The small leaf size and the pointed apex can be found in many populations of C. diversifolia, especially in Cuba, the Bahamas and the Florida Keys. This is the first time that this common West Indian species has been recognized in Central America and Mexico. México. QUINTANA Roo: Lago San José, Frére Arséne s.n. (B). SAN Luts Potosi: Tamazunchale, Edwards 921 (¥). Guatemala. ALTA VERAPAz: Cerro Cinaja, Steyermark 45669 (A, F, MICH). IzABAL: Bay of Santo Tomas, Stever- mark 39232 (¥), 39351 (F, MICH), 39357, 39363 (F). British Honduras. Corozat District: Location unspecified, Gentle 231 (F, MICH, NY); Consejo, Lundell 4945 (¥, micH). District unspecified: Jacinto Hills, Schipp 1200 (micu-holotype of C. lancifolia; a, GH, w). Coccoloba emarginata Jacquin, Enum. Syst. Pl. 37. 1760. This species was described by Jacquin in 1760 and illustrated in his Observationum Botanicarum in 1764 (tab. 9). It is based on a specimen from the herbarium of Mygind but no locality is given. The species was compared with the Hispaniolan Coccoloba leoganensis Jacq., which I have recognized (Jour. Arnold Arb. 39: 28-30. 1958), but which is better known in that flora as C. rotundifolia Meisn., a later homonym. Lindau was un- able to place the Jacquin species and so listed it as of uncertain identity. The situation here is similar to that of Coccoloba barbadensis. The speci- men which Jacquin described must have come from the Yucatan Peninsula of Mexico. The illustration, although of a sterile plant, is an excellent one and is obviously the same as Neomillspaughia emarginata (Gross) Blake (Bull. Torrey Club 48: 84. 1921), the basionym of which is Podopterus emarginatus Gross (Repert. Sp. Nov. 12: 218. 1913). Gross’ species is 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 197 based on Seler 5600 from Yucatan, Mexico. Since the combination has already been made, I can only add Coccoloba emarginata Jacquin to the synonymy of Neomillspaughia emarginata (Gross) Blake and thereby dis- pose of a troublesome and misplaced ancient epithet in the genus Coccoloba. Coccoloba goldmanii Standley, Contr. U.S. Nat. Herb. 23: 245. 1922. This distinctive species from the drier areas of northern Mexico has not been recollected in the past sixty years. Cited below are the Goldman collection made in 1898 and the Hartman collection made in 1891. The latter was misidentified as Coccoloba orizabae. Coccoloba goldmanii is recognized by the nearly orbicular leaves which are persistent short pubes- cent below and which possess a conspicuous network of veins on the lower surface. The flowers are unknown. The ellipsoidal fruits are not coronate and the lobes of the fruiting perianth cover the upper half of the achene, the Vea the lower half. ico. CHIHUAHUA: Apajcachi, Hartman 535 (GH). StnaLoa: Valley of Rio eet. Goldman 245 (us-holotype; GH). Coccoloba hondurensis Lundell, Bull. Torrey Club 66: 591. 1939. Coccoloba marginata J. Donnell Smith, Enum. Plant. Guat. 6: 36. 1903, not Benth. In describing this species, Lundell states, ‘the specimens show consider- able variation, especially in leaf shape, inflorescence, and fruits but all appear to be referable to a single species. . . . Collections of C. hondurensis have been referred to C. Schiedeana Lindau, C. marginata Benth., C. bar- badensis Jacq., and C. leptostachya Benth. I have not been able to asso- ciate the species with any of these.” It is unfortunate that Lundell did not state how he distinguished C. hondurensis from the other species he lists. Standley and Steyermark (Fieldiana Bot. 24: 116. 1946) refer C. hondurensis to the synonymy of C. schiedeana Lindau in their treatment of the genus for Guatemala. I regard C. schiedeana Lindau as identical to C. barbadensis Jacq. I hesitate to follow Standley and Steyermark in reducing C. hondurensis, for there is a problem here which I cannot solve, even with abundant herbarium material. I prefer to keep C. Aondurensis as a distinct species and to call this problem to the attention of field workers for further study. Material cited below as C. hondurensis is easily recognized, but difficult to define or to separate in a key. For the present, C. hondurensis can be recognized when dry by the dark brown or black upper leaf surface which contrasts with the lighter-colored lower surface. The inflorescences are usually twice as long as the subtending leaves. The fruiting pedicels are slightly more tenuous and seem relatively longer than those of C. barbadensis and the apex of the fruit, while coronate, is more obtuse than is general in C. bar- badensis. Usually such distinctions are not valid in this difficult genus and I am maintaining this as a distinct species for temporary convenience only. 198 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL Further study should be rewarding, for C. hondurensis may prove to be either a variety of C. barbadensis or of hybrid origin. México. Oaxaca: Tehuantepec, L. Williams 9906 (vy). Veracruz: Fortuno, Coatzacoalcos River, L. Williams 8701 (vy). Guatemala. ALTA VERAPAZ: Rio Sebol, near Carrizal, Steyermark 45785 (¥). IzaBaL: Quirigué, Standley 23867, 23988, 24547, 24601 (GH), 72249 (F); Puerto ese Standley 72143 (F); Salomén Creek, south of Bananera, Steyermark 38934 (FP). Petén : Rio Can- cuén, paca 45925 (A, F); Rio Machaquilla, north of El Cambio, Steyer- mark 45928 (F), Honduras. ATLANTIDA: Vicinity of Tela, Bangham 212 beg Chickering 233 (F, MICH), Howard, Briggs et al. 452, 503, 504, 505, 506, 507 (A), Salvoza 810 (a), Standley 53716, 54476, 54481, 54752, 50598, ee vi F); La Ceiba, Vuncker, Koepper & Wagner 8295 (Ff, GH). CoLon: Cuyamel, Carleton 598 (us). SANTA BARBARA: San Pedro Su la, J. Donnell Smith 5433 (A, F, GH, M). Dept. unspecified: Puerto Sierra, Wilson 74 (Ny), 562 (¥, GH, NY British Honduras. BeLizE District: Quamina Creek, Manates: Gentle 3437 (A, MICH); Cornhouse Creek, Manatee River, Bartlett 11275 (r, MICH); Sibun River, Bartlett 11361 (cu, MICH): Northern River, Gentle 897 (A, F, MICH). EL Cayo District: Little Cocquericot, Belize Ric. Lundell 3995 (micH), 3996 (mIcH-holotype; F), 3997 (GH, MICH). STANN CREEK District: Sarawee Pine Ridge, Gentle 2699 (a, F, MICH), 2748 (A, F, MICH); Mullins River Road, Gentle 2731A (A, F, MICH); Can Reserve, Gentle 2734 (A, F, MICH), S.m. (A, MICH, WIS); Stann Creek Valley, Gentle 2733 (A, F, MICH ), 2735 (A, F, MICH, WIS), 2737, 2738 (A, F, MICH), s.m. (MICH); Middlesex, Gentle 2820, 2874, 2878, 2887 (A, MICH), 3011, 3018 (A, F, Mic); Silk Grass Creek Reserve, Gentle 2989 (A, MICH); Big Eddy Ridge, Gentle 3348 (a, mic): Mt. Cow Vayyel, Gentle 3538 (a, MicH); Silk Grass Creek Reserve, Record BH-20 (vy); Middle- sex, Schipp 453 (A, F, GH, MICH). ToLEDO Distercr : Swasey Branch, Monkey River, Gentle 3716, 3584 (A, MICH). District unspecified: Manatee Lagoon, Peck 255 (cu). Coccoloba humboldtii Meisner, DC. Prodr. 14: 163. 1856. Coccoloba orizabae Lindau, Bot. Jahrb. 13: 189. 1890, Coccoloba humboldtii var. longipedicellata Gross. Repert. Sp. Nov. 12: 219. 1913 Caries petrophila Brandegee, Univ. Calif. Publ. Bot. 10: 404. 1924. Meisner cited two collections in the original description of this species, Humboldt 4484, which is in the Willdenow Herbarium as No. 7705. and Berlandier 105, collected near Tampico, Although the Humboldt specimen in the Willdenow herbarium at Berlin bears no specific locality beyond “Mexico,” this should be selected as the lectotype. A second specimen col- lected by Humboldt is in the Paris herbarium and bears the number 4484, with the locality given as ‘Vera Cruz.” Unhappily, this specimen was iden- tined and annotated by Lindau as Coccoloba nutans, an entirely different species ( ‘ocraliia orizabae Lindau was distinguished by Lindau in a key to the species on the shape of the leaf blade, especially the narrowed base and the 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 199 acuminate apex. The type of this species was Bourgeau 2822 from Orizaba, Veracruz Gross distinguished his new variety Coccoloba humboldtii var. longi- pedicellata by the longer pedicels and the inflorescence rachis as long as the leaves. Gross failed to appreciate the dioecious character of the species and based his variety on a staminate plant, in contrast to the holotype, which is a pistillate plant. The variety is clearly included within the range of variation of the Humboldt type collection and, as more recent col- lections demonstrate, should be included within the species. Coccoloba petrophila was described by Brandegee in a report on Purpus collections from Mexico. Purpus 8872 from Cameron, Veracruz, was se- lected as the type. No discussion was given of the distinctions or affinities of the species. Another collection, Purpus 8736, made at the same time and location, was not seen, or at least not aed by Brandegee. Subse- quently Purpus collected additional material assigned to this species from near Los Conejos (16425, 16430) and Rancho Remudadero (10967). The last number was suggested to be a wide-leaved variety in Brandegee’s herbarium notes. These Purpus collections, together with Seler 4482, show the extremes of variation to be expected in C. humboldtii, as well as the intermediate forms which made a continuous series from former concepts of C. humboldtii to C. petrophila and C. orizabae. Several of the Purpus collections also show thin-textured immature leaves of varying sizes and shapes. It is in such specimens as Purpus 16425 that transitions to the thin-textured acuminate-tipped leaves of C. orizabae can be found. In his treatment of the trees and shrubs of Mexico, Standley places both Coccoloba orizabae and C. humboldtii in a division of his key headed “rachis of the inflorescence glabrous.” In all of the material cited below, the young inflorescence rachis is subtomentose, but at maturity is pu- berulent or glabrate. The same is true of the ocreae. The original descrip- tions of the three species and the variety fail to mention pubescence. México. SAN Luts Potost: Las Canoas, Pringle 5111 (A, GH). VERACRUZ: Tam- pico, Berlandier 105 (A, B, LE, P), 201 (LE), Palmer 322 (GH, Ny); Veracruz, ge 109 (Br), Juzepczuk 1116 (F, LE), Gorrin sn. (LE); Colipa, Karwinsky . (LE); Laguna Tamiahua, LeSueur 100 (GH), 101 (F); Orizaha, Bourgeau 2822 ee of C. orizabae; P); Panuco River near Ebano, LeSueur 102 (F, GH); Pacho, Liebmann s.n. (LE); Coatzacoalcos, Orcutt 3162 (ps, F, GH); Pueblo Viejo, Palmer 443 (F, G, GH, NY); Camerén, Purpus 8736 (GH, NY, US), 8872 (GH-isotype of C. petrophila) Rancho Remudadero, Purpus 10967 (B, F); a oat 12022, in part (ps); Los Conejos, Purpus 16425 (A, F, LE), 16430 ( - Puerto de Alvarado, Seler 4482 (B-holotype of C. humboldtit var. epebedse aot. cH); Mocambo, Troll 59 (m); Alvarado, Schubert & Rojas 1844a ee location unspecified, Humboldt 4484 (s-Herb. Willd. 7705, lecto- type; Coccoloba lasseri Lundell, Contr. Univ. Mich. Herb. 6: 10. 1941. This species is regarded as endemic to Panama and is perhaps not dis- tinct from Coccoloba tucrckheimii Donn. Sm. For the present, C. lasseri 200 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL may be recognized by the paniculate inflorescence of few branches, the leaves with short, stout petioles and the blades which, while narrowed below the middle, are truncate or abruptly rounded at the base. Finally, the fruits known for C. lasseri are rounded at the base. In contrast, C. tuerckheimii has a paniculate inflorescence of many more branches, leaves with longer, tenuous petioles and blades cuneate at the base with fruits narrowed to a short stipe at the base. Panama. Cocié: Aguadulce, Pittier 4989 (r): Penonomé, Williams 357 (NY). Herrera: Chitré, Allen 1111 (r, cu, MO); Mangle Bejuco, Steyermark, Allen & Dodge, s.n. (F). PANAMA: Bejuco, Allen 2542 (a, F); Panama Nat. Highway, Las Lajas Bridge, Bartlett & Lasser 16638 (micH): Rio Pacora, Bartlett & Lasser 16946 (micH-holotype); Balboa to Chamé, Dodge, Hunter, Steyermark & Allen 16732 (mo): Punta Paitilla. Standley 26265 (A). Province unspecified: Duchassaing s.n. (P). Coccoloba liebmannii Lindau, Bot. Jahrb. 13: 189. 1890. Coccoloba lapathifolia Standley, Contr. U.S. Nat. Herb. 23: 245. 1922. This species is quickly and easily recognized by the persistent pubes- cence on the lower leaf surface. Most of the leaves show a tendency toward an undulate or wavy margin to the blade, while in some specimens this tendency is so exaggerated that the marginal folds overlap in the dried specimens. The fruit is globose and the perianth lobes are imbricate in fruit. The achene is not coronate. Standley distinguished his new species, Coccoloba lapathifolia, from C. liebmannii on the basis of a “glabrous or very minutely puberulent” inflorescence rachis, in contrast to one densely short-pilose. Examination of the material cited below, however, including both types, indicates quite clearly that C. lapathifolia is the same as C. liebmannii and should be placed in synonymy. México. CoLtima: Manzanillo, Ferris 6061 (A, ps, F), Palmer 965 (GH, NY); between Galera and Pochutla, Liebmann s.n. (c-holotype of C. liebmannii; B). GUERRERO: Acapulco, Palmer 206 (us-holotype of C. lapathifolia; GH); Placeres, Hinton 9101 (Ny). Oaxaca: Distrito de Tamiltepec, Conzatti 4406 (GH); Llano Grande & Pinotepa, Nelson 2341 (cH): Cerro de Picacho, Purpus 7735 (GH, NY); Tequisistlan, Seler 1719 (B). Coccoloba lehmannii Lindau, Bot. Jahrb. 20 (Beibl. 49) : 7. 1895. Uvifera lehmanni Lindau, in footnote, l.c. Coccoloba lehmanni Lindau, Repert. Sp. Nov. 1: 156. 1905. Coccoloba changuinolana Standley, Publ. Field Mus. Bot. 8: 9. 1930. Coccoloba williamsii Standley, Publ. Field Mus. Bot. 11: 148-149. 1936. Coccoloba allenii Lundell, Contr. Univ. Mich. Herb. 6: 8. 1941. Lindau described this species in two papers. In the earlier publication the specific name is spelled “/ehmannii” and in the later one, “lehmanni.” Two collections (Lehmann 6109 and 7560) were cited in the first publica- tion, but in the second description only Lehmann 7560 was cited. In both 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 201 descriptions this new species is compared with Coccoloba gracilis of the section Campderia. Lindau’s two descriptions of Lehmann 7560 are com- parable; thus one may assume that Lindau overlooked the first publica- tion which described both of these collections. It seems desirable to select Lehmann 7560 in the Berlin Botanic Garden Herbarium as the lectotype. Coccoloba changuinolana was based on three collections from Panama with Dunlap 499 being selected as the type. An examination of these specimens shows Standley’s original description to be in error on the im- portant character of “‘perianthii tubo accrescente inclusus.” The perianth lobes, not the perianth tube, increase in size to surround the mature achene, and the species as represented by the collections cited by Standley belongs to the section Campderia. Although Standley stated that the new species was related to C. schiedeana, I can find no basis for comparison between the two. However, C. changuinolana is comparable in all char- acteristics with C. lehmannii, from Columbia, and is assigned in synonymy to that species. I have seen only the type collection, Llewellyn Williams 4803, of the several specimens that Standley cited in his original description of C. williamsti. This Peruvian species must also be placed in the synonymy of C. lehmannii. Williams 4803 is in fruiting condition and the longer fruiting pedicels are in contrast to the shorter pedicels found on the flowering specimens of the Lehmann collection which is the type of the species. Additional collections which I have seen are intermediate. Stand- ley indicated a relationship of C. williamsii with C. caracasana, both species being assigned to the section Campderia, but these are clearly distinct in the leaf shape and nature as well as in the abundant pubescence. Still a third species, Coccoloba allenit Lundell is assigned to synonymy here. Lundell correctly noted the alliance of C. allenii with C. williamsit, but distinguished C. allenii on the basis of a furfuraceous-lepidote indu- ment and large apiculate spathaceous ochreolae. Additional specimens col- lected by Allen in Panama show many intermediate conditions in both characteristics. Coccoloba lehmannii is currently known to occur in Columbia (type), Panama, Peru, Costa Rica and Venezuela. Costa Rica. La Palma, Suxaola Valley, Dunlap 467 (¥, y). Panama. Bocas DEL Toro: Daytonia farm, Region of Almirante, Proctor Cooper 421 . a Changuinola Valley, Dunlap 499 (r-type of C. changuinolana; B, Y). : El Valle, Allen 2181 (micH-type of C. alleni; a, F), Allen & Alston oe (F, GH, MO). Coccoloba lindaviana, sp. nov. rutex, 8 m., ramulis teretibus, leviter striatis, glabris; ocreis coriaceis, dense ae vel glabratis, 9-12 mm. longis; netiolis supra basem ocre- arum gerentibus, 8-11 mm. longis, glabris; laminis oblongis, anguste oblongis vel oblongi-ellipticis, apice acutis vel breviter acuminatis, basi rotundatis, vel truncatis, 15 & 4.5 vel 22 x 8.5 cm. longislatisque, cori- 202 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XL aceis, glabris, nerviis primariis 11-15, adscendentibus; inflorescentiis ter- minalibus, racemosis 12—15 cm. longis, rachis glabris, bracteis triangulari- bus, minutis, minusque 0.5 mm. longis, glabris, ocreolis membranaceis, glabris vel ciliatis ad marginem, bracteis aequaliter; floribus ignotis; pedi- cellis fructiferis 1.5-2 mm. longis, glabris; fructu ovato, 8 mm. longo, 6 mm. diametro, ad apicem obtuse coronato; acheniis castaneis, levibus, nitidis. Honduras. Yoro: Aguan river valley, vicinity of Coyoles, Yuncker, Koepper & Wagner 8032 (cH-holotype; F); Cortés, canyon of Rio Piedras, 3 km. from San Pedro Sula, Molina 3430 (¥); El Encanto in Montana La Cumbre, Molina 3577 (F); Montana de Rio Piedras, Molina 3555 (¥). The Molina collections bear an unpublished name attributed to Stand- ley and L. Williams. I agree with them that a new taxon is represented here, even though this species is currently known only from old and fruit- ing specimens. Coccoloba lindaviana is, therefore, distinct on the basis of the leaf shape, the position of the leaves, which are borne above the base of the ocreae, and by the oval fruit which is slightly coronate at the apex. It is also similar to C. hondurensis, differing in the venation of the leaves and the position of the petiole base. All of the fruits I have examined had fully-formed seeds. Nevertheless, there is a possibility of a hybrid origin of this species. The species is named in honor of Gustav Lindau who published a mono- graph of the genus Coccoloba in 1890. Coccoloba lindeniana (Bentham) Lindau, Bot. Jahrb. 13: 82. 1890. Campderia lindeniana Bentham, in Benth. & Hook. Gen. Pl. 3: 103. 1880. This species is known only from the type collection made by J. Linden (1602) in Tabasco, near Teapa, Mexico, in May, 1840. An altitude of 300 m. is given on the collector’s label. Specimens have been seen in the herbaria at Geneva, Leningrad and Paris, as well as a fragment in the Berlin herbarium. Coccoloba lindeniana seems distinctive among Mexican and Central American species in the very pubescent leaves and inflorescence axis and the pubescence of the bracts and ocreolae. It may prove to be nothing more than the pubescent form of C. acuminata which Lindau described as C. acuminata var. pubescens. Coccoloba acuminata is widely distributed through Central America and adequately represented in herbaria, although only a few collections are available from Guatemala. The pubescent phase, which I do not recognize as a distinct taxon, is found throughout the range of the species. The abundance and persistence of the pubescence varies, The inflorescence of Coccoloba acuminata is characteristically inter- rupted, with the axis swollen below each flower cluster. The clusters there- fore appear at an acute angle to the inflorescence axis. In C. lindeniana 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 203 the axis is not swollen and the flower clusters appear to be produced flat on the axis. The final position of Coccoloba lindeniana will await the collection of more material from Mexico. It is to be hoped that some of this will be in fruiting condition. Coccoloba & lundellii Standley (pro sp.) See lundellii Standley, Publ. Field Mus. Bot. 8: 138. 1930. loba suborbicularis Lundell, Lloydia 2: 84. 1939. In the original description of Coccoloba lundellii Standley stated. “Evidently a relative of the widespread C. schiedeana Lindau, but dif- fering conspicuously in its very broad leaves, somewhat suggestive of those of the common sea grape, C. uvifera.” A few years later Lundell described Coccoloba suborbicularis and re- ported that ‘among the species of the Yucatan Peninsula, C. suborbicularis is unique in having the leaf conspicuously unequal at the base.” The type of Coccoloba lundellii bears a few fruits, two of which were opened and found to be sterile. The type specimen of C. suborbicularis bears a few staminate flowers. Both holotypes show many of the character- istics of C. uvifera. The few fruits, the clustered flowers, the puberulence and the asymmetrical leaf base are characteristics of C. uvifera. The broad, orbicular leaves are dominant on all specimens seen, but a few smaller oblong leaves, with a strong development of secondary veins when dry, are suggestive of some other influence. Several hybrids of Coccoloba uvifera and neighboring species have been found or suggested in the flora of the West Indies and it appears that a parallel situation is present here. It is not clear from the material on hand whether the second parent is C. hondurensis, which may, in turn, be a variation of C. barbadensis, or C. reflexiflora. Both species are known from the area of the hybrid. British Honduras. ORANGE WALK District: Honey Camp, coastal region, Lundell 649 (¥-holotype; DS, GH). STANN CREEK District: Stann Creek Rail- way, Gentle 2687 ie holotype of C. suborbicularis; A, F); Stann Creek Valley. Gentle 2716 (a, MICH); Middlesex, Gentle 2877 (A, MICH), 2890 (A, MICH). (To be concluded) JOURNAL OF THE ARNOLD ARBORETUM VoL. XL JuLy 1959 NUMBER 3 STUDIES IN THE GENUS COCCOLOBA, VII. A SYNOPSIS AND KEY TO THE SPECIES IN MEXICO AND CENTRAL AMERICA * RicHaArp A. HowArpD Coccoloba macrophylla Sessé & Mocino, Fl. Mex. ed. 2. 96. 1894. Although Coccoloba macrophylla appears to be published as a new epithet in the “Flora Mexicana,” the name is a later homonym of C. macro- phylla Desf. ex Hook. (see Jour. Arnold Arb. 38: 225-227. 1957), a syno- nym of C. rugosa Desf. There are no specimens cited by number in the work of Sessé and Mocifio and all of their collections I have seen which bear the name are to be referred to C. rugosa. Coccoloba rugosa is endemic to Puerto Rico and the species is attributed to Mexico in error. Coccoloba manzanillensis Beurling, Prim. Fl. Portobello in Kongl. Vetensk. Akad. Hand]. 142. 1854; Lindau, Bot. Jahrb. 13: 209. 1890. Campderia nematostachya Grisebach, Bonplandia 6: 4. 1858. Coccoloba nematostachya Lindau, Bot. Jahrb. 13: 208. 1890. The original description of Coccoloba manzanillensis cited Billberg 234 as the type collection and Lindau reported seeing specimens from the Berlin herbarium and from Stockholm. The fragment at Berlin which Lindau annotated is Billberg 230, although the Stockholm specimen, presumably the origin of the fragment, is correctly numbered Billberg 234. The col- lection was made on the island of Manzanilla near the port which Lindau called Puerto Belo, known on modern maps as Porto Bello. The name of this species was originally spelled “manzinellensis’” by Beurling, but the spelling was changed to ‘‘manzanillensis” by Lindau. While the island where Billberg collected the type specimen is not on modern maps, Beurling’s intention in describing the species is obvious. “Manzanillo,” or “little apple’ in Spanish, is commonly applied to the widespread and infamous strand plant Hippomane manchineel and although Lindau failed to comment on his correction of an orthographic error or to cite the original spelling, it seems desirable to allow this correction to stand. * Continued from vol. XL, p. 203. 206 JOURNAL OF THE ARNOLD ARBORETUM [ VOL, XL Campderia nematostachya Griseb., transferred to Coccoloba by Lindau, is based on an unnumbered Andersson specimen from Panama. A specimen from the Meisner herbarium, now at the New York Botanical Garden, appears to be identical with the Gottingen specimen and part of the Andersson collection. A label bears Grisebach’s handwriting, but unfortu- nately it carries the legend, ‘“Coccoloba fagifolia Guadeloupe? coll. Duchas- saing.” The label is obviously incorrect. Lindau recognized both Coccoloba nematostachya and C. manzanillensts. He distinguished them by placing C. manzanillensis in a group of species having glabrous leaves, while associating C. nematostachya with species having persistent pubescence. However, examination of the specimens cited shows the distinction which Lindau established to be untenable, since it is based on the age of the specimen. Coccoloba nematostachya, therefore, must be considered to be the same as C. manzanillensis. A long foliar ocrea distinguishes this species which is further characterized by having long, almost silky, hairs. The Shattuck specimen cited carries a plant of Oryctanthus cordifolius as a parasite. Bailey 42 is the first mature specimen with fruit. ma. CANAL ZONE: Barro Colorado Island, L.H. & E.Z Bailey 42 (¥), Shattuck 425 (F); between Gatun and Lion Hill, Pittier 2570 (Ny). COoLon: Porto Bello, Billberg 230 (8). DARIEN: La Palma, Pittier 6599 (GH). PANAMA: Pacora, Allen 3450 (A, BRUX, F, MO). Province unspecified: location unspecified, Andersson s.n. (GOET-type of C. oe B, NY), Billberg 234 (¥-photo; s). Coccoloba matudai Lundell, Contr. Univ. Mich. Herb. 7: 8. 1942. Further study of populations in the field may prove this species to be a hybrid, possibly of Coccoloba belizensis and C. montana. However, the col- lections assigned here show an even greater range of leaf size, shape and texture than is normally found in either species. Lundell described both staminate and pistillate flowers, the former apparently from Matuda 2002 and the latter from Matuda 4315. No fruits were formed. The species at present is recognizable only through the character of the branched inflores- cence México. CHIAPAS: Finca Olvido, Mapastepec, Matuda 2002 (A, MICH, NY, us); Saxchanal, Sierra oni Matuda 4315 (micu-holotype; A); Cascada, Siltepec, Matuda 5144 (¥, GH) Coccoloba montana Standley, Jour. Wash. Acad. 13: 368. 1923. Coccoloba escuintlensis Lundell, Phytologia 1: 213. 1937. Coccoloba schippii Lundell, Bull. Torrey Club. 66: 594. 1939. Coccoloba steyermarkii Standley, Publ. Field Mus. Bot. 22: 138. 1940. It is unfortunate that the oldest name to be applied to this taxon is based on a sterile, fast-growing shoot. In describing Coccoloba montana Standley suggested that the ‘leaves of this Coccoloba are so distinct from 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 207 those of other Central American species that it seems desirable to give it a name for purposes of reference.”’ In the Flora of Guatemala (Fieldiana Bot. 24: 115. 1946) Standley and Steyermark suggested that the material represented is from “C. escuintlensis or perhaps of one of the other species listed.” The specimens cited below include sterile shoots, as well as fertile specimens, but unfortunately no collections of both sterile, fast-growing shoots and fertile branches have been made from a single plant. Nevertheless, it appears certain that only one species is represented here. A pubescence in the axils of the veins and along the midrib appears in some of the specimens cited, but the leaves are usually glabrous when mature. Coccoloba schippii was described by Lundell and the suggested relation- ship was with C. guvanensis Meisner. This was apparently made on the basis of the sessile flowers which were in bud, appearing in a “‘spicate” in- florescence. In their key to the Guatemalan species, Standley and Steyer- mark distinguished C. schippii on the characters of sessile flowers, con- spicuous nerves and short racemes, but the type of C. schippii is obvi- ously a branch with an immature inflorescence. Comparable inflorescence branches can be found on many of the other specimens cited. I believe the mature inflorescence of C. schippii as defined by Lundell would be elongated and the flowers would be borne on longer pedicels. Coccoloba schippii, with C. escuintlensis, seems properly referred to the synonymy of C. montana. In describing Coccoloba steyvermarkii Standley noted “the species is re- markable for its narrow leaves, more elongate and narrower than those of any other Central American species except C. acuminata HBK., to which it is not closely related. It is similar to C. escuintlensis Lundell |C. montana], of the Pacific slope of Guatemala and of Chiapas, but that has broader leaves, almost concolorous and with much fewer nerves.” I have examined sufficient material of C. montana to be certain that the type speci- men of C. steyvermarkii can safely be included in C. montana. The fe specimens which have been assigned to C. stevermarkii differ at most in a slightly more coriaceous texture to the leaves and less venation. In all other characteristics I have been able to compare, these same specimens agree well with material here considered as C. montana. Coccoloba montana appears to be similar to C. padiformis which was de- scribed from South America and is represented by recent collections from Panama and Costa Rica. In all specimens seen the apex of the leaf is acute or acuminate, while C. padiformis has a rounded or obtuse leaf apex. México. CHIAPAS: Escuintla, Matuda 413 (micH-holotype of C. escuintlensis; A, F, GH); Acacoyagua, Matuda 17410 (F); Esperanza, sagan Matuda 4143, 4145 (a, GH, MICH), 17466, s.n. (F). British Honduras. ToLepo District: at the British Honduras Guatenals boundary, Schipp 5-687 (F-holotype of C. schippit). Guatemala. EscuINTLA: Rio Guacalate, Standley 60190, 89302 (¥); between Rio Jute and Rio Pantaléon, Standley 63464 (F, MICH), 63488 (F); a Lajas, Standley 64787 (¥). IzABaL: Puerto Barrios, Standley 25075 (us). QUEZALTE- 208 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL NANGO: Colomba, Skutch 2022 (A, F); Finca ‘‘Pireneos” below Santa Maria de Jess, Standley 68207 (F); between Finca “Pirineos” and Patzulin, Standley 86954 (F), 86964 (mIcH); between Santa Maria de Jesus and Calahuaché, Steyermark 33677 (¥); Calahuaché, Steyermark 35473 (); Rio Dulce, west of Livingston, Steyermark 39533 (-type of C. steyermarkii), 39404 (F). RETAL- HULEU: between Retalhuleu and Nueva Linda, Standley 87219 (F, us), 87291, 88516 (F); Ajaxa, east of Santa Cruz Mulua, Standley 88220, 88232 (¥F); vicinity of Retalhuleu, Standley 88604, 88807 (¥, MICH); between Asintal and Colomba, Standley 87876 (®). SAN Marcos: Rio Mopa, below Rodeo, Standley 68770 F); Finca Vergel, near Rodeo, Standley 68949, 68952 (F); Finca El Porvenir, Steyermark 37 yah (F, MICH), 52336 (F). SUCHITEPEQUEZ: Pueblo Nuevo, Standley 66936 (¥ El Salvador. haere Sierra de Apaneca, region of Finca Colima, Stand- ley 20061 (us-holotype of C. montana; GH). LA LIBERTAD: near Comasagua between Finca Germania and Finca San Antonio, Carlson 238 (F). Coccoloba nicaraguensis Standley & L. Williams, Ceiba 3: 198. 1952. This species is based on material which is inadequate for an accurate analysis. All the specimens available to me were collected in the mountains at an altitude of 1050-1350 m. west of Jinotega, Department of Jinotega, Nicaragua. The four numbers cited in the original description are Stand- ley 9768, 10194, 10233 and 10409, the last of which is the holotype in the herbarium of the Chicago Natural History Museum. A fifth number, Standley 10318, was also referred to this species by the original authors by annotation, but was not cited in their paper. The type, Standley 10409, and one other, Standley 9768, bear young developing leaves which have turned black on drying. Another collection, Standley 10233, bears both mature leaves as well as young blackish leaves comparable to the type. The other sheets are of mature branches. Only the type sheet bears flowers and this specimen has four immature inflorescences. On the basis of the sterile specimens with mature foliage, the material would be referred un- hesitatingly to Coccoloba diversifolia, for it is strictly comparable to ma- terial described by Lundell as C. dancifolia and later referred by Standley and Steyermark to C. laurifolia, a species which is here called C. diversifolia. However, as indicated in the original description, the young inflorescence axis is minutely puberulent when seen under a microscope. In this char- acteristic the species is similar to C. cozumelensis. Steyermark 45669, from Cerro Chinaja, Department of Alta Verapaz, Guatemala, referred to C. diversifolia (C. laurifolia Standley & Stevermark), is an accurate matca except for a glabrous inflorescence rachis. Coccoloba novogranatensis Lindau, Bot. Jahrb. 13: 192. 1890. Coccoloba caribaea Urban, Symb. Ant. 5: 337. 1907. Coccoloba waittii Johnston, Sargentia 8: 122. 1949. The typification of this species has been discussed in an earlier paper where the lectotype was designated as Triana 978 (Pp), Bogota, Colombia (Jour. Arnold Arb. 40: 85-87. 1959). For over sixty years no collections 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 209 have been assigned to Coccoloba novogranatensis. This is a wide-ranging species found at lower elevations and along the seacoast. The sole exception is R. S. Williams 945, a collection of very young material from an altitude of 2800 feet in Panama. Additional material from this area may prove this to be C. obovata, for in both species the leaves are attached above the base of the ocrea. In C. novogranatensis the leaf blades are broadest above the middle, while those of C. obovata are broadest below the middle. Leaves of adventitious shoots are nearly indistinguishable. The species are clearly distinct, however, in the stages with mature flowers or fruit. The Steyer- mark collection from Guatemala, also sterile, is referred here, again with some question. Johnston does not make any specific comparison in describing Coccoloba waittii, noting only that “it is a well-marked species.” Although all the specimens I have seen from Panama are from the Pacific side of the Isthmus, C. novogranatensis is to be expected to be on the Atlantic coast, as well. This species is known from the Lesser Antilles, Trinidad, Venezuela, Colombia and Peru, in addition to the specimens cited below from Guate- mala and Panama. Guatemala. PETEN: Between Cerro Ceibal and Ceibal, Steyermark 46178 (F). Panama. HERRERA: Pesé, Allen 802 (F, GH, MO, NY). PANAMA: Perlas Islands, Pedro Gonzalez, Allen 2602 (Mo); San José Island, Erlandson 377 (GH), 550 (GH, NY), 216 (GH); Harlow 41 (GH); Johnston 213 (GH, MO), 230 (GH-type of C. waittii; MO), 975, 1306 (GH). DariEN: Cana, Williams 945 (Ny, US). Coccoloba obovata HBK. Nov. Gen. 2: 176. 1817. Coccoloba coriacea Willdenow ex Lindau, Bot. Jahrb. 13: 194. 1890, not Sagra. Coccoloba goudotiana Weddell, Ann. Sci. Nat. Bot. III. 13: 260. 1849. Coccoloba riparia Lundell, Contr. Univ. Mich. Herb. 6: 11. 1951. Coccoloba obovata HBK. was based on a Humboldt specimen collected in flower in June, 1805, at Honda in Colombia. Although the specimen in the Willdenow Herbarium has been considered authentic, I wish to designate a sheet in the herbarium at Paris as the lectotype since the latter sheet is more complete, in better condition and bears a label with more adequate data than does the specimen in Berlin. The name Coccoloba coriacea, attributed to Willdenow by Lindau (not C. coriacea Sagra), was not published by Willdenow but occurs on the label of the lectotype as the first two words of a long polynomial written in one line. On the sheet which Lindau saw in the Willdenow herbarium the label is smaller and the name ‘“‘Coccoloba coriacea’’ appears as one complete line, the two words being underlined. It is unfortunate that Lindau chose to pay any attention to this binomial which has no standing. Lundell described Coccoloba riparia from Panama but did not discuss the species or indicate affinities or differences. Coccoloba riparia as repre- sented by the type specimen is clearly the same as C. obovata. The majority of the specimens cited below from Central America had been identified either as Coccoloba marginata or C. leptostachya. I con- 210 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL sider the former species to be the same as C. nitida and C. guianensis from Trinidad and northeastern South America. The latter species, C. lepto- stachya, is referred to the synonymy of C. barbadensis. Coccoloba obovata is known from Costa Rica, Panama and Colombia. The type is Humboldt s.n. from Honda, province of Tolima, Colombia. The type of Coccoloba goudotiana is Goudot s.n. from San Luis in Colombia. Costa Rica. Peninsula Osa near Puerto Jimenez, Brenes 12218 (Fr); Cufodontis 91 (#); San Rafael de = Ramon, Brenes 22020 (¥, NY); Santo Domingo de Golfo Dulce, Tonduz 7122 (¥, GH, US), 9934 (BR, LE, M). Panama. CANAL ZONE: Barro Colorado Island, Aviles 998 (Ff), Salvoza 984 (a), Shattuck 1063 (¥), Starry 298 (¥); F. L. Island, Bangham 593 (A, F); Sala- manca Hydrographic Station, Dodge, Steyermark & Allen 16983 (BR, DS, 6, K, MO), 16983a (BR, K, MICH, MO); Woodson, Allen & - 1572 (A, F, MO); northwestern part of Canal Zone, Johnston 1519 (mo); between Tumba Vieja & Salamanca, Steyermark & Allen 16758 (BR, MO). pasos Cerro Galera Chorcha, Gualaca, Allen 5020 (A, Mo); San Félix, Allen 3655 (aA, BR, MO); be- tween San Félix and Cerro Flor, Allen 1917 (¥, GH, MO); Progreso, Cooper & Slater 270 (us); San Bartolomé, Woodson & Schery 947 (¥, MO); between Rio Chiriqui and Remedios, Woodson, Allen & Seibert 1182 (A, ¥, MO). COCLE: El Valle, Allen 2229 (micH-holotype of C. riparia). Coccoloba padiformis Meisner, DC. Prodr. 14: 166. 1856. Coccoloba roseiflora Standley & L. Williams in Allen, Preliminary Index to the Trees of the Golfito-Palmar area, 37. 1952; Rain Forests of Golfo Dulce, 177. 1956 Coccoloba padiformis Meisner is very similar to C. densifrons Martius and perhaps is not distinct. At the present time the distinguishing charac- teristics of C. densifrons are the larger leaves, broadest above the middle, and the longer inflorescence rachis. Lindau did not include C. densifrons in his key; hence it is difficult to determine his basis of separation. Coccoloba candolleana Meisner is likewise similar, if not identical, and further material is needed for study before drawing the final limits of any of these species. Certainly the specimens of ‘““Goudot 4” cited by Lindau in extending the range of C. candolleana from Bahia, Brazil, to Colombia are better referred to C. padiformis. The type of C. candolleana, Blanchet 1818, has a nearly orbicular and umbonate leaf blade. If this is true of the species and not just a growth form, C. candolleana can be maintained as a distinct species. Coccoloba roseiflora, credited to Standley and L. Williams by Allen in the latter’s book, ‘‘The Rain Forests of Golfo Dulce” is not validly pub- lished, since it lacks a description in Latin. The holotype in the herbarium of the Escuela Agricola Panamericana at Tegucigalpa cannot be sent on loan, but I have seen duplicates of the two collections cited by Allen and on the basis of these, refer the species to C. padiformis. Allen 5964 consists of branches with young leaves indicating a completely deciduous 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII raul plant. Mature foliage of this species is represented le Allen 5944 and the Tonduz collection from the same area of Costa Rica. The species is currently known from Costa ci Panama, Venezuela and Colombia. The type is Moritz 377, collected near Caracas in Venezuela. Costa Rica. PUNTARENAS: Palmar Norte de Osa, Allen 5944 (F-isotype of - roseiflora), 5964 (F); Santo Domingo de Golfo Dulce, Tonduz 7119 (A, F, H)<7120 (A, ¥,.6H). nama. CANAL ZONE: South of Fort Sherman, Johnston 1737, 1814 (mo); Quebrada Ancha, oo & Allen 17101 (mMiIcH, MO); Rio Indio de Gatun, Pittier 2785 (GH Coccoloba parimensis Bentham in Hooker, Lond. Jour. Bot. 4: 626. 1845 Coccoloba parimensis var. schomburgkii Meisner, Fl. Bras. 5(1): 35. 1855. Coccoloba excelsa var. glabra Lindau, Bot. Jahrb. 13: 171. 1890. Coccoloba bracteolosa Meisner, Fl. Bras. 5(1): 30. 1855. Coccoloba paraensis Meisner, Fl. Bras., 5(1): 38. 1855. The habit of this plant has accounted in part for the confusion surround- ing its description in botanical literature. With regard to the collections studied here, the specimens have been described as a liana, a slender climber, a small tree, a tree, or a tree with scrambling branches. As for the varying leaf aspects which have been described, many of which are cited below, there is no question in my mind that they can be associated with the growth habits of the plant. Other species in northern South America also show the habit of a woody, tree-like plant with scrambling branches. The branches are often so thin and tenuous as to be liana-like in character. Collections made from older plants or those growing in isolated situations show a tree-like habit. Collections from the ends of branches or from plants growing in thickets are likely to be considered liana-like. Most of the specimens I have seen are in fruiting condition. One speci- men, Maxon & Harvey 6700 from Panama, is in flower and has been cited by Lundell as Coccoloba bracteolosa Meisn. (Contr. Univ. Mich. Herb. 6: 9. 1941). The type specimen of C. bracteolosa was a staminate flowering specimen and all specimens I have seen referred to this species have likewise been in flowering condition. The same is true for C. pari- mensis. Coccoloba paraensis, however, was based on fruiting material and only fruiting material has been referred to this species since. The numerous specimens from Barro Colorado island were all collected near the laboratory and are either sterile or in fruit. Nevertheless, the series is elaborate and complete, showing the variation in leaf shape, size and texture between the flowering material called Coccoloba bracteolosa and the fruiting material called C. paraensis. The thin-textured leaves associated with the flowering condition indicate a tendency towards de- ciduousness in this species. The mature foliage associated with the fruit- ing specimens is coriaceous and the leaf blades are umbonate or bullate. Although this evidence is convincing from the specimens on hand, Cocco- 212 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL loba bracteolosa was based on specimens from Bahia and Alagoas on the east coast of Brazil. It is possible that further collections from that area will show that C. bracteolosa is better assigned to the synonymy of C. ochreolata. The latter species is distinctive in the shape of the fruit. Coccoloba parimensis was based on a Schomburgk collection, without number, from Rio Parime. Lindau assigned this species to the synonymy of C. excelsa. I cannot agree completely with this conclusion and am recog- nizing these as distinct species to call attention to the problem. Field study and mass collections from one plant are needed. For the present, C. excelsa is distinct in being more pubescent on the lower leaf surface and on the inflorescence axis. Coccoloba excelsa also appears to have larger hairs which break off and leave clear hair bases or black-colored hair bases, either of which frequently appear as punctations. Coccoloba parimensis leaves, in contrast, are only puberulent with minute hairs. When Lindau assigned C. parimensis to the synonymy of C. excelsa he recognized a glabrous state which he called C. excelsa var. glabra. Lindau also recog- nized C. paraensis and distinguished between this species and C. excelsa on a key character of pubescence being present in C. excelsa and absent in C. paraensis. Lindau’s C. excelsa var. glabra, therefore, becomes an anom- aly in his key. Meisner divided Bentham’s Coccoloba parimensis, recognizing two vari- eties. Coccoloba parimensis var. schomburgkii contained the type and is assigned here to C. parimensis. Meisner’s second variety, C. parimensis var. hostmanni, is better referred to C. excelsa. Coccoloba paraensis Meisner was based on an unnumbered Martius specimen from Iquapémirim in the Province of Para and a Spruce collec- tion, also without number, from Barra in the province of Rio Negro. The latter collection has been distributed with labels written “Coccoloba /2/” and printed ‘In vicinibus Barra, Prov. Rio Negro, coll. R. Spruce, Dec.— Mart. 1850-51.’ The Spruce specimens appear in many herbaria and all are unidentified. No type was designated by Meisner but the Field Museum has distributed photographs of the Martius collection from the Munich herbarium and indicated this as the lectotype. I have seen material of this species from Brazil, British Guiana, Colom- bia, Peru and Panama. Panama. CANAL ZONE: Barro Colorado Island, Aviles 15 (¥), L.H. & E.Z. Bailey 219 (Ff), 654 (GH), Bangham 447 (a, F), Salvoza 904 (A), Shattuck 1123 (F), Standley 41099 (a), Wetmore & Woodworth 859 (a). PANAMA: Juan Diaz region, near Tapia river, Maxon & Harvey 6700 (F). Coccoloba pubescens Linnaeus, Syst. Nat. ed. 10. 1007. 1759. This is a West Indian species which I have seen, studied in the field, collected many times and discussed in a previous paper (Jour. Arnold Arb. 38: 229-231. 1957). Lindau described the species (Bot. Jahrb. 13: 202. 1890), citing, in addition to West Indian collections, a Schiede speci- men from “herb. Petrop.” credited to Mexico without specific location i959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 213 and a specimen from “‘Guyana batava” which Meisner also cited. I have not seen Kegel 1339 which Meisner cited from Dutch Guiana, nor was it present in the Prodromus Herbarium at Geneva. Coccoloba pubescens was also credited to the Mexican flora by Standley in his ‘‘Trees and Shrubs of Mexico” (page 245), although he uses a later homonym, C. grandifolia Jacq., for the species. No recent collections are cited by Standley and none is available to me. This species is conspicuous wherever it occurs and it is difficult to believe that it would be overlooked or not be collected in a flora where sterile material has been the basis of many species. I have on loan three sheets credited to Schiede from Mexico and iden- tified as Coccoloba pubescens. Two of these sheets, one from the Berlin herbarium and one from Paris, were not cited by Lindau and do not bear his annotation label. These sheets bear a number “60” in the same script as the generic name, although the specific name and the collector and locality of “Mexico” are in a different hand and possibly may be different from each other. The sheet from the herbarium at Leningrad bears Lin- dau’s annotation label, as well as a label indicating that the specimen was from Fischer’s herbarium and was collected at “la Cuesta grande de Chi- conquraos.”” There is no indication of the collector (i.e., Schiede), and “Mexico” has been added in a different script. Schiede’s travels in Mexico have been described in Linnaea, volumes four and five, in 1829 and 1830. There are several references to Coccoloba. One is referred to as a seacoast plant known as “uva”’ and this is probably C. uvifera. The other two refer to plants collected between Rio de Tecoluta and Rio de Nantla and at Cambre del Obispo. Neither of these locations fits the data cited above for Schiede 60. Lindau refers to three unnumbered Schiede collections, but all are cited by him as having come from “Mexico” without further locality. All three specimens called “C. pubescens” consist of a single leaf and a short piece of branch with a terminal bud. The leaves are large (30 28 cm. and 38 & 32 cm.) and the branch stout (10 mm. diameter), sug- gesting an adventitious shoot or one of vigorous growth. The leaves are suggestive of Coccoloba pubescens and the pubescence distribution and hair type can be matched in recent West Indian collections. However, the terminal bud is not that of the West Indian C. pubescens but is that of the Mexican C. liebmannii Lindau. Only one collection of C. liebmannit (Ferris 6061 from the vicinity of Manzanillo, Colima, Mexico) is available showing the large leaves of vigorous and adventitious shoots. In the three examples of Ferris 6061 cited elsewhere in this paper, the specimen con- sists of a fertile branch in fruit with small leaves (average 9 * 3.5 cm.) and a single detached larger leaf (23 & 13 cm). No stem section is avail- able for these larger leaves. The pubescence is similar to that of the Schiede specimens; the terminal bud is the same, though the leaves are smaller. The Fischer herbarium, particularly in reference to Coccoloba specimens, is primarily of fragments from cultivated plants; adventitious shoots and associated leaves are common. It seems likely that the Schiede specimen 214 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL cited by Lindau is from a plant cultivated in a greenhouse and that the country of origin given on the label is faulty. However, if the data with the specimen are correct, it is probable that this specimen should be iden- tified as C. liebmannii and the record of C. pubescens be removed from the Mexican flora. Additional field study of C. liebmannii is needed to deter- mine variation in size and shape of the leaves and the terminal buds of vigorous and adventitious shoots, Coccoloba reflexiflora Standley, Publ. Field Mus. Bot. 4: 203. 1929. Coccoloba colonensis Lundell, Contr. Univ. Mich. Herb. 6: 9. 1941. This species is readily distinguished by the basically obovate-oblong leaves, broadest above the middle, which are borne on pale grayish petioles which, in turn, are borne above the bases of the ocreae. The margin of the leaf is invariably revolute and the upper surface is darker than the lower surface upon drying. The inflorescence is rarely longer than the leaves and the staminate flow- ers appear to be reflexed, a character on which Standley based the specific name. However, the fruiting pedicels are at right angles to the inflores- cence axis and are 1—1.5 mm. long, exceeding the ocreolae. The fruits are smaller than those of other Central American species and the perianth lobes are imbricate, not coronate, over the obtuse apex of the achene. Coccoloba colonensis was based by Lundell on sterile material. He sug- gested an affinity with C. undellii, which I have concluded is a hybrid of C. uvifera and an unknown parent, possibly C. reflexiflora. The type of C. colonensis is from an adventitious shoot. In the characteristics of the ocreae and the point of attachment of the petiole, as well as in the color and venation of the leaf blade, the specimen is clearly identical with com- parable vegetable adventitious shoots of C. reflexiflora. México. CAMPECHE: Villahermosa, Lundell 1139 (¥, MICH). YUCATAN: Chichén Itza, Steere 1600 (¥, MICH). Guatemala. PETEN: Between Uaxactun & San Clemente, Bartlett 12815 @, MICH); Carmelita, Egler 42-244 (F). British Honduras. BeLizE District: Northern River, ni tle 990 (A, F, LE, MICH); Maskall, Gentle 1223 (A, F, GH, MICH), 1274 (A, DS, F, GH, MICH), O’Neill 8557 (A, F, MICH); Sibun Road, southwest of Belize, 0 Neill 8570 (Mich), Corocar District: San rare apres S-614 (¥, GH). ANGE WALK District: Honey Camp, Lundell 488 (a, F), Meyer 45 (F); La Winzerling I-17 (vy). STANN CREEK District: Silk ‘Grass Creek Reserve, Gentle 2959 (A, MICH). TOLEDG District: Monkey River, rien — (A, MICH). District unspecified: Tower Hill, Karling 15 (r-holotype; us); Manatee Lagoon, Peck 92 (GH). Honduras. CoLon: Guarunta, Wispernini C amp, C. & W. von Hagen (ny-isotype of C. colonensis). Coccoloba spicata Lundell, Bull. Torrey Club 66: 594. 1939. “Coccoloba coronata Jacquin?.” Millspaugh, Publ. Field Mus. Bot. 1: 294. 1896 Lundell compares this species with Coccoloba mayana Lundell which I have placed in synonymy with C. barbadensis Jacq. The specimens cited 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 215 below, with one exception, are those also cited by Lundell in the original publication. Coccoloba spicata was based on staminate flowering material. Only one fruiting collection is known. The species appears distinct in the sessile flowers and densely flowered inflorescence. In fruit the pedicels are scarcely developed and, if present, are less than one-fourth the length of the ocreolae. In contrast to C. barbadensis, the apex of the fruit is obtuse to rounded and not at all coronate. The perianth lobes are orbicular and represent the upper third of the fruiting perianth. In sterile condition it is almost impossible to distinguish C. spicata from C. barbadensis. México. QUINTANA Roo: Coba, Lundell 7813 (A, F, MICH). YUCATAN: Chichén Itza, Bequaert 103 (A, F); Valladolid, Flores 1935 (F); Kancabdzonot, Gaumer & sons 23894 (A, F); Izabal, Gaumer 819 (¥, GH); Location unspecified, Gaumer 23976 (¥, GH, MICH), Stegeerda 1B (f); Chichén Itza, Sacred Cenote, C.A. & A.A. Lundell 7325 (micH-holotype; A, F). British Honduras. Corozat District: Santa Rita, Lundell 4951 (A, F, GH, MICH). ORANGE WALK District: Honey Camp, Lundell 560 (DS, F, GH). Coccoloba standleyana Allen, The Rain Forests of Golfo Dulce 177, 409. 1956 The holotype of this species, Allen 6645, is in the herbarium of the Escuela Agricola Panamericana, Tegucigalpa, Honduras. Since the school does not send specimens on loan, I have been unable to study the material; duplicates are not available in other institutions. Coccoloba standleyana is based on sterile material collected in the Esquinas forests of Costa Rica, at an elevation of 200 feet. Allen suggests that the species is ‘‘probably related to C. hirsuta Standl. and C. belizensis Standl. but amply distinct in the much smaller ocreae and other characters.” I regard C. Airsuta as the same as C. belizensis and I suspect that Cocco- loba standlevana will ultimately be assigned to the synonymy of C. belizen- SiS. Coccoloba swartzii Meisner, DC. Prodr. 14: 159. 1856. Coccoloba corozalensis Lundell, Bull. Torrey Club 66: 588. 1939. Coccoloba gentlei Lundell, Bull. Torrey Club 66: 591. 1939. Coccoloba swartzii has not been recognized previously in the flora of Central America, although it is a common element in the vegetation of the West Indies, occurring from Cuba and Jamaica to Curacao and Aruba and having a great variation in leaf size and shape on all the islands of the Lesser Antilles. The collections cited can be compared favorably with many other collections from the West Indies, especially those populations occurring in the Virgin Islands, the Leeward Islands and the French islands of Guadeloupe and Martinique. I am not satisfied that the Central American specimens have character- istics of sufficient value to create even a geographical variety. Only the 216 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL slightly more tenuous inflorescence axis, with its fewer flowers, appears to differ from the West Indian representatives of the species. The type of Coccoloba gentlei Lundell is Percy Gentle 56. The meager material appears to be from a vigorous shoot, perhaps even an adventitious shoot. If considered as the latter, the leaves are comparable to collections I have made on St. Kitts and Montserrat (R.A. & E.S. Howard 11882, 11867, 11938). In describing the species Lundell put undue emphasis on the branched inflorescence which in reality is a spike with one anomalous branch developing from the base. This is not an unusual condition either in the genus, or in the West Indian specimens of C. swartziit, Standley and Steyermark (Fieldiana Bot. 24: 108. 1946) place this species in a section of their key with other species having much-branched panicles. Coccoloba corozalensis Lundell is clearly the same as C. swartzit. British Honduras. BELIzE District: Belize-Sibun Road, Gentle 56 (MICH- holotype of C. gentlei); Pine Ridge, Maskall, Gentle 1113 (A, F, GH, MICH). CorozaL District: Xiabe, Lundell 4908 (MicH- type of C. corozalensis; A, GH). ORANGE WALK District: Honey Camp, Lundell 347 (Ds, F). TOLEDO D1stRICT: Monkey Ridge, Gentle 3651, 3656 (A, MICH). District unspecified: Cornhouse Creek, Bartlett 11281 (mtcH, NY); Mullins River at Stann Creek Road, Gentle 3359, 3368 (A, MICH). Honduras. Point Triumfo, Puerto Sierra, Wilson 33 (NY). Coccoloba tuerckheimii Donnell Smith, Bot. Gaz. 37: 213. 1904. Coccoloba guatemalensis hort. ex Lindau, Bot. Jahrb. 13: 226. 1890. Coccoloba latifolia Goyena, Flora Nicarag. 2: 707. 1911, not Lam. According to the collectors’ labels, several botanists have believed Cocco- loba tuerckheimii to be similar to, if not identical with, C. latifolia Lamarck from Trinidad and northern South America, although John Donnell Smith made no comparison of these two species in his original description. In a footnote to his monograph of the genus, Lindau assigned cultivated material carrying the unpublished horticultural name Coccoloba guatemalensis to C. latifolia. I have seen this material which has been attributed to Guate- mala and believe it is more correctly referred to the present species than to C. latifolia. Coccoloba latifolia is similar to C. tuerckheimii in having stout, generally hollow stems, petioles borne slightly above the base of the ocrea, and a pa- niculate inflorescence. The leaf base is generally rounded or cordate, while the blades are bullate between the veins. In contrast, C. tuerckheimii has solid stems with leaf blades tapering to a cuneate base and appearing to be flat. The fruits of C. tuerckheimii appear to be larger and more stalked than those of C. latifolia. Guatemala. ALTA VERAPAZ: Finca Santa Inés, Record & Kuylen G-92 (vy); Tucurt, Standley 70724 (¥); between Campur and Socoy6, Standley 91708 (¥F); Cubilgiiitz, Stevermark 44674 (F), Tuerckheim 8493 (¥-isotype; GH); Cerro, Chinaja, Steyermark 45568 (¥). IzaBaL: Rio Dulce, Livingston, Steyermark 39552 (F). Honduras. ATLANTIDA: Lancetilla Valley near Tela, Standley 55812 (A, F); 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 217 La Ceiba, Yuncker, Koepper & Wagner 8603 (¥, GH, MICH); Tiquitapa, Howard, Briggs, et al. 560 (A). Yoro: Progreso, Hottle 92 (F); Subirana, C. & W. von Hagen 1058 (¥). Nicaragua. CHONTALES: La Libertad, Standley 9064 (F). Costa Rica. ALAJUELA: San Carlos, Suere, A. Smith H-1694 (F, MICH). Panama. Bocas DEL Toro: Changuinola Valley, Island of Potrero, Dunlap 181 (¥, Y). Coton: Dos Bocas, Rio Fato Valley, Pittier 4197 (F). Cultivated. Kalbreyer sn. as C. guatemalensis (B). Coccoloba umbilicata Sessé & Mocino, Fl. Mex. ed. 2. 96. 1894. No specimens are cited in the original description in which the species is attributed to Mexico. I have seen the Madrid specimens on loan to the Field Museum, and the sheets labelled ‘“‘C. umbilicata” are all to be referred to Coccoloba pyrifolia Desf., a species limited to Puerto Rico. Coccoloba uvifera Linnaeus, Syst. Nat. ed. 10. 1007. 1759. Polygonum uvifera Linnaeus, Sp. Pl. 365. 1753. A common strand plant known to all botanical collectors and, as a con- sequence, one which is poorly represented in herbaria. Presumably the species has a wider distribution in Mexico and Central America than is indicated by the collections cited below México. CAMPECHE: Lerma, Seler 4019 (a, F); Champoton, Steere 1968 (F, MICH). SINALOA: Mazatlan, J. Gonzalez Ortega 7203 (BRUX). TAMAULIPAS: Moron, LeSueur 103 (F); Tampico, Palmer 477 (F). YucaTAN: Chichancanab, Gaumer 1897 (F); Sisal, Gaumer 23235 (F); Cozumel Island, Goldman 673 (F) aetna 1590 (F); Progreso, Goldman 603 (Fr); C.L. & A.A. Lundell 8060 (M : Location unspecified, Gaumer 638 (F), Millspaugh 6 (F). Guatemala: ak "Bay of Santo Tomas, Steyermark 39248 (¥, MICH). British Honduras. Betize District: Turneffe Island, Egler 42-42 (F); Belize, Gentle 27 (A, F MICH). CorozaL District: Corozal, Lundell 4927 (F, MICH). STANN CREEK District: Stann Creek, Gentle 3119 (micH), 3120 (MICH), 3131 (A, MICH), 3123 (micH). District unspecified: All Pines, Schipp 796 (A, F, MICH). El Salvador. La LipertabD: La Libertad, M.C. Carlson 575 (Ff). SAN SALVADOR: San Salvador, Calderén 2624 (¥). Honduras. ATLANTIDA: Tela, Chickering 51 (micH), Stand- ley 53056 (A, F), Yuncker 4655 (A, F, MICH); Ceiba, Yuncker, Koepper & Wagner 8235 (¥, MICH). CoMAYAGUA: Comayagua, Standley & Chacoén 5486 (F). Department unspecified: Puerto Sierra, P. Wilson 23 (Ff). Swan ISLANDs: Nelson 22 (GH). Nicaragua. ZELAYA: Bahia de Bluefields, Rio Escondito, A. Molina 1281 (¥). Costa Rica. ALAJUELA: San Roque de Barbe, Ledn 10 (F). MON: Limon, Costa de Portete, Quirds 508 (F); Talamanca, Puerto Viejo, Tonduz 9423 (BRUX). PUNTARENAS: Los Loros and Tivives, Brenes 22661 (F). Panama. Bocas DEL Toro: Almirante. G. Proctor Cooper 558 (¥, Y); Chan- guinola Valley, Cooper & Slater 82 (vy); Chiriqui Lagoon, Von Wedel 2826 (mo). Province unspecified: Isthmus of Chagres, Fendler 287 (LE, MO). Coccoloba venosa Linnaeus, Syst. Nat. ed. 10. 1007. 1759. Campderia floribunda Bentham, Bot. Sulphur 159. tab. 52. 1846. Campderia mexicana Meisner, DC. Prodr. 14: 171. 1856. Coccoloba alagoensis Weddell var. major Meisner, DC. Prodr. 14: 163: 1856. — 218 JOURNAL OF THE ARNOLD ARBORETUM [VOL, XL Coccoloba floribunda Lindau, Bot. Jahrb. 13: 217. 1890. Coccoloba molinae Standley & L. Williams, Ceiba 3: 198. 1952. The synonymy given for Coccoloba venosa is new, C. floribunda and C. molinae and Campderia mexicana being reduced here for the first time. An examination of the type specimens alone would make this decision a ques- tionable one, except for C. molinae which is clearly the typical West Indian expression of C. venosa. Standley and Williams described C. molinae in 1952 but were incorrect in believing that their new species was unlike any material then known from southern Central America. In his Flora Nicaragiiense (2: 707. 1911), M. Ramirez Goyena recognized Cocco- loba nivea from Nicaragua and gave the same common name reported for C. molinae. Coccoloba nivea Jacq. is a synonym of Coccoloba venosa L. The specimens which Standley and Williams cite for C. molinae are ob- viously variable, since two phases of coriaceous leaves are shown even in the two specimens of the type number which I have seen. Coccoloba molinae is undoubtedly the same as C. venosa L. Coccoloba floribunda described by Bentham was based on material col- lected on the voyage of the Sulphur. The type came from Tiger Island in the Gulf of Fonseca, Honduras. Campderia mexicana was based on Andrieux 115, a specimen from Tehuantepec, Oaxaca, Mexico, and since there is no difference between these, Lindau placed Campderia mexicana in synonymy. Recent collections from the same area are also cited. Coccoloba floribunda as defined by Lindau and as generally recognized in current floras is strikingly different from C. venosa as generally recog- nized in the West Indies. In drier areas the leaves are thick and almost rigidly coriaceous. The leaf apex is obtuse, rounded or even truncate and the base slightly narrowed but rounded or obtuse. Bentham described the leaf base as emarginate to slightly peltate, although the type specimen and the illustration fail to show the latter condition. The leaves are much smaller and the inflorescence generally shorter and thicker than C. venosa. At greater altitudes and in wetter areas, according to the data on the speci- mens studied, the differences are less obvious between C. floribunda and C. venosa. ‘The leaves become thinner, longer, the apex pointed, the in- florescence axis longer and more tenuous in comparison. A full range of variation between the extremes has been found in the specimens cited below to indicate a variable population over the geographic range of the species. I have been unable to find any stable characteristics which would allow the segregation of the smaller, coriaceous-leaved form as a geographic variety. Lindau separated C. floribunda from C. venosa (his C. excoriata) in a key to the species (Bot. Jahrb. 13: 130. 1890) by placing C. floribunda in a group in which the leaves are more or less pubescent below or at least barbellate in the vein axils, while C. venosa reportedly had glabrous leaves. Comparable pubescence along the midrib and in the axils of the veins is found in the West Indian specimens, so Lindau’s distinction does not hold. Additional study of this species in the field is needed to establish leaf shape, size and texture variations on single plants in Central America, espe- 1959] HOWARD, STUDIES IN THE GENUS COCCOLOBA, VII 219 cially since none of the collections I have seen represents adventitious or fast-growing shoots. In their treatment of Coccoloba floribunda for the Flora of Guatemala (Fieldiana Bot. 24: 113. 1946), Standley and Steyer- mark described the species as ‘‘a densely branched shrub or tree, some- times 9 meters high, with a broad spreading crown, the low trunk often gnarled and twisted, sometimes a meter in diameter.” This description fits well many of the plants seen in the Antilles. Variations in leaf texture from membranaceous and chartaceous to coriaceous have also been en- countered in C. venosa in the Lesser Antilles, although a reduction in leaf size or an adjustment of shape comparable to that found in many of the specimens cited has not been encountered Among the specimens cited is Standley 74249 from Guatemala, repre- sented by two sheets from two different locations. Coccoloba venosa is under cultivation in Cuba and has been reported from Jamaica in old literature, but no specimens referable to the island have been seen. It is common in Hispaniola, Puerto Rico and the Virgin Islands south through the Lesser Antilles to Trinidad. I have also seen two collections from Venezuela. México. Curapas: Belem, Mapastepec, Matuda 16741 (¥, GH); Tapachula, Fisher 35437 (¥, NY); Huixtle, Matuda 16795 (F); Las Garzas, Acapetagua, Matuda 2683 (A, F, MIcH); Mojarra, Tonala, Matuda 17125 (¥). Couta: Manzanillo, Ferris 6057 (ps), 6245 (a, Ds). GUERRERO: Acapulco, MacDaniels 9 (F, MICH). JALIsco: Mazatlan, Liebmann s.n. (B). Oaxaca: Salina Cruz, Deam 121 (GH, Micu), L. J. Williams 9708 (¥, MicH); Tehuantepec, Zuccarini SM, (ut). Province unspecified: Haenke 1608 (F); Sessé & Mocino 5432, Guatemala. CHIQUIMULA: Quebrada Shusho above Chiquimula, Standley 74249 (F), GUATEMALA: Location unspecified, Aguilar 550 (¥). JuTIAPA: Ju- bee Standley 75193 (#, MICH). RETALHULEU: Champerico, Bernoulli & Cario 663 (LE), Standley 66507 (A, ¥), 87547 (F, MicH). SANTA Rosa: between oe & El Ahumado, Standley 79541 (F). Zacapa: Zacapa, oe 74249 (¥); 74379 (Ff, MiIcH); Rio Motagua west of Teculutan, Steyer 29195, 42167 (¥). Department unspecified: Location unspecified, Friedrichsthal sm. (MICH-type of C. alagoensis var. major; F). E] Salvador. La Lizertap: La Libertad, Standley 23231 (GH, us), Calderén 2399 (F). LA UNION: La Union, Grant 722 (a, F), Standley 20644 (cx). SAN MIGuEL: Laguna de Olomega, Standley 21015 (GH); San Miguel, Standley 21133 (GH). SAN SALVADOR: Hacienda Santo Tomas, Carlson 1124 (F). D eee Standley & Padilla 3297 (¥). SONSONATE: Acajutla, Calderén 1662 , ; Izalco, Standley a (GH, Us). Department unspecified: location unset Choussy 2054 (vu . VALLE: San en Rodriguez 3299 (¥); Isla Tigre, near Amapala, oo. 20728 (GH, US); El Tigre volcano above Amapala, West 3534 (cH); Jicaro Galan, Williams & Molina 15002 (¥F); Nacaome, Williams & Molina 16721 Nicaragua. Chinandega: Corinto, Standley 11545 (F). CHONTALES: Juigalpa, Standley 9220 (F). GRANADA: Granada, Baker 203 (ps, cH), Lévy 1352 (F). MANAGUA: Managua, Garnier 320 (w); Sierras de Managua, Grant 1097 (A, F). 220 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL ZELAYA: Rio Grande, Molina 2291 (¥, GH); Toumarin, Rio Grande, Molina 2436 (¥-isotype; GH); El Recreo on Rio Mico, Standley 19600 (F). Department unspecified: Asseradores Island, Baker 81 (ps, GH); Location unspecified, Lévy 1103 (8), Wright s.n. (GH, NY). Costa Rica. GUANACASTE: Libano, Standley & eons 44900 (fF). NICOYA: Nicoya, Tonduz 13779 (F, GH, MICH). PUNTARENAS: Los Loros, Brenes 22470 (F); between Los Loros & Tivives, Brenes 22680 ee Caldera, Echeverria 4155 (Ff); Isla de Chira, Orozco 233 (F); Valerio 1466 (F). 1959 | STEEVES & BARGHOORN, POLLEN OF EPHEDRA 221 THE POLLEN OF EPHEDRA MARGARET WOLFE STEEVES AND ELSo S. BARGHOORN THE ORIGIN OF THE GNETALES and their relationship to other fossil and living gymnosperms is a problem which is nearly as baffling as that of the origin of the flowering plants; and indeed the two problems have often been linked. On the study of the Gnetales, paleontology has contributed very little because of the extreme paucity of fossil remains (Pearson, 1929 Consequently any early fossil material of this group, and particularly of the most primitive member, Ephedra, may be regarded with great interest by students of plant evolution. The first reference to fossil remains of this gnetalian genus is found in Die im Bernstein befindlichen organischen Reste der Vorwelt, a description of the Oligocene Baltic Amber flora (Berendt and Goppert, 1845). In their study, twigs bearing ovuliferous structures were designated Ephedrites johnianus. A translation of their Latin description is as follows: ~— Stem jointed, joints cylindrical with longitudinal striations ending in an annular sheat Aments subglobose, single, composed of imbricated bracts, and elevated on the apex of thickened peduncles. The peduncles are verticillate, in the axis of the sheath and non-articulate. As a result of Menge’s subsequent discovery of a preserved male strobilus from the same flora and further study of Géppert and Menge’s material, the close similarity between the fossil remains and modern Ephedra became apparent. For this reason the genus was redescribed as Ephedra (Goppert, 1853; Goppert and Menge, 1883). Later workers have described Ephedra fragments from the Miocene of Germany and Switzerland (Unger, 1851, 1870), and the Miocene of France (Saporta, 1889, cited by Wodehouse, 1934) but there has been some dispute concerning the validity of these identifications. Heer (1855), in publishing a report of Ephedrites sotzkianus Unger in his Flora Ter- tiaria Helvetiae, conceded that there was a superficial similarity between these articulate, striated, fossil stems and those of Casuarina, an isolated, Australian genus of angiosperms. However, Schimper (1890) saw no simi- larity with the genus Casuarina but questioned all previous reports of the genus from the Tertiary on the grounds that they could belong to poorly preserved specimens of either Equisetum or the conifer Callitris. As a result of this confusion it was not until the application of pollen analysis to older sediments that remains of Ephedra were uncontestably demonstrated in the Tertiary. Furthermore, with the discovery of Ephedra pollen by Wodehouse from the Green River Formation (1933), the past geographic distribution of the genus was broadened to include North America. Likewise, its geologic record was extended back to the Eocene. 222 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL Although Tertiary and Pleistocene records in the Western Hemisphere have become more complete with the discoveries of megascopic remains of Ephedra from the Florissant beds of Colorado (Wodehouse, 1934), the Pleistocene of California (Axelrod, 1937), the Pleistocene of Nevada and Arizona (Laudermilk and Munz, 1934, 1938), micropaleontological evi- dence has been surprisingly rare both in Europe and the United States. Thiergart (1940) reported the presence of fossil Ephedra pollen grains from Tertiary beds in Germany, Cookson (1956) has recorded them from Australian Tertiary sediments, and, as a result of recent pollen studies, it has been shown that Ephedra persisted as a late- and postglacial desert steppe plant in protected areas in Europe (Christensen, 1949, cited by Iver- sen, 1954; Lang, 1951; Iversen, 1954; Welten, 1957) as well as in South America (Auer, 1933) and the United States (Anderson, 1954). In view of the facts that Ephedra possesses a high degree of structural aden. and also that early Tertiary Ephedra exhibits pollen struc- tures essentially comparable with that of modern species it is apparent that the genus or its close ancestral type had an extended pre-Tertiary history. An extension of this pre-Tertiary history has been found during an analysis of a series of cores from the Cretaceous Raritan and Magothy formations from east-central Long Island, New York. In these sediments a number of intact, as well as fragmented, pollen grains assignable to Ephedra were observed. This discovery extends the known fossil record of the genus in North America back to the lower Upper Cretaceous period. The first part of this paper will deal with a description of the fossil grains found on Long Island; in the second part the results of a detailed survey of the pollen morphology of the living members of the genus will be presented with a view to interpreting the fossil forms and their phylo- genetic significance. FOSSIL EPHEDRA FROM CRETACEOUS SEDIMENTS OF LONG ISLAND In the process of exploratory sub-surface geologic studies at the Brook- haven National Laboratory, two deep well cores were recovered from two borings approximately one mile apart. The cores extended from the sur- face Pleistocene drift through the Cretaceous and into the underlying bedrock. The cores were recovered nearly intact and represent an un- usually complete section of the Cretaceous formations underlying eastern Long Island. One of us had the opportunity to sample both of these cores throughout, and this study comprises one aspect of an intensive micro- paleontological study still in progress. The cores, designated Well No. 1 (S 6409) and Well No. 2 (S 6434) were drilled to a depth of 1,568 feet and 1,294 feet, respectively. In both ‘Samples of a cores were made available in May 1950 by the Ground Waters Division of the U. S. Geological Survey and were transmitted through the courtesy of the Paeontoly and Stratigraphy Branch of the Geologic Division of the U. S Ge siogical Surve 1959] STEEVES & BARGHOORN, POLLEN OF EPHEDRA oe cores, which were drilled through the Magothy and Raritan formations, the sediments recovered were all non-marine, Coastal Plain deposits, compris- ing variegated and lignitic clays, white, fine-grained and coarse arkosic sands, and gravels. Occasional lignite beds were dispersed throughout the section. The entire sequence rested on crystalline igneous rocks of pre- sumed pre-Cambrian age. The Magothy formation is regarded as of Upper Cretaceous age and has been correlated with the Austin chalk of the Gulf Coast region. Al- though it rests unconformably on the Raritan, rarely is the Magothy clearly differentiated from it. According to Spangler and Peterson (1950), the two formations are comparatively easy to differentiate in the field but insufficient contacts of the two have been found to map them satisfac- torily. Data available from well logs are usually too uncritical and in- complete for stratigraphic purposes. The Raritan formation of New Jersey was considered by Spangler and Peterson (1950) as equivalent to those non-marine sediments in Delaware, Maryland and Virginia referred to as “Raritan,” Patapsco, Arundel and Patuxent. This series, they be- lieve, should be considered as a single unit and comprises both Upper and Lower Cretaceous beds. Furthermore, the beds called ‘Raritan’ in Mary- land-Delaware are only part of the sediments called Raritan in New Jersey. However, they point out that the contact of the Upper and Lower Cretaceous occurring within the Raritan formation in New Jersey has not been recognized in outcrop and sub-surface sediments have not been studied. This interpretation has been refuted by Dorf (1952) who, after reviewing the faunal and floral evidence, reaffirmed the assignment of the Raritan formation to the early Upper Cretaceous. Hence, in his view, the Potomac beds (Arundel, Patapsco and Patuxent) are pre-Raritan and assignable to the Lower Cretaceous. One of the purposes of this study was to examine the possibility of separating the Raritan from the Magothy on the basis of evidence from plant microfossils. Such a separation might render stratigraphic correlation possible in other parts of the Coastal Plain where the two formations are at present unsatisfactorily separable. Small samples (1-2 gms.) taken at intervals along the length of both cores were prepared according to techniques modified from Erdtman (1943), Faegri and Iversen (1950). The material was first demineralized by a treatment of cold 50% HF for 24-48 hours, then boiled in 10% KOH in a water bath for six minutes, washed, and dehydrated with glacial acetic acid. The t remaining cellulose was removed by acetylation for 11% minutes at 100° C. By this sequence of chemical treatments such constitu- ents as quartz and clay minerals, lignin and cellulose are removed and the cutinized and suberized pollen grains and spores are concentrated. In our material, evidence for the existence of a species of Hphedra was provided by the presence of ridged pollen grains found at the 603-613 foot level of Well No. 2. To these grains, which could be related to modern > forms of Ephedra, the designation ‘“‘Type A” was given.” —” This terminology is based on a study of modern Ephedra and will be clarified in the second part of this paper. 224 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL A total of eight intact grains of Type A and a number of fragments of similar type were found. A list of the intact fossil grains and their meas- urements is shown in TABLE I. The Type A grains are prolate, their shape class index * varies from 1.5 to 1.9, the average being 1.7. The polar axis measures 47—53 y, the average being 49 »; the equatorial diameter varies from 26-33 yp, the average being 29 » TABLE I. A list of the fossil grains and their measurements. NUMBER Ephedra LOCATION SIZE OF 8.C.L* RIDGES Al 45-85.5 (S1) 47 X 29 p 5 1.6 A 2 45-85.5 (S1) 49 X 26 pu 6 1.9 A 3 39.8-88 (S3) 53 X 33 p 6 1.6 A 4 27.5-88 (S2) 47 X 30 w 6 1.5 A 5 41-97 (S82) dia. 32 » 6 A 6 37.5-103(S1) 50 X 27 p» 6 1.8 A 7 40-87.5 (S1) 53 X 29 p 6 1.9 A 8 42-95.5 (S1) 47 X 30 p 6 1.5 Average 49 X29 6 1.7 * Shape class index. Characteristic features of these pollen grains are the narrow, undulating ridges which extend meridionally from pole to pole. These ridges fuse at the poles so that there are actually three continuous, intersecting ridges circumscribing the grain. The exine is double and intectate. It consists of a thick endexine and a highly variable ektexine. In thickness the endexine is constant, averaging between 0.5 and 1 p» over the entire surface of the grain. However, the ektexine fluctuates considerably in thickness, averaging 0.5—-1 » along the region between the ridges and 3-4 » at the central peak of the ridge at the equator. Furthermore, ridge height increases poleward to 4—5 p» as a result of the fusion of the ridges. In surface view, the highly undulate wave-like pattern of the ridge is evident. The wave height measures 2—3 » and the wave length averages 2—4 ». In the majority of cases the ridges are continuous. However, in a few cases the ridges are composed of short, individual, angular undulations which overlap one another but are not fused. In width, the ridges average between 0.8 » and 1.5 » at the equator becoming thinner at the poles. With high magnification two slender, hyaline lines, averaging less than half a micron in width can be seen to run along the surface of each ridge bor- dering its edge. A central strip 0.2 » wide, which remains dark, separates these two undulate and highly refractive borders. With changing focal levels a reversal of the light-dark relationship of the central strip and “Shape class index is the ratio of the long axis to the short axis (Erdtman, 1943). 1959 | STEEVES & BARGHOORN, POLLEN OF EPHEDRA 225 the bordering lines is produced, so that the central strip appears as a single hyaline line. In optical section, the ridges appear to be composed of individual bacu- late projections which extend out from the grain at right angles to the endexine. However, these are not true baculae, which are defined by Faegri and Iversen (1950) as “‘radial projections with the height of the element greater than the greatest diameter of projection and the upper end of the element not thicker than the base.” Rather, the projections represent pro- file views of the ridge which is seen optically sectioned at various angles as a consequence of its highly irregular form. TEXT-FIGURE I. Ephedra stapfii Steeves & Barghoorn, n. sp. 1, High focus view of the undulating ridge. 2, Same grain at mid-focus, showing the outline of the distinct, thick endexine. 3, A view slightly beyond mid-focus showing the thickness as well as the baculate appearance of the ektexine in optical section. 4, High focus view of the reticulate sculpturing found on the ss of the exine between the ridge crests on the opposite side of the grain. (All & 1000.) 226 JOURNAL OF THE ARNOLD ARBORETUM [ VOL. XL The grains as a whole appear slightly angular in polar aspect with the ridges forming the angles. In this view the ridges themselves appear rounded and slightly constricted at their bases. The surface sculpturing of the grain is rugulate-reticulate with muri 0.8—1 » wide and 0.5 » high extending out from the ridge crest down the ridge flanks enclosing lumina 2—4 » in diameter. Ephedra stapfii, n. sp. TEXT-FIGURE I. Dracnosts: Grains prolate, the polar axis varying from 47 to 53 p, aver- aging 49 »; equatorial diameter varying from 26 to 33 ym, averaging 29 p. Ridges 5 or 6, highly undulate, extending from pole to pole. Shape class index varying from 1.5 to 1.9, averaging 1.7. Locaity: Brookhaven, Long Island, New York; U.S.G.S. Well No. 2, 603—613-foot level. GEOLOGIC OCCURRENCE: Raritan formation, lower Upper Cretaceous. MATERIAL: Eight intact pollen grains and a number of fragments. Hototype: No. 56011, Palaeobotanical Collections, Harvard University. POLLEN MORPHOLOGY OF MODERN SPECIES OF EPHEDRA Stimulated by the discovery of ephedroid pollen of considerable geologic age, an intensive study was made of the pollen of existing species of Ephedra with a view to determining the possible phylogenetic significance of the fossils. Modern Ephedra pollen was first described extensively by Wodehouse (1935), who recognized basic differences within the genus based on his work with FE. equisetina and E. foliata. In his description of E. equisetina he states that the pollen is characterized by having “high few ridges 5-8 with their crests blade like and arching from end to end of the grain, and in each of the grooves between the ridges is a hyaline line which follows a serpentine course its full length.” A description of EZ. foliata notes that “when the ridges are more numerous, 11 to 15, they are not so high and the hyaline lines in the grooves are absent or only represented by a faint streak.” His descriptions of these two species represent a few of the many variations in form that characterize pollen of the genus. Although the grains are all similar in their possession of ridges, they differ in size, in shape, in number of furrows and ridges and in exine structure. Recently, Welten (1957) studied extensively the species of the Mediterranean and western Asian regions in order to identify pollen of Ephedra from north- ern and central European late- and postglacial sediments. Although the taxonomic treatment of the genus is somewhat unsatisfac- tory, forty-three of the currently recognized forty-eight species of Ephedra were examined in the present study. A list of the species, as well as their geographic source, is given in TABLE IT. In the course of the work it be- came apparent that the pollen of Ephedra falls into certain major mor- phological categories. These we have arbitrarily designated A, B, C, and D. The characteristics of the four designated pollen types will be con- 1959 | STEEVES & BARGHOORN, POLLEN OF EPHEDRA 227 sidered in detail as illustrated by descriptions of individual species. It would be of much interest to examine the relationship of these divisions, based on pollen morphology with those based on other organs and parts of the plant. Although such information is not available on a wide range of species it would add to the phylogenetic significance of the fossil forms. POLLEN TYPE A The grains are perprolate to subprolate in shape, their shape class index varying from 1.1 to 2.6 but generally averaging 1.8. The grains are sculp- tured by an average of 5 to 9 heavy, hyaline, ektexinous ridges which ex- tend meridionally from pole to pole. The least amount of variation is observed in Ephedra sinica where the ridge number ranges between 4 and 6 and averages 5.8. The greatest number is found in EF. pachyclada and E. funera. In the former species the ridge number varies from 5 to 12 and averages 7.9, in the latter it ranges from 6 to 13 and averages 8.3. (See Cart II which is based on an average of 50 grains per species.) In the polar view the grains are polygonal with the alternating ridges and deep concave furrows forming an angular outline. A narrow and ser- pentine colpus * is situated at the base of each furrow. In most cases the colpi are highly undulate. The undulation of the colpus may be either rounded or highly angular and the frequency of undulation may vary, as well as the degree. The colpus may divide forming lateral branches which extend up the ridge where they may occasionally divide again. In this manner the ektexine forms a reticulate pattern, such as that found in E. distachya. In a few cases, as in E. clokeyi, the colpus scarcely divides. Also. the width of the colpus may vary, as well as the depth to which it cuts into the ektexine. The ridge itself may be relatively smooth or highly verrucate. The term “verrucate” in the manner in which it is used here may be open to question but it is chosen for the sake of simplicity. The true “verruca” has been defined as a sculpturing element in which “the greatest diameter of the radial projection is greater than the height of the element” (Faegri and Iversen, 1950). In Ephedra pollen, however, the ridge crest appears verrucate in optical section because of the cutting of the ektexine by the branching co!pi which extend up the ridge. In all species possessing Type A pollen, except E. sinica, the ridges are triangular in shape in the polar aspect with bases which vary in width from 0.5 to 18 » and central axes which range in height from 3 to 8 ». As was recognized by Welten (1957), the ridge in this view is composed of an inner, central, hyaline core and a thin, *Colpus (pl. col). germinating furrow, the area on the grain forming or surround- ing the normal pie e of emergence of the po len tube, with a length-breadth ratio higher than 2. The ektexine is reduced, even absent (Faegri and Iversen, 1950; Ing- in this case a distinction between the two will be made. The term used to refer to the region between the ridges; the term “colpus” will be used to describe the thin, longitudinal, serpentine grooves located in the middle of the furrow area and formed by the absence of or the thinning of the ektexine. 228 JOURNAL OF THE ARNOLD ARBORETUM [VOL. XL TABLE II. Species of Ephedra examined in this study.* SPECIES SOURCE COLLECTION ripe alata Dene. Algeria Bornmiller (A) . altissima Desf. var. Morocco Balls 2487 (A) nee ee Stapf E. americana Humb. & Bolivia Cardenas 3422 (GH) Bonpl. ex. Willd. Poepp. & Endl. — Chile Joffuel 2751 (GH) E. antisyphilitica Ber- W. Texas E. Palmer 1292 (GH) land E. te Engelm. ex. S. E. cme ani E. breana ; E. californica S. Wats. E. chilensis Miers E. clokeyi Cutler E. coryi Reed var. vis- cida Cutler E. distachya L. E. equisetina Bge. E. foliata Boiss. & Kot- Vv olss. E. foliata Boiss. var. ciliata (C.A.Mey.) Stapf E. fragilis Desf. E. fragilis Desf. var. campylopoda (C.A. E. frustillata Miers . funera Cov. & Mort. - gerardiana Wall. . helvetica C. A. Mey. . intermedia Schrenk ey. var. schrenkii Stapf E. major Host var procera Aschers. & Graebn New Mexico Chile Prov. pea aren Chile San Diego Co., California Valparaiso, Chile Washington Co., Utah Rock Point, Arizona Germany Turkestan Punjab, India Prov. Almeria, Spain Palaestine Rio ach Argentina Stn _ Atacama Chile Soiteerland Teheran Talas Ala-Tau branch of the Tien-Shan Mountains in N. W. Kirghiz USSR. Wright 1851 (GH) Jegons (GH) neni visi (GH) Epling Go sere 4592 (GH) Gould 1526 (GH) Cutler 2768 (GH) Flora Germanica Exsic- catae 7513 (GH) Korshinsky, Herb. Inst. Bot. Ac. Sc. USSR 6384 (A) Parker 3304 (A) Aitchison 537 (GH) Reverchon 611 (A) Bornmiuller 1746 (A) Parodi 11858 (GH) Hitchcock 329 (GH) G. L. Webster 5830 Johnston 6208 (GH) Baumier (GH) Reliquiae Mailleanae 7046 (GH) Bornmiiller (a) det. Florin Herb. Hort. (A) Bot. Univ. Asiae Mediae * Pollen obtained from herbarium specimens in the herbarium of the Arnold Arbo- d the Gray Herbarium (cH). The dried herbarium pollen rains were f minutes in boiling water, washed in three changes of distilled water and mounted in glycerine jelly. 1959] STEEVES & BARGHOORN, POLLEN OF EPHEDRA 229 SPECIES SOURCE E. major Host var. villarsi Stapf E. monostachya L. E. multiflora Phil. ex. Stapf nana Dusen nevadensis S. Wats. ex. 5 peninsularis I. M. regeliana Florin rupestris Benth. sinica Stapf strobilacea Bge. triandra Tul. trifurca Torr. torreyana S. Wats. Beh Seah & Baas — ° E. tweediana Fisch. & Mey . A. Mey. E. wraithiana I. M. ohnston E. viridis Cov. pedunculata Engelm. Amasya, Turkey Saknain-Nor, Tibet Argentina Patagonia New Mexico Argentina Afghanistan Mexico Mexico Skardu, Baltistan Bolivia Prov. Suiyuam, China Turkmenistan, USSR Bolivia New Mexico Rocky Mt. Region, Uinta Co., Utah Uruguay Chile Mt. Pinos, Kunto, California COLLECTION Bornmiiller 3544 (A) a) Johnston 6286 (GH) Donat 42 (GH) Tower (GH) Fisher 15 (GH) Aitchison 1122 (GH) Stewart 2265 (GH) Anthony 281 (GH) G. L. Webster 5950 GH ) Licewtak 13523 (A) Herb. Inst. Bot. Ac. Sc. USSR (A) Fiebrig 2151 (GH) Hortman 642 (GH) Rollins 1719 (GH) — 5683 (GH) Johnston 4705 (GH) Howell 3824 (GH) Aitchison 40 (GH) outer, non-hyaline layer. The ridge peak is not pointed, but rounded, and averages 1.5 to 4» in width. The ridges tend to decrease slightly in height at the poles and the verrucae gradually disappear. The exine is always double and may measure up to a maximum of 8 yu in thickness. It is composed of a relatively thin endexine which extends continuously and evenly around the grain, varying in thickness from 0.8 to 2 », and an ektexine which averages | to 2.5 » in thickness in the furrow region and 3 to 6 » along the central axis of the ridge. The ektexine is double. One component, the inner, hyaline layer forms the triangularly shaped central core of the ridge. This layer extends completely around the grain becoming thinnest at the colpi and accounts for the hyaline appear- ance of the colpi at some levels of focus. The thinner, outer layer is con- tinuous or discontinuous, depending upon the depth of the colpi. The group as a whole ranges in size along the polar axis from 51 yp, the average in E. vulgaris, to 69 p, the average in E. funera, with considerable overlap between species. To a certain extent a distinction between species with large pollen and species with small pollen in the Type A group can be CHART I EPHEDRA SIZE FREQUENCY MEASUREMENTS 210 28.0 35.0 42.0 49.0 56.0 630 700 770u 0%, mm 2 40 SPECIES /s 34% -10 % E. NEVADENSIS O% E. REGELIANA E.RUPESTRIS E. ALTISSIMA E. AMERICANA E.FOLIATA E. TRIANDRA E.MULTIFLORA E. TWEEDIANA ______ E.FOLIATA CILIATA ) Ad / Aww — 24% Fan A/V AY m OGHREATA E. STROBILACEA E. GERARDIANA E.NANA Vane E. FRUSTILLATA mp N : E. BRACTEATA : Z x E. CHILENSIS , Ay E. PEDUNCULATA Z — E. FRAGILIS Jk CAMPYLOPODA yi E. TRIFURCA E. MONOSTACHYA | CHART I (Continued) 23% A IM E. PENINSULARIS \ m . TORREYANA A m . WRAITHIANA f - E. ASPERA L E. ANDINA E. PACHYCLADA E. SINICA E. HELVETICA E. CLOKEY! E. FRAGILIS E. GRACILIS /¥ E. INTERMEDIA _ A’ E. ALATA als E. ANTISYPHILITICA E. DISTACHYA ¢ E—. MAJOR PROCERA SiS