Annals ofth e Missouri Botanical "" Garden Volume XLII 19 5 5 With 34 plates, 27 figures, and 60 maps Published quarterly at Galesburg, Illinois, by the Board of Trustees of the Missouri Botanical Garden, St. Louis, Mo. Entered as second-class matter at the post-office at Galesburg, Illinois, under the Act of March 3, 1879. W*KC 0- rO Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the Missouri Botanical Garden and the Henry Shaw School of Botany of Washington University in affiliation with the Missouri Botanical Garden. Information times during the calendar year: February, May, September, and November. Four numbers constitute a volume. Subscription Price $10.00 per volume Single Numbers 2.50 each THE Company. W. Wilson TABLE OF CONTENTS Lichenological Notes on the Flora of the Antarctic Continent and the Subantarctic Islands. I— IV Comparison of Juniper Populations on an Ozark Glade and Old Field Tassel Modifications in Xea Mays A New Species of Doryopteris from Surinam PAGE Selaginella rupestris and Its Allies Rolla M. Tryon 1- 99 A New Pellaea from South Africa -Alice F. Tryon 101-102 Studies on Asiatic Relatives of Maize _ Nalini Nirodi 103-130 Carroll W. Dodge and Emanuel D. Rudolph 131-149 Three New Annonaceae from Panama Robert E. Fries 151-152 The Botanical Catalogues of Auguste de St. Hilaire John D. Dwyer 153-170 rMarion Trufant Hall 171-194 Norton H. Nickerson and Ernest E. Dale 195-212 Karl U. Kramer and Rolla M. Tryon, Jr. 213-214 A Revision of the Genus Celastrus Ding Hou 215-302 The Food of a Hindu Village of North India (Reprinted from Bulletin No. 2, Bureau Statistics and Economic Research, United Provinces, Allahabad, 1937) Charlotte Viall Wiser, Ph.B., M.Sc. 303-412 General Index to Volume XLII .4 13-418 MISSOURI BOTANICAL GARDEN STAFF Emeritus Director George T. Moore Director Edgar Anderson Assistant Director Hugh C. Cutler Carroll W. Dodge, Mycologist Robert E. Woodson, Jr., Senior Taxonomist Henry N. Andrews, Paleobotanist Rolla M. Tryon, Assistant Curator of the Herbarium Hugh C. Cutler, Curator Museum of Economic Plants George B. Van Schaack, Acting Curator of Herbarium Julian A. Steyermark, Honorary Research Associate Frederick G. Meyer, Dendrologist Alice F. Tryon, Research Associate John D. Dwyer, Research Associate Nell C. Horner, Librarian and Editor of Publications BOARD OF TRUSTEES President John S. Lehmann Vice-President Daniel K. Catlin Second Vice-President Eugene Pettus Leicester B. Faust Dudley French Henry Hitchcock Richard J. Lockwood Henry B. Pflager A, Wessel Shapleigh Rorert Brookings Smith EX-OFF1CIO MEMBERS Arthur C. Lichtenberger, Bishop of the Diocese of Mi ssouri James F. Morrell, President of the Board of Education of St. Louis Stratford Lee Morton, President of the Academy of Science of St. Louis Ethan A. H. Shepley, Chancellor of Washington University Raymond R. Tucker, Mayor of the City of St. Louis Hugh C. Cutler, Secretary Volume XLII Number 1 nnals of the Missouri FEBRUARY, 1955 Selaginella rupestrxs and Its Allies Rolla M. Tryon 1- A New Pellaea from South Africa Alice F. Tryon 101-102 PUBLISHED QUARTERLY AT GALESBURG, ILLINOIS BY THE BOARD OF TRUSTEES OP THE MISSOURI BOTANICAL GARDEN, ST LOUIS. MI8SOURL -el, tte ice at G*l*»bw», IIILmoU, March I UN** Annals of the Missouri Botanical Garden Missouri Quarterly Journal containing Scientific Contributions tanical Garden and the Henry Shaw School the >tany Washington University in affiliation with the Missouri Botanical • r^i Information Z^ft??". * "■ ¥?»™i Botanical Garden appears four times Four J.r!«rr"i j ^ksoum botanical Garden appears ion Subscription Pri< Single Numbers 0.00 per volume 2,50 each GAi? E N t Tre terf^T ^ °i f *"i ^ NALS ° F raE Mlssouw Botanical C^« v m ^ A*"* 1 *""! *"»«. published by the H. V. Wilson Annals of the Missouri Botanical Garden Vol. 42 FEBRUARY, 1955 No. 1 SELAGINELLA RUPESTRIS AND ITS ALLIES* ROLLA M. TRYON, JR. These are small plants, hardly exceeding a few inches in height, yet they are a striking part of the vegetation of the semi-arid and subalpine regions of western North America, covering barren soil and festooning rock ledges. The forty-three species included in this study comprise the section Tetragonostachys of subgenus Selaginella. Thirty- two species occur in America, eight in Africa and Madagascar, and the remainder in Asia. Superficially they appear similar and nondescript but under magnification they reveal structures of beauty in design and symmetry which amply reward the student who takes more than a casual interest in them. Selaginella rupestris and its allies occupy a unique place in xeric ecology for they are vascular plants adapted to being completely desiccated and reviving a few hours after moisture becomes available. Study of their physiological response to arid conditions would be of interest. In a number of ways the group offers data of value in geographic and evolutionary studies. S. rupestris is of particular cyto- logical and geographic significance because it has both sexual and apogamous races, the latter occupying a much larger territory than the former. In this species, as well as the others, details of the life-cycle, especially the means of fertilization and of dispersal, are poorly understood. Aside from botanical collectors and browsing by deer in lean seasons, they are relatively unaffected by animals or man and the habitats which most of them occupy are relatively little affected by fire, lumbering, grazing or cultivation. Studies of the group may be of comparative value in relation to species in which the activities of man have played a critical role. The principal effort in the present study has been directed toward a definition of the species and groups of species within a coherent framework of characters, to place the nomenclature on a firm foundation through reference to holotype material and to provide keys and illustrations to facilitate accurate identification. It is not to be expected that the keys and descriptions can be used by a student of the vascular plants without some introductory study of a few species, and indeed this Issued March 24, 1955. (i) [Vol. 42 2 MISSOURI is true of most genera. One can not successfully divorce the use of a monograph from specimens and it is not my intention that this treatment be wholly inde- pendent of such material. The Present Classification The taxonomy of the Selaginella rupestrh group is beset with a number of difficulties which must be recognized and resolved before an adequate classification can be developed. The first, and perhaps most important, is the relative plasticity of the characters, not only of the vegetative parts, but of the strobili, sporophylls and spores as well. This does not mean that taxonomically useful characters do not exist but that considerable caution must be used in choosing those of specific value. Relatively few species possess distinctive traits; rather most of them must be defined in terms of a particular combination of characters. There is evidence that there has been a great deal of parallel development in the group and that the potential variations of different structures have been realized many times. This situation has made it troublesome to find the morphological coherence of a species and the differentiating characters. Another difficulty is the very numerous char- acters that present themselves for consideration. In many groups, students prob- ably feel the characters are too few and wish for more but a superabundance of them can be equally trying. The inadequate nature of most herbarium specimens has been a source of confusion for nearly all of the early collections and many contemporary ones are fragmentary. Complete mats of some fifteen species were collected in the field and notes were made on local variations of habit and habitat. With the information obtained from these studies, it was possible to understand the relation of herbarium fragments to the living plants. The minute nature of the characters has undoubtedly resulted in many of them being overlooked or misin- terpreted. For example, the deciduous setae of S. eremophila were not detected until many years after the species was described. A relatively high magnification is needed to see the structures accurately. For general observation of the leaves and sporophylls 30 diameters is recommended, 40 diameters for the smaller details. Inasmuch as the species must largely be defined in terms of combinations of characters rather than distinctive ones, a set of specific characters for each species does not exist by itself but only in relation to other species. One set of characters may distinguish a species from its close relatives, additional ones, or another set, may distinguish it from less closely related species. It is obvious, therefore, that before a species can be worked out, the groups of species must be defined to estab- lish the circle within which character comparisons will be made. The following method was used to form a basis for grouping the species into natural units. Some twenty-five presumed species of which complete material was available early in the study were scored for about thirty characters. These included some from all parts of the plant. Their value was not known although only qualitative ones were used. Each species was compared with the others for the number of 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 3 characters in common. Those that showed definite affinities with each other were grouped together. An analysis was then made for characters which would dis- tinguish between the groups. After a certain amount of modification and rear- rangement, the series and their characters were established. Such a method has a number of limitations that might make the results obtained inaccurate but when used with discretion it undoubtedly has value. The final test of the classification was the integration of some fifteen species not originally included. These took natural places within the system in regard to their geography, morphology and phylogeny. Although many characters had been used by previous authors for the definition of species these were at first disregarded to remove all prejudice as to their possible value. The numerous available characters were reduced to a manageable number in two ways. It was assumed that lineal measurements should be avoided, if at all possible; that qualitative differences should be used for the primary characters of species. Quantitative characters of real value could be used as expressed by ratios. Structures that were variable on individual stems and within individual mats would not be reliable for taxonomic use and many characters were rejected for this reason. When the number of possibly useful characters was reduced to convenient propor- tions, several species were chosen to be analyzed for the characters of specific value. These were species such as S. cinerascens, S. Bigelovii, S. oregana and S. tortipila of which I had ample material and which have a distinctive morphology, a compact and natural range and a rather defined ecology. Each was examined to determine the structures that were sufficiently stable to be used as specific characters. The species were then compared with each other in terms of this final set of characters and it was found that the entire set was usable to express species differences. Such characters were then used to define the species within each series. Most of the species thus derived were found to have the following attributes: a set of suffi- ciently stable characters so that the "population" showed morphological cohesion; complete morphological distinction from other species; a degree of morphological cohesion and distinction that is qualitatively and quantitatively comparable to other related species; a natural geographic distribution, allopatric or largely so from related species; and a definite local ecology although this might be rather broad within the whole range. These characteristics were used as criteria for determining the category in difficult and perplexing cases. The "good" species of the section, therefore, defined themselves, their attributes were determined and the other species were defined comparatively in similar terms. Subspecific categories were defined in relation to the species. The category subspecies is used in S. arenicola where the major variants have not completely developed the properties of species. The variety is used in three species but only one of these may be considered a proper example of the use of that category. That is in S. mutica where the two variants, although definite, have poo Wight varieties are recognized largely for practical reasons, the material available being 4 [Vol. 42 MISSOURI BOTANICAL GARDEN insufficient to establish the proper category. In S. densa a most interesting but quite unsatisfactory taxonomic condition exists. The three variants are quite dis- tinct morphologically and geographically over large areas in the northern part of the range of the species. Here they would certainly rank as subspecies and perhaps even as species although the characters that separate them are not quite as well developed as in other species in the series. But the varieties all grow in the large central portion of the range where they exhibit such complete intergradation in characters that only one highly variable species could be recognized. I have chosen to use the category variety for them, placing perhaps a bit more emphasis upon their intergradation than upon their distinctiveness. The species descriptions have all been taken directly from the keys and are uniform only within each series. All of the characters mentioned in the species key are included in each description. Although this is a departure from the usual procedure, it is necessary in order to emphasize the set of characters important within an evolutionary line. A number of characters used in this treatment are peculiar to the genus or at least not commonly used in vascular taxonomy. For these I have tried to use suitable descriptive terms or phrases often without intend- ing precise definition. I quite agree with Mr. Weatherby's views that: "An attempt to find other characters which, even if seemingly incapable of altogether definite statement, could be more easily seen and better relied upon, seemed worthwhile." 1 I have used many of the characters that he introduced and some new ones. Although it was not my original purpose to enter into the general classification of the genus in any way, it finally did seem necessary to give formal recognition to the major groups of species. The erection of the series has caused in turn some slight readjustments within the subgenus. These changes are presented in the synopsis of the subgenus. The Background of the Study L. M. Underwood was the first botanist to give serious attention to the Selag- pes iris His first paper, in 1898 2 , clearly expressed the condition U1 " 1C ""wncaoon at tnat time: "The variations of the species of Selaginella with many-ranked leaves have long been a puzzle to botanists Two clearly marked species from North America have been separated from the tangle already; there remain .... the widely varying forms that for the past forty years have found an unsatisfactory resting place under S. rupestris." Underwood described a total of nine new species in unraveling "the tangle"; I maintain seven of them. He was in correspondence with Hieronymus at Berlin and they exchanged specimens of many species in the group. Hieronymus did the Selaginellaceae for Engler & Prantl's 'Natiirlichen Pflanzenfamilien' 3 and prior to and following this work described many new species in the genus. In the S. r. pestris *Jour. Am. Arb. 25:409. 1944. *Bull. Torr. Bot. Club 25:125-133. 1898. 3 Nat. Pflanz. 1 4 :621-715. 1901. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 5 posed twenty-eight species 4 of which I recognize eleven. Hieronymus deserves credit for his general position that a multiplicity of species existed and for his attempts to work them out. His work suffered for reasons that were largely either beyond his control or else in harmony with the times. His material was scrappy and inadequate, sometimes hopelessly so. Also his concepts of the natural distribu- tion of species, particularly those of the United States, were not well developed. Finally, he relied largely upon quantitative characters of the leaves, setae and cilia, many of which have proved to be of little value. The considerable number of his species that are valid is probably due to the fact that he did have material of many species. The last worker to describe many species was Maxon 5 who published eight new ones from the United States. All but three of these I recognize as valid and two of those three are recognized as varieties. Maxon not only described most of the new species remaining in the United States but in his treatments of Pteridophyta for various floras of the western United States he supplied keys for their identifica- tion. The work of Underwood and Hieronymus was largely consolidated and systematized by Maxon and a knowledge of most of the species of the United States was placed on a firm foundation by him. His knowledge and judgment of the species were excellent but probably because his work in the genus was primarily floristic he separated the species on the most convenient characters and did not develop a set of characters useful for the whole group. Also, he relied in his descriptions too strongly, in my opinion, on quantitative characters. Prior to the studies of Underwood, Hieronymus and Maxon, the species were poorly understood. The most common policy was to place all material under the name of Selaginella rupestris; a few authors segregated varieties or forms of it; a very few segregated species. This early work is reviewed although it did not play an important role in the classification of the group. The species named Lycopodium rupestre by Linnaeus 6 was, as many he treated, previously well known. Among several earlier figures of the species, that of Dillenius 7 is the best. It presents an excellent illustration of the plant from Virginia. Linnaeus combined Virginian and Siberian plants in the same species and this broad range was enlarged upon by later authors until finally S. rupestris was credited with essentially a world-wide distribution. At that time it included many diverse species. Beauvois 8 segregated Selaginella and other genera from Lycopodium in 1805. He placed Lycopodium rupestre in his genus Stachy- gynandrum. Spring, Milde and A. Braun were the principal monographers of Selaginella in the nineteenth century and they recognized, to a greater or lesser extent, the variability within S. rupestris. Spring 9 grouped his material on the 4 Most of them in Hedwigia 39:290-320. 1900. 5 Most of them in Smiths. Misc. Coll. 72 s . 1920. 6 Sp. PI. 2:1101. 1753. 7 Hist. Muse. t. 63, fig. 11. 1741. 8 Prodr. Aetheog. 101. 1805. 9 Nouv. Mem. Acad. Roy. Belg. 24:57. 1850. [Vol. 42 6 ANNALS OF THE MISSOURI BOTANICAL GARDEN basis of habit, the rather short-stemmed northern species forming his S. rupestris var. borealis and the generally elongate and lax southern ones his var. tropica. Milde 10 used an essentially geographic segregation, recognizing as formae of S. rupestris material from several countries and regions. Of the ten forms he lists, I recognize seven as species. A. Braun 11 also recognized variants of S. rupestris as they occurred in Africa, using some obvious characters of the leaves, cilia and setae. Primarily due to floristic work, by the time of Underwood's treatment the following species were usually recognized as distinct from S. rupestris: S. Dregei Presl, S. tortipila A. Br., S. oregana D. C. Eaton and S. echinata Baker. The present authority on the genus as a whole, A. H. G. Alston, has treated a number of species in the S. rupestris group, especially as they have occurred in his regional treatments. Outside of the United States and Mexico the species are rather few and he nowhere has had to deal with a sufficiently large number to seriously engage the problem of specific characters. His work may be ranked with that of Maxon, placing our knowledge of the species of South America, Africa, Madagascar, India, China and northeastern Asia on a firm foundation. It will be noted that the species of Mexico had not been treated. This was probably due to the feeling that there was not sufficient material available to make it possible to deal successfully with the many species suspected of being present. The first worker to accept this difficult problem was C. A. Weatherby 12 ' 13 . In his two papers he was the first, from the point of view of this treatment, to derive a general set of specific characters and to start grouping the species into natural units; he was also the first to understand the special nature of the S. rupestris group and to adapt his classification to it. Mr. Weatherby's publications on Selaginella epitomize rather well his publications in systematics as a whole. His output was not voluminous but he had a strong predilection for the poorly under- stood groups and in these his abilities resulted in treatments of basic and permanent value. It will be apparent to students of his papers that the present study is largely an extension of his work. This is fitting for my interest in the group stemmed from his. Morphology and Life History Most to form a mat of more or less distinctive type. The stems of a mat may be inter- laced and intricate as in S. XJnderwoodii (fig. 40) and S. cinerascens or discrete as in S. densa (fig. 47). The branching may produce a very compact mat as in Watsonii (tig. 49) or a loose, open one as in S. mutica. Growing conditions affect the form of the mat so that it is not stable in most species. Edges of cliffs, steep slopes, boulders and seepage crevices modify the habit of growth. In addi- 10 Fil. Europ. Atlant. 260-263. 1867. 11 Kuhn, Fil. Afr. 212-214. 1868. "Amcr. Fern Jour. 33:113-119. 1943. 13 Jour. Am. Arb. 25:407-419. 1944. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 7 tion, some species may have the central stems crowded and rather erect and the peripheral ones closely prostrate and widely creeping. In S. tortipila the two types were described as separate species. A similar condition exists in S. Hansenii although fortunately they were not segregated. In these species the size of the mat is in- creased by growth of the apical buds. The older portions of the stems die pro- gressively toward the apical buds and strobili. The dead portions may be rather evenly distributed throughout the mat or, especially in the species with discrete branches, they may be in the center or at one side (fig. 47). A strobilus-bearing branch dies completely since the strobili, with rare exception, are determinate. Vegetative apices are theoretically indeterminate although actually they do occas- ionally die. The species of Arenicolae, with rhizomes or basal branch buds, form various shaped clumps or mats depending upon the extent to which the substrate permits or hinders the growth of the rhizomes or basal buds. S. arenicola, which usually grows in open sand, has quite symmetrical mats. In the species with rhizomes, as S. Weatherbiana, the rhizomes grow forward, occasionally branching or producing erect, aerial stems. These aerial branches produce a branch-system on which strobili are eventually produced and a short time later the whole branch-system dies. Most of the apices die due to the determinate strobili produced although the vegetative tips also die. This may be due to the death of the portion of the rhizome beneath the aerial branches or because of the distance the apices have grown from the nearest roots. New aerial branches produced by the younger parts of the rhizome perpetuate the plant. The leaves of the rhizome are generally similar to those of the aerial stems (compare figs. 7 and 8) but are usually broader, thinner, have a shorter seta and cilia, have a less-developed dorsal groove and the living ones are pale tan or a very pale green* S. arenicola and S. rupincola lack rhizomes but have short buds at the base of the erect aerial stems (figs. 1, 5). These buds become active upon the death of the branch-system; each developing into a new branch- system. This results in a rather bushy habit unless, as is often the case in S. rupincola, growth is modified by rock crevices. The stems often branch dichotomously but this is not the predominant type of branching in any of the species. The branching pattern is formed by laterals variously subordinate to the main stem, distance between branches and the angle at which they depart from the stem. Although this is obviously different in some species the degree to which it may be modified by growing conditions makes it an unreliable taxonomic character. A branch will remain subordinate to the stem upon which it is borne until the two apical buds are sufficiently removed from each other and then it will grow more rapidly. This seems to occur at about the time the stem and branches are separated by the death of the older portions. Branches of the second and third order are also subordinate to their respective main stems. In three species, S. utahensis, S. leucobryoides and S. asprella, the stems are fragile when dry and are easily broken even with careful handling. The anatomical nature of this character has not been investigated. [Vol. 42 8 ANNALS OF THE MISSOURI BOTANICAL GARDEN During periods of unfavorable growing conditions the stems become dormant. It is probable that many species have the ability to remain alive in the dormant state for many months. A specimen of S. Watsonii was planted three months after it had been pressed and dried and it put forth new growth. Part of a mat of S. densa var. dens a was planted six months after it had been collected and stored as an herbarium specimen and it also grew. Various changes take place when the stem becomes dormant. In many species the branch tips curl upward; in Eremo- philae they become involute. A considerable portion of the stem is involved and the dormant stems form ringlets in S. oregana. This curling of the stems recalls the behavior of S. lepidophylla and its relatives, the commercial Resurrection Plant. The leaves also change position, becoming more or less closely appressed to the stem. These changes are particularly marked in the strongly dorsiventral species in which the spreading lateral leaves almost completely enfold the upper ones. However, the erect upper leaves do not perceptibly change in position. In some species as S. tortipila the leaves are usually rather closely appressed in the growing state and change position very little when desiccated. The difference in position of the leaves of a species in the growing and dormant states is not sufficiently constant for gen- eral application. This behavior, however, has a distinct advantage to the taxonomist for it renders herbarium material as useful and as valid as living material for most characters of the plant. Conclusions as to the relation of living to herbarium material were confirmed by comparisons of specimens collected from species grow- ing in the greenhouse. There is no difference between the living dormant plant and the herbarium specimen prepared from it. Soaking the specimen in water will restore it in the same way as watering the living plant. Specimens prepared from actively growing plants will assume the characters of the dormant state upon drying with the exception that portions subject to pressure in pressing will be held in more or less the original position. I have not measured the growth of the stem of any species in its native habitat but measurements of material grown in the greenhouse indicate that it usually amounts to 1-4 cm. a year. During 100 days in the greenhouse, S. arenicola ssp. Kiddellii grew 30 mm. and S. viridissima grew 15 mm. although this may not reflect the relative growth rates of the species in their natural habitats. A speci- tortip, 13 mm. on some of the stems. appears to show annual growth of The leafy stem is radially symmetrical in many species, that is, the leaves are disposed in the same manner on all sides of the stem at a given place and they are of the same size, shape, color and texture. This is true of the species that have erect stems and many of those that are prostrate. Most of the species with stems rather closely appressed to the ground are more or less dorsiventral. Extreme dorsiventral species such as S. Landii (figs. 62, 63) and S. echinata have the under leaves tightly appressed to the stem, the upper erect, the lateral ascending-curved, the under leaves longer than the upper, different in shape, thinner, tan or brownish, and the setae also differ. All intermediate conditions occur, some species having 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 9 one or more dorsiventral characters so that there is not a sharp line of distinction between radial and dorsiventral symmetry. Leaves are the most useful part of the plant for purposes of taxonomy and a number of characters of the position, base, apex, shape, texture, cilia and seta have been used in this treatment. The base of the leaf, in some species, is abruptly adnate to the stem and distinct from it in color. In other species the leaf-base is decurrent, gradually blending into the stem in form and color. Red leaves often occur in S. Sartorii, S. Steyermarkii and S. Hansenii, especially in the area between the oldest living leaves and the youngest dead ones. (fig. 23, the dark portions of the leaves are red). These are often a bright red or tinged with purple. They perhaps develop color in a manner similar to the sugar maple and other deciduous trees in northern regions. A few double leaves have been observed. The leaves are firm in texture with the exception of S. oregana in which they are sufficiently soft and thick so that they partly collapse in drying. The usual sequence of leaf development appears as follows. At the beginning of the growing season relatively few green leaves are present at the apex of the stem but renewed growth of the apical bud produces new leaves and a conspicuous zone of green leaves develops. After a time the oldest of these die; those that were green at the start of the growing season die first. Under rather uniform conditions the death of the older leaves proceeds at a rather constant rate and probably at a rate similar to the production of new leaves so that a zone of green leaves of rather uniform length is maintained. During 100 days in the greenhouse, leaves died on 11 mm. of stem in S. mutica and on 39 mm. in S. arenicola ssp. Riddellii. Toward the end of the growing season, or with the advent of a dry period, the leaves con- tinue to die, although no new ones are produced, until sometimes only a few green leaves near the bud remain. The apparent inhibition of growth of lateral branches has been discussed. In some species the leaves are affected in a similar manner. In S. cinerascens, for example, the leaves on the main stem are the longest and those on the primary, secondary and tertiary branches progressively shorter. However, once the branch apex is separated from the stem apex by the death of the older portions, the leaves that are produced are as large as those on a main stem. The strobilus is nearly always determinate but vegetative growth from its apex does occur. This is not uncommon in S. arenicola ssp. Riddellii and I have seen two other examples of the condition: S. X neomexicana {Slater J, US) and S. arizonica (A. & R. A. Nelson 1158, MO). This condition was noted in S. densa planted in the greenhouse in a dormant condition that produced vegetative tips oh two strobili upon renewal of the growth of the mat. I have seen one strobilus that was dichotomous (Soxman 351, US), a specimen of S. X neomexicana. In most species the sporophylls die progressively upward after the death of the leaves on the leafy branch beneath. In S. arenicola ssp. arenicola and ssp. acanthonota, and to some extent in S. Dregei and S. tortipila, the apical sporophylls die first and the basal last, and this occurs prior to the death of the leaves beneath so that the death of the strobilus is independent of the progressive death of the leaves. [Vol. 42 10 MISSOURI BOTANICAL GARDEN The sporophylls are hinged by their stalks to the strobilus axis and when the spores mature and the sporangia open they move outward presumably in response to their water content. This movement evidently aids in the dispersal of the spores. Although I am not certain, the two valves of the sporangium may also move. I have observed in S. rupestris that megaspores will frequently remain in the basal sporangia after all of the others are shed, apparently because the leaves directly beneath do not allow the sporophylls to move sufficiently. Occasionally spores are found in dead strobili, and thus their liberation from the sporangium is not always effected. I have seen sporelings in only a single collection and germinated megaspores were never observed on the specimens. Megaspores were germinated on filter paper in a petri dish and then were dried. The triradial split in the spore coat and the gametophyte were still evident; thus if germinated megaspores had occurred in the collections they could have been recognized. Lyon 14 reports germination of the megaspores within the sporangium, and fertilization of the gametophyte prior to dispersal from the sporangium. I have observed nothing to support this but since the species do have natural ranges, many of them over considerable areas, it would seem that the normal life cycle is completed. Possibly germination and fertilization occur at rather rare intervals when the microclimate is particularly favorable. This may occur in the sporangium, as Miss Lyon observed, or after the spores have been shed. Since the stems of most species are quite tough and difficult to break, and are usually anchored at frequent intervals by the rhizophores and roots it does not seem likely that vegetative reproduction could account for the distribution of the species. In the group of species with fragile stems, S. utahensis, S. leucobryoides and S. asprella, it is not unreasonable to consider that fragments might be dispersed by the wind although actually they are all rather local endemics. Their restricted ranges, however, may be due to other causes and they may still possess an effective means of dispersal. In S. rupestris, which is apogamous throughout much of its range, fertilization is not a factor and the megaspores may be effectively dispersed immediately upon release from the sporangium. S. rupestris has the widest distribution of any species in recently available areas and its range may be taken to illustrate the effectiveness of megaspore dispersal. I have seen sporelings growing in the soil among the stems of a mat of S. sibirica (Calder # Billard 2Q95, MO). These are about 1 cm. long and inside the megaspore a foot can be seen from which are produced one to three roots and an erect stem. This stem does not bear leaves for the first third to half of its length. The first leaves are thinner, more elongate at the base and more widely spaced than the adult leaves. Those at the tip of one of the sporelings are of the adult type in aspect and disposition; the other two sporelings have only juvenile leaves. Two rhizophores had been produced on each sporeling, the first in the axil of the second or third leaf. 14 Bot. Gaz. 32:124-141. 1901. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 11 The chromosome number has been determined in S. Weatherbiana (Try on & Tryon 5077, 2n = 18), S. mutica (Tryon & Tryon 5073, n = 9, 2n = 18) and in S. Bigelovii (Tryon & Tryon 5054, n = 9) . The sporophytic counts were made from the tips of rhizophores and the gametic ones from the contents of the mega- sporangia at meiosis. I am indebted to Dr. Amy Gage Skallerup for the consider- able work involved in making these counts. These numbers agree with the ones for S. selaginoides (S. spinulosa), S. helvetica and S. denticulata given by Miss Manton 15 as n = 9. Phylogeny The phyletic chart portrays the general relationships of the series and the species within them. There is not sufficient evidence of the actual course of evolution to more than indicate in very general terms the origin of the series. The characters that are of particular importance in determining a primitive or advanced phyletic position are the nature of the leaf-base, abruptly adnate or decurrent, the presence or absence of rhizomes, stolons or basal buds and the symmetry of the leafy stem. Arenicolae is placed as primitive because it exhibits some characters of Section Selaginella, the most primitive section of the subgenus. The comparison is par- ticularly effective between S. arenicola and S. rupincola of Arenicolae and the two species of Section Selaginella, S. selaginoides and S. deflexa. These four species have erect strobilus-bearing branches and have rhizophores borne only at the base of the stem. S. rupincola is not entirely uniform in the latter character in that occasional prostrate stems may have rhizophores borne throughout. S. deflexa is similar to pincola resume P does not mean that they are actually ancestral to the species of the series or that they represent the transition from Section Selaginella to Section Tetragonostachys; but rather that they retain the ancestral characters to a greater degree than the other living species. Eremophilae, on the basis of its leaf-base characters, is con- sidered to be the most advanced and Sartorii and Rupestres occupy a mid-position. Due to the limitations of space and the diversity of Sartorii it has not been possible to portray accurately the phyletic level of Rupestres. The radially symmetrical species of Sartorii are considered as more primitive than Rupestres, those with dorsi- ventral leafy stems and symmetrical strobili essentially on the same level and the species with dorsiventral strobili more advanced. It will be noted that the habit of growth is uniform in the Sartorii, Rupestres Eremoph Eremoph The leaf-base characters ve in Sartorii, Rupestres :s Sartorii and Rupestres are regarded as differentiating from Arenicolae or its ancestors, the characters of 15 Manton, I. Problems of cytology and evolution in the Pteridophyta. p. 259. 1950. 12 MISSOURI [Vol. 42 25. ECHINATA * 28. DREGEI 43. LANDII 42. PARISHII 24. CAFFRORUM 27. PROXIMA 23. NJAMNJAMENSIS 41. EREMOPHILA 40. ARIZONICA 22. INDICA 26. NIVEA STROBILUS DORSIVENTRAL STROBILUS SYMMETRICAL 39. PERUVIANA 21. CARINATA 20. HANSEN] I EREMOPHILAE (LEAF BASE ABRUPTLY ADNATE ON UPPER SIDE OF STEM, STRONGLY DECURRENT ON UNDER SIDE) 19. STEYERMARKII 18. WRIGHTII LEAFY STEM DORSIVENTRAL 17. WALLACEI 38. ASPRELLA 16. SHAKOTANENSIS * LEAFY STEM RADIALLY SYMMETRICAL 17. LEUCOBRYOIDES 36. UTAHENSIS 35. WATSONII 34. SIBIRICA 15. MUTICA 13. MACRATHERA 14. EXTENSA 10. WIGHTII 33. DENSA /,, 12. ARSENEI CINERASCENS 9. SARTORII 8. SELLOWII 31. OREGANA 32. RUPESTRIS SARTORII (LEAF BASE ABRUPTLY ADNATE) 30. UNDERWOODII 29. VARDEI CONTINUED GROWTH BY AERIAL STEMS RUPESTRES (LEAF BASE DECURRENT) 7. TORTIPILA 6. VIRIDISSIMA CONTINUED GROWTH BY RHIZOMES, STOLONS OR BASAL BUDS 5. WEATHERBIANA 4. BALANSAE 3. ARENICOLA 2. BIGELOVII 1. RUPINCOLA ARENICOLAE (LEAF BASE ABRUPTLY ADNATE OR DECURRENT) Phyletic chart of subgenus Selaginella section Tetragonostachys. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 13 the leaf-base becoming fixed in each series and the same type of growth developing independently. The Eremophilae were derived from the Sartorii by further special- ization of the leaf-bases. The development of the dorsiventral leafy stem has occurred independently in Sartorii and Rupestres. S. Dregei, with its radially symmetrical leafy stem and unilateral strobilus, poses a problem. Perhaps the radial symmetry is primitive and S. Dregei should be placed apart from S. echinata and S. proxima, or perhaps, as the chart implies, it may be derived from species with dorsiventral leafy stems. The evolutionary relation of the species is usually not known although in a few cases it is fairly clear that one is ancestral to another, or more likely, that the two have had a common ancestor. Within each series the most specialized species are considered advanced. The species that exhibit few if any special structures and which may have characters in common with a less advanced series are considered primitive. Geography and Ecology Section Tetragonostachys (Map 1) occurs in temperate and tropical North and South America; it is absent from the Amazon basin, Central America and, with few exceptions, the colder regions. Its distribution is not as extensive in the Old World J Siberia. In Africa there is a single species in the region of the Atlas Mountains; the others occur south of the Sahara, with a better representation in eastern than in western Africa. Three species grow in Madagascar. The section is notably absent from the Malaysian-Australian region, from the Pacific Ocean and from Eurasia. Series Arenicolae (Map 2) extends more or less laterally across the southern United States and adjacent Mexico with an isolated species in northwestern Africa a distribution not unlike that of the genus Plat anus. Sartorii (Map 3) has a range much like that of the whole section except for its absence from boreal regions and from northwestern Africa. Rupestres (Map 4) is predominantly North American with a single species in northeastern Asia and one in the Himalayan region. It is the only series that has a distribution of any extent in areas of Pleistocene continental glaciation. The ranges of S. densa (Map 43), S. sibirica p p more glaciated territory than the others and it is undoubtedly significant that there it is apogamous. Eremophilae (Map 5) is New World with essentially an Andean distribution. It is absent in the American tropics from northern Peru to southern Mexico. Most of the forty-three species of the section grow in the United States and Mexico. The two countries have a total of thirty species which represent all but two of the species of the New World. The United States has twenty-one species of which ten are endemics, while Mexico has sixteen species and five endemics. The southwestern United States— Texas to southern California— is the richest area 14 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 ■- ■ t 1. SECTION TETRAGONOSTACHYS 5. SERIES EREMOPHILAE 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 15 with a total of seventeen species. Ten species, four of them endemic, grow in California and eight species in Texas. There are three species in Canada, two in Alaska, one in Cuba, one in Guatemala and four in South America. Africa has five species, four of them endemics, Madagascar three endemic species, India and the Himalayan area three species, two endemic, and northeastern Asia two species, one of them endemic. Most of the species have continuous ranges, minor discontinuities probably being due to insufficient collecting. The important and apparently authentic examples of discontinuous range are the following. Both S. Sellowii (Map 17) and S. Sartorii (Map 18) are disjunct between Mexico and northern South America and the former species has an isolated station in Cuba. S. peruviana (Map 52) is similar but the disjunction is wider. S. Wightii (Map 19) grows in southern India pestr Greenland. Most of the species of each series are allopatric in their distribution and, with two exceptions, the relatively few cases of sympatric species are ecologically sep- arated. The examples I know in which two species of the same series actually grow pestris in Canada and S. peruviana The There are two important ecological features of the species. They are mostly plants of dry habitats and they usually do not thrive in the presence of other plants. Consequently they are most abundant in regions of moderate rainfall, or in locally dry habitats in moist areas, and they grow in rocky, gravelly or sandy places, or on cliffs or barren soil where there is a minimum of other vegetation. One species, S. oregana, is usually an epiphyte and another, S. externa, occasionally is. Many of the species grow in acid soil or on acid rocks; a few of them will also grow on Wrighti There is little evidence that they are important pioneers, being notably few in glaciated areas and in rocky places recently disturbed by fire or lumbering. Rather their chief role seems to be to fill the niche of locally transient but regionally permanent open, xeric habitats. The plants are not especially adapted either for the prevention of water loss or for its storage. The minor adaptations that are present in the form of fleshy and cutinized leaves are not sufficient to prevent frequent drying. The ability to survive desiccation undoubtedly resides in unusual physical and chemical properties of the cell contents. The extensive network of roots, borne close to the surface, enables the plants to utilize rapidly even small amounts of moisture. The species probably can grow in relatively moist places, as a few of them do, but because of their inability to compete with other plants, these habitats are not often available. They are restricted to the open habitats, which in a given area are usually the most xeric, where their special ability to survive desiccation allows them to be successful. [Vol. 42 16 ANNALS OF THE MISSOURI BOTANICAL GARDEN Distribution Maps A solid dot is used on the maps to indicate the location of a collection seen; in a few cases a circle is used to represent a specimen from a general region. In S. arenicola, half-dots are used to represent intermediates between the subspecies. Literature records, indicated by an X, have been used when they significantly aug- ment the range and when there is no doubt about the identity of the species. Except for S. Balansae, all of the literature records have been taken from the various papers of A. H. G. Alston. Acknowledgments I am indebted to the curators of the herbaria of the following institutions for the generous loan of material. Botanisches Museum, Berlin-Dahlem, British Museum (Natural History), Chicago Natural History Museum, Gray Herbarium, Department of Botany, University of Minnesota, New York Botanical Garden, Museum National d'Histoire Naturelle, Paris, United States National Herbarium and Department of Botany of Yale University. The abbreviations for these and other herbaria are taken from Index Herbariorum, part I (Regnum Vegetabile, vol.2). I wish to thank Mr. A. H. G. Alston and Mr. J. A. Crabbe of the British Museum, Mr. F. Ballard of Kew, Mme. Tardieu-Blot of Paris, Dr. H. Reimers of Berlin, and Mr. C. V. Morton of the United States National Herbarium for courtesies extended during the course of this study. Systematic Account SYNOPSIS OF SELAGINELLA SUBGENUS SELAGINELLA 16 Selaginella Beauv. Prod. Aethog. 101. 1805, conserved name. Type species: Selaginella selaginoides (L.) Link. Subgenus Selaginella (Section Homoeophyllae Spring, in Mart. Fl. Bras. 1:118. 1840. Homotropae A. Br. Ind. Sem. Hort. Bot. Berol. 1857: Append. 11. 1857, without rank. Subgenus Etiselaginella Warb. Monsunia 1:100. 1900. Subgenus Homoeopbylhm (Spring) Hieron. in Engl. & Prantl, Nat. Pflanz. 1 :669. 1901.). Sporophylls uniform or similar, leaves uniform or gradually dimorphic. Type species: Selaginella selaginoides (L.) Link. Section Selaginella (Cylindrostachyae A. Br. loc. cit., without rank. Section Cylindrost achy s (A. Br.) Hieron. loc. cit.). Strobilus cylindrical, leaves spirally arranged. Type species: Selaginella selaginoides (L.) Link. species: L S. selaginoides (L.) Link, 2. S. deflexa Brack. 16 Partly adapted from Walton & Alston, in Verdoorn, Man. Pterid. 503. 193 8. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 17 Section Tetrastichus (A. Br.) Tryon, stat. nov. (Tetrastichae A. Br. loc. cit., without rank). Strobilus tetragonous, at least the lower leaves decussate. Type species: Selaginella pumila (Schlecht.) Spring, species: 1. S. pygmaea (Kaulf.) Alston (S. pumila (Schlecht.) Spring), 2. S. gracillima (Kze.) Alston (S. Preissiana Spring), 3. S. uliginosa (Labil.) Spring. Section Tetragonostachys (A.Br.) Hieron. loc. cit. (Tetragonostachyae A.Br. loc. cit., without rank). Strobilus tetragonous, leaves spirally ar- ranged. pestris Arenicolae Tryon. Type species: Selaginella arenicola Underw. ■pine oh ovn Underw., 3. S. arenicola Underw., 3a. ssp. Riddellii (Van Eselt.) Tryon, 3b. ssp. arenicola, 3c. ssp. acanthonota (Underw.) Weatherbia Weath Series Sartorii Tryon. Type species: Selaginella Sartorii Hieron. species: 8. S. Sellowii Hieron., 9. S. Sartorii Hieron., 10. S. Wightii Hieron., 10a. var. Wightii, 10b. var. Phillipsiana, 11. S. cinerascens A. A. Eaton, 12. S. Arsenei Weath., 13. S. macrathera Weath., 14. S. shakotanensis (Franch. ex Takeda) Miyabe & Kudo, 15. S. Wallacei Hieron., 16. S. tnutica D. C. Eaton ex Underw., 16a. var. limitanea Weath., 16b. var. mutica, 17. S. externa Underw., 18. S. Wrightii Hieron., 19. S. Steyermarkii Alston, 20. S. Hansenii Hieron., 21. S. carinata Tryon, 22. S. indica (Milde) Tryon, 23. S. njamnjamensis Hieron., 24. S. caffrorum (Milde) Hieron., 25. S. echinata Baker, 26. S. nivea Alston, 27. S. proxima Tryon, 28. S. Dregei (Presl) Hieron. Series Rupestres Tryon. Type species: Selaginella rupestris (L.) Spring, species: 29. S. Vardei Lev., 30. S. oregana D. C. Eaton, 31. S. Underwoodii Hieron., 32. S. rupestris (L.) Spring, 33. S. densa Rydb., 33a. var. scopulorum (Maxon) Tryon, 33b. var. densa, 33c. var. Standleyi(Maxon) Tryon, 34. S. sibirica (Milde) Hieron., 35. S. Wat- sonii Underw., 36. S. utahensis Flowers, 37. S. leucobryoides Maxon, 3 8. S. asprella Maxon. Series Eremophilae Tryon. Type species: Selaginella eremophila Maxon. species: 39. S. peruviana (Milde) Hieron., 40. S. arizonica ^phila Greenm. & Pf eiflf. 18 [Vol. 42 MISSOURI BOTANICAL GARDEN THE SERIES OF SECTION TETRAGONOSTACHYS a. Stems erect or ascendent with rhizophores produced only at or near the base, rarely one or a few stems of a group prostrate with rhizophores pro- duced nearly throughout; subterranean rhizomes or ground-level stolons or basal branch buds present Series Arenicolae, p. 18 a. Stems prostrate to decumbent, or irregularly ascendent, with rhizophores produced at or near the apex, at least of the primary stems (the branches may rarely be erect) , or epiphytic and long-pendent with rhizophores pro- duced only at the base; rhizomes and stolons absent, basal branch buds rarely and irregularly present, b. b. Leaves with the base abruptly adnate and distinct from the stem in color 17 on all sides of the stem, or especially on the branches, those on the main stem rarely decurrent Series Sartorii, p. 31 b. Leaves with the base decurrent on all sides of the stem 17 or at least those on the underside strongly decurrent. c. c. Upper leaves with the base usually decurrent; leafy stem radially symmetrical to definitely dorsiventral, the zone of green leaves 18 about equal on all sides of the stem; branch tips straight or slightly curled in the dormant condition Series Rupestres, p. 58 c Upper leaves with the base abruptly adnate, distinct from the stem in color (rarely a distinct stem-ridge may be present); leafy stem strongly dorsiventral, the zone of green leaves 18 well developed on the upper side of the stem, very short to absent on the under side; branch tips involute in the dormant condition Series Eremophilae, p. 76 Series Arenicolae Tryon, ser. nov. Rhizomata vel gemmae brevia simplicia ad bases caulium aeriorum hae saepe stolones producentes praesentia. Caules erecti vel ascendentes. Apices ramorum recti vel leviter curvati statu inerte. Caules frondosi radialiter symmetricales. bo* base decurrente vel abrupte adnata cum caule distincta colore. Typus: arenicol Plants terrestrial; rhizomes present, their leaves different from those of the aerial stem, thinner, with shorter setae, generally broader, with shorter cilia, with a less-developed dorsal groove, the living ones pale tan to a very pale green, or short, simple buds present at the base of the aerial stems, these sometimes elongated to form stolons or both rhizomes and basal buds present. Stems erect or ascendent, with rhjzophores produced only at or near the base, rarely one or a few stems (especially m S. ru P incola) prostrate with rhizophores produced generally through- ° Ut; StCms Sh ° rt to m °derately so, forming distinct groups or in the species with 17 18 from tl- ^L rtM „ T : -•»—■■ "*"u urancnes DacK trom the apex of the i Th,s character may be obscured by death of the leaves during a Ion/ dormancy: 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 19 rhizomes usually forming tall, loose mats. Branches short to moderately long, usually remote, discrete or intricate; branch tips straight or slightly curved in the dormant state. Leafy stems radially symmetrical, the leaves equal in position, length, shape and texture on all sides of the same portion of the stem, rarely slightly dorsi ventral in position on prostrate stems; zone of green leaves equal on all sides of the stem or rarely slightly longer on the upper side. Leaves with the base abruptly adnate and distinct from the stem in color, to decurrent; dorsal groove evident from base to apex, or in S. tortipila absent or poorly developed; setae, if present, relatively straight and stout, without a modified tip, persistent, or in S. tortipila with a filiform, tortuous, usually deciduous tip. Arenicolae is a homogeneous series although the species are quite distinct and show considerable diversity. The species are held together by the characters of the limited growth of the aerial branches and the rhizome or bud growth at the base of the branches. The leaf-base which is such an important character in the other series is variable here. In S. rupincola and S. Bigelovii, the only two species that show a close af- finity, the leaf-base is abruptly adnate to the stem and distinct in color. In S. Balansae, S. Weatherbiana and S. viridissima it is decurrent toward the base of the branches and tends to be abruptly adnate toward the tip. In S. arenicola and S. tortipila the leaf-bases are decurrent. The least specialized species appear to be S. rupincola and S. arenicola, while the most specialized, those with modified setae, are certainly S. tortipila and S. viridissima. No species shows an evident relation to a species of any other series. KEY TO SPECIES a. Leaves with the base abruptly adnate and distinct from the stem in color on all branches, b. b. Rhizomes absent; cilia of the leaves of a branch- system usually predomi- nantly or entirely piliform, about l / 4 as long as the width of the blade or longer, rarely only a few piliform; leaf -apex flat to rounded. Ari- zona to Texas, Tamaulipas to Sonora and southward — 1. S. rupincola, j b. Rhizomes present; cilia of the leaves of a branch system usually pre- dominantly or entirely dentiform, about l /% to % as long as the width of the blade, rarely some or most piliform and longer; leaf-apex usually carinate to broadly rounded. California and Baja California _„2. S. Bigelovii, p. 22 a. All or at least the basal leaves of the aerial branches and of the rhizomes (when present) with the base decurrent. c. c. Setae of the leaves stout, relatively straight, without a modified tip, persistent, or setae absent; dorsal groove evident from the base to the apex. d. d. Leaves setate. e. [Vol. 42 20 ANNALS OF THE MISSOURI BOTANICAL GARDEN e. Rhizomes absent, short stolons occasionally present. Texas to Florida and North Carolina 3. S. arenicola, p. 23 e. Rhizomes present, f. f. Leaf -apex flat to slightly rounded; setae tawny to whitish. Northwestern Africa 4. S. Balansae, p. 27 f. Leaf -apex strongly carinate; setae lutescent to greenish-white. New Mexico and Colorado 5. S. Weatherbiana y p. 28 d. Leaves muticous 6. S. viridissima, p. 28 c. Setae of the leaves usually with a long, filiform, tortuous, often partly or wholly deciduous tip, rarely the tip only irregularly flexuous, or the setae without a modified tip; dorsal groove absent or developed only in the mid-portion of the blade, rarely moderately developed but not extending to the base or apex 7. S. tortipila, p. 29 1. Selaginella rupincola Underw. in Bull. Torr. Bot. Club 25:129. 1898. (Lectotype: Woo ton 124 NY! marked by Underwood as type although he did not cite it as such. Para types: Palmer Q2 K; Tourney US!; Wright 21 16 NY!, US! cited by Underwood as 2 1 06). Figs. 1,2. Map 6. Selaginella Chrismarii Hieron. in Hedwigia 39:299. 1900, as Chrismari. (Holotype: von Chrismar B!; Purpus 3156 F, GH, MO, US is identical). Selaginella Chrismarii var. Karwinskyana Hieron. in Hedwigia 39:300. 1900. (Lectotype: Palmer 554 B! chosen because of its wide distribution although at least at GH and US it is mixed with S. peruviana. Paratypes: Karwinsky B!; Schaffner II B!). Selaginella Chrismarii var. Neeana Hieron. in Hedwigia 39:300. 1900. (Holotype: Nee B!). Selaginella rupestris (L.) Spring var. rupincola (Underw.) Clute, Fern Allies, 142. 1905. Rhizomes absent; aerial branches with buds present at their base, erect or ascendent, sometimes decumbent at the base, rarely one or some of the branches of a plant prostrate. Leaves with the base abruptly adnate, distinct from the stem in color; margins ciliate, the cilia usually piliform, l / 4 as long as the width of the blade or longer, rarely some or most leaves with the cilia predominantly dentiform and shorter; apex flat to rounded to rarely carinate; setae milk-white and opaque or sometimes with a greenish-lutescent base, to whitish-lutescent and translucent. pale orange. ooth pincola is closely related only to S. Bizelovii the characters in the key. The long spreading cilia on the leaves are most char- acteristic but infrequently they may be rather short. In a very few specimens that I have seen the stems are prostrate but these may be distinguished from species of other series by the basal buds. A park supervisor in Arizona informed me that in times of drought this plant was browsed by deer. Dry open places, on ledges and in crevices of cliffs, on rocky slopes or in talus, or in gravel, usually from 1000 to 2000 m., rarely up to 3500 m. It is apparently confined to igneous rocks. 19SS] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 21 *•* Texas to Arizona, south to Puebla. Representative specimens: United States, texas: Cory 48388 (GH); Warnock 614 (GH, US), 20893 (F> MO), new Mexico: Organ Mountains, Dona Ana Co., 5800 ft. July 10, 1897, Wooton 124 (K, MO, NY, US); Wright 21 16 (sometimes as 2106) (GH, K, MO, NY, US). Arizona: Ferriss 26 (NY, US); Maguire 10021 (GH, US); Santa Catalina Mountains, April 3, 1894, Tourney (US); R. M. tf A. F. Tryon 5050 (ARIZ, B, BM, DS, F, FI, GH, K, MICH, MO, NY, P, RM, UC, US), 5051 (CU, GH, MIN, MO, PH, POM, US, WS). Mexico. Nee (B) ; 1827, Karwinsky (B). coahuila: Johnston & Muller 30J (GH). chihuahua: sw. Chihuahua, £. Fainter 92 in 1885 (GH, MO, US) ; £. Palmer 38 in 1908 (B, F, GH, MO, US), sonora: Gentry 1194 (F, GH, MO, US); Wiggins tf Rollins 450 (GH, MO, NY, US), tamaulipas: Bartlett 10529 (F, US), durango: city of Durango and vicinity, £. Palmer 554 in 1896 (B, BM, F, GH, MO, US); E. Palmer 485 in 1906 (BM, F, GH, MO, US). Guanajuato: Guanajuato, 1848, von Chrismar (B). Jalisco: Nov. 16, 1930, M. £. Jones (BM, GH, MO, US), distrito federal: Schaffner II (B, GH, US), puebla: Purpus 3156 (BM, F, GH, MO, US). la. Selaginella X neomexicana Maxon, in Smiths. Misc. Coll. 72 5 :2. 1920, as o. neomexicana. fragment GH!. fragment MO!' S. mutica X rupincola). ratvoes: Wooton in 190: Wooton 1909 US! 1907 US! Map 7. The recognition of S. X neomexicana as a putative hybrid is based on its abortive spores and intermediate morphological characters. The sporangia and [Vol. 42 22 ANNALS OF THE MISSOURI BOTANICAL GARDEN spores are variously abortive in all of the specimens examined. No mature mega- spores have been seen and in only one collection (Wooton in 1904) have I seen microspores of a size that might indicate maturity. These are probably also abortive, however, since they average some 22 /x in diameter, while the microspores of S. rupincola and S. mutica 19 range in size from 38 to 64 fi and from 30 to 53 /i respectively. Although abortive sporangia and spores might be due to a number of causes, specimens with this character are intermediate in various degrees between the two proposed parental species. The erect to ascendent branches with buds at their base are characters of S. rupincola while the commonly linear-lanceolate leaves usually with a strongly carinate apex suggest S. mutica. Most specimens are rather inter- mediate but some, such as Wherry in 1925, are rather close to S. mutica while others, such as Wooton on January 9 and on September 28, 1909, are close to S. rupincola. The distribution of the hybrid is within the range of both parent species and it is known to grow with both at some localities. On open or shaded igneous rocks at about 1000-2000 m. Texas to Arizona. Representative specimens: United States, texas: Slater 2, 3 (US) ; Knoblock 5557 (US) ; Soxman 351 (US) ; April 26, 1925, Wherry (US) ; Hinckley 3404 (GH, US), new Mexico: March 3, 1907, Wooton (F, MO, US); Organ Mountains, Dona Ana Co., 6000 ft., Jan. 9, 1909, Wooton (GH, US); Sept. 28, 1902, Wooton (NY, US); Sept. 11, 1904, Wooton (US). Arizona: Paradise, Cochise Co., 1904, Ferriss (GH, US). 2. Selaginella Bigelovii Underw. in Bull. Torr. Bot. Club 25:130. 1898. (Holotype: Bigelow NY!. Para types: Coville & Funs ton IO I NY!; Henshaw NY! US!; Parish & Parish 671 NY!; Underwood NY!). Fig. 3. Map 8. Rhizomes widely creeping; aerial branches with buds often present at their base, erect, rarely some decumbent at the base. Leaves with the base abruptly adnate, distinct from the stem in color; margins ciliate, the cilia usually dentiform, about y s to y 6 as long as the width of the blade, less often some or most of the leaves with the cilia piliform and longer; apex strongly carinate to broadly rounded, or predominantly so; setae usually milk-white and opaque, to whitish, lutescent- or greenish-white, translucent. Megaspores rugose to rugose-reticulate, rarely nearly smooth on the outer face, lemon-yellow to pale orange. lovii is closely related to S. rup key. important times This species usually grows in open places, on serpentine, sandstone or igneous rocks, in crevices or shallow depressions, on cliffs, at the base of boulders or in 2Q Tr 7 OI1 » A - F. in Ann. Mo. Bot. Gard. 36:418, 419. 1949: Wiggins, I. L. in Amer. Fern Jour. 22:15. 1932. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 23 gravelly soil, less often in shaded or moist situations or in clay soil; it occurs from about sea level up to 2000 m. California and Baja California. Representative specimens: United States. California: Abrams 3126 (GH, MO, NY), 3402 (F, GH, MO, NY, US); Lt. Whipple's Exped., March 16, 1854, Bigelow (GH, NY, US, YU); Clokey 5165 (F, GH, NY); Coville tf Funston 101 (GH, NY, US); Fosberg 697 (F, MO, NY, US); Heller 7276 (GH, MO, NY, US), 7610 (F, GH, MO, NY, US), 8604 (F, GH, MO, NY, US) ; April 3, 1893, Henshaw (NY, US) ; Moxley 876 (F, GH, NY, US) ; E. Palmer 434 in 1875 (F, MO, NY); Parish tf Parish 671 (F, NY); R. M. tt A. F. Tryon 5053 (BM, MO, NY, P), 5054 (F, K, MO, UC), 5056 (B, GH, MO, US); Jan. 1889, Under- wood (NY). Mexico, baja California: Ferris 8470 (NY, US) ; Wiggins 4212 (F, GH, NY, US). 3. Selaginella arenicola Underw. in Bull. Torr. Bot. Club 25:541. 1898, based on S. arenaria Underw. Map 9. Rhizomes absent; aerial branches with buds present at their base, which are sometimes elongated to form stolons, strictly erect, ascendent to decumbent at the base. Leaves with the base decurrent, blending into the stem in color or sometimes rather abruptly decurrent on the apical portion of the stem; margins rarely eciliate, usually ciliate, the cilia piliform, about % as long as the width of the blade, less often shorter or dentiform; apex nearly flat to carinate; setae usually milk-white to tawny and opaque, rarely whitish with a lutescent base and translucent. Mega- spores rugose, rugose-reticulate to rugose-tuberculate on the commissural face, less prominently marked to smooth on the outer face, white to pale orange. The three subspecies differ in characters of essentially specific value but there are too many intermediates to allow them full recognition. These characters are those mentioned in the key to subspecies. In addition, the subspecies also tend to differ in various leaf characters but these are too variable within each subspecies to be of taxonomic value. Ssp. Kiddellii has the longest leaf-base and ssp. acanthonota has the shortest; ssp. acanthonota has the longest setae and ssp. Kiddellii the shortest; ssp. acanthonota has the stoutest leafy stem with the most leaves; ssp. Kiddellii has the most slender stem with the fewest leaves; and ssp. acanthonota has the most cilia on a leaf while ssp. Kiddellii has the fewest. Subspecies arenicola occupies an intermediate position in the range of variability of all of these characters. Subspecies Kiddellii is evidently the most primitive of the subspecies. The manner of death of the strobilus is the same as that in all other species of the series and its spores are less distinctive. Ssp. acanthonota, with aerial rhizophores, is the most advanced. S. arenicola differs from S. rupincola, the other member of the series that lacks rhizomes, in having the base of the leaves, at least on the basal portions of the stems, decurrent rather than abruptly adnate. It grows in open habitats or in light shade, in sand or in weathered soil pockets of acidic rocks. It probably has considerable ability as a colonizer; at least I have seen it in Florida growing on the graded sides of relatively recent roads. North Carolina to Texas. [Vol. 42 24 ANNALS OF THE MISSOURI BOTANICAL GARDEN KEY TO SUBSPECIES a. Aerial branches erect; all rhizophores subterranean, branched into roots near the base. b. b. Basal sporophylls of a strobilus die after the leaves beneath the strobilus; outer face of the megaspores usually rugose 3a. ssp. Riddellii, p. 24 b. Basal sporophylls of a strobilus die before the leaves beneath the strobilus; outer face of the megaspores usually smooth 3 b. ssp. arenicola y p. 26 a. Aerial branches erect to decumbent at the base; all or many of the rhizo- phores aerial and unbranched toward the base; basal sporophylls of a strobilus die before the leaves beneath the strobilus; outer face of the megaspores usually smooth 3 c. ssp. acanthonota, p. 26 3a. Selaginella arenicola ssp. Riddel lii (Van Eselt.) Tryon, comb. nov. Map 10. Selaginella Riddellii Van Eselt. in Contrib. U.S. Nat. Herb. 20:162. 1918. (Holotype: Thurow 7 US!. Paratypes: Drummond 352 NY!; Holmes & Fetberolf US!; Jermy 342 MO!, US!; Undheimer 76 MO!; Long in 1900 NY!, in 1901 NY!; Plank on Aug. 10, Thurow Riddell Basal buds sometimes elongated to form stolons; aerial branches erect; rhizo- phores subterranean, branched into roots near their base. Strobilus usually borne on a long leafy branch, often with vegetative apical growth; basal sporophylls dying after the leaves beneath the strobilus. Megaspores with the outer face usually rugose to rugose-reticulate, rarely smooth. The following material is intermediate between ssp. Riddellii and ssp. arenicola (Map 10): Louisiana: Dorman in 1935 (GH, US); Wherry in 1937 (US). Alabama: Harper 37 (GH, MO, NY, US), 3116 (MO, NY, US); Wolf in 1933 (US); Leeds in 1934 (NY); Gattman in 1938 (F). Georgia: Harper 1800 (F, GH, MO, NY, US) ; Pyron & McVaugb 3051 (US) ; Duncan 8683 (MO). All of the material cited above from Alabama is from Eight-mile Creek and St. Bernard, Cullman Co., and has most of the megasporangia abortive. This is Co., but it may be due to other causes. pestris oods or shallow depressions of granite, sandstone or conglomerate outcrops, or at the base of boulders. Alabama to Oklahoma and Texas. Representative specimens: United States. Alabama: Harper 3403 (GH, MO, NY, US), 3762 (GH, MO, NY, H?" , L ° u f A ? A: Red R iver, Hale (GH, US). Oklahoma: 1928, Draper (g Gould (US); Y« W - H - L °»S (NY); E. }. Palmer 9540 (MO .US) 13411 (MO, US), 33364 (GH, MO); Aug. 10, 1892, Plank (NY); Aug. 18, 1892, Plank (NY); Reverchon 1632 (F, GH, MO, NY, US); Riddell 16 (NY); 19 55] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 25 -*fc ^. ARENICOLA SSP. —i * A :-; -•» o*v RIDDELLH "?"%' j « \ V-l-o '.* ■ INTERMEDIATE RIDDELLH - ARENICOLA * S. ARENICOLA SSP. ARENICOLA v S. ARENICOLA SSP ACANTHONOTA • * \ \ -Jt- - / • * -.4 - -\j INTERMEDIATE •* .1 ARENICOLA - ACANTHONOTA [Vol. 42 26 ANNALS OF THE MISSOURI BOTANICAL GARDEN W I hurow 7 (US) ; Thurow . (BM, F, P, MO, NY, US). 3b. Selaginella arenicola ssp. arenicola. Fig. 4. Map 11. Selaginella arenaria Underw. in Bull. Torr. Bot. Club 25:129. 1898, not Baker 1 88 3. (Lectotype: Underwood 1355 NY! was marked by Maxon as type as to six of the specimens (now marked x) ; the other specimens and all of those on the sheet of the same collection at US are intermediate between ssp. arenicola and ssp. acanthonota. Paratypes: Underwood 1353a NY!; Chapman NY!. Nash 1449 is excluded as a type, it is ssp. acanthonota) . Basal buds not forming stolons; aerial branches erect; rhizophores subterranean, branching into roots near their base. Strobilus usually borne on a long leafy branch, very rarely with vegetative apical growth; basal sporophylls dying before the leaves beneath the strobilus. Megaspores with the outer face smooth to rarely rugose. Specimens that are intermediate with ssp. Riddellii have been mentioned under that subspecies. The following specimens are those intermediate with ssp. acanthonota (Map 11): Georgia: Pyron tf McVaugh 3101 (MO, US) , 31 1 4 (US) , 3175^ (US). Florida: A. A. Eaton in 1903 (US); Underwood 1355 (NY in part, US). Subspecies arenicola grows in open places among shrubs in white sand or rarely in crevices or on flat exposures of sandstone or granite. Georgia and Florida. Representative specimens: ,MO N £v t S t T q A x TES "^ E ?^^ A - H ' Curth$ 6 7'4 < GH > MO > NY > US ) I Harper 1854 Sin i^* T h l86 ° ( >' GH ' M °' NY ' US > ; P y ron * McVaugh 3048 (US), 3 I 75 N?' JS 5 r"^ 35I ° ( JJ°;Ii S) - FL ° RIDA: 184 °' Cb *P™* < NY > i JulyV 1934, iii« (NY, US); Tryonsoi4 (B, BM, GH, MO), 5 oi6 (F, MO, NY, P, US), 5 0I 7 MICH > MIN > MO, NY, P, PH; POM, RM, UC US, WS); Sept. 1901, Underwood (NY); Underwood d Selby 2 (NY) l E Z^?°A- *"*» '7817 (US), 17980 (US); Drouet Richards 3337 $ '.MO ) NY Usf • 9WZ eek '/ a , TT MiSUe i ?- 800 ° ft " Wy 25 ' 1908 > Standley 4558 (F, MO NY, US); Standley 4158 (US), 4626 (MO, NY, US), 14692 (F, US). 6. Selaginella viridissima Weath. in Jour. Am. Arb. 24:326. 1943. (Holo- type: Stewart 2204 GH!. Paratype: Johnston 8683 GH!). Fig. 9. Map 15. 3 *t££^! A " Amer ' Fem J ° Ur - 36:5L I946 ' (Holot ^ e: C "y 44831 GH! Rhizomes widely creeping; aerial branches with buds rarely present at their base erect or ascendent, occasionally some decumbent or prostrate. Leaves with the base decurrent and blending into the stem in color on the basal portion of the stem, less decurrent, essentially distinct in color on the apical portion; margins ciliate cilia dentiform, % to % as long as the width of the blade; apex broadly rounded to carinate, muticous. Megaspores rugose, pale orange. The muticous leaves afford the best character to distinguish S. viridissima from the other species of the series. An additional character is the very slender stems. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 29 From additional material it is evident that S. Coryi can not be maintained on the basis of habit; both the Mexican and Texan material have erect or ascendent branches and some prostrate ones in the same mats. S. viridissima forms large, loose mats, rarely partly pendent, on shaded or locally sheltered cliffs and ledges of igneous rock. Southwestern Texas and Coahuila. Specimens examined: United States. Texas: Correll 13627 (MO); Chisos Mountains, Brewster Co., July 10, 1944, Cory 44831 (GH, US) ; May 24, 1928, Murray (F) ; R. M. & A. F. Tryon 5015 (ARIZ, B, BM, DS, F, FI, GH, K, MICH, MO, NY, P, PH, RM, UC, US) ; Warnock 21694 (US). Mexico, coahuila: Johnston 8683 (GH); Kenoyer # Crum 2620 (GH) ; Canon de Calabasa, north wall of Sierra Maja, Oct. 27, 1941, Stewart 2204 (GH). >-^ l \ ■ • i L ■ ■ \ \ t * \ I * * X X* « - , *"-» » * '••-••*^ ...,*- * im \ - ^ f * • ' I f ■ •yv * * \ ■ * *m ' *'•* * - r w* .1 ;•• - 15. S. VIRIDISSIMA e* m 9 i -•- 14. S. WEATHERBI AN A \ > 7. Selaginella tortipila A. Br. in Ann. Sci. Nat. V. 3:271. 1865. (Lecto- type: Kugel B!. Paratype: Curtis B!). Fig. 10. Map 16. Selaginella rupestris (L.) Spring var. tortipila (A. Br.) Underw. Our Nat. Ferns, ed. 3, 140. 1888 Selaginella Sherwoodii Underw. in Torreya 2:172. 1902. (Lectotype: Sherwood in 1902 NY!. Paratypes: Donnell-Smith in 1882 NY!; comm. Sherwood in April, 1902 NY!; Sherwood in 1901 NY!. Donnell-Smith in 1881 NY! is excluded as a type, it cor- responds to the type of S. tortipila). Selaginella rupestris var. Sherwoodii (Underw.) Clute, Fern Allies, 142. 1905. 30 [Vol. 42 MISSOURI Rhizomes widely creeping; aerial branches with buds sometimes present at their base, erect or ascendent, occasionally some decumbent. Leaves with the base decurrent and blending into the stem in color; dorsal groove absent or developed only in the mid-portion of the blade; margins eciliate to ciliate, the cilia dentiform to piliform, up to % as long as the width of the blade; apex carinate to strongly carina te; setae with a filiform, tortuous to irregularly flexuous, often deciduous, tip, rarely lacking a modified tip, lutescent- to greenish- white, translucent to sub- opaque. Megaspores rugose to tuberculate, rarely smooth on the outer face, pale yellow to pale orange. The two distinctive characters of S. tortipila, the poorly developed dorsal groove and the tortuous, filiform seta tip, are somewhat variable. Rarely on some leaves of a stem the dorsal groove may be quite well developed or the setae may be nearly straight and stouter than usual. The strobili are usually very short, about 5 mm. long or less. The basal sporophylls of a strobilus usually die after the leaves beneath; occasionally they die before. S. arenicola is the only other species in the series in which the sporophylls die before the leaves beneath the strobilus, this being acanthonota j the rhizomes, a character only suggested by the many other collections available. Wherry 2 ^ pointed out, on the basis of field observation, that specimens de- scribed and identified as S. Sberwoodii were portions from the center or the upper part of a mat, or from mats growing in unusually exposed situations. The type and similar specimens of S. tortipila, on the other hand, represent material from the outer or lower portions of a mat or from a mat growing in a shaded place. S. tortipila usually forms large, compact mats on exposed granite or granite- gneiss cliffs or rocks; less often in thin soil in depressions of rocks or in shaded localities. It grows from 900-1500 m. in North Carolina to as low as 90 m. in Georgia. North Carolina to Georgia. Representative specimens: United States, north Carolina: D. S. 6 H. B. Correll 7807 (GH, US)- 1882, o^oIgh m G n' ^ N J' US); 1898 ' Huger (B ' NY); ^elell 94 78 (GH, US) H ;/ V a \Y h °Z" m * 364 ° (MO); Broad River > 22 J" 1 ?' "41, Rugel (B, NY, US) Highlands, Macon Co 1902, Sherwood (NY); April, 1902, comm. Sherwood (NY) , , ~„^„ VJ .„ x , , Jiyiu, i^v/ii, comm. znerwt south Carolina: Carte (B); 1881, DonmU-Smith (F, GH, MO, NY US1 Georgia- $£■*"»/<&£ (GH> MO); "'""""" ,0, ° 8 US); Peninsula Ind. orientalis, Wight us on 211 (US): Thwaites I did (P): W 10b. Selaginella Wightii var. Phillipsiana Hieron. in Hedwigia 39:320. 1900. (Holotype: Phillips B!). Selaginella Phillipsiana (Hieron.) Alston, in Jour. Bot. 77:222. 1939. The localities taken from the literature (Map 19) are from Alston, in Jour. Bot. 77:222. 1939. On stones and rocks, 1200-1500 m. British Somaliland to Tanganyika. Representative specimens: British Somaliland: /. B. Gillett 4681 (K), 4681 A (K); Glover & Gilliland 480 (BM), 771 (BM); Godman 6 (BM) ; Darra-as, 5000 ft., June, 1895, Mrs. Lort Phillips (B,K,P). Kenya: Napier 2054 (&)• Tanganyika: Uhlig 845 (BM). 11. Selaginella cinerascens A. A. Eaton, in Fern Bull. 7:33. 1899. (Holo- Fig. 14. Map 20. Selaginella bryoides Underw. Our Nat. Ferns, ed. 6, 13 8. 1900, not (Kaulf.) Hieron., type: Kimball GH!). times 1901, not Lycopodium bryoides Nutt. ex Baker, 1887; considered as based on Selag- inella cinerascens A. A. Eaton. Stems with the older primary branches once to usually twice to rarely three Leaves herbaceous to herbaceous-papyraceous, none red, linear, acuminate to broadly lanceolate; base abruptly adnate, distinct from the stem in color or sometimes rather decurrent on the leader stem, usually glabrous, sometimes pubescent; margins rarely eciliate, usually ciliate, the cilia usually piliform, or dentiform only toward the apex, the longest cilia up to % as long as the width of the blade; apex acute to nearly obtuse, flat to slightly rounded, plane in profile; setae absent, the apex not or slightly modified, or if present, arising rather abruptly in form and col the blade, stout, whitish, opaque to translucent. Sporophy % very Megaspores slightly rugose on the commissural face, more prominently marked on the outer face, with an equatorial ring, pale yellow. 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 41 S. cinerascens is quite closely related to the next species, S. Arsenei. The two may be distinguished by the cilia which are predominantly piliform in S. cinerascens and dentiform in S. Arsenei. There is a close resemblance to S. nivea of Madagascar but this does not extend beyond characters of habit and general aspect. This species affords the best illustration of the difference in length of leaves on the stem and branches, a character presumably due to some type of inhibition of growth. The leaves on the main stem are the longest, those on the primary branches are shorter and those on the secondary branches are the shortest. S. cinerascens is one of the few species that is not closely associated with rocky habitats. It grows on hillsides and slopes, among or beneath shrubs and small trees, often in clay soil, sometimes in rocky soil, from sea level to about 200 m. California and Baja California. Representative specimens: United States. California: Abrams 3399 (F, GH, MO, NY, US); April 12, 1918, Bethel (GH, MO, US) ; National City (San Diego), May, 1897, L. F. Kimball (GH, NY) ; Parish 10765 (NY, US). Mexico, baja California: Ferris 8474 (NY, US); Wiggins 4213 (GH, NY, US). 12. Selaginella Arsenei Weath. in Jour. Arm Arb. 25:417. 1944. (Holotype: Arsene 10641 US! fragment GH!. Paratypes: Arsene 9983 US!, 10643 US!). Fig. 15. Map 21. Stems with the older primary branches once to twice pinnate; leafy stems radially symmetrical. Leaves herbaceous, none red, ligulate to ligulate-long- triangular; base abruptly adnate, distinct from the stem in color, glabrous to pubescent; margins ciliate, the cilia dentiform, the longest cilia about l /% as long as the width of the blade; apex acute to acuminate, rounded, plane or nearly so in profile; setae absent, the apex slightly modified, or if present, arising evenly in form and usually in color, about % as long as the blade, stout, greenish- to lutescent-whitish, subopaque. Sporophylls with the margins very short-ciliate toward the base; apex broadly rounded. Megaspores slightly and finely rugose- reticulate, lemon-yellow. The differences between S. Arsenei and the closely related S. cinerascens are discussed under S. cinerascens. S. Arsenei grows in rocky places, at one locality at 2700 m. Central Mexico. Specimens examined: Mexico, san luis potosi: T. C. # E. M. Frye 2983 (F). queretaro: Queretaro, 1914, Arsene 10641 (GH, US), 9983 (US), 10643 (US), hidalgo: Hitchcock & Stan- ford 7248 (US). Guerrero: T. C. & E. M. Frye 3149 (US). 13. Selaginella macrathera Weath. in Jour. Arn. Arb. 24:326. 1943. (Holo- John Fig. 16. Map 22 stems snort, witn tne oiaer primary orancucs unwc i-u **xw 7 leafy stems radially symmetrical. Leaves herbaceous-papyraceous 42 [Vol. 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN ulate; base abruptly adnate, distinct from the stem in color, glabrous to slightly pubescent; margins ciliate, the cilia dentiform, or short-piliform only at the base, the longest cilia % or less as long as the width of the blade; apex acuminate to bluntly acuminate, broadly rounded, plane or nearly so in profile; setae arising abruptly in form and color, % to usually % as long as the blade, filiform, white or whitish, subopaque to translucent. Sporophylls with the margins short-ciliate toward the base; apex broadly rounded to carinate. Megaspores slightly to mod- erately finely rugose on the commissural face, moderately rugose on the outer face, yellow-orange. 20. S. CINERASCENS 23. S. SHAKOTANENSIS 24. S. WALLACEI Although known from only a single collection there is no doubt that this is a valid species. The dentiform cilia of the leaves and the very long setae distinguish it from all others. Collected on ledges on north-facing lava cliffs in Chihuahua, Mexico. Specimens examined: Mexico, chihuahua: Sierra del Virulento, 2-3 miles east of Rancho Virulento, ca. 65 miles south of Ojinaga, Aug. 11, 1941, J. M. Johnston 8067 (GH, MO). 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 43 14, Selaginella shakotanensis (Franch. ex Takeda) Miyabe & Kudo, in Jour. Fac. Agric. Hokkaido Imp. Univ. 26 (Fl. Hokkaido & Saghal. I):64. 1930. Fig. 17. Map 23. Selaginella rupestris (L.) Spring var. shakotanensis Franch. ex Takeda, in Bot. Mag. Tokyo 23:237. 1909. (Holotype: Faurie 9895 SAP). Stems with the older primary branches once to twice pinnate; leafy stems some- what dorsiventral in the position of the leaves. Leaves fleshy to fleshy herbaceous, none red, ligulate to ligulate-lanceolate; base abruptly adnate, distinct from the stem in color, or sometimes rather decurrent on the leader stems, glabrous; margins ciliate, the cilia piliform, or dentiform only toward the apex, the longest cilia % to % as long as the width of the blade; apex convexly acute to obtuse, rounded to carinate, abruptly beveled to truncate in profile; setae arising abruptly in form and color, about */5 as long as the blade, attenuate, whitish, translucent. Sporophylls with the margins eciliate to long-ciliate toward the base; apex narrowly carinate. Megaspores finely and very slightly rugose-reticulate, orange. S. shakotanensis may be distinguished from S. sibirica of Rupestres, which overlaps its range, by the shorter setae and the abruptly adnate leaf-bases. The setae are about % as long as the blade in this species while they are % as long or longer in S. sibirica. The main stem may have some of the leaves with a decurrent base but on the branches the leaf-base is abruptly adnate. S. shakotanensis bears a rather close resemblance to S. mutica var. limitanea from which it may be sep- arated by its slightly dorsiventral leafy stem. Rocky alpine habitats; one collection was made at 2000 m. Japan. Specimens examined: Japan: Aug. 20, 1898, Faurie (P), 1 577 (P), Shakotan, Yezo Island, Shiribeshi, June 9, 1893, Faurie 9895 (P, MO), 10543 (P, MO), 13151 (P, MO) 5 U. S. Nat. Herb, sheet 1095388. 15. Selaginella Wallacei 1900, maintained over S. montanensis by Broun, Index N. Am. Ferns, 159. 1938. (Holotype: Walla B! fragment NY!). Fig. 18. Map 24. Selaginella montanensis Hieron. in Hedwigia 39:293. 1900. (Holotype: A. tf A. Krause B! fragment NY!). Selaginella rupestris (L.) Spring f. Wallaces (Hieron.) Clute, in Fern Bull. 16:52. 1908. Selaginella rupestris f. montanensis (Hieron.) Clute, in Fern Bull. 16:52. 1908, as montaniensis. Selaginella rupestris var. columbiana M. E. Jones, in Univ. Mont. Bull. Biol. Ser.- 15:8. 1910. Weiser Selaginella rupestris var. Wallacei (Hieron.) Frye, Ferns Northwest, 33. 1934. Selaginella RosendahlH Hieron, ex Frye, Ferns Northwest, 33. 1934, in synon. (Evidently based on Rosendahl 864). Selaginella Wallacei f. columbiana (M. E. Jones) Broun, Index N. Am. Ferns, 159. 1938. Selaginella rupestris var. montanensis (Hieron.) [incorrectly attributed to Frye by") Broun, Index N. Am. Ferns, 159. 1938, in synon. [Vol. 42 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN btems with the older primary branches once to twice, rarely three times pinnate; leafy stems radially symmetrical to rather dorsiventral in position, length and shape of the upper and under leaves. Leaves herbaceous to fleshy, none red, ligulate to ligulate-long- triangular; base abruptly adnate, distinct from the stem in color, sometimes rather decurrent on the leader stems, glabrous or pubescent; margins eciliate or ciliate, the cilia dentiform to usually piliform, the longest cilia Y 5 to l / 4 as long as the width of the blade; apex acute to obtuse, broadly rounded, abruptly beveled to truncate in profile, or predominantly so; setae arising abruptly in form and color, % to %, rarely %, as long as the blade, rather stout, milk-white and opaque to whitish or greenish-white and translucent. Sporophylls with the margins eciliate to short-cilia te toward the base; apex broadly rounded to subcarinate. Megaspores rugose to rugose-reticulate, usually with an equatorial ring, pale orange. Wallacei It is know and to a considerable extent its variability is correlated with its ecology. Specimens from very damp, shady situations have long stems that form loose mats and the leaves are somewhat distant. Those from dry, exposed rock or gravelly turf have short stems that form a compact mat and the leaves are close. Of the species that grow within its range it sometimes resembles S. densa var. scopulorum of Rupestres. It may be separated by the remote branches, S. densa having approximate ones, and by its abruptly adnate rather than decurrent leaf- ba ses. Suskdorf species inella Wallacei gravelly soil, or on moist, shaded rocks, bluffs and cliffs, in crevices or covering various types of igneous and sedimentary rocks, from sea level to 2000 m. Western Montana to southern British Columbia and California. Representative specimens: Canada. British Columbia: Hitchcock & Martin 7367 (MO, NY) ; Macoun 86172 <™' NY ' US >' 86 3? 6 < GH ' NY » US), 86377 (GH, NY, US); Rosendabl 864 (BM, MIN, MO, NY, US). United States. Montana: Aug. 13, 1908, M. E. Jones (BM, MO, US); "Crossing", Northern Pacific Railroad, Oct. I, 1882, A. tf A. Krause (B, NY), idaho: Kirkwood & Severy 1718 (GH, US); H. T. Rogers 763 (GH, MO); Sandberg et al. 54 (GH, US). Washington: Constance fif Rollins 998 (F, GH, MO, NY, US); Otis 1149 (GH MO NY, US) ; Suskdorf 8563 (GH, MO, US) ; Thompson 8747 (F, MO, NY) ; S. M. a? E B. Teller 748 (GH, MO, US), oregon: Thompson 4002 (MO, NY, US), 4020 (MO NY VS),4W2 (MO, US); Oregon, Capt. Wallace (B, NY, YU); Wilkes Exped. "Wash. & ?™ ?/A x ( £ H ' t ? Y ' US ,', YU) • CALIFORNIA: Baker 2746 (GH, NY, US) ; Beetle 1566 (GH, MO, NY, US); Heller 5024 (GH, MO, NY, US). 16. Selaginella mutica D. C. Eaton ex Underw. in Bull. Torr. Bot. Club 25:128. 1898. (Lectotype by Weath. in Jour. Am. Arb. 25:413. 1944: Meehan YU! fragment NY!. Paratypes: M. A. C. Uvermore NY!; A. G. Compton NY!; E. J. Spence NY!; Tourney on Aug. 28, 1894 NY!; Brandegee 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 45 NY!. Bigelow NY! and Mex. Bound. Survey NY! are excluded as types; they are intermediates between var. limitanea and var. mutica) . Stems with the older primary branches once to twice pinnate; leafy stems radially symmetrical. Leaves fleshy to fleshy-herbaceous, none red, ligulate- lanceolate to lanceolate-elliptical to lanceolate-ovate; base abruptly adnate, distinct from the stem in color, or slightly decurrent on the leader stems, glabrous or pubescent; margins ciliate, the cilia entirely piliform, the longest cilia l /± to % as long as the width of the blade, to entirely dentiform and shorter; apex obtuse to convexly acute, strongly carina te, subtruncate in profile; setae absent, the apex being scarcely modified, or if present, arising abruptly in form and usually in color, up to % as long as the blade, stout, whitish to greenish-white, subopaque to translucent. Sporophylls with the margins eciliate to long-ciliate toward the base; apex carinate to narrowly carinate. Megaspores rugose-reticulate to coarsely rugose on the commissural face, less marked to smooth on the outer face, pale orange. The variation of S. mutica, as pointed out by C. A. Weatherby, 23 is rather parallel to that of S. Underwoodii which has a very similar range. However, in S. mutica the material of Texas, southern New Mexico and southern Arizona is more distinctive and the line of demarcation stronger than in S. Underwoodii, so that I am recognizing two varieties. The characters of seta length and ciliation of the sporophylls correlate to a high degree and separate the major components of the species. The southern var. limitanea is itself variable but the extremes, one with short, dentiform, ascending cilia on the leaves, and the other with long, piliform, patent cilia, are too often connected by intermediates and the characters vary considerably in some individual mats. The muticous incurved leaves of var. mutica give the stems a reptilian aspect under low magnification. In Colorado I have seen S. mutica growing with S. Underwoodii in the same niche. From a number of such cases it appeared quite clear that although the mats of the two species were originally nearly equal in size, S. mutica gradually occupied more and more of the niche and finally entirely replaced S. Underwoodii. Rarely a strobilus may be slightly dorsiventral. Colorado and Utah to Texas and Arizona. KEY TO VARIETIES a. Leaves setate; sporophylls with the cilia mostly or entirely ascending, all or some dentiform 16a - var - limitanea, p. 46 a. Leaves muticous or nearly so; sporophylls with the cilia entirely or pre- dominantly spreading, piliform 16b. var. mutica, p. 46 23 Jour. Arn. Arb. 25:415. 1944. 46 [Vol. 42 MISSOURI 16a. Weath. in J 1944. (Holotype: Goodding US! fragment GH!. Paratypes: Ferriss GH!; Hinckley 1155 US!; Ingram 2723 US!; Moore 8 Steyermark 3046, in part, GH!; E. J. Palmer 31951 US!; Slater US!; Standley in 1906 US!; Wooton in 1899 US!, in 1903 US!, in 1906 US!, in 1909 US!). Map 25. Selaginella mutica var. texana Weath. in Jour. Am. Arb. 25:414. 1944. (Holotype: Moore & Steyermark 31 96 GH! fragment US!. Paratypes: Moon? # Steyermark 3046, in part, GH!; E. J. Palmer 30871 GH!; Sperry 428 US!. Tracy & Earle 273 GH! US! and Mex. Bound. Survey YU fragment NY! are excluded as types; they are inter- mediate between var. limitanea and var. mutica) . Leaves short-setate; cilia mostly or entirely ascending, piliform to dentiform. Sporophylls with the cilia mostly or entirely ascending, all or some dentiform. Sheltered or moist cliffs and rocky hillsides, on igneous rocks, 1300-2400 m. Southwestern Texas to southeastern Arizona. Representative specimens: United States. TEXAS Steyermark 3046 (GH, MO, US) ; ridge south of Emory Peak, Chisos Mountains, Brewster Co., 2300 m., June 22, 1931, Moore tf Steyermark 3196 (GH, MO, US); £. /. Palmer 30871 (GH, MO, US), 31951 (MO, US); Sperry 428 (US), new Mexico: mountains south of Demmg, Oct. 4, 1937, Goodding (GH, US); Dec, 1924, Slater (US); Sept. 10, 1899, Wooton (US), April 18, 1903 (NY, US), June, 1906 (US), Jan. 9, 1909 (US). Arizona: March, 1904, Ferriss (GH). 16b. mutica Fig. 19. Map 26. Watsonii 1905. Leaves muticous or nearly so; cilia entirely or predominantly spreading, pili- form. Sporophylls with the cilia entirely or predominantly spreading, piliform. The following specimens are intermediate between var. mutica and var. limitanea: Tracy tf Earle 275 (BM, F, GH, MO, NY, US) ; Mex. Bound. Survey, ow Exposed or sheltered rocky bluffs, cliffs and ledges, in crevices or soil pockets or usually covering rocks, on igneous rocks, sandstone or rarely limestone, 1450- 4300 m. Colorado and Utah to Texas and Arizona Representative specimens: United States. Colorado: 1874-1878, Bn 2423 Compton (NY) x^t w^ 1, T ' MeehaU (NY ' YU) ; E ' h SpenCe (NY) ; K - M - d A - F > Try™ 507 1 (CU, MIN (ARIZ DS, FI, K, MICH, MO, RM) ; E. P. Walker 365 (GH, NY, US), 'ut'ah: Graham 7592 ( US ) , 10004 ( US ) ; Maguire 18450 ( GH ) . TEXAS - o» 5083 (BM, GH, MO, US) ; Waterfall 4811 (GH, MO, NY), W 2 ri GH ' M °' NY) ; Wri * bt 2I15 (GH ' NY >- NEW Mexico: Arsene & Benedict 16643 (F, US); Drouet e? Richards 3311 (F, GH, MO, NY, US); Standley 5199 (MO, NY, US). Arizona: Harrow 3631 (GH); 1896, Tourney (NY, US), 1894, Tourney CNY). 19SS] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 47 S. EXTENSA ^ 25. S. MUTICA VAR UMITANEA 26. S. MUTICA VAR. ' MUTICA S. WRIGHTII 17. Selaginelxa Extensa Underw. in Bull. Torr. Bot. Club 25:131. 1898. (Holotype: Pringle 3QOO NY. Paratype: C. Mullet NY). Fig. 20. Map 27. Selaginella rupestris mexicana Conzatti, Fl. Tax. Mex. 1:150. 1939, without Latin descrip- tion, not var. mexicana Milde, 1867. (Based on Pringle 3900). Plant terrestrial or sometimes epiphytic. Stems very long to long, branches long, remote, with the older primary branches once to twice pinnate; leafy stems radially symmetrical or nearly so. Leaves herbaceous, none red, subulate to ligu- late-long-triangular to ligulate-lanceolate; base abruptly adnate, distinct from the stem in color, usually glabrous, sometimes pubescent; margins cilia te, the cilia dentiform, the longest cilia less than l / 4 as long as the width of the blade; apical portion usually revolute; apex acuminate, rounded, plane or nearly so in profile; muticous, the apex subulate, greenish-hyaline. Sporophylls with the margins eciliate to short-ciliate toward the base; apex carina te. Megaspores finely reticulate on the commissural face, the equatorial region strongly and irregularly papillate, granular on the outer face, pale orange. S. extensa is casually epiphytic and rather similar in habit to the pronounced epiphyte S. oregana of Rupestres. The revolute apical portion of many of the leaves, the muticous essentially unmodified leaf-apex, the habit and the unusual rugose megaspores combine to make this a very pend 870-1700 m. [Vol. 42 48 ANNALS OF THE MISSOURI BOTANICAL GARDEN Eastern Mexico and Jalisco. Specimens examined: Mexico, tamaulipas: Sharp et al. 50255 (MO, US), san luis potosi: McVaugh 10446 (US); Pennell 1/94/ (US); Las Canoas, Aug. 21, 1891, Pringle 3900 (F, GH, K, MO, US); Sharp 46264 (US). Veracruz: Bourgeau 2541 (GH, YU) ; Mohr 12 (YU); Purpus 6052 (F, GH, MO, US); Sharp 44168 (US), hidalgo: T. C. 6 E. M. Frye 2537 (NY, US) ; Hoogstraal tf Chase 7308 (F, MO, US) ; Kenoyer & Crum 3910 (GH) ; Sharp 45864 (US), puebla: Sharp 45375 (US), jalisco: McVaugh 11829 (US). 18. Selaginella Wrightii Hieron. in Hedwigia 39:298. 1900. (Holotype: Wright 828 m). Fig. 21. Map 28 Selaginella rupestris (L.) Spring f. Wrightii (Hieron.) Clute, in Fern Bull. 16:52. 1908. Stems with the older primary branches once to twice pinnate; leafy stems usually strongly to rarely moderately dorsiventral in position, length and shape of the upper and under leaves. Upper leaves ligulate-long-triangular to long-tri- angular; apex acuminate to convexly acute. Under leaves herbaceous, none red, longer to only slightly longer than the upper, linear; apex acuminate, flat to slightly rounded, plane or nearly so in profile. All leaves with the base abruptly adnate, distinct from the stem in color, or the under leaves with the base rather decurrent, pubescent or glabrous; margins cilia te, the cilia piliform, or dentiform only toward the apex, the longest cilia % to l / 2 as long as the width of the blade; setae absent, the apex being hardly modified, or if present, arising very evenly in form and usually in color, rarely up to % as long as the blade, stout, whitish- lutescent to lutescent, translucent to opaque. Sporophylls with the margins long- cihate toward the base; apex broadly rounded to carinate. Megaspores prominently rugose, pale to bright orange. ventral. The Wright ferentiated setae are the primary characters of the species. This is the only species for which I have sufficient information to class definitely as a calciphile. Exposed or shaded ledges and pockets of limestone, 800-2300 m. Texas and New Mexico to southern Mexico. Representative specimens: United States, texas: E. J. Palmer 11389 (GH, MO, NY, US) ; Rose & Fitch 17973 (MO, NY, US) ; Tbarp 43.514 (MO, NY, US) ; R. M. & A. F. Tryon 5034 (BM, GH, MO, P, US), 5035 (B, BM, F, GH, MO, NY, P, UC, US), 5037 (ARIZ, FI, GH, K, MO, RM, US); hills, Turkey Creek (near Cline, Uvalde Co.), June 25, 1849, Wright 828 (B, GH, NY, US), new mexico: Standley 40389 (GH, US). Mexico, tamaulipas: Bartlett 10561 (F, US), nuevo leon: Barkley 16144U (GH, MO); Harvey 988 (GH, US); C. H. Muller 2638 (GH, US), coahuila: John- ston 8607 (GH, MO), 8854 (GH), 9106 (GH, MO, US), chihuahua: Harvey 1 507 (GH, US) ; E. Palmer 455 in 188 5 (GH, US, YU). san luis potosi: A. A. Lundell (US). HIDALGO 9678 (G san luis potosi: A. A. Lundell 96 kwood (or Kirkwood) 20 (MO, US), 142 (GH, MO). puebla: E. B. Copeland 1 74 (US). Guerrero: Hinton le & Conzatti 1390 (GH). 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 49 19. Selaginella Steyermarkii Alston, in Ann. & Mag. Nat. Hist., XII, 7:638. 1954. (Holotype: Steyermark 50 50 1 BM!. Para types: Steyermark 471 1 5 BM!, 5I2Q2?M\). Figs. 22, 23. Map 29. Stems with the older primary branches once to twice pinnate; leafy stems strongly dorsiventral in position, length and shape of the upper and under leaves. Upper leaves subulate to subulate-long-triangular; apex long-acuminate. Under leaves papyraceous to herbaceous-papyraceous, usually some red, longer than the upper, subulate-lanceolate to ligulate-lanceolate; apex long-acuminate, flat or very slightly rounded, plane in profile. All leaves with the base abruptly adnate, distinct from the stem in color, pubescent to rarely glabrous; margins ciliate, cilia piliform, % to y 2 % to y a strongly attenuate, white to whitish-lutescent, translucent to subopaque. Sporo- phylls with the margins eciliate to short-ciliate toward the base; apex rounded to carinate. Megaspores finely to moderately rugose-reticulate, to nearly smooth on the outer face, more prominently and coarsely marked in the equatorial region, pale orange. This species and the next, S. Hansenii, are the only ones that usually have some red leaves on the stem. In S. Sartorii such leaves are present infrequently. They tend to occur particularly among the oldest living leaves and the youngest dead ones and often are more common on the under side of the stem. The long-acuminate apex of the upper leaves and the thin under leaves afford an ample separation from S. Hansenii, This species has the most robust leafy stems of any in the series and material collected in the growing state presents particularly attractive sprays. It is one of the very few species I know to be casually culti- vated; the Standley collection cited below was from the garden of Mariano Pacheco. Rocks and banks, 1630-2000 m. Guatemala and adjacent Mexico. Specimens examined: Mexico, chiapas: Matuda 1846 (US). Guatemala: Aguilar 1429 (F, US); Hatch tf Wilson 283 (US), 325 (US); Haupt J OS (US) ; Maxon d Hay 3410 (US) ; Standley 63102 (F, US) ; Steyermark 47115 (BM) , between San Sebastian Hacienda and large penasco above town, Dept, Huehuetenango, 2000-2200 m., 1942, Steyermark 50501 (BM), Steyermark 51292 (BM); Tuerckheim 8844 (GH, US); L. O. Williams 14299 (US). 20. Selaginella Hansenii Hieron. in Hedwigia 39:301. 1900, as Hanseni; maintained over S. Bolanderi. (Lectotype: Hansen 878 B!. Paratypes: Hille- brand B!; A. A. Eaton B!). Fig. 24. Map 30. Selaginella Bolanderi Hieron. in Hedwigia 39:300. 1900. (Holotype: Bolander 4511 B! fragment NY!). Selaginella rupestris (L.) Spring f. Hansenii (Hieron.) Clute, in Fern Bull. 16:52. 1908. Selaginella rupestris f. Bolanderi (Hieron.) Clute, in Fern Bull. 16:52. 1908. Selaginella rupestris var. Hansenii (Hieron.) Jepson, Man. PI. Calif. 42. 1923. Selaginella rupestris var. Bolanderi (Hieron.) Jepson, Man. PI. Calif. 42. 1923. [Vol. 42 50 ANNALS OF THE MISSOURI BOTANICAL GARDEN Stems with the older primary branches once to usually twice to rarely three times pinnate; leafy stems strongly dorsiventral in position, length and shape of the upper and under leaves. Upper leaves long- triangular to ligulate to lanceolate- long- triangular; apex convexly acute to acute. Under leaves herbaceous, usually some red, longer than the upper, linear-lanceolate to lanceolate; apex convexly acute to acuminate, flat to slightly rounded, plane to gently beveled in profile. All leaves with the base abruptly adnate, distinct from the stem in color, glabrous or pubescent; margins ciliate, the cilia piliform, or dentiform only toward the apex, the longest cilia l / 4 to rarely nearly l / 2 as long as the width of the blade; setae arising rather abruptly in form and in color, % to nearly J4 as long as the blade, stout to attenuate, milk-white to whitish, translucent to opaque. Sporophylls with the margins eciliate to short-cilia te toward the base; apex broadly rounded to carinate. Megaspores rugose to rugose-reticulate on the commissural face, less prominently rugose-reticulate to smooth on the outer face, with an equatorial ring, lemon-yellow. S. Hansenii is related to the previous species, S. Steyermarkii, by the dorsiventral leafy stem and the commonly occurring red leaves. These characters, however, may be the result of parallel evolution. The red leaves tend to be more common on the under side of the stem and in the area where living and dead leaves adjoin. The branches at the center of a mat tend to be rather erect and compact, those on the periphery prostrate, long-creeping and remote. This variation within a mat, together with its associated characters of number and spacing of the leaves, is similar to that found in S. tortipila of Arenicolae. Open or lightly shaded bluffs or cliffs and surfaces of igneous rocks, 330— 1350 m. ■ Central California. Representative specimens : United States, californ Co (B, MO, NY, US); 1892, A. A. Eaton (B, NY); Fisher's Cabin, Mokelumne River, Amador and Calaveras counties, April, 1893, G. Hansen 878 (B, MO, NY, US); Heller 2583 (F, GH, MO, NY, US), 5466 (GH, MO, NY, US), 8173 (F, GH, MO, NY, US), 10715 (GH, MO, US), 11802 (F, GH, MO, US); 1863, Hillebrand (B); R. M. (g A. F. Tryon 5067 (ARIZ, B, BM, CU, DS, F, FI, GH, K, MICH, MIN, MO, NY, P, POM, RM.UC, US, WS). 21. Selaginella carinata Tryon, spec. nov. Figs. 25, 26. Map 31. Caules breves, ramis brevibus propinquis; caules frondosae positione foliorum superorum et inferorum valde dorsiventrales sed forma minus. Folia carnosa, apice carinato vel valde carinato, aspectu laterali truncato vel abrupte declivi, base abrupte adnata cum caule distincta colore, setis abrupte productis colore sensim diversis. Sporophylla marginibus ad basem eciliatis. Megasporae laeves flavae. Typus: Rose, Pachano & Rose 2303Q (US). Stems short; branches short, approximate; older primary branches once to usually twice, sometimes three times pinnate; leafy stems strongly dorsiventral in position and less so in the shape of the upper and under leaves. Upper leaves 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 51 ligulate-long- triangular; apex convexly acute. Under leaves fleshy, none red, lanceolate to ligulate-lanceolate; apex acuminate to convexly acute, carinate to strongly carinate, truncate to abruptly beveled in profile. All leaves with the base abruptly adnate, distinct from the stem in color, glabrous; margins cilia te, the cilia piliform toward the base, dentiform above, the longest cilia % or less as long as the width of the blade; setae arising abruptly in form and evenly in color, ]/ 4 to l /z as long as the blade, rather stout at the base, tapering evenly to the tip, greenish- to lutescent-whitish and subopaque at the base, milk-white and opaque toward the tip. Sporophylls with the margins eciliate toward the base; apex strongly carinate. Megaspores smooth, yellow. S. carinata is a distinctive species and not evidently related to any others. The very short stems and branches and the fleshy leaves with carinate and truncate apex set it off from the other species of this series. Some strobili are slightly dorsi- ventral, particularly toward the base, apparently a development parallel to that in the African and Madagascar species. Central Ecuador. Specimens examined: bon '-W -' t • % * / ^"n • ^ i fc* ^* 'i /.-**. r^ U m j^*r*jf ^W. * 1 X * M • \ * ior\ X £}>M\ \j\ mfr ^T * ^fc*-^^J l3) ^^^^ <^i» WWN> »» *#•** .7^ * l r *y !l" >*** / ■ v * «. f Y\* *» . ■ ■ "f '• -^ •* * ■ _ ■ S. STEYERMARKII S.HANSENII S. CARINATA 'd S. NJAMNJAMENSIS [Vol. 42 52 ANNALS OF THE MISSOURI BOTANICAL GARDEN 22. Selaginella indica (Milde) Tryon, comb. nov. Fig. 27. Map 32. Selaginella rupestris (L.) Spring f. indica Milde, Fil. Europ. Atlant. 262. 1867. (Holo- ty pe : Hooker f. & Thomson ) . Selaginella longipila of authors, not Hieron., for example, of Alston, in Proc. Nat. Inst. Sci. India 11:214. 1945. Stems with the older primary branches once to twice pinnate; leafy stems strongly dorsiventral in position, length and shape of the upper and under leaves. Upper leaves subulate-long-triangular to long- triangular; apex acuminate. Under leaves papyraceous, none red, longer than the upper, long-triangular to ligulate- long- triangular to subulate-lanceolate; apex acuminate, flat, plane in profile. All leaves with the base abruptly adnate, distinct in color, glabrous to slightly pubes- cent; margins ciliate, the cilia piliform toward the base, dentiform above, the longest cilia */4 to l /z as long as the width of the blade; setae usually arising evenly in form, less so in color, % as long as the blade, strongly attenuate, tawny to lutescent-whitish, translucent. Sporophylls with the margins eciliate to short- ciliate toward the base; apex rounded to carinate. Megaspores rather finely rugose- reticulate on the commissural face, less prominently marked on the outer face, pale orange. This species was known as S. longipila from the time of the description of that species by Hieronymus, since the specimens were cited as from the Himalayas. However, the labels on Hieronymus' material were erroneous, the specimens being North Wight its dorsiventral leafy stem and thin under leaves. The localities taken from the literature (Map 32) are from Alston, in Proc. Nat. Inst. Sci. India 11:214. 1945. Cliffs, rocky banks and ledges, 700-2800 m. Nepal, Bhutan, northern and western India. Specimens examined: Nepal: Polunin 14.78 (BM). Bhutan: Griffith 2871 (BM) ; Ludlow et ai. 17039 (BM), 18606 (BM). India: Duthie 3727 (BM, US); Khasia, 5000 ft., /. D. Hooker tf T. Thomson (GH, NY, P, YU) . 23. Selaginella njamnjamensis Hieron. in Hedwigia 39:312. 1900. (Holo- f Fig. 28. Map 33 Stems with the older primary branches once to usually twice pinnate; leafy stems dorsiventral in position, length and slightly in the shape of the upper and under leaves. Upper leaves lanceolate to long-triangular; apex acuminate. Under leaves herbaceous, none red, longer than the upper, linear-lanceolate to lanceolate to long-triangular; apex acuminate, flat to slightly rounded, plane in profile. All leaves with the base abruptly adnate, distinct from the stem in color, glabrous to pubescent; margins ciliate, the cilia usually piliform toward the base, dentiform above, the longest cilia l / 4 to nearly l / 2 as long as the width of the blade; setae arising evenly in form and color, ]/ 4 or less as lone as the blade, stout to attenuate, 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 53 tawny to lutescent-whitish, subopaque to opaque. Sporophylls on the upper side of the strobilus lanceolate, acuminate, on the under side ovate-long-triangular to narrowly ovate, acuminate, slightly longer and broader; margins short-ciliate toward the base; apex broadly rounded to carina te. Megaspores granular to granular-tuberculate, yellow. This species is the least dorsiventral of the related species in Africa and Mada- gascar. Although sufficiently distinct from S. caffrorum, both species tend to overlap in the range of variation of the individual characters. The herbaceous under leaves with the base abruptly adnate and distinct in color and the poorly differentiated setae are the primary characters of S. njamnjamensis. On and between exposed rocks, 800-1200 m. Central and southeastern Africa. Specimens examined: Anglo-Egyptian Sudan: Dandy J/o (BM); Hoyle 520 (BM). 35S7 (BM). 6668 6095 furth 24. Selaginella caffrorum (Milde) Hieron. in Hedwigia 39:313. 1900. Fig. 29. Map 34. Selaginella rupestris (L.) Spring f. caffrorum Milde, Fil. Europ. Atlant. 262. \%67. (Holotype: Terra Caffrorum, Bunge). Selaginella rupestris var. incurva A. Br. in Kuhn, Fil. Afr. 213. 1868, category taken from annotation of Welwitsch 169 by A. Br. (Lectotype: Drege, Lycopodium rupestre a B. Paratypes: Ecklon S? Zeyber 7 B; Drege, Lycopodium rupestre aa B; Weltuitsch 169 B!; Quar tin-Dillon; Steudner). Selaginella rupestris var. incurva f. abyssinica A. Br. in Kuhn, Fil. Afr. 213. 1868, nomen nudum. A single description was provided for "forma capensis et abyssinica". I have seen Quar tin -Dillon, P, but not Steudner, the other specimen cited. Selaginella rupestris var. incurva f. angolensis A. Br. in Kuhn, Fil. Afr. 213. 1868. (Holo- type : Welwitsch 169BI). Selaginella rupestris var. incurva f. capensis A. Br. in Kuhn, Fil. Afr. 213. 1868, nomen nudum (see f. abyssinica). Selaginella capensis Hieron. in Hedwigia 39:314. 1900, epithet from S. rupestris t capensis A. Br. (Lectotype: Drege, Lycopodium rupestre a B. Paratypes: Ecklon # Zeyher 7 B; Drege, Lycopodium rupestre aa B; K. Baur IIIO B; Sonder B!; Goldschmid B; Kehmann 3924 B, 4005 B; Sutherland B). Selaginella Quartiniana Cufodontis, in Phyton 4:178. 1952, nomen nudum, based on S. rupestris var. incurva f. abyssinica A. Br., nomen nudum. Stems with the older primary branches once to usually twice to rarely three times pinnate; leafy stems dorsiventral in position, length and shape of the upper and under leaves. Upper leaves linear-long-triangular to ligulate-long- triangular; apex acuminate to convexly acute. Under leaves papyraceous to papyraceous- herbaceous, none red, longer than the upper, lanceolate to ligulate-lanceolate to triangular; apex acuminate to convexly acute, flat to slightly rounded, plane or nearly so in profile. All leaves with the base abruptly adnate, distinct from the [Vox.. 42 54 ANNALS OF THE MISSOURI BOTANICAL GARDEN stem in color, or the under leaves with the base slightly decurrent and blending in color, glabrous or rarely pubescent; margins cilia te, the cilia piliform, or sometimes dentiform toward the apex, the longest cilia 54 to J4 as ^ on S as t ^ ie width of the blade; setae arising rather abruptly in form and color, % to Yz a $ l° n g as the blade, rather stout, whitish to whitish-tawny or whitish-lutescent, translucent to subopaque. Sporophylls on the upper side of the strobilus ovate-lanceolate to long- triangular, on the under side lanceolate-ovate to broadly ovate, acuminate, longer and broader; margins short- to long-ciliate toward the base; apex rounded to carinate. Megaspores rugose-reticulate to granular-rugose on the commissural face, more finely and less prominently marked on the outer face, lemon-yellow. The thin under leaves and well-differentiated setae are characteristic of S. caffrorum and set it off from the related S. njamnjatnensis. In addition, S. caffrorum usually has some of the under leaves on the main stem with decurrent bases. Dry or moist rocky places, 600-2000 m. Anglo-Egyptian Sudan to Angola and Union of South Africa. Specimens examined: Anglo-Egyptian Sudan: Jackson 1116 (BM); MacLeay 114 (BM). Tanganyika: Greenway 2431 (BM). Angola: Antunes 343 (P) ; DekJndt 543 (K, P); Lopolo, Huilla, 5000 ft., April, 1860, Welwitsch 169 (B, K, NY, P). Union of South Africa, natal: Kudatis 8 $7 (P). orange free state: Kehmann 3924 (P), 3949 (NY, P). basutoland: Dieterlen iij (P). cape of good hope: T. Cooper 662 (NY) ; Katberg, 4000-5000 ft., Lycopodium rupestre a, Drege (GH, K, P, US) ; Ecklon ti Zeyher (B, P) ; Sander (B). 1 25. Selaginella echinata Baker, in Jour. Linn. Soc. 22:536. 1887. (Holo- type: Baron 4226 K!). Figs. 30, 31. Map 35 primary pinnate; leafy stems dorsiventral in position, length, shape and in setae characters of the upper and under leaves. Upper leaves ligulate-long-triangular; apex acute; setae arising abruptly, mostly soon deciduous, usually substellate, divided into two to several filiform branches, whitish, translucent to opaque. Under leaves her- baceous-papyraceous, none red, longer than the upper, ovate-lanceolate; apex acuminate Va opaque to subopaque. All leaves with the base adnate to abruptly adnate, distinct from the stem in color, glabrous; margins ciliate, the cilia piliform, the longest to usually x /% as long as the width of the blade. Sporophylls on the upper Va very ovate-triangular, acuminate, longer and broader; margins long-ciliate toward the base; apex broadly rounded to carinate. Megaspores tuberculate, yellow. S. echinata is the most distinctive of the three species of Madagascar. It is the most strongly dorsiventral species and in addition has the truly unique character of substellate setae on the upper leaves. Also some leaves and sporophylls have 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 55 a unique whitish to tawny, long-appressed pubescence on the back, which among the specimens I have seen is most pronounced in Perrier 8262. Madagascar, 500-2000 m. Specimens examined: Madagascar: Baron 4226 (BM, K) ; 1879, Cowan (BM) ; Humbert 2823 (P), 3000 ( p )> 3538 (P), 1 1778 (P); Humbert & Swingle 4791 (P); Perrier 1168 (P), 8261 p are mentioned under var. densa. Usually in rocky alpine tundra, also cliffs, talus slopes, on ledges, among boulders or in thin soil over rocks, on igneous or sedimentary rocks, 700-4660 m. Alberta to British Columbia, south to Texas, Arizona and northern California. Representative specimens: Canada, alberta: Scamman 2790 (GH); Rosendahl 1074 (US). British Columbia: Heacock 235 (BM, GH, MO, NY, US); Hitchcock tf Martin 7435 (GH, MO, NY); Shaw 398 (BM, GH, MO, NY, US), 902 (BM, GH, NY, US), 1060 (BM, GH, MO, NY, US). United States. Montana: Standley 15318a (US), 15598 (US), vicinity of Cracker Lake, Glacier National Park, 1740-1920 m., July 15, 1919, Standley 1 57 32 (US), Stand- ley 16216 (US), 16255 (US), 16288 (US), 16378 (US), 17055 (US), 17164a (US), 17979 (US), 18050 (US), 18185 (US); Aug. 25, 1917, Ulke (US); Vmbach 856 (F, MIN, NY, US). Wyoming: Goodd'tng 483 (F, GH, MO, NY, US); Mearns 4274 (US); Merrill & Wilcox 1218 (GH, NY, US). Colorado: Cox 290 (F, MO); Murdoch 4787 (F, US), utah: E. B. tf L. B. Payson 4037 (GH, MO). Texas: Moore & Steyermark 3245 (US), new Mexico: Metcalfe 1 172 (BM, F, GH, MIN, MO, NY, US). Arizona: 1913, Ferriss (US), idaho: Epling 7011 (F, MO, US). Washington: Flett Suksdorf 81 Holmgren 2 3036 (US). S (F, MO, US). Washington: Flett 3092 (US); Oregon: Ferris & Duthie 941 (US) ; Maguire & fornia: J. T. Howell I1S2I (US) : L. C. Wheeler 33b. Selaginella densa var. densa. Figs. 44, 45, 47. Map 45. Selaginella longipila Hieron. in Hcdwigia 39:291. Dec. 28, 1900. (Lectotype: Bourgeau in 1858 (Herb. Kew Exsicc. no. 1239, wrongly labeled "Himalaya") B!. Paratype: "Herb. Griffith, Bhotan" B! fragment NY!, collector and locality unknown). Selaginella rupestris (L.) Spring f. longipila A. Br. ex Hieron. in Hedwigia 39:291. 1900, in synon. Selaginella Engelmannii Hieron. in Hedwigia 39:294. Dec. 28, 1900. (Holotype: Engel- mann B!). 19SS] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 69 45. VAR. 46. S. DENSA VAR. STANDLEYI 70 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Selaginella Haydenii Hieron. in Hedwigia 39:296, Dec. 28, 1900, as Haydeni. (Lecto- type: Hay den B!. Paratype: Lyall B! fragment NY!). Selaginella rupestris var. densa (Rydb.) Clute, Fern Allies, 142. 1905. Selaginella rupestris f. Engelmannii (Hieron.) Clute, in Fern Bull. 16:52. 1908. Selaginella rupestris f. Haydenii (Hieron.) Clute, in Fern Bull. 16:52. 1908. Apex of leaves plane to abruptly beveled in profile; setae milk-white and opaque to whitish and subopaque. Sporophylls with the cilia dentiform to piliform and ascending, rarely laxly ascending, toward the apex; apex plane to abruptly beveled in profile. misapplication of S. longipila mixture i been discussed under S. indica. The type specimens are actually S. densa var. densa but fortunately Rydberg's well-known name has a few months priority. specimens under var pulorum are mentioned under that variety and imens 26, 1896, Holm (MO) (US). July 6, 1937, Wherry >P Me GH An example of var. densa with a tendency toward var. Standleyi is: Empire, Colorado, Aug. 27, 1874, Engelmann (B, MO, US). The following specimens are rather intermediate between all three varieties: Wyoming, Hermann 4665 (MO) ; Colorado, Ewan 12699 (Ewan) ; New Mexico, Arsene & Benedict 18056 (F) . Prairies, alpine meadows or dry rocky places, acidic rocks or sandy soil, 1100- 4000 m. Southeastern Manitoba to British Columbia, south to New Mexico and Ar Representative specimens: lzona. Canada Manitoba: Macoun ef Her riot 70373 (F,GH,NY). Saskatchewan: Boivin fif GtUett 8686 (MO); Fort Carlton, March 29, 1858, Bourgeau (Herb. Kew Exsicc. no. 1239) (B, NY, P, US, YU) ; Cowles 43 (F, MO) ; Gillett 6022 (MO), alberta: Breitung 5553 (MO); S. Brown 128 (GH, MO, US); Make tf Watson 1229 (GH). British Columbia: Calder & Savile 7677 (MO), 9221 (MO). United States. "Oregon", 49 N. Lat., 1858-59, Lyall (B, NY), north Dakota: Aug. 12, 1908, Lunell (GH, MIN, US), Sept. 10, 1908 (NY, US), south Dakota: Black Hills, 1853-54, F. V. Hayden (B, MO, NY, US); Hay-ward 315 (F, NY). Montana: Little Rocky Mountains, Sept. 1889, Havard (NY, US); Newberry (NY)- Rydberg & Bessey 3517 (F, GH, MIN, NY, US); R. S. Williams 534 (US). Wyoming: A. Nelson 6590 (NY) ; L. tf R. Williams 2126 3898 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 71 ^33c. Selaginella densa var. Standleyi (Maxon) Tryon, comb. nov. Fig. 46. Map 46. Selaginella Standleyi Maxon, in Smiths. Misc. Coll. 72 5 :9. 1920. (Holotype: Standley 17228 US!. Paratypes: Brown 95 US!; Standley 15363 US, 16970 US!, 17483 US!, 18 1 36 US!; Ulke on Aug. 25, 1917 US!). Apex of leaves predominantly or entirely truncate in profile; setae whitish- to greenish-lutescent to lutescent, translucent. Sporophylls eciliate or with the cilia dentiform or piliform and ascending toward the apex; apex truncate, rarely abruptly beveled in profile. Intermediates of var. Standleyi and var. densa and material of var. densa with a tendency toward var. Standleyi are mentioned under var. densa. Specimens of var. scopulorum with a tendency toward var. Standleyi are mentioned under var. scopulorum. Specimens of var. Standleyi with a tendency toward var. densa are such as: ndley 16970 (US), Van Schaack 2786 l / 2 (MO); Colorado, Johnston Jc%>p(GH). pulorum J. H. Schaffner (BM) ; Colorado, Underwood (MO) 500-4660 Southern Alaska to Colorado. Representative specimens: Alaska: A. & A. Krause 162 (B). Brown Holway & Butters (MIN, US) ; Raup & Abbe (GH, NY). United States. Montana: Barkley 1725 (GH, MO, US); Standley 15363 (US), vicinity of Sexton Glacier, Glacier National Park, Aug. 7, 1919, Standley 17228 (US), Standley 17483 (US), 18136 (US); Aug. 25, 1917, Ulke (US). Colorado: Johnston 3903 34. Hedwieia 39:290. 1900. Fig. 48. Map 47. Selaginella rupestris (L.) Spring f. sibirica Milde, Fil. Europ. Atlant. 262. 1867. (Lecto- type: Unalaska, Chamisso. Paratypes: Ajan, Tilling; Dahuria ad flumen Ingoda, Pallas). Selaginella rupestris f. amurensis Milde, Fil. Europ. Atlant. 262. 1867. (Holotype: Amur, Maximowicz) _:. manchu Khalkyli, Maack) . 1867. (Holotype: - j w j ^ - ^— — w w— — w -mm m w Selaginella Schmidtii Hieron. in Hedwigia 39:292. 1900. (Lectotype: Schmidt B! frag- ment NY!. Paratype: Chamisso B! fragment NY!). Selaginella Schmidtii var. Krauseorum Hieron. in Hedwigia 39:293. 1900. (Lectotype: A. & A. Krause 53 B! fragment NY!. A. ti A. Krause 162 B! is excluded as a type- it is S. densa var. Standleyi) . Selaginella rupestris i. Schmidtii (Hieron.) Clute, in Fern Bull. 16:52. 1908. Stems long to moderately long, forming open, spreading mats; branches long and remote to moderately short and subapproximate, intricate, not fragmenting * m . _- . « » • 1 . 1 - 1 _ _ J when dry. bequal in length, or the under [Vol. 42 72 ANNALS OF THE MISSOURI BOTANICAL GARDEN slightly longer, on the same portion of the stem; leaves linear to ligulate-long- triangular; base usually glabrous, less often pubescent, that of the upper leaves adnate to the stem for % their length; margins ciliate, the cilia spreading to laxly ascending toward the apex; apex of the upper leaves fleshy, broadly rounded to carinate, subtruncate to usually truncate in profile; setae forming conspicuous tufts at the dry branch-tips, % to % as long as the blade, scabrous, milk-white to white to tawny, opaque to translucent. Sporophylls eciliate or with the cilia piliform to dentiform and ascending toward the apex, the broadest 3 to usually 4 times as broad as the leaves; seta base usually strongly broadened and flattened. Megaspores rugose to rugose-reticulate, pale yellow to pale orange. The Alaskan and Yukon material is relatively uniform while that from Asia is less so. In addition to the typical form with short, milk-white setae a phase occurs in Asia with longer and tawny setae, and also occasional specimens bear strobili having 1 to 2 megaspores in a sporangium. This is quite parallel with S. densa var. densa, the reduced number of megaspores being so rare that it is not possible to be certain of apogamy. S. sibirica is most similar to S. densa var. Standleyi from which it may be sep- arated by the white to tawny rather than lutescent setae and the intricate rather than discrete branches. It is separated from S. shakotanensis, which also grows in Japan, by the leaf-bases that place the two in different series. The seta length likewise distinguishes these species — in S. sibirica they are %— % as long as the blade while in S. shakotanensis they are I/5 as long. The only sporelings observed in this study were seen in the soil of a mat of S. sibirica collected by Calder & Billard (2995 MO). These are discussed in some detail in the introduction. ocky Representative specimens: Japan ton 4454 (MO), 460 Alaska: unalaska, Chamisso (B, NY) ; Flett 1529 (NY, US) ; A. E. & R. T. Porsild 690 (GH, MO, US); Scamman 198 1 (GH, MO). Behring Straits: Emma Harbour, Sept. 21, 1881, A. # A. Krause 53 (B, NY); C. Wright (GH, NY, US, YU) . Union of Soviet Socialist Republics, kamtchatka: Eyerdam (F) ; Novogablenov 541 (US), sachalin: Fr. Schmidt (B, GH, NY, US), trans Baikal: Ingoda river (Dahuria), Fischer (GH, P). 1 100 (P), 7279 (P) 9 S4g6 (MO, P), 9687 (MO, P), 13903 35. Selaginella Watsonii Un Watsoni. (Holotvoe: Watson 1898, as Watson Hansen 879 NY!; Coville & Funs ton 2071 NY!; Brewer 2103 NY!). Figs. 49-51. Map ming subapproximate to moderately long and remote, intricate, not readily fragmenting dry Upper and under leaves equal in 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 73 length, or the under slightly longer, on the same portion of the stem; leaves ligulate to broadly ligulate to ligulate-long- triangular; base glabrous to slightly pubescent, that of the upper leaves adnate to the stem for % to % their length; margins eciliate, rarely with ascending cilia toward the apex; apex of the upper leaves fleshy, narrowly to broadly carinate, subtruncate to truncate in profile; setae not or hardly forming conspicuous tufts at the dry branch-tips, */$ to %, rarely x fe y as long as the blade, usually smooth, sometimes slightly scabrous, greenish-white to greenish- lutescent to whitish-lutescent, translucent to subopaque. Sporophylls eciliate toward the apex, the broadest about 2 times as broad as the leaves. Megaspores rugose to rugose-reticulate, pale orange. S. Watsonii is characterized by its fleshy leaves which are usually eciliate and short-setate. In habit it forms dense rounded mats with intricate branches. The fragile nature of the stem of the next three species is poorly developed in S- Watsonii. The stems can be easily broken by hand but specimens do not fragment in packets or with ordinary handling. It seems probable that the next three species have been derived from S. Watsonii or a common ancestor. Exposed or shaded cliffs, talus slopes or rocky alpine meadows, in crevices or on boulders or among rocks, usually related to igneous rocks, rarely to limestone, 1800-4300 m.. usuallv 3000- 3700 m. Southwestern California. S. ASPRELLA 74 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Representative specimens: United States. Montana: T. D. Howe 66 (US), utah: G. /. Goodman (GH, MO, NY, US) ; M. £. Jones 1246 (NY, US) ; Maguire tf Richards 13166 (GH, MO) ; E. B. (j L. B. Payson 4917 (GH, MO, US), 5074 (GH, MO, NY, US) ; Rydberg 8 Carlton 6566 (GH, NY, US); Cottonwood Canon, 9500 ft., July, 1869, S. Watson 1370 (NY, YU). Nevada: Heller 11054 (F, GH, MO, NY, US); Maguire 2IOQ0 (GH, NY, US); Sept., 1868, S. Watson 1370 (NY, US). Oregon: July 1, 1931, Wherry (US). California: Brewer 2103 (NY, YU); Coville tf Funston 2071 (NY, US); Sept. 5, 1923, C. C. Hall (F, MO, NY, US) ; Hansen 879 (BM, MO, NY, US) ; Heller 7167 (GH, MO, NY, US) ; R. M. ef A. F. Tryon 5060 (ARIZ, B, BM, CU, DS, F, FI, GH, K, MICH, MIN, MO, NY, P, PH, POM, RM, UC, US, WS) . 36. Selaginella utahensis Flowers, in Amer. Fern Jour. 39:83. 1949. (Holo- type: Cottam 5644 UT fragment US!. Paratype: Cottam 8817 UT). Fig. 52. Map 49. Stems moderately short, forming rounded to flat cushion mats; branches mod- erately short to moderately long, approximate to nearly remote, intricate, readily fragmenting when dry. Upper and under leaves equal or subequal in length on the same portion of the stem; leaves subulate to ligulate-lanceolate to ligulate-long- triangular; base usually glabrous, sometimes pubescent, that of the upper leaves adnate to the stem for l / 4 their length; margins eciliate to ascending-ciliate toward the apex; apex of the upper leaves slightly fleshy, broadly carinate, gently to rather abruptly beveled in profile; setae not forming conspicuous tufts at the dry branch- tips, less than % as long as the blade to essentially absent, smooth, whitish to greenish- to lutescent-whitish, subopaque. Sporophylls eciliate or with the cilia dentiform and strongly ascending toward the apex, the broadest 2 to 3 times as broad as the leaves. Megaspores slightly rugose-reticulate to slightly rugose, yellow-orange. S. utahensis may be separated from the next two species, that also share the character of readily fragmenting stems, by its leaves which are muticous or have short, smooth setae. The anatomical basis of the fragmenting stems has not been investigated. In mounted material the parts will be held together if glue or soil holds the roots and rhizophores firmly but in packets or in unmounted material ordinary handling of the sheet will soon reduce the stems to small pieces. Ledges and crevices of sandstone cliffs, 1500-2300 m. Southern Nevada and southwestern Utah. Specimens examined: United States, utah: Lady Mountain, Zion National Park, 6,500 ft., April 5, 1931, W. P. Cottam 5644 (MO, US) ; Degener & Peiler 16982 (NY) ; Eastwood ef Howell II 59 (US); Flowers 3249 (MO). Nevada: June 28, 1930, E. Jaeger (US). 37. Selaginella leucobryoides Maxon, in Smiths. Misc. Coll. 72 5 :8. 1920. (Holotype: Munz & Harwood 3789 US!. Para types: Coville & Funston 628 US!; Munz, Johnston d Harwood 4226 US!). Fig. 53. Map 50. Stems short, forming flat or rounded cushion mats; branches short, approxi- mate, intricate or discrete, readily fragmenting when dry. Upper and under leaves 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 75 equal in length, or the under slightly longer, on the same portion of the stem; leaves linear to ligulate-ovate; base glabrous to pubescent, that of the upper leaves adnate to the stem for l / 4 or less their length; margins eciliate to ascending-cilia te toward the apex; apex of the upper leaves herbaceous, slightly to broadly rounded, gently to abruptly beveled in profile; setae not forming conspicuous tufts at the dry branch-tips, % or less as long as the blade, scabrous to slightly scabrous, milk- white and opaque, rarely white and translucent. Sporophylls eciliate or with the cilia dentiform and ascending toward the apex, the broadest about 3 times as broad as the leaves. Megaspores slightly rugose to rugose-reticulate, to nearly smooth on the outer face, pale orange. The short, white, scabrous and usually opaque setae are characteristic of S. leucobryoides. It shares the peculiar character of fragile stems with S. utahensis and S. asprella. Rocky slopes, in crevices or among rocks, 900-2300 m. Southeastern California. Specimens examined: United States. California: Coville & Funs ton 628 (NY, US); Coville & Oilman III (US); /. T. Howell 3989 (F, GH, MO, US); Bonanza Mine, Providence Mountains, 2800 ft., March 30, 1920, P. A. Munz & R. D. Harwood 3789 (F, GH, NY, US) ; Munz, Johnston tf Harwood 4226 (US). 38. Selaginella asprella Maxon, in Smiths. Misc. Coll. 72 5 :6. 1920. (Holo- type: Johnston 18 15 US!. Paratypes: Johnston 1595 US!, 1807 US!). Fig. 54. Map 51. Stems short to moderately short, forming rounded or flat cushion mats; branches moderately short and remote to short and subapproximate, intricate, rather readily fragmenting when dry. Upper and under leaves essentially equal in length on the same portion of the stem; leaves linear-lanceolate to lanceolate-ovate to lanceolate-long-triangular; base pubescent, rarely glabrous, that of the upper leaves adnate to the stem for about % their length; margins eciliate to ascending-ciliate toward the apex; apex of the upper leaves herbaceous, broadly rounded to carinate, nearly plane to truncate in profile; setae forming conspicuous tuhs at the dry branch-tips, % to over l / 2 as long as the blade, scabrous, white and translucent to rarely milk-white and opaque. Sporophylls eciliate or with the cilia piliform to dentiform and strongly ascending toward the apex, the broadest about 3 times as broad as the leaves. Megaspores prominently and coarsely rugose-reticulate, most prominently marked in the equatorial region, pale yellow to pale orange. S. asprella is characterized by its long, scabrous, white, usually translucent setae. Of the three species with readily fragmenting stems, it is the furthest re- moved from the presumed ancestor, S. Watsonii. Open areas in rocky soil or in crevices at the base of boulders, igneous rock, rru Southern California. Specimens examined: [Vol. 42 76 ANNALS OF THE MISSOURI BOTANICAL GARDEN United States. California: /. T. Howell 5031 (US); Jaeger 276 (US); /. M. Johnston 1595 (MO, US), 1807 (MO, US), West end of Ontario Peak, San Antonio Mountains, 6000 ft., March 25, 1918, /. M. Johnston 1815 (MO, US) ; Munz 7612 (NY), 9683 (US), 17165 (GH); Sept. 14, 1921, Saunders (US); April, 1906, Streeter (NY, on WS) Series Eremophilae Tryon, ser. nov. Rhizomata et stolones et gemmae nulla. Caules prostrati. Apices ramorum involuti statu inerte. Caules frondosi valde dorsiventrales. Folia supera base abrupte adnata caule distincta colore, folia infera base valde decurrente. Typus: Selaginella eremophila Maxon. Plants terrestrial; rhizomes, stolons and basal buds absent. Stems prostrate or assurgent at the branch tips, with rhizophores produced generally throughout; stems long to rather short, forming open or usually rather compact, rounded or usually flat mats. Branches moderately long to short, remote to subapproximate, usually intricate, branch tips involute in the dormant state. Leafy stems strongly dorsiventral, the under leaves appressed, the upper erect, the under longer, usually thinner and different in shape from the upper; zone of green leaves much longer on the upper side of the stem, very short to absent on the under side. Upper leaves with the base abruptly adnate, distinct from the stem in color, under leaves with very Eremoph, arizonica is increasing specialization in the dorsiventral habit. S. peruviana is stron ventral in position, length and shape of the upper and under leaves; in S. they are also different in texture. In S. eremophila the setae are highly 1 m and the leaves of S. Parishii are muticous. Finally, in S. Landii, the under leaves are unusually long in relation to the upper and they are setate while the upper leaves are muticous. The species are all quite distinct with the exception of the peruviana KEY TO SPECIES a. All leaves setate (sometimes only in the bud) ; under leaves less than twice as long as the upper, b. b. Setae mostly or entirely persistent, stout, straight, c. c. Under leaves subulate, acuminate, broadest at or very near the base, rarely linear-lanceolate, not appreciably thinner than the herbaceous upper leaves, setae 0.5-1.0 mm. long. Oklahoma to New Mexico; Mexico; Peru to Argentina 39. S. peruviana, p. 77 c. Under leaves lanceolate to lanceolate-ovate, broadest above the base, papyraceous, the upper fleshy, setae 0.1-0.3 mm. long. Texas to Arizona; Baja California 40. S. arizonica, p. 78 b. Setae mostly early- deciduous (sometimes present only in the bud) , deli- form >phila 19S5] TRYON— SELAGINELLA RUPESTRIS AND ITS ALLIES 77 a. Upper leaves, or all leaves muticous. d. d. All leaves muticous; under leaves about twice, or less, as long as the upper; upper leaves lanceolate to lanceolate-triangular... 42. S. ?amhn ? . 80 d. Upper leaves muticous; under leaves predominantly short-setate, the setae persistent; under leaves over twice as long as the upper; upper leaves long-deltoid or with parallel sides toward the base 43. S. Landu, p. 81 39. Selaginella peruviana (Milde) Hieron. in Hedwigia 39:307. 1900 Figs. 55, 56. Map 52. Lycopodinm ciliatum Ruiz ex Spring, in Nouv. Mem. Acad. Roy Belg. 24 (Monog. Fan, Lycopod.):55. 1850, in synon. (evidently based on R«,z p^B). SrlaJnelU rupestris (L.) Spring f. Peruvian, i Milde, Fd. Eu~P- Atl^t « 3 1867 . £Holo type: Ruiz 9 8 B!; the lower right-hand specimen of W tikes Exped. J, Peru, Gttl Sel^neUatpestris var. mexicana Milde, Fil. Europ. Atlant. 263. 1867. (Aschenbom 65 B! is so annotated by Milde or A. Br.). .-,,-. T ,t, r u <>2-417 1896 Stla^lla rupees va, peruvun. (Mildo, H.eron ■ ,*« . Bo, £££f '£,!£„ Srl.gwrHa Aichenbornii Hieron. in Hedwig.a MiJOS. 1? 00 ' l "'™ £ A4 ,.„„ ,,, B!; 05 B!; PaW 554 GH! and US- (62778) ,s '^""''Vv Ja^/b Tolut, r™ gW «1 £U B, H* i. ..« B, Meaicc , C, y, SrW* B, Toluu, . . Chrismar B. Vancouver Island, /. ti. nenry d snouiu Wallacei) Bang III B!. Dombeyana Hieron. in Hedwigia_39:308. 1900- (Lectjype: Ba*g 171 B!. Paratypes: M<.»do» <3 4 *»5 ^"-^ *£ -• — ; ' . , lg74 1892, O. Kuntze B; Argentina, Hi,ra«^«s tf Lorentz 162 B, Argentma, Jan. Lorentz & Hieronymus V) . _ . M , w1 ,010 as sheldoni. (Holo- Srb*f«tfe SfeMmti Maxon, in Proc. Biol. Soc. Wash. 31:171. 1918 a . U*ao ^ type: Sheldon 2 33 US!. Paratypes: F. C Grm,, on June 10, 1918 US., M 7 1918 US!; Jenny 343 US!; HW US!; Wooton US!). Upper leaves herbaceous, or fleshy only at the base, linear, acuminate to U» **- long-triangular; under leaves herbaceous, subulate, acummate broadest at or vej near the bL, to rarely linear-lanceolate, a little longer than the upper on th same portion of the stem; leaves setate in the bud, setae percent (part may be ^abraded) , stout, straight, those of the under leaves 0.5-1.0 mm. long, rarely some 0.3 «£ long. Megaspores rather coarsely rugose-reticulate on the commissural face, less prominently marked on the outer face, yellow to bright orange r J -«..i__ • ~~A tr\r rhi<; reason it 5. peruviana is the least specialize «-- — „ rttT ,;„^ ( 54^4 (20); 6296 (15); 6711, 7355> 74S6 (2); 7723, IOO70 (20). Abrams & McGregor 760 (35). Abrams & Wiggins 207, 341 ( 2 ) . Aguilar 1429 (19). Alexander & Kellogg 2062 (2). Allard 3570, 4570b, 4610a, 5488, 6475 (32). Amable 1 38 3 (39). Amortegui A6l (9). Anderson, A. P. 1894, 1975 (7). Anderson, J. P. Andrieux 2 (9). Anect 166 (39). Antunes 1427 543 (24). Antunez & Dekindt 3040 Archer 6764, 7129 (35). Archer & Gehrt 1 18 (8). Arechavaleta 472 (8). Arse ne 1020$ 10641 (18); 9983 (12); 10639 (*); 10640 (i) 10642 (1); 10643 (12) 17777 (33 a-b, 33b); 1 78 1 7 (5); 17929 (5); i 793 o (31); 17931 (5); 17945' I7903, 17972, 17973, 17974, 17975, 17976, 17977, 17978 (31); l798o (5); 17982, 18049 (31); 180 m (39); 18052 18342 18666 20174 (39); 20985 (31); 20987, 20988, 20989, 20990, 20992 (39); 21107 (33a-b); 2IIl8 (16b); 21120 (33a-b). EXSICCATAE Arsene & Benedict 15872, 15873, 15874, 15875 (31); IS&7& ( 5 ) ; 15877, 15878, 1603I, 16033, 16152 (33b); 16370 (33a-b); 16402 (33b); 16643 (16b); 16644 (33a); 1 6645, 16838 (39); 16884 (31); 18053 (3D; 18054 (33a-b) ; 18055 (33a); 18056 (33a, 33a-b-c). Aubineau 25 (31). Austin 241 (20). Bachmann 9 (28). Bacigalupi 1015 (16b); 2750 (15). Baker, C. F. 2(31); 16, 2746 (15); 5234 (2). Baker, M. F. 1,5 (3b). Baldwin 5/J3<5 (31). Clausen 3674, 3679 (7). Clausen & Trapido 3663 (32). Clements, F. E. & E. S. 279 (31); 532 (33b). Clokey ^9, 57^, 5765 (2). Clokey & Anderson 6508, 6532 (2). Colby 3364 (32). Cole 5135 (32). Conard 774 (30); 246 (15, 30); 7-930 (32). Constance & Rollins 998 (15). Cook & Gilbert 1 58, 240, 244, 248, 305, 608 (39). Cooper, J. G. 462 (2). Cooper, T. 662 (24). Copeland 774 (18). Core 3869 (32). Correll 6169 (3c) ; 6615 (32) ; 6711 (7) ; 10857 (32); 73627 (6); 73 38438, 39454 (18); 40232 (40); 47099 (3a); 42743 (18); 44*37 (6); 405^ (1). Cottam 3674, 3754, 4458 (35); 564/ (36). Coville 727 (2); 186 (20); 850 (30). Coville & Funston 707 (2); 628 (37); 2077 (35). Coville & Gilman 777 (37). Cowles 43 (33b); 428a, 455 (15); 579 (30). Cox 290 (33a). Crandall 7054 (31). Cronquist 4955 (3c); 5647 (15). Culbertson 4790 (20). Curtiss, A. H. 3796 (32); 6774 (3b). Cutler 7007 (40); 2580 (32); 3220 (3a). Dandy Daniels Darker 5IO (23). '5' (31). 3579 (15); 5861 (35). Darrow 2847 (35); 3637 (16b). Darrow & Phillips 2777, 278 1 (31) 19SS] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 85 64053 16982 4.964 3337 3^35, 1046^ Davis, R. J. 941 (15). Deam 2 1 58 4 (32); 58692, 63923, (3c). Degener & Peiler 1 61 1 4 (33b); (36). Dekindt 543 (24). Demaree 3251B (32). Diehl 2517, 2565 (39). Dieterlen 1 17 (24). Dixon 303 (16b); 335 (35). Dodge, C. K. 335 (32). Dombey 14 (39). Doty 1222 (15); 1868 (30). Drake JOJJ (22). Drouet & Cohen 4959 (33a-b); (16b). Drouet, F. & R. 4051 (2). Drouet & Richards 33 1 1 (16b); (5); 3780 (1). Drouet, et al. 3358, 3359, 3633, 3669, 3687 (1); 4097 (16b). Drushel J 1 128 (32). Drummond 352 (3a). Dudley 7495 (35). Duncan 7867 (3c); 17775* (1). Eastwood & Howell JI59 (36); 3524 (33a) 18808 (2). Edwards 4/p (18). Egeling 26 19 (28). Eggleston 2 1 99, 2905 (32); /odtfp (15); 17487 (39). Ehlers 7274 (32). Epling 7033, 7284, 10418 (33a). Evers 32650, 32668 (32). Ewan 2751 (3 5); 4240 (2) ; JI293 (33b) ; 1/665, II730, 1 1827 (33a); 720^7 (33a-b); 12659 (33b); 72<5pp (33a-b-c) ; 'l?4J>'55<5p(33a); 13267 (Hb); 14048 01*); 14370 (5); /4P50, 75J02 (33b); /547/tf (33a-b); 18377, 18395 (33a); /(fyfo (33c); 18586 (33a). Exell & Mendonca 169 (28). Eyerdam 2^7p 39); &*V, £255 (1). Gerard 1943 (28). Gibbs 4 (28). Gibert 1 323 (8). Gillett, J. B. 4681, 468 1 A (10b). Gillett, J. M. 6022 (33b); 6046 (15). Gilman 25 (31). Glaziou 3306 (8). Gleason p^07, 94 lS > 9 6l 3 ( J 2). Glover & Gilliland 480, 77* O ob )- 86 [Vol. 42 MISSOURI Godfrey GH1657 (28); 398 1 (32); 4677 Haynes 2648, 2649, 27 1 1 (41); 2785, (3c). Godman 6 (10b). Godman, et al. 690 (33a). Goldberg 19100 (28). Gomez & Sousa 85 (28). Goodding 2 (41); 4 (2); 5 (31); 56 (16b); IOI, 102, 103 (40); 173, 178 (16b); I** (1); 213 (31); ^5 (33a); 722 (40); 2444 (1); 6/4/ (39); 6142 (16b). Goodman do> (33b); 1 8 58 (35); 2JJ7 (39). Goodman & Hitchcock 1 514 (35). Gorman 675, 4594 (15). Gossweiler ^2p O0 (3b); 1957,1987 (3c) ; 3//<5 (3a-b) ; 3403 (3a, 32) ; 3762 (3a). Harris, B. B. 708 ( 1 8 ) . Harris, C C. & S. K. 3329 (15); 5555 (30). Harrison 9160 (35). Harrison & Larsen 7634 (35); 7697 (33a) Hartman 56,327 (1). Harvey ^o*, J 773, 720> (18); 7454 (39); 7454* (i); '5 5<5, 94884, 99600 (8). Hildebrandt 2363 (28). Hill, R. T. 75 (18). Hinckley 7077 (31); 7755 (16a); 7756 (31); 7757(39); 7756 (31) ; 2747 (40) ; 3566 8423 Hinton 3453 ( 9 ) ; 697b , , . Hitchcock, A. S. 22546 (39). Hitchcock, C. L. & Martin 4688 (33a); 4705 (15); 736> (15, 33a). Hitchcock, C. L. & Muhlick 6*725 (33a); 8499, 0048, 9099 9Sl3 (33a). Hitchcock, C. L. & Stanford 6780 (40) ; 7248 (12). Hodgdon 6ll (32). Hoogstraal & Chase 7235 ( 1 Hoover 7399 (35); 4770 ( 5085 (15); 5398 (35). Horner 607 (15). Hosie, et al. 4, 1 46 (32). Hotchkiss 161 4 (32). House l6S2, 4707, 7308 I 4900 _. 4274, 4307 (32) ; 490*2 496*7 (15); 70279, 76*726*, 21482, 22671, 27974, 28004 ( 32 ) • Howe, M. A. 999 (30). Howe, T. D. 66 (35). Howell, J. T. 2834, 2886 A (2); 2946" (11); 396*9(37); 4557(15); 4692(20); 5037 (38); 7600 (15); 73527 (33a); 16761. 26066 (35} : 19SS] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 87 Howell, T. 696 (30). Hoyle 520 (23). Humbert 2823, 3000, 3538 (25); 7089US (27); Il6l2 (26); 11778 12439, 12706, 13338, 14.158 (27) Humbert & Swingle Jones, M. E. & Eggleston 19596 (2) Jorgensen 4073 (8). (>SS7, Julio IOO (39). (25); Jiirgens 29, 308, 309 ( 8 ) . 4791 (26); 5581, 5705 (27). Hunnewell 8365 (15); 9490 (15); 14518 (32). Jermy 542 (3a); 343 (39). Johnston, I. M. 7044 (2); 1047 (41); I595> 1807 (38); 1808 (2); 1815 (38); 1 82 4, 1825, 1S26, 1827, 1828 ( 2 ) ; 2420 (5); 2421 (5, 31); 2422 (5); 2423 (16b); 2424, 2425 (31); 3004 (11); 5005 (2) ; 3870 (31); #7/ (16b) ; 3872 (5); J*7J, J*74 (33b); 3875 (33a); i?7<5 (33c); 5577, 3878, 3879 (33b); J&?/ (33a); 3881a (33c); Jo^te (33a); jtffetf, 3883 (33c); Jcftfy (33a); 3^4^ (33c); 3885 (33a); J&W, j&?7 (33c); 3887a (33b); jm<* (33c); 3888b, 3889 (33b); j 59^9 (39). Knopf JJJ (2). Knowlton, F. H. 82, 103 (33a) ; /4/ (31). Krause, A. & A. 53 (34); 702 (33c). Kreager 294 (15). Lacas 512 (18). Larsen 5998 (16b); OOotf (35). Lawrence, W. E. 7 (16a); 8007 (16b); /0446, Napier 2054 (10b). //o*29 (17); /4660 (35). Nash 1449 (3c). 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 89 Nelson, A. 1345 (33a, 33b); 2448 (33b); 6368 (33a); 7811 (33a-b, 33b); 8781 (33b); 10676 (16b); JI217, 11217a (40). Nelson, A. & E. 6368 (33a). Nelson, A. & R. A. 1 1 58 (40); 1 448 ( 1 ) J 5765 > 5766, 5782a, 5782b (33a). Nelson, J. C. 3876 (15). Nielsen 3423 (32). Nil 105 (42). Norton 1 405 (32). Novogablenov 547 (34). Nuttall, L. W. 553 (2). I O'Neill 75 IH$> 1149, H50, Pease 1 1 5, 1 29 50, 1 7967 (32); 22298 (33a); 22300 (15); 22^/p (33a); 22499 (33c); 24rp5 (32). Pease & Bean 23334, 23581, 26103, 26118 (32). Pease & Ogden 25070, 251 14, 25140(52). Peattie 899, QOO (7). Peck 15996a (15). Peebles & Harrison 6984 (1). Peebles, et al. 480 (40); 4657 (1). Pegler £?2 (28). Pennell J6890 (18); 77/p# (9); 17272 (42); 77470" (39); 77753 (9); J7P47 (17); /o^d (39); 18178, 1 8646, 1 9533 (1). Perrier 7/6 (3a); 10146, 1 1 389 (18); JJ^// (3a); 27399, 27461 (3c); 2^^, 2 p/ (32); 30543, 30661 (39) ; 3087J (16a) ; 3J<5op (32); 31951,31952 (\Sz); 32800 (32); 555&* (3a); 35561, 36810 (32); 37964 05); jtfj/o' (3c). Palmer, L. J. 235 (34). Palmer, fm. & King /tfd, 21 4 (32). Parish, S. B. 1200 (41); 4724'5 (3c); 3'0/ (3b-c); 3™7 (3b) 3774 (3b-c); 377p (3c); 3'5* (3b) 3755 (3c); 3175 (3b); 3775*(3b, 3b-c) Quer 7 (4). 90 [Vol. 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN Raup 630S, 641 1, 6607, 6929, 8 1 84 (32). Raup & Abbe 3910 (33c); 4451, 4608 (32). Rawlinson 284 (32). Reed, F. 6180 (2). Reed, F. M. 2244 (2); 2466 (35); 4070 (2). Reeve IOO (32). Rehmann 3924 (24) ; 3949 (24, 28) ; 4005 (2*)-> 4333, 5576 (28). Reko 5280 (1). Rendle 50/ (28). Reverchon 1 5 (18); I^5<5 (39); 1632 (3a). Richards & Drouet 363 (31). Riddell J<5, //p7 (3a). Riedel 7 (8). Robinson, C. B. p/5 (32). Robinson, T. R. & Maxon 1 5, 32 (32). Rodgers, L. 404 (7). Rodin 4403 (28). Rogers, F. A. 5683 (28). Rogers, H. T. 763 (15). Rolland 6131 (32). Rollins 51227 (35). Rose, J. N. 2879 (1). Rose, J. N. & Fitch 17973 (18); -^7/ (39). Rose, J. N. & Painter 6536 (39); 6694 (1); 6835 (S);7499 (43). Rose, J. N. & J. S. 1 1 199 (I). Rose, J. N. & Mrs. 18845, 19466, 19468 (39). Rose, J. N. & Russell 19937 ( 8 ). Rose, J. N. et al. 8268 (39) ; 9614, 12857 (I); 23039 (21). Rose, L. S. 34495 > 400l6 (2) ; 45039 (20) ; 45342 (41). Rose-Innes 41026 (18). Rose-Innes & Warnock 783 (3a). Rosendahl 448 (32); #64 (15); 1074 (33a); 2<5&2 (32). Rosendahl & Butters 4564 (32). Rosendahl, et al. 1 51 5 (15). Rousseau 26929 (32). Rudatis £57 (24). Ruiz 98 (39). Ruth 104, 706, 749 (32). Rydberg 1 184 (33b). Rydberg & Bessey 3517, 3518, 3519 (33b). Rydberg & Carlton dj/o, dj<5<5, 7/ 18050 (3 3 a) ; 18 1 36 4826 (32); 5008, 5009, 5010, 501 I, 5012, 5013 (3c) ; 5014, 5016, 5017 (3b) ; 5018, 5019, 5021, 5024, 5026, 5027, 5093 (3c). (33c); 18185, 18606 (33a); 40J<%>(18); Tryon, R. M. & A. F. 5005 (32); 50/5 52533 (3c); 63702 (19); 73579, 92638 (3c). Standley & Bolman 10830 (31); 77963 (32). Stork & Horton 9388, 10719 (39). Suksdorf 4 (15); 47, 2.25 (33b); 20S 05) ; 753 (20) ; 1 3 64, 2085, 2644, 2645, 2646 (15); 6735, 6736, 68 1 1 (30); 7412,8563 (15); 8834 (33a). Summers 2209 (30). Svenson & Fassett 302 1 (32). Swynnerton Ml 7 a (28). Tamayo 2p I2 90 (15); 2014 (33a-b); 2015, 2023 (33a); 2085 (15); 9794 (33a); 46024 (15). Taylor, T. M. C. & Pillsburt 461 18 (15). Taylor, T. M. C, et al. 71, 1 44 (32). Tharp 1792 (9)1 43Si3> 43 SH (I 8 ); 44027 (3a). (6); 5030 (3a); 5031, 5032 (39); 5033 (3a); 5034, 5035, 5037 (18); 3039 (39) ; 5043 (40) ; J79 (5); 5080 (39, 40); 30^1/4 (40); 5080^/2, 5082 (39); 30 4484O0); 4844a, 5263, 5478, 5695 (15); 5779 (33a); 3972 (15); 5932 (30); 6040 (15); 6241 (30); 6333 (15); 6673 (33a); 6797, 7458, 8 1 28, 8747 (15); 9399 (30); 70724 (15); 703 614 O); 924 (18); 20893 (1); 21694 (6). Waterfall 4622 (39); 4811,5392 (I6b); 6358 (39); 6426a (3a); <*090, <*737 (39). Watson, P. 30OO (28). Watson, S. 4^9 (15); 7370 (35). 92 MISSOURI BOTANICAL [Vol. 42 Webster 1 3 (15). Weintraub 72a (32). Welsh 18089 (24). Welwitsch 48, 49 (28); 169 (24). Wetherill 563 (16b). Wheeler, L. C. 2981, 3036, 3122 (33a); 3158, 3419 (15). Wheeler, W. A. 672, 1058 (32). Wherry & Adams 2739 (32). White J9&? (1). Whitford 47 (33b). Wiegand & Manning 50 (3c). Wiggins ^2/2 (2, 11); 42 J? (11); 6^5 (40); 6764 (35); ?2p5, 7407 {1)18174, 8330, 8694 (40); 9937 (2); 9981 (41); 7/Jc?p (35). Wiggins & Rollins 450 (1); 4*4, 561 (35). Wight 74, 2r»T« of subgeneric categories are in capitals. Arenicolae 17 is CY LIN DROST ACHY AE "" [ u CYLINDROSTACHYS 16 Eremophilae 17 76 EUSELAGINELLA \ \ 6 HOMOEOPHYLLAE " 16 HOMOEOPHYLLUM .... _ 16 HOMOTROPAE 1 6 Ly co podium 40, 67 — 8 2 bryoides carinatum _ ciliatum Dregei laxum 77 57 82 rupestre 64 82 struthioloides 8 1 , Rupestres 17 58 Sartorii 17 3 1 Selaginella (Section) 16 16 Selaginella (Subgenus) ... Selaginella acanthonota amazonica Arechavaletae 26 34 ~ 34 arenaria 26 arenicola 23 __ 26 - 26 ssp. acanthonota ssp. arenicola ssp. Riddellii I 24 arizonica 7$ Arsenei 4 1 Aschenbornii 77 asprella 7 5 Balansae 27 Bigelovii 22 Bolanderi 49 Bourgeaui 64 bryoides 40 caffrorum ___. 53 capensis __ 53 carinata 50 Ch • • rismaru 20 var. Karwinskyana 20 var. Neeana . 20 cinerascens 40 columbiana .. 67 Coryi __. deflexa . 28 16 1955] TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 93 densa „_ 66 var. densa 68 var. scopulorum 67 var. Standleyi 71 denticulata 1 1 Dregei 57 var. Bachmanniana 57 var. Hildebrandtiana 57 var. Petersiana 57 var. pretoriensis . 57 var. Kehmanniana .... 57 var. Welwitschiana 57 echinata 54 Engelmannii 6 8 eremophila 8 extensa 47 Fendleri 62 florid ana 26 funiformis 2 6 gracillima 1 7 grisea 57 Hansenii 49 Haydenii 76 helvetica 11 Hintonii 37 humifusa 26 indica 5 2 Landii 81 lepidophylla 8 leucobryoides — 74 longipila _52, 68 macrathera 4 1 Mildei 34 montanensis 43 montevideensis 34 mutica 44 var. Hmitanea 46 var. mutica 46 var. texana __„ 46 mutica X rupincola - 21 X neomexicana 21 nivea 5 5 njamnjamensis ~~ 52 oregana 61 Parishii 80 pauciciliata 34 peruviana 77 var. Dombeyana 77 Phillipsiana 40 porrecta 37 Preissiana 1 7 proxima — . 56 pumila 17 py gmaea 1 7 Quartiniana . 53 Riddellii „ 24 Kosendahlii _ 43 rupestris — _ 64 f. amazonica ... 34 f. amurensis _. 71 f. Bolanderi .. ..... 49 f. Bourgeaui 64 f. brasiliensis 34 f. caffrorum «. 53 f. Dregei _ 57 f. Engelmannii ... ~ 70 f. Hansenii 49 f . Haydenii — 70 f. indica 52 f. longipila _. 68 i, manchuriensis — 71 f. montanensis 43 f. peruviana ... - 77 f. Schmidtii - 71 f. sibirica 71 f. Wallacei ...-_ 43 f. ssp Wrightii - 48 . Balansae 27 var. acanthonota var. var. var. 26 27 Balansae Bolanderi 49 borealis 8 1 _ 34 var. brasiliensis brevipita __ columbiana var. densa Fendleri ~. var. var. .... 81 ..... 43 var. ... 70 .. 62 var. Hansenii . 49 var. incurva - 53 f. abyssinica 53 f. angolensis 53 f. capensis - 53 var. longipila 81 subvar. glaucina - 81 subvar. viridis „ - - 81 ._ 47 var. mexkana 77 var. montanensis 43 mextcana var. peruviana var. recurva f. Dregeana — f. Welwitschiana .- 77 . .... 57 57 57 94 [Vol. 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN var. rupincola 20 var. Sartorii 37 var. shakotanensis 43 Sherwoodii 29 29 var. var. tortipila var. rupincola Sartorii . var. tropica 82 Wallace! 43 20 36 var. oregonensis 36 var. venezuelensis 36 Schmidtii 7 1 var. Krauseorum 71 scopulorum 67 selaginoides 1 6 Sellowii shakotanensis 34 43 Sbeldonii 77 Sherv/oodii 29 sibirica spinulosa 1 1 Standleyi 71 Steyermarkii 49 stru thioloides 8 2 tortipila 29 uliginosa 1 7 Underwoodii 62 var. dolichotricha 62 utahensis 74 61 Vardei viridissima 28 Wallacei _ 43 f. columbiana 43 Watsonii 7 2 mutica Weatherbiana .. 46 - 28 . 39 Wightii var. Phillipsiana 40 var. vetusta 39 var. Wightii 39 Wrightii 48 Stachygynandrum rupestre 64 71 TETRAGONOSTACHYAE 17 TETR AGONOSTACHYS 1 7 TETRASTICH AE 17 Tetrastichus 1 7 Urostachys carinatus 82 Explanation of Plate PLATE 1 Series Arenicolae Fig. 1. S. rupincola, buds at ground level, Tryon tf Tryon 5051 (MO), X 10. Fig. 2. Wooton Fig. 3. S. Bigelovii, branching rhizome and basal portion of aerial stems, Tryon & Tryon 5053 (MO), X 1.75. Fig. 4. S. arenicola ssp. arenicola, stems with basal buds and rhizophores, Tryon 5016 (MO), X 1.75. Fig. 5. S. arenicola ssp. acanthonota, basal portion of stems, Tryon 5009 (MO), X 1.75. Fig. 6. S. Balansae, portion of aerial stem, Balansa (P), X 10. Fig. 7. S. Weatberbiana, portion of rhizome, Tryon # Tryon 5077 (MO), X 10. Fig. 8. S. Weatherbiana, portion of aerial stem, Standley 4558 (MO), X 10. Fig. 9. S. viridissima y portion of aerial stem, Stewart 2204 (GH), X 10. Fig. 10. S. tortipila, portion of aerial stem, Rugel (US), X 10. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 1 TRYOyi-SELACINELLA RUPESTRIS AND ALIJES Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 2 TRYON — SELAGINELLA RUPESTR1S AND ALLIES [Vol. 42, 19551 TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 95 Explanation of Plate PLATE 2 Series Sartorii Fig. 11. S. Sellowii, portion of stem, upper side, Venturi 1866 (US), X 10. Fig. 12. S. Sartorii, portion of stem, upper side, Liebmann (GH), X 10. Fig. 13. S. Wightii var. Wightii, portion of stem, upper side, Ferguson (US), X 10 Fig. 14. S. cinerascensy portion of stem, upper side, Kimball (GH), X 10. Fig. 15. S. Arsenei, portion of stem, upper side, Arsene 1 064.1 (US), X 10. Fig. 16. S. macrathera, portion of stem, upper side, Johnston 8067 (GH), X 10. Fig. 17, S. shakotanensis, strobilus, Faurie 9895 (MO), X 10. Fig. 18. S. Wallace!, portion of stem, upper side, Wallace (NY ex B), X 10. Fig. 19. S. mutica var. mutica, portion of stem, upper side, Meehan (YU), X 10. Fig. 20. S. externa, portion of stem, upper side, Pringle 3QOO (GH), X 10. Fig- 21. S. Wrightii, portion of stem, under side, Try on & Try on 50 3 4 (MO), X 10. Fig. 22. S. Steyermarkii, upper leaves, Hatch tf Wilson 325 (US), X 10. Fig. 23. S. Steyermarkii, under leaves, the dark areas are colored red, Hatch & Wihon 325 (US), X 10. 96 MISSOURI [Vol. 42, 1955] Explanation of Plate PLATE 3 Series Sartorh Fig. 24. 5067 Fig. 25. S. carinata, stem with branches, under side, Rose et al. 23039 (US), X 1.25 Fig. 26. S. carinata, portion of stem, upper side, Rose et al. 23039 (US), X 1.25. $ Thomson (I Schweinfurth Fig. 27. Fig. 28. S. n)ar Fig. 29. S. caffrorum, portion of stem, under side, T. Cooper 662 (NY), X 10. Fig. 30. S. echinata, upper leaves from stem apex, Terrier 8313 (BM), X 14. Fig. 31. S. echinata, under leaves from stem apex, Perrier 8313 (BM), X 14. Fig. 32. S. nivea, portion of stem, upper side, Humbert & Swingle 5518 (US), X 10. Fig. 33. S. proxima, portion of stem, upper side, Humbert ef Swingle K70K (US), X 10. Fig. 34. S. proxima, sporophylls from upper side of strobilus, Humbert & Swingle 5705 (US), X 10. Fig. 35. S. proxima, sporophylls from under side of strobilus, Humbert ef Swingle 5705 (US), X 10. 4403 Ann. Mo. Bot. Gard., Vol. 42 % 1955 Plate 3 TRYON — SELAGINELLA RUPESTRIS AND ALLIES Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 4 TRYON — SEL AGIN ELLA RUPESTRIS AND ALLIES [Vol, 42, 1955 J TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 97 Explanation of Plate PLATE 4 Series Rupestres Fig. 37. S. Vardei, portion of stem, upper side, £. H. Wilson 54.II (P), X 10. Pig. 38. S. oregana, portion of stem, Kautz (YU), X 10. Pig- 39. S. Underwoodii> portion of stem, under side, Fendler 1024 (MO), X 10. Fig. 40. S. Underwoodii, portion of mat, Arsene 17963 (F), X 1. Fig. 41. S. rupestris, portion of stem, upper side, Leeds 76 (NY), X 10. Fig. 42. S. rupestris, basal portion of strobilus and leaves, /. K. Small on Aug. 18-27, 1894 (NY), X 10. Fig. 43. S. densa var. scopulorum y portion of strobilus, Standley 1 57 32 (US), X 10. Fig. 44. S. densa var. densa, portion of stem, upper side, Boivin & Gitlett 8686 (MO), X 10. Fig. 45. S. densa var. densa, portion of stem, under side, Boivin & Gttlett 8686 (MO), X 10. Fig. 46. S. densa var. Standleyi, portion of strobilus, Standley 17228 (US), X 10. 98 [Vol. 42, 1955] MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 5 Series Rupestres Fig. 47. S. densa var. densa, portion of mat; the dark area to the right is a region of dead vegetative branches, Van Schaack d Freytag 2934 (MO), X 1. Fig. 48. S. sibirica, portion of stem, upper side, Calder # Billard 4454 (MO), X 10. Fig. 49. S. Watsonii, portion of mat, Tryon & Tryon 5060 (MO), X 1. Fig. 50. S. Watsonii, portion of stem, upper side, Watson 1 370 (NY), X 10. Fig. 51. S. Watsonii, portion of strobilus, Watson 1 370 (NY), X 10. Fig. 52. S. utahensis, portion of stem, upper side, Flowers 3249 (MO), X 10. Fig. 53. S. leucobryoides, portion of stem, lateral view, Munz & Harwood 3789 (NY) , X 10. Fig. 54 S. asprella, portion of stem, upper side, Johnston 18 1 5 (US), X 10. Ann. Mo. Bot. Gaud., Vol. 42, 1955 Plate 5 TRYON — SELAG1 SELLA RUPESTR1S AND ALLIES Ann. Mo. Bor. Gard., Vol. 42, 1955 Plate 6 TRYON — SEL AGIN ELLA RUPESTRIS AND ALLIES I Vol. 42, 1955 TRYON SELAGINELLA RUPESTRIS AND ITS ALLIES 99 Explanation of Plate PLATE 6 Series Eremophilae Fig. 55. S. peruviana, tip of stem in dormant state, Tryon 6 Try on 5031 (MO), X 10. Fig. 56. S. peruviana, portion of stem, under side, Soukup 188 5 (US), X 10. Fig. 57. S. arizonica, portion of stem, upper side, W. S. Phillips 2588 (US), X 10. Fig. 58. S. arizonica, portion of stem, under side, Shreve (US), X 10. Fig. 59. S. eremophila, leaves from apical bud of stem, Johnston J047 (US), X 14. Fig. 60. S. Parishii, portion of stem, upper side, and strobilus, £. Palmer 306 (GH) , X 10. 306 Fig. 62. S. Landii, portion of stem in dormant state, upper side, Pringle 10823 (GH), X 10. Fig. 63. S. Landii, portion of stem, under side, Pringle 10823 A NEW PELLAEA FROM SOUTH AFRICA* ALICE F. TRYON Pellaea rufa sp. nov., A. F. Tryon. Rhizoma gracile repens dichotomum paleis concoloribus rufis elongato-lanceolato-triangularibus acumine filiformi basi cordata. Folia 12—30 cm, longa stipite rufo rachide rufa apicem versus flexuosa. Laminae elonga to- triangulares bi-tripinnatae segmentis ellipticis vel ovatis vel nonnullis ternatis coriaceis nervis immersis. Sporangia brevissime stipitata. Sporae pallide luteolo-fuscae triplantae laeves. Typus: Compton 16402 (US), Rhizome slender, cord-like, dichotomously branched, long creeping. Scales of the rhizome and stipe-base concolorous, ruddy-tan, elongate lanceolate- triangular, sparsely dentate, the tip filiform, the base cordate, scales surrounding the meristem pinkish or red. Fronds 12-30 cm. long, approximate, the buds paleaceous. Stipe and rachis convex or plane on the upper surface, nearly glabrous, ruddy-tan to red becoming darker and gray with age, the upper portion of the rachis usually flexuous. Blade 10 cm. long and 2 cm. broad to 25 cm. long and 5 cm. broad, bi-tripinnate, elongate-triangular, the pinnae ascending at a broad angle to the rachis, the rachises somewhat flexuous. Segments 3 mm. long and 2 mm. broad to 10 mm. long and 6 mm. broad, elliptical or oval (some ternate), retuse, coriaceous, the veins im- mersed and obscure, the margin reflexed or revolute, border lutescent, crenulate, the young segments reddish! Sporangia with short stalks less than one- fourth the capsule length. Spores 64 per sporangium, tetrahedral, pale yellowish- brown, essentially smooth. Specimens examined: Cape Province: Rock crevices, Ngaap Kop, Laingsburg Dist., Nov. 6, 1944, R. H. Compton 164.02 (US) ; Same locality, Dec. 1, 1941, R. H. Compton 12619 (US); Foothills of Witteberg, April 19, 1925, R. H. Compton 2961 (K) ; Rocky hillsides among succulents, The Great Karoo, near Matjiesfontein, Jan. 24, 1948, Robert /, Rodin 3342 (UC, K, MO) ; Karoo, Groote Fontein, Mr. Dickson, Lady Barkly reed. 5/75 (K) ; Locis rupestribus montosis, siccis in dumetris, Kendo, alt. 3000-4000, 1838, Drege (BM, K, MO, P). Additional localities from the Cape Province have been cited by Compton 1 from — Whitehill Ridge; Klipbank, Beaufort West Div.; and by Alston and Schlepe 2 from — Sutherland: Klein Roggeveld, Schietfontein. This fern of the South African Karroo has had an elusive record for some hun- dred years. It is not surprising for the species is strange among African ferns Thes Wfoh ^Trans. Roy. Soc. South Africa 19:270. 1931. Jour. South African Botany 18:170. 1952. *This paper was prepared in connection with a grant from the Bache Fund of the National Academy of Sciences for the study oi desert ferns. Issued March 24. 1955. (101) [Vol. 42 102 ANNALS OF THE MISSOURI BOTANICAL GARDEN that they were long considered as a single species. Another member of the genus, P. ovata, with a widespread Cordilleran range is more distantly related. Affinities of these species are easily seen in the unique, reddish color of the stipe and rachises and of the segments particularly on the immature frond. They are likewise ex- pressed in the elongate- triangular form and generally tripinnate division of the blade, in the elliptical, retuse form of the segments and elongate, basally cordate scales. P. and romedae folia of California and P. intermedia of southern Arizona, New Mexico, Texas and adjacent Mexico are distinguished from the Chilian and African plants in having the rhizome scales with a sclerotic central stipe. The rugose spores and compact, multicipital rhizome of P. myrtillifolia distinguish the Chilian plant from P. rufa. The first collection, made by Drege was included by Kunze 3 under Allosorus andromedaefolius along with collections from California and Chile. Hooker 4 retained these under Pellaea andromedaefolia. The Chilian species was segregated in Kuhn's publication on the Mettenius manuscript. 5 . Later authors regarded the African record as erroneous and it was excluded until reinstated by Marloth 6 , Compton and more recently by Alston and Schlepe as P. andromedaefolia. The range of the species is local largely within a radius of 40 miles in the yicinity of Whitehill. Several of the collections, including the type, were made by Professor Compton whose regard for this area has resulted in an extensive flora of the Whitehill District and a Botanical Reserve for the preservation of Karroo plants. The species occurs on dry, rocky hillsides at an elevation of 3,000-4,000 feet under 5-6 inches of rainfall. It is apparently able to make rapid growth during the January rains and to tolerate long periods of desiccation. Adaptations to this environment in other genera as Mesembryanthemum, Crassula, Helichrysum and Stapelia have resulted in the striking xeromorphic flora of the Karroo. Related species of Pellaea in Chile and the Southwestern United States exist under environmental conditions of essentially the same extremes. The distribution of such closely allied species in three distant regions is not a unique one. Such relationships particularly between South Africa and South America have been treated at considerable length in studies on the origin of the flora of South Africa as well as in systematic treatments of several plants and animals as the earthworms AcanthodriluSy the rosaceous genus Acaena y and Menodora of the Oleaceae. The range of the latter 7 is nearly identical to these species of Pellaea. Evidence from these closely related species of ferns supplies additional data on the floristic relation- ships of three continents. 8 Linnaea 10:503. 1836. 4 Species Filicum 2:149. 1858. 5 Linnaea 36:85. 1869. 6 Das Kapland in Wiss. Ergeb. Deutschen Tiefsee-Exped. auf "Valdivia" 2 3 :276. 1908. 7 Steyermark, in Ann. Mo. Bot. Gard. 19:100. 1932. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 7 D Telia exposed ea MISSOURI BOTANICAL GARDEN STAFF Emeritus Director George T. Moore Director Edgar Anderson Assistant Director T-Ttinw C. Cutler Carroll W. Dodge, Mycologist Robert E. Woodson, Jr.. Curator of the Herbariua Henry N. Andrevs, Paleobotanist Rolla M. Tryon, Assistant Curator of the Herbarium Hugh C. Cutler, Curator fuseum of Economic Plant- George B. Van Schaac^ Honorary Curator of Grasses Juuan A. Steyermark, Honorary Research Associate Frederick. G. Mi £R, Dendrologist Alice F. Tryon, Research Associate JOH D. DWYER, Research Associate Neix C Horner, librarian and Editor of Publications Gerald Ulrici Business Manager BOARD OF TRUSTEES President John S. Lehmann Vice-President Daniel K. Catlin Second Vice-President Eugene Pettus Leicester B. Faust Dudley French Henry Hitchcock Robert Brookings EX OFFICIO MEMBERS Richard J. Lock wood A. Vessel Shapleigh Ethan A. H. Shepuby Smith Arthur C Lichtenberger Bishop of the Diocese of r si Ethan A. H. Chancellor t; versity Stratford Morto Stolar Pre t the Aca ■ ■.,..- - .-' St. Louis P j tn Lot Raymond R. Tucker Gerald I ««- Secretary Volume XLII Number 2 Annals of the Missouri Garden MAY, 1955 103-130 Studies on Asiatic Relatives of Maize Nalinl Lichenological Notes on the Flora of the Antarctic Continent and the Subantarctic Islands. I-IV - . - • • • * • « 'j-L.*.' i3i_t40 Carroll W. Dodge and Emanuel D. Kuaolpn «i-iw ..- ' „ ,-, T?'--hert E F •".• 5 151-152 Three New Annonaceae from Panama - - • *od«t ^. he Botanical Catalogues of Auguste de St. HiIaire Johr D Dwye ; 15Z _ 170 Comparison of Juniper Populations on an 0«rk Glade^d^ ^ m _ m Old Fields F^ISHKD QUARTER AT O^SS^SSSSSL «««* BY THE BOARD OF TRUSTEES OF TH1 «*«£ 3T* LOUIS, MISSOUK** Entered as »econa-c i&c zh Annals of the Missouri Botanical Garden A Quarterly Journal containing Scientific Contributions from the Missouri Botanical Garden and the Henry Shaw School of Botany of Washington University in affiliation with the Missouri Botanical Garden- Information The Annals of the Missouri Botanical Garden appears four times during the calendar year: February, May, September, and November. Four numbers constitute a volume. Subscription Price $10-00 per volume Single Numbers 2.50 each Contents of previous issues of the Annals of the Missouri Botanical Garden are listed in the Agricultural Index, published by the H. W. Wilson Company. Annals of the Missouri Botanical Garden Vol. 42 MAY, 1955 No. 2 STUDIES ON ASIATIC RELATIVES OF MAIZE NALINI NIRODI 2 Introduction The five oriental genera of the Tripsaceae — Coix, Trilobachne, Polytoca, Scler- achne and Chionachne — are especially worthy of an investigation more detailed than has hitherto been accorded them, not only because of their close relationship to maize but in their own right. They form a series of primitive forage plants, grain crops, and weeds, concerning which very little, aside from taxonomic descrip- tions, has appeared in print. As thorough as the study of maize has been, it is not improbable that a better knowledge of its Old World relatives will go some way towards helping to solve the big problem it still presents, that of its origin. Coix is a cereal so ancient that it is mentioned in Vedic literature, was culti- Ary (W; Watt infers, judging by the amount of attention the plant received in early botanical writings such as those of Pliny, Rumphius, Loureiro, and Gerarde, that formerly it was more extensively cultivated. Wester (1920), however, believes that though the hard-shelled forms of Coix were known in Europe as far back as the time of Pliny, the soft-shelled kinds do not appear to have been known to European writers until the seventeenth century. The Arabs are responsible for its introduction into Spain and thereby to the West in general, and also for the picturesque name by which it is commonly known in Europe, Job's Tears (Watt, 1904). poo understood taxa, largely because of the occurrence of numerous intergrading forms. Therefore, those who have worked on the genus (Watt, Vallaeys, and Mimeur) do not agree too closely about which taxa should be allotted specific rank. In Table I are listed in chronological order as many basonyms in the genus Coix, as *An investigation carried out in the Henry Shaw School of Botany of Washington University and submitted as a thesis in parti.il fulfillment of the requirements for the degree of Doctor of Philosophy. Corn Co (103) 104 MISSOURI [Vol. 42 TABLE I NOMENCLATURE IN THE GENUS COIX 1. 2. 3. 4. 4a. 4b. 5. 6. 7. 7a. 8. 8a. 8b. 8c. 9. 10. 10a. 10b. C. Lacryma Jobi L. Sp. PL 972. 1753. C. Lacryma Jobi Thunb. FL Jap. 37. 1787. C. arundinacea Lam. Encyc. Method. 3:422. 1789. C. agrestis Lour. FL Cochin. 2:551. 1790. C. Lacryma-Jobi 1. ^r«/w (Lour.) Backer, Handb. FL Java 2:33. 1928. C. Lacryma-Jobi forma agrestis (Lour.) [incorrectly attributed to Backer by] Vallaeys in Bull. Agr. Congo Beige 39:255. 1948. C. pendula Salisb. Prod. 28. 1796. C. exaltata Jacq. Eclog. Gram. 60. 1813-1820. C. gigantea Koen. ex Roxb. Hort. Beng. 66. 1814. C. Lachryma-Jobi var. gigantea (Roxb.) Stapf ex Hook. FL Brit. Ind. 7:100. 1896. C. aquatica Roxb. in FL Ind. 3:571. 1832. C. gigantea Koen. ex Roxb. var. aquatica (Roxb.) Watt, Agr. Ledger 11:198. 1904. C. Lacryma-Jobi 4. aquatica (Roxb.) Backer, Handbk. Fl. Java 2:33. 1928. C. Lacryma-Jobi forma aquatica (Roxb.) [incorrectly attributed to Backer by] Vallaeys, I.e. 1948. C. chinensis Tod. in Index Sem. Hort. Bot. Pan. [Palermo], 5. 1861. C. Ma-yuen Roman, in Bull. Soc. Nat. Acclimat. II, 7:442-444; 1881: C. Lacryma-Jobi var. "Ma-yuen" (Roman.) Stapf ex Hook. FL Brit. Ind. 7:100. 1896. 11. 11a. lib. 12. 13. 14. 15. 15a. 16. 17. 18. 18a. 18b. 19. 20. 21. C. Lacryma-Jobi forma "Ma-yuen" (Roman.) [incorrectly attributed to Stapf by] Vallaeys, I.e. 254. C. Lachryma var. stenocarpa Oliver in Hook. Ic. PL III, 8: pi. 1764. 1888. C. stenocarpa (Oliver) Balansa in Jour, de Bot. [Paris] 4:77. 1890. C. Lacryma-Jobi forma stenocarpa (Oliver) [incorrectly attributed to Stapf by] Vallaeys, I.e. 255. C. puellarum Balansa, 1. c. C* tubnlosa Hack, ex Warb. in Bot. Jahrb. 13:260. 1891. C. lingulata Hack, in Oest. Bot. Zeitschr. 41:5. 1891. C. Lacryma-Jobi var. monilifer Watt, I.e. 214. C. Lacryma-Jobi forma monilifer (Watt) [incorrectly attributed to Watt by] Vallaeys, 1. c. 254. C. Lacryma-Jobi var. typica Watt, 1. c. 206. ouwehandii Koord. in Bull. Jard. Bot. Buitenz. Ill, 1:191. 1918. C. palustris Koord. 1. c. C. Lacryma-Jobi 3. palustris (Koord.) Backer, I.e. Lacryma-Jobi forma palustris (Koord.) [incorrectly attributed to Backer by] Vallaeys, I.e. 255. C. poilanei Mimeur in Rev. Internat. Bot. Appl. 31:199. 1951. C. lacryma-jobi var. major Mimeur, L c. 202. C. lacryma-jobi var. minor Mimeur, 1. c. 204. C. C. S 1 1 1 7? 1 1 7 7 8a 8a • •.* 10a 10a 10a 11 11 15 11 7 15 16 • • * 4b 4b 8c 8c 8c 8c 10b 10b 10b 10b lib lib • • * » 15a 15a u 3 E 1 7 8a 8a 10a 11 15 *-•• 16 .... 17 18b MM 18b .... 18b .... 19 .... 20 21 *Here spelled Coyx. delimited here, as were found. These bear the unmodified numbers 1 to 21. The list includes also a selection from the many combinations which have been made in this genus. These follow their respective basonyms and bear the same number modified by a letter. A number occurring in a column under one of the names, 1955] NIRODI ASIATIC RELATIVES OF MAIZE 105 Watt, Vallaeys, or Mimeur, indicates that the corresponding author equates the opposite name to that bearing the number given. The best-known and most widely spread species is Coix Lacryma Jobi L. Originally native to southern and eastern Asia, it has today virtually run wild throughout the tropical and warm temperate regions of the world. The other species have a comparatively narrow distribution. C. gigantea Koen. ex Roxb. is found chiefly in the eastern and central parts of India, in Ceylon, Burma, and in Japan. Its variety aqua t tea has been reported in Indochina, western and southern India, Burma, and Japan. C. ouwehandii Koord. is endemic to Sumatra. Another endemic species is the one set up by Mimeur, C poilanei, found in Laos. Vallaeys (1948) suggests the Malay Archipelago as the center of origin of CoiXy taking into consideration the multiplicity of varieties present there, some of which are not found elsewhere in the wild state. Accordng to Burkill (1935), "it seems early to have become a cereal of some importance in the hills of Indochina and southern China, and it is probable that the beginnings of the cultivated races should be sought for in that part of the world." As an article of diet, Coix used to be known as being most important during times of famine, helping to tide over seasons when the rice crop failed. Its nutritive importance, however, has impressed many investigators, and of recent years it has been gaining in popularity as a cereal and has come to be more generally cultivated. Wester was responsible for popularizing it in the Philippines where it is known as Adlay. Since then similar campaigns have been set in motion in Java, Ceylon, the Belgian Congo, Guatemala, Nicaragua, Costa-Rica, Bolivia, Peru, Chile and Brazil. Analyzed along with certain other famine foods, its energy value was found to be the highest of all those studied (Paton and Dunlop, quoted by Watt, 1904). Wester (1920) records the results of analysis of the Bukidnon form of ad lay as closely approximating wheat in starch and protein content and exceeding it in fat. Comparing it with two other great cereals of the world, he says, "with a greater protein and fat content than either rice and corn it is a more complete human food than either of these cereals." An idea of its productiveness is obtained from Pieris' statement (1936) that a crop of adlay from an acre of land will provide sufficient food for a family of four people for about one year. Duthie (1888), speaking of its value as a fodder, says it is largely eaten by cattle and reported to be very fattening. Vallaeys (1948) writes of its curious use as "the object of practices and beliefs held in magic" and also of the preparation of a refreshing drink from the grain which acts as a diuretic and a depuratory. In hospitals in China it is said to form a good and cheap substitute for barley in the making of gruel. It is known to be used in the elimination of stones from the bladder (Burkill, 1935). A decoction of the roots is given to children as a vermifuge (Ridley, 1906). It is best known, however, for its use for decorative purposes, in the making of bead-baskets, rosaries, and various articles of personal adornment. Seeds of the wild forms with the hard, lustrous shells are the ones thus employed, usually those [Vol. 42 106 ANNALS OF THE MISSOURI BOTANICAL GARDEN belonging to var. stenocarpa, which has elongated, cylindrical false fruits, and those of var. typica, with false fruits of the characteristic tear-drop shape. This list of the uses of Coix could be appropriately closed with a quotation from Vallaeys. Wherever man has lived it is possible to discover vestiges of ancient cultures and of plants whose use is lost. Speaking of the regrettable tendency to limit the number of products useful as food, he writes, "Cette limita- tion, cette 'standardisation' ne peuvent etre considerees comme un progres. Voila pourquoi dan certain pays, on regard la coix comme une 'solution* a des problemes d'importance capitale d'ordre dietetique, economique et social". Both Polytoca and Chionachne are generic names which owe their origin to Robert Brown (Bennett, Horsfield & Brown, 183 8). Many species have been described for both genera as a result of which the literature has become cluttered with synonyms. In "A contribution to the knowledge of the Indian Maydeae" (1931), Henrard made an attempt to clarify their taxonomy. The following table is based on Henrard's and Pilger's treatments. It is constructed in the same way as Table I. l. la. 2. 3. 4. 4a. 5. 5a. 6. 7. 7a. 8. 1. 2. 3. 3a, 4. 5. 5a 6. 7. TABLE II NOMENCLATURE IN THE GENUS POLYTOCA Apluda digit ata L. f. Suppl. 434. 1781. Polytoca digitata (L. f.) Henr. in Meded. Rijks. Herb. Leid: 67:10. 1931 Coix heteroclita Roxb. Fl. Ind. 3:572. 1832. Polytoca bracteata R. Br. in Bennett & Brown, PI. Javan. Rar. 20. tab. 5. Cyathorhachh Wallichiana Nees ex Steud. Syn. Pi. Gram. 1:403. 1854. Polytoca Wallichiana (Nees) Benth. in Jour. Linn. Soc. 19:52. 1881. Sclerachne cyathopoda Muell. Fragm. Phytog. Austr. 8:116. 1873. Polytoca cyathopoda (Muell.) F. M. Bailey, Queensl. Fl. 6:1849. 1902. Polytoca macrophylla Benth., 1. c. 1881. Chionachne Massii Balansa in Jour, de Bot. [Paris] 4:78. 1890. Polytoca Massii (Balansa) Schenck ex Henr. I.e. 67:9. 1931. Polytoca javanica Henr. in Blumea 3:241. 1939. 1838 NOMENCLATURE IN THE GENUS CHIONACHNE Coix arundinacea Koenig ex Willd. Sp. Pi. 4:203. 1805. Coix barbata Roxb. Fl. Ind. 3:569. 1832. Coix Koenigii Spreng. Syst. 1:239. 1825. Chionachne Koenigii (Spreng.) Thwaites, Enum. Pi. Zeyl. 3 57. 1864. Coix crypsoides C. Muell. in Bot. Zeit. 19:334. 1861. Polytoca semiteres Benth. ex Hook. Fl. Brit. Ind. 7:101. 1896. Chionachne semiteres (Benth. ex Hook.) Henr. in Meded. Rijks. Leid. 67:16. 1931. Chionachne biaurata Hack, in Philipp. Jour. Sci. Suppl. 1:263. Chionachne Hubbardiana Henr. in Blumea 3:162. 193 8. Herb 1906. la la la la 4a 4a 5a 5a 6 7a 7a 8 s PC 3a 3a 3a 3a 3a 5a 5a 6 7 la la la 4a 4a 5a 6 7a 8 | s 3a 3a 3a 5a 6 7 1955] NIRODI ASIATIC RELATIVES OF MAIZE 107 The species of both Polytoca and Chionachne range in distribution from India to Australia. Wall J cyathopoda is confined to northern Australia and Queensland; P. macro phylla grows in the Malay Archipelago, New Guinea, and the Louisiade Archipelago; P. javanica has been reported only from Java by Henrard. Chionachne Koenig Q land. Q The only economic use reported for these grasses is as fodder. Chionachne Koenigii, which is widely spread over India, is used as a fodder when young (Duthie, 1888). Materials All the plants used in this study were grown from seed at the Missouri Botanical Garden, both outdoors and in the experimental greenhouse. Through the Plant Introduction Station, Beltsville, Maryland, seeds of various varieties of Coix were obtained from India, Ceylon, the Philippines, Central America, and the Belgian Congo, and large numbers of plants were raised. Since Coix forms a polyploid series with haploid chromosome numbers of 5, 10 and 20, it was decided to choose one representative of each of these numbers for cytological study, as follows: (1) A species with a haploid chromosome number of 5, bearing small, hard, chalk- white false fruits, kindly procured for our purpose from Cuttack, India, by Dr. N. Krishnaswami, Cytogeneticist, Agricultural College, Coimbatore, India. (2) "Blue Adlay", a variety of Coix Lacryma Jobi L., with a haploid chromo- some number of 10, obtained from Trinidad through the courtesy of Dr. William L. Brown, Geneticist, Pioneer Hi-Bred Corn Company, Iowa. (3) A species with a haploid chromosome number of 20, bearing greenish- brown false fruits, collected on the Western Ghats (the mountain range that ex- tends for 800 miles along the southwest and west coast of India), also sent by Dr. N. Krishnaswami. Specific or varietal names have not been appended to (1) and (3), because their morphological characters do not conform exactly to any of the descriptions of Coix species met with in the literature. One species of Chionachne — Ch. Koenigii (Sprengel) Thwaites (n =. 10), sent by Dr. N. Krishnaswami from the Millet Breeding Station, Coimbatore, India, is included in this study as is also one species of Polytoca — P. macrophylla Benth. (n = 20) Will Full descriptions of the five taxa are given below, and inclusive herbarium specimens are being prepared for distribution. No members of the genera Sclerachne and Trilobachm were available for cyto- logical or morphological study. [Vol. 42 108 ANNALS OF THE MISSOURI BOTANICAL GARDEN The following is a key to distinguish the three genera: I. Female spikelets enclosed in stony covering formed by modified indurated leaf sheath Coix II. Female spikelets enclosed in hardened covering chiefly formed by indurated glume I. a. Terminal spikes male, lateral spikes mixed or female Volytoca b. All spikes with flowers of both sexes Chionachne ( 1 ) Coix from Cuttack, India. — Root-system very much branched with extremely fine roots. Culm erect with numerous basal leaves. Leaf -blades linear, very narrow, up to 5 mm. broad, the longest leaves attaining a length of about 60 cm.; prominent white midrib; at union with sheath a well-marked suture which is narrow in the middle and fans out on either side; upper leaf surface sparsely set with hairs each arising from a colorless, crateriform gland. Sheaths ribbed, dotted on the outer face with gland- ular hairs similar to those of the blade. Inflorescences axillary, on triangular peduncles, as many as 8 arising from one node; false fruit small compared to most other varieties of Coix y 8 mm. long and 5 mm. in diameter, flattened on one side, often constricted towards the top, turbinate, chalk-white, hard. Male portion of the inflorescence up to 3.5 cm. long. Glume I 1 cm. long, lanceolate-ovate, slightly falcate, emarginate, the margins reflexed, the glume thereby 2-keeled, en- closing glume II, the keels winged, one of them throughout, the other only in its upper half, the wings serrate. Glume II about the length of glume I, acute, entire, margins infolded. This description comes fairly close to that given by Mimeur (1951) for her species, Coix poilanei y endemic to Laos. The finely branched root-system, the num- erous basal, narrow, linear leaves, and especially the dimensions and color of the false fruits, coincide with her description of C. poilaneL She says, however, that "les feuilles sont en continuite directe avec les gaines dont elles ne se distinguent pas", whereas in our plant the blades are separated from the sheaths by a distinct suture. The glands on the sheath are colorless and not brown as stated by Mimeur. Also, for the male spikelets she reports "une seule fleur par epillet," while the spike- lets of our plant exhibit the 2 -flowered condition. In referring to her species, she remarks that it is "tres proche de Coix aquatic a 1 (C. gigantea var. aquatica Watt) . :ussion is n = 5 , which is the same as that determined for var. aquatica by Mangelsdorf and Reeves (1939). (2) "Blue Adlay", Coix Lacryma Jobi L., from Trinidad. This is the wild "Coix Lacryma-Jobi proper" of Watt to which he gave the varietal name typica and of which he speaks as "the normal form" referring to the false fruits. Stem erect. Leaf -blades lanceolate, broad, up to 3.5 cm. in width, in length variable, up to 56 cm., auriculate, glabrous, "except for the double row of ascending teeth along each of the veinlets of the upper surface — a peculiarity that gives the texture of the leaf the appearance of being embroidered and makes it backwardly hispid" (Watt, 1904). Writing further about this character, Watt says: "all the 1955] NIRODI ASIATIC RELATIVES OF MAIZE 109 forms of this species are at once recognized by the texture of the leaf and its peculiar scabrosities." Ligule membranous; union of blade with sheath marked by a broad suture. Sheaths quite smooth. fruits 11 mm. long, porti 7 mm. in diameter, hard, smooth, polished, bluish- white. Male florescence up to 3.5 cm. long, spikelets in two's or three's. Glume I 7 mm. long, lanceolate-ovate, emarginate, strongly keeled, winged in the upper half, wings serrate. Glume 11 about the length of glume I, acute, margins infolded, entire. Text-fig. 1. Coix poilanci Mimeur; fig. 2 All about y$ natural size. (3) Coix from the Western Ghats, Stem perennial, erect. Leaf "Blue Adlay"; fig. 3. Coix gigantea Koen; ex Roxb. broad and up to 30 both each arising from a crateriform gland; margins serrate. Sheath also w,th crater,- [Vol. 42 110 ANNALS OF THE MISSOURI BOTANICAL GARDEN form glands bearing hairs, but here the glands of a dark reddish hue; union of blade with sheath marked by a broad suture. Ligule membranous; just above ligule, at base of leaf, a tuft of stiff hairs similar to those on the rest of the leaf. Inflorescences axillary, 2 or more spikes emerging from the sheath of a short leaf, the spike emerging first, being borne on a longer peduncle than that of the others; peduncles flattened on one side, glabrous; false fruits elongate, pyriform, 1 cm. long, 5-6 mm. in diameter, mouth oblique, pale green with brown markings, slightly flattened on one side. Male spikes up to 4 cm. long, drooping, spikelets mostly in 3's, the lateral ones sessile and the middle one pedicellate, each 2-flowered, Glume I 9 mm. long, ovate-oblong, emarginate, the margins reflexed, the glume thereby 2-keeled, enclosing glume II, the keels winged throughout, margins of the wings serrulate. Glume II winged, acute, margins entire. The description above corresponds in general to that given for C. gigantea by Roxburgh (1832) and that by Watt (1904). The important differences are discussed below: 1. Roxburgh describes the sheaths as "smooth", and Watt as "quite glabrous and eglandular". In the specimens under discussion, however, the sheaths are sprinkled with sharp hairs arising from crateriform glands of a dark reddish color. 2. The leaves are not "narrow" and "linear" as in Watt's description but "about I inch broad" as in Roxburgh's. They are evidently subject to variation, for Watt speaks of ". . . . Sir J. E. Smith's sheet of C. gigantea . . . . which has gland- dotted broad leaves . . 3. With regard to the false fruits, Watt describes them as "flattened on one side, and with 2 furrows longitudinally on the flat face . . . ." While the ones described here are flattened on one side, no furrows are present, nor are any furrows mentioned by Roxburgh or by Mimeur (1951). The name Coix gigantea is retained for the entity used in this study. The un- certainty of its correct application is indicated by a question mark. The chromo- some number of "2n = 40" tallies with that reported for C. gigantea by E. K. Janaki Ammal (Darlington & Janaki Ammal, 1945). (4) Polytoca macro phylla Bentham. Stem robust, erect. Leaf -blades up to 55 cm. long, between 2.5 and 3.5 cm. broad, glabrous, lanceolate, acuminate, cordate at base; margins dentate; midrib prominent; union with sheath marked by a broad suture. Sheath glabrous. Ligule narrow, membranous. Female spikelets 1 cm. long, variable in glume I which may be entire or toothed at summit or have 2 lateral teeth, sometimes 1 . Male spikelets in pairs, each pair made up of 1 pedicellate and 1 sessile spikelet. Glume I up to I I mm. long, acute, asymmetrical in the pedicellate spikelet, margins infolded with a broad wing along the fold on one side, wing serrate, other side merely folded slightly, lanceolate-ovate in the sessile spikelet, margins infolded, winged along the fold symmetrically on both sides. Glume II laterally compressed, folded along the 99 J955] NIRODI ASIATIC RELATIVES OF MAIZE 111 Text-fig. 4. Polytoca macrophylla Bench- About % natural size. middle, keeled along the fold in the pedicellate spikelet; keel winged in upper half, very The chromosome number determinat 40. (5) Chionachne Koenigii (Sprengel) Thwaites. Stem perennial, erect, freely branching at the base. Leaf-blades about 40 cm. long (said to attain a length of 2 feet, Ranga Achariyar, 1921), up to 1.2 cm. broad, linear, with a prominent midrib; union with sheath well-marked by a suture. Sheaths with scattered tubercle-based hairs at the base, sheath margins hairy. Ligule membranous; nodes densely ringed with hairs. Inflorescences axillary, most of them clustered at the ends of very long peduncles which are flattened or slightly concave on one side; leaf, in the axil of which several peduncles arise, with a sheath about 2.5 cm. long, the blade, although shorter than normal, fairly well-developed; each spike in a cluster enclosed in its lower portion by a bladeless awned sheath, 1-1.5 cm. long; 1, sometimes 2, of the lowermost spikelets female, borne in pairs of which one spikelet is sessile and the other pedicellate, the pedicel partly fused with the rhachis; rhachis pubescent. Female spikelets oblong, 5.5 mm. long, en- [V«l. 42 112 ANNALS OF THE MISSOURI BOTANICAL GARDEN Fig. 5. Cbionachne Koenigii (Spreng.) Thwaites. About % natural size. closed in glume I which becomes shiny and hard at maturity; at the union of the edges of the glume a groove through which the rhachis passes, Male s pikelets 2- flowered. Glume J up to 8 mm. long, acute, pubescent, the margins infolded, the glume winged along the fold on both sides in the sessile, but only on one side in the pedicellate, spikelet; wings serrate. Glume II acute, laterally compressed and folded along the middle with a pronounced keel along the upper half of the fold in the pedicellate spikelet, only slightly keeled in the sessile. 1955] NIRODI ASIATIC RELATIVES OF MAIZE 113 Cytology A number of cytologists have definitely established the existence of the chromo- World Maydeae, Euchlaena and Zea. Tripsacum is a well-known exception; all the and varieties have 18 pairs of chromosomes or multiples of this number. Among the Oriental Maydeae the gametic or reduced number most often encountered is also 10; this is evident from the following table: 1. Coix Lacryma J obi L. 2. Polytoca macrophylla Benth. 3. C. Lachryma stenocarpa Oliver 4. Sclerachne punctata Brown 5. C. Lacryma- J obi var. "Ma-yuen" (Roman.) Stapf 6. Polytoca barbara Stapf [Chionachne Koenigii (Spreng.) Th wakes] 7. C. aquatica Roxb. 8. C. gigantea Koen. ex Roxb. 9. C. poilanei Mimeur (?) Kuwada, 1915 Avdulov, 1931 Reeves and Mangelsdorf, 1935 Reeves and Mangelsdorf, 1935 Mangelsdorf and Reeves, 1939 Mangelsdorf and Reeves, 1939 Mangelsdorf and Reeves, 1939 E. K. Janaki Ammal, 1945 Nirodi, 1955 The numbers in all the above taxa except 3, 4, and 7 have been confirmed in this study. C. aquatica (Roxb.) and C. poilanei Mimeur (?) are exceptional in having the gametic number 5, which is the lowest found in the Gramineae and reported so far only for the unrelated genera Briza and Anthoxanthum and for five species from two of the sub-genera of the Sorghastreae. METHODS For cytological study entire male spikes were fixed in a solution of glacial acetic acid and absolute alcohol 1:3. Two days to a week later they were transferred to 70 per cent alcohol and stored under refrigeration. Young spikes that had barely begun to emerge from the ensheathing spathe were found to be at the right stage for meiotic studies. Temporary smears were stained either with aceto-carmine or with aceto-orcein. A few were made permanent by using the dry-ice freezing technique described by Conger and Fairchild (1953). While pachytene smears of Coix were being examined to investigate the possi- bilities for making chromomere analyses, stray cells from somatic anther-tissue were noticed showing well-stained and well-spread diploid chromosomes. These were assumed to be tapetal cells. Proof of the assumption was obtained when young anthers fixed in a solution of 3 parts absolute alcohol to 1 of glacial acetic acid were embedded in paraffin, sectioned and stained with crystal violet. [Vol. 42 114 ANNALS OF THE MISSOURI BOTANICAL GARDEN Smears made for the observation of tapetal chromosomes were treated in a slightly different way from smears made for meiotic studies. The contents of an anther were squeezed out in a drop of stain and well stirred with an iron needle. The debris was allowed to remain. After the placing of the cover-slip the slide was heated over a spirit-lamp and pressed vigorously between blotters to release the tapetal chromosomes from the nuclear membrane and spread them evenly. CHROMOSOME MORPHOLOGY The tapetal cells of Coix, Polytoca and Chionachne go through the process of endomitosis. At the time the pollen mother cells are undergoing pachytene the tapetal cells have already passed from the uninucleate to the binucleate condition, and smears made at this stage show many of them in endoprophase. Speaking of endoprophase stages in tomato tapetum, Brown (1949) remarks that they are "valuable for a study of somatic chromosome morphology". The same can be said of the three genera of grasses dealt with in this paper. Observations on chromo- some morphology were based almost entirely on tapetal cells from anther smears. At about the middle of endoprophase, the chromosomes reach a particularly favor- able degree of contraction and stain deeply enough to be identified individually. Sectioned material shows them surrounded by a nuclear membrane but in smearing the membrane may be lost. Coix poilanei Mimeur (?). — This species, together with Coix gigantea var. aquatica Watt, occupies the lowest position in the polyploid series formed by Coix. Its somatic chromosome number determined both from root-tips and tapetal cells was found to be 10. At very early endoprophase the chromosomes are very much attenuated and are up to five times as long as they are at the end of this phase. They are very loosely coiled, and though they often overlap they can easily be traced throughout their length (pi. 8, fig. 1). A characteristic of tapetal chromosomes in Coix is their striking differentation into regions of different staining capacity. Extremely lightly staining areas occur which are very constant in their location at a particular stage. These stand out in good contrast to the deeply staining portions, thus making the chromosomes easy to identify. Usually, and especially when they occur at the ends of the chromo- somes, these achromatic regions are visibly split, with the daughter chromatids often divergent. A point to be noted at this stage is that quite often along the achromatic regions the chromomeres are visible, and two of these may be seen side by side in each diverging half of a chromosome indicating that it is made up of four chromatids (pi. 8, fig. 3A). "Achromatic" as used here is a relative term. These regions appear pale in comparison with the remaining portions of the chromo- some. The chromomeres in the achromatic regions often stain clearly enough to be distinctly visible. The five chromosomes of the haploid set have been designated A, B, C, S and D, S standing for the satellite chromosome (pi. 8, figs. 2-6). Chromosome A has a 1955] NIRODI ASIATIC RELATIVES OF MAIZE 115 very long achromatic region at one end, and Chromosome B has shorter achromatic regions at both ends. Chromosome C has two achromatic regions, one in each arm midway between the end and the centromere. Chromosome S is recognized by the presence of the satellite. In its elongated state it also exhibits an achromatic region in either arm but this may disappear on contraction. Chromosome D is very like Chromosome S but devoid of the satellite and shorter. Figures 2—6 of pL 8 show tapetal chromosomes in varying stages of contrac- tion. Even when quite strongly contracted they can be told apart — Chromosome A and B by the difference in number and location of the achromatic regions, Chromosome S by the satellite, Chromosomes C and D by their difference in length, Chromosome D being shorter. When the achromatic regions do not occur at the ends of the chromosomes their contraction leaves wide gaps in the body of the chromosomes. Figure 6 of pi. 8 shows a late endometaphase with most of the chromosomes divided throughout but held together at the centromeres. The diverging arms give them a characteristic cruciform appearance. Owing to the fact that there may be two or even three places along the chromo- somes that show clear spaces and look like centromeres, the true position of the spindle-attachment region was ascertained from this stage and from root-tip squashes. Root-tip chromosomes examined at anaphase were found to be bent into V*s with equal or slightly unequal arms indicating that all the chromosomes had median or sub-median centromeres. Plate 10, fig. 2, shows one large tapetal nucleus containing the tetraploid number of chromosomes after the completion of one endomitosis. Since the chromosomes are in early endoprophase it is probably entering a second cycle. It was not ascertained whether a second endomitosis is completed nor what degree of ploidy a tapetal cell is capable of attaining. When the chromosome number for Coix aquatica Roxb. was reported by Mangelsdorf and Reeves (1939) to be 2n == 10 they observed that "no variation in the number was found". In C. poilanei Mimeur (?), though the tapetal cells most often showed 10 chromosomes, occasionally cells were found where 12 and rarely 1 1 were present. Though the nuclear membrane is ruptured on smearing, it is possible to recognize the contents as belonging to a single nucleus. Thus the possibility that the extra chromosomes might have been displaced from a neighbor- ing nucleus due to the pressure exerted while smearing is eliminated. Besides, when pollen mother cells were examined, occasionally 6 bivalents were found at diakinesis; fig. 1 of pi. 1 1 shows an MI telophase where 6 chromosomes have been distributed to either pole. Due to the way in which the anthers were collected it was not pos- sible to say with certainty whether this variation in number was between plants or between spikelets of one plant. Coix Lacryma Johi L.— The tapetal cells of three kinds of this species were smeared to see if varietal differences were manifest in the chromosomes. The first was "Blue Adlay", i.e. var. typica from Trinidad (described above). The other two were two kinds of var. "Ma-yuen". Cultivated forms of C. Lacryma jobi have been placed under this varietal name. Cultivation makes itself evident by [Vol. 42 116 ANNALS OF THE MISSOURI BOTANICAL GARDEN turning the stony capsule spathe soft and brittle and striated. One of the two kinds mentioned came from the Philippines and has chalk-white, brittle, and very faintly striated capsular spathes. The shape is that of var. typica. The other has dark brown, brittle and strongly striated capsular spathes, somewhat elongated and with a mouth larger than usual. For convenience these shall be referred to as var. "Ma-yuen (1)" and var. "Ma-yuen (2)". A general similarity is obvious in the chromosome sets of these three kinds of Coix. On the whole, the chromosomes are smaller than in the 10-chromosome species. Another difference is that achromatic regions, when they occur, are usually found only towards the ends of the chromosomes. Therefore, the pronounced gaps that are left in the body of the chromosomes after a certain degree of contraction described in the 10-chromosome species are rarely met with here. The attempt to discern homologues among the somatic chromosome sets of var. typica, var. "Ma-yuen (1)" and var. "Ma-yuen (2)" was confined to those chromosomes provided with pronounced achromatic ends as they helped as markers. They have been marked by small letters of the alphabet (pi. 8, figs. 7 and 8; pi. 9, fig. 1 ) . The nucleolus chromosomes are very similar in all three, with distal ends achromatic and split. In var. typica and var. "Ma-yuen (2)" both nucleolus chromosomes in a diploid set possess satellites but in var. "Ma-yuen ( 1 ) " a satellite was seen on only one of them. Its homologue is seen to be in connection with the nucleolus but no satellite was observed. As in the 10-chromosome Coix, the centro- meres are all median or sub-median which agrees with Taylor's observations (1925) on root-tip chromosomes. This is fairly clear in the tapetal chromosomes (pi. 8, figs. 7 and 8; pi. 9, fig. 1) but much more evident in pi. 11, fig. 7, where meiotic chromosomes of var. typica are seen in second metaphase. The chromatids prior to separation are held together at the centromere which stands out markedly. Coix gigantea Koen. ex Roxb. (?). — Here the chromosomes, in spite of being twice as numerous as in the varieties of C. Lacryma ]obi y are no smaller. In fact, two pairs approach A and B in size, the two largest chromosome pairs in the 10- chromosome Coix. Coix poilanei? (pi. 8, fig. 3), the varieties of C Lacryma J obi (pi. 8, figs. 7 and 8), and C. gigantea (?) (pi. 8, fig. 2) all show their respective tapetal chromosomes in endoprophase at more or less the same degree of contrac- tion. By comparing these figures it can be seen that many of the chromosomes of C. gigantea (?) are intermediate in size between those of C. poilanei (?) and of the 20-chromosome Coixes. The centromeres of C. gigantea chromosomes stand out more clearly than those of any of the other kinds described so far. Again, all of them are either median or sub-median. This is confirmed by fig. 1 of pi. 13, where the chromosomes are in prophase of the second meiotic division. The chromatids are held together only at the centromeres and the four diverging arms of each chromosome are seen to be approximately equal. Due to the much larger number of chromosomes present in C. gigantea? it was difficult to obtain cells which showed them spread at an appreciable distance from The achromatic ends 19SS] NIRODI ASIATIC RELATIVES OF MAIZE 117 one another. The cell shown in fig. 2 of pi. 9 was stained in aceto-orcein and does not show the nucleolus. Diakinesis stages clearly indicate the presence of two pairs of nucleolus chromosomes. From observations of carmine-stained tapetal cells it was seen that two of the nucleolus chromosomes were similar to the ones indicated by arrows (pi. 9, fig. 2). Although only one pair of chromosomes has been indicated (by X's) in var. "Ma-yuen (2)" and in C. gigantea? as being prob- ably homologous, a general similarity between the genomes of the 20- and the 40- chromosome Coixes is very much in evidence. The characteristic achromatic regions are found in practically all the chromosomes of C. gigantea? and, as in the varieties of C. Lacryma Jobi, are mostly confined to the ends. are visibly split and often divergent. Occasionally, tapetal cells were noticed which had two extra chromosomes. It was not ascertained how often this occurred in the tapetal cells but when studying the meiosis, counts were made in a hundred pollen mother cells at diakinesis. Of these only two were found with an extra bivalent and one with one bivalent missing. The variation seems to be within plants. Polytoca macrophylla Benth.— The tapetal chromosomes of this species differ from those of Coix in that there is no conspicuous differentiation into chromatic and achromatic regions. They stain more or less uniformly, as can be seen in pi. 10, fig. 3, where they are in endoprophase. Two nucleolus chromosomes with satellites can be observed in the same figure. Figure 4 of this plate shows them at endometaphase, more condensed and already divided as shown by the longitudinal split apparent along each chromosome. Conspicuous constrictions indicate the position of the centromeres. Plate 10, fig. 5, shows a late endometaphase. The divided halves of each chromosome are seen to diverge and are held together only at the centromere. p fig. 6). Such cruciform configurations due to union only at the centromeres are sometimes characteristic of colchicine- treated plants undergoing c-mitoses. Levan (1938) describes it in Allium. Brown (1949), speaking of endomitotic cycles m the tapetum of tomato, says that in some respects they resemble mitoses upset by treatment with colchicine. No cruciform stage, however, nor any ev.dence of protracted union at the centromere was observed by him. It is not known whether the union at the centromere in Polytoca macrophylla and Coix poilanei? ,s much protracted but the cruciform configurations form a definite stage in the endo- mitotic cycle of the tapetum in these two instances. In the contracted condition at endometaphase (pi. 10, fig. 4) , Polytoca chromo- somes bear a likeness to somatic chromosomes of corn treated with paradichloro- benzene. This produces effects very similar to those of colchicine, one of which is to shorten the chromosomes much beyond their normal length. Chionachne Koenigii (Sprengel) Thwaites.-In this species of Chionachne the „«~,1 ,I,, nmf «n mM . like those of Polytoca macrophylla, stain more or less uni- [Vol. 42 118 ANNALS OF THE MISSOURI BOTANICAL GARDEN formly. Plate 9, fig. 3, shows them in endoprophase. The presence of satellites on the nucleolus chromosomes is doubtful. What seemed like satellites were only- observed in one cell in which the chromosomes were in early endoprophase and were thin and long. They were not observed either in root-tip chromosomes or in pollen mother cells at diakinesis. The centromeres, though not marked by con- spicuous constrictions as in Polytoca, are easy to locate. At endometaphase, the chromosomes are very much contracted and the beginnings of a longitudinal split can be observed in some of them. MEIOSIS Observations were made on pachytene smears of Coix, Polytoca and Chionachne. The pachytene stage in Coix was found to be markedly different from that of Polytoca and Chionachne in that conspicuous knobs were observed in Coix that were absent in the other two genera. The part played by chromosome knobs in the determination of relationships between maize and its relatives is only too well known. Considering this feature of chromosome morphology, it could be assumed that of the three genera Coix is the most closely related to the New World Maydeae. Longley (1941) pointed out the similarity between the chromosomes of Xea and those of Coix. Comparing the bivalents of corn and Coix, he showed that the variation in size in the two sets of bivalents was very similar. A general similarity seems even more apparent on comparing the pachytene stages of a 20-chromosome Coix with those of corn. A glance at fig. 7 of pi. 10, however, shows that of the six knobs observable only one is internal and the rest terminal, a condition at variance with that in corn where the majority of knobs are internal. In this respect, therefore, Coix seems to resemble Euchlaena or Tripsacum more than corn (knobs in Euchlaena are both internal and terminal and in Tripsacum the majority are terminal). Figure 7 shows a pachytene smear from a variety of C. Lacryma ]obi obtained from the Sugar Cane Breeding Station, Coimbatore, India. From a study of its gross morphology this plant is classified as var. typica. Yet the number of its knobs differs widely from that of var. typica from Trinidad ("Blue Adlay"). This variation is quite in accordance with the situation in corn. Longley (1939), writing about such variation in knob number in corn, observes that it is "frequently very pronounced when plants of varieties from different geographical regions are compared". In corn, knob-forming centres can exist but not be evident unless a knob is formed at the knob-forming point, an occurrence which depends on several factors, one of them being the amount of knob material available (Longley, 1939). Figure 6 of pi. 10 shows a single chromosome of "Blue Adlay", with an elongated, terminal knob. In appearance and position it is exactly like one of the terminal knobs of var. typica from Coimbatore. It may well be that the position and number of knob-forming centres are similar in the two plants but that knobs make their appearance at more centres in var. typica from Coimbatore than in var. typica from Trinidad. The knobs of 1955] NIRODI ASIATIC RELATIVES OF MAIZE 119 var. typica from Coimbatore vary in appearance from large, somewhat elongated ones to smaller, more rounded ones (pi. 10, fig. 7). The pachytene chromosomes of C. poilanei? were in too tangled a state to allow proper observation. The knobs in C. gigantea? (pi. 12, fig. 1) are small, rounded and mostly terminal. No knobs were observed in Polytoca macrophylla and Chionachne Koenigii. If a chromomere analysis of Coix chromosomes were undertaken it would seem best to start with varieties of C. Lacryma Jobi. The pachytene stages are easy to handle and the chromosomes spread out well enough to be examined. The chromo- some number being the same as in corn, similarities with and differences from corn chromosomes would be easier to identify. Only late diplotene stages of C. poilanei?, "Blue Adlay" and Polytoca macrophylla were observed. These showed a maximum of three chiasmata per bivalent in the two Coixes and up to five in Polytoca (pi. 9, fig. 4 ; pi. 1 3 , fig. 3 ) . In C. poilanei? the chromosome pairs could be identified at this stage and have been marked in the figure with the same letters used to identify tapetal chromosomes. By late diakinesis in C. poilanei (?) (pi. 9, fig. 5), all ex- cept the satellite bivalent are completely terminalized and lie side by side or are connected at the ends by slender filaments. The bivalent of satellite chromosomes is still connected at the satellite end and free at the other, forming a V. All five bivalents could easily be identified even in this contracted phase and have been indicated by letters as in previous figures of C. poilanei (?). In "Blue Adlay", five rings and five V's were observed to make up the ten bivalents (pi. 11, fig. 2); the achromatic ends mentioned in describing the tapetal chromosomes are still noticeable and are especially clear in the distal end of the satellite bivalent. Diakinesis in C. gigantea (?) deviates somewhat from the perfectly normal for- mation of bivalents exhibited by C. poilanei (?) and "Blue Adlay". One association of 4 chromosomes was seen in some pollen mother cells and some showed two rings of 4 (pi. 12, fig. 5) but the majority had 20 bivalents. These three kinds of be- havior were found in pollen mother cells from the same spikelet. Evidently the two sets of 4 chromosomes that tend to form rings possess some incipient homology. C. gigantea (?) presumably has two pairs of nucleolus chromosomes. Two bivalents were often seen in association with the nucleolus at diakinesis and one pollen mother cell was observed which had two distinct nucleoli each associated with a bivalent. But the prominent satellites characteristic of Coix were only seen to be displayed by one of the bivalents (pi. 12, fig. 6). Both Polytoca macrophylla and Chionachne Koenigii behave normally at diakinesis, regularly forming 20 and 10 bivalents respectively (pi. 13, fag. 4; pi. 14, fig. 3 ) . The former has satellites to its two nucleolus chromosomes but none were observed in Chionachne. Formation of metaphase plates and separation at anaphase I take place in the normal fashion. There is an orderly migration to the poles; and no formation of [Vol. 42 120 ANNALS OF THE MISSOURI BOTANICAL GARDEN bridges, lagging chromosomes, nor any other irregularities were observed in any of the five plants (pi. 9, fig. 6; pi. 11, fig. 3; pi. 12, fig. 7; and pi. 14, fig. 4). At telophase I the chromosomes are seen to have formed two compact groups. By the time they emerge from telophase I and enter prophase II, the arms of the component chromatids of each chromosome are seen to diverge from the centro- mere with the result that the chromosomes appear cruciform. A nucleolus has been re-formed in each group. This behavior is similar in all five taxa under dis- cussion (pi. 12, fig. 3; pi. 11, fig. 6; pi. 13, fig. 1; pi. 14, fig. 1). At this stage, the second meiotic prophase, a curious body was seen to occur regularly in each of the two groups of chromosomes in C. gigantea (?). It ap- peared to be composed of a few darkly staining granules embedded in a lightly staining matrix (pi. 14, fig. 1). It varied in shape, being sometimes small and rounded, sometimes larger and lens-shaped. The number of granules also varied. An attempt to investigate the nature of this body was made by halving an anther containing pollen mother cells in second prophase and staining one half with aceto- carmine and the other with aceto-orcein. With aceto-carmine both nucleolus and the body showed up clearly, while the orcein-stained preparation stained neither the nucleolus nor the matrix. The granules showed extremely faintly. It would seem from this evidence that the body was in some way connected with the nuc- leolus. It was not seen in contact with any of the chromosomes or with the nucleolus, and so far as was observed was not seen to be in later stages. The remaining stages of the second meiotic division leading to tetrad formation are completed in normal fashion. In C. gigantea? and Ch. Koenigii at anaphase II separation of chromatids and migration along the spindle may not take place simultaneously in both cells (pi. 13, fig. 2; pi. 14, fig. 6). Relationship between Chromosome Number and Gross Morphology Coix Poilanei?, with 10 somatic chromosomes, is essentially a small plant and even when grown outdoors did not exceed a height of three feet. The varieties of C. Lacryma-Jobi, 2n = 20, of which several forms of var. "Ma-yuen" and var. typica from different regions were raised, showed a wide variation in size, some attaining a height of 3 feet, others growing 9 feet tall. Most of them averaged 7-8 feet. Coix gigantea?, with the highest chromosome number (2n = 40), exhibits characteristics typical of polyploids. Roxburgh (1832) describes it as attaining a height of 8-15 feet. Our plants of C. gigantea? were only between 5 and 6 feet high, but the leaves were coarser than in the two other species and the plants came into flower much later. Wh tionship between Chromosome Number and Geographical Distribution Species of Coix seem to have reached an optimum in the somatic number of 20 Jobi 195 J] NIRODI ASIATIC RELATIVES OF MAIZE 121 Discussion On finding that accurate chromosome counts could be made from tapetal cells, a study of their chromosome morphology was undertaken. While tapetal chromo- somes at the right stage for study were being sought it gradually became apparent that they undergo what is essentially a process of endomitosis. Endomitosis in tapetal cells was first reported by Witkus (1945) in Spinacia. Brown (1949) showed that tapetum endomitosis in tomato differed in some respects from the process in Spinacia. In both cases the process ends in the formation of polyploid nuclei but whereas in Spinacia the nuclear membrane remains intact throughout and the chromosomes are irregularly distributed, in tomato there is a breakdown of the nuclear membrane and the chromosomes form clumps or are arranged in plates at endometaphase. In spite of the breakdown of the nuclear membrane, Brown retains the term "endomitosis," explaining that "lack of movement of the chromosomes on a spindle rather than persistence of a nuclear membrane would seem to be the better criterion of endomitosis." In the present study the behavior of tapetal cells follows a somewhat similar pattern to that described in tomato. The cells are uninucleate to start with but soon become binucleate following a normal mitosis, i. e., with spindle formation but without the formation of a cell wall. The chromosomes become visible as sep- arate entities at endoprophase and contraction commences. At endometaphase the chromosomes are very much contracted and the beginning of a longitudinal split, prior to anaphase separation, is distinctly noticeable (pi. 8, fig. 9; pi. 10, fig. 4). However, the cruciform configurations at late metaphase described in C. poilanei? and Polytoca and presumably present in the other three taxa are reported by Brown to be absent in tomato. At endoanaphase the chromosomes form dense clumps. The fact that the chromosomes fall apart at this stage was only gathered from the small size of stray chromosomes at the periphery of the clumps. That the completion of endomitosis results in tetraploid nuclei was concluded from the larger size of many resting nuclei and from instances like the one por- trayed in pL 10, fig. 2, where a tetraploid nucleus of C. poilanei} is shown with its chromosomes in endoprophase of a second endomitosis. Whether this is con- cluded with the formation of an octoploid nucleus is not known Except in C. Poilanei} no tetraploid nuclei were observed in a state in which the chromosomes could be counted. For this to be possible the chromosomes would have to emerge from the resting stage and start on a second endoprophase, and this does not seem to happen in the other four taxa, and was only rarely observed in C potlana} Presumably then, endomitotic activity stops after one division and the formation of tetraploid nuclei, i. e., two tetraploid nuclei per tapetal cell. Brown reports three endomitotic cycles as normal to tapetal cells in tomato. . Sectioned material at endoprophase showed the nuclear membrane intact and fig 9 of pi. 8 shows an endometaphase in van "Ma-yuen ( 1 ) " with the membrane 122 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 still intact. Whether it disappears at a later stage is not definitely known but it seems unlikely. Assuming that it does not, the endomitosis in Coix, Polytoca and Chionachne resembles Spinacia regarding this point. According to Brown, in tomato the membrane disappears toward late endoprophase. From what study has been made on Coix chromosomes it would appear that the various species and varieties dealt with in this paper form a fairly homogeneous group, with 5 as the basic chromosome number. C. poilanei? would then be con- sidered a diploid species, C. Lacryma Jobi and its varieties tetraploids, and C. gigantea? an octoploid. Morphologically, the octoploid is similar to the diploid in an outstanding characteristic, the presence of crateriform glands bearing hairs. (Watt considers this characteristic as one of the more important ones that sep- arate what he terms "the gigantea- aquatic a series" from the Lacryma-Jobi series). In its chromosomes, however, the octoploid is more like the tetraploids. Since no crossing experiments were undertaken for this study any hypothesis regarding the possible origin of the tetraploid and octoploid should be made with the necessary caution. It could be surmised that the tetraploids arose by a crossing of two similar diploids and a doubling of the chromosome number of the hybrid. They function like true diploids forming no multivalents. In the octoploid two nucleoli have been observed, and there are two sets of nucleolus chromosomes. Multivalent configurations are found. It probably originated from the tetraploid as an autopolyploid. Since its origin, its genomes have undergone changes so that it has come to function essentially as a diploid but the occasional formation of tetravalents indicates its autopolyploid beginnings. On the basis of its pachytene chromosomes, Coix as a genus seems closer to the New World Maydeae than to Polytoca and Chionachne. In staining reactions, as previously described, Polytoca and Chionachne behave differently from Coix. Morphologically, they are similar in having the walls of their fruit-cases formed from the hardened first glume, whereas Coix has hardened, modified leaf- sheaths. The probability must not be overlooked that investigation of species of Polytoca and Chionachne other than those treated here might disclose the presence of pachytene knobs. The resemblance of endometaphase chromosomes of Polytoca to those of artificially shortened somatic chromosomes of corn has been mentioned. On the basis of the present evidence, it would seem that Polytoca and Chionachne bear a closer affinity to each other than to either Coix or to the New World Maydeae. Bibliography Avdulov, N. P. (1931). Karyo systematische Untersuchung der Familie Gramineen. Bull. Appl. Bot., Genet. & Pi. Breed. (Leningrad), Suppi. 43:1-428. Backer, C. A. (1927). Handboek voor de Flora van Java 2:32-33. Bailey, F. M. (1902). Queensland Flora 6:1848. Balansa, B. (1890). Catalogue des Graminees de PIndo-Chine Franraicp. Bennett, J. London. ne Francaise. Jour, de Bot. 4:76-84. Plantae Javanicae Rariores. pp. 15-19. 1955] NIRODI ASIATIC RELATIVES OF MAIZE 123 Bentham, George (1878). Flora Australiensis 7:515-516. , (1881). Notes on Gramineae. Jour. Linn. Soc. Bot. 19:14-134. •, and J. D. Hooker (1883). Genera Plantarum 3:1112-1115. Berger, C. A., and E. R. Witkus (1943). A cytological study of c-mitosis in the plant Spinacia oleracea with comparative observations on Allium cepa. Bull. Torr. Bot. Club 70:457-467. Bor, N. L. (1940). Flora of Assam 5:453-459. Brown, Spencer W. (1949). Endomitosis in the tapetum of tomato. Amer. Jour. Bot. 36:703-716. Burkill, I. H. (1935). A dictionary of economic products of the Malay Peninsula 1:629-631. Conger, Alan D., and Lucile M. Fairchild (1953). A quick-freeze method for making smear slides permanent. Stain Technol. 28:281-283, Cooper, D. C. (1933). Nuclear divisions in the tapetal cells of certain angiosperms. Amer. Jour. Bot. 20:35 8-364. Darlington, C. D., and E. K. Janaki Ammal (1945). Chromosome atlas of cultivated plants. 397 pp. London. Duthie, J. F. (1888). The fodder grasses of northern India. 90 pp. Roorkee. Gamble, J. S. (1934). Flora of Madras 3:1705. Hackel, E. (1891). Descriptiones Graminum novorum. Oest. Bot. Zeitschr. 41:5—9. , (1906). Notes on Philippine Gramine.ie. II. Philipp. Jour. Sci. Suppl. 1:263—269. Henrard, J. Th. (1931). A contribution to the knowledge of the Indian Maydeae. Meded. Rijks Herb. Leiden 67:1-17. (1938). On a new Chionachne from Queensland. Blumea 3:23 8-242. (1938). On a new species of Polytoca from Java. Ibid. 238-242. Heyne, K. (1927). Die nuttige planten van Nederlandsche Indie 1:150-152. Holland, T. H. (1926). Adlay — its uses and prospects. Dept. Agr. Ceylon, Yearbook 1926:60-61. Hooker, J. D. (1896). Flora of British India 7:100. Jacquin, J. F. von (1820). Eclogae Graminum. p. 60. Koorders, S. H. (1918). Beschreibung einer von Dr. Ouwehand im Toba-See, in Sumatra, entdeckten neuen Art von Coix. Jard. Bot. Buitenzorg, Bull. Ill, 1:190-191. Kuwada, Y. (1915). Ueber die Chromosomenzahl von Zea Mays L. Bot. Mag. Tokyo 29:83-89. Lamarck, J. B., A. P. M. de (1789). Encyclopedic Methodique. 3:422. Levan, Albert (1938). The effect of colchicine on root mitoses in Allium. Hereditas 24:471-486. Linnaeus, C. (1753). Species plantarum. pp. 972-973. Linnaeus, filius (1781). Suppl. PI. Syst., Veg., Gen. PL, et Sp. Pi. p. 434. Longley, A. E. (1939). Knob positions on corn chromosomes. Jour. Agr. Res. 59:475-490. i (1941). Chromosome morphology in maize and its relatives. Bot. Rev. 7:263-289. Loureiro, J. de (1790). Flora Cochinchinensis 2:550-551. Mangelsdorf, P. C, and R. G. Reeves (1939). The origin of Indian corn and its relatives. Texas Agr. Exp. Sta. Bull. 574:1-315. Merrill, E. D. (1906). An enumeration of Philippine Gramineae with keys to genera and species. Philipp. Jour. Sci. Suppl. 1:320-321. Mimeur, Genevieve (1951). Systematique specifique du genre Coix et systematique varietale de Coix Lacryma-J obi. Morphologie de cette petite cereale et etude de sa plantule. Rev. Internat. Bot. Appl. Agric. Trop. 31:197-211. Miiller, Carolus (1861). De graminibus novis vel minus cognitis. Bot. Zeit. 19:332-335. Mueller, F. von (1875). Fragmenta Phytographiae Australiae 8:116-117. Pieris, H. A. (1936). Adlay. Trop. Agr. (Ceylon) 86:217-219. Pilger, R. (1940). Gramineae. III. Unterfamilie Panicoideae. Engler & Plantl's Nat. Pflanzenfam. 14 e :184-201. Ranga Achariyar (1921). A handbook of some South Indian grasses, pp. 140-142. Reeves, R. G., and P. C Mangelsdorf (1935). Chromosome numbers in relatives of Zea Mays L. Amer. Nat. 69:633-635. Ridley, H. N. (1906). Malay drugs. Agr. Bull. Straits & Fed. Malay States 5:193-206. Roxburgh, W. (1832). Flora Indica 3:568-572. Royen, A. Van (1740). Flora Leydensis. p. 72. Salisbury, R. A. (1796). Prodromus stirpium in Horto ad Chapel Allerton vigentium. p. 28. Smith, F. H. (1933). Nuclear divisions in the tapetal cells of Galtonia candicans. Amer. Jour. Bot. 20:341-347. Sprengel, Curt. (1825). Systema vegetabilium 1:238-239. Stapf, O. (1SSS). Job's Tears (Coix Lachryma L. var. stenocarpa). Kew Bull. Misc. Inf; 1888:144-145. Steudel, E. G. (1854). Synopsis Plantae Graminum 1:403. Taylor, William Randolph (1925). Chromosome constrictions as distinguishing characters in plants. Amer. Jour. Bot. 12:238-244. [Vol. 42, 1955] 124 ANNALS OF THE MISSOURI BOTANICAL GARDEN Thunberg, C P. (1784). Flora Japonica. p. 37. Thwaites, G. H. K., and J. D. Hooker (1864). Enumeratio plantarum zeylaniac. p. 357. Vallacys, G. (1948). Le "Coix Lacryma-Jobi*\ Bull. Agric. Congo Beige 39:247-304. Warburg, O. von (1891). Beitrage zur Kenntnis der papuanischen Flora. Bot. Jahrb. 13:230-272. Watt, George (1889). A dictionary of the economic products of India. 2:492-500. , {1 904). Coix spp. or Job's Tears. A review of all available information. Agr. Ledg. 11:189-229. Weatherwax, P. (1926). Comparative morphology of the Oriental Maydeae. Indiana Univ. Studies 73:1-18. Wester, P. J. (1920). Notes on Adlay. Philipp. Agric. Rev. 13:217-222. Willdenow, C. L. (1805). Species Plantarum 4:202-203. Witkus, E. R. (1945). Endomitotic tapetal cell divisions in Spinacia. Amer. Jour. Bot. 32:326-330. Explanation of Plate PLATE 8 Figs. 1—4. Coix poilanei Mimeur. Tapetal chromosomes in endoprophase. Fig. 5. Coix poilanei. Tapetal chromosomes in endometaphase. Fig. 6. Coix poilanei. Tapetal chromosomes in late endometaphase. Fig. 7. "Blue Adlay" (C. Lacryma Jobi var. typica Watt), from Trinidad. Tapetal chromosomes in endoprophase. Fig. 8. Var. "Ma-yuen (1)" (C. Lacryma Jobi L. var. "Ma-yuen" Stapf). Tapetal chromosomes in endoprophase. Fig 9. Var. "Mayuen (1)". Tapetal chromosomes in endometaphase. Ann. Mo. Bot. Gard., Vol. 42, 1955 Pl 8 ^%m v? i i .;■.■ • ^ A* . .. «** S** - B C 2 « A ST* 3 IQM .>■-- a ^ / t 0. ~ i ^5 '■ , '••*■ >«& > - fc • • ■ - ■ ■■-" ■: r >- f ■ 2ft * -■ ■ *s & ■ .* j s *g 1 ,r& ■a ■ 4%v ,,,, • . r, •- -; 3 ¥ ft • ' ■ . . ■• 5; <**' '-r .<-.- T - . * - ■ * -^ ... * ^ - ■r - ^ - * *t !*$* V, • vv if r_l itf * % *€ pi fe. Jtts. £' 2 ■ - - % ; ^' • > y;^ -;'■:;>> . ** .;-'■■•. , 3 'c' ■■'■■;'■■ '.-■'■■• c 4 '^T ;.. v . ^.. D % , v-; . MM *-.■. ■ r. * t i S*"f 6 5 * . . ■ B tOM XOlt NIRODI _ASIATIC RELATIVES OF MAIZE [Vol. 42, 1955 J 126 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 10 Fig. 1. Coix poilanei Mimeur. Showing twelve tapetal chromosomes in endoprophase. Fig. 2. Coix poilanei. Tapetal cell with tetraploid number of chromosomes under- going a second endomitotic cycle. Fig. 3. Poly toca macrophylla Benth. Tapetal chromosomes in endoprophase. X 13 50. Fig. 4. Poly toca macrophylla. Tapetal chromosomes in endometaphase. X 13 50. Fig. 5. Poly toca macrophylla. Tapetal chromosomes in late endometaphase. Fig. 6. "Blue Adlay." Pachytene. X 13 50. Fig. 7. Coix Lacryma J obi var. typica, from Coimbatore. Pachytene. X 13 50. Ann. Mo. Bot. Gard., Vol. 42, 195 5 Plate 10 — ,■.. Miiiiu^w^iinigp.«spw ■ € **i I m S 1 I s I I Si; i ■ miiw^^ ■Eiiiiimm *"jvr*^ ** MB ^«* * - -***-p ^MMi * • MIRODI— ASIATIC RELATIVES OF MAIZI. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 1 1 2 .*& ■•*• • ■-■■-*,%.•* t« ■ ■ 1 / ^ « S ■ 1 I w= » Star- ts 1 3 rife £#> / -*3V •••~ ,- & - - 7 ,- ■ •£* ate? 2P ;' : •WSWi4'. 4 5 *|m « % w f?".s i<-*r. W »* r , Y ff? m 7 4fe .".:< .V. - ?*' %J* 6 NIRODI— ASIATIC RELATIVES OF MAIZE I Vol. 42, 1955) NIRODI ASIATIC RELATIVES OF MAIZE 127 Explanation of Plate PLATE 11 Fig. 1. Coix poilanei Fig. 2. "Blue Adlay Fig. 3. "Blue Adlay Fig. 4. "Blue Adlay Fig. 5. "Blue Adlay Fig. 6. "Blue Adlay Fig. 7. "Blue Adlay Mimeur. Telophase I of a 12 -chromosome cell »» Diakinesis. " Anaphase I. M Telophase I. »» Late Telophase I " Prophase II. M Metaphase II IVol 42, 19551 128 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 12 Fig. 1. Coix gigantea Koen. ex Roxb. Pachytene. X 13 50. Fig. 2. Coix poilanei Mimeur. Telophase I of normal 10-chromosome cell. Fig. 3. Coix poilanei. Prophase II. X 1150. Fig. 4. "Blue Adlay." Telophase II. Fig. 5. Coix gigantea. Diakinesis showing ring formation. X 13 50. Fig. 6. Coix gigantea. Diakinesis showing satellite bivalent. X 1800. Fig. 7. Coix gigantea. Anaphase I. X 13 50. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 12 «» &* *• ^ i \ % % \ %>" •* • % fe 1 ■ •» m. % \ % * % t • •• • • m f* #*» •* «# # 'mmmp N I RODI— ASIATIC RELATIVES OF MAIZI Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 13 * ■ : i j * ■**- c # *fc ■■ ■ ** ■'-•'■ " "-*-: i-:-: 1 ■ .-. -- ■ ■ ■ . — ■ NIRODI— ASIATIC RELATIVES OF MAIZE [Vol. 42, 1955] NIRODI— ASIATIC RELATIVES OF MAIZE 129 Explanation of Plate PLATE 13 Fig. 1. Coix gigantea Koen. ex Roxb. Prophase II showing extra bodies. X 990 Fig. 2. Coix gigantea. Anaphase II. X 1170. Fig. 3. Polytoca macrophylla Benth. Diplotene. X 875. Fig. 4. Polytoca macro p by lla. Diakinesis. X 13 50. Fig. 5. Polytoca macrophylla. Telophase I. X 1100. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 14 * ^U M^o a w »*< M^ 4 . . « .*fc ■ *rr*wy * NIRODI— ASIATIC RELATIVES OF MAIZE 130 [Vol. 42, 1955] ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 14 Fig. 1. Poly toca macrophylla Benth. Prophase II. X 1170. Fig. 2. Poly toca macrophylla. Telophase II. X 800. Fig. 3. Chionachne Koenigii (Spreng.) Thwaites. Diakinesis. Fig. 4. Chionachne Koenigii. Anaphase I. Fig. 5. Chionachne Koenigii. Telophase L Fig. 6. Chionachne Koenigii. Anaphase II. X 15 50. Fig. 7. Chionachne Koenigii. Telophase II. X 1350. LICHENOLOGICAL NOTES ON THE FLORA OF THE ANTARCTIC CONTINENT AND THE SUBANTARCTIC ISLANDS. I-IV CARROLL W. DODGE and EMANUEL D. RUDOLPH* In 1947-1948, the Australian National Antarctic Research Expedition estab- lished weather stations at Heard Island and Macquarie Island and somewhat later at Mawson in MacRobertson Land, on the Antarctic Continent. Through the kindness of Dr. H. W. Jessep of the National Herbarium, Melbourne Botanic Garden, and Dr. A. M. Gwynn, Medical Officer and Biologist of the Australian National Antarctic Research Expedition (A.N.A.R.E.), we have had the privilege of studying the lichen collections. L Additions to the Lichen Flora of Heard Island Previous collections were reported by Dodge (1948) based on the British Australian New Zealand Antarctic Research Expedition (BA.N.Z.A.R.E.), No- vember 27-December 2, 1929, all between Atlas Cove and Corinthian Bay. The weather station was established at Atlas Cove in December 1947 (Scholes, 1951) and closed in 1955. The earlier collections received in 1949 were by D. Alan Gilchrist, Medical Officer; the collector of the later specimens was not recorded on the herbarium labels and are cited: A.N A.R.E. The island was more accurately mapped in 1948 (A.N.A.R.E. 1949). Most of the southern part of the island is covered by glaciers and ice fields, so that the lichen collections have come from the northwestern part, especially the Cape Laurens peninsula on the northwest corner. Thirty- two species are represented, of which three are new and fifteen have not been previously reported although known from Kerguelen Island to the northwest, making a total of 52 species known from Heard Island. Thelidium heardense Dodge, B.A.N.Z.A.R.E. Rept. B. 7:44. 1948. The thallus is lighter (vinaceous buff) than the type, but it agrees micro- scopically. North of Cape Laurens on volcanic rock, A.N.A.R.E. 75. Thelidium praevalescens (Nyl.) Zahlbr., Deutsche Sudpolar Exp. 8:51. 1906. Verrucaria praevalescens Nyl., in Crombie, Jour. Linn. Soc. Bot. 15:192. 1876. As in most previous collections, all of our material is sterile although the thallus has a characteristic appearance. West Bay, A.N.A.R.E. 743; north of Cape Laurens, on broken lava, in cave, A.N.A.R.E. 67, 72. * Graduate Student, Henry Shaw School of Botany of Washington University. Issued June 23, 1955. (131) 132 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Microglaena kerguelena (Nyl.) Zahlbr., Deutsche Siidpolar Exp. 8:51. 1906. J , Bot. Brit. For. 14:22. 1876. Wertbii Zahlbr. New to Heard Island. Xanthopyrenia heardensis Dodge & Rudolph, n. sp. Type: Heard Island, north of Cape Laurens at base of black cliffs, A.N.A.R.E. 76- Thallus areolatus, areolis 0.3-0.5 mm. diametro subconvexis substipitatisque, marginibus liberis crenulatis, humectatis obscure flavo-viridibus, siccatis flavo- citrinis, homoeomericus; algae Xantbocapsa, coloniis sphaericis aut oblongis, vaginis flavo-brunneis, cellulis 7—8 xx diametro; hyphae 2-3 /x diametro inter colonias algarum. Perithecia immersa, 1—3 in quaque areola, subsphaerica, 200—250 xx diametro, ostiolo minuto; parathecium obscure brunneum, 15-20 /x crassitudine, cellulis polyhedricis; asci 12-15 X H5 /x, leptodermei; ascosporae octonae, imbricatim monostichae, late fusiformes, 19-24 X 8-9 tx, tenui cum halone dum in asco, dein liberae 24 X 10-11 /x, biloculares, cellula superiori majori, hyalinae, septo con- strictae. Thallus areolate, areoles 0.3-0.5 mm. in diameter, slightly convex above, sub- stipitate below, margins free, crenulate, dark yellow-green when moist, old gold to buffy citrine when dry, homoeomerous ; algae Xantbocapsa, colonies rounded to oblong, of 4-16 cells with a thick yellow-brown sheath at first, becoming densely packed in a homogeneous gel with abundant hyphae and cells more rounded, 7-8 /x in diameter, each with its own sheath about 2 /x thick; hyphae 2-3 /x in diameter, filling most of the interstices between the algal cells and colonies. Perithecia immersed or nearly so, 1-3 per areole, showing as minute dark brown to black points; subspherical, about 200-250 /x in diameter, ostiole small; wall dark brown, 15-20 /x thick, of polyhedral cells; asci 115 X 12-15 /x, thin-walled, 8- spored; ascospores imbricately monostichous, broad fusiform, 19-24 X 8-9 /x, with a thin halo while still in the ascus, 24 X 10-12 /x when free, bilocular, the upper cell larger, hyaline, slightly constricted at the septum. Steinera glaucelxa (Tuck.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:66. 1948. Pannaria glaucella Tuck., Bull. Torrey Bot. Club 6:57. 1875. Growing with Placopsis bicolor (Tuck.) B. de Lesd., on broken lava, north of Cape Laurens, A.N.A.R.E. 8o. New to Heard Island. Steinera Werthii Zahlbr., Deutsche Siidpolar Exp. 8:43. 1906. D. Alan Gilchrist 5 and unnumbered specimen. Lichina Antarctica Crombie, Jour. Bot. Brit. For. 14:21. 1876. On crystalline rock (sandstone?), A.N.A.R.E. New to Heard Island. 19SS] DODGE & RUDOLPH LICHENOLOGICAL NOTES, I~IV 133 ? Siphulastrum cladinoides Dodge, B.A.N.Z.A.R.E. Rept. B. 7:69. 1948. "We have doubtfully referred our material to this species. It forms compact hemispheric cushions up to 2.5 cm. in diameter. The habit resembles young dense tufts of Sphaerophorus fragilis Pers. from the Arctic and Subarctic but it is com- pletely different in structure. The thalli are stiffer than in the type. Although the tips of the ultimate branches become very dark brown, we have been unable to find any reproductive structures. It may represent a new species. North of Cape Laurens on broken lava, A.N.A.R.E. 78. New to Heard Island. PI. 15, fig. 3. Collemopsidium pyrenuloides Dodge & Rudolph, n. sp. Type: Heard Island, north of Cape Laurens, on humus in cave, A.N.A.R.E. 65. Thallus gelifactus, flavidus, siccitate fragillimus, membrano-foliosus, semi- pellucidus, lobis irregularibus sinibus rotundatis, erectis, subrugosis; homoeomericus; algae Xanthocapsoideae, cellulis 5-6 ^ diametro; hyphae sparsae, tenues. Apothecia lecanorina immersa, 0.4-0.5 mm. diametro; amphithecium non bene evolutum; hypothecium ca. 25 ^ crassitudine, hyphis periclinalibus intertextis; thecium 100 ^ altitudine; paraphyses 3-4 /* diametro, pachydermeae; asci clavati, juventute apice incrassati; ascosporae octonae, brunneae, late fusiformes, biloculares, septo constrictae, pachydermeae, 25-30 X 10-14 /a. Thallus a yellowish gel, very fragile when dry, yellow, foliose-membranous, semipellucid, lobes very irregular with somewhat rounded sinuses, erect, subrugose; homoeomerous ; algae Xanthocapsoid, cells mostly singly dispersed in the gel with- out evident sheath, 5-6 ^ in diameter, occasionally in small subsphencal colonies up to 40 M in diameter with thin sheaths about each cell and a somewhat thicker sheath surrounding the colonies, best seen in the amphithecium; hyphae very slen- moss (?) fragments. Apothecia lecanorine, immersed or nearly so, about 0.4-0.5 mm. in diameter; amphithecium not clearly differentiated, a zone of Xanthocapsoid colonies with more abundant subvertical hyphae; hypothecium about 25 ^ thick of interwoven periclinal hyphae; paraphyses 3-4 M in diameter with thick walls, occasionally branched; thecium 100 ,* tall; asci clavate, thickened at the tip when young, dif- fluent, 8-spored; ascospores brown, broad, fusiform, sometimes flattened on one side, bilocular, constricted at the septum, wall thick, protoplasts rounded, rarely 2 distinct protoplasts in each cell, resembling the spores of Pyrenula, 25-30 X ThT thallus suggests the Pyrenopsidaceae, but the algal cells mostly occur singly without an evident sheath and the gel is much softer when moist so that we have been unable to secure good sections. It is possible that the fungus is a parasitic Tichothecium, but the mycelium below the hypothecium is continuous with that of the thallus and it lacks a parathecium. Although C. Pyrenulodes is foliose, it seems more closely related to Collemopsidium than to any other genus of the Pyrenopsidaceae* [Vol. 42 134 ANNALS OF THE MISSOURI BOTANICAL GARDEN North of Cape Laurens, on humus in cave, A.N.A.R.E. 64, type, 67 sterile. Pannaria dichroa (Hook. f. & Tayl.) Crombie, Jour. Linn. Soc. Bot. 16:220, 1876. Lecanora dichroa Hook. f. & Tayl., London Jour. Bot. 3:643. 1844. Probably owing to the very rough surface of the lava, the lobes are somewhat ascendant and imbricate. The thallus is darker, light brownish olive and not stained with iron. On another very dense rock without locality, nearly covered by Bias tenia keroplasta Zahlbr., the thallus is quite typical. D. Alan Gilchrist 3; north of Cape Laurens on broken lava, A.N.A.R.E. 74' Lecidea Auberti B. de Lesd., Ann. Crypt. Exot. 4:99. 1931. D. Alan Gilchrist 8. Lecidea assentiens Nyl., in Crombie, Jour. Bot. Brit. For. 13:334. 1875. North of Cape Laurens, on volcanic rock, A.N.A.R.E. 75. New to Heard Island. Lecidea subassentiens Nyl., in Crombie, Jour. Bot. Brit. For. 14:21. 1876. D. Alan Gilchrist 4. Rhizocarpon kerguelense Dodge, B.A.N.Z.A.R.E. Rept. B. 7:116. 1948. Jacka Valley, on cliffs, A.N.A.R.E. 33. Rhizocarpon Mawsoni Dodge, B.A.N. Z.A.R.E. Rept. B. 7:115. 1948. D. Alan Gilchrist 7. New to Heard Island. Rhizocarpon Johnstoni Dodge, B.A.N.Z.A.R.E. Rept. B. 7:118. 1948. On moraine of Schmidt Glacier, A.N.A.R.E. 242. New to Heard Island. Cladonia phyllophora (Tayl.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:132. 1948. Cenomyce phyllophora Tayl. in Hook. f. & Tayl., London Jour. Bot. 3:652. 1844. North of Cape Laurens, A.N.A.R.E. 70. Cladonia Johnstoni Dodge, B.A.N.Z.A.R.E. Rept. B. 7:135. 1948. Some podetia are nearly clothed with coarse granules, rarely almost isidioid, thus somewhat resembling C. pyxidata (L.) Fr. Among mosses on broken lava, north of Cape Laurens, A.N.A.R.E. 6q. New to Heard Island. Argopsis cymosa (Crombie) Stzbgr., Ber. Th'atigk. St. Gall. Naturw. Ges. 1889-90:231. 1891. Stereocaulon cymosum Crombie, Jour. Linn. Soc. Bot. 15:182. 1876. On broken lava, north of Cape Laurens, A.N.A.R.E. 71. New to Heard Island. Ureceolina kergueliensis Tuck., Bull. Torrey Bot. Club 6:58. 1875 Jacka Valley, 600 ft — ^ -- — — — — -j — - - — — - _____ ___ , on cliff, A.N.A.R.E. 34. New to Heard Island. 1955] DODGE & RUDOLPH LICHENOLOGICAL NOTES, I~IV 135 Aspicilia lygomma (Nyl.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:164. 1948. Lecidea lygomma Nyl. in Crombie, Jour. Bot. Brit. For. 13:3 34. 1875. On boulders, 20 ft., West Bay, A.N.A.R.E. 743, growing with Thelidium praevalescens (Nyl.) Zahlbr. and Kuttlingeria crozetica (Zahlbr.) Dodge. New to Heard Island. Aspicilia disjunguenda (Nyl.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:167. 1948. Lecanora disjunguenda Nyl. in Crombie, Jour. Bot. Brit. For. 15:105. 1877. D. Alan Gilchrist 6, 7, p. 1 Aspiciliopsis macrophthalma (Hook. f. & Tayl.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:175. 1948. Urceolaria macrophthalma Hook. f. & Tayl., London Jour. Bot. 3:640. 1844. D. Alan Gilchrist 10. Placopsis bicolor (Tuck.) B. deLesd. Ann. Crypt. Exot. 4:100. 1931. Placodium bicolor Tuck., Bull. Torrey Bot. Club 6:57. 1875. The specimens are much paler than usual, probably from less iron in the rocks, and cephalodia are very rare. North of Cape Laurens, on broken lava, A.N.A.R.E. 8o, growing with Strinera glaucella (Tuck.) Dodge; on cliff, Jacka Valley, 600 ft., A.N.A.R.E. 31. Usnea Taylori Hook, f., London Jour. Bot. 3:657. 1844. Moraine of Baudessen Glacier, 800 ft., A.N.A.R.E. 250. New to Heard Island. Usnea insularis (Lamb) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:211. 1948. Neuropogon insularis Lamb, Jour. Linn. Soc. Bot. 52:215. pi. 8, fig. 17. 1939. Only a few plants, beginning to form cupulate apothecia. Mt. Aubert de la Riie, 300 ft., A.N.A.R.E. 4. Usnea trachycarpa (Stirton) Mull. Arg. Nuovo Giorn. Bot. Ital. 21:37. 1889. 1881 Neuropogon trachycarpus Stirton, Scottish Nat. 6:105. D. Alan Gilchrist 2. New to Heard Island. Blastenia keroplasta Zahlbr., Deutsche Siidpolar Exp. 8:28. 1906. Two specimens without locality, A.N.A.R.E. New to Heard Island. Kuttlingeria 1948. Calopl 1906. West A.N.A.R.E. 743; on cliff in Jac christ I. New to Heard Island. bplicata Nyl. in Crombie, J Jahrb. [Engler] 5:138. 1884. Rnt. For. 13:334. 1875. J [Vol. 42 136 ANNALS OF THE MISSOURI BOTANICAL GARDEN Rinodin a aspicilin a Zahlbr., Deutsche Siidpolar Exp. 8:50. 1906. Two small specimens without locality, A.N.A.R.E. DEUTEROLICHENES (LICHENES IMPERFECTI) Occasionally lichenologists have encountered conidial fructifications on lichen thalli, sometimes associated with apothecia, sometimes not. Miiller-Argau (1881) described an otiform conidial structure which he named a campylidium. Vainio found the same structure on a thallus of Lopadium perpallida (Nyl.) (1890) pallida Nyl.) and referred it to Cyphella Spegazzini (1909) ocyphella P Cyphella Karst. and described several new varieties. Mameli Calvino (1930) proposed the name Deuterolichenes to include Cblorocyphella, and Cengia Sambo (1937, 1941) and Rizzini (1952) have reported species of Cblorocyphella not associated with apothecia. Malme (1935) considered C. aeruginascens (Karst.) Keissl. to be a conidial stage of Lopadium per pallidum (Nyl.) Zahlbr., and Dodge (1953) de- scribed a campylidium as a conidial state of L. Deightoni Dodge. Campylidia have also been observed on Sporopodium sp. Miiller-Argau (1890) described another type of fructification, the orthidium, which resembles an apothecium except that the thecium is replaced by conidiophores, and the senior author has seen a similar fructification on a foliicolous thallus from Jamaica. Our material contains an orthidium-bearing lichen from Heard Island, resembling Ephelis Fr. of the Excipulaceae, which has unicellular spores, while those of our lichen become septate as do those of Ephelis trinitensis Cooke & Massee, the imperfect state of Balansia trinitensis Cooke & Massee on Panicum palmifolium in W EPHELIDIUM Dodge & Rudolph, n. gen. Ethel Thallus crustosus, indeterminatus, sorediosus, ecorticatus, heteromericus ; algae protococcoideae. Orthidium sessile, concavum, margine persistente; conidiophorae simplices; conidia singulatim disposita, terminalia, acicularia, septata. Thallus crustose, indeterminate, sorediose, ecorticate, heteromerous; algae protococcoid. Orthidium sessile, concave with a persistent margin, resembling a lecanorine (or biatorine as the algae die above) apothecium; conidiophores un- branched; conidia single, terminal, acicular, long remaining unicellular but finally multiseptate. Ephelidium heardense Dodge & Rudolph, n. sp. Pi. 15, fig- !• Type: Heard Island, Atlas cove at foot of Poa mound, A.N.A.R.E. 147. Thallus crustosus, indeterminatus, 0.7-1.2 mm. crassitudine, citrinus, sore- diosus; ecorticatus; stratum algarum ca. 280 ft crassitudine, cellulis protococcoideis, 8.7-12.2 /x diametro; medulla crassa, hyphis 1 /x diametro dense intertextis, nubilatis. 1955] DODGE & RUDOLPH LICHENOLOGICAL NOTES, I-IV 137 Orthidium sessile, basi constrictum, orbiculare, 0.6—1.5 mm. diametro, sub- ochraceum, margine 250-280 /x crassitudine; conidiophorae tenues, unicellulares, 13—17 fi longitudine; conidia singulatim disposita, acicularia, hyalina, 30—44 X 2 /x, primum unicellulares, dende ad 7-septata, recta aut subcurvata. Thallus crustose, indeterminate, 0.7-1.2 mm. thick, buffy citrine, sorediose, K orange brown, C — ; ecorticate; algal layer about 280 /x thick, cells protococcoid, spherical to somewhat polyhedral from mutual pressure, 8.7-12.2 /x in diameter; medulla thick, of closely woven hyphae about 1 /x in diameter, somewhat nubilated with granules and including pieces of roots etc. from the substrate. Orthidium sessile, constricted at the base, circular, 0.6-1.5 mm. in diameter, ochraceous buff, disc very concave; margin 250-280 /x thick, finally undulate; algae in a discontinuous layer on the outside of the layer of conidiophores, tending to die out above and forming a continuous layer below; the medulla around and between the algal colonies is formed of compactly woven hyphae about 1 /x in diameter; conidiophores arising from the medullary hyphae, forming a continuous layer 13-17 tt thick; conidia borne singly, acicular, ; septate until late, finally up to 7-septate, straight or slightly curved, slightly taper- ing at the ends. 0-44 II. Additions to the Lichen Flora of Macquarie Island Previous collections were reported by Dodge (1948). Most of the present collections are from the northern half of the island, collected mostly by Norman R. Laird and by N. M. Haysom of the A.N.A.R.E. Taylor (1954) has discussed the problem of distribution of the flowering plants which apparently have much wider ranges than the lichens. Twenty-seven species are reported from the present collection, three of which are new and three not previously reported bringing the total species of lichens to forty-four. Several other possible new species will be reported in a later number of these Lichenological Notes. In the following list MI/49/ has been omitted from N. M. Haysom's numbers of collections. Microthelia macquariensis Dodge, B.A.N. Z.A.R.E. Rept. B. 7:48. 1948. Z8 J Orange red when fresh, drying olive buff. The Trentepohlia filaments have few corticating hyphae. From glacial moraine above Sandy Bay, 500 ft., very few corticating hyphae. N. M. Haysom Z2. Psoroma versicolor (Hook. f. & Tayl.) Mull. Arg., Flora 71:53 8. 1888. Lecanora versicolor Hook. f. & Tayl., London Jour. Bot. 3:642. 1844, non Ach. The squamules are less well developed than in previous collections but the apothecia agree microscopically. Norman R. Laird 2, 2a, 2c; Featherbed Terrace, growing over decaying hepatics and other vegetable debris, A.N.A.R.E. [Vol. 42 138 ANNALS OF THE MISSOURI BOTANICAL GARDEN Pseudocyphellaria glabra (Hook. f. & Tayl.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:79. 1948. Stricta glabra Hook, f, & Tayl., London Jour. Bot. 3:647. 1844. South end of Plateau, on moss, N. M. Haysom Z114; north end of Plateau, on soil and over decaying mosses, Norman R. Laird; Featherbed Terrace, A.N.A.R.E.; Wireless Hill, N. M. Haysom Zi; north of Lusitania Bay, N. M. Haysom Z134. Peltigera Lairdi Dodge & Rudolph, n. sp. Type: Macquarie Island, growing on soil over decaying grasses, mosses and hepatics, Norman R. Laird 4, A.N.A.R.E. Thallus foliosus, in herbario isabellinus, ad 10 cm. diametro, 400 /x crassitudine, lobis sterilibus ca. 1.5 cm. latitudine, 2 cm. longitudine; tomentum crassum, verru- cosum in partibus junioribus; superficies inferior arachnoideus inter venas; venae sparsae, elevatae, arachnoideo-tomentosae, rhizinis simplicibus vel semel dichotome ramosis, ca. 5 mm. longitudine, concoloribus; cortex ca. 90 /x crassitudine, pseudo- parenchymaticus, cellulis 12—15 /x diametro leptodermeis; stratum algarum 50—55 /x crassitudine, cellulis nostocaceis, 5 /x diametro; medulla 270 it crassitudine, hyphis periclinalibus pachydermeis, 4 xt diametro, parte superiore compacta, inferiore laxi- ore, hyphis ad 8 ft diametro, lumine 5 /x. Apothecia in lobis erectis non revolutis, 1 5 mm. altitudine, 5-6 mm. latitudine, sita; discus nigro-brunneus, 4 mm. diametro; sine amphithecio nee parathecio, stratum algarum sub hypothecio crassiore, ad 120 /x crassitudine; hypothecium 30 /x crassitudine, brunneum, superne obscurius, pseudoparenchymaticum, hyphis periclinalibus; thecium 120-130 ju altitudine; paraphyses simplices, ca. 2 /x diametro; asci 100-105 X 11—12 /x, juventute apice incrassati; ascosporae octonae, hyalinae, aciculares, multiseptatae, ca. 56 X 3 /x. Thallus foliose, becoming isabella color in the herbarium, up to 10 cm. in diameter, 400 /x thick, margin lobed, sterile lobes about 1.5 cm. wide and 2 cm. long, somewhat lacerate on drying, ends of lobes with thick, verrucose tomentum above, becoming smooth but surface dull in the older portions as the tomentum weathers away; underside pinkish buff, arachnoid between the sparse veins; veins elevated, arachnoid- tomentose; rhizinae not abundant, simple or once-dichotomous near the tips, about 5 mm. long, concolorous; cortex about 90 /x thick, pseudo- parenchymatous, cells rather thin-walled, 12-15 /x in diameter; algal layer 50-55 fi thick, of N os toe colonies, cells about 5 xt in diameter in an inconspicuous gel; medulla 270 /x thick, of closely packed periclinal hyphae, thick-walled, about 4 /x in diameter, less compact and more irregularly arranged below, forming the arach- noid underside with hyphae up to 8 xx in diameter, lumen 5 /x. Apothecia on erect lobes, not revolute, 15 mm. tall, 5-6 mm. wide, disc very dark brown, 4 mm. in diameter; no amphi thecium nor para thecium differentiated; algal layer thicker below the hypothecium, up to 120 /x thick; hypothecium 30 /a thick, lower half pale brown, upper half very dark brown, pseudoparenchymatous from periclinal hyphae; thecium 120—130 tx tall; paraphyses about 2/x in diameter, unbranched, tips ending in the dark brown epithecial gel; asci 100-105 X 11—12 /*> 1955] DODGE & RUDOLPH LICHENOLOGICAL NOTES, I-IV 139 tips slightly thickened when young, 8-spored; ascospores hyaline, acicular, 56 X 3 /x, thin-walled, many septate. This species somewhat resembles Peltigera dilacerata Gyelnik from Auckland, New Zealand, which has longer and narrower lobes with dilacerate margins and is about half as thick. Lecidea subglobulata Knight, Trans. Proc. N. Zeal. Inst. 8:314. 1875 [1876]. South slopes of Wireless Hill, A.N.A.R.E. New to Macquarie Island. Catillaria (Eucatillaria) Rudolphi Dodge, n. sp. Type: Macquarie Island, without locality, but the rock types closely resemble those from the south slopes of Wireless Hill, A.N.A.R.E. Thallus crustosus, albidus, continuus aut ad centrum areolatus crassior, mar- gine irregulariter lobatus; cortex fastigiatus subdecompositus, erosus, ad 30 /x crassitudine; stratum algarum 65 /x crassitudine, cellulis protococcoideis 6-8 xt diametro; medulla 250-300 /x crassitudine, hyphis tenuibus, granulis brunneis nubilatis, dense contexta. Apothecia subimmersa aut sessilia, ad 2.5 mm. diametro, orbicularia, aut mutua pressione angularia, margine elevato, disco piano aut subconvexo nigro; para- thecium carbonaceum, in margine 50 /x crassitudine ad 125-160 /x sub thecio; hypothecium non bene evolutum; thecium 90-120 /x altitudine; paraphyses tenues, septatae, super ascos ramosae, apicibus clavatis brunneis, ca. 1.5 tt diametro; asci 50-60 X 7-8.5 /x, cylindrico-clavati, leptodermei; ascosporae octonae, hyalinae, biloculares, uniseriales, ellipsoideae, 13—15 X 6—7 /x. Thallus whitish, sometimes stained ferruginous from iron in the underlying rock, thin, continuous at the margin, thicker toward the areolate center, K yellow then orange, margin irregularly lobate, distinct; cortex fastigiate, somewhat de- composed and eroded, up to about 30 /x thick; algal layer 65 /x thick, cells pro to- coccoid, 6-8 /x in diameter in a continuous dense layer; medulla 250-300 xi thick, of densely woven slender hyphae, heavily nubilated with dark brown granules. Apothecia slightly immersed to sessile, up to 2.5 mm. in diameter, abundant, circular or angular from mutual pressure, margins elevated, disc plane to slightly convex, black; parathecium carbonaceous, 50 /x thick at the margin, 125-160 /x thick below the thecium; hypothecium not clearly differentiated; thecium 90-120 it tall; paraphyses slender, septate, branched at the level of the tips of the asci, tips slender, clavate, brownish, about 1.5 /x in diameter, asci 50-60 X 7-8.5 it, cyiindric- clavate, thin-walled, 8-spored; ascospores hyaline, bilocular, mostly uniseriate, ellipsoidal, 13-15 X 6-7 /i. The germinating ascospore produces an extensive black hypothallus. As con- tact is made with algal cells, thin assimilative areoles develop and finally fuse, covering the hypothallus as a continuous thallus which gradually thickens with age. Without locality but probably from south slopes of Wireless Hill, A.N.A.R.E., type; west side of Wireless Hill, N. M. Haysom Z152; south end of Plateau, N. M. 140 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Haysom Z137; Plateau, 900 ft., young, A.N.A.R.E.; north end of Plateau, N. M. Haysom Z86; without locality, Norman R. Laird 8. Cladia aggregata (Sw.) NyL, Bull. Soc. Linn. Normand. II, 4:167. 1870. Lichen aggregatus Sw., Nova Gen. Sp. Pi. Ind. Occ. 147. 1788. West coast, fluviatile swamp, 50 ft., A.N.A.R.E. Cladonia floriformis Dodge, B.A.N.Z.A.R.E. Rept. B. 7:134. 1948. Wireless Hill, 300 ft., Norman R. Laird id; Nuggets Creek, Norman R. Laird la, lb, young; without locality, Norman R. Laird I; only a few young plants among mosses, A.N.A.R.E. $c. Mawsoni 1948. Nuggets Valley, 100 ft., A.N.A.R.E.; ? primary thallus only, mouth of cave, south end of isthmus, A.N.A.R.E. Cladonia sarmentosa (Tayl.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:129. 1948. Cenomyce sarmentosa Tayl. in Hook. f. & Tayl., London Jour. Bot. 3:651. 1844. Interior of cave, 35 ft., A.N.A.R.E.; near Brothert Point, 250 ft., A.N.A.R.E.; near Nuggets Creek, Norman R. Laird ic, id, 2c; Lusitania Valley, east coast, 150 ft., A.N.A.R.E. Stereocaulon argodes NyL, Compt. Rend. Acad. Sci. [Paris] 83:88. 1876. Stereocaulon Argus Th. Fr., Nova Acta R. Soc. Sci. Upsal. Ill, 2 1 :333. 1858; Hook. f. & Tayl., London Jour. Bot. 3:653. 1844, pro parte min. For discussion of nomenclature, see Dodge (1948, pp. 142-144) sub Argopsis megalospora. In the present collection the dimensions of the ascospores are some- what greater, 45-53 X 5.5-7 /x, than those given by Th. M. Fries and Nylander. On coastal vegetation north of Lusitania Bay, N. M. Haysom Z128; Plateau, N. M. Haysom Zi. Stereocaulon pulvinare Dodge, B.A.N.Z.A.R.E. Rept. B. 7:139. 1948. Near Nuggets Creek, Norman R. Laird. * Stereocaulon submollescens NyL, Compt. Rend. Acad. Sci. [Paris] 83:88. 1876. Featherbed Terrace, A.N.A.R.E.; north end of Plateau, N. M. Haysom Z85. Pertusaria tyloplaca Nyl. Compt. Rend. Acad. Sci. [Paris] 83:90. 1876. South end of Plateau, N. M. Haysom Z122; Plateau, 900 ft., A.N.A.R.E.; north end of Plateau, N. M. Haysom Z59, Z62; growing over Menegazzia circum- sorediata Santesson on radio mast erected by A.A.E. in 1911 on Wireless Hill, A.N.A.R.E. Placopsis perrugosa (Nyl.) NyL, Lich. Nov. Zelandiae, 57. 1888. Lecanora 1865. For complete description, see I. M. Lamb, Lilloa 13:268-272. 1947. 1955] DODGE & RUDOLPH LICHENOLOGICAL NOTES, I-IV 141 On rocky banks of Lusitania Creek, N. M. Haysom Z123. New to Macquarie Island. macquariensi Type: Macquarie Island, Wireless Hill, on cliff subject to water seepage, A.N.A.R.E. Thallus fruticosus erectus aut subdecumbens, pulvinos hemisphaericos ad 3 cm. diametro formans, di- aut tri-chotome ramosus, torulosus, cinnamomeo-alutaceus, inferne ad 0.8 mm. diametro, superne tenuior, ad 1 cm. altitudine; ecorticatus; algae flavo-virides, cellulis sphaericis vel mutua pressione polyhedricis, 8-11 p. diametro; hyphae medullares ca. 3 fi diametro, conglutinatae, inter cellulas algarum. Apothecia ad 1 mm. diametro in lateribus ramorum sessilia, plana dein convexa emarginataque, ca. 250 fi altitudine, disco nigro; amphithecium 250 fi crassitudine; parathecium non bene distinctum, margine ad 40 /* crassitudine; hypothecium obconicum, ca. 190 fx. altitudine, hyphis verticalibus nigro- brunneis; thecium 135 /t altitudine; paraphyses 2.5-3 /a diametro, septatae, cellula terminali clavata 4 X 5.5 ju., pachydermea nigro-brunnea ; asci ca. 65 X 12 /*, clavati, juventute apice incrassati; ascosporae octonae, subdistichae in ascis 16 X 8 /*> biloculares, brunneae, septo tenui constrictae, liberatae, 12 X 6 /*, nigro-brunneae. Thallus fruticose, erect or subdecumbent, forming dense, depressed hemispheric cushions up to 3 cm. in diameter, branching di- or trichotomous, torulose, pinkish buff to cinnamon buff, up to 0.8 mm. in diameter below, more slender above, circular in cross-section, about 1 cm. tall; ecorticate; algae yellow-green, spherical or somewhat polyhedric from mutual pressure in the outer portion, 8-11 /1 in diameter with a thick sheath of periclinal medullary hyphae, more scattered throughout the medulla; hyphae about 3 fi in diameter, compact and conglutinate between the algal cells. Apothecia up to 1 mm. in diameter, sessile on the sides of the branches, not on the ultimate branches, flat at first becoming convex and emarginate, about 250 /* tall, disc black; amphithecium about 250 fi thick, similar in structure to that of the thallus but the medullary hyphae more vertical; parathecium not well differ- entiated, about 40 fi thick at the margin, scarcely differentiated from the para- physes, continuous below with the dark brown obconic hypothecium about 190 » tall at the center, of vertical dark brown hyphae, not sharply differentiated from the thecium above; thecium 135 /1 tall; paraphyses 2.5-3 fi in diameter, septate, terminal cell broadly clavate, about 4 X 5.5 /*, thick-walled, dark brown above; asci about 65 X 12 /*, clavate, tips very thick when young, 8-spored; ascospores subdistichous, 16 X 8 /t, bilocular, brown, ellipsoid, slightly constricted at the thin septum, shrinking to 1 2 X * A* when free from the ascus and very dark brown. The specimens have been cut from the underlying rock, but the cut ends sug- gest that the base may expand into a circular holdfast. The systematic position of this species is not clear. The bilocular brown spores with a thin septum, the para- physes and lecanorine apothecium suggest Rinodina sect. Beltrammta; the very dark hvoothecium suggests Melants pi cilia Vainio. From both of these it differs in 142 ANNALS OF THE MISSOURI [Vol. 42 its fruticose thallus. The structures of the thallus and of the apothecia are wholly different from those of the fruticose species of Anaptychia, but do resemble those of T hamnolecania from which it differs in a very dark brown hypothecium and broader brown bilocular spores. Its yellowish green algae suggest a relationship to Thelidea Hue, which is foliose with biatorine apothecia and hyaline ascospores. It is not a parasite as the medullary hyphae are continuous with those at the base of the hypothecium and the amphithecium is well developed; hence it cannot be considered a Karschia sp. on a Sipbula. Parmelia Turgidula Bitter, Hedwigia 40:246. 1901. Featherbed Terrace, over mosses. A single sterile plant. New to Macquarie Island but previously known from New Zealand. Parmelia sublugubris Dodge, B.A.N.Z.A.R.E. Rept. B. 7:188. 1948. Featherbed Terrace, A.N.A.R.E.; Norman R. Laird 3; north end of Plateau, N. M. Haysom Z.61. Parmelia tenuirima Hook. f. & Tayl, London Jour. Bot. 3:645. 1844. Featherbed fluviatile terrace on northwest slopes, growing over mosses, Norman R. Laird; from rock face, Half Moon Bay, west coast, A.N.A.R.E. Menegazzia circumsorediata Santesson, Ark. f. Bot. 30:11:14. 1942. Thallus 9 cm. in diameter. Wireless Hill, on radio mast erected by the A.A.E. in 1911, A.N.A.R.E.; coastal rocks north of Lusitania Bay, N. M. Haysom Z127. Usnea arida v. muscicola Dodge, B.A.N.Z.A.R.E. Rept. B. 7:207. 1948. Wireless Hill, on planks of A.A.E. radio hut, N. M. Haysom Z53. Usnea contexta Motyka, Lich. Gen. Usnea Stud. Monog. 2:436. 1937. Featherbed Terrace, A.N.A.R.E.; north end of Plateau, N. M. Haysom Z83. Usnea torulosa (Mull. Arg.) Zahlbr., Cat. Lich. Univ. 6:594. 1930. dasypogoides Our 1883. North of Lusi- tips thickened when young; ascospores numerous, ellipsoid, about 3.5 X 1-8 /x (not seen free from ascus). Spermogonia immersed, about 200 /x in diameter, wall hyaline, 15 /x thick, of thick-walled, conglutinate, periclinal hyphae; layer of spermatiophores invaginated, forming cerebriform cavities, spermatiophores 12-15 /x long, somewhat branched (Cladonia type) ; spermatia hyaline, narrow ellipsoid, about 3 X 1 /*• On weathered reddish granite, A. M. Gwynn Lt2i, Lt2y. Lecanora griseomarginata Dodge & Baker, Ann. Mo. Bot. Gard. 25:572. 1938. Growing on moss cushion, A. M . Gwynn Li20. New to MacRobertson Land. Lecanora exsulans (Th. Fr.) Dodge & Baker, Ann. Mo. Bot. Gard. 25:570. 1938. Lecanora chrysoleuca v. melanophthalma f. exsulans Th. Fr., Nyt Mag. Natur- vidensk. 40:208. 1902. A. M. Gwynn Li2l, Li24 y Li^S, U31 (a single young plant not sectioned) . J 1948. 1906. Growing over weathered rocks and mosses, A. M. Gwynn Li2i, U22, L123, IJ25, Li 28, Lis I. Alectoria congesta (Zahlbr.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:195. 1948. Parmelia pubescens v. congesta Zahlbr., Deutsche Siidpolar Exp. 8:! On rock, A. M. Gwynn Li2Q. Protoplastenia citrina Dodge, B.A.N.Z.A.R.E. Rept. B. 7:222. The color is more orange than in the type and all the material is steril the thallus agrees microscopically. Growine over moss cushions, A. M. Gwynn Li20, Li2l, Li25. 1948. 146 [Vol. 42 MISSOURI Gasparrinia Harrissoni Dodge, B.A.N.Z.A.R.E. Rept. B. 7:23 5. 1948. Thallus about 5 cm. long and 2 cm. wide, apparently starting near the edge of the rock fragment, so probably capable of forming a circular thallus about 5 ■ cm. in diameter when the substrate permits. A. M. Gwynn U26, L131, a smaller thallus. Xanthoria Mawsoni Dodge, B.A.N.Z.A.R.E. Rept. B. 7:236. 1948. M John 1948. A. M. Gwynn U28, associated with Lecanora exsulans (Th. Fr.) Dodge & Baker, as in the type. New to MacRobertson Land. Rinodina frigida (Darb.) Dodge, B.A.N.Z.A.R.E. Rept. B. 7:259. 1948. Buellia frigida Darb., Brit. Nat. Antarct. "Discovery" Exp., Nat. Hist. 5:Lich.:7. 1910. A. M. Gwynn Li2l, with lecanorine apothecia nearly immersed; L$2f $ U33, larger older thallus up to 6 cm. in diameter. Nostoc sp. Several thalli growing on a moss cushion, akinetes abundant. A few young perithecia or cleistothecia were seen which may be parasites, or perhaps the thalli are those of a Pyrenidiaceous lichen. Very few fungus hyphae were seen in the gel between the algal filaments but in one or two places there is a faint suggestion of a cortex. Some thalli contained a few bacterial colonies. We hope it will be found again in a more mature condition, if it be a lichen. A. M. Gwynn Lt20 y along with Lecanora griseomarginata Dodge & Baker and sterile Protoblastenia citrina Dodge. BIBLIOGRAPHY Australian National Antarctic Research Expedition (1949). [Map of] Heard Island from survey by Australian National Antarctic Research Expedition, 1948, scale 1:83,000. Cengia Sambo, Maria (1937). Lichenes Africae Orientalis Italicae. Nuovo Giorn. Bot. Ital. N.S. 44:456-473. 3 figs, [see p. 470]. , (1941). Fragmenta lichenologica III. Ibid. 48:495-516. [see p. 500-501], Dodge, Carroll W. (1948). Lichens and lichen parasites. British Australian New Zealand Antarctic Research Expedition (B.A.N.Z.A.R.E.) Rept. B. 7:1-276. , (1953). Some lichens of tropical Africa. Ann. Mo. Bot. Gard. 40:271-401. [see pp. 362-363]. Keissler, K. v. (1927). Systematische Untersuchungen iiber die Flechtenparasiten und lichenoide Pilze. V. Ann. Naturhist. Mus. Wien 41:157-170. [see p. 157-163]. Malme, G. (1935). Das Kampylidium, ein verkanntes Fortpflanzungsorgan der Flechten. Svensk Bot. Tidskr. 29:302-3 05. Mameli-Calvino, Eva (1930). Ricerche su una forma singolare di Deuterolicheni: Chlorocyphella subtropica Speg. Nuovo Giorn. Bot. Ital. N.S. 37:369-379. pi. 19: Miiller-Argau, J. (1881). Lichenologische Beitrage. XII (Schluss). Flora 64:100-112. [seep. 111- 112]. — — — , (1890). Lichenologische Beitrage. XXXIII. Ibid. 73:187-202. [see p. 202]. Rizzini C. Toledo (1952). Lichenes in Horto Botanico Fluminis Januarii crescentes. I. Arq. Jard. Bot. Rio de Janeiro 12:187-202. 4 pi. Scholes, Arthur (1951). Fourteen men, the story of the Australian Antarctic Expedition to Heard Island. London front. 5 p. 1. 273 pp. 10 pi. 1955] DODGE & RUDOLPH LICHENOLOGICAL NOTES, I-IV 147 * Spegazzini, C. (1909). Mycetes Argentinenses. Series IV. An. Mus. Nac. Buenos Aires 19:257-458. 40 fig. [see p. 279]. Taylor, B. W. (1954). An example of long distance dispersal. Ecology 35:569-572. Vainio, E. A. (1890). Etude sur la classification naturelle et la morphologie des lichens du Br£sil. [Thesis Univ. Helsingfors]. Acta Soc. Fauna Fl. Fenn. 7:1-247; Pars II, 1-256. [see part 2 p. 27]. [Vol. 42, 1955] 148 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 15 Fig. 1. Ephelidium heardensis Dodge & Rudolph. A. Cross-section of thallus. B. Cross-section of orthidium. C. Conidiophores. D. Conidiospores. Fig. 2. Acrospora Gwynni Dodge & Rudolph. A. Cross-section of apothecium. B. Cross-section of thallus showing spermogonium C. Ascus. D. Spermatiophores. Fig. 3. Collemopsidium pyrenuloides Dodge & Rudolph. A. Ascospores. B. Algal colonies. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 15 B 3 DODGE & RUDOLPH-LICHENOLOGICAL NOTES, I-IV Ann. Mo. Bot. Gard., Vol. 42, 195 5 Plate 16 1 l 2 DODGE H RUDOLPH— LICHENOLOGICAL NOTES, I-IV [Vol. 42, 1955] DODGE & RUDOLPH LICHENOLOGICAL NOTES, I-IV 149 PLATE 16 Fig. 1. Heard Island, Laurens Peninsula. Fig. 2. Heard Island, Coast of Atlas Cove. (Photographs taken by the Australian National Antarctic Expedition.) THREE NEW ANNONACEAE FROM PANAMA 1 ROBERT E. FRIES Anaxagorea Allenii R. E. Fr. n. sp. — Arbor circ. 6 m. alta; ramuli novelli brevissime rufo-tomentelli. Foliorum petiolus 10-12 mm. longus, crassus; lamina papyracea, (exsiccata) flavo-viridis, supra glaberrima laevisque, subtus demum glaber et pallidior, oblongo-lanceolata, basi rotundata, apicem versus sensim long- eque angustata, 30-35 cm. longa et 8-10 cm. lata; costa supra subplana, subtus valde prominens, teres, rugosus; nervi laterales I. utrinque circ. 15, sub angulo circ. 80° exeuntes, deinde sursum curvati, sicut venulae subtus paulo elevatae. Pedicelli florum breves (3-5 mm. longi), crassi, sicut sepala et petala rufo-tomentelli, juxta calycem bractea vaginaeformi instructi. Sepala ovato-orbicularia, apice rotundata, non recurvata, 7-10 mm. longa. Petala crassa, exteriora oblonga, apicem rotund- atum versus haud angustata intus apicem versus carinata, ad 12 mm. longa et 5 mm medium intus carinata. Stamina linearia, 3 mm. longa; appendix connectivi tenuis, circ. 0, 7 mm. longa, apice rotundato-truncata. (Fructus ignotus). Canal Zone: Quebrada Lopez, alt. 30 m., in flower, Feb. 11, 1940, P. H. Allen 2143 (Type in Herb. Missouri Bot. Garden) . This species is very like the Anaxagorea crassipetala Hemsl., found in Guate- mala and Nicaragua, but differs from this especially in its leaves being larger, rounded at the base, and with much-elongated, almost thread-like apices; in its pedicels being short and thick, in its sepals being longer and of rounded oval, not triangular-shape, and its outer petals longer and not tapering toward their apices. it is also similar to Anaxagorea dolichocarpa Sprague and Sandwith, from which it differs by sepals rounded at the tips and not retroflexed, short inner petals, shorter stamina, etc. Ramuli novelli pilis brevibus fer- tn^^^^^prtsm Y&^U. Foliorum petiolus teres vel supra planus tomen- panamensis 4-5 primo lorum glabrescens, oblongo-oblanceolata, supra medium latissima, basi breviter acuta, apice abrupte contracta et 1-1.5 cm. longe cuspidata, 15-20 cm. longa et 4-5 cm. lata; costa utrinque elevata; nervi validiores utrinque circ. 15, arcuatim adscen- dentes et 1-2 mm. a margine conjunct!, tenues et supra elevati; reticulum venularum in foliis adultis haud conspicuum. Inflorescentiae ex axdhs f delapsorum evolutae; pedicelli florum crassi, ferrugineo-sencei breves (7 mm. longi vel ultra?), paulo supra basin articulati. Sepala inter se libera, rotundato- ovata, acutiuscula, circ. 2 mm. longa. Petala rigida, exsiccata nigrescentia, ex- teriora rotundata et extra ferrugineo-sericea, circ. 7 mm. in diam., intenora minora, medium ferrugineo-tomentosa. Stamina 1.8-2 mm. longa, nervum issued June 23, 1955. (151) [Vol. 42 152 ANNALS OF THE MISSOURI BOTANICAL GARDEN apice truncata. Ovaria 1, 5 mm. longa, prismatica et breviter strigosa, stigmate globoso-ovoideo coronata; ovula 3—4, lateralia. (Fructus ignotus). Panama: Cerro Campana, trail from Campana to Chica, alt. 600—800 m., with flowers Aug. 10, 1941, Allen 2647 (Type in Herb. Missouri Bot. Garden). The material is unfortunately scanty, with no fruits at all. The number and lateral positions of the ovules, however, prove it to be well differentiated from the Central American species U. Pittieri Saff. and Schippii R.E.Fr. Flowers of the third and last Central American species, costaricensis R.E.Fr. (Acta Horti Bergiani 13:105. 1941) are unknown so far, but its fruit structure with one single oblique to horizontal seed favours the view that, structurally, the flower agrees with that of Pittieri-Schippii, and accordingly differs from panamensis. Unonopsis panamensis also differs from costaricensis in having much larger and differently shaped leaves. In my survey of the species of this genus (Acta Horti Bergiani 12:237) panamensis should be ranged among the guatteriodes y obovata, and allied species, but cannot be joined to any of these. Guatteria inuncta var. caudata R. E. Fr. nov. var. — Dif f ert a typo f oliis pro rata angustioribus, basi longe cuneato-angustatis, apice acumine 2-3.5 cm. longo caudiformi acutissimo ornatis, pedicellis florum tenuioribus. Bocas del Toro, vicinity of Chiriqui Lagoon, Old Bank Island, von Wedel 2108 (Type in Herb. Missouri Bot. Garden) . THE BOTANICAL CATALOGUES OF AUGUSTE DE ST. HILAIRE JOHN D. DWYER* The recent essay of Jenkins 1 serving as an introduction to Auguste de St. Hilaire's Esquisse de mes voyages au Bresil et Paraguay focused considerable atten- tion on this French naturalist who collected plants and animals in southeastern Brazil during 1816-1822. Few of the 30,000 plant specimens collected by St. Hilaire 2 have found their way into other botanical institutions from the herbaria of the Museum National d'Histoire Naturelle in Paris or the Universite de Mont- pellier 3 . This has contributed to the failure of many modern monographers of Brazilian plants to cite St. Hilaire material. While studying at the Museum National d'Histoire Naturelle in Paris I came upon the ten handwritten books constituting the catalogues of plants collected by St. Hilaire 4 . It is my hope to demonstrate the importance of these catalogues and also to rectify some errors made in the handling of the St. Hilaire collections. I wish to express my thanks to Prof. H. Humbert, Directeur Laboratoire de Phanerogamie, Museum National d'Histoire Naturelle, Paris, France, for permission to examine and to have microfilmed the catalogues of Auguste de St. Hilaire, as well as to use the facilities of the Herbarium. I owe especial gratitude to Monsieur J. Leandri, Sous-Directeur of the same institution, for his substantial assistance. To Mr. George Swett, Examiner of Questioned Documents, Clayton, Mo., I extend my heartfelt appreciation for his comparison of handwritings found in and related to the catalogues. ST. HILAIRE IN BRAZIL St. Hilaire set foot on Brazilian soil at Rio de Janeiro on June 1,1816. He was destined to be the first French naturalist to penetrate into the interior of south- eastern Brazil. Rio de Janeiro, the infant capital of Portuguese Brazil, was from the time of Vellozo in 1780 the hub of botanical activity in South America. Three years before the arrival of St. Hilaire, Georg von Langsdorff, a German, found the capital an ideal center for his botanical work. Two English collectors who botanized in Rio de Janeiro and in Sao Paulo, Brazil, James Bowie and Alan ■ 1 Jenkins, A. E., in Chron. Bot. 10 1 :5-4L 1946. 2 Rapport sur le voyage de M. Auguste de Saint -Hilaire dans le Bresil et les Missions du Paraguay, lu a 1'Academie Royale des Sciences, de Jussieu Rapporteur. 8 pp. Paris, 1823. This rare reprint may be found in the library of the Missouri Botanical Garden, St. Louis, Missouri, as well as the Oliveira Lima Library, Catholic University, Washington, D. C. 3 In America, for example, there are less than 100 sheets of St. Hilaire material deposited in the Smithsonian Institution (U.S. National Herbarium), Washington, D. C, and less than 40 sheets in the Missouri Botanical Garden, St. Louis, Mo. 4 1 wish to express my gratitude to the National Science Foundation, Washington, D. C, for the grant of money to defray in part the expenses of travelling to the Vlllth Botanical Congress held in Paris in the summer of 1954; at the Museum d'Histoire Naturelle I had the opportunity to ex- amine the catalogues first hand. I prefer to speak of the books as catalogues, despite the fact that it may be more accurate to designate the ten books collectively as a catalogue. * Missouri Botanical Garden and St. Louis University. (153) 154 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Cunningham 5 , in their unpublished (Brazilian) diary speak of visiting "Mr. Langs- dorfPs garden" on January 10, 1815, Four fellow-countrymen of Langsdorff had already penetrated into the interior of southeastern Brazil: Sellow, von Chamisso, Eschscholtz, and Maximilian, Prinz zu Wied-Neuwied. The most illus- trious German systematist of this era, von Martius, was soon to arrive (July, 1817). Urban's account of St. Hilaire's five "voyages" informs us that St. Hilaire met two of these botanists, Langsdorff and Sellow 6 . One of the most concise and accurate summaries of the travels and activities j mi travels presented to the French Academy by Baron Cuvier and five other members. This represents the acceptance by the Academy of St. Hilaire's Apergu d y un voyage dans Vinterieur du Bresil of the previous year 7 . I have taken the liberty of trans- lating a section of this "summary" from the French. Wherever possible, the spelling of geographical names has been changed to that of modern usage. After speaking of St. Hilaire's arrival in Brazil we read: In the following December he proceeded northwards, in the province of Minas [Gerais], a territory as vast as France, where he sojourned for 15 months in order to explore its various parts. At first he headed in the direction of Villa Rica, penetrating into old and very extensive forests, covering in turn open country with a different flora, with different birds and insects. There he discovered a "quinquina" unlike the Peruvian kind, but used in the same way, and serving as an indicator of iron. Likewise he observed a grass sticky to the touch, growing luxuriantly in habitats formerly covered with virgin forests. After an illness of a month's duration in Villa do Principe, he travelled through other forests, situated to the east of this town and studied the survivors of several ancient Indian tribes. He noted in the "district" of Minas Novas a change in the flora, with woods composed of only small shrubs. Further on, other trees of greater stature appeared, called Catingas, which are leafless during the dry season. He spent fifteen days on the banks of the Jequitinhonha in the midst of the Botocudos, warlike Indians who live in the forests, naked and without definite dwelling places. From there he turned toward the mighty San-Francisco river, followed along its banks and entered the District of the Diamonds [Diamantina] and visited the areas where he had previously noted their methods of extracting gold from the mountains and rivers. Returning to Villa Rica by a long detour, he again headed for Rio (de) Janeiro arriving in March of 1818, whence he sent to the Museum d'Histoire Naturelle a shipment of 200 birds, some quadrupeds, 800 species of insects, and two hundred packets of seeds, as well as two botanical papers which have been printed in the publication of the Museum. A second voyage of several months duration brought him to the maritime districts, to the north of Rio de Janeiro, as far as the Rio Doce, and yielded just about the same materials as those collected in the woods adjacent to the capital with the exception of those markedly different ones found on the sandy expanses. He visited Cabo Frio, the town of San Salvador de Campos, the entire province of Espirito Santo, the unhealthy banks of the Rio Doce, and having embarked at Villa da Victoria, he returned by sea to Rio de Janeiro, whence he sent a new shipment of animals and of seeds to the Museum. His third voyage, the longest and most important, begun in January of 1819, was at first toward the northwest in the eastern section of the province of Minas [Gerais] where he saw a beautiful waterfall at the source of the Sao Francisco, and the sulphur springs of Araxa whose waters cattle drink unhesitatingly. To arrive at Goiaz he followed a barren plateau 5 Bowie and Cunningham were to depart from Rio de Janeiro in October of 1816, the former to collect at the Cape of Good Hope, the latter in Australia. Their unpublished Brazilian journal is at the Royal Botanic Gardens, Kew, England. 6 Urban, I., in Martius* Flora Brasiliensis l 1 :93-98. 1906. 7 St. Hilaire, Auguste de, in M£m. Mus. d'Hist. Nat. 9:307-380. 1822. ► 1955] DWYER BOTANICAL CATALOGUES OF ST. HILAIRE 155 whose opposite sides form the sources of the Sao Francisco river and the Dos Tocantins both entering the sea from different directions. The vegetation struck him as being little different from that seen on the banks of the Sao Francisco. The pastures were at times entirely barren, at times covered with small stunted trees. M. de St-Hilaire after having visited Villa Boa, capital of the province of Goiaz, spent a few days among the Coyapos Indians, then went to the border of Matto Grosso; from here he took a southeastern route to Sao-Paul. He arrived there in December 1819, after a long and arduous journey marked by great heat and dryness, and here he deposited his various collections, packed with care, to permit him to continue southwards. At first he skirted the western side of the vast chain of mountains which separate the interior of the country from the seashore, and saw in succession the following villages: Sorocaba, Itu, Porto Feliz and their environs; he noted the boundaries of the most important colonial industries. Then he entered Campos Gerais, such delightful country because of its diversified countryside. He made a goodly collection of plants which indicated a more temperate climate. There he observed the erroneously named her be du Paraguay; this he correctly identified as a tree of the genus Ilex. He also identified the local quinquina which is in reality a Solanum. Crossing over a chain of mountains at a difficult spot called Serra de Paranagoa, he found himself at the sea coast at a much lower altitude and hotter than the interior plateau whose crops approximate more those of the tropics even as far south as 27° latitude. He was able to visit the islands of St. Francis and of St. Catherine and their whaling-posts. Here the ground beyond is a sandy and dry stretch enlivened only by the presence of a prodigious number of aquatic birds. The province of Rio Grande, which he soon entered, is very fertile, with a more vigorous population. He observed that the culture of sugar-cane stops at about 30° while cotton- growing extends to 31°, palms to 34°, just at noted in New-Holland [Australia]. He spent the winter, already severe in these areas, at Porto Alegre and at Rio Grande, departing in October 1820, that is in the springtime of this country. The plants which flourish at this time belong to the same genera as those constituting the spring flora in Europe, such as the carex [sedges], anemones, centunculus [chatfweeds], arenaria, etc. He recognized, too, the influence of climate on vegetation: thus, in the coldest season, the trees retained their foliage, up to a degree north of Porto Alegre; a third lose their leaves at Rio Grande, and at two degrees further south scarcely a tenth have retained their leaves. M. de St. Hilaire then entered the Spanish possessions, passed through several towns and beautiful estates along the banks of the Rio de la Plata, arriving at Montevideo where the vegetation appeared to him to be so closely related to that of Europe that he scarcely found fifteen plants which could not have been assigned to European families. He also noted around certain towns that several of our own plants were well acclimated; he called attention to this especially near Montevideo, where vast waste-lands were covered with our cardan [thistle] growing wild and used solely as fuel. Proceeding then in the direction of the Rio Negro, a river rushing headlong into Uruguay, he found beyond a country considerably less populated and with customs agreeing with the descriptions of Azzara. His journey became more dif- ficult between Belem and the province of the Missions, and he spent thirteen days in an uninhabited desert, populated only by autruches [rhcas], cerfs [deer], and jaguars, against which he had to defend his own life, and especially that of his horses; unfortunately he could not save all of the horses. It was in this district that curious accidents befell him and his two companions after having eaten several spoonfuls of a honey made by a species of wasp. He visited the seven small villages which are all that remain of thirty which constituted in time past the ancient missions of Paraguay. Observing these carefully, he could naught but groan at finding only ruins and the sight of such misery in a fertile land whose former splendor our writers have not exaggerated. Then re-entering Brazil proper and traveling in the rainy season, he frequently ran the risk of losing his notes and collections; he returned at the end of a year by a detour to Porto Alegre, embarked on Lake Patos for Rio Grande; returned by sea to Rio Janeiro, where he arrived toward the end or 1821, after an absence of nearly three years. St. Hilaire has elaborated on his travels in his own Esquisse (loc. cit.) and his Voyages. These are in need of careful study by contemporary scholars. The recent remarks of the Brazilian P. R. Reitz (Anais Botanicos do Herb. Barbosa Rodrigues 1:86. 1949) on St. Hilaire reflect the warm affection of the Brazilians for the French naturalist. I have taken the liberty of translating freely the same from the Portuguese: [Vol. 42 156 ANNALS OF THE MISSOURI BOTANICAL GARDEN He had a kind heart, adapting himself to the environment in which he lived and like- wise to the uneducated people with whom he often traveled, treating with the manners of a gentleman those who helped him, censuring amiably those who treated him badly. He described objectively the customs of our lands. St. Hilaire stands in marked contrast to those pseudo-investigators who, as tourists, describe our country as a grand carnival from the seats of gambling halls or even from the summit of Corcovado 8 where they sit to write their observations, their memoirs of the people, of the Brazilian way of life and customs. Only a St. Hilaire, a Martius or a Bonpland who came to Brazil with great sacrifice, who traveled through inhospitable areas, could sense the pulse of the Brazilian heart not only in the cities but in the back country and small villages, could see in their habitats great diversity of plants, animals, and minerals. Only these or others of their calibre could describe, in the role of investigators, our natural riches and our people. There is an excellent map of the itinerary of St. Hilaire's five "voyages", pre- j the Lntos for Chronica Botanica (loc. cit. p. 12), from St. Hilaire's back of his Voyage a Rio Grande do Sul. A large map found in the first volume of Martins' Flora Brasiliensis labelled "Itinera Botanicum" with inked- in itineraries of the majority of the principal botanists who collected in Brazil from 1780 to 18 52 is unfortunately inaccurate, at least with respect to St. Hilaire. For example, the delineator indicates that St. Hilaire collected as far north on the eastern coast of Brazil as Villa de Porto Seguro (Bahia) ; this is almost 225 miles north of the Rio Doce, the most northeastern limits of St. Hilaire's collections. THE CATALOGUES OF ST. HILAIRE The catalogues are made up of separate books in St. Hilaire's own hand- writing. Approximately 8900 collections are entered in the 1575 pages 9 . My count from the catalogues stands in marked contrast to Urban's statement (Martius 9 Flora Brasiliensis l 1 :92-98. 1906) of the number deposited in three herbaria, that: "Collectio princeps (cr. 7600 numeri) in herbario musei historiae naturalis Parisiensis. Dupla in Montpellier, parca in museo Berolinensi". Urban's estimate may have been the result of adding up the collection numbers in the catalogues without considering that often single collection numbers were subdivided one to several times, or from St. Hilaire's own statement (Esquisse, p. 61) that: "Le nombre des plantes que j'ai recueillies s'eleve a environ sept mille ....". In the catalogues there are approximately 835 collection numbers which are subdivided a la "bis", rt ter", "4", etc. In one instance (in series C 2 ) the collection num- ber 2791 is subdivided thirteen times and applied to eleven different families of angiosperms. The failure on the part of monographers or copyists to add "bis", "ter", etc. to a specific collection number has caused considerable confusion in the citing of specimens. Following each collection number is (usually) the specific plant family (written in French) to which the collection belongs. Occasionally only the generic name is given. For approximately 550 collection numbers St. Hilaire fails to supply any taxa although usually one finds an "analysis" or a few words. In more than one 8 A high peak overlooking Rio de Janeiro easily reached in our times by street-car. 9 In a letter recently received, M. Leandri writes: "Our catalogue of entries bears actually 7692 gatherings [collections] from St. Hilaire; maybe the others have been lost or kept by A. de Jussieu and Cambessedes (and others)." 1955] DWYER BOTANICAL CATALOGUES OF ST. HILAIRE 157 tc third of the entries a detailed description ("analysis") of the plant is given in French. In some families, for example in the Orchidaceae, these descriptions may take up three full pages. St. Hilaire paid particular attention to listing the com- mon names of the plants as well as their economic uses. The place of collection is usually found at the end of the "analysis". 10 The "analyses" are models of pre- cision and critical observation. St. Hilaire himself states (Esquisse, p. 61) je les ai toutes analysees sur les lieux-memes, et me suis principalement attache a la dissection des parties dont la connoissance repand le plus de lumieres sur les rapports naturels". Dreuzy, a great-nephew of St. Hilaire, in an appendix to the Voyage a Rio Grande do Sul, refers to the difficulties encountered by Saint Hilaire in preparing his "journal botanique", the name by which Dreuzy apparently designates St. Hilaire's diary and his catalogues collectively: Nous donnons ci-dessous quelques extraits de ce volumineux journal pour faire comprende au lecteur quel travail s'imposait l'auteur au soir de chacune de ses fatigantes journees de voyage, assis sur ses malles sous un rancho, eclaire par la lueur tremblante d'une chandelle fumeuse, devore par les moustiques qu* attirait la lumiere. Dreuzy then supplies two samples of St. Hilaire's "analyses" from catalogues C 2 , Section 2, one collection number 2682, and the other 278 f l \ To my knowl- edge these are the only "analyses" which have been published in toto. In fact, I have found no reference of any length in the literature concerning St. Hilaire's mav use Dreuzv's sample 2682 as typical of an "analysis" of catalogues. We St. Hilaire: 11 Estancia de Santo-Rei, 1" mars, n° 2682. Graminee. Sur chaque dent de 1'epi sont 2 fleurs, Tune sessile, l'autre pediculee. Epillets sessiles, 1 fl. herm., lanceoles, sous-falciformes, glumes 2-valves, a valves ext. herisee, lanceolee, tantot simplement aigue, tantot courtement 1 ou 2 arretees a valv. int. un peu plus courte que l'autre lanceolee-aigue, pubescence. Glum- elles 2-valv., a valv. membraneuses sous-violettes, sous-pubescentes presque egales, obtusiuscules, 2 paleoles tronquees; 3 et 2 styles setaces, 2 stig., complets, continus, plumeux, ovaire glabre ovoide. Epillets pedicelles males. Glumes 2 valv. a valve ext. pubescente, lanceolee-lineaire- aigue, plane, a valve int. lanceolee-aigue, presque egale a l'autre pubescente carenfce, 2 paleoles tronquees. 2 et. glumelles 2 valv. a valves lanceoUes, hneaires, aigues, presque egales, ciliees, membraneuses, un peu violettes. Toute la plapte repand une forte odeur de citron. Nom. vulg. capim Ihnao. Unfortunately, St. Hilaire made no great effort to list binomials in the cata- logues. In the more than 8900 collection numbers there are approximately 160 binomials, the majority in St. Hilaire's handwriting and a few in Spach's. Most of 10 The catalogues do not list as many specific localities as does Urban in his elaborate list of the itineraries of St. Hilaire in Martins* Flora Brasilivnsis (be. at). Here more thar ,800 local- given arranged in proper sequence as well as according . tc ; the • J-- ^^ "^es five ities are "voyages". It is probable that Urban gleaned this detailed itinerary from St. H.la.re s own pub- lished reports of his trips. "Unfortunately, in 2682 Dreuzy terminates this "analysis" of a Graminee with the binomial Ecbites guaranitlc* Aug. de S.-H. This binomial, despite the fact that ,t appears ,n the catalogues to be written into this "analysis" at the upper right-hand corner obviously refers to co lee .on number above, 2681 (Apocynacee). To make matters worse, the second sample analyse 1S labelled "Composee 2787**'." From the catalogues the collection number is simply Z787. [Vol. 42 158 ANNALS OF THE MISSOURI BOTANICAL GARDEN the binomials are addenda to the "analyses"; those written by St. Hilaire belong as a rule to families on which he published extensively: e. g., Violaceae (including cer- tain genera now assigned to the Ochnaceae) , Droseraceae, etc. For some reason "analyses" of certain families, e. g., Malvaceae, Malpighiaceae, etc. are often cross- hatched in ink. In the case of the Rutaceae the numbers of the family were designated by a scrawled name which looks like "Oranger" (pi. 17, under 2?QI 12 ). Figures 1 and 2 show the number of collections made in specific families (with the exception of the ferns) according to the catalogues of St. Hilaire. Figure 1 lists the families represented by more than 50 collection numbers; fig. 2 lists those represented by less than 50 and more than 20 collection numbers. The family Leguminosae is considered in sensu lato to include the Mimosaceae and Caesal- piniaceae. The Ochnaceae includes certain genera which St. Hilaire related to the Violaceae, e. g., Sauvagesia, Lavradia. To avoid confusion I have combined the Liliaceae with the Amaryllidaceae. NUMERICAL SEQUENCES OF ST. HILAIRE COLLECTIONS In preparing a collection of plants for herbarium deposit the most convenient sequence of collection numbers is a continuous linear one which is correlated with the date of collection and the itinerary of the collector. In examining the more than 8900 collections of St. Hilaire we find the following numerical sequences 12 in the ten books: Book 1 — 1- 758 Book 6 — 86- 96 Book 2- } A f+2 f??n Book 7- ) *~ 60 < not in Su Hilaire,s handwritin S> Book 4 Book 5 1- 758 464- ■ 650 1B- 132B 1- 63, 103- -2085 1- 112 2101- -2489 1- • 408 Book 3— 1- 63,103-2085 $1176-1899,2000-2797 Book 8-j 1-219 Book 9 — 692- 709, 1858-1869, 2004-2120 Book 10 — 1- 818 Despite the patch-work pattern of these sequences of numbers definite cor- relations exist between sets of numbers and the itineraries of St. Hilaire's five voyages. It is not impossible that St. Hilaire prepared the catalogues following his return to Paris in 1822. I have seen certain individual "analyses" attached to herbarium sheets in the Paris Herbarium, e. g. to one of several of St. Hilaire's un- numbered collections of Lavradia capillaris (Ochnaceae). It may be that the "analyses" were written on separate sheets of paper in the field, inserted separately with the unmounted specimens and then removed in Paris to be gathered together and recopied into a catalogue. The mounting of his 30,000 specimens in Paris may have resulted in some confusion in the rc-organization of the "analyses" in the catalogues. Occasionally, one finds the date 1822 at the end of an analysis, sug- gesting that an addendum was made to the original more-abbreviated "analysis". The "analyses", too, arc generally written in a neat (but often difficult to decipher) 12 In a given run of a number in the catalogues it is not uncommon to find occasional numbers missing or out of place. For the sake of brevity I am not specifying these, although the absentees have been taken into account in computing the sum-total of collection in the catalogues. 19SSJ DWYER BOTANICAL CATALOGUES OF ST. HILAIRE 159 COMPOSITE U€GU»MO*AE *U8U€EAE HCLASTOUACCAC ClAMlMEAf EUPHOffBUCEAE lAtUTAC ORCmioacCaE RfRN* MTtTACtAC SOtAWACEAE AC AM TH ACE AC APOCYMACCAf HAtVACCAC POCYOAtACEAC CT*fRA«Af VCBBEMACCAf CONVOCVUtACCAC SCROPMULARIACCaE MAtHlOHtACEAC amaramthaCEaE • ICHONIACCAt ASCtE^IAOACCAE BORRAQNACEAE AAURTtLIOACEAE Lf THRACCAC QEriTlAHACEAE U*BELLlFERAf 100 200 500 400 500 600 700 800 900 1000 1100 Fig. 1. Approximate numbers of collections (except for Ferns) in the botanical catalogues of collections in excess of 50 are shown. listed under specific Aug. de St. Hilaire. families of Angiospermae Only those families with CA»Y0HnrUAC|Ae WfTTIMftAE OtUGtACEAt OCHHACEAC ERICACEAE tmoCAULONAClAE URTICACCAE LOBELIACCAE 5A^!MDACCAf CUCURBfTACEAf LEMTIBUtARIACEAC 'ASSIPLORACCA1 OXAtlDACEAC ^'PERACCAE LAURACEAE RUTACEAE ftmiOMACfAf ROSACEAE TURWfRACEAE IRIDACEAE oioscorcaceac MTRStMACCAf AMHOMAClAt 20 25 30 35 40 45 50 Fig. 2. Approximate numbers of collections listed under specific families of Angiospermae in the botanical catalogues of Aug. de St. Hilaire. Only those families with at least 20 and not more than 49 collections are shown. [Vol. 42 160 ANNALS OF THE MISSOURI BOTANICAL GARDEN hand. Perhaps legibility is too much to expect of a scientist working in the evening after an exhaustive day in the field. In Book 5 Weddell writes that certain notes for collection numbers 2085-2101 were not sent to the Museum for insertion. Several points argue against this possibility, however. In the first place, it is more logical, if not more safe, to insert the "analyses" initially into a field book; again we note, especially in the early books, that St. Hilaire often fails to complete an "analysis" before beginning the next collection number but occasionally completes the "analysis" on another page and consequent to a different collection number. It is difficult to believe that he would do this if he were preparing the catalogues in Paris. The fact that volume 6 has only 1 1 collection numbers sandwiched in with what is obviously a segment of his diary written in the field is another substantial point of evidence. St. Hilaire and his contemporaries used these numbers in the catalogues of col- lections and for herbarium sheets, and not for purposes of citation in the literature. J 'dionalis, Jacque No doubt, they did not consider individual collections to be as important as do our botanists of today who adhere to the principles of the type concept. Obviously, if a collection of St. Hilaire had been cited, e.g.,No. 26, the reader would wonder to which sequence to assign it unless the locality were given. Wed- dell 13 , no doubt stimulated by the pioneer efforts of certain authors in Martins' Flora Brasiliensis in citing collectors and collection numbers, attempted (probably in the 1860's) to organize the catalogues in a manner more suitable for citation and for labelling of specimens. Unfortunately, Weddell did not publish his "Key" to the catalogues 14 . The facts that Spach, who was curator of the herbarium at the Museum d'Histoire Naturelle, employed Weddell's system on printed labels and that certain monographers have used the Weddellian system give some status to the key. We shall discuss the popularity of Weddell's key in following para- graphs. 13 H. A. Weddell (1819-1877), an "aide-naturaliste" to Adrien de Jussieu, collected extensively in Brazil; in addition he was a distinguished taxonomist at the herbarium of the Museum d'Histoire Naturelle, Paris. 14 In recent years some members of the staff of the Museum d'Histoire Naturelle, Paris, prepared a typewritten copy of Weddell's key. Weddell's handwritten key may be found scattered in the front of the catalogues of St. Hilaire (see pi. 18). An accurate handwritten summary (not figured here) of the sequences of numbers (accompanied by broad geographical data) of St. Hilaire's botanical collections occurs in the initial two pages of Book 8 of the catalogues. Mr. Swett, Ex- aminer of Questioned Documents, Clayton, Missouri, expresses the probability that this was prepared by a person (or persons) other than St. Hilaire, Weddell, or the author or authors of the page shown in pi. 19. It was obviously not prepared by Spach. I have elected to credit Weddell with the establishment of the key to St. Hilaire's collections. That Weddell inserted notes in numerical sequences and geography in many of the books of St. Hilaire seems certain. M. Leandri of the Laboratoire de Phanerogamic Museum D'Histoire Naturelle, Paris, writes me that the typewritten key was^copied from notes in the books [St. Hilaire's catalogues] partly of the handwriting of Weddell". While Weddell may have based his insertions on the well-organized summary found in book 8, it seems safer to suppose that the latter was prepared from his notes. 1955] DWYER BOTANICAL CATALOGUES OF ST. HILAIRJE 161 WEDDELL'S KEY TO THE CATALOGUES OF ST. HILAIRE Weddell divided the book series: A (books 1, 2, 3, and 4) ; B (books 5 and 6) ; C (books 7, 8, and 9) ; and D (book 10). Series A is divided into two volumes, B, into two volumes, C, into three volumes, and D consists of one volume. The following revised outline is Weddell at Paris: SERIES A (TWO VOLUMES) i A*j comprising numbers I— 758. Plants collected during the year 1816, some in the vicinity of Rio de Janeiro (1—469 and 651- 758), others in the vicinity of Uba, on the Rio Paranhyba about 25 leagues from Rio de Janeiro, and in the same province (470—650). A 2 , in two sections. Section 1, comprising numbers 464-650, which are duplications of the same numbers in Volume A 1 . Section 2, comprising numbers 1 B -130 B (or 132 B ). Plants from the environs of Rio de Janeiro collected in 1818. SERIES B (TWO VOLUMES) B 1 , comprising numbers I—2085 (December 1816 to March l8l8j. — Plants collected in the province of Minas Gerais, with the exception of 1-25, which are from the province of Rio de Janeiro. B 2 , in two sections. Section 1, comprising the numbers 2101 b,s to 2493; these follow in order the series oi plants of Minas Gerais which constitutes the three previous volumes. Section 2, comprising the numbers 1-408. Plants collected during a "voyage" on the seacoast of Brazil, from Rio de Janeiro to the Rio Doce in the provinces of Rio de Janeiro and Espirito Santo, from August to November, 1818. SERIES C (THREE VOLUMES) C 1 , comprising numbers I-II75. Plants collected in the following districts: 1- 10 — Province of Rio de Janeiro. 1819. 11- 625 — Province of Minas Gerais. 1819. 926- 985 — Province of Minas Gerais. 1819. 626- 920— Province of Goiaz. Sept. 1819 to May, 1820. 986-1175 — Province of Sao Paulo. Sept. 1819 to May, 1820. C 2 , divided into two sections. — Section 1, comprising numbers 1176-2797. Plants collected in the following regions; 1176-1704 — Province of Sao Paulo. May, 1820. 1705-1806— Province of Sta. Catarina. May to June, 1820. 1807-1874— Province of Rio Grande do Sul. 1820. 2569-2797— Province of Rio Grande do Sul. 1820. 2000-2260 I - From San Pedro de Rio Grande to Montevideo. 1820. 2261-2568 — From Montevideo to the Brazilian frontier on the banks of the Uruguay. 1820. Section 2, comprising the numbers 1-2 19c. Plants collected in the environs of Rio de Janeiro, in July, 1821. Comprising three small series of numbers (692-709, 1858-1870, and 2004-2020) tercalated in the large series of numbers contained in volumes C and C . 15 Weddell referred to these as volumes. [Vol. 42 162 ANNALS OF THE MISSOURI BOTANICAL GARDEN SERIES D (ONE VOLUME) D, comprising the numbers 1—818. — Plants collected in 1821 and 1822 during the last voyage of St. Hilaire, in the following places: 1_ 30 — Province of Rio de Janeiro. 31— 587 — Province of Minas Gerais. 588- 818 — Province of Sao Paulo. Certain elements of confusion exist in the interpretation of these numbers. In Book 2 the numbers 464-650 are simply duplicates ("double emploi") of the same numbers in the 1—758 sequence of Book 1. The numbers 1—112 in Book 4 are duplicates of those found in Book 3 except for 64-102 which are missing in Book 3. The numbers 86-96 in Book 6 (if one should refer to this as Book 6 in- asmuch as it is apparently one of St. Hilaire's diaries) belong to the sequence of numbers in Book 5 from which they have been omitted. In Book 7 the initial sequence of numbers 1—60 are not in St. Hilaire *s handwriting 16 . In Book 8 num- bers 1900-1999 are missing, apparently due to carelessness. Weddell (?) used paper labels for all the specimens of the collections except numbers 1-758 in catalogue A 1 (Rio de Janeiro), 1-25 in B 1 (Rio de Janeiro), and 25-2085 in B 1 (Minas Gerais), for which parchment labels were used. All paper labels, except for numbers 2101 b! " to 2469 of B 2 section 1 (Minas Gerais), and 2470 to 2493 B 2 section 1 (Rio de Janeiro) bear, or should bear, symbols. This is summarized as follows: TABLE I Labels Numbers Catalogues Localities Paper b 1- 240 B 2 Sect. 2 Rio de Janeiro 241- 408 B 2 Sect. 2 Espiritu Santo Paper B 1- 130 A 2 Sect. 2 Rio de Janeiro Paper C 1- 219 C 2 Sect. 2 Rio de Janeiro Paper D 1- 30 D Rio de Janeiro 31- 587 D Minas Gerais 588- 818 D Sao Paulo Paper O 1- 10 C 1 Rio de Janeiro 11- 625 C 1 Minas Gerais 626- 691 c 1 Goiaz 692- 709 c 3 Goiaz 710- 920 c 1 Goiaz 926- 985 c 1 Minas Gerais 986-1175 c 1 Sao Paulo 1176-1704 C 2 Sect. 2 Sao Paulo 1705-1763 C 2 St a. Ca tar in a 1764-1775 C 2 Sao Paulo 1776-1806 C 2 Sta. Catarina 1807-1857 c 2 Rio Grande do Sul 1858-1870 c 3 Rio Grande do Sul 1871-1874 c 2 Rio Grande do Sul 1875-2003 c 2 Banda Oriental del Uruguay 2004-2020 c 3 J— fc Banda Oriental 2021-2568 c 2 Banda Oriental 2569-2797 c 2 1 Rio Grande do Sul 16 These 60 numbers have neither families nor "analyses" appended, although the localities of the collections are given. Some of the localities (e. g., Barbacena, Rio Ouro Branco) suggest the begin- ning of St. Hilaire's second voyage to Minas Gerais in 1816. On this voyage he was accompanied by I. Gomes and G. H. de Langsdorff. Perhaps these are collections of one or both of these men. 19S5] DWYER BOTANICAL CATALOGUES OF ST. HILAIRE 163 THE ROLE OF THE CATALOGUES OF ST. HILAIRE IN TAXONOMIC MONOGRAPHS Although only a few botanists such as Wcddell, Spach, Dreuzy, etc., are known to have consulted the catalogues of St. Hilaire, the various check-marks and binomials inserted suggest that they may have been studied by others. Prior to an arbitrary date of 1860 monographers in citing St. Hilaire collec- tions, in imitation of St. Hilaire, Cambessedes, and de Jussieu failed (usually) to supply collection numbers but gave clear-cut geographical data. Later mono- graphers often used St. Hilaire collection numbers, and, almost without exception, they furnished only a fraction of collections as listed in the catalogues. Mueller, for example, in his classic work on the Apocynaceae (Mart. Fl. Bras. 6 1 : 1-195. 1860) gave collection numbers for only 23 of the 56 collections of St. Hilaire which he cites. These 56 collections represent about one-third of this family as listed in St. Hilaire's catalogues, all of which have specific collection numbers. In general, in Martins' Flora Brasiliensis it is the exception to find collection numbers when St. Hilaire material is cited; collecton numbers are missing, for example, in Radlkofer's monograph of the Sapindaceae (1895), Bennett's work on the Poly- galaceae (1874), Koehne's work on the Lythraceae (1887), etc. (see Table II). No doubt much of the confusion and inconsistency encountered in the citation of St. Hilaire's collections, especially by the 19th century monographers, arise from several circumstances: (a) localization of the bulk of the St. Hilaire collections in the Paris herbarium; (b) the fact that the duplicate collections often have labels bearing only a binomial and the geographical data (often undecipherable) 17 ; (c) the lack of interest in collection numbers by the contemporaries of St. Hilaire; (d) awkward and overlapping sequences of numbers prior to Weddell's groupings; (f ) the subsequent failure to apply the Weddellian system to many of St. Hilaire's collections. Table II, based on an examination of twenty monographs (1832 to the present) which include St. Hilaire's collections, reveals many important points 18 . In the Anacardiaceae (D.C. Monograph. Phanerogam. 4. 1883) Engler cites eleven col- lection numbers of St. Hilaire, none of which has catalogue numbers; in attempt- ing to check them against the catalogues it appears from the blank spaces in the latter that St. Hilaire may not have been certain of the characters of the Anacardiaceae and thus never entered the family name. In Cogniaux's Cucur- bitaceae (D.C. Monograph. Phanerogam. 3. 1881) the number of St. Hilaire's collections cited exceeds the total number estimated to be entered in the catalogues under this family. This disparity is a reflection of the number of blanks left by 17 Of the 59 sheets of St. Hilaire reported by Dr. Lyman B. Smith, of the Smithsonian Institute (Washington, D. C), to be in that herbarium only 34 bear collection numbers; of these only two have the Weddellian series appended to the number. I have been unable to locate ten of the num- bered specimens with certainty in the catalogues. . 18 When the collection numbers are not listed in a monograph the author may be including more than one collection (of St. Hilaire) for the species in question without making the same clear. This makes an accurate statistical analysis of the number of collection cited impossible. 164 ANNALS OF THE MISSOURI [Vol. 42 Specific Monograph Malpighiaceae A. de Jussieu Melastomaceae Naudin Apocynaceae Mueller Polygalaceae Bennett Erythroxylaceae Peyritsch Alismaceae Micheli Commelinaceae Clarke Cucurbitaceae Cogniaux Turneraceae Urban Anacardiaceae Engler Lythraceae Koehne Rubiaceae Tribes VII-XIX Schumann Guttiferae Vesque Sapindaceae (Serjania) Radlkofer Bromeliaceae Mez Bignoniaceae Bureau Ochnaceae (Ouratees- Orthospermees) Van Tieghem Passifloraceae Killip Vochysiaceae (Salvertia & Vocbysia ) Stafleu Labiatae (Hyptis) Epling Date 1832 1849-1852 1860 1874 1878 1881 1881 1881 1883 1883 1887 1889-1889 1893 1896 1896 1897 1902 1938 1948 1949 TABLE II 0> ^ 6/3 4'i 64 51 58 30 8 7 11 36 15 11 24 8 4 19 19 2 7 17 8 100 497 r* x WO O 3 O | .2 64 51 35 30 8 1 24 7 19 1 1 1 8 250 I u o o I c/5 E G O u n 23 7 11 36 14 11 1 4 19 1 6 17 7 92 249 V5 C .2 '■M s c •i e to CO CO CO o — 0* E 3 6 35 9 1 10 4 2 67 c o — -H o v- o c ElS ■ 1 * CO fM^ 0% o o o Ut v-» "<« o • M 9 X -Q m £ 3 *-* C/5 Z c 97 378 147 122 6 18 35 28 62 386 46 35 14 91 41 22 15 244 1878 1955] DWYER BOTANICAL CATALOGUES OF ST. HILAIRE 165 St. Hilaire after collection numbers in the catalogues, some of which are un- doubtedly of the Cucurbitaceae. In Naudin's monograph of the Melastomaceae 19 , we find very few of St. Hilaire's collections cited despite the fact that Naudin must have been well acquainted with them 20 . "What is even more remarkable is the complete absence of St. Hilaire material in the "second half" of the monograph (i. e., vol. 16, after p. 87. 1851), even though such common Brazilian genera as Clidemia, Miconia, etc. are included 21 . In the several sample monographs from Martins' Flora Brasil- iensis we note that few collections of St. Hilaire are cited as compared with the total number of collections of the specific families estimated to be in the catalogues. Killip, in his recent treatment of the Passifloraceae (Field Mus. Nat. Hist. Bot. 19:1-613. 1938), cites six numbers which do not check with the corresponding ones in the catalogues. Of these, three (717, 751, and 970) were probably erroneously copied from labels bearing the numbers, 719, 753, and 976 respectively, all of which are entered in the catalogues under Passifloraceae; a fig. 9 may be easily mistaken for a 7, a 1 for a 3, etc. Epling, in his monograph of Hyptis (Rev. Mus. de La Plata, n. s. Secc. Bot. 30:153-497. 1949) cites more St. Hilaire material than any contemporary mono- grapher 22 . Unfortunately, only two of the hundred collections cited bear collec- tion numbers with Weddell's series added. This suggests, of course, that the Weddellian system has not been applied in any great degree to the labelling of specimens. Despite the quantity of St. Hilaire material cited by Epling, approxi- mately 6 5 collection numbers of Hyptis were not included in his work. In checking this genus in the Paris Herbarium 13 of the 92 collection numbers (St. Hilaire) cited by Epling, were not located; but eight of these thirteen can be located readily when "bis", "ter", etc. are appended. (Some copyist of course may have failed to insert "bis", "ter", etc. on the labels.) Of the remaining numbers, prob- ably 578 was misread for 518, and 1155 for 1133. In those monographs using the Weddellian system the collection numbers are usually matched with ease in the proper volume in the catalogue. "Ann. Sci. Nat. Bot. Ser. Ill, 12:196-284. 1849; 13:25-39 126-159 273-303 .347-362. ,850; 14:53-76, 118-165. 1850; 15:43-79, 276-345. 1851; 16:83-246. 1851; 17:305-382. 1852; i g • g 5 i 54 258—294. 1852. "St. Hilaire and' Naudin were co-authors of a paper entitled "Revue de la Flore du Brttl Meridionals which appeared in three sections (Ann. So. Nat. Bot. Ser II, J'- 2 *- 5 *. 20 »""S. 1842). In addition they were co-authors of a new genus, Augustine*, of the Melastomaceae (Ann. Sci. Nat. Bot. Ser. Ill, 2:145. 1844). 21 1 hope to discuss in another paper the stormy history of the Melastomaceae collected by St. Hilaire. , "Hyptis is one of the few cases where St. Hilaire wrote a genus name with fair consistency into his catalogues in his "analyses" instead of the family name "Lab.ee". However, ,n checking Epling's monograph against the catalogues, it is obvious that St. Hda.re d.d on occas.on use the taxon of family in place of Hyptis. 166 [Vol. 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN A PARTIAL LIST OF THE GENERA COLLECTED BY ST. HILAIRE In the herbarium of the Museum d'Histoire Naturelle, Paris, there exists an unpublished list of about 2000 entries of St. Hilaire's collection numbers; these are arranged according to genera (occasionally binomials are listed) and cover eighteen families or tribes. The author of this list is unknown 23 . Fortunately, the collection numbers of St. Hilaire are entered according to the Weddellian system. The presence of geographical data increases its value considerably (cf. pi. 19). In the list the entries from numbers 1 to 364 are of the Graminee only, divided according to 1 3 tribes. Among the Dicotyledoneae we find the following families and tribes listed in the order given (entry numbers 1-1589): Chloranthaceae (entries 1-2) Piperaceae (entries 3-44) Celtideae (entries 45-53) Morcac (entries 54-85) Artocarpiees (entries 85 bl8 -95) Urticacees (entries 96-100) Salicineae (entries 101-103) Lacistemeae (entries 103 bU -103 5 ) Polygoneae (entries 104-149) Nyctagineae Chenopodeae Amarantaceae Aristolochieae Proteaceae Monimieae Laurinees Compositae Campanulacees (entries (entries (entries (entries (entries (entries (entries (entries (entries 150-182 bU ) 183-192) 193-307) 308-322 bls ) 323-328) 329-3 37) 338-390 bU ) 391-1572) 1573-1589) It is obvious that such a list may be of great value to specialists of the above groups working on material from southeastern Brazil As an example of its use- fulness, I was able to ascertain in the model analysis (cf. p. 157) that Elionurus is the genus of the grass family described by St. Hilaire. SUMMARY Weddell the collection numbers of St. Hilaire's material with the specific catalogue series and section, probably due to the absence of these numbers from the majority of the labels of the specimens. The The Weddellian system answers the problem of the possible overlapping of collection numbers, and monographers should be encouraged to adopt it. Being aware of the existence of the catalogues of St. Hilaire, they are now able to check collection numbers against St. Hilaire's "analyses" 24 . The cata- logues themselves represent an original source of reference for the 8900 collections of Aug. St. Hilaire in southeastern Brazil. The discovery of a list (author unknown) of some of the genera collected by system, should be of great assistance to monographers. Wedd There are 17 collection numbers of St. Hilaire and accompanying data at the end of the list which are obviously in Spach's handwriting. This permits one to give the date of the list as some time prior to 1879, the year of Spach's death. Mr. George Swett, from a comparison of one sheet of the above list (pi. 19, a photo from microfilm) with photographs of the writings of St. Hilaire and Weddell (cf. pis. 17 and 18) expressed the opinion neither prepared the list; he qualifies this, however, by stating that "no opinion is held as to whether the writers (St. Hilaire and Weddell) executed the word Andropogoneae in the above sample sheet." The author, through the Missouri Botanical Garden, will be happy to supply sets of collection numbers of St. Hilaire for specific families to specialists. 1955] DWYER BOTANICAL CATALOGUES OF ST. HILAIRE 167 APPENDIX Through a letter received during the proofing of this paper I learned that the Bibliotheque Centrale (Mus. Nat. Hist. Naturelle) has no manuscripts of Aug. de St. Hilaire except several fragments of the Flore du Brisil Meridionale and four of his letters. Apparently there are no St. Hilaire manuscripts in the Library of the Division de Phanerogamie of the Museum except the catalogues herein described, and a single letter. In addition, I learned that the editor of St. Hilaire's Voyage a Rio Grande do Stil is Charles Roland de Dreuzy (1837-1899), a native of Nivernais. He was the grand-nephew of St. Hilaire, heir to his estate at Turpiniere, and grandson of a sister of St. Hilaire. At his own expense he published the Voyage a Rio Grande do Sul, in accordance with the last wishes of his "parent". The chateau at Tur- piniere is located near Ferte Saint-Aubin, Sennely (Loiret). The present owner is M. Robert de Dreuzy, son of the editor of Voyage a Rio Grande do Sul. The grass described in the model "analysis" (cf. p. 157 of text) was determined by Miss Agnes Chase as Elionurm candidus var. bisetosus Hack, and Linden. 168 [Vol. 42, 19551 MISSOURI Explanation of Plate PLATE 17 A portion of one page of Catalogue C 2 of Auguste de St. Hilaire. This is in his hand- writing and covers numbers "2791" Hypoxis, 279 1 10 , Labiee, 279 1 11 , Solanee, 2791", Oranger", and "Suite a 2791°." The last is a continuation of the "analysis" of Hypoxis at the top of the page. u < ON Q a: < • H O CQ O z «R/ ■HA* 4iJp*(^«P>pMF PP^^^" - ^ 1 & 3*4** DWYER— BOTANICAL CATALOGUES OF ST. HILAIRE Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 18 % h t* / / '-L •■ to to ( 4K MS" m i /> 14+&1 * — e\&£* u u 4* QtS fit 6 20 St II t l»- */** Cj- > i * 0- / ~j* * J- *4 i/: : / m C § *♦ I '/t/j m. j-r - ^;< t "** rtV ^ 4|K*v- «Wr i^ **-» .*» ( ' DWYER— BOTANICAL CATALOGUES OF ST. HILAIRE [Vol. 42, 1955] DWYER BOTANICAL CATALOGUES OF ST. HILAIRE 169 Explanation of Plate PLATE 18 A portion of the "key" to the botanical catalogues of Auguste de St. Hilaire. This is in Weddell's handwriting and is from Series C, catalogue C 1 . The handwriting at the top and bottom of the page is that of the librarian. 170 [Vol. 42, 1955] MISSOURI Explanation of Plate PLATE 19 A page from a list of approx'mately 2000 collection numbers of Auguste de St. Hilaire. Note that the Grasses are entered according to tribes and genera. The fourth collection entered under the genus Elionurus (C 2 No. 2682) is that upon which the "model analysis" (see p. 157) is based. Thelist is deposited in the Herbarium of the Museum National d'Histoire Naturelle, Paris. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 19 m . $$& ■ ■ m.3si JJt 'jf'fterm* tta/***t // >£? m m$Wt ffijom* "ut^iwcj */ *r*/ «£ <« "t t Yi'\ if iRf . ••mttou Jm V ' mtm/«u** <^/«m> it* t/J4 4 /r ' ft t /ft;*£'2*« y ■ ■ /'/L^ <3. 4' »'**4 < m- ft 4* t>$t MI0UV«MU44l Ayl/ i*if**J ► ft?c C* »• f*3t DWYER— BOTANICAL CATALOGUES OF ST. HII AIRE COMPARISON OF JUNIPER POPULATIONS ON AN OZARK GLADE AND OLD FIELDS MARION TRUFANT HALL* INTRODUCTION Using comparative morphology as criteria, I presented evidence in an earlier paper (Hall, 1952) that Juniperus virginiana L. has differentiated into five geo- graphic races, and two of these races, called Typica and Ozark, were described in detail. Typica is known to the horticulturist as the Eastern form and may be characterized as the very tall, narrowly pyramidal tree found from Virginia to southern Illinois. Within its present distribution Typica is concordantly variable, and at least so homogeneous that it seems to be carrying no introgressant genes from other species. The influence of Typica in the juniper populations of the Ozark Plateau is not evident even though isolated plantings of this race are to be found on the western banks of the Mississippi River from St. Louis southward, usually about old farmsteads. The only other race of red cedar which occurs occasionally in the St. Louis area is the Northern, which is characterized by high frequencies of the spire-like form known as var. crebra> by crooked fruiting peduncles, and a preference for well-drained, sandy soils with a slightly acid reac- tion. I postulated that this Northern race may be the consequence of a few genes from /. horizontalh becoming stabilized in red cedar. The Northern race, with its tendency toward suppressed lateral branching, follows the glacial drift down the Mississippi River, and where the periglacial loess occurs these junipers are abundant. South of the loess deposits, particularly in the vicinity of St. Louis, juniper popula- tions closely resembling the Northern race occur sporadically. The Ozark race is indeed the most abundant and important in the St. Louis area and represents red cedar throughout the Southwest. The Ozark race, which is the result of introgression by genes of /. Ashei Buchholz into /. virginiana L., includes a highly variable assemblage of junipers distributed from central Texas across eastern Oklahoma and the Boston Mountains of Arkansas through the Mis- souri Ozarks to the Mississippi and Missouri rivers. Introgression by genes of /. Ashei into red cedar has probably occurred repeatedly, at least during the Pleisto- cene, and current mixing is evident in restricted areas where extensive hybrid swarms exist. The purpose of this investigation was to determine the differences between the habitats in relation to the differences in the variability within and between popula- tions of these Northern and Ozark races of junipers. The habitats studied are close together, less than a mile apart, and are surrounded by many hundreds of popula- tions of junipers both large and small. It was thought that these habitats may select from this large gene pool adaptive genotypes which may vary from popula- tion to population proportionately to the habitat differences. * Cranbrook Institute of Science, Bloomfield Hills, Michigan. 171 172 [Vol. 42 MISSOURI ACKNOWLEDGMENTS In the spring of 1952 I was granted a leave from the Cranbrook Institute of Science to be a guest lecturer in Genetics and Evolution at the Henry Shaw School of Botany, Washington University, St. Louis, Missouri, and to take over Dr. Edgar Anderson's class during his absence. This study was part of a project undertaken by Dr. Anderson's class under my direction. I am most grateful to Dean Henry N. Andrews for his advice and cooperation and to Dr. Anderson for his enthusiasm in regard to field studies as well as his helpful criticism. The study was carried out in the native tracts of the Missouri Botanical Garden Arboretum, Gray Summit, Missouri, which is approximately forty miles southwest of St. Louis on US Highway 66. I am especially grateful to Mr. August P. Beilmann, Superintendent of the Arboretum, for his cooperation and detailed knowledge of land use in the Ozark region. GENERAL DESCRIPTION OF THE AREAS The areas studied are within the Missouri Botanical Garden Arboretum, which is situated 40 miles southwest of St. Louis, Missouri, just south of the junction of Highway 66 and Highway 50 near Gray Summit, a hilly region principally of weathered dolomitic limestones in the northern portion of the Ozark highland. The climate is designated by Thorn thwaite as BB^ (humid, mesothermal, with precipitation adequate at all seasons). Records in St. Louis for the past hundred years show such variations in precipitation that no general trends are evident, so that it is difficult to judge whether climatic amelioration or deterioration is in process. The Ozark Plateau is a part of the Oak-Hickory Forest Region described by E. Lucy Braun (1950) as the Interior Highlands, Southern Division, which lies essentially south of the glacial border. Braun wrote: "The Southern Division is characterized by the prominence of southern species of oaks and hickories, as Quercus stellata, Q. marilandica, Q. Shumardu (and var. Schneckii), and Carya Buckleyi (var. arkansana and var. villosa) , and in the bottomlands, Q. nigra, Q. lyrata, Carya Pecan, C. myristicaeformis, and C. aquatica . . . . Prairie openings, limestone glades, and balds locally interrupt the forest cover." Those who know the Ozark landscape consider these local interruptions of forest cover as the particular mark of beauty of the uplands, and the "openness" was a major feature in selection for settlement in pioneer days. Visitors to the Ozarks from foreign countries usually are most impressed by the red cedar glades which are island playgrounds in the present-day dense oak woods. In the proper season these open areas serve as edge for much wildlife activity, and perhaps the summer booming target The soils indicate a long history of open conditions where prairie vegetation, probably maintained by fire and grazing, played a very important role in the development of soil horizons. Soils in the area are characteristic of prairie-forest 1955] HALL OZARK JUNIPER POPULATIONS 173 transition regions and are much like the nut-structured prairie-forest Grood soils of continental sub-humid steppes, being intermediate between chernozems and podsols. Regardless of its history of prairie development, the Ozark region has a forest climate and probably has had since the end of the xerothermic period. Thus the climax plant formation, which is of course subject to continual change, may be the oak-hickory forest which has been spreading extensively throughout Missouri roughly in the last hundred years. Beilmann and Brenner (1951) and Etter (1953) report this trend in their studies of Ozark vegetation. It is not likely that any region has ever had a vegetational mantle entirely in equilibrium with climate alone, for other factors affecting balance are legion — e. g. nature of geological substrate, grazing and browsing, fire, and man. Rather, as Sauer (1950) has pointed out, "Plant associations are contemporary expressions of historical events and processes, involving changes in environment and biota over a large span of geologic time." However, Braun (1950) stated that knowledge of these historical events and processes is very scant even though general trends seem to be quite clear. The history of land usage in the Gray Summit area is not known in detail, but titles to land were granted beginning about 1850, nearly 50 years after the first settlers, mostly French, began to locate along the major waterways in what is now Franklin County, and over 80 years after St. Louis was founded, also by the French. Evidence of timber cutting, heavy grazing and burning variously from the time of settlement until the Missouri Botanical Garden acquired the property in 1925 is clear and common. Apparently, in the days before the settlement of the region most of the timber was found on the headwaters of the creeks and was scattered, giving a prairie- or park-like aspect presumably as a result of a somewhat drier climate than today and the effects of grazing, fire, etc. The timber, where dense enough to be called woodland, was usually situated in the rough lands or dissected headwaters of the streams where it was largely safe from fire. Today the Ozarks are well covered with close forests, and the once open areas show signs of rapid forest encroachment; old fields develop rapidly to forest, but the glades show little invasion by woody species even when protected from fire or grazing. Steyermark (1940) made a vegetational survey of Missouri which cov- ered fifteen years of extensive observation, and even though the survey was strictly qualitative it was detailed enough to permit rough comparisons with present-day conditions. He stressed the role of rock strata and soil, i. e., edaphic factors, in determining the distribution of plant associations. He described six plant associa- tions and many more associes based on differences in soil, slope, and water avail- ability; in his conclusions he favors the polyclimax concept subscribed to by Domin, Gams, Gleason, Tansley, and Du Rietz. He indicated that the glade flora was the result of edaphic factors, and in most sites would eventually support a maple- white oak association which in the Ozarks seems to be restricted to slopes and bluffs with neutral to alkaline soils underlain by limestone. The Glade. — The Glade, which is one of several within the boundaries of the Missouri Botanical Garden Arboretum, may be characterized as a small open space 174 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 (open glade), approximately 60,000 square feet in extent, covered by grasses and junipers surrounded by a closed area (closed glade), or buffer, approximately fifty feet in width, chiefly composed of junipers and chinquapin oaks and situated on a west-facing slope (upward ENE, 10°20'). The area studied constituted a rectangle 400 X 150 feet, approximately the total extent of the open glade. In this study the term Glade refers to the open glade within the enclosure studied. This same glade was described by Erickson, Brenner, and Wraight (1942), who demonstrated a direct correlation between the structure and composition of the glade flora and the occurrence of a dolomite stratum at or near the surface. The glade soil is very shallow (averaging about 4 inches), organically rich and derived from a dolomitic rock which under the open glade is a thin, porous dolomite locally known as "cotton-rock/' an upper stratum of the Cotter Formation, and lies un- conformably on the Jefferson City Formation, both in the Canadian Series of the lower Ordovician. Under the closed glade a less porous dolomite with lower magnesium content is covered by a fairly thin mantle of Union Silt Loam. The dense oak-hickory upland forest on the Union Silt Loam gives way to red cedar- chinquapin oak forest at the periphery of the glade, but wherever the "cotton- rock" occurs it drains the area sufficiently that the red cedar-chinquapin oak forest is fairly well stabilized. In winter and spring, as a consequence of greater precip- itation and less evaporation, the open glade ("cotton-rock" stratum) is wet, often with standing water, while in summer and fall, when precipitation effectiveness is least, the porosity of the rock permits rapid drainage which results in rather extreme dryness in the upper few feet of the stratum. I have been frequently observing several of these glades since 1947, and, besides the rather obvious seasonal cycle of wetness and dryness, there are fluctuations every few years which markedly affect the population density of the species and the cover. In order for the peren- nials, in particular, to be successful in the glade community, they must be able to withstand fairly great fluctuations of the environmental conditions. The Old Field. — The Old Field is situated on a gently sloping knoll below the orchid greenhouse. The upward slope (5° 10') is NNW facing SSE. The knoll is covered by Union Silt Loam and probably will eventually support an oak-hickory forest. On this knoll a plot 60,000 square feet in extent was laid out as a rectangle 300 X 200 feet. Prior to 1925 the area was under cultivation in the traditional "Ozark fashion." A few acres of beans, peas, and corn were planted, rows running uphill, and tilled by means of the mouldboard plow and mule driven by a tenant farmer. In 1925, when the property was acquired by the Arboretum, this field was badly eroded and almost bare of vegetation, but since 1925 it has not been grazed or burned, and the revegetation has been accomplished in the last 28 years. At present, the arborescent growth on the Old Field consists of scattered juni- pers, black oak, white ash, and thickets of red elm. Wisteria is an abundant shrub. 1955] HALL OZARK JUNIPER POPULATIONS 175 having invaded the field from plantings nearby. The upper edge of the sloping field supports a growth of sugar maple, white ash, black oak, shingle oak, post oak, black cherry, smooth sumac, buckbrush, and poison ivy. The middle portion of the field consists of thickets or clumps of red elm saplings with a maximum height of about ten feet, and a thick growth of smooth sumac and wisteria. On the lower slope which drains into a tiny creek are growing black oak, white ash, northern ood summer The oldest arborescent growth is made up of persimmon, post oak, black oak, and shingle oak, which apparently invaded the field approximately five years after its abandonment. Then northern red oak, sugar maple, white ash and red cedar came in six to eight years after abandonment. At any rate, the older specimens of red cedar were twenty years old, eight years having elapsed since abandonment of the field. The rate of invasion by red cedar depends on the balance of several variables such as the number and proximity of seed-producing plants, condition of the invaded area, amount of cover, slope, pH, etc. Drew (1942), in his studies of abandoned crop-land in the Cedar Creek area, Boone and Callaway counties, Mis- souri, found that five years after abandonment sassafras, persimmon, shagbark hickory, post oak, and shingle oak invaded; after six to seven years, American elm; after eight to ten, northern red oak, sugar maple, white ash, honey locust, and black walnut; after eleven to twelve years, red cedar and black oak. In the herb layer the general cover consisted of fairly evenly distributed clumps of broom-sedge with Kentucky blue-grass in swales and level places bordered by clumps of little blue-stem. On the up-slope, broom-sedge, tall red-top, Canada goldenrod, and old field goldenrod were conspicuous. From midway to the bottom of the slope were these same species plus dewberry (Rubus flagellaris) and Aster pilosus. Here and there were almost bare sandy slopes, a few feet in extent, con- sisting of stands of triple awn-grass with a few lichens and mosses. Directly beneath the crowns of most of the larger trees were areas covered by lichens and mosses. Cedar Hill. — Another old field, here called Cedar Hill, is an exclosure with a gentle slope from west to east mowed annually with a sickle-bar and covered with well-spaced red cedar, black oak, black cherry, honey locust, and red elm. Th shrub layer consists of dewberry (Rubus flagellaris), smooth sumac, trumpet-vine, buckthorn, aromatic sumac, large clumps of buckbrush, partridge pea, and seedlings of sassafras, shingle oak, and persimmon. The herb layer consists principally of Canada blue-grass, triple awn-grass, tall red-top, Panicum capillare, and frequent clumps of broom-sedge, scattered clumps of little blue-stem, and occasional clumps of mountain mint and ironweed. Dewberry (Rubus flagellaris) was scattered fairly evenly throughout the field. The junipers average about 30 feet in height and thirty years of age, somewhat younger reprod [Vol. 42 176 ANNALS OP THE MISSOURI BOTANICAL GARDEN negligible. Although there were abundant seedlings, particularly at the edge of the crowns of those mature junipers bearing berry-cones, all were below sickle-bar height and usually under three years of age. Just outside the exclosure and in the general area juniper reproduction was very good, and ages up to thirty-five years were represented. Species composition was not analyzed because of the artificial situation. In the meadow the juniper population in an area of 60,000 square feet was sampled. In general features this meadow is closely similar to the Old Field. The soil is Union Silt Loam but much deeper since Cedar Hill lies above an old river channel. The Cedar Hill population was included in this study because of its intermediate position morphologically between the Glade junipers and those of the Old Field. METHODS OF STUDY Several factors were studied for the Glade and Old Field in order to make com- parisons. Species composition, distribution of juniper seedlings, and the population structure of junipers were studied, and the data appear in Tables I— II, figs. 1—3, and pis. 20—21. Frequency was used as a quick method of measuring species com- position. The Glade and the Old Field were divided into four plots by staking out the diagonals. Along the diagonals alternating from one side to the other, 20 quadrats (1 meter in size) were laid out, equally spaced from the center. Prelim- inary work was done in late spring of 1952, but other surveys were made in late spring and late summer of 1953. The presence of species in each quadrat was tabulated and the results for the Glade and Old Field appear in Table I. The areas studied seemed sufficiently homogeneous for the use of 1 -meter quadrat size, specif- ically to determine the species composition of the herb layer. A sample area about 0.5 per cent of the total was used. A compromise in favor of a small sample area was made because of the limited time the class could be in the field. Variations of the juniper populations were measured by pictorialized scatter diagrams as demonstrated in Juniperus in a previous paper (Hall, 1952). Five characters were scored: gland length- width ratio; length of whip (longshoot) leaf; length of terminal whip; length of lateral whip; on secondary shoots, the per cent of tertiary branches which have the decussate leaf arrangement. Two of these characters were also analyzed by use of the parameters of the log-normal curve which the data fit. These data, when plotted on logarithmic probability paper, result in a straight line. In an attempt to get some estimate of seed-dispersal pattern, the distribution of seedlings was studied. Observations were made to determine the role that birds, mammals, and gravity may play in seed dispersal. SPECIES COMPOSITION Rough comparisons were made of species composition of vascular plants for the Glade and the Old Field. Only one quadrat size was used and no attempt was made t o obtain quantitative data on density or cover. For convenience a frequence index after the method of Raunkiaer (1918) was made up for the Glade and the 1955] HALL OZARK JUNIPER POPULATIONS 177 Old Field, The percentage of frequence for a given species is the percentage ratio the plots on which the species occur bear to the whole number of plots sampled. Raunkiaer found that the greatest number of species had the least frequence; and that as the frequence increased the number declined steadily, until at the highest frequence the number increased slightly. Raunkiaer used five frequence groups, A, B, C, D, and E, designating the species of frequence 1-20 per cent, 21-40 per cent, 41-60 per cent, 61-80 per cent, 81-100 per cent, respectively. His law of dis- tribution, determined from quadrat studies in Europe was A>B>C — DB>C = D : =E. This index is similar to that found by Kenoyer in his studies for America and indicates a fairly high proportion of sporadic species and a very few acting in a "dominant" role. The Glade has uniformly distributed patches of Sporobolus neglectus, and the size of each patch fluctuates somewhat from year to year. When the patches are smaller, more Bouteloua curtipenJula and forbs are in evidence. In good years the Glade has a good cover contributed principally by a very small number of species even though there is an appreciable number of sporadic species between the clumps of the cover plants. The Old Field index was 22, 3, 8, 5, or A>BD>E. This somewhat irregular result is a product of the heterogeneous nature of the Old Field flora. There are a large number of sporadic species, a fairly large number of species inter- mediate in frequence, and no "dominants." The floral composition and vegeta- tional structure are markedly different from those of the Glade. Factors of competition and selection are less delicately balanced in the Old Field, and these factors undoubtedly cause the population structure of the junipers there to be different from that of the Glade. Table I lists the species and their frequencies found in sampling 0.5 per cent of the area studied. Only the under-story is represented. Woody plants in the list were seedlings or sprouts in size class 1 (0-0.9 feet) of Weaver and Clements (1938). The Old Field had seven species of woody plants represented in this size class for the area surveyed, which compares with nine species of similar size found by Drew (1942) to be common in old fields twenty-five years after abandonment. In the Old Field all sizes of woody plants were abundant up to size class 4 (3.6-9.5 inches D.B.H.), but on the Glade very few specimens larger than size class 1 were seen, suggesting a high seedling mortality. The frequency data in Table I show that the Glade and Old Field are quite different in species composition and suggest a difference in the pattern of succession. The Glade is a "prairie" association with Andropogon scoparius and Rudbeckia missouriensis contributing most to its aspect and with Andropogon scoparius and Sporobolus neglectus contributing most to the cover. Even though there is some fluctuation in abundance from year to year, there are definite dominants, a good variety to the flora, and a stability in its over- all composition which suggests that succession is very slow and that the Glade [Vol. 42 178 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE I SPECIES LIST WITH FREQUENCIES FOR UNDER STORY OF GLADE AND OLD FIELD* Glade Frequencies 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 3 6. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. Sporobolus neglectus Car ex Crawei Andropogon scoparius Houstonia langifolia 48. 49. 50. 51. 52. 53. 54. 55. Euphorbia corollata Rudbeckia missouriensis Sisyrinchium cam pest re Hypericum prolificum Gaillardia pule hell a Echinacea pallida Kuellia humilis Oenothera missouriensis Heliotropium tenellum Viola pedata Selaginella rupestris Juniper us virgin tan a Viola papilionacea Coreopsis lanceolata var. villosa JJlmus rubra Mosses Eleocharis camp res sa Aster oblongifolius Zizea aurea Monarda citriodora Cercis canadensis Rhamnus caroliniana var. Comandra Richardsiana Lotus americanus Orobanche uni flora Acalypha graciliens Panicum virgatum Triodia flava Lichens Bumelia lanuginosa var. oblongifolia Fraxinus am eric ana Monarda Russeliana Acer sac char um Anemonella thalictroides Bouteloua curtipendula Silphium laciniatum Psoralea esculenta Quercus Muehlenbergii m ollis 1.00 .90 .80 .75 .70 .70 .65 .60 .55 .55 .50 .50 .45 .45 .45 .45 .40 .40 .40 .3 5 .35 .35 .30 .25 .25 .25 .20 .20 .15 .15 .15 .15 .10 .10 .10 .10 .10 .10 .10 .10 .10 .10 .10 .10 .05 .05 .05 .05 .05 .05 .05 .05 .05 .05 .05 Old Field Poa pratensis Rubus flagellar is Andropogon virginicus Aster pilosus Desmodium glabellum Rumex hastatulus Oxalis stricta Panicum virgatum Lichens Erigeron strigosus Taraxacum officinale Hieracium Gronovii Campsis radicans Triodia flava Achillea millefolium Parthenocissus quinque folia Mosses Ulmus rubra Celtis occidentals var. pumila Hypericum punctatum Prunus serotina Cercis canadensis Rhus aromatic a Juniperus virginiana Asplenium pi aty neuron Ambrosia artemisifolia Allium vineale Solanum carolinense Carex Haydenii Eragrostis spectabilis var. sparsihirsuta Veronica arvensis Asclepias syriaca Acer saccharum Phy satis pruinosa Pentstemon pallidus Sporobolus neglectus Daucus pusillus Antennaria fallax Potentilla recta Galium aparine Poa campressa 1955] HALL OZARK JUNIPER POPULATIONS 179 Glade Frequencies Old Field 56. Pellaea atropurpurea 57. Vimbristylis Drummondii 5 8. Petalostemum purpuretim 59. Symphoricarpos orbiculatus 60. Pentstemon pallidus .05 .05 .05 .05 .05 Rhus Toxicodendron Nomenclature after Gray's Manual of Botany, 8th edition, 1950. community is one of real integrity. Because of the uniformity in distribution of Andropogon, Sporobolus, and Bouteloua the quality of the cover on the Glade can be determined fairly well with frequency data. In a good year the clumps of Andropogon are larger, denser, and more frequent. The Sporobolus effectively fills in between the clumps. In a dry or unseasonable year the Andropogon may lose ground, the Sporobolus to a lesser extent, but the Bouteloua may increase con- siderably. Thus, one can tell when the Glade is in good condition, so far as climatic fluctuations are concerned, by the relative frequencies (or preferably cover) of Andropogon scoparius and Bouteloua curtipendula. I think the Glade's general aspect and its severe selection against trees, shrubs, and woodland herbs, particularly as evidenced by selection against all but very Ashei-like junipers, support the con- clusions of Erickson, Brenner, and Wraight (1942) that these glades are edaphic sub-climaxes. The frequency data show the Old Field to be fairly typical for the Ozarks. There are several abundant species, but none of these contribute decisively to the field's aspect or cover. Poa pratensis, Andropogon virginicus, Panicum virgatum, Triodia flava, and Rubus flagellars are conspicuous. While the number of species is much the same for the two areas, the Old Field has fewer species with high frequencies or very low frequencies, and more with average frequencies. The Old Field is obviously in a rapid stage of succession toward woodland. Several other species were not found in the quadrats sampled. Mostly, these are not common plants in these areas but in some cases may occur with frequencies of 5 or 1 per cent in localized areas of the Glade or Old Field. The following species occurred on the Glade but did not occur in the quadrats: Ulmus rubra, Celtis occidentalism Physocarpus opulifolius, Amelanchier arborea, Rosa Carolina, Prunus serotina, Rhus glabra, R. aromatica, Ilex decidua, Celastrus scandens, Acer saccharum, Cornus Drummondi, Diospyros virginiana var. pubescens, Fraxinus quadrangulata, Viburnum rufidulum, Sporobolus heterolepis, Elymus virginicus forma hirsutiglumis, Allium stellatum, Menispermum canadense, Psoralea psora- lioides var. eglandulosa, Linum sulcatum, Petalostemum purpureum, Croton monanthogynus, C. capitatus, Asclepias tuberosa ssp. interior, Triosteum perfoli- atum, Vernonia crinita, Kuhnia eupatorioides var. corymbulosa, Hypericum per- foratum, Amsonia illustris, Isanthus brachiatus, Ruellia humilis, Solidago rugosa var. aspera, S. rugosa var. celtidifolia, S. nemoralis, and Parthenium hispidum. A [Vol. 42 180 ANNALS OF THE MISSOURI BOTANICAL GARDEN total of 80 species of vascular plants was collected on the Glade (45 from the quadrats; 35 by systematic search). Eighteen other species, mostly seedlings of the woody species already mentioned, were found in the Old Field but not in the quadrats. According to experience and estimates, it seems likely that slightly over 100 species may be found in similar areas the size of the Glade and Old Field. Since these studies did not run continuously through the seasons involved, a considerable number of species was probably missed. It is also obvious from the data that a sampling area 0.5 per cent of the total is too small for best results; an area between 1 and 2 per cent would have been effective. >sed glade is a tension zone between the open glade and the oak-hickory woodland. oodland most of the cover in the arborescent layer is from juniper and chestnut oak. The transition from one community to the other seems abrupt since it occurs over a short distance, usually from 50 to 100 feet. From transect studies in the closed glade I listed 75 species, 39 of which were frequent species in the oak-hickory woodland and 43 were frequent in the open glade (Glade). Plate 20 shows the aspect of the Glade with close-ups of two quadrats (1 meter in size) on the up-slope and two on the down-slope. Plate 21 shows the Old Field aspect with two quadrats up-slope and two down-slope. DISTRIBUTION OF JUNIPER SEEDLINGS The diagonals and sides of the Old Field and the Glade were used as transects for seedling counts. All juniper seedlings were counted within one-half meter on either side of the diagonals and one meter to the inside of the area boundaries. Seedlings were divided into three classes as follows: Class 1, 18 inches or less; Class 2, 18.1-36 inches; Class 3, 36.1-72 inches. Total number of seedlings re- ported is based on counts in approximately one twentieth of the total area. In the Old Field, a total of 621 seedlings was estimated, of which 276 were in Class 1, 207 in Class 2, 138 in Class 3. The actual number is probably somewhat higher since the smaller seedlings are rather easily overlooked. The seedlings were not distributed at random, but were either clumped about the mature (fruiting) female trees (with over 90 per cent of them on the down-slope side) , or were more or less aggregated at the bottom of the slopes or in other areas where materials carried by run-off water are deposited. Birds play an important role in the distri- bution of seedlings in the Old Field. Many young seedlings were clustered below the branches of shrubs and trees other than junipers. A very few were found on isolated knolls and in the open where they may have been deposited by birds in flight or by ground-feeding birds. The Glade supported a great many more seedlings under four years of age than did the Old Field. The estimate was: Class 1, 6,672; Class 2, 10; Class 3, 12. The total estimate was 6,764 seedlings for the 60,000 square feet of glade exclosure. These were distributed much as in the Old Field, but even more uniformly because of the greater uniformity of the 10° slope characterizing the Glade. The seedlings 1955] HALL OZARK JUNIPER POPULATIONS 181 were either clumped on the down-slope side of the mature (fruiting) female trees or were aggregated at the bottom of the Glade slope where the Sporobolus neglectus, because of its increasing cover toward the lower edge, acted as a brake to tumbling berry-cones carried by run-off water. There were numerous seedlings in the open, probably from seeds deposited by birds. Great flocks of robins and cedar waxwings are a common sight on the glades, particularly in the fall after the crop of berry- cones has ripened. Fortuitous fluctuations in climate probably add to the seedling mortality. In very wet seasons the seedlings with more genes of red cedar are probably favored only to be wiped out in a later year by drought. Likewise, seedlings possessing a strong /. Ashei component are favored in drought seasons but may suffer defeat in wetter periods. POPULATION STRUCTURE OF JUNIPERUS The variations in the junipers (Jtiniperus virginiana) of the Glade, Old Field, and Cedar Hill were studied, and five morphological characters were measured and plotted as pictorialized scatter diagrams. Two of the five characters, measured, gland length-width ratio and lateral whip length, were used on the ordinate and abscissa of logarithmic paper (1X1 cycle) and the other three as rays on the scatter diagrams. The data for the three characters plotted as rays were grouped as follows: (1) whip-leaf length: long-bar = 8 mm. or more, half-bar = 5-7 mm., no bar = 2-4 mm.; (2) per cent decussate: long-bar = 25-100 per cent, half-bar = 6-24 per cent, no bar = 0-5 per cent; (3) terminal whip length: long-bar =80 mm. or more, half -bar =± 31-79 mm., no bar = 0-30 mm. The scatter diagrams for the Glade, Cedar Hill, and the Old Field populations, as well as for populations typical for the species, are shown in fig. 1. Population means and ranges were obtained from the grouped data (pictorialized scatter diagrams) in the manner described in an earlier paper (Hall, 1952, p. 53). The numbers given have no absolute value but represent a graded scale or index of characteristics. The scale is made up in such a way that low scores belong to /. Ashei, high scores to J. virginiana, and intermediate scores to morphological intermediates or introgressants of the two species. From these scatter diagrams the following population means on this arbitrary scale were obtained: Glade — 4.35; Cedar Hill — 5.48; Old Field — 10.0. Typical J. Ashei has a population mean of 2.0-2.3; typical /. virginiana, Eastern race, has a mean of 9.0-10.0. The two characters plotted on the abscissa and the ordinate for the pictorialized scatter diagrams were also analyzed by using ordered values. The data were ex- pected to fit a log-normal distribution which was confirmed, for the data produced a straight-line function when the ordered values (from smallest to largest) were plotted on logarithmic probability paper. I followed the practical short-cut recom- mended by George W. Thomson of the Ethyl Corporation, which used the range as a measure of dispersion instead of the standard deviation. Statistical techniques using the range are particularly efficient with small samples. This is an excellent method of checking the confidence of data where only an average and the range are available. The confidence limits a, b = x ± ts/v n > 182 [Vol. 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN LIMITS OF VARIATION IN TYPICAL RACES OF THE SPECIES OLD FIELD POPULATION OF Juniptrut vlrqlnlono L. 7 3 Juniptrut virgin! • Juniptrut Athti ! I J I i 1 i L w •oo tOO JOO 400 500 900 7C0 #00 *X> LATERAL WHIP LENGTH # WHLP.LCAT UNtTH # nilMrtUk. *HlP LCNCTM # %OCCUSiATC LfATe* too JOO *oo UTCAAL WHI* LENGTH 9O0 100 100 IO0 CEOAR HILL POPULATION OF Juniperus Virginia na L t • 7 5 A Id o z 5 *o WO # WHIP-LCAF LCMTN TfAMtHAL WMtf UMTN \| % DCCUSMTC LCAVCS tOO JOO 400 LATERAL WHIP LENGTH 300 «00 TOO #00 900 OLAOE POPULATION OF Juniptrut virginiano L. )t i )t V i 4U ¥ V V ¥ i i r i , »0 100 # WHtP-k.C«P LCMCTM • %OtCU»tAT| UAVC* JOO JCO UTERAt WHIP UNfiTM Fig. 1. Scatter diagrams showing the variations in five characters of a typical race of Juniperus virginiana from southwestern Virginia, a typical race of /. Ashes from the Edwards Plateau of Texas, and of three populations of /. virginiana at Gray Summit, Mo. The diagram at the upper left for the typical races shows the means (large dots), the extremes (small dots), the limits of variation (the curved lines). (The symbol in the space at the upper right of this diagram should have had three long arms.) In the diagrams for the three Gray Summit populations each dot rep- resents a single individual. 1955] HALL OZARK JUNIPER POPULATIONS 183 root lation variance, and n is the number in the sample. Or, by using the short-cut the 95 per cent confidence limits are equal to the range multiplied by the proper value 1 (0.1064 for n = 25, 0.0720 for n minus imi (log values) for two characters of each of the three populations and typical popu- /. Ashei and / Figure 2 represents the ordered values plotted on logarithmic probability paper. The slopes of the lines depend on the amount of variation within each group; the greater slopes indicate the more heterogeneous populations. The point of intersection of each line with the 50 per timated points correspon of the 25 per cent and 75 per cent values with each line delimit the range where half the values for that character will probably lie. In fig. 3 rectangles are made about the means to correspond to the 5 per cent probability level, and the farther two rectangles are separated the greater the probability that the difference between ling means is not due to chance. The dotted lines indicate the 95 per cent confidence limits of the means from Table II. These five populations were compared to determine the probability that they represent the same population with equal means by means of the t-test. If one assumes the means of two populations to be equal, the sample difference is tested to see whether it is no more than sampling variation from the hypothetical dif- ference, zero. Each of these comparisons led to a large t value, indicating a low probability that they represent samples from a single population. Welch's modifi- cation of the t-test was used for the comparison because the variances were not homogeneous (Comparison of /»i— /i 2 regardless of 13 C. Glumes well developed D. Glumes long (average 11.5 mm.) Tu — 14 D. Glumes near normal length (average 8-10 mm.) E. No pollen shed „ m5l — 10 E. Pollen shed F. Up to l / 4 or y$ of branch tips sterile; pedicellate spikelets often with several sets of florets bd — 11 F. Branch tips never sterile; pedicellate spikelets not branched Club — 12 A. Tassel of abnormal proportions B. Tassel with silks C. Half or more of tassels pistillate; staminate spikelets confined to middle areas or outer halves of branches; spikelets with no more than two florets each D. Branches indeterminate, mostly ending in immature pistillate structures Ts 3 — 3 D. Branches otherwise E. Spikelets loosely spaced on branches f S2 — 2 E. Spikelets crowded on branches ts± — 1 C. Tassels either staminate, pistillate, or mixed; if mixed, staminate spikelets often with more than two florets o- 1955] NICKERSON & DALE modifications in Zfti May 197 D. Extra florets developed on both sessile and pedicellate spikelets; average peduncle length 10 cm. or less Tst — 7 D. Extra florets developed on pedicellate spikelets only; average peduncle length more than 10 cm. E. Seed set very sparsely if at all; average glume length 6 mm to — 4 E. Seed set in noticeable quantities; average glume length 8.5 mm to* — 5 B. Tassel without silks C. Branches horizontal or slightly lax with reference to central culm ra\ — 15 C. Branches upright, close to central culm tai — 16 Certain measurements have been made on at least fifteen and often twenty-five plants of each mutant form, and averages of these data are given in Table I. Peduncle length is the distance from the uppermost leaf-bearing node to the node at which the lowermost tassel branch arises. Tassel length refers to distance from the node at which the lowermost tassel branch arises to the tip of the central culm. Branching area is that distance along the central culm from the lowermost branch node below the central spike; it is included within the tassel length. Primary branches are axes of the second order, when the main culm of the plant is con- sidered to be an axis of the first order. The percentage values for peduncles and branching areas were obtained by dividing these respective lengths by the sum of peduncle and tassel lengths. These same data are presented graphically in fig. 1 by means of a pictorialized scatter diagram (Anderson, 1949). The diagram is meant only to indicate the average extent to which each of the mutants departs from average measurements of Standards. Limits were chosen so that Standard averages would always be char- TABLE I AVERAGE MEASUREMENTS OF CERTAIN MORPHOLOGICAL FEATURES IN TASSELS OF NORMAL AND MUTANT MAIZE FORMS (1 (2 (3 (4 (5 (7 (8 (9 (10 (11 (12 (13 (14 (15 (16 Standard Tassel-seed Tassel-seed Tassel-seed Tassel-seed Tassel-seed Tassel-seed Tassel-seed Tassel-seed Tassel-seed Form 1 (to) 2 (to) 3 (Ts 3 ) 4 (to) 4 a (T5 4 a ) 5 6 7 8 (Ts 5 ) (to) (to) (msi) Male-sterile 1 Branched-silkless (bd) Club (club) Vestigial glume (Vg) Tunicate (Tu) Ramosa 1 (rai) Ramosa 2 (ra 2 ) s u -a !? 17.1 6.3 4.9 3.9 16.2 14.4 13.0 10.0 17.9 18.0 19.0 15.1 13.9 14.9 13.7 14.6 12.0 J3 c o £ to U H.S 35.0 23.4 26.4 22.5 25.2 24.8 33.5 18.5 33.1 31.9 33.3 34.4 30.3 35.3 33.3 33.8 27.7 6C C IS c u PQ £ c v. tt 10.4 8.5 9J 9.0 19.1 16.3 12.4 12.0 15.0 8.8 10.8 10.7 10.8 11.7 12.1 28.0 21.7 S c U O Xi a u 6 £ 14 16 21 26 34 43 24 36 23 14 18 18 20 20 22 72 67 o J! v G rt 3 « ^1 33 21 16 15 39 37 28 35 35 36 36 31 31 30 29 30 30 2 c g" 5 M * 20 29 31 34 46 42 27 42 29 18 21 22 24 23 26 58 55 198 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 acterized by long flags. Thus, similarity to Standard in both graph position and flag lengths means that a particular mutant departs relatively slightly from a normal form. The converse is also true; the farther away in position and the greater the flag deviations from Standard, the more a given mutant differs in form. SO 15 3E u 10 z *. rf* TS, is 4 a V V Q Ts Q. V CLUB V T$ 20 TASSEL LENGTH IN CM 30 TS, STA HOARD 2*V Va PERCENT PEDUNCLE NO. PRIMARY BRANCHES PERCENT BRANCHING AREA BRANCHING AREA IN CM. 25 20 -25 % 10 4 26-31 21-29 26-34 11-14 32 30 35- 15- K.r ~ L t Pkt ° rialized « ct V er di3gram ° f "«»**» contained in Table I, showing relationships 19551 NICKERSON & DALE Tea Mays 199 DESCRIPTIONS i Standards, L317 and CC5 (pi. 22, fig. 1), Various mixtures of these two corn-belt inbreds were used. The history of Lancaster Surecropper, from which M. T. Jenkins developed L317, has been re- viewed by Anderson (1944b). Wisconsin Inbred CC5 has an ancestry in the open-pollinated maize which was common to that section. Anderson and Brown (1952a, 1952b) showed that these corn-belt forms were various mixtures of Northern Flint and Southern Dent races. Nickerson further pointed out (1953) that these two races had previously been mixed in some degree, and that the excellent hybrid vigor manifest in present-day crosses, such as those between these two standards, is based on rather small differences. Ears are 14- to 18 -rowed, 8 to 10 inches long, and bear bright yellow, dented kernels. 1. Tassel-seed 1 (7s J (pi. 23, fig. 1), First described as "tassel-ear" by Emerson (1920), this mutant was classified as Tassel-seed 1 by Emerson et al (1935), and assigned by them to Chromosome 2. Emerson originally described the inflorescence as being completely pistillate, com- pact and distinctly ear-like, with "glumes and palae short, broad and rounded, in all respects much like those of true ears." He noted that these characters could be observed in both mature and immature tassels as well as in poorly pollinated mature ones. It may be added that each pair of pistillate spikelets is associated with a well-developed cupule (Nickerson, 1954). The branches are thickened so that they appear as wide green bands in adaxial view. In our cultures, as in Emer- son's, seed set was good and very often both florets of a spikelet develop a caryopsis. This development obscures any evidence of regular rowing in the central spike. Emerson noted the general plant weakness in comparison to sibs, profuse silk production in the tassel, a general lack of elongation in upper internodes, and early assumption of a pendant position by the tassel. In our stock, this pendant position is due to bending of the culm; in Tassel-seed 2, likewise a pendant form, it is attributable mostly to bending of the tassel proper. These differences are mani- festations of the degree of condensation (Anderson, 1944a), at least in part. Spikelet pairs of Tassel-seed 1 are generally closely spaced, resulting in rigid branches, while those of Tassel-seed 2 are more widely spaced, resulting in lax branches. Emerson mentioned these relative spikelet-pair densities, but he did not call attention to their differing effects on the area of bending. However, his figures 10 and 11 show each form quite well. In our stocks, branch tips most commonly end in sterile zones characterized by closely overlapping glumes. Occasionally at the very tip, these glumes simulate those of male spikelets; they sometimes even contain stamens, but no pollen shedding has been observed. 200 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 The ear of Tassel-seed 1 is slow to develop, and, as in Emerson's strains, fails to set fruit unless the tassel is either removed or poorly pollinated. It is normal except that, just as in the tassel, both florets of each spikelet often develop, obscuring evidence of regular rowing. 2. Tassel-seed 2 (ts 2 ) (pi. 23, fig. 2). Emerson (1920) originally described this mutant, located on Chromosome 1 (Emerson et al y 1935), as "tassel-seed," and noted that the tassels were loose, like those of normal plants, with individual spikelets more or less separated. Our material never showed individual spikelets, but always paired spikelets. In Emer- son's specimens, spikelet pairs sometimes occurred more densely, "but not ear-like in any way." Although it rarely occurred, he noted that staminate flowers might develop with pistillate ones throughout the entire tassel, but he did not determine whether these male florets were functional. He also observed that glumes and "palae" (paleas and lemmas) of such male spikelets were long, narrow, and pointed, as in normal tassels, while in female spikelets these parts were shorter, broader, and more rounded. In our material, spikelets were either female or perfect, no male spikelets having been observed. On perfect spikelets, glumes were elongate and near normal. Also, each spikelet pair was associated with a cupule which was often elongated above spikelet pairs located on central spike tips. Kernels developed on nearly all spike- lets; in many, as with Tassel-seed 1, both florets were functional and two kernels were formed. Each tassel branch is thin, with spikelet pairs more widely spaced in its distal portion, but the base may be as thickly set as any found in Tassel-seed 1. Tassels of this mutant are pendant, but differ from those of Tassel-seed 1 plants in that the bending is mostly accomplished in the tassel rather than in the supporting culm. This point is discussed further with regard to Tassel-seed 1 above. Ears of Tassel-seed 2 develop to a degree depending on the fate of the tassel. If little or no successful pollination takes place, ears are developed rapidly, but always later than on normal sibs. If tassels are removed soon after their appear- ance, ears are formed about the same time as in normals. These results are in accord with the findings of Emerson. Like most other tassel-seeded forms, second florets in many spikelets set fruit, after the manner characteristic of Country Gentleman sweet corn (Weatherwax, 1916), so that regular rowing may be obscured. When ears develop, they are of normal size. 3. Tassel-seed 3 (Ts 3 ) (pi. 24, fig. 1). The first published mention of this dominant mutant was made by Phipps (1928), who stated that "a third type of tassel seed designated as Tassel-seed 3 bee of Emerson's. 1935) also listed Tassel-seed 3 as being an unpublished They noted that this form was similar to Tassel-seeds inflorescence was generally mixed pistillate and staminate ually pollen could be obtained. The 1955] NICKERSON & DALE Tea Mays 201 Tassel-seed 3 tassels are easily recognizable, but, like other forms, they are variable in appearance. Branches bearing distichously arranged pistillate spikelets resemble flat green ribbons on their adaxial surfaces; both branches and central spikes most often end in tapering sterile rudimentary pistillate spikelets. Often, these sterile spikelets have an appearance of being whorled rather than distichously arranged. About three-fourths of our specimens p soon oduced some spikelets from which the florets shed pollen. These staminate spikelets are imperfect, most com- monly produced on proximal portions of branches and central spikes and sometimes separated by sterile zones from pistillate regions. An occasional branch is entirely staminate and ends with a zig-zag axis, but most branches which bear staminate spikelets end in the characteristic pistillate tips described above. Tassels also are as they are exserted. The upper three to five internodes of Tassel-seed 3 plants do not elongate to the degree common in Standards, and as a result a rather tight collar of leaf sheaths is formed covering the lower sixth of each tassel. One tassel in our culture was infected with smut (Ustilago zeae) , a probable reflection of the extreme growth and lasting succulence of tassel parts. Emerson (1920) observed, and Dr. E. G. Anderson agreed, that Tassel-seeds 1 and 2 were highly susceptible to smut, but this year we noted no infections in any of our stocks except Tassel-seed 3. There were often instances in which a spikelet pair consisted of one staminate and one pistillate spikelet; in these mixed pairs, the staminate one is always pedi- cellate. No perfect florets were observed, and only one floret in any particular pistillate spikelet would form a caryopsis. Cupules are well- developed adaxial to both pistillate and mixed spikelet pairs. The ear of Tassel-seed 3 shows very few secondary florets, although Emerson et al (1935) mentioned that such florets do develop. In other respects, the ear resembles that of a Standard plant. 4. Tassel-seed 4 (tsj (pi. 24, fig. 2). First reported by Phipps (1928), and also as Sorghum Tassel (Hayes and Brewbaker (1928), ts± was assigned by Emerson et al (1935) to Chromosome 3. Phipps described the tassel as being tassel-like in structure but predominantly pistillate-flowered, and stated that when it emerged it was a mass of silks. Our material showed tassels much shorter than normal, with an abnormal number of weak branches held upright by a tangled growth of silks. In Phipps' stock, mature tassels were studded with kernels, usually densely packed together, especially on the central axis of the tassel, and not arranged in regular rows. Our specimens varied from setting fruit to this extent to setting no fruit at all. Phipps interpreted this irregular rowing as a result of development of a second floret in each spikelet; with this finding we are in agreement. In both his material and in ours, tassel branches were short and had distichously arranged spikelets, the spikelets on the proximal branch and the lower central spike producing stamens which contained functional poll papery, but we cannot 202 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Phipps that they were similar to glumes of a normal ear; we observed no glume induration. We noted also that sessile spikelets rarely branched, while pedicellate ones often formed short branches bearing two to six spikelets. Our standard stock has an average condensation index of 1.3; spikelets may occur at a given node in pairs, in 4*s, in 3's and 5's and occasionally in 6's. As in Phipps' stock, both florets of each spikelet always developed, and thus each node had at least four and oc- casionally as many as twelve florets in addition to possible extra ones formed on branches of proliferated pedicellate spikelets. Phipps mentioned that careful study of his material showed "a few multi-flowered spikelets." Silks were produced only from pistils, and such florets were always imperfect. No glumes were modified into silks, as Phipps reported. It is doubtful if his ob- servation is accurate, because he noted "the basal parts of such glumes were normal." The silk-like structures were most likely awns, and if so, they were probably produced not on glumes but upon lemmas. We observed no such struc- tures in this mutant, but in a mutant combination not reported on here, awns are developed on the lemmas. In some tassels, spikelets were sterile, producing neither silks nor stamens. These sterile spikelets were most common on branches between the proximally located obviously staminate areas and the distally located obviously pistillate areas. Characteristic of our Tassel-seed 4 stock were both its generally light set of fruit in the tassel and production of staminate and pistillate florets in widely varying amounts. No cupules were developed adaxial to any spikelet pairs of the tassel. Ears of Tassel-seed 4 were often shorter than normal because of failure of the tip to mature; they developed better if tassels were removed early. In our stocks, no excessive development of silks was noted, but Phipps reported that some glumes were modified into silks. Development of second florets, referred to by Phipps as common, was apparently much less common in our stock, and regular rowing was not obscured. Glumes were usually less indurated than in normal ears. 5. Tassel-seed 4* (tsf). This allele of Tassel-seed 4 was found by E. G. Anderson in a background stock of CC5 X L317 which had been exposed to radiation at Eniwetok. Crosses made by the junior author between this mutant and Tassel-seed 4 showed the two types to be allelic. The tassel of Tassel-seed 4 a resembles that of Tassel-seed 4, but the quantity of seed set is much greater. Pollen is much more freely produced than in Tassel- seed 4. Pistillate florets which set fruit are always borne on pedicellate spikelets and are generally imperfect, but some perfect florets have been found which always appeared to set fruit. In general, pistillate florets are confined to the outer third of branches. The central spike is sometimes all staminate, with several extra florets produced on short branches which are proliferations of pedicellate spikelets. Cupules are sometimes weakly developed on the culm of the central spike, but their presence is not universal. Average lengths of glume are about half again as great as those of Tassel-seed 4. 19S5] NICKERSON & DALE TASSEL MODIFICATIONS IN Tea Mays 203 In Tassel-seed 4 a a few pedicellate spikelets have a tendency to proliferate slightly in the upper (distal) third of the ear. After kernels are removed, the right- angled insertion and excessive lengths of the glumes, paleas and lemmas of this allele give the cob a markedly different appearance from that of Tassel-seed 4. 6. Tassel-seed 5 (Ts 5 ) (pi. 25, fig. 1), Although this mutant is attributed to Emerson (1932), the only mention of Tassel-seed 5 in his paper was its inclusion on Chromosome 4 of a linkage map (see his fig. 1, p. 145). Emerson et al (1935), in a brief description, noted that it contained both silks and anthers but was not compacted as in Tassel-seed 4. In our material, tassels were very close to normal proportions, but they were never exserted as far as in normal sibs. Apparently this condition was due to lack of elongation of the internode directly below each tassel, as the plants appeared to be normal in height (Table I and fig. 1). Silks are much shorter and more scattered than in previously discovered tassel-seed forms. Their occurrence varies from very few being located in small branch areas (either basal or terminal), to being uniform over the entire tassel. Spikelets may bear either staminate, pistillate, or perfect florets, but these types occur in specific places. Imperfect pistillate spikelets are usually located in proximal parts of lower branches, and when they occur on central spikes, it is only in their lower regions. Cupules are well developed adaxial to such spikelets, which most often occur as sessile members of spikelet pairs; the pedicellate ones are always both imperfect and staminate. The imperfect pistillate florets form most of the caryopses. Perfect spikelets are never associated with well-developed cupules, and seldom set fruit unless there are few or no imperfect pistillate spikelets present. Even here, however, fruits are formed predominantly on sessile spikelets. If a fruit is set in such a floret, the stamens are generally not exserted, but examination shows their anthers to be full of pollen and occasionally dehisced inside the palea and lemma. Imperfect staminate spikelets occur wherever the other two forms are absent; they never have any trace of silks, and their stamens shed pollen copiously. Silks not pollinated quickly withered. The ear of Tassel-seed 5 in our strain developed very few second florets, but Emerson et al noted that "secondary florets develop in ears." Otherwise, the ear closely resembles that of our Standard stock. 7. Tassel-seed 6 (Ts 6 ) (pi. 25, fig. 2). The first apparent reference to this dominant mutant was made by Emerson (page 14 of the Maize Genetics Cooperation Newsletter for January, 1933). In subsequent 'Newsletters', he established that this mutant is located on Chromosome 1 , but no description of its morphology has been published. foot appear Standards. In general 204 ANNALS OF THE MISSOURI [Vol. 42 appearance, the tassels resemble those of Tassel-seed 4, but they usually set a considerably greater number of fruits. Branches are numerous, short, thin, and lax, with no cupules developed adaxial to any spikelet pairs. They end in many small florets which rarely form fruits and which may be borne on whorled instead of distichous spikelets. Spikelet pairs are closely set on both branches and central spike. Scarcely any distinction exists between pedicellate and sessile individuals, as each one forms several closely packed florets. Short branches with 4 to 12 spike- lets are developed on the central culm, giving it a thickened appearance. Glumes are hyaline, without chlorophyll, and extremely short, averaging about 5 mm. in length. Tassels of this mutant have two general forms. Approximately half of them are entirely pistillate; the others produce staminate spikelets on proximal branch areas and on the lower third of their central spikes. Tassels of this latter sort may have their remaining spikelets constructed in one of three different orders which occurred in our samples with equal frequency: (1) florets bearing silks might also bear stamens and thus be perfect; (2) they may bear no stamens and be imperfect, or (3) a zone of perfect florets of variable length may grade off into a tip which is >erfect. Stamens in perfect florets often neither extruded their anthers nor shed pollen, but those in staminate florets invariably shed good quantities of pollen. Silks are produced only from pistils, but an occasional lemma near a branch tip may develop a short awn. Ears of Tassel-seed 6 are not well developed unless the tassel is removed early. There is the same characteristic production of second florets found in other tassel- seeded forms, so that regular rowing is often obscured. In other respects, ears are comparable with those of Standard plants. lm 8. Tassel-seed 7 (ts 7 ) (pi. 26, fig. 1), This mutant was found by E. G. Anderson at Pasadena in material which had been exposed to radiation at Bikini. (See Anderson et al, 1949, for a further dis- cussion of radiation effects.) It is not known to which chromosome it should be assigned. tions and dimensions of normal ones, and, like Tassel- ristillate florets. Silks are generally short and evenly varies from almost none to profuse. Each silk arises ret; florets which do not produce silks are imperfect and staminate, producing pollen freely. Only two florets are borne on each spike- functional occurrence let. the tassel. The number twelve per tassel, and on about half the tassels they are not formed at all. The ear of Tassel-seed 7 commonly bears up to six short branches at its base. These branches are produced from pistillate spikelets, and an adaxial cupule is present above each one. The sessile spikelet of such a pair is generally not fully developed. These basal branches do not form any viable kernels. Silks form only slightly if at all, even though the branches remain meristematic after cessation of growth elsewhere throughout the plant. Occasionally, two or three small 1955] NICKERSON & DALE TASSEL MODIFICATIONS IN Zft* Mays 205 branches are found on the shank below an ear. These sub-ear structures have always been sterile, and none of them developed beyond the point at which pistillate spikelets could be easily recognized. 9. Tassel-seed 8 (ts 8 ) (pi. 26, fig. 2). This mutant, found by E. G. Anderson at Pasadena, was segregating in a cul- ture of chromosome translocations. It is known to be linked to white endosperm (y 1 ). Both ts 8 and ms 1 tassels are male-sterile, in chromosome 6 and linked to y 19 but the two forms are not allelic. Further, since Ts s sets no fruit but produces silks and since Emerson et al (1935) mentioned no male-sterile forms which so behaved, the material is here described as an eighth tassel-seed form. In over-all appearance, plants resemble normal sibs. Tassels are of standard proportions and size, but glumes are occasionally shorter than normal and the spikelets never swell as time for anthesis approaches. Silk production varies from none to a profusion as great as may occur in Tassel-seed 7. Four-fifths of our population of this mutant had tassels in which no stamens were formed; in the remaining one-fifth, they were formed but their anthers were empty. No stamens are ever exserted. Silks arise from abortive pistils. Three florets per spikelet may send out silks, but no fruits are ever set regardless of ample exposure to pollen. Branch tips are often undeveloped. The ear of Tassel-seed 8 produces a great profusion of silks; 4 to 6 are formed by each spikelet. The first silk formed (from the lowest flower) is of greatest diameter, and its pistil forms a caryopsis. Each of the others sends out a silk which generally is exserted, but no fruits are formed. Kernels are arranged in rows, and, except for the fact that many extra silks protrude between the tightly packed kernels, a mature ear resembles that of a Standard plant. 10. Male-sterile (msj (pi. 27, fig. 1). J occurring j tassels of male-sterile plants were more slender than normals, that anthers failed to extrude, and that no pollen was shed. They also reported that Beadle found meiosis essentially normal but that the haploid nucleus did not go through the first mitotic division, and by the time pollen should be shed it was difficult to find even of spores. Emerson et al noted further that shriveled anthers were exserted much later than in normal sibs. In our material, no stamens were exserted. In other respects, tassels of ms 1 resembled those of our Standards. There was a tendency for the tassel to emerge sooner than with normals, and this tendency was reflected in slightly longer remnants time sibs. [Vol. 42 206 ANNALS OF THE MISSOURI BOTANICAL GARDEN The ear is normal. Singleton and Jones reported that ms 1 is closely linked to the white endosperm locus ( y x ) with about 5 per cent recombination. Our material behaved in the same fashion; white kernels almost invariably produced male-sterile plants. 11. Brancbed-silkless (bd) (pi. 28, fig. 2.) First described by Kemp ton (1934), this mutant is listed by Emerson et al (1935) as occurring on Chromosome 7. Kempton noted that the division of the tassels into a central spike and branches is as definite as in normals, but that the tassels have a thickened appearance suggesting those of Tunicate plants. He at- tributed this thickening to development of short branches in place of paired spikelets, a condition which also occurred in our material. He also mentioned the fact that this branching was most common on central spikes, and that on the branches there was more tendency toward retaining a pedicellate-sessile form. Kempton made no mention of sterility, nor did his illustrations show any, but in our material sterile areas occurred at branch tips and sometimes at the tip of the central spike. Not uncommonly lowermost branches were sterile throughout. In sterile areas, spikelets were less developed as distance from the central culm in- creased. Kempton noted that many plants shed pollen in good quantity, but stamens often were not fully exserted. Ears of Branched-silkless are of two forms, a fact which Kempton likewise noted; his figures show each type clearly. In one type, each spikelet primordium develops into a short branch bearing female spikelet rudiments which never develop beyond the earliest stages. In the other type, such branch development is con- fined to the basal quarter of the ear, the upper three-fourths being composed of sterile spikelets with elongated and sometimes thickened glumes. That these branches develop from spikelet primordia on an otherwise normal cob can be easily shown by removal of the branches; underlying each pair adaxially is a well- developed cupule. It is this latter type which may occasionally produce a scatter- ing of silks too short to become exserted. Kempton stated that his material was wholly without silks, but Emerson et al mentioned that the ear occurs "often without silks. y* 12. Club (club) (pi. 27, fig. 2). This mutant was first noted by E. G. Anderson and segregated from non- irradiated genetic stocks at the Caltech farm in Arcadia. Its chromosome length is unknown. The tassel is shorter than normal, with branches held at angles approximating 45°. Branch ends do not droop and their thickly set spikelet pairs are fertile throughout. The central spike is exceedingly thick, and it is on account of this character that the mutant was named. This thickening is brought about by a shortening of internodes, so that the spikelets stand nearly at right angles to the central culm. Further, these closely spaced spikelets occur singly or in pairs; there is no pedicellate branch development such as is found responsible for creation of thickened portions of the central spike of Branched-silkless. 1955] NICKERSON & DALE Tea Mays 207 The ear of Club generally matures a week later than the tassel. Its silks are profuse, and there are 4-6 silks per spikelet, each from a rudimentary pistil — a condition also found in ears of Tassel-seed 8. However, Club ears are fascia ted, the branching strongly resembling a type listed by Kemp ton (1923) as Bearsfoot (see his fig. 1 for a clear illustration) . It is apparently caused by one or more incomplete divisions of the growing point when the ear is partly formed, followed by simultaneous development of each new point into a more or less independent ear tip. The upper half of such an ear is generally hollow, and no spikelets are formed on the inside walls. 13. Vestigial glume (Vg) (pi. 28, fig. 1). First reported by Sprague (1939) , Vestigial glume was found to be a dominant mutation on Chromosome 1. It is easily noted in the tassel, where the hyaline outer glumes range from awl-shaped vestiges to nearly one-third normal length. Sprague pointed out that "flowering glumes" (lemmas) and "palets" (paleas) were also reduced in size, with the result that the stamens were nearly completely exposed. In our material, as in his, anthers generally dried up before dehiscence, but occasionally a plant shed viable pollen. Many specimens in our stock had sterile areas (i.e., no stamens were formed) at tips of branches and central spikes, in some cases involving the outer sixth of all branches. Generally, over-all tassel size was smaller and tassel construction, as exemplified by thickness of central culm and branches, was lighter than in normal sibs. The ear of Vg is easily recognized, as Sprague pointed out, after removal of kernels. There are no chaffy upper and lower glumes, paleas or lemmas present, only short pedestals, each with a low, hard ridge on both upper and lower sides. Above each spikelet pair, in an adaxial position, is a cupule with somewhat reduced rachis-flaps (Lenz, 1948). 14. Tunicate (Tu) (pi. 22, fig. 2). According to Collins (1917a), this mutant has been reported from several sources and its origin is not known. Recently, it has been cited in connection with theories of maize origin (Mangelsdorf, 1948; Mangelsdorf and Smith, 1949). Cutler (1944) suggested that its widespread occurrence among Indian tribes was because of mystical significance attached to it. Emerson et al (1935) listed the character as occurring on Chromosome 4. Tassels of heterozygous Tunicate plants have normal measurements in all visible parts except the glumes. The glumes are one-fifth to one-fourth longer than those on Standard tassels and enclose paleas and lemmas similarly elongated. The stamen length being normal, the anthers, especially in the central spike, are not fully exserted and therefore are not pendant at anthesis. Pollen is shed abundantly. Ears of heterozygous Tunicate plants are easily classified. Collins (1917a, b) stated that "the glumes of the female inflorescence, or ear, are developed so that each seed [fruit] is entirely enclosed." Emerson et al (1935) likewise attributed [Vol. 42 208 ANNALS OF THE MISSOURI BOTANICAL GARDEN these covers to glumes. In our stock, paleas and lemmas were also elongated, and in many basal spikelets, paleas and lemmas of both upper and lower florets were noted. No fruit occurred in these lower florets, however, and regular rowing was externally apparent. The rachis of a Tunicate cob is rather flexible. Cupule development adaxial to each spikelet pair was not excessive in comparison with Standards; but in the Tunicate ear of Guarani maize sent to the senior author by Dr. P. C. Mangelsdorf, the cupules are much more apparent than is normally the case. Homozygous Tunicate plants, according to Emerson et al (1935), are usually female-sterile. No homozygous plants were available in our collection. An allele of Tunicate, tu h , was reported by Mangelsdorf (1948) as being present in Maiz chapolote. How this allele would behave with our Standards as background was not determined. 15. Ramosa 1 (raj (pi. 29, fig. 1). This form was originally described by Gernert (1912) under the name of Zea ramosa, and was listed by Emerson et al (1935) as Ramosa-ear 1, located on Chromosome 7. Gernert's description of the tassel was brief. In his stocks, tassels were slightly smaller than normal, invariably much branched and cone-shaped. Collins (1917a, b) noted that branches were much more numerous than normal. Kemp ton (1921) reported 400 as an extreme number; the branches gradually de- creased in size upward, the transition from branches to pairs of spikelets being imperceptible. A short central spike was characteristic in our specimens — a fact which Kempton had likewise noted in his plants but which Collins did not mention. It was not possible, according to Collins, to distinguish between plants heterozygous for Ramosa 1 and normals. However, in our material, it was quite simple to sep- arate them; in five F x families of Ramosa 1 X Standard, four with 10 plants and one with 9 plants, separation was easily made into 24 normal and 25 heterozygous individuals. Resemblance of tassels of homozygous Ramosa 1 plants to grass panicles is striking (Kempton's plate 13 is typical). Tassels of plants heterozygous for this character were intermediate between a pyramidal paniculate tassel and a normal one, identical to one figured by Kempton in his plate 14. The ear of Ramosa 1 was described by Gernert as being much branched, with- out male florets, covered with husks, and composed of a mass of kernels borne on numerous irregular branches. His description applies to our specimens as well, except for two other points: (1) branches on a Ramosa 1 ear are about as num- erous and have much the same irregular whorled arrangement as do those of the tassel; (2) there are no adaxial cupules at the junctions of ear branches and central culm. Cupules were present, however, adaxial to each spikelet pair of the branches. No other ear among the mutants in this collection is so organized. It was also a common occurrence for tips of those branches originating near the upper part of the ear to be sterile. 1955] NICKERSON & DALE 7*ea Mays 209 16. Ramos a 2 (ra 2 ) (pi. 29, fig. 2). The only published reference to this mutant is found in Emerson et al (1935), which simply listed Ramosa 2 as being located on Chromosome 3 and credited its discovery to Brink. Ramosa 2 has a tassel characterized by stiff upright branches which remain closely appressed to the central culm. The central spike region does not bear short- pedicelled spikelet pairs; instead the spikes are borne on branches which decrease in length from base to apex. There is a pronounced transition from well-developed tassel branches bearing many pairs of spikelets to smaller branches bearing fewer spikelets, some of which appear to occur singly. As one proceeds acropetally, these multi-spikeleted branches are replaced by stalks each bearing only one pair of spikelets which are themselves borne on pedicels longer than normal. In the adaxial area of each of these branches and stalks, a small cupule-like depression in the central culm can be observed which is generally of a different color from the surrounding surface. Although these depressions and short stalks are not confined to Ramosa 2, they are here most common and best developed. The spike- let pairs in proximal areas of well- developed lower branches are likewise stalked and also have abnormally long pedicels, but there is no adaxial depression. Pollen is shed freely, but much of it remains within the confines of the tassel because close proximity of branches does not allow unrestricted air passage. Each spikelet has two male florets, and each of these has three functional stamens. Tassel-seed 4 a might possibly be confused with Ramosa 2, as its tassels sometimes show the same tendency to have numerous stiff, erect branches and stalked spikelet pairs in the central culm region, with an abundance of pollen produced. It can be separated from Ramosa 2 by the lack of stalked spikelet pairs on lower branches and the fact that each of its branches end in a few sterile undeveloped spikelets. The ear of Ramosa 2 sometimes ends in a staminate structure looking like a normal central spike. Most commonly, it produces scattered branches on the upper (distal) half of the cob, which in turn bear female spikelets in pairs. These branches are nearly always found in younger ontogenetic stages than the rest of the plant and consequently set seed only rarely. A branch originates by growth of what would normally be the pedicellate spikelet of a pair of kernel-bearing spikelets. No instances were observed in which the sessile spikelet of such a pair formed anything but a normal caryopsis. The spikelet pairs themselves are borne on short stalks; the cupule is bent at nearly right angles, and is adherent half to the stalk and half to the cob. In addition, these stalks are spaced rather widely apart, so that a cob of Ramosa 2 from which the chaff has been removed resembles a similarly treated cob of Coroico maize (Cutler, 1946). SUMMARY Morphological studies have been made on tassels and ears of sixteen mutant forms of maize. Four of these forms are reported for the first time, and the others, for which descriptions were either inadequate or non-existent, are here [Vol. 42 210 ANNALS OF THE MISSOURI BOTANICAL GARDEN described. Each genetic form was introduced into a standard background derived from the inbreds CC5 and L317. The mutants studied included nine tassel-seeded forms: ts 19 ts 2 , Ts 3 , ts 4 , /s 4 a , Ts 5 , Ts 6 , ts 7 and ts 8 . Others studied were Male- sterile (msj, Branched-silkless (bd), Club (club), Vestigial glume (Vg) f Tuni- cate (Tu), Ramosa-1 (raj and Ramosa-2 (ra 2 ). LITERATURE CITED Alava, Reino O. (1952). Spikelet variation in Zea Mays L. Ann. Mo. Bot. Gard. 39:65-96. Anderson, Edgar (1944a). Homologies of the ear and tassel in Zea Mays. Ann. Mo. Bot. Gard. 31:325-344. , (1944b). The sources of effective germ -plasm in hybrid maize. Ibid. 3 55-3 61. , (1949). Introgressive hybridization. 109 pp. New York. , (1951) The sacred plume. 24 pp. Publ. Pioneer Hi-Bred Corn Co. Des Moines. , and W. L. Brown (1952a). The history of the common maize varieties of the United States corn belt. Agric. History 26:2-8. » ■ (1952b). Origin of corn belt maize and its genetic significance. In: Heterosis, pp. 124-148. Edited by John W. Gowen. State College, la. Anderson E. G., A E. Longley, C. H. Li, and K. L. Retherford (1949). Hereditary effects pro- duced in maize by radiations from the Bikini atomic bomb. I. Studies on seedlings and pollen of exposed generation. Genetics 34:639-646. Collins, G. N. (1917a). Hybrids of Zea ramosa and Zea tunicata. Jour. Agric. Res. 9:3 83-395. - , (1917b). Hybrids of Zea tunicata and Zea ramosa. National Acad. Sci. Proc. 3:345-349. Cutler, H. C. (1944). Medicine men and the preservation of a relict gene in maize. Tour. Hered. 3 5:291-294. J , (1946). Races of maize in South America. Harv. Univ. Bot. Mus. Leafl. 12-2 57-292 Emerson, R. A. (1920). Heritable characters of maize. II. Pistillate-flowered maize plants. Tour. Hered. 11:65-76. r J , (1932). The present status of maize genetics. Sixth Internat. Cong. Genetics Proc. 1:141-152. Co "'iP;r W * ? Cadle ' 3nd A ' C Fraser < 1935 )- A summary of linkage studies in maize, nell Univ. Agric. Exp. Sta. Mem. 180:1-83. Gernert, W. B. (1912). A new subspecies of Zea Mays. Amer. Nat. 46:616-622. Hayes, H. K., and H. E. Brewbaker (1928). Heritable characters in maize. XXXIII. Sorehum tassel. Jour. Hered. 19:561-567. Kempton, J. H. (1921). Inheritance of ramose inflorescence in maize. U. S Dept Aer Bull 971:1-20. 5 " -, (1923). Heritable characters of maize. XIV. Branched ears. Jour. Hered. 14:243-252. , (1934). Heritable characters in maize. XLVII. Branched silkless. Ibid. 25:29-32. Lenz, L. W. (1948). Comparative histology of the female inflorescence of Zea Mays L Ann Mo Bot. Gard. 35:353-376. Mangelsdorf, P. C. (1948). The role of pod corn in the origin and evolution of maize. Ann. Mo. Bot. Gard. 3 5:377-406. , and C. E. Smith, Jr. (1949). New archaeological evidence on evolution in maize Harv. Univ., Bot. Mus. Leafl. 13:213-247. Nickerson, N. H. (1953). Variation in cob morphology among certain archaeological and ethno- logical races of maize. Ann. Mo. Bot. Gard. 40:79-111. — - , (1954). Morphological analysis of the maize ear. Am. Jour. Bot. 41:87-92. Phipps, I. F. (1928). Heritable characters in maize. XXXI. Tassel-seed 4. Jour. Hered. 19:399-404. Singleton, W. R and D. F. Jones (1930). Heritable characters in maize. XXXV. Male sterile. Jour. Hered. 21:266-268. Sprague, G. F. (1939). Heritable characters in maize. L. Vestigal glume. Jour. Hered. 30:143-145. Weatherwax, Paul (1916). Morphology of the flower of Zea Mays. Bull. Torrey Bot. Club 43:127- 143. 1955J NICKERSON & DALE TASSEL MODIFICATIONS IN Z*0 Mays 211 Fig. 2. Drawings to same scale of spikelet pairs found on lowermost primary tassel branches of certain mutants showing characteristic features: it% 9 adaxial view of thick, ribbon-like branch showing cupules and a developed second floret (silks removed). Ts? y , pedicellate (right) and sessile (below) spikelets with hyaline awnless glumes; pedicellate spikelet forms numerous naked small pistils, bd, both spikelet axes form extra spikelets. r*2, an extra interncde is found between the primary axis and point of departure of the sessile spikelet. Vg, glumes do not develop, so stamens are left naked. ts 4i sessile spikelet left has one male and two female florets; pedicellate spikelet (right) forms numerous spikelets, with many of the thin glumes ending in soft awn-like tips. rs 3 , sessile female and pedicellate male in one spikelet pair; pedicellate axis adheres to edge of cupule for part of its length (silks removed). Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 22 3 > J, > r O Q 1 O >fl n > C/5 X 2 1 to ~r. 1 Co i'O t -'' ■* A afc-i^* *•• Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 24 . h ■ J? t W < * X > * * V J£ 2 * » / c\> o I o 1/3 (/) H (N tc >3 m a o C/3 I to H N Z z < u, Q S w < 1 a o z Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 25 Z > CO m 8 -n n > CO 2 co c5* C/i l o vi •n cr5* C/) i en ON C5 A3 M Ann. Mo. Bot. Gard., Vol. 42, 195 5 Plate 26 Cv? MOT I rs H 00 N CO o U n< Q O w on en i Q < z » o ~0> CO W5 rf rt w H * y • Z w> * i— • fJH rH Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 27 2 O O > o o > r 1 j, > o I riT to ►n > o . z 2 crc* K) N til c/3 8 <> ■*■ ...- . ■■JSS. " Ann. Mo. Bot. Gard., Vol. 42, 195 5 Plate 28 fc, - .^ met ■'" *« o C/5 u 3 I -a M q> WD - -J Q Z o r- ^ E * 2 Z rH Ann. Mo. Bot. Gard., Vol. 42, 195 5 Plate 29 Z o > o > I > f ^i n > o on 2 1 s5' » o Z £ J! erg* K> 3 o i\3 A NEW SPECIES OF DORYOPTER1S FROM SURINAM* KARL U. KRAMER and ROLLA M. TRYON, JR. In the course of examination of specimens in the fern collection of the Botanisch Museum en Herbarium, Utrecht, the senior author discovered a speci- men of Doryopteris that seemed to be different from any species treated in the junior author's revision of the genus 1 . Further examination has proved this to be the case and it is here described as new. Doryopteris conformis, spec. nov. Rhizoma modice crassum breviter repens, squamis elongatis angustissimis dense vestitum, partium hyalinarum cellulis latitudine maxime partem quintum longi- tudinis aequantibus; stipes obscurus laevis vel leviter rugosiusculus, fasciculis vascularibus duobus; lamina fertilis sterili similis coriacea, venatio libera; sterilis suborbicularis — quinquangularis, profunde bipinnatifida; fertilis conformis, pro- funde bi- vel tripinnatifida; receptaculum intramarginale plus minusve continuum; sporangia breviter pedicellata. Typus: Gonggrijp & Stahel (B. W.) 5699; Surinam: Mt. Hendriktop, alt. 1080 m.; moist, sunny rocks; in Herb. Utrecht. Rhizome moderately stout, short-creeping; scales of the apex of the rhizome very long and narrow, the cells of the hyaline portions at least five times as long as broad; stipe dark purple to black, naked or slightly scaly at the base, glabrous, smooth or minutely and irregularly roughened, with two vascular bundles at the base, terete; fertile and sterile blades similar, without proliferous buds, coriaceous; sterile leaf about 25 cm. long; blade 7 cm. long, suborbicular-pentagonal, deeply bipinnatifid with about 12 oblong, broadly rounded, entire or partially crenulate ultimate lobes; margin with a pale brown cartilaginous border; venation free; hydathodes prominent on the upper surface; fertile leaf about 25 cm. long; blade about 8 cm. long, suborbicular-pentagonal, deeply bi- to tripinnatifid, with num- erous ultimate segments; primary segments broadly decurrent and surcurrent, the bases forming wings along the rachis with symmetrically concave sides; ultimate segments oblong-lanceolate, narrowly rounded, entire; soral lines continuous around the sinuses; sporangia short-stalked, i.e. the stalk somewhat shorter than capsule (which is 340—360 fx long), borne on a more or less continuous vascular commis- sure; spores subglobose, triplanate, slightly rugose, pale yellow-brown, about 55 fi. Doryopteris conformis is most closely related to D. lomariacea from which it differs in the non-dimorphic fertile and sterile leaves, the segments of the fertile being much broader than in D. lomariacea, in the short-stalked sporangia and coriaceous texture of the blades; in D. lomariacea, the sporangia are long-stalked and the leaf-tissue is herbaceous. This species is notable in that it is the only local endemic in the genus outside of southeastern Brazil. However, the Guiana Highlands represent a weak sec- 1 Tryon, Rolla M. A revision of the genus Doryopteris. Contr. Gray Herb. 143:1-80. 1942. ♦Issued November 17, 1955. * ' [Vol. 42, 1955J 214 ANNALS OF THE MISSOURI BOTANICAL GARDEN ondary center for the genus in South America. This is the sixth species known from there, not including the widespread and somewhat doubtfully allied D. concolor; a concentration of species exceeded only in the southeastern Brazilian Highlands. It is also significant that all the other five species, D. lomariacea, sa git ti folia, collina, v avians, and pedata var. multipartita are variously disjunct between southeastern Brazil and the Guiana Highlands. Whether D. conformis represents a local offshoot of D. lomariacea or whether it was evolved, as all other Guiana species evidently were, in southeastern Brazil and migrated via the Andes to the Guianas, can onlv be determined if it is eventuallv discovered in Uracil. Explanation of Plate PLATE 30 Fertile and sterile leaf of Doryopteris conformis, from type. X % Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 30 KRAMER AND TRYON— DORYOPTER1S CONFORM1S A REVISION OF THE GENUS CELASTR US* DING HOU** Introduction There are about 500 species which have been referred in the past to the genus Celastrus; of these half are African plants. In 1942 Loesener 1 , in a review of the genus for Engler & PrantPs Tflanzenfamilien', transferred some of the African species to the genus Gymnosporia and some to the allied genus Maytenus. In indicating the geographical distribution of Celastrus only in Asia, America, Australia, and Madagascar, he seems to have been aware of the fact that the so- called species of Celastrus of Africa proper did not truly belong to that genus. Loesener has expressed the opinion that the confusion regarding species limits in Celastrus and the synonymy indicate the need for a more precise study. I have attempted in this treatment to define the generic limits of Celastrus and its rela- tionship with other closely related genera, to review and to check all the published binomials of Celastrus, and to clarify the complicated synonymy. I have also made a study of the morphological characters and geographical distributions of the dif- ferent species. I have used the data to separate species or groups of species and to show their inter-relationships (fig. 1). All the (f Celastrus y> species from Africa proper have been excluded in this treatment, as they are referable either to Gymno- sporia or Maytenus. I have tried to delimit definitely the three genera, Celastrus, Gymnosporia, and Maytenus, which have been very much confused. It is my intention in the near future to make a detailed study of these last two genera. This treatment of Celastrus includes two subgenera with thirty-one species and five subspecies. The generic name comes originally from Theophrastos, who, however, desig- nated with the Greek word KeXaorpos (Kelastros) an evergreen tree (Phillyrea), that has nothing to do with our genus 2 . It is feminine in gender as used by Theo- phrastos, but Linnaeus in adopting it, made it masculine. It has already been pointed out by Airy Shaw 3 that under the International Rules the masculine gender must be retained. Hence, in this treatment, Celastrus is treated as masculine and I have standardized all the epithets accordingly. 1 Loesener, Th., in Engler, A. & PrantI, K. Die Natiirlichen Pflanzenfamilien 2 Aufl. 20b: 131, 134. 1942. 2 Loes. loc. cit. 2 Aufl. 20B:132. 1942. 3 Airy Shaw, H. K. in Curtis's Bot. Mag. 158:/. 9394. 1935. ♦An investigation carried out in the Henry Shaw School of Botany of Washington University and submitted as a thesis in partial fulfillment for the requirements for the degree of Doctor of Philosophy. ** Arnold Arboretum of Harvard University, Cambridge, Mass. Issued November 17, 1955. (215) I Vol. 42 216 ANNALS OF THE MISSOURI BOTANICAL GARDEN History The genus Celastrus was founded and described by Linnaeus 4 in his ^Genera Plantarum* (1737). Later, in his 'Species Plantarum' 5 (1753), he described three American and two South African species. Only one of these, Celastrus scandens, is now retained in the genus Celastrus. In 1824, Kunth 6 pointed out that some of the Celastrus species, for example, Celastrus buxifolia L., C. montana Roxb., C. trigynus Lam., etc., have peculiar characters and might well constitute a new genus. He gave a detailed description for that group of species, but provided no name for it. In 1834, Wight and Arnott 7 studied the Indian species of Celastrus, and, chiefly on the basis of ovule or seed characters, divided the genus into two sections, eucelastrus and gymno- sporia. They gave each section a very concise and clear description. In the sec- tion eucelastrus, the ovary is free from the disc, the ovules have a cup-shaped aril at their base, and the seeds are surrounded by an entire fleshy aril. In the section gymnosporia, which corresponds to Kunth's unnamed genus, the ovary is half- immersed in the disc, the ovules are naked at the base, and the seeds are apparently without an aril or with a very short, imperfect, and membranaceous one at the hilum. In 1862, Bentham and Hooker 8 elevated the section gymnosporia to the rank of genus. At present the limits of the genus Celastrus are identical with those of section eucelastrus of Wight & Arnott. Generic Relationships porta They often have been confused with Celastrus both in the literature and in the herbaria. porta Maytenus, on the presence in the former of either thorns or inflorescences borne on short shoots, appears artificial. He says: "There seems little point in keeping the two genera distinct." After examining all available specimens of Celastrus, Gymnosporia, and May- tenus, I have concluded that they are distinct though very closely related genera. The following table for distinguishing the three genera is based on a consideration of several characters taken together and not any one single character taken by itself (pi. 31). 4 Linnaeus, C. Genera Plantarum. ed. 1. 59. 1737. 5 - ■• Species Plantarum. ed. 1. 195-197. 1753. 6 Kunth, C. S. in HBK. Nov. Gen. & Sp. Pi. 7:64. 1824. 7 Wight, R. & Arnott, G. A. W. Prodr. Fl. Ind. Orient. 152. 1834. 8 Bentham, G. & Hooker, J. D. Genera Plantarum 1:364. 1862-67. 9 Exell, A. W., in Kew Bull. 1953:103. 1953. 10 Loes. loc. cit. 2 Aufl. 20b:109. 1942. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 31 CELASTRUS MAYTENUS HABITS TWINERS PHYLLOTAXY LENTICELS ALTERNATE DISTINCT DISTINCT DICHASIA OR RARELY SOLITARY INFLORESCENCES SHRUBS, SMALL TREES OR TREES GVMNQSPORIA ALTERNATE RARELY OPPOSITE OBSCURE FASCICLED OR RARELY SOLITARY SHRUBS OR TREES ALTERNATE OBSCURE DISTINCT DICHASIA OR FASCICLED AXILLARY BUD AND BUD SCALES DISTINCT OBSCURE FLOWERS ARTICULATIO N ON THE PEDICEL POLYGAMOUS RARELY BISEXUAL DISTINCT M^ BISEXUAL OBSCURE BISEXUAL OVARY AND DISC NO. OF OVULES IN EACH CELL OVARY FREE FROM THE DISC, 3- CELLED; DISC USUALLY CUP-SHAPED, RARELY MERELY CONCAVE, MEMBRAN- OUS, RARELY SUBFLESHY NONE OVARY SEMI -IMMERSED IN THE DISC, USUALLY 2-CELLED, RARELY 3-CELLED; DISC FLESHY FLAT, RARELY CONCAVE DISTINCT OVARY SEMI -IMMERSED IN THE DISC, USUALLY 3-CELLED, RARELY 2-CELLED; DISC FLESHY FLAT, RARELY CUP-SHAPED USUALLY 2 (1 IN CENTRAL AMERICAN SPECIES) SHAPE SUBGLOBOSE. FURROWED USUALLY 2, RARELY 1 ALWAYS 2 ELLIPSOID OR OBOVOID , SMOOTH SUBGLOBOSE ANGULAR CO NO. OF VALVES NO. OF SEEDS 3-VALVED USUALLY 3-6-SEEDED, RARELY 1-SEEDED SHAPE UJ UJ USUALLY OVOID OR ELLIPSOID; ROUND IN CROSS-SECTION USUALLY 2-VALVED. RARELY 3-VALVED USUALLY 1-SEEDED, RARELY 2-4-SEEDED. USUALLY 3-VALVED, RARELY 2-VALVED USUALLY 3-6-SEEDED o USUALLY OVATE; OVATE OR PLANO-CONVEX IN CROSS-SECTION ARIL ARIL OF THE OVULE DISTRIBUTIONS COMPLETE COMPLETE OR INCOMPLETE CUP-SHAPED ARIL CHIEFLY IN ASIA, SOME IN AMERICA & MADAGASCAR USUALLY OVOID OR ELLIPSOID ROUND IN CROSS -SECT I ON O AT THE BASE OR INCOMPLETE NONE OR SLIGHTLY AT THE BASE NONE CHIEFLY IN S. I C. AMERICA FEW IN AUSTRALIA AND AFRICA CHIEFLY IN S. AFRICA AND ASIA FEW IN PACIFIC ILS. Celastrus Maytenus, and Gymnos porta contrasted. 218 ANNALS OF THE MISSOURI BOTANICAL GARDEN [ X. 42 Morphology habit: — Plants of Celastrus are scandent, from 1 to 50 meters in height. Th branches are hardy, rapid-growing and vigorous. There are very few available records of the diameter of older branches for all the species, but two records of p The branches and branchlets usually are terete, except in Celastrus angulatus, and are glabrous except on young branchlets, for example, of Celastrus hirsutus. Most older stems have orbicular to ovate lenticels. In Celastrus hindsii, lenticels are usually absent on the current year's growth. They are fine and dense in Celastrus angulatus and C. lenticellatus 9 but in C. hirsutus and C. c ase ariij olius they are large, elevated, ovate, or orbicular, and are found on the peduncles and pedicels. This feature may well serve as a supporting character to assist in identification of some species. vegetative buds: — Celastrus gemmatus bears axillary buds which are conical and characteristically large. All other species bear small, depressed and ovoid ones. The inflorescences of the species of series axillares (except Celastrus monospermus y C. monospermoidesy and C. hindsii) and one species of the subgenus racemo- celastrus, C. panamensis 9 are associated with accompanying vegetative buds, while the inflorescences of other species are not associated with vegetative buds. The outermost bud scales are usually deciduous, occasionally persistent, falcate, and spiny. These persistent bud scales are called "stipules" by some authors. Usually in Celastrus aculeatus and in some specimens of other species, there are four or more outermost bud scales as opposed to the usual two. These are nearly triangular, spiny, and semi-persistent or persistent. leaves: — The leaves are extremely variable, in the same species, and even on the same plant, the result of environmental conditions and the age of the plant. Previous authors mainly have used leaf characters for distinguishing species and varieties in the genus, but such characters are very unreliable. Most of the species are deciduous, but there are evergreen ones, for example, Celastrus monospermtis, C. monospermoides, and C. hindsii. The leaf shapes vary from elliptic to oblong, or from broadly ovate to orbicular. The apex and the base vary from acute to obtuse or rotund. The margins usually are serrate, serrulate or subentire; in Celastrus flagellaris they appear finely cilia te. In size, the leaves range from 2.0 cm. long and 0.8 cm. wide in Celastrus punctatus to 16 cm. long and 16 cm. wide in C. angulatus. Texture usually varies from delicately to firmly membranous. In general, the texture of the leaves in flowering and fruiting stages of the plant is different: for example, in Celastrus angulatus the leaves of flowering specimens are delicately membranous, while those on the fruiting specimens are firmly membranous (so- called "coriaceous" of some authors). The leaves are glabrous or pubescent on the veins below. Sometimes they are pubescent in the juvenile stage and later become glabrous. In Celastrus hirsutus there is a dense brownish pubescence on both surfaces. i 1955] HOU REVISION OF THE GENUS CELASTRUS 219 The leaves of all the species have pinnate and netted veins. The midrib usually is elevated below, and distinctly or slightly elevated above- The primary lateral veins are four to nine pairs on each side of the midrib, arcuate towards the apex. The veinlets though usually visible, are obscure on both surfaces in Celastrus aculeatus, and while usually loosely reticulated are densely reticulated in C. hindsii and C. gemmatus. The phyllotaxy always is alternate. The petioles are canaliculate, from 0.5 to 3.0 cm. long. In Celastrus flagellaris the petiole is half as long as the blade while in the other species it is less than half as long. floriferous branches and inflorescences: — Diagrams of the floriferous branches of the current season's or year's growth of most of the species of Celastrus have been drawn from herbarium specimens (pis. 32 and 33). These diagrams are arranged according to the sequence of the species in the key, and are not in- tended to show any evolutionary paths. They will make it convenient for a com- parison of the relationships among the species. They also reveal that the male and female plants of the same species have different inflorescence patterns. A "flower-cluster" has been thought of as a whole inflorescence or as a part of the inflorescence. Parkin 11 has commented: "It seems futile to quibble over the question whether the inflorescence means the mode of floral branching or the flower-group itself. Custom has sanctioned the latter meaning." I have adopted the customary approach of considering a "flower-group" as an inflorescence only as a matter of convenience in this treatment of the genus Celastrus. Dichasium and Aggregate Dichasium: — The fundamental type of inflorescence in the genus Celastrus is an axillary, solitary dichasium. A dichasium in its simplest form consists of three flowers — one terminal and two lateral, the terminal flower blooming first — and is organized as follows: First, there is the primary peduncle terminating in three secondary peduncles to each of which the pedicels of the flow- ers are articulated. What appears at first to be the entire flower stalk therefore is really a part of the peduncle and the whole pedicel. The dichasium usually is com- pound, and sometimes the branches are much multiplied to form a terminal panic- uliform aggregate dichasium. The aggregate dichasium differs from the thyrse in all axes being determinate and lacking any association with vegetative buds. In the literature the paniculiform floriferous branches are described usually as "panicles." Each dichasium may or may not be subtended by a foliage leaf or bract, but in the series axillares it is always associated with a dormant bud in the axil of the subtending appendage (pis. 31 and 32). All the species of subgenus racemo- celastrus have axillary dichasia with or without distinct primary peduncles. The Latin American species of racemocelastrus, in addition, usually have dichasia arranged to form a racemiform floriferous branch on the current year's growth (pi. 33). In this case the racemiform branch may be compounded and be located terminally on the branchlets to form a pseudopaniculiform aggregate dichasium ii Parkin, J. The evolution of the inflorescence. Jour. Linn. Soc. Bot. 42:512—563. 1914. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 32 ftali.4* 79 if) EMl V (Fr. J C. PANICUIATUS Cl«Mn« bl&2 C. SCANDENS *%^ ■ i llliM 36* f.) C. ANGULATUS - I ^r'- *%%*% aeaoo Ul "0 |f> C. HINDSli VlUto Pr.J MtftMMi, f» £) r»m«t »40t *#) s.r.iu aoow »rr.> C. MONOSPERMUS C. GIAUCOPHYUUS C. VANIOT I rr. » ■urn* iat» i£i ■ii«o« *♦« irr.) C. HYPOLEUCUS Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 33 Porrast 16M1 CMftg 20682 <-£) C. GEMMATUS C. PUNCTATUS C. ACULEATUS Cull, to M.B.O. 5*»ortso* f^) Vllton 7«12 f C. 0R6ICULATUS rr.l C. ROSTHORNIANUS 8 C. KUSANOI r eta lot 80M fcj) Till 60499 tFr.l C. HIRSUTUS Hanry MM (fl tl 1181 IPr.) a C.STYLOSUS SSP.GLABER -I 'tiara SaSO ( 7956o (A); Cheng-kango, Monghui, Yu 16787 (A); Fo-Hai Hsien, 74966 (A); Jenn-yeh Hsien, Wang 80588, 80194 (A); Shunning, Yu l62Q2 (A); Szemao, Henry I2I22 (MO), I2I22*, 12122^ (US), 12572* (A), IIQ72 A (US), II993 (A, MO), 11993 Thomp Indo-China: Prov. Sonla, 4049 (A, UC) ; Bao-chiang, Prov. Bien-hoa, Pierre 2794 (lectotype of Celastrus laotica Pitard, P) ; Phuoc-than, Tborel s. n. y 1862-66 (paratype of C. laotica Pitard, P) ; Laos, Phon-thane, Spire 289 (paratype of C. laotica Pitard, P). Indonesia: Celebes: Minahassa, Lam 2466 (L). java: Bantam, Kuhl & v. Hasselt 18 (L) ; Madiun, Ngebel, Koorders 2980B (L) ; Parve-sosa, Backer 37252 (L) ; Pehalongan, Koorders 27372B (L); Soerabaja, Dorgelo 2290 (L). Sumatra: without precise locality, Junghubn s. n. (type of Alsodeia glabra Burgdk., L). Malay Peninsula: Pahang: Kwala Sunbelung, Mat s.n., 1893 (SING); Sea Pardens, Ridley s.n., 1892 (SING) lerrill 9311 (A, F, L, MO, SING, US). SING); Bitsenuloke, Groff 6132 (UC); Kin Sayo, "Wanjek 441 (L, SING) ; without precise locality The wide distribution and the variable leaf forms of this species have been the cause of the many synonyms. Because of the inadequacy of flowering material, I selected the middle leaf of a branch of current season's or year's growth for com- parative statistical study, following Woodson's 27 method of measuring leaf char- acters in Asclepias. A study of the herbarium material according to the geo- graphical distribution and measurable differences among the populations has demon- strated that this species includes three distinct populations, which I am treating as three subspecies (pi. 34, upper diagram). Celastrus paniculatus ssp. paniculatus has leaves which are usually suborbicular or broadly obovate. It is chiefly distributed in India, but it extends to Burma, Yunnan multifl rerlaps ssp. paniculatus. It is also distributed in Hainan, Formosa, Siam, Indo-China, Malaya, Indonesia, and Palawan IsL, in the westernmost Philippines. The leaves of ssp. serratus are usually broadly ovate. This subspecies is chiefly distributed in the Philippines, and is less common in Java where it overlaps ssp. multtflorus. 27 Woodson, R. E., Jr. Some dynamics of leaf variation in Asclepias tuberosa. Ann. Mo. Bot. Gard. 34:353-432. 1947. Ann. Mo. Bot. Gard., Vol. 42, 1955 Plate 34 C. ZA 105 100 95 90 85 80 PANICULATUS SSP. PANICULATUS •19 SSP. SERRATUS O. O O o o o o o o o o O SSP. MULTIFLORUS o o o o o o Q v O O O o o 75 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.3 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 LEAF LENGTH/WIDTH RATIO Frequency distribution of three subspecies of Celastrus paniculatus: large dots, up. paniculatus; small dots, ssp. serratus; circles, ssp. multiflorus. Explanation in the text. [Vol. 42 234 ANNALS OF THE MISSOURI BOTANICAL GARDEN I have not seen any specimen of C. paniculata collected in Borneo, but I think that it may be found there from the distribution pattern. It is interesting to note from the geographical distribution of these three subspecies that the specimens collected in Palawan Island and Formosa belong to the ssp. multiflorus instead of the Philippine ssp. serratus. Geologically, Palawan Island and Formosa belong to the Malaysian continent. The leaves borne on a branch of current season's or year's growth are slightly variable. If a leaf picked at random from a plant of one subspecies is compared with that of another subspecies the same size and form can be found. Following Anderson's 28 method, I have traced the leaf forms on a current season's or year's branch for each subspecies, and the differences in leaf patterns were easily observed. 2. Celastrus novoguineensis Merr. & Perry, in Jour. Arn. Arb. 22:260. 1941. (T.: Clemens 5152, A!). Scandent shrubs; branchlets terete, glabrous, brown to reddish-brown; the lenticels numerous, elliptic, elevated, white; axillary buds conoid, obtuse, about 2 mm. long. Leaves oblong, ovate to elliptic, the apex acute or obtuse, the base obtuse or rotund, the margins minutely and remotely serrulate, 12-16 cm. long, 5.0—7.5 cm. wide, firmly membranous, glabrous, the primary veins 6 pairs, elevated below, distinct, plane or slightly elevated above, the veinlets obscure below, plainly visible above; stipules linear, about 2.5 mm. long; petioles 1.0-1.5 cm. long. Inflorescences terminal, paniculiform, usually thrice compound, about 15-20 cm. long, the floriferous branchlets divaricate, the peduncles glabrous, the primary peduncles about 6—10 mm. long; flowers green, dioecious, the pedicels about 1 mm. long, the articulation varying from the lower to upper half of the stalk. Male flowers: calyx lobes imbricate, subreniform, slightly erose, about 1 mm. long; petals oblong, rotund, subentire, scarious-marginate, about 2.5 mm. long and 1.5 mm. wide, subfleshy, brownish-punctate; disc cup-shaped, membranous, the lobes obscure, truncate or mucronate; stamens subsessile or with very short flat filaments, the anthers narrow-ovoid, obtuse, apiculate, brownish-dotted; sterile pistil 1 mm. long. Female flowers: calyx lobes, petals, and disc as in the male; sterile stamens 1 mm. long; pistil about 1.4 mm. long, the ovary subglobose, wider than long, the style obsolete, the stigmata 3-lobed, each bifid. Fruits subglobose, the valves broadly ovate, about 9—12 mm. long and 7-10 mm. wide, 3- to 6-seeded; seeds ellipsoid, 5—8 mm. long and 2.5—3.5 mm. wide, reddish-brown, smooth. In thickets, at altitudes 1,220-1,830 m.; New Guinea, flowering from February to September. Guinea W (A); Yoangen (Yunzaing), /. & M. S. Clemens 3523, 6606, 7230 (A), papua: Lala River, Carr 15608 (A). Netherlands new guinea: Meervlakte, Motor bivouac, f. Leeuwen 1 1 008 (L). 28 Anderson, E. Concordant versus discordant variation in relation to introgression. Evolution 5:133-141. 1951. 1955] HOU REVISION OF THE GENUS CELASTRUS 235 Map 1. species PANICULATI. [Vol. 42 236 MISSOURI This is an endemic species of New Guinea. It is distinctive and can be dis- tinguished from other species by the thrice compound aggregate dichasia and main floriferous axis with branchlets extending out at nearly right angles. The stamens in the male flowers usually are sessile or with very short, flat filaments. The pistil in the female flower has a very short style. 3. Celastrus subspicatus Hook. Icon. PL 5:t. 482. 1842. (T.: Collector un- known, s. n. y cultivated at Kew, K!). Celastrus australis Harv. & Muell. in Trans. Phil. Soc. Vict. 1:41. 1855; Benth. Fl. Austral. 1:398. 1863, ex char. (T.: Mueller s.n.). Celastrus papuana Warb. in Engl. Bot. Jahrb. 13:366. 1891, ex char. (T.: Warburg s.n.). Celastrus paniculatus Willd. var. Balansae Loes. in Engl. Bot. Jahrb. 39:160. 1906, ex char. (T.: Balansa 3029). Scandent shrubs; branchlets terete, brownish-pubescent, the lenticels elliptic, distinct, and white; axillary buds deltoid, spiniform, about 1-2 mm. long, the outermost scales acuminate. Leaves elliptic, ovate, suborbiculate, or ovate- lanceolate, the apex acuminate, the base obtuse to acute, the margins subentire to remotely serrulate, 4-12 cm. long, 1-6 cm. wide, membranous, glabrous, the primary lateral veins 5—8 pairs, the veins and veinlets slightly elevated on both surfaces; stipules linear, about 2.5 mm. long; petioles usually 4—7 mm. long. In- florescences terminal, once compound, rarely twice compound in the male, 2-9 cm. long, occasional in the axils of the uppermost leaves, the peduncles usually brownish-puberulous ; flowers white, the pedicels about 3 mm. long, the articula- tion toward the base of the stalk. Male flowers (young) : anthers ovoid, obtuse. Female flowers: calyx lobes imbricate, orbicular, broad, cilia te, about 1 mm. long; petals ovate-oblong, rotund, slightly erose, about 2.5 mm. long and 1.0-1.5 mm. wide; disc cup-shaped, the lobes obscure, truncate; sterile stamens 0.5 mm. long, the filaments very short; pistil ovoid, 2 mm. long, the style short, the stigmata 3-lobed, spread. Fruits subglobose, the valves broadly elliptic, about 5—10 mm. long and 5—8 mm. wide, 3- to 6-seeded; seeds ellipsoid, 5 mm. long and 2.5 mm. wide, reddish-brown, the areolae obscure. In the thickets; Australia, New Caledonia, and Papua. Australia: Queensland: Cedar Beenleigh, White s.n., May 1920 (SING); head of Burnett R., Miiller s.n. (GH); Benarkin, Cameron s.n., April 1924 (A); Mt. Glorious, White Unungar (UC); without definite locality: Strathdickie 1 124. (GH), Clemens 43683 (A), new south wales: Success Hill, Caley s.n., 1807 (A, UC, US); Tweed River, Moore s.n. (GH) ; cultivated at Kew, collector unknown s. n. (holotype of Celastrus subspicatus Hook., K). New Caledonia: Balansa 1 8 70 (A); De planch 92 (A). New Guinea: papua: Kanosia, Carr 11260 (A, L). Celastrus subspicatus was published by W. J. Hooker, based on a plant long cultivated in the Royal Botanic Gardens at Kew. Its history and habitat are unknown. In the type specimen, borrowed from Kew Gardens, the leaf forms, texture, short petioles, distinct and elevated lenticels, acuminate outermost 1955] HOU REVISION OF THE GENUS CELASTRUS 237 nflorescen This simi- larity leads me to believe that Celastrus subspicatus of Kew Gardens was grown from seeds obtained from Australia. Specimens collected from New Caledonia and lowland Papua, New Guinea, are morphologically similar and geographically related to the Australian ones except in the leaf shape. Meanwhile, not enough specimens are available for a population study to find out whether they really differ from the Australian species and whether they should be put into subspecific or varietal category, so they are treated tenta- tively as one species. 4. Celastrus richii A. Gray, in Wilkes U. S. Expl. Exped. (Bot. Phanerogam., pt. 1):289. 1854. (T.: Wilkes Exped, s.n., GH!). Scandent shrubs; branchlets terete, glabrous, light to dark brown, the lenticels distinct and numerous, white and elliptic, rarely lacking on the young branchlets; axillary buds small, conic, about 1 mm. long. Leaves elliptic, the apex and the base acute, the margins subentire or remotely serrulate, thinly membranous, shining above in dry condition, glabrous, the primary veins 5-7 pairs, elevated on both surfaces, the veinlets distinct on both surfaces, densely reticulate; stipules laciniate, filiform, about 1 mm. long; petioles 2-9 mm. long. Inflorescences terminal, once or twice compound, 2—7 cm. long, the peduncles glabrous, the primary peduncles 3.5-5.5 mm. long; flowers white, the pedicels 1.0-2.5 mm. long, accrescent, up to 15 mm. long on the fruits, the articulation toward the base of the stalk. Male flowers: calyx lobes imbricate, subreniform or semi-orbicular, about 1 mm. long, ciliate; petals oblong, entire, 3 mm. long and 2 mm, wide; disc concave, mem- branous, the lobes obscure, mucronate, the margins confluent with the base of the filaments; stamens about 2 mm. long, the filaments short, flat, broad at the base, gradually attenuated toward the top, glabrous, the anthers ovoid, slightly cordate and apiculate; sterile pistil conoid, 1.5 mm. long. Female flowers: calyx lobes and petals as in the male; disc cup-shaped, membranous, the lobes subreniform; sterile stamens 0.8 mm. long; pistil 2 mm. long, the ovary subglobose, the style short, the stigmata 3-lobed, each bifid. Fruits subglobose, the valves suborbicular or obovate, about 9 mm. long and 7-8 mm, wide; seeds ellipsoid, about 4.5 mm. long and 3 mm. wide, reddish-brown, distinctly areolate. In sunny forest; at altitudes 30-900 m.; Fiji Isl.; flowering from December to May. Fiji: fulanga: Smith 1122 (GH, NY, S, UC, US), vm levu: Lautoka, Greenwood OOO (A) ; Mba, Smith 4301 (A, US), 5827, 6085 (A) ; Tholo, Degener 14798 (NY, US), 14.940, 14973 (A, NY, US), 14978 (A, MICH, NY), 13050, 153*5 (A, MICH, NY, S, US) ; Gillespie 4182 (GH, NY, UC) ; Nandi, Degener 15330 (A, NY, S, US) ; Ra, Degener 15352 (A, NY), 15424 (A, MICH, NY, S, US), 15427 (A, NY), without precise locality: Home 1135 (GH); Mrs. Parham 3 (GH); Wilkes Exped. s.n., 1838-42 (GH, holotype) . 238 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 This is an endemic species of Fiji, Its geographical distribution and accrescent pedicels make this species easily separable from other species. A specimen collected by Degener (15330) in an "arid patch of forest along coast" near Nandi has obovoid fruits while all others have subgloboid ones. 5. Celastrus madagascariensis Loes. in Notiz. Bot. Gart. Berlin 13:215. 1936, ex char. (T.: Perrier de la Bat hie 2067). Scandent shrubs; branches terete, glabrescent, brown, the lenticels elliptic or orbicular, slightly elevated; axillary buds globose to deltoid, about 1 mm. long. Leaves usually elliptic to ovate-oblong, the apex acute to shortly acuminate, the base cuneate or obtuse, the margins crenate-serrulate, 4-7 cm. long, 2.5—3.5 cm. wide, chartaceous to firmly membranous, glabrous, the primary lateral veins 5-7 pairs, elevated on both surfaces, curved toward the apex, the veinlets distinct and slightly elevated on both surfaces; stipules laciniate, about 2 mm. long; petioles 5-11 mm. long. Inflorescences terminal, paniculiform, once compound, 1-3 cm. long, the peduncle puberulous, the primary peduncles 2-3 mm. long; flowers dioecious, the pedicels about 2.5 mm. long, the articulation toward the base of the stalk. Male flowers: calyx lobes imbricate, ovate, obtuse, slightly erose, about 1 mm. long; petals elliptic to ovate, obtuse and subentire, about 3 mm. long and 2 mm. wide; disc membranous, the lobes inconspicuous, cuspidate; stamens arising from the margin of the disc, about 3 mm. long, the filaments shorter than the anthers, glabrous, the anthers ellipsoid, distinctly apiculate; sterile pistil columnar, about 1 mm. long. Female flowers: calyx lobes, petals and disc as in the male; sterile stamens about 1 mm. long; pistil about 2.5 mm. long, the ovary globose, the style shortly columnar, the stigma discoid. Fruit unknown. In forests; at altitudes 100-1,500 m.; endemic in Madagascar; flowering from October to November. Madagascar: Valley of the Mandrare, d'Alleizette 1288 (L) ; Basaltic near Betafo, d'Alleizette s.n., Sept. 1905 (L). This species quite clearly belongs to the series paniculati by its distinct terminal paniculiform inflorescences. It is closely related to the Australian Celastrus subspicatus, rather than to the Asiatic C. paniculatus, by its short and once compound inflorescences. It can be distinguished from those two species by its apiculate anthers, cuspidate disc-lobes, and discoid stigma. Loesener has pointed out that the Celastrus species of Madagascar enumerated by Drake 29 are Gymno- sporia. The excellent illustrations reveal that Loesener's view is quite right. Thus Celastrus madagascariensis is the only remaining Celastrus species described from Madagascar. 2'J Drake del Castillo, M. E., in Grandidier's Hist., Phys., Nat. et Pol. Madag. 36 (Hist. Nat. Pi. Atlas y):ph. 280, 280*, 280*. 1896. 1956] HOU REVISION OF THE GENUS CELASTRUS 239 Fig. 4. Celastrus angulatus Maxim 6. Celastrus angulatus Maxim, in Bull. Acad. Sci. St. Petersb. Ill, 27:455. 1881 (as angulata); in Mel. Biol. Acad. St. Petersb. 11:199. 1881, ex char. (T.: Piaseski s.n.). Celastrus latifolim Hemsl. in Jour. Linn. Soc. 23:123. 1886; Oliver, in Hook. Icon. Pi. t. 2206. 1894, ex char. & ill. (T.: Henry j. /*.). Scandent shrubs up to 10 m. tall; branches subterete, the branchlets usually angular, glabrous, both reddish to dark brown, shining, densely lenticellate, the lenticels small, orbicular or ovate, white; axillary buds short-ovoid, about 2-5 mm. long. Leaves broadly elliptic, broadly ovate to suborbicular, the apex abruptly margins cm. pubes primary lateral veins usually 6-7 pairs, prominently elevated below, plane or slightly elevated above, the secondary lateral veins slightly elevated below, plane and visible above, nearly parallel to one another; stipules filiform, tufted; petioles 0.6-3.0 cm. long. Inflorescences terminal, densely paniculiform, up to 20 cm. berulo ped 240 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 tion just beneath the flower. Male flowers: calyx lobes open, ovate or oblong-ovate, obtuse, subentire, about 1.5 mm. long; petals elliptic-oblong, obtuse, ciliate to slightly erose, reflexed, about 2.5 mm. long and 1.2 mm. wide; disc subfleshy, nearly flat, plate-like, the lobes usually deprcssed-subquadrate; stamens attached just under the margin of the disc, about 3 mm. long, the filaments filiform, glabrous, the anthers ovoid, obtuse, apiculate; sterile pistil ovoid, about 1 mm. long. Female flowers: calyx lobes, petals and disc as in the male; sterile stamens about 1 mm. long; pistil about 3—4 mm. long, the ovary subglobose, the style distinctly columnar, the stigmata 3-lobed, each bifid, recurved. Fruits subglobose, the valves broadly elliptic to suborbicular, about 6—9 mm. long and 5-9 mm. wide, 3- to 6- seeded; seeds ellipsoid, about 3-5 mm. long and 1.5-3.0 mm. wide, brownish-red, shining, the areolae usually obscure, rarely distinct. In open places, light woods and thickets; at altitudes of 400-3,600 m.; China; flowering from May to June. China: honan: Hia Hsien, Hers 1831 (A); Hwei Hsien, Hers 772 (A); Lushih, Hiung-eul Shan, Hers 861 (A); Sung Hsien, Hers 1288, 1 314 (A); Teng-feng Hsien, Hers 26Q (A) ; Tsi-yuan Hsien, Hers 1712 (A) ; Yungning, Hers 417 (A), hupei: Chang- yang, Wilson 753A (A); Enshih Hsien, Ho-ch'ang Chow 200 5 (A, NY); Fang Hsien, Wilson 364% 7 S3 (A, F, GH, MO, US) ; Ho-long-teng, Silvestri 5099 (A); Ichang, Chun tf Chien 8050 (UC), Henry 2084, 3405, 3883 (GH), 5925 (A, GH, NY), Wilson 364 (A, F, MO, US); Ma-pan-scian, Silvestri 1336 (A); Nanto, Wilson 783 (A); Paokan, Wilson 467 (A, NY); Patung Hsien, Ho-ch'ang Chow 495, 597 (A, NY); Sin-yeh-sie, Chun 3724 (A), 4066 (US) ; Tzu-kwei Hsien, Ho-ch'ang Chow 405 (A, NY), kansu: Tienschui, Fenzel 2817 (A), kwangsi: Ch'uan Hsien, Tsang 27666 (US), kwangtung: Loh-ch'ang, Tsang 20813 (A, MO, NY, SING, UC), Tso 21152 (NY), kweichow: Kiangkou, Steward et al 541 (A, F, NY, S, US) ; Kweiyang, Tsiang 8653 (A, NY) ; Nganping, Handel-Mazzetti 132 (A) ; Tsingchen, Teng 90315 (A) ; Tsunyi, Steward et al 97 (A, F, NY, S, US); Tuhshan, Tsiang 699 s (A, NY); Tungtze, Tsiang 4945, 4970 (A, NY), shensi: Hua-Shan, Hers 3097 (A); Shih Tan, Hers 2442 (A); Tai-pei Shan, Purdom 946 (A, US) , Fenzel 7Juna. Buncombe, Iredell, and Mitchell counties, north Dakota: Pembina, Ramsey, and Stark [Vol. 42 242 ANNALS OF THE MISSOURI BOTANICAL GARDEN counties, ohio: Clinton, Cuyahoga, Gallia, Hamilton, Harrison, Jackson, Lawrence, Meigs, Richland, Sciota, and Wood counties. Oklahoma: Blaine, Cleveland, Johnston, Kay, Murray, Payne, and Woodward counties. Pennsylvania: Adams, Allegheny, Beaver, Bedford, Berks, Blair, Bucks, Cambria, Carbon, Center, Chester, Clarion, Clearfield, Clinton, Crawford, Cumberland, Dauphin, Delaware, Erie, Fulton, Huntingdon, Indiana, Jefferson, Juniata, Lancaster, Lebanon, Lehigh, Luzerne, Lycoming, Monroe, Montgomery, Montour, Northampton, Northumberland, Perry, Philadelphia, Pike, Schuylkill, Snyder, Somerset, Sullivan, Susquehanna, Tioga, Union, Venango, Wyoming, Westmoreland, and York counties, rhode island: Providence County, south Dakota: Brookings, Brule, Grant, Lawrence, and Pennington counties. Tennessee: Crockett, Dyer, Gibson, Hamb- len, Obion, and White counties, texas: Armstrong, Brewster, Culberson, Jeff Davis, and Sutton counties. Virginia: Bedford, Caroline, Craig, Fauquier, Giles, Roanoke, Shenan- doah, Surry, and Isle of Wight counties. Vermont: Caledonia and Rutland counties. west Virginia: Cabell, Marshall, Summers, and Wyoming counties. Wisconsin: Brown, Buffalo, Door, Iowa, Kenosha, Lafayette, Outagamie, Polk, and Walworth counties. Wyoming: Crook County. This is the only widely distributed Celastrus species in North America. It is also commonly cultivated as an ornamental plant in the United States. This species is easily recognized by its paniculiform inflorescences, open calyx lobes, cup- shaped disc, and especially by the crimson arillate seeds of the opening fruits. Since it is widely distributed and the only commonly known species of Celastrus in North America, it will not be confused in the field or in the herbarium. I have seen many specimens of this species all of which are identified correctly. In the present treatment, I have not cited individual collections, but just listed the counties where they are found. All of the collectors and their field numbers can be found in the Index to Exsiccatae at the end of this paper. The variations of morphological characters and geographical distribution of this species have been discussed under the section on general morphology. Series 2. axillares Rehd. & Wils. in Sarg. PL Wils. 2:355. 1915. Sempervirentes Maxim, ex Rehd. & Wils. in Sarg. PL Wils. 2:357. 1915. KEY TO THE SPECIES A. Flowering branches terminal or rarely in some species those of the female plants intercalary. B. Flowering branches terminal both in male and in female plants; inflorescences both lateral and terminal. C. Evergreen species, the current season's branches usually subtended by foliage leaves; inflorescences usually without accompanying vegetative buds (except some specimens of Celastrus hindsii); fruits 1 -seeded. D. Young branches with distinct lenticels; flowers numerous in dense umbelli- form cymes; discs subfleshy or fleshy, flat; leaves not densely reticulate. E. Fruit conspicuously stipitate; seeds about 15 mm. long, 8 mm. wide. India, Burma, Siam, Indo-China, southwestern and southeastern China. 8. C. monospermus EE. Fruit not stipitate; seeds about 9 mm. long, 5 mm. wide. Malaya, Indo- nesia, Philippines and New Guinea 9. C. tnonospermoides DD. Young branches rarely with lenticels; flowers few in simple lax cymes, or solitary; discs membranous, cup-shaped; leaves densely reticulate; seeds about 6 mm. long, 5 mm. wide. Southeastern Asia 10. C. hindsii CC. Deciduous species, the current season's branches not subtended by foliage leaves; inflorescences with accompanying vegetative buds; fruits 3- to 6-seeded. F. Flowers fasciculate or solitary, without distinct peduncles or with very short ones on the terminal inflorescences; calyx lobes imbricate; petals thin and 1955] HOU REVISION OF THE GENUS CELASTRUS 243 with distinct venation, 2.5-3.0 mm. long, 1.0-1.5 mm. wide; anthers of the * male flowers 0.7-1.0 mm. long; fruiting inflorescences erect, the stalks of fruits not accrescent after anthesis, about 1.0—1.5 cm. long; seeds ellipsoid or ovoid. G. Terminal inflorescences usually 2-6 cm. long; axillary inflorescences usually several- to many-flowered, fruiting frequently. H. Veinlets of leaves usually immersed above and elevated beneath; seeds smooth. I. Anthers of male flowers not apiculate; disc lobes subquadrate to ob- long, obtuse or slightly emarginate; leaves usually paler beneath. Southwestern and central China and Indo-China 11. C. glaucophyllus II. Anthers of male flowers apiculate; disc lobes subreniform, mucronate; leaves concolorous. India, Burma, and China 12. C. hookeri HH. Veinlets of leaves elevated on both surfaces; seeds highly rugose. India. 13. C. membranifolius GG. Terminal inflorescences usually 6-18 cm. long, axillary inflorescences few- flowered or suppressed, fruiting infrequently. Southwestern China 14. C. vanioti FF. Flowers in distinctly pedunculate dichasia on the lower two-thirds of the terminal inflorescences; calyx lobes valvate; petals relatively thicker and without distinct venation, about 4 mm. long and 2 mm. wide; anthers of the male flowers 1.5 mm. long; fruiting inflorescences pendulous, the stalks of the fruits accrescent after anthesis, about 2-4 cm. long; seeds slightly lunate. Central and western China 15. C. hypoleucus BB. Flowering branches terminal in male plants only; inflorescences both lateral and terminal in male plants, lateral only in female plants. J. Inflorescences dichasial, conspicuously pedunculate; pedicels of fruits green. K. Vegetative buds relatively large, ovoid, acuminate, 5-1 1 mm. long; leaves usually broadly ovate, rarely broadly elliptic, the secondary veinlets densely reticulate and elevated; anthers apiculate. China south of the Yangtze Rj ver 16. C. gemmatus KK. Vegetative buds relatively small, conoid, obtuse, 1-3 mm. long; leaves usually broadly elliptic, or broadly obovate to nearly orbicular, the secondary vein- lets slightly reticulate and elevated; anthers not apiculate. China north of the Yangtze River, Korea, and Japan - 17. C. orbiculatus JJ. Inflorescences usually fasciculate or subracemiform, usually sessile, rarely very shortly pedunculate; pedicels of fruits brown. L. Disc lobes broader than long, about one-fourth to one-fifth as long as the disc proper, the articulation usually at the middle or lower half of the stalk. China south of the Yangtze River - IS C rosthornianus LL. Disc lobes longer than broad, about half to one-third as long "the. dl * c proper, the articulation usually at the upper third or upper half of the stalk. Southeastern China, Formosa, southern Japan, and Ryukyu Islands 19. C. punctatus AA. Flowering branches usually distinctly intercalary. M. Flowers usually in pedunculate cymes, axillary or extra-axillary; leaf margins ser- rate, the petioles less than half as long as the blades. N. Seeds slightly lunate, attenuate at both ends; sepals usually erose or glandular; secondary veinlets of leaves conspicuous or elevated on both surfaces. O. Leaves usually broadly elliptic to orbicular; disc subfleshy, flat or slightly concave. # ., , P. Plants glabrous to sliehtly pubescent on youne flowering branchlcts and the leaves beneath; leaves usuallv suborbicular or orbicular: disc lobes inconspicuous. China: Kwangsi, Kwangtung, Hainan, and Formosa 20. C. kusanos PP. Plants densely yellowish brown-pubescent, especially on both surfaces of the leaves; leaves usually broadly ovate or broadly obovate; disc lobes con- spicuous, arcuate. Burma, China, and Indo-China t' 2l 'j ^rsutus OO. Leaves usuallv elliptic or elliotic-oblong; disc membranous, cup-shaned. Eastern India,' Indonesia, and China 22 ' C * 5 '> / 79776 (A); Chen Kang, Yu 17495 (A); Fo-hai, Wang 73579, 73590, 73825> 73^39, 74430, 74907, 77*02* (A); Jenn-y eh Hsien, Wang 80312, 80781 (A); Lan-tsang Hsien, Wang 73415 (A); Lung-ling, Tsai 55679 (A); Lu-se, Tsai (A); 56917 (A) ; Mengtze Henry 10446 (A, NY), 10955 (A, MU), 1 1399 (A, MO, INT, between Mohei and Maokai, Rock 2900 (A, UC, US) ; Nan-chiao, Wang 75183 between Palut and Nam-dip, on the trail from Raheng to Mesawt, Rock 1024 (US); Ping-pien Hsien, Tsai 55076, 55139, 60152, 60459, 61680, 62295 (A); Shweli-Salwin, Forrest 1 1874 (UC) ; Si-chour Hsien, Feng II 445, ^53°, I2 4&9 (A); Sunning, Hila, Yu 16508 (A) ; Szemao, Henry 11972 (A, NY), 11972* (A, MO, NY), Rock 2770, 2823 (A, UC,US) ; Tsing-pian, Tsai 52477,52552 (A) ; Tung-ting, Feng 13472,13749 (A) ; Wen-shan Hsien, Tsai 51546, 51591, 72857 (A); without precise locality, Wang & Ying Liu 82917, 82947 (A). Hongkong: Chun 40143, 40276 (NY); Chun 4831 (VC),4943 (A, UC) ; Horn Fung 161 (NY); Taam 1 792 (A, US), 2087 (US); Lanto, Tsiang 720 (UC) ; Wright 9 (L), 92 (GH). _ . India: Assam: Darrang, Upendranath Kanjilal 3801 (CAL); Jawai, Dr. Kmgs col- lector s.n, (L); Khasi Hills, Kurz 182 (CAL); Kohima, Dr. King's collector 178 (U, [Vol. 42 246 ANNALS OF THE MISSOURI BOTANICAL GARDEN UPS); Konoma, Dr. King's collector 233 (A); Namchung, Luckimpore, Clarke 37949 s (CAL); Naga Hills, Jagarmani 499 (A, US); Dr. King's collector 795 (A, CAL), 1082 (CAL); Silhet (as Sillet), Desiloa s.n. (NY, type of Celastrus monospermus) ; Hooker & Thomson s. n. (CAL, GH, L, S) ; Khasia, Liam, Clarke 451 52* (CAL); without precise locality, Griffith 666 (CAL). bengal: Sinchula, Biswas 2033 (A, NY); Darjeeling, Lace 2229 (CAL) ; Darjeeling, Anderson 107 (CAL) ; Wangiet, Kurz s. n. (CAL) ; Runquo Valley, Anderson 108, 109 (CAL); Sureil, Darsteri 17 (CAL). Indo-China: Chapa, Tonkin, Petelot 3184 (UC), 5931 (A), 6368 (A, NY, US). Pakistan: Griffith 1992 (CAL, GH, L, NY, S). This species is distinguished readily by stipitate fruit bases, the one-seeded capsules, the large cylindric seeds, the characteristic fleshy annular disc, and the attachment of the stamens immediately beneath the margin of the disc. The synonymy of this species is due to the failure, first, of Bentham to appre- P Wilson variability of that species. My present interpretation is that of Lawson (in Hook. Fl. Brit. Ind. 1:618. 1875). Generally speaking, plants of Yunnan, Burma, and India have oblong or elliptic- oblong leaves, while those of Kwangtung and Kwangsi have broadly elliptic or broadly ovate leaves. At first I attempted either to maintain these plants as two species or to unite them into one species with two subspecies having differing leaf characters. Since then, I have measured the length and width and angles A (apical) and B (basal) 30 from the different areas, and found that the frequency curves of the ratio of length and width, and the angles A and B overlap. Since leaf shape is an extremely variable character and the floral structures of these plants are so similar, I feel justified in considering them as representing a single species. 9. Celastrus monospermoides Loes. in Lorentz, Nova Guinea 8:280. 1910. (T.: Versteegh 1643, L!). 7 onti No. 8:353). 1896, as to spec, cited, non Benth. Celastrus apoensis Elmer, Leafl. Philipp. Bot. 7:2579. 1915. (T.: Elmer 11411, A!). Celastrus malayensis Ridl. in Jour. Roy. Asiat. Soc, Str. Branch 75:18. 1917. (T.: Ridley 13538, SING ! ) . Scandent shrubs up to 17 m. tall; branches terete, smooth to striate, glabrous, black or reddish-brown; lenticels lacking or obscure; axillary buds conoid, about 2 mm. long. Leaves ovate to elliptic, the apex acute to acuminate, the base shortly cuneate, obtuse to rounded, the margins serrate to nearly entire, 5-16 cm. long, 2.5-7.5 cm. wide, firmly membranous, lucid and usually blackish-brown on both sides or yellowish- green below in dry condition, the primary lateral veins 4-7 pairs, curved toward the apex, conspicuous and elevated below, plane above, the veinlets distinct and slightly elevated below, immersed and obsolete above; stipules laciniate, about 1 mm. long; petioles 6-13 mm. long. Inflorescences axillary, soli- 50 Woodson, R. E., Jr. Some dynamics of leaf variation in Asclepias tuberosa. Ann. Mo. Bot. Gard. 34:353-432. 1947. 1955] HOU REVISION OF THE GENUS CELASTRUS 247 Fig. 6. Celastrus monospermoides Loes. mm tary or in clusters of three, aggregate dichasia, frequently fasciculate-racemiform, once to thrice compound, up to 10 cm. long, the peduncles glabrous, the primary peduncles obscure to 16 mm. long; flowers dioecious, white, the pedicels about 4 >art of the stalk. Male flowers: calyx lobes imbricate, suborbicular, entire, about 0.7 mm. long; petals oblong, the apex rounded, about 2 mm. long and 1 mm. wide, entire, pink-punctate; disc fleshy, plane, the lobes obscure, truncate; stamens attached slightly under the margin of the disc, about 2.5 mm. long, the filaments filiform, glabrous, the anthers ovoid and obtuse; sterile pistil conoid, 2 mm. long. Female flowers: calyx lobes, petals, and disc as in the male, sterile stamens about 0.7 mm. long; pistil flask-shaped, 1.2 mm. long, the ovary globose, the style slender, distinct, the stigmata discoid. Fruits the valves ovate or obovate to broadly mm 12 mm. angular-ovoid, obovoid to subglobose, < elliptic, about 12-17 mm. long and 7-1 1 long and 5—10 mm. wide, blackish-brown, smooth. In dense woods, at altitudes 100-3,300 m,; Borneo, Indonesia, Malaya, New Guinea, and Philippines; flowering from January to September. Borneo: western Koetai, Endert 3846 (L). British north Borneo: Kini Taki River, /. tf M. S. Clemens 31827 (A, L), 32052 (L); Marai Parai, /. £* M. S. Clemens 32860 (L), 33182 (L, UC) ; Penibukan, /. tf M. S. Clemens 32120 (L), 50253 (L, UC) ; Upper Kinabalu, /. # Af. S. Clemens 27870, 20010 (L), 50641 (A), 51295 (A, L, UC). JL. 42 248 ANNALS OF THE MISSOURI BOTANICAL GARDEN Sumatra: Gaju & Alas Land, van Steenis 8717 (L), 96 jl (A, L); old jungle near the Ack Kanopan, Loendoet concession, Koeale, Bartlett 6864 (F, MICH, NY, US) ; Laboebom Batoe, Kota Pinang, Toroes 4009 (A, NY, UC, US). Malaya: selangor: Ridley 8228 (SING), negri sembilan: Sutu, Alims 1 605, 2103 (SING), penang: Government Hill, Curtis 404 (SING), Ridley 404 (SING, para- type of Celastrus malayensis) ; Telom, Ridley 1 3538 (SING, paratype of C. malayensis) ; top of Penang Hill, Ridley 7127 (SING), perak: Batu Petch, Wray TO31 (SING, para- type of C. malayensis); Hijan, Haneeff & Jan 2466 (SING); Larut, Dr. King's collector 5690 (L, SING, UC, UPS), 6928 (L, US); Maxwell's Hill, Curtis 2005 (SING), Ridley s.n., 1890 (A); Tapa, Wray 175 (L). Singapore: Sungei Loyang, Mat s.n., 1894 ( SING ) . New Guinea: dutch new guinea: Bali, Stresemann s.n., May-Aug. 1911 (L) : s. w. of Bernhard Camp, Idenburg River, Brass 1 2990 (A, L) ; Dalman, 45 km. inward of Nabire, Kanehira & Hatnsima T2J02 (A) ; southern New Guinea, via Resi Mts. to HeJlwig Mts., Versteegh 1643 (L, type of Celastrus monospermoides) . northeast new guinea: Ogeramnang, Morobe, J. # M. S. Clemens 4581, 5084, 6253 (A), papua: Boridi, Cart 13036 (A). Philippines: Mindanao: Davao, Mt. Apo, Elmer 11411 (A, type of Celastrus apoensis 3 GH, L, MO, NY, US); Mt. McKinley, Edano 900 (A, L, SING, UC) ; BukMnon, Mt. Dumalucpihan, Ramos & Edano 38961 (A, L, US); Mt. Katanglad, Sulit 3188, 3421, 9917, 10086 (L). The wide distribution, within geographically isolated areas, and the morpho- logical variations of this species are the causes of several names. Specimens col- lected in the Malay Peninsula from lowlands up to 1,330 m. altitude have leaf forms varying from ovate to elliptic. They have thick and large leaves and long in- florescences. A specimen {van Steenis 9651, A, L) collected at altitude 3,000 m. from northern Sumatra has small and thin leaves and short inflorescences — about 1 cm. long. The type specimen of Celastrus apoensis Elmer from Mt. Apo (alt. 2,000 m.), Mindanao, Philippines, has unusually small leaves. In addition, two other speci- mens (Sulit 9917 and 10086, L) collected at altitudes 1,700-1,800 m. from Mt. Katanglad, Mindanao, have larger leaves. The shapes and textures of leaves and inflorescences are similar to those from Malaya, Borneo, and New Guinea in similar environments. There are several collections from Mt. Kinabalu, North Borneo, at altitude about 3,000 m. The leaf form is ovate with rounded base which is very distinctive and might be thought as a form of this species. In addition, there is a specimen (Endert 3846, L) collected at altitude 1,200 m. from east-central Borneo having elliptic leaves, acute at both ends. I think it may be possible, in addition to those mentioned above, to find more specimens with various leaf forms from different environments. The leaves of the New Guinea specimens are usually elliptic, acute at both ends. They are quite similar to those from other localities in similar environments. I have made some studies and measurements of leaf characters. Except those from Mt. Kinabalu, Borneo, they can hardly be separated into groups. The shapes and sizes of fruits are variable also. As far as the shapes of fruits are concerned, those of the Malayan specimens seem to be broadly elliptic while those of the Philippine specimens are obovate. However, they display intermediate forms as well. 1955] HOU REVISION OF THE GENUS CELASTRUS 249 Map 2. Distribution of seven species of Celastrus, Subgenus celastrus, Series II. AXILLARES. 10. Celastrus hindsii Benth. in Hook. Jour. Bot. 3:334. 1851, ex char. (T.: Hinds s. n. ) . Catha monosperma (Roxb.) Benth. in Hook. Lond. Jour. Bot. 1:483. 1842, quoad descr. & spec, cit., non sensu Roxb. Celastrus venulosus Wall. Cat. no. 4321. 1831, nom. nud. Celastrus racemulosus Hassk. Hort. Bogor. Descr. 1:1915. 1858, ex cnar. Flueggea serrata Miq. Fl. Ind. Bat. 1:356. 1859. (T.: Junghuhn s.n., L!). Celastrus monosperma sensu Benth. Fl. Hongk. 63. 1861, non Roxb. Celastrus cantonensis Hance, in Jour. Bot. 23:323. 1885. (T.: Hance 22I 9 I, A, a photo and a fragment of a leaf!). , 4A1 . tn-467 Ofcsrras Wwi Benth. var. W>i Loes. in Engl. Bot. Jahrb. 29:444. 1901, 30.467. 1902. (T.: Henry 3495, A!). [Vol. 42 250 ANNALS OF THE MISSOURI BOTANICAL GARDEN Celastrus oblongifolia Hay. Icon. PL Formos. 3:58. 1913. (T.: Hayata s.n., A!). Celastrus marianensis Koidz. in Bot. Mag. Tokyo 30:400. 1915, ex char. (T.: Koidzumi Celastrus axillaris Ridley, in Jour. Roy. Asiat. Soc. Malay. Branch, 1:56. 1923, ex char. Celastrus approximata Craib, in Kew Bull. Misc. Inf. 1926:349. 1926. (T.: Kerr 9941, SING!). Celastrus merrillii Tardieu, in Bull. Soc. Bot. France 95:180. 1948; in FL Gen. l'lndo- Chine Suppl. 1:805. fig. 98, t. 1,2 & J. 1948. (T.: Poilane 28/14, P!). Fig. 7. Celastrus hindsii Benth Scandent shrubs up to 19 m. tall; branches terete, glabrous, brown, rarely lenticelled, the branchlets smooth and green, the lenticels usually lacking on the current season's growth; axillary buds deltoid, about 1 mm. long. Leaves elliptic- oblong, obovate-oblong, elliptic or rarely broadly elliptic, the apex abruptly acute to acute, the base cuneate to obtuse, the margins remotely and obscurely serrate to serrate, 4.0-14.5 cm. long, 1.5-6.0 cm. wide, coriaceous, shining, glabrous, the primary lateral veins 6-9 pairs, the veins and veinlets usually distinctly elevated and densely reticulated on both surfaces, rarely slightly elevated or obsolete above; stipules laciniate, about 1 mm. long; petioles about 0.5-1.0 cm. long. Inflorescences usually axillary as well as terminal, sometimes axillary only, solitary, usually 3- to 5 -flowered, or sometimes racemiform and longer than the subtending leaf, the 1955] HOU REVISION OF THE GENUS CELASTRUS 251 peduncles glabrous, the primary peduncles almost obsolete to 5 mm. long, rarely up to 25 mm. long; flowers dioecious, white to pale yellow, the pedicels almost obsolete to 5 mm. long, the articulation on the upper half of the stalk. Male flowers: calyx lobes semi-orbiculate, imbricate, obtuse, ciliate, about 1.5 mm. long; petals oblong to obovate, obtuse, glandular-ciliate, 2-5 mm. long and 1.3-2.5 mm. wide; disc cup-shaped, the lobes usually deltoid, rarely slightly oblong, shortly acute, rounded, or slightly dentate; stamens arising from the margin of the disc proper, about 2.3 mm. long, the filaments subulate, glabrous, the anthers ovoid, obtuse, cordate; sterile pistil ovoid, about 1.5 mm. long. Female flowers: calyx lobes, petals, and disc as in the male, but the petals smaller; sterile stamens about 1.5 mm. long, the ovary subglobose, the style columnar, the stigmata 3-lobed. Fruits ovoid to subglobose, the valves broadly ovate to suborbicular, about 8-11 mm. long and 7-9 mm. wide, usually 1 -seeded; seeds cylindric, 5-8 mm. long and 5 mm. wide, reddish-pink, smooth. Chiefly in thickets, at altitudes from 232-1,800 m.; Bonin Islands, Burma, China, India, Indonesia, Indo-China, and Siam; flowering from January to October. British North Borneo: Penibukan, /. 6 M. S. Clemens 30493 (A, L, UC). — _„ — — ^ ^ _ — J — Lap., Toppin 6154 (CAL). China: fukien: Amoy 4292 485 1002 5601 (A); Changchow, Chung 1174, 1206 (A, UC) ; Chuanchow, Chung 1084 (UC) ; Foochow and vicinity, Sin Ging Tang 7039 (A), 7 17 5 (MICH); Kuliang and vicinity, Sin Ging Tang 6991 (MO, UC). Hainan: Bak Sa, Lau 26227 (A); Ching-mai Hsien, Lei 129 (L, NY, SING, UC, US), 440 (NY, SING, UC, US); Hung-mo Shan and vicinity, Lai area, Tsang & Fung 241 (A, US) ; Nodoa, McClure s. n. (A, MO) ; Yaichow, How 70368 (A, F, NY) ; without precise locality, Levine s. ». (A, F, MO, US), 484 (A, F, GH, MO), hupei: Chang-lo Hsien, Wilson 561 A (A); Enshih Hsien, Ho-Cb'ang Chow 1944 (A, NY); Ichang, Wilson 561 (A, F, GH, MO, US); Patung, Ho-Ch'ang Chow 548 (A), 705 (A, NY), Wilson 450 (A, NY); without precise locality, Henry 3241 (NY, type of Celastrus hindsii var. benryi), 3495 (A, GH, US), 3856, 349S A (GH), v. Kosthorn 9 (A), kwangtung: Au-tsai, McClure 3544 (NY) ; Canton and vicinity, Hance 22191 (A, photo of type of Celastrus cantonensis, with a piece of leaf), Levine s.n. (C.C.C. nos. 180—A, UC; 44 I ~^ F '_ GH « MO _i US; 1236— A, GH, MO, US), Merrill IOO70 (UC), Keng-p. wang, Kang-ping To et al 889 (US); Kochow, Tsiang Hsien, Tsang 21422 (A) ; Pok-Lor, Horn Fung A-511 (NY); Mn-tung risien, 1 aam 507, 759 (A) ; Sun-wui Hsien, Tso tf Tsiang 2024 (A, F) ; Tai-o, Wang 3196 (NY) ; Tseng- shing, Tsang 20209 (NY); Tsung-hwa Hsien, Tsang 20604 (MO, NY), 25019 (A); Wung-yuen, Lau 693 (A, NY, SING), 2672 (A); Ying-tak Hsien, Tsui 391 (A, MO, NY, UC, US), Tsiang 1919 (SING, UC), Tsang tf Kam-chow Wong 2661 (UC) ; Wung- yuen, Lau 693 (A, SING), kweichow: without precise locality, Cavalerte 3976 (A). szechuan: Kiating Fu, Wilson 3324 (A); Omei Shan, Chiao & Far, 1 458, 8639,11712, l 49 8l y 15221 (A); Hu 7380 (A), Lee 4551 (A), Wilson 2307 (A, US) <, Wang 8096 (A). t AI wan: Aderu, Heito, Suzuki 11162 (TAI) ; Arisan, Faurie 1376 (A, TAI)^ Hayata s n (A, photo of type of Celastrus oblongifolius). yunnan: Che-h Hs.en, Wang 7584b, 79467, 79470 (A) ; Chen-pien Hsien, Ko 56089 (A) ; Jenn-yeh Hsien, Wang 80136 80181, 80232 (A); Lu-se, Tsai 56342 (A); Luh-shuen Hsien, Wang 8ll 7 7 (A); Mar-h-po, Tung-ting, Feng 13068, 13210, 13387, 13454, 134*4, 13*47 (A) ; Mengtze, Henry 10559 (A, MO, NY, US), Tsai 5*375 (A); Si-chour Hsien, Feng Il6S 5> {^62 (A)- Szen^o, Henry 11972* (US); Tsing-pien, Tsai 5*545 (A), monokono: Chun 4840 ■ 59824 (A). Indo-China: Chapa, Tonkin, Petelot 5941 (A, NY). 12. Celastrus hookeri Prain, in Jour. Asiat. Soc. Bengal 73:179. 1904; Novic. Ind. 418. 1905. (T.: Hooker s. n., CAL!). Wall W Suppl. 1:806. W Scandent shrubs up to 10 m. tall; branches terete to slightly striate, glabrous, brown to reddish-brown, the lenticels sparse or lacking on the current year's growth, suborbiculate to elliptic; axillary buds conoid, 2-4 mm. long, the bud 1955] HOU REVISION OF THE GENUS CELASTRUS 255 scales accrescent, persistent. Leaves broadly elliptic to ovate, the apex acute, the base obtuse, the margins serrate, 6-12 cm. long, 4-7 cm. wide, thin-membranous on flowering branches, glabrous, the primary lateral veins 5-6 pairs, slightly ele- vated below, immersed and distinct above, the veinlets visible below, obscure above, relatively thick-membranous on fruiting branches; stipules laciniate, about 1.5 mm. long; petioles 8-15 mm. long. Inflorescences axillary as well as terminal, the axillary ones 3- to 5 -flowered, the primary peduncles glabrous, about 1.5-3.5 mm. long, the terminal inflorescences very short-pedunculate dichasia, fasciculate or solitary, about 4 cm. long, the axils of the peduncles associated with vegetative buds (except the uppermost) ; flowers dioecious, green, the pedicels 1.5-2.4 mm. long, the artic- ulation on the upper half of the stalk. Male flowers: calyx lobes deltoid, imbricate, cilia te, about 1.5 mm. long; petals oblong to obovate, slightly glandular on the margins, about 3.5 mm. long and 1.5 mm. wide; disc cup-shaped, the lobes sub- reniform; stamens arising from the margin of the disc proper, about 2 mm. long, the filaments linear, glabrous, the anthers apiculate; sterile pistil 1.5 mm. long. Female flowers: calyx lobes, petals, and disc as in the male, the sterile stamens 1.5 mm. long, the pistil 4 mm. long, the ovary globose, narrowed into a distinct style, the stigmata 3-lobed, each lobe bifid, linear. Fruits subglobose, the valves broadly elliptic, about 10 mm. long and 7 mm. wide, 3- to 6-seeded; seeds ellipsoid or ovoid, about 4 mm. long and 2 mm. wide, black, the areolae distinct. Chiefly in thickets, at altitudes from 1,525 to 3,050 m.; Burma, India, China, and Pakistan; flowering from March to May. Burma: Adung Valley, Kingdon Ward 9455 (A); Bhamo, Lapyeka to Sinlum Kaba, Lace 5775 (CAL); without precise locality, Prazer s. n., 1890 (CAL), Lace 5753 (CAL). China: yunnan: between Chienchuan Plain and the Mekong drainage basin to Lachining, Rock 8615 (A, UC, US); Wei-se Hsien, Tsai 59901 (A); without precise locality, Yu 8491 (A). f , ' India: assam: Khasia, Griffith 605 (CAL). bengal: Darjeeling, Clarke 6#2^ 9 27046®, 35758* (CAL), Gamble 1926^ (CAL); Jalubaham, King s.n., May 9, 1876 (CAL). sikkim: Lachung, Gammie 1179 (CAL) ; Lachen, Dr. King's collector s. n May 1885 (CAL); Lebong, Kurz s.n. (CAL); Pauree, Dr. Prain's collector 289 (CAL); Lachen, Clarke 46569 (CAL). without precise locality: temp, region, 8,000-10,000 ft., Hooker s. n. (type of Celastrus hookeri, CAL), Kurz s. n. (CAL). Pakistan: Griffith 1993 (CAL, GH). Rehder and Wilson 81 have cited two Indian specimens, Hooker d Thomson, "Khasia, alt. 4-5,000 ped." and C. G. Rogers, "Sikkim, January 1900", as Celastrus hookeri Prain. Wall. As it was stated by be filaments in Celastrus stylosus, but glabrous ones in Hooker's species". In addition, the intercalary inflorescences and slightly lunate seeds of Celastrus stylosus are termina 13. Celastrus membranifolius Prain, in Jour. Asiat. Soc. Bengal 73 : 197. 1904; Novic. Ind. 418. 1905. (T.: Mann s. n., CAL!). Branches terete, the branchlets slightly angular on the dry specimens, glabrous, :il Rehder, A. t and Wilson, E. H., in Sarg. PL Wils. 2:352. 1915 [Vol. 42 256 ANNALS OF THE MISSOURI BOTANICAL GARDEN the lenticels scattered, ovate or orbiculate; axillary buds conoid, acute, about 1.5 apex the margins serrate, 7.5-14.0 cm. long, 3.5-5.5 cm. wide, membranous, pale green in dry condition, glabrous, the primary veins 4-5 pairs, elevated below, immersed or slightly elevated above, the veinlets prominent on both surfaces; petioles 1.0-1.5 cm. long. Flowers unseen. xillary short-pedunculate or fasciculate, associated with axillary buds, the stalks 7-15 mm. long, the articulation on the upper half of the stalk. Fruits subglobose, the valves broadly elliptic, about 7 mm. long and 5 mm. wide, 3- to 6-seeded; seeds elliptic, (immature?) mm India: assam: Khasia and Junteah Hills, Mann s.n., 1877 (CAL, type; MO); Khasia Hills and Brahmaputra plains, Kurz s. n. (CAL). The few specimens available are fruiting specimens in which the seeds are highly wrinkled, indicating that they might be immature. However, I have seen no immature seeds with such wrinkling in any other species of Celastrus. 14. Celastrus vanioti (Levi.) Rehd. in Jour. Arn. Arb. 14:249. 1933. Saurauja vaniota Levi. Fl. Kouy-Tcheou, 415. 1915. (T.: Bodinier s. n., A!). Celastrus spiciformis Rehd. & Wils. in Sarg. PI. Wils. 2:348. 1915. (T.: Wilson 2U2, A!). astrus spicif 1915. (T.: Wilson u?6, A!). Scandent shrubs up to 10 m. tall; branches terete or slightly striate, glabrous, light brown to reddish or fuscous, the lenticels scattered, orbicular or elliptic, some- times lacking on the current year's growth, slightly elevated on the older branches; axillary buds globose, about 2 mm. long. Leaves ovate or elliptic, the apex acute acuminate serrate, the teeth incurved, glandular-mucronate, 5-13 cm. long, 3.5-7.5 cm. wide, membranous, glabrous or slightly puberulous on the veins below, the primary lateral veins 6-7 pairs, slightly elevated on both surfaces, the secondary lateral veins parallel; stipules laciniate, filiform, about 1.5 mm. long; petioles 1-2 cm. long. Inflorescences axillary as well as terminal, shortly pedunculate or fasciculate, the peduncles almost obsolete to 5 mm. long, the axillary inflorescences dichasia or very sometimes axillary acemiform articulation at the middle or on the lower half of the stalk. Male flowers: calyx lobes deltoid, imbricate, obtuse, glandular-ciliate, about 1 mm. long; petals oblong, rotund, slightly erose, about 3 mm. long and 1.5 mm. wide; disc cup-shaped, membranous, the lobes acute; stamens arising from the margin of the disc proper, about 2.5 mm. long, the filaments linear, glabrous, the anthers ovoid, obtuse, slightly apiculate; sterile pistil conic, about 1 mm. long. Female flowers: calyx 1955] HOU REVISION OF THE GENUS CELASTRUS 257 lobes, petals, and disc as in the male; sterile stamens 1 mm. long; pistil 3 mm. long, the ovary ellipsoid, gradually attenuate into the style, the stigmata 3-lobed and spreading. Fruits subglobose, the valves broadly ovate, about 9 mm. long and 8 mm. wide, 3- to 6-seeded; seeds ellipsoid, obtuse at both ends, areolae distinct, about 4 mm. long and 2 mm. wide, black. In mixed woods, at altitudes from 500 to 2,000 m.; China; flowering from May to July. China: hunan: Wukang, Te-hui Wang 1 13 (A), hupei: Fang Hsien, Wilson 2215 (A, NY); Heing-shan Hsien, Wilson 2312 (A, type of Celastrus spiciformis; GH, MO) without precise locality, Henry 5935 (A, GH, US), Wilson 2215* (NY), kweichow Bin Long, Miu Shan, Luchen, Ching 6057 (NY); Tuhshan, Tsiang 6655, 6993 (NY), without precise locality, Bodinier 2287 (A, type of Saurauja vanioti). szechuan: Loo- shan Hsien, Wang 23582 (A) ; Ma-pien Hsien, Wang 23WI (A) ; Omei Shan, Chow 7094 (A), Tang 18958 (A), Sun tf Chang 149, 286 (A), Wang 8035 (A, US); Wa-shan, Wilson 1 176 (A, holotype of Celastrus spiciformis var. laevis), 1 38 3 (A, NY), yunnan: Menetze. Henry IIO06 (A, MO, NY): Tone Tch'ouan, Maire 6227 (UC); Wcn-shan 52090 Wil somewhat intermediate in character between C. angnlata Maximowicz and C. hypoleuca Warburg". It appears to be a hybrid between those two species. Super- ficially, the type of C. spiciformis is similar to C. angulatus, except for the in- florescences which are both terminal and axillary. On examination, the floral structures were found to be similar to those of Celastrus hypoleucus. 15. Celastrus hypoleucus (Oliv.) Warb. apud Loes. in Engl. Bot. Jahrb. 29:445. 1900 (as hypoleuca). Erythrospermum hypoleucum Oliv. in Hook. Icon. Ill, 9:/. 1899. 1889, ex char. & ill. (T.: Henry 5877). Celastrus hypoglaucus Hemsl. in Ann. Bot. 9:150. 1895. (T.: Henry 2837, GH!). Celastrus hypoleucus forma a genuina Loes. in Diels, in Engl. Bot. Jahrb. 29:445. 1900 (based on Erythrospermum hypoleucum Oliv.). Celastrus hypoleucus forma P argutior Loes. loc. cit. 1900. (T.: Henry 6771, A!). Scandent shrubs 3-5 m. tall; branches slightly striate, glabrous, darkish brown, the lenticels sparse to dense, rarely lacking on the current year's growth; axillary buds orbicular, about 2 mm. in diameter. Leaves elliptic to elliptic-oblong, the apex acute, the base obtuse, the margins remotely serrate to serrulate, 3.0-9.5 cm. long, 2.0-5.5 cm. wide, usually delicately membranaceous, glabrous, sometimes puberulous on the veins below, glaucous below, pallid above, the primary lateral veins 5-6 pairs, slightly elevated below, immersed above, the veinlets prominent below, obscure above; stipules filiform, about 2 mm. long; petioles 1-2 cm. long. Inflorescences axillary as well as terminal, the terminal portion usually up to 1 cm. long, usually short-pedunculate, the peduncles glabrous, accrescent, the mm the pedicels accrescent, about 2-8 mm. long, the articulation at the upper half of the stalk, after anthesis usually only the flowers of the terminal inflorescences maturing into fruits, and the stalk accrescent up to 4 cm. long, spreading and 258 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 pendulous, Male flowers: calyx lobes valvate, deltoid to oblong, subentire to slightly erose, about 1.5 mm. long; petals oblong or rarely elliptic, usually entire, rarely obscurely erose, about 4.3 mm. long and 2 mm. wide; disc cup-shaped, the lobes obscure, truncate to mucronate; stamens arising from the margin of the disc proper, about 4 mm. long, the filaments filiform, glabrous, about 3 mm. long, the anthers ovoid, slightly apiculate, about 1.5 mm. long; sterile pistil about 2 mm. long. Female flowers: calyx lobes, petals, and disc as in the male; sterile stamens about 1.5 mm. long; pistil flask-shaped, about 3 mm. long, the ovary ellipsoid, the style cylindric, distinct, the stigmata 3-lobed, flat, spreading. Fruits subglobose, the valves broadly elliptic or suborbicular, about 8-12 mm. long and 8—10 mm. wide, irregularly pink-punctate within, 3- to 6-seeded; seeds more or less plano- convex to slightly lunate, attenuate at both ends, about 6 mm. long and 2 mm. wide, black-brown, the areolae distinct. Chiefly in thickets, at altitudes from 1,700 to 2,745 m.; China; flowering from June to July. China: anhvei: Wang Shan, Ching 3065 (UC). hupei: Fang Hsien, Wilson 362 (A, F, GH, MO, US); Hsun-tien-tsze, Chun 4427 (A); Ichang, Wilson 1063 (A, NY); without precise locality, Henry 2837 (GH, paratype of Celastrus hypoglaucus) , 5887 (GH, NY, US), 68 1 1 (GH), 6771 (A, type of Celastrus hypoleucus forma P argutior; US), shensi: Kin-qua Shan, Giraldi s. n., July 10, 1897 (A) ; Qua-in Shan, Giraldi s. n., July 16, 1897 (A); Tai-pei Shan, Giraldi s. n., July 15-20, 1897 (A); Tsinling, Fenzel 927 (A), szechuan: Kuan Hsien, Wilson 2306 (A, US). This species is easily recognized by its long terminal pendulous racemiform dichasia of long accrescent stalked fruits and by the slightly lunate seeds. The leaves are usually glaucous below, but this character is extremely variable. 16. Celastrus gemmatus Loes. in Engl. Bot. Jahrb. 30:468. 1901 (as gemmata) . (T.: Henry 9872^ MO!). Embelia esquirolii Levi, in Fedde, Rep. Sp. Nov. 10:374. 1912. (T.: Esquirol 4, A!). Celastrus lokchongensis Masamune, in Trans. Nat. Hist. Soc. Formosa 25:15. 193 5. (T.: Tsiang 1 3 46 ', A ! ) . Scandent shrubs, 3-7 m. tall; branches terete or slightly striate, light to darkish brown, the lenticels scattered, suborbicular or oval, usually elevated, white; axillary buds conoid, acuminate, about 4-11 mm. long. Leaves usually broadly elliptic to elliptic-ovate, the apex gradually acuminate to acute, the base obtuse, rotund or truncate, the margins serrate, 5-15 cm. long, 2-8 cm. wide, firmly membranous, usually glabrous, rarely yellowish-puberulous on the veins below, the primary lateral veins 5-7 pairs, curved toward the apex, the veinlets densely retic- ulate, the veins and veinlets elevated on both surfaces; stipules filiform, tufted, about 1.5 mm. long; petioles 1.0-2.5 cm. long. Inflorescences axillary, usually 3- to 7-flowered, the peduncles glabrous, the primary peduncles 4-7 mm. long; flowers dioecious, white or yellowish-green, the pedicels 3-8 mm. long, the articu- lation at the lower half of the stalk. Male flowers: calyx lobes valvate, ovate- deltoid, obtuse, glandular-ciliate, about 1 mm. long; petals subequal, obtuse, 19S5J HOU REVISION OF THE GENUS CELASTRUS 259 Fig. 9. Celastrus gemma tus Loes slightly erose, about 3.5-4.5 mm. long and 1.5-2.0 mm. wide; stamens inserted between the disc lobes, about 3 mm. long, the filaments more or less complanate, glabrous, gradually tapered toward the apex, the anthers oblong-ellipsoid, obtuse, usually apiculate, pink-punctate; disc thin, cup-shaped, the lobes acute; sterile pistil columnar, 1.5-2.5 mm. long. Female flowers: calyx lobes and petals as in the male; sterile stamens about 1 mm. long; disc thin, cup-shaped, the lobes in- conspicuous; pistil flask-shaped, about 4 mm. long, the ovary ovoid, the style columnar, the stigmata 3-lobed, each bifid, reflexed. Fruits usually 1-3 on each dichasium, the pedicels green, the fruits subglobose, the valves broadly oval or suborbicular, 7-15 mm. long and 7-15 mm, wide, 3- to 6-seeded; seeds ellipsoid or ovoid, about 4.5 mm. long and 2.5 mm. wide, reddish-brown, smooth. Common in thickets, valleys or open slopes, at altitudes from 400 to 3,000 m.; widely distributed in China: south of the Yangtze River; flowering from April to October. China: anhwei: Chu-hwa Shan, Cbing 2638 (A, UC), 7566 (UC); Wang Shan, Ling 1159, 1246 (UC); Wu-yuen, Ling 1 349 (UC). chekiang: east Tien-mu, Hu 1 585 (A, UC). fukien: Dingschou, Tienhwa-schan, Wang 399 (A), hupei: Fang Hsien, Wilson 919 (A, NY), 2215 s - (A) ; Sin-shan Hsien, Wilson 363 (A, MO, US), 502 (A, F, [Vol. 42 260 ANNALS OF THE MISSOURI BOTANICAL GARDEN GH, MO, US); Patung Hsien, Ho-Ch'ang Chow 901 (A, NY); Wan-tsao Shan, Chun 3884 (A), 4229 (US); without precise locality, Henry 761 4 (GH). kiangsi: Hsin-feng Hsien, Hu II 19 (A); Lu-shan, Chung & Sun 592 (A, NY), Steward 1234 (A), 5470 (A, UC, US), Wilson 1 519 (A, MO, US), kiangsi et fukien: in monte Dunghwa-schan inter Schitscheng et Ninghwa, Wang 299 (A), kwangsi: Tunghua Shan, Thwang, Tsiang IOO47 (NY); Tzu-yuen Hsien, Chung 83605 (A); Huangtung, Yao-shan, Sin 11835 (NY), kwangtung: Jen-hwa Hsien, Tsang 26404 (A); Lokchong, Chun 42959 (A), Tsang 20855 (A, L, MO, NY, SING, UC, US), Tsiang 1 346 (A, isotype of C. lokchong- ensis, UC), 1419 (A, UC) ; Tsungfa, Tsang 25189 (A); Yang-shan Hsien, Tsui 763 (MO, NY), kweichow: Nan-kan, Cheng-feng, Tsiang 4661 (A, NY) ; Pinfa, Kweiting, Tsiang 5360 (A, NY); Tsingchen, Teng 9° 2 54 (A); Tungtze, Chun 5013 (A, S, UC), Tsiang 501 3 (NY); Yao-ren Shan, Sanhoa, Tsiang 6450 (A, NY); without precise locality, Esquirol 4 (A, type of Embelia Esquirolii). sikang: Kangting (Tachien-lu), Smith 12898 (S), Wilson 1302 (A, MO, US), szechuan: Kuan Hsien, Fang 2124 (A, NY); Mo-tien-ling, Wang 22451 (A); Monkong-ting, Wilson 2305^ (A); Mow Hsien, Fang 5529 (A); Nan-chuan Hsien, Fang 593 (A); Omei Shan, Chow 9906 (A), Fang tf Lee 3575 (A, US), Yu-shih Liu 1147, 1423 (A), Tai 1113, 1120 (A); O-pien Hsien, Sun 781 (A); Pao-Hsing Hsien, Chu 3761 (A); Wen-chuan Hsien, Wilson 2305 (A, MO, US) ; Yun-ching Hsien, Sun 1247 (A), yunnan: Chien-chuan-Mekong Divide Forrest 21464, 22311 (A, US); Chungtien, Feng 2868 (A), Rock 24679 (A, MO, NY UC, US); in the mountain above Dashao, Handel-Mazzetti 13074 (A); between Kambaiti and Tengyueh, via Kuyung, Rock 7568 (A, UC, US) ; Kunming, Wang 62990 (A) Likiang Ching 20682 (A), Rock 3799 (A, UC, US), 4195 (A, US); Mengtze, Henry 9782 (A, NY), 9782^ (A, type of C. gemmatus; MO), 10531 (A, paratype of C. gemmatus; MO, NY, US), 11471 (A, paratype of C. gemmatus; MO, NY) ; Muli Wachin, Yu 14375 (A); Ping-pien Hsien, Tsai 61004 (A); south of Red River from Manmei, Henry 9679^ (A); between Shweli and Yengyueh valleys, Forrest 8704 (A, S), II 173 f 12050 (A) ; Suen-oui, Maire 14 (A) ; Tali, Rock 6809 (A, US) ; Tai-pon, Maire 102 (A) ; Wei-si Hsien, Tsai 59523, 59799, 63116 (A), Wang 63709 (A) ; without precise locality, Ducloux 132 (NY, UC), Forrest 16241, 14510 (A), Yu 5562, 5771 (A). Celastrus gemmatus often has been confused with C. orbiculatus Thunb., since the two are closely related. After a careful examination of specimens of the two species, it was found that although their floral morphology is similar, C. gemmatus has larger conic axillary buds, firmly membranaceous and densely reticulated leaves, and, in addition, a distinct geographical distribution which separates it from Celastrus orbiculatus and other related species. Celastrus gemmatus is chiefly distributed in southwestern China and is rarely found in southeastern China; some plants are recorded as far north as Hupei which may represent its northern limit. I have seen the types of Embelia esquirolii Levi, and C. lokchongensis Masamune, They both bear large conical axillary buds and densely reticulated leaves, and are within the geographical range of Celastrus gemmatus. Hence, I consider these species as synonyms of Celastrus gemmatus. 17. Celastrus orbiculatus Thunb. Fl. Jap. pp. xlii, 97 (errore ft articulatus yy ). 1784; Gmel. Syst. Veg. 406. 1796; Lam. Tab. Encycl. et Meth. Bot. 2:94. 1797; Airy Shaw, in Curtis's Bot. Mag. l5S:tab. 9394. 1935. (T.: collector unknown s. n., A, photo! ) . Celastrus articulatus Thunb. Fl. Jap. 97. 1784 (sub. Celastrus orbiculatus, quoad nomen apud Thunb. in Fl. Jap. p. xlii); Maxim, in Bull. Acad. Sci. St.-Petersb., Ill, 27:456. 1881; Mel. Biol. Acad. St.-Petersb. 11:200. 1881. tatarinowii 18 57, ex char. 1955J HOU REVISION OF THE GENUS CELASTRUS 261 Celastrus punctatus sensu Regel, in Gartenfl. 9:407, tab. 312, f. 6. I860, non Thunb. Celastrus articulatus Thunb. var. pubescens Makino, in Bot. Mag. Tokyo 7:102. 1893, ex char. (T.: Makino s.n.). Celastrus orbiculata Thunb. forma £. microphylla Loes. in Engl. Bot. Jahrb. 30:469. 1901, ex char. (T.: Henry 3827). Celastrus orbiculata Thunb. forma 7. maior Loes. loc. cit. 30:469. 1901, ex char. (T.: Ciraldi 237) . Wils. in Sare. Pi. Wils. 2:350. 1915. ( T. : Wilson 2308,Al). Celastrus lancifolia Nakai, in Bot. Mag. Tokyo 37:3. 1923, ex char. (T.: Nakai s. «.). Celastrus insularis Koidz. in Bot. Mag. Tokyo 39:22. 1925, ex char. (T.: Koidzumi s. ».). Celastrus strigillosus Nakai, loc. cit. 40:492. 1926, ex char. (T.: Nakai s.n.). Celastrus articulatus Thunb. var. stephanotiifolius Makino, in Jour. Jap. Bot. 3:24. 1926, ex char. (T.: Makino s.n.). Celastrus stephanotiifolius (Makino) Makino, loc. cit. 3:46. 1926. Celastrus orbiculata var. aureo-arillata Honda, in Bot. Mag. Tokyo 45:422. 1931, ex char. (T.: D. Shimisu 50) . Celastrus jeholensis Nakai, apud Nakai & Kitagawa, in Rept. 1st. Scientif. Exped. Man- choukuo, Sect. IV, 1:6, pi. I. 1934, ex char. (T.: Nakai et al. s. «.). Celastrus articulatus var. papillosus Nakai ex Hara, in Jour. Jap. Bot. 10:84. 1934, ex char. (T.: H. Hara s.n.). astrus articulata var. orbiculata (Thunb.) "Wang 1936. Celastrus orbiculatus var, papillosus (Nakai) Ohwi, Fl. Jap. 736. 1953. Celastrus versicolor Nakai, in Bull. Sci. Mus. Tokyo 33:16. 1953, ex char. (T.: Nakai 6 Maruyama s.n.). Celastrus orbiculatus var. pilosus Nakai, loc. cit. 33:16. 1953, ex char. (T.: Hozawa- Sigeo s.n.). Celastrus orbiculatus f. papillosus Nakai ex Hara, Enum. Spermat. Japon. 80. 1954. Scandent shrubs up to 10 or 12 m. tall; branches terete to slightly striate, glabrous, light to darkish brown, the lenticels inconspicuous or sparse; axillary buds small, depressed, ovoid to subglobose, 1-3 mm. long, sometimes the outermost scales becoming deltoid, sharp-spinose, 1-2 mm. long. Leaves extremely variable in size and shape, usually obovate to suborbicular, ovate, or oval-oblong, the apex rounded, shortly cuspidate to acute, the base cuneate to obtuse, the margins crenate- serrate, 2-12 cm. long, 1.5-8.0 cm. wide, delicately membranaceous, glabrous above, rarely puberulent below, the primary lateral veins 3-6 pairs, distinctly elevated below, slightly elevated above, the veinlets slightly reticulated, prominent and slightly elevated below, immersed and obscure above; stipules filiform, tufted, about 1-2 mm. long; petioles 1-3 cm. long. Inflorescences axillary, rarely also terminal in the male plant, usually 3- to 7-flowered, occasionally solitary, often rather congested along the shoots of the current year, the subtending leaf some- times not developed, distinctly pedunculate, the peduncles subequal, glabrous, the primary peduncles 3-8 mm. long; flowers yellowish-green, regularly dioecious, but occasionally monoecious in late growth, the pedicels about 2-3 mm. long, accres- cent, the articulation usually at the base or lower third of the stalk. Male flowers: calyx lobes open, ovate-deltoid, subacute to obtuse, glandular-ciliolate, about 1.5 mm. long; petals narrowly oblong to oblanceolate, obtuse, subentire to slightly ciliolate or erose, usually 3-5 mm. long and 1-2 mm. wide; disc thin, cup-shaped, the lobes usually acute; stamens arising from the margin of the disc proper, about 262 [Vol. 42 MISSOURI 3 mm. long, the filaments filiform, glabrous, the anthers oblong-ellipsoid, obtuse, pink-punctate; sterile pistil columnar, about 2 mm. long. Female flowers: calyx, petals, and disc as in the male, sometimes the petals smaller; sterile stamens about 1.3 mm. long; pistil flask-shaped, about 4 mm. long, the ovary ovoid, nar- rowed into a columnar style, the stigmata 3-lobed, the lobes flat, recurved, some- times each slightly bifid. Fruits subglobose, bright yellow, the valves broadly elliptic to suborbicular, about 6-8 mm. long and 5-7 mm. wide, 3- to 6-seeded, the pedicels usually green in dry condition. Seeds usually ovoid or ellipsoid, about 5 mm. long and 3.5 mm. wide, minutely areolate, pinkish-brown. Common in lowland slopes or in thickets, at altitudes from 100 to 1,400 m.; J Jap China: chekiang: Hai-wei, southern I-shing, Ching & Tso 580 (A), harbin: (Pinkiang), along a river, Skvortzov s.n., Dec. 7, 1929 (A), honan: Siashih, Hers 345 ?J ; ™ n 8 hsien ' Shih-tze-miao, Hers 1272 (A) ; Teng-feng Hsien, Yu-tai Shan, Hers 2701 (A); Tsi-yuan Hsien, Hers 1758, 2705 (A); Yungning, Yu-tze-ping, Hers 828 (A); without precise locality, Hers 26 (A), hopei (chili): Ch'o-K'ou-tien, 50 km. sw. Peking, Bohlin 104 (S, UPS); Chang-li, Dorsett tf Morse 7152 (US); western hills, Dorsett & Morse 7206 (UC, US); Prince Park, Wang 176 (NY); mont. a l'ouest Peking pagoda, Pi-yun-sen, Janet 1639 (UPS); Haiao-wu-tai Shan, Smith 740, 959 (S, UPS); Kuan-tso-hng, Liu 1176 (UC) ; Ming Tombs, Nankou, Chiao 2126 (US) ; Tsing-ling-chiao, Chtao 21 16 (NY), Dorsett gf Mow 7074 (US); near eastern Tombs, Liou 385 (NY); western Tombs, Liu 398 (UC) ; without precise locality, Li 11169 (NY), hupei: Ichang, Chun 3505 (A) , Wilson 2308 (A, holotype of Celastrus articulata var. cuneata Rehd. & Wils.); Noh-chen Hsien, Chow 369 (NY); Lo-kia Shan, Wuchang, Sun 108 (NY); western Hupei, without precise locality, Wilson 181 (GH, US), jehol (gehol): David W (GH, US), kangsu: Mien Shan, Lin-shih Hsien, Tang 889 (A); Tienschui, Fenzel 2827 (A), liaoning: Liaoyang, Mukden, Li 69 (UC). port Arthur: Wilson ,, e ( tt2: N j VNKING: Lin S H02 (UC). shansi: Chioh-hsin Dist., Smith S745 AS, UPS) , 7697 (A UPS); Chin-yuan, Ling 1607, 1664 (UC); Fu-ping, Tao-ho-tze, Janet A34S (UPS) ; Fu-ping Ta-wu-tai Shan, Janet A349 (UPS) ; Hweihsien, Shansi ~™ e x r * 5 P " 7 °? ( ^. } ; Tsin S- hu /« /""' A 79 (UPS) ; Dao-hui-gou, Wenshui, Wang 1 74 (GH); Yuan-chu Dist., Yang-shu-ling, Smith 6154 (UPS), shangtung: Meng Shan, Fei Hsien, Cheo ef Yen 193 (A); Tsing-tao, First Park, Chiao 2395 (A, NY, US), 2612 (A 'x F ; ?\\y^^J Tsin S- tao ' Lioent *3262 (A) ; Tsinan, Ching-lung Shan, Chiao 3010 (A, NY, UC, US), shensi: Lao-y-san, Giraldi s.n., June 5, 1897 (A, UC); Tai- pe,-shan, Purdom 944 (A, US), 945 (A); Tsingling-shan centr. inter Mei et Liupa, Fenzel 502 (A). Japan: Honshu: Chiba Pref. Mobara, Walker 5717 (US) ; Kanagawa, Hakone in Sagami, Ohwt ? Okamot 504, 506 (L, S, UC, US); Hakone. Sawada 2t6t 455009 UC, Ub); Hakone, Sawada 2161 (S); Mt. ), Shiota 3899, 3905, 9643, 9766 (A) ; Prov. 44 (US); Prov. Nagano. Kimura ?07 (US). s. «., 1914 (A) ; Otake-gawa, Shinano Prov., Wilson s. n., 1914 (A) ; Suruga Prov., Wilson 0907 (A, MO, US) ; Oku-shiobara, Tochigi, Suzuki 444020 (UC) ; Tokyo, Faurie 6134 (A), Yamazala 44-A (US); Kamohura, Sagami Prov., Wilson 6627 (A, US); KumashH, Toman , Toyama, Bergman 439 (S) ; Prov. Ugo, Mizushima 1929 (MO); Mitsutooge, Yamanashi Pref., Suzuki 486024 (UC); Mt. Hoowoo, Yamanashi, Suzuki 491027 (UC). hokkaido: Prov Iburi, Zezo, Takenauchi s.n., July 17, 1916 (S) ; Sapporo, Faurie 6133 A), Jack s. n., Aug. 22, 1905 (A, GH), Miyabe s. «., June 1880 (A, GH), Wilson 74 J2 (A) ; vicmity of Obihiro, Tokachi, Dorsett d Morse 1070 (A, US) ; southern Hokkaido, 19SS] HOU REVISION OF THE GENUS CELASTRUS 263 without precise locality, Brooks 44 (A, UC), 544 (UPS), kurile islands: Shikotan, Ohwi 436 (UPS), without precise locality: Thunberg s.n. (A, type photo of C. orbiculatus, pi. 961). Korea: Port Chusan, Keisho Nan, Wilford s.n., 1859 (GH, S); Port Hamilton, Wil- fords. «., 1859 (GH, S) ; Keijo (Seoul), Mills s. «., Oct. 2, 1915 (UC) ; Kongo-san, Prov. Kogen, Wilson 10461 (US); Pyongang, Wilson 9203 (US); Taiyudo Prov., Hsian Hoku, Wilson 861 1 (US) ; without precise locality, Gilbert 6 (UC). Thunberg describes this species (Fl. Jap. p. 97, 1784) as Celastrus articulatus. However, in his index, p. xlii, articulatus is omitted and orbiculatus is listed. In 1796, Gmelin 32 used the name orbiculatus and referred to Thunberg's Fl. Jap. p. 97, but he did not mention articulatus. Later, in 1881, Maximowicz pointed out 33 that the name articulatus was due to a printer's error, having been changed from C. orbiculatus, as shown in the 'Flora Japonica' and the figures of Thunberg's manuscript. However, Maximowicz chose Celastrus articulatus as the species name because it was universally accepted at that time. Airy Shaw 34 states in 1935 that "The form articulatus is a 'vain tradition', and the sooner it is dropped the better— but tradition dies hard!" I agree that it is best to use the originally correct name for this species, even more so because "orbiculatus" refers to the leaf shape. This is a very widely distributed eastern Asiatic species. It is found in northern Japan The be specimens that have been collected south of the river ( northern Chekiang ) . comm eastern United States where it was introduced at the New York Botanical Garden in 1891 from seeds secured from the Royal Botanic Gardens, Kew 35 . The fruit- clusters are axillary and remain attractive through the winter after the leaves have fallen; the fruiting branches are as showy as those of the paniculiform Celastrus scandens. Birds are fond of the ripe seeds 36 and help to propagate them. I have observed two living plants of this species cultivated in the Missouri Botanical Garden. In the spring both have male flowers; all the floriferous branches bear both axillary and terminal inflorescences. In June, the male flowers fall, and the terminal inflorescences fall away from the plant; some of the axillary buds or sometimes bear onlv female flowers. Some which in these plants are scarce. The leaves are variable. Due to the variation caused by these characteristics, there are many synonyms for this species. '"Syst. Veg. p. 406. 1796. "Mel. Biol. Acad. St. Petersb. 11:201. 1881. 34 Curt. Bot. Mag. 158:/**. 9394- 1935. 35 Nash, C. V. Celastrus articulatus. Addisonia 4:9-10, pi. J 25. 1919. 86 Burbidge, F. W. Japanese tree-strangler, or staff tree (Celastrus arttculatus) . Card. Cnron. Ill, 23:28, /. //. 1898. 264 [Vol. 42 GARDEN Map J. Distribution of ten species of Celastrus, Subgenus celastrus, Series II. AXILLARES. 1955] HOU REVISION OF THE GENUS CELASTRUS 265 18. 1900. (T.: Ros thorn 1572, A, photo!). Jah Celastrus stylosa sensu Levi. Fl. Kouy-Tcheou, p. 69. 1914, non Wall. Celastrus loeseneri Rehd. & Wils. in Sarg. PL Wilson. 2:3 50. 1915. (T.: Wilson 357*, A!). Celastrus cavaleriei Levi, in Monde des PL II, 18:31. 1916. (T.: Cavalerie 496, A, photo!), non Levi. 1914. Scandent shrubs up to 7 m. tall; branches terete, occasionally striate, glabrous, gray-brown to brown-red, densely to sparsely lenticelled, the lenticels elevated, ovate; axillary buds ovoid, about 3 mm. long. Leaves elliptic, oval, or obovate- oblong, the apex acute, the base cuneate to obtuse, the margins remotely serrulate to serrate, usually 4-11 cm. long, 2-6 cm. wide, thinly membranaceous on the flowering specimens, firmly membranaceous on the fruiting specimens, glabrous, the primary lateral veins 4—7 pairs, prominent, slightly elevated on both surfaces, the veinlets obscure on both surfaces; stipules filiform, tufted; petioles 0.5-1.5 cm. long. Inflorescences axillary as well as terminal in the male plants, axillary only in the female, the terminal portion sometimes up to 5 cm. long, usually fascicular, rarely solitary; peduncles glabrous, the primary peduncles very short to nearly obsolete in the fascicular inflorescences, 2—5 mm. long at the solitary ones; flowers dioecious, yellowish-green, the pedicels 2-5 mm. long, the articulation at the or lower half of the stalk. Male flowers: calyx lobes valvate, ovate to middle mm stamens arising from the margin of the disc proper, 2.5 mm. long, the filaments filiform, glabrous, the anthers ovoid, obtuse, slightly brownish-punctate; sterile pistil subulate columnar, about 1.5 mm. long. Female flowers: calyx lobes, petals, bout bglobose each deeply bifid, filiform. Fruits subglobose, the valves broadly elliptic, about mm long and 2 mm. wide, the areolae distinct, yellowish-brown. Chiefly in thickets, at altitudes from 400 to 3,400 m.; China; flowering from April to May. China: hupei: Bo-moh-ping, Chun 3 570 (A); Changyang, Wilson 688 (A), 706 (NY); Hsing-shan Hsien, Wilson 357* (A, type of Celastrus loeseneri Rehd. & Wils.; 3H, MO), 560, 2309 (A, US), Chen logo (UC); Kui, Wilson 5. w., June Patung Hsien, Ho-chang Chow (A, US) ; without precise locality, Henry 315 (A, GH), 5909 (GH, NY, US), hunan: Hsikwang Shan, Handel-Mazzetti 583 (A); Hsinhwa, Handel-Mazzetti 8 JO (A); Wukang, Te-Hui Wang 1 14 (A), kwangsi: Lin-yuin Hsien, Steward & Cheo 429 (A, NY); Ling Wun, Lau 28207 y 28763 (A); Ling-yun Hsien, Lau 28558 (A), kwangtung: Suny, Wang 32148 (A) ; Yang-shan Hsien, Tsui 542 (A, MO, UC, US), kweichow: Anlung, Tsiang 9354 (A, NY); Kiang-kow, Tsiang 7499 (A, NY); Kweiting, Tsiang 5443 (A, NY), 5614 (A); Lang-tai, Yeh-tin, Tsiang 9513 (A, NY); Wai-ho, near Sio-chang, Tsiang 5601 (A); without precise locality, Cavalerie 496 (type of Celastrus cavaleriei, photo, A), sikang: near San-tao-chiao, 40 li from I-tung, Chiao 1720 (A); Konting (Tachien-lu), Wilson 4187 (A); Tienchuan, Tang 3424 (A) ,3486 (A, NY), szechuan: Chao-hua Hsien, Hopkinson 297 (S) ; Chengtu, Chien 5920 (A), Feng 19884 (A) ; Chin- tine Shan, Wilson 211 1 (A); inter Kuapie et Tahao-ko, Schneider 1 37 5 (A); Mian-ning 266 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Hsien, Yu I 7 56 (A) ; Mow Hsien, Fang 5589 (A, NY) ; Nan-chuan, Ros thorn 1 57 2 (type, photo of Celastrus rosthornianus, A), 1753 (A, paratype of C. rosthornianus) ; Ning-yuen, Handel-Mazzetti 1305 (A); Telipu, Schneider 1131 (A); Tien-chuan, Yu-shih Liu 1319 (A); Wen-chuan Hsien, Wilson 1 175 (A, US), 4159 (A); inter flumina Yalung et Nganning-ho, Handel-Mazzetti 2008 (US); Omei Shan, Chow 5707 (A), Chow 11742, 12013 (A), Chu370l (A), Fang 16419, 17795 (A), Lee 2805 (A), 3069 (A, US), Wil- son 4781 (A); without precise locality, Faber 227 (NY), Henry $640 (A, GH, MO), 5734 (GH, NY), yunnan: A-tun-tze, Wang 69361 (A), Yu 7978 (A); Chung- tien, 2428, 2479, 3262 (A); La Kou, Maire 3633 ser. B (A, NY, UC); Lapping Hsien, Tsai 56096 (A); Liang-shan, Tsai 51340, 51343 (A); Likiang, Ching 22119 (A) .Forrest 21201 (A, US), Rock 3885 (A, US), 8312 (NY, UC, US) ; Mar-li-po, Feng 13124, 13369 (A); Mekong, Rock 6946 (A, UC, US); Mienning, Poshang, Yu 17893 (A); Muli, near Lama-sery, Yu 14848 (A); Pan-pien-kai, Maire 7256 (UC) ; Pin-chuan Hsien, Tsai 52901 ( A) ; Ping-pien Hsien, Tsai 60781 (A) ; Si-chour Hsien, Feng 11720, 12084 (A); Shun-ning, Wang 71910 (A); Suen-oui, Maire 81 (A); between Sung-kweh and Tengchuan, Schneider 2683 (A, GH, NY) ; Tso-si, Maire 254 (A) ; Wei-se Hsien, Tsai 57974, 5954$* 61753, 63059, 64147, 68293 (A); without precise locality, Forrest 7812, 16226 (A) ; Tsai 57253, 57271, 57343, 57510, 57769, 60933 (A) ; Yu 5185, 7204, 5$ 45 (A). This is a common species chiefly distributed in southwestern and central China. Since the leaf characters are quite variable, this species has sometimes been mis- identified. However, the characters of the flower, fruit, and seed are constant. The racemiform inflorescences bear clustered, usually sessile flowers; the disc-lobes are subquadrate, broader than long, and mucronate; and the seeds are ovoid or ellipsoid and usually yellowish-brown. These characters can distinguish Celastrus rosthornianus from related species. Leveille first identified a specimen collected by Cavalerie (no. 496) from Kweichow as Celastrus stylosa Wall., and later described it as a new species C. cavaleriei (non Leveille 1914) although he had previously published the same name based on another specimen. The earlier name subsequently was shown by Render 37 to be a Myrsine. I have seen the type photo of the Cavalerie 496 specimen, and it compares closely with others collected in Kweichow. They are all fruiting speci- mens, and clearly belong to Celastrus rosthornianus. These Kweichow plants have grayish-brown, striate-fissured branches, and remarkably uniform, elliptic, firmly membranous leaves. This uniform population perhaps could be considered as an ecotype. 19. Celastrus punctatus Thunb. Fl. Jap. 97. 1784. (T.: Collector unknown, 5. n. y A, photo!). Celastrus scandens sensu Thunb. in Trans. Linn. Soc. 2:332. 1794, non L. Celastrus punctulatus in Abh. Bayer. Akad. Muench. II, 4:150. 1845 (apparently an error for punctatus). Celastrus crispulus Regel, in Gartenfl. 9:407, tab. 312, f. 7-5. i860, ex char. &ill. Celastrus striatus Miq. in Ann. Mus. Bot. Lugd.-Bat. 2:210. 1865-66, non Thunb., ex char. Celastrus kiusianus Franch. & Sav. Enum. PI. Jap. 2:314. 1879, ex char. (T.: Savatier 3528 ) . Celastrus orbiculatus var. punctatus (Thunb.) Rehd. in Bailey, Cycl. Am. Hort. 1:267. 1900. Celastrus articulatus Thunb. var. punctatus (Thunb.) Makino, in Bot. Mae. Tokyo 21:138. 1907. * ' 37 Jour. Arn. Arb. 15:292. 1934. 1955] HOU REVISION OF THE GENUS CELASTRUS 267 Celastrus gracillimus Hay. Icon. Pi. Formosa. 5:24. 1915. (T.: Hayata s. n. A, photo!). Celastrus leiocarpus Hay. loc. cit. 5:22. 1915. (T.: Mori s.n., A, photo!). Celastrus longe-racemosus Hay. loc. cit. 5:23, pi. 3. 1915. (T.: Hayata s.n., TAI!). Celastrus geminiflorus Hay. loc. cit. 5:25, /. 9. 1915. (T.: Nagasama s.n., A, photo!). Celastrus elevativenus Hay. loc. cit. 6:14. 1916. (T.: Faurie s.n., A, photo!). Celastrus punctatus Thunb. var. microphyllus Li & Hou ex Hou, in Taiwania 1:172. 1950. (T.: Yamamoto 802, TAI!). Scandent shrubs, about 2-3 m. tall; branches terete, glabrous, hazel-brown, the lenticels rounded or oval, scattered and elevated; axillary buds deltoid, about 2 mm. long, the outermost bud scales sometimes sharply spinose. Leaves usually elliptic, the apex acute, the base usually cuneate, the margins remotely serrate, 2-7 cm. long, 0.8-3.0 cm. wide, membranous, glabrous, the primary lateral veins 4-5 pairs, slightly elevated below, immersed above, the veinlets obscure to distinct; stipules filiform, tufted, about 2 mm. long; petioles about 0.5-1.0 cm. long. Inflorescences axillary as well as terminal in the male plant, axillary only in the female, fascicu- late, racemiform, or solitary, the peduncles obsolete, glabrous, the primary peduncles suppressed to about 5 mm. long; flowers dioecious, white or pale green, the pedicels 1.0-1.5 mm. long, the articulation at the middle or upper third of the stalk. Male flowers: calyx lobes open, deltoid to oblong, obtuse, glandular-ciliate, about 1 mm. long; petals usually oblong to slightly obovate-oblong, cilia te to slightly erose, 2-5-4.5 mm. long and 1.3-1.5 mm. wide; disc cupuliform, the lobes erect, oblong, about half to one- third as long as the disc proper, obtuse; stamens arising from the margin of the disc proper, the filaments filiform, glabrous, the anthers broadly ovoid, obtuse, the base divided to above the middle, pink-punctate; sterile pistil subconical-columniform, slightly 3-lobed, about 1.5 mm. long. Female flower unseen. Fruits globose or subglobose, usually solitary, fasciculate, sessile or oc- casionally short-pedunculate, pale yellow when dried, the valves suborbicular, about 6 mm. in diameter, 3- to 6-seeded; seeds broadly ellipsoid, about 2 mm. long, smooth, pinkish-brown. Hillsides and in thickets, at altitudes from 100 to 2350 m.; southeastern China, Riukiu Islands, and southern Japan; flowering from March to August. China: anhwei: Chu-hwa Shan, Ching 2661 (UC). chekiang: Changhua, Keng 567 (A, UC) ; south of Siachu, Ching 1677 (MO, NY, UC, US), fukien: Amoy, Chung 493 (SING), 1437, *447> 149* (A, UC), 151 5 (UC), 4604 (A, NY), 5957 (A); Diongloh, Ping En Chen 2447, 2701 (UC), Siu Ging Tang 13657 (UC) ; Foochow, Siu Ging Tang 1677, 7085 (A, UC) ; Kushan, Dunn 44 (A) ; Mt. Useke, Nagasawa s. n. (A, photo of type of C. geminiflorus). Taiwan: Arisan, Gressitt 1 41 (A, L, NY, S, U), 185 (A, L, NY, S, U), Hayata s. n. y April 26, 1914 (TAI, photo of type of C. longe- racemosus; A), 5. n. (A, photo of type of C. gracillimus) ; Nanko-taisan, Sasaki s. «., July 22, 1922 (A, photo); Kwarenko, Faurie s.n. (A, photo of type of C. elevativenus), Nakamura 3618 (TAI) ; Shinchow, Prov. Karenko, Wilson IIO96 (A, US) ; Taito, Mori 5. n. (A, photo of type of C. leiocarpus) ; Mt. Taito, Yamamoto 802 (type of C. punctatus var. microphyllus, TAI). Japan: kyushu: Mt. Kirishima, Tashiro s.n., May 5, 1917 (A); Kudsi Kadsura, Tsuro Ome (?) 500 (A); Nagasaki, Oldham 161 (GH, L, S), 162 (GH), Wilson 6304 (A); Satuma, Suzuki s.n., Aug. 13, 1930 (TAI); Tanegashima, Kita-tane, Mori s.n., Aug. 13, 1934 (TAI); Tanaga-shima, Wilson 6l2I (A). Riukiu Islands: Wright 54 (GH). [Vol. 42 268 ANNALS OF THE MISSOURI BOTANICAL GARDEN The type of Celastrus longe-racemosus is a flowering specimen whereas the type of C. punctatus var. microphyllus is a fruiting specimen. Until now no fruiting or flowering specimens, respectively, for these two species have been found. On further examination of additional specimens, it was observed that the fruiting branchlets of the type of C. punctatus var. microphyllus are the same as the so- called long-racemose inflorescences of C. longe-racemosus; consequently, they appear to belong to the present species because they cannot be distinguished from it. On comparing material from Fukien and southern Japan, this interpretation is confirmed by both morphological characters and geographical distribution. In general appearance, the present species differs conspicuously from Celastrus orbiculatus and more closely resembles C. rostbornianus. However, the deeply cleft, oblong disc lobes, the articulations located at the upper third or half of the stalk, as well as the distinct geographical distribution distinguish this species from related ones. 20. Celastrus kusanoi Hayata, in Jour. Coll. Sci. Imp. Univ. Tokyo 30:60. 1911; Icon. PI. Formos. 1:137. 1911; 5:20. t. 8. 1915, ex char. & ill. (T.: Kusano $. «.). Scandent shrubs up to 18 m. tall; branches terete, sometimes striate, brown, the branchlets glabrous to brownish-pubescent, both sparsely or rarely densely lenticelled, the lenticels orbicular or ovate; axillary buds deltoid, about 2.5 mm. long. Leaves broadly elliptic to nearly orbicular, rarely elliptic, the apex rounded to shortly cuspidate, the base broadly truncate, rarely obtuse to cordate, 5.0-10.5 cm. long, 5-11 cm. wide, chartaceous, glabrous or pubescent on the veins below, the primary lateral veins 5-7 pairs, slightly elevated on both surfaces, the veinlets distinct below, obsolete above; petioles usually 1.5-2.5 cm. long, rarely up to 5 cm. long; inflorescences axillary and cauline at the basal position of the current year's growth, rarely also terminal in the male plant, usually 3- to 7-flowered, the peduncles usually pubescent, sometimes yellow, the pedicels about 3-5 mm. long, usually pubescent, the articulation nearly toward the base of the stalk. Male flowers: calyx lobes deltoid, obtuse, entire, about 1 mm. long; petals obovate to oblong, rounded, about 4 mm. long and 1.5 mm. wide, ciliate, sometimes pubescent within on the lower portion; disc subfleshy, flat, the lobes obscure, truncate; stamens arising from the margin of the disc, about 3 mm. long, the filaments fili- form, papillose-tuberculate, the anthers ovoid, subcordate; sterile pistil conical, about 1 mm. long. Female flowers: calyx, petals, and disc as in the male; sterile stamens about 1.5 mm. long; pistil about 3.3 mm. long, the ovary subglobose, th style columnar, the stigmata 3-lobed, reflexed. Fruits globose or subglobose, the valves suborbicular or broadly elliptic, about 8-10 mm. long and 5-7 mm. wide, 3- to 6 -seeded; seeds slightly lunate, minutely areola te, dark brown. Chiefly in thickets, at altitudes from 100 to 1,000 m.; southeastern China; flowering from February to April. CTO ^ H T T A ^ Bak v S u La Z 26l8 ± i 66<>6 (A l> Chin S-mai Hsien, Lei 3 OQ (A, NY, SING, UC, US); Fan Yah, Chun & Tso 44028 (A, F, NY, US), kwangtung: 1955] HOU REVISION OF THE GENUS CELASTRUS 269 Fig. 10. Celastrus kusanoi Hayata Lokchong, Tsiang 1 393 (A, UC), Tso 20/88, 20796, 21026 (NY); Tsinleong Shan, Mei Hsien, Gressitt 1240 (A, MO) ; Lung-t'au Shan, Iu village, Kang-peng To 131 (UC), 442 (MO, UC); Nam Shan, Ho-yuen Hsien, Tsang 28927 (A); Ying-tak, Kang-peng To & Kam-chow Wong 2784, 2928 (UC). Taiwan: Arisan, Faurie 1 374 (A) ; Kosyun, Kudo tf Suzuki 16048 (TAI), Suzuki 6195 (TAI) ; Kuaru, Yatnada 121 5 (TAI) ; Mt. Naier, Bansyoryo, Kawakatni & Mori 31 45 (TAI) ; Nanto, Wilson 998 1 (A) ; Taipei, Wilson 1 121 4 (A); Takou, Apes Hill, Henry 1893 (A, NY, TAI). Henry 38 first regarded the Formosan specimen, collected by himself (no. 18Q3) , as a variety of Celastrus articulatus Thunb., but did not give it a name. Later Hayata described a specimen collected by Kusano as Celastrus kusanoi Hayata, but he did not cite Henry's specimen. He stated that Celastrus kusanoi is near C. articulatus (C. orbiculatus) but differs from it in having more rounded leaves and 38 A List of Plants from Formosa, p. 17. 1895. 270 MISSOURI [Vol. 42 carpels transversely wrinkled in the dried condition. His original description was based on a fruiting specimen; later 39 he obtained a flowering specimen, and then amplified his description. Celastrus kusanoi bears pubescent filaments, lunate seeds and usually nearly rounded leaves, hence it is easily distinguished. miens and sometimes even cordate leaves, whereas those from Hainan and Kwangtung usually bear suborbicular or broadly obovate leaves without cordate bases. Merrill and Chun 40 have said that Kwangtung records of Celastrus articulatus Thunb. (C. orbiculatus Thunb.) actually belong to C. kusanoi Hay. and that Thunberg's species does not extend as far as Kwangtung. Of the many Kwang- tung specimens labelled Celastrus articulatus which I have examined, none of them actually is that species; they are either Celastrus kusanoi Hay. or C. gemmatus Loes. 21. Celastrus hirsutus Comber, in Notes Roy. Bot. Gard. Edinb. 18:233. 1934, ex char. (T.: Forrest 17963). Scandent shrubs up to 12 m. tall; branches terete or slightly compressed, some- times more or less sulcate, densely brownish-pubescent or glabrescent, the lenticels elevated, large, ovate, elliptic or orbicular, rarely lacking on the current year's growth; axillary buds ovoid, usually about 3 mm., rarely 5 mm., long, the bud scales accrescent, sometimes persistent. Leaves broadly ovate to obovate, the apex round, shortly cuspidate, the base rotund or broadly cuneate, the margins crenate- serrate, 7-14 cm. long, 4-10 cm. wide, membranous, densely pubescent on both surfaces especially so on the veins when young; primary lateral veins 6-7 pairs prommently elevated below, slightly elevated above, the veinlets distinct below, immersed to obscure above; stipules laciniate, about 1 mm. long; petioles pubescent, about 1.5-3.0 cm. long. Inflorescences axillary and cauline at the basal portion of the flowering branch, usually 7- to 1 4-flowered, the peduncles hirsute, the primary peduncles 5-15 mm. long; flowers dioecious, creamy-white to greenish-yellow the pedicels 3 mm lobes Male flowers: » -* / »"« f«uc;>L.ciii. wnnouc, aDOUt 1 mm. long; petals obovate-oblong, puberulous at the basal portions of both sur- mm. long and 2 mm lobe, the disc lobes, about 2.5-3.5 mm. long, the filaments subulate, papillose- tuberculate, the anthers ovoid, obtuse, cordate; sterile pistil small, ovoid, about 1.3 mm. long. Female flowers: calyx lobes, petals, and disc as in the male; sterile stamens about mm long; pistil 3-5 mm. long, the style columnar, the stigmata conspicuously 3-lobed. Fruits subglobose, the valves broadly elliptic, about 8-10 mm. long and mm 39 Icon. Pi. Formos. 5:20. 1915. 40 In Sunyatsenia 5:111. 1940. 1955] HOU REVISION OF THE GENUS CELASTRUS 271 Fig. 11. Celastrus hirsutus Comber In thickets, at altitudes from 1,400-2,500 m.; China, Burma, and Indo-China; flowering from March to April. Burma: Adung valley, 27°30' to 28°30' lat., and 97°30' to 98°30' long., Kittgdon Ward 9362 (A). China: yunnan: Kengma, Yu 17294 (A); Liukiang, Muchietu, Yu 2I0I2 (A); Ma- kwan Hsien, Tsai 51945 (A); Mar-li-po, Sze-tai-po, Feng 13765 (A); Mienning, Yu 17952 (A); Ping-pien Hsien, Tsai 55421, 60261, 60499, 62396 (A); Shang-pa Hsien, Tsai 54799 (A); without precise locality, Forrest 17496 (A). Indo-China: Chapa, Tonkin, Petelot 5936 (A, NY). The particularly dense pubescence on both surfaces of the leaf makes Celastrus hirsutus easily recognized and distinguished from other species of the genus. On imilar to Celastrus kusanoi r has point Specimens sim 272 [Vol. 42 MISSOURI BOTANICAL leaf and lenticel characters are distinct. For these reasons I am considering them as two separate species. 22. Celastrus stylosus Wall Wall Cat. 4313. 1831; Lawson, in Hook. Fl. Brit. Ind. 1:618. 1875 (pro parte); photo!). J 1904. (T.: Wallich 3-4 branchlets, light brown to dark brown, sparsely lenticelled, the lenticels small, 11 • -44 * 1 - - mm long. Leaves usually elliptic-oblong, ovate to obovate, the apex acute, the base acute or obtuse, the margins serrate, 6-1 5 cm. long, 3-9 cm. wide, membranous to firmly membranous, glabrous or rarely pubescent on the veins below, the primary lateral veins usually 5-7 pairs, distinctly elevated below, plane or slightly elevated above, the veinlets prominent below, obscure to visible above; stipules filiform, about 1 mm. long; petioles usually 1-2 cm. long. Inflorescences axillary and cauline, usually 3- to 7-flowered, at the basal part of the current year's growth (rarely also terminal in the male plant of Celastrus stylosus ssp. stylosus), distinctly pedunculate; th peduncles puberulous to glabrous, the primary peduncles 5-11 mm. long; flowers dioecious, green or pale green, the pedicels 2-5 mm. long, the articulation toward the base of the stalk. Male flowers: calyx lobes imbricate, oval to oblong, obtuse, slightly erose to entire, about 1.5 mm. long; petals obovate, obtuse, slightly erose, 1.5 mm. wide; stamens arising between the disc-lobes, about 2.5 mm. long, the filaments filiform, fleshy, glabrous to papillose-tuberculate, the anthers ovoid, cordate; disc membranous, cup-shaped, the lobes distinctly arcuate or depressed- quadrate; sterile pistil about 1.5 mm. long. Female flowers: calyx lobes, petals, and disc as in the male; sterile stamens about 1 mm. lone: oistil 2—4 mm bout 3 mm. long, the ovary J-lobed, each lobe bifid, flat bo blackish lunate ;xed. Fruits subglobose, the valves 10 mm. wide, 3- to 6-seeded; seeds mm KEY TO THE SUBSPECIES A. Staminal filaments always densely papillose-tuberculate; the disc-lobes depressed- quadrate; leaves elliptic, membranous. India % i AA. Staminal filament, usually glabrous, rarely S^^^a^i^u^^' "^ lobes arcuate; leaves usually elhpt.c-oblong, firmly membranous. China..: 22b. ssp. glaber 22 a. Celastrus stylosus ssp. stylosus. Celastrus neglecta Wall. Cat. 4341. 1831, nom. nud Gymnosporia neglecta Wall ex Lawson, in Hook. Fl. Brit. Ind. 1:619. 1875- Prain in Jour. Asiatic Soc. Bengal 73:198. 1904, Novic Ind 419 19ns (LLa r l \ neglecta WM. y Wallicb 434 i). 419 ' 19 ° 5 ' (based on Celastrus In thickets, at altitudes from 1,000 to 2,745 m.; India; flowering from March to July. ° 1955] HOU REVISION OF THE GENUS CELASTRUS 273 s Fig. 12. Celastrus stylosus ssp. stylosus 1:62. 1936; in Contr. Bot. Surv. N. W India: assam: Khasi, Kurz 171 (CAL), Clarke 18678° (CAL). bengal: Darjeeling, Boswastore 73 (CAL), Clarke 26701* (CAL), Kurz s. n. (CAL). sikkim: Duphla Hills, Lister 1 57 (CAL) ; Lachen, Smith & Cave 964 (CAL) ; Moughoo, Lister s. n., March 1878 (CAL) ; Punkabari, Lister s. n., April 1878 (CAL) ; Singolila forest, Rogers s. n., Jan. 1900 (CAL); Sureil, Dr. Train's collector 439 (CAL). without precise locality: Anderson 104 (CAL); Meebold 1 5912 (S). 22b. Celastrus stylosus ssp. glaber Ding Hou, stat. et nom. nov. Celastrus hypoleucus forma 7. puberula Loes. in Engl. Bot. Jahrb. 29:445. 1900. (T.: Kosthorn /55<5\ A, photo!). Celastrus crassifolia Wang, in Chin. Jour. Bot. China 1:62. 1939. (T.: Yu 445> A!). Chiefly in thickets, at altitudes from 800 to 2,000 m.; China and Indo-China; flowering in April. China: anhwei: Kimen, Ip 41 (UC) ; Ching 3165 (UC) ; Wong Shan, Ling 1 1 04 (UC). kwangst: Chuen Yuan, Tsoong 82041 (A); Kwei-lin Hsien, Tsang 38365 (US) ; Ling-yuin Hsien, Steward & Cheo 87 (A, NY, S, SING), 396 (NY, S) ; Nanning, Seh-feng-dar Shan, Ching 8205 (NY, UC, US) ; Shang-sze Hsien, Tsang 24131 (A, MO) ; Pin-lam, Ko 55544 (A); Tsin-hung Shan, Hin Yen, Ching 7080 (A, US), kweichow: Mapo, Pingchow, Tsiang 6832 (A, NY), sikang: Tien-chuan Hsien, Tat & Feng 5238 (A), szechuan: Nan-chuan, v. Kosthorn 1556* (A, photo of tvne of C. hypoleuca 3168 445 (A, type of C. crassifolia); Wa W — - (A) ; without precise locality, Henry 5559 (A, GH). yunnan: Chungtien, Feng 3365 (A); Hokin, Feng 747 (A); Lung-ling Hsien, Tsai 54563, 54578 (A); Mar-li-po, Chun*-dzia, Feng 12727 (A) ; Mengtze, Henry 10^22, 11267 (A, MO) ; Ping-pien Hsien, Tsai 60420,60919, 61156, 61629 (A); Shang-pa Hsien, Tsai 54929 (A); Shunning, Wumulun*, Yu 16592 (A); Si-chour Hsien, Feng 12260, 12283 (A); without precise locality, Forrest 9396, 15980 (K), Li 1 18 1 (A), Tsai 62841 (A). Indo-Chtna: Chapa, Tonkin, Vctelot 5829, 5946 (A, NY, US). These two subspecies are superficially similar except for the staminal filaments, the leaf shapes, and the geographical distributions as shown in the key. [Vol. 42 274 ANNALS OF THE MISSOURI BOTANICAL GARDEN Subspecies glaber is widely distributed in China; its extra-axillary inflorescences and firmly membranous elliptic-oblong leaves make it easy to separate from other related species. The staminal filaments are usually glabrous; however, a few speci- mens collected from Kwangsi and Yunnan are slightly papillose-tuberculate. Because of the intermediate leaf forms, staminal filaments, and distinctly geograph- ical distributions, I consider the Chinese population as a subspecies of C. stylosus. Under the provisions of the International Rules, priority of publication operates only within individual taxa. It would therefore be rather inappropriate to employ Loesener's epithet "puberula" in this instance, since Loesener chose the epithet with reference to the puberulence of the leaves of Kosthorn 1556*, which actually is an abnormal condition for the other known specimens of Celastrus stylosus ssp. glaber. Thus, I have chosen the epithet "glaber" with reference to the more significant character of the staminal filaments. 20092, NY!). Jour. 13:37. 1934. (T.: Tsang cited, in part, non Prain. W specimens Celastrus oblanceifolia Wang & Tsoong, in Chin. Jour. Bot. 1:65. 1936, ex char. (T.: Tsoong 2443 ) . Scandent shrubs up to 10 m. tall; branches terete, glabrous, the young branch- lets sometimes brownish-pubescent, brownish-red to dark brown, both branches and branchlets lenticellate, the lenticels orbicular, sparse to dense. Axillary buds ovoid, about 2.5 mm. long, the outermost scales persistent, usually deltoid and spiny, acute to acuminate, accrescent, up to 5 mm. long. Leaves elliptic to ob- lanceolate, the apex acute, the base cuneate to obtuse, the margins remotely serrulate, 3-10 cm. long, 1.5-6.0 cm. wide, membranous, usually glabrous, rarely pubescent on the veins below, the primary lateral veins 4-5 pairs, arcuate toward the apex, immersed, distinct to slightly elevated below, obscure above, the veinlets usually obsolete on both surfaces; stipules laciniate, filiform, about 1 mm. long; petioles 7-12 mm. long. Inflorescences axillary and cauline at the lower part of the flowering branch, shortly pedunculate, the primary peduncles almost obsolete to 3 mm. long, usually 3 -flowered, puberulous; flowers dioecious, greenish-yellow, subsessile, the articulation just below the flower. Male flowers: calyx lobes im- bricate, ovate-deltoid to oblong, obtuse, entire, subglabrous without, accrescent, about 1.0-1.5 mm. long; petals oblong to oblanceolate, obtuse, slightly undulate, puberulent at the basal parts on both surfaces, about 3.5-4.5 mm. long and 1 mm. wide; disc fleshy, annular, entire; stamens arising from the margin of the disc proper, about 3 mm. long, the filaments filiform, densely papillose-tuberculate; sterile pistil ovoid, about 2 mm. long. Female flowers: calyx, petals, and disc as in the male; sterile stamens 1.5 mm. long, papillose; pistil subglobose, the style columnar, distinct, the stigmata discoid or slightly 3-lobed. Fruits subglobose, the valves broadly elliptic, about 8 mm. long and 7 mm. wide, 3- to 6-seeded; seeds 19S51 HOU REVISION OF THE GENUS CELASTRUS 275 Fig. 13. Celastrus aculeatus Merr. arcuate or semi-annular, about 4.5 mm. long and 1.5 mm. wide, brown, slightly wrinkled, distinctly areolate. In open fields or in thickets, from lowland up to about 900 m. elevation; south- eastern China; flowering from March to April. China: chekiang: Sia-chu, Ching 1664 (A, US); Taichow, Ching 1336 (US); Tsi- shun, Keng 285 (A); Tsingtien Hsien, Keng 73 (A), fukien: Amoy, Chung 1 514 (A, UC), 1693 (A, SING, US), 1773 (US), 4774 (A), 4777 (A), 5954, 6235 (A); Baek- liang and vicinity, Siu-Ging Tang 1 5609 (UC) ; Chuanchow, Chung 3085 (UC) ; Diongloh and vicinity, Ku Tai Lin 11677 (UC) ; Foochow, Norton 1 346 (UC), Chung-Chang Tang esf Shan En Ma 2936 (US); Kuliang, Chung 7264 (A, F); Kushan and vicinity, Cheng 1573 (UC); Kutien, Chung 4036 (A); Minhow Hsien, Chung 2065 (UC) ; Pu- cheng, Ching 2508 (A, UC, US) ; Tsze-chook-Hang, central Fukien, Dunn s. n. (HK, A) ; Yenping, Kuang-Han Chou 821 5, 8460 (UC) ; Yuen-fu gorges, Dunn s. n. (HK, A). hunan: Yi-chang Hsien, Tsang 23574 (A, US), kiangsi: Kaoan, Tsiang 10460 (NY); Kiennan Hsien, Lau 4367 (A, S, US); Lungnan Hsien, Lau 4801 (A, S, US); Lu Shan, Chiao 18768 (US); Swe-chuen, Hu 889 (A); Yung-shing, Hu 784 (A); without precise locality, Hu s.n. f 1920 (UC). kwangtung: Ho-yuen, Tsang 28843 ( A ); Kowloon, Wang 3060 (NY, SING) ; Mt. Lung-t'au, near lu, Kang-peng To et al 631 (US); Lung- tung, Tso 21640 (NY); Mei Hsien, Tsang 21416 (A, NY, S); Mui-nen Hsien, McClure & Shang 4 (UC) ; Naam-kwan-shan, Lung-men, Tsang 25268 (A) ; Sin-fung Hsien, Taam 276 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 724, 1041 (A); Kakchieh, Swatow, Gressitt 1767 (MO); Ta-ching, Chun 5539 (A); Tapu, Tsang 21189 (A, NY, S); Tsengshing Hsien, Tsang 20092 (NY, holotype of C. aculeatus; A, isotype) ; Wai-yeung, Tsui 154 (A, MO, NY, UC, US); Yao-shan, Sin U8S9 (NY). This species is well characterized by the arcuate seeds, the papillose-tuberculate filaments, the fleshy and annular disc, and the obsolete veinlets of the leaves. Celastrus aculeatus is a very distinctive species of southeastern China. Most of the specimens which are cited here have been identified as Celastrus hookeri Prain. Celastrus hookeri, however, is easy to separate from Celastrus aculeatus by its glabrous filaments, its cup-shaped disc, and its ovoid seeds. A specimen collected by Ching (2508) from northern Fukien has puberulent veins and slightly oblanceolate leaves. It matches the description of Celastrus oblancei folia Wang & Tsoong from southern Anhwei which also has oblanceolate leaves. The veinlets described for Celastrus oblanceifolia are reported to be obsolete here and the seeds are said to be curved, both typical characters of Celastrus aculeatus. Because of these morphological similarities and the identical geographical distribution, Celastrus oblanceifolia is placed in synonymy. 24. Celastrus flagellars Rupr. in Bull. Acad. Sci. St. Petersb. II, 15:357. 1857, ex char. Celastrus ciliidens Miq. in Ann. Mus. Bot. Lugd.-Bat. 2:8 5. 1865-66, ex char. Celastrus clemacanthus Levi., in Fedde, Rep. Spec. Nov. 8:284. 1910. (T.: Taquet 632, A!). Scandent shrubs; branches and branchlets terete to slightly striate, glabrous, brown to dark brown, the lenticels small, oval, sparse, the young branchlets cling- ing by a line of filiform, branched aerial roots; axillary buds small, depressed-ovoid, usually protected by two prominent, broadly falcate, spiny scales, acuminate, about 3 mm. long. Leaves broadly elliptic to suborbicular, the apex obtuse to shortly acuminate finely cm. long, 2-5 cm. wide, delicately membranous, glabrous or puberulous on the veins below, the primary lateral veins 4-6 pairs, distinctly elevated below, slightly elevated above, the veinlets distinct, sometimes elevated below, obscure above; stipules laciniate, filiform, about 5 mm. long; petioles 1.5-2.5 cm. long. In- ary the peduncles about 2-5 mm. long; flowers dioecious, white or yellowish- green, the 1-4 upper half of the stalk. Male flowers: calyx lobes imbricate, oblong, obtuse, ciliate, about 2 mm. long; petals obovate-oblong to elliptic-oblong, ciliate to slightly erose, about 4 mm. long and 1.3 mm. wide; disc cup-shaped, the lobes incon- mm ments fleshy, filiform, glabrous, the anthers ovoid, obtuse, apiculate, cordate; sterile about 1 mm. long. Female flowers: calyx lobes, petals, and disc umnar mm angular-cordate; pistil about 2.8 mm. long, the ovary subglobose, the style 1955] HOU REVISION OF THE GENUS CELASTRUS 277 columnar, distinct, the stigmata trilobed, each lobe deeply bifid, reflexed. Fruits globose, the valves broadly elliptic to suborbicular, 6-7 mm. long and 5-7 mm. wide, 3- to 6-seeded; seeds shortly ellipsoid, about 3.5 mm. long and 2 mm. wide, brown, the areolae obscure. Lowland thickets, at altitudes up to 1,000 m.; China, Korea, and Japan; flower- ing from May to August. China: chekiang: Yu-tsien, Hu 1631 (A, UC). northeastern china: ad fl. Amur, Maack u *., 1855 (GH); sinus Possict, Maximowicz s.n., 1860 (GH). Korea: Chulla, Mrs. R. K. Smith s. n. (A) ; Gyouhfeng, Taquet 2723 (A) ; Quelpaert, scandens in muris agrorum Haouen, Taquet 632 (A, type of Celastrus clemacanthus) ; Hongo-san, Prov. Hogen, Wilson 10425 (A); Ping Yang, Jack s.n., Sept. 18, 1908 (A); Puk Han, Seoul, Jack s.n., Sept. 25, 1905 (A); Nam-san, Seoul (Heijyo), Prov. Heiki, Wilson 8450 (A, US). Japan: Kawagishi, Naganoken, Uno 21823 (A, NY); Musatre, Mitake, Hayakawa s.n., May 20, 1910 (S) ; Thinano, Suwa, Sakurai s.n., May 21, 1913 (A); Thinano, Nagano, Sakurai s.n., June 19, 1913 (A); near Lake Yamanaka, Dorsett & Morse 6lQ (A, US). This species is easily recognized by its characteristic ciliate-serrate leaves and the two persistent spiny outermost bud scales. Its branches sometimes produce aerial roots which function as attachment organs on smooth surfaces. The flowers are usually clustered on young shoots which occasionally elongate during the fruiting stage. Subgenus II. Racemocelastrus Ding Hou, subgen. nov. Frutices scandentes. Flores hermaphroditi; ovario triloculare, in quoque loculo ovulo singulo. Capsula semine singulo, ovulis manifeste abortivis binis. America centralis et australis. Type species: Celastrus racemosus (Reiss.) Loes. KEY TO THE SPECIES A. Inflorescences paniculiform, obviously compound. B. Inflorescences thrice- or multi-compound, up to 14 cm. long; lenticels dense. Mexico: Chiapas 25. C. lenticellatus BB. Inflorescences usually once or twice compound, up to 6 cm. long; lenticels scattered. C. Inflorescences usually clustered in the leaf axils; peduncles not associated with vegetative buds; leaves elliptic. Brazil, Venezuela, Costa Rica, British Hon- duras, Guatemala, El Salvador, and Veracruz, Mexico . 26. C. racemosus CC. Inflorescences solitary in the leaf axils; peduncles associated with vegetative buds; leaves ovate. Panama 27. G panamensis AA. Inflorescences racemiform, obscurely compound. D. Pedicels of the flowers 1-2 mm. long. E. Anthers distinctly apiculate; fruit valves about 12-14 mm. long, the septa 1.0-2.5 mm. wide; leaves closely serrate. Central Mexico 28. C. pringlei EE. Anthers obscurely apiculate; fruit valves about 14-18 mm. long, the septa 4.0-6.0 mm. wide; leaves remotely serrate to entire. Chiapas, Mexico; Guate- mala; Honduras; El Salvador 29. C. vulcanicolus DD. Pedicels of the flowers obsolete or less than 1 mm. long. F. Leaves oblanceolate-oblong; lenticels large, dense and elevated; anthers apiculate. Colombia 30 - C - caseariifolius FF. Leaves broadly elliptic or ovate; lenticels small, sparse and obscure; anthers not apiculate. Colombia and Venezuela 31. C. meridensis 278 [Vol. 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN Map 4. Distribution of seven species of Celastrm Subgenus racemocelastrus. 25. Celastrus lenticellatus Lundell, in Bull. Torr. Bot. Club 67:616. 1940. (T.: Purpus 73?o, US!). Scandent shrubs; branches glabrous, black-brown, the lenticels small, dense, white, slightly elevated; axillary buds conoid, acute, about 2 mm. long. Leaves broadly elliptic, the apex abruptly short-acuminate, the base cuneate to rotund, the margins remotely serrulate, 10-20 cm. long, 5.0-7.8 cm. wide, chartaceous, glabrous, the primary lateral veins distinctly elevated below, plane or slightly im- pressed above, the veinlets prominent below, obscure to visible above. Stipules about 1 mm. long; pet Inflorescences axillary as well as terminal, solitary or fasciculate, thrice or multicompound, paniculiform, up to 14 cm. long, much branched at the base, the peduncles glabrous, the primary peduncles 1-5 mm. long; flowers bisexual, the pedicels about 1 mm. long, the articulation at the lower part of the stalk. Calyx lobes valvate, ovate, mm\ _*^ mm- ^ _ J --^ ^ _ — _ _ m— _ I _ _ ^\ P* *x ^*» 4 _-. _ rotund oj r ~~* v ** & , *viuuu, V punctate, 1.8 mm. long and 1 mm. wide; disc fleshy, flat, about mm 1955] HOU REVISION OF THE GENUS CELASTRUS 279 diameter, the lobes depressed-rectangular; stamens attached slightly under the margin of the disc, about 2 mm. long, the filaments filiform, glabrous, the anthers ovoid, obtuse, slightly apiculate; pistil short, ovoid, about 1 mm. long, the style stout, the stigmata obscure. Fruit unknown. Mexico: Chiapas; flowering in June. Mexico: Chiapas, Finca San Cristobal, Purpus 7370 (F, type; MICH; US, holotype). The paniculiform inflorescences and the densely lenticellate branches make this species easily separated from other Latin American Celastrus species. 26. Celastrus racemosus (Reiss.) Loes. in Engl. Bot. Jahrb. 24:199. 1898, (as racemosa). US!). astrus liebm MO!). / 4871 Celastrus pachyrachis Lundell, in Lilloa 4:3 82. 1939. (T.: Jabn 476, US!). Celastrus mainsiana Lundell, in Lloydia 2:99. 1939. (T.: Lundell 6307, MICH!). Scandent shrubs up to 50 m. tall; branches terete, glabrous, blackish-brown, the lenticels scattered to dense, white, elliptic to orbicular; axillary buds conoid, acute, about 1 mm. long. Leaves elliptic to ovate, the apex acute, the base cuneate to rotund, the margins crenulate-serrulate, 5-12 cm. long, 3-5 cm. wide, firmly membranous, glabrous, the primary lateral veins 7-9 pairs, slightly elevated below, plane and visible above; stipules subulate, erose, about 1 mm. long; petioles 3-10 mm. long. Inflorescences axillary, 1- to 3 -branched, once compound, up to 4 cm. long, the primary peduncle glabrous, about 3 mm. long; flowers bisexual, greenish- white, the pedicels 2-3 mm. long, accrescent, the articulation toward the base of the stalk. Calyx lobes imbricate, deltoid, obtuse, subentire, brownish-punctate, about 1 mm. long; petals oblong, obtuse, subentire, about 1.5 mm. long and 1 mm. wide; disc fleshy, flat, about 1.5 mm. in diameter, the lobes subreniform; stamens attached slightly beneath the disc margin, about 1.5 mm. long, the filaments linear, glabrous, the anthers ovoid, obtuse, rarely apiculate; pistil ovoid, about 1 mm. long, the style columnar and blunt. Fruits ellipsoid, the pedicels 2.5-5.0 mm. long, the valves broadly elliptic, 14-21 mm. long and 8-11 mm. wide, the septa 2.5-6.0 mm. wide, 1 -seeded; seeds cylindric, about 10-13 mm. long and 7-9 mm. wide, pinkish-brown, areolae obscure. In wet forests or thickets, at altitudes 1,400-2,400 m.; Mexico, British Hon- duras, Guatemala, Costa Rica, Venezuela, and Brazil; flowering from March to August. Brazil: Rio de Janeiro, Riedel s.n. (US, type of Maytenus racemosus). British Honduras: El Cayo Dist., Lundell 6307 (MICH, holotype of C. mainsiana). Costa Rica: Heredia, Standley & Valerio 52024 (US). Guatemala: Tactic, Standley 71 349 (F). Mexico: vera cruz: near Jalapa, Pringle 8133 (K, MO, NY, S, UC, US); Mirador, Liebmann 1 48 7 1 (F, type of C. liebmannii; MO), 14872, 14873* H$74> *4875*> H^75 h (F), 14875 (GH, UC, US), Purpus 8926 (MO, NY, UC, US); Zacuapan, Purpus 7094 (F, MO, NY, UC), 8080 (MO, NY, UC, US). 280 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 Venezuela: east of El Junquito, Steyermark 57015 (F, NY, US) ; Macarao, Jahn 476 (US, holotype of C. pachyrachis) . This is a widely and disjunctively distributed species of the Latin American Celastrus. It is found in the rain forests or thickets from Vera Cruz, Mexico, southward to Rio de Janeiro, Brazil. Lundell 41 has stated, "from examination of the type photograph of C. racemosus (Reiss.) Loes., the similarity of that species and the type of C. liebmannii is striking," I have seen the types of Celastrus racemosus and C. liebmannii; in addition, I have examined specimens collected from Vera Cruz, British Honduras, Guatemala, Costa Rica, and Venezuela. They are all similar and excessively difficult to separate into two species. This species is closely related to Celastrus pringlei Rose and C. vulcanicolus Donn. It can be distinguished from them by the distinctly once compound, aggregate dichasia and longer pedicels. The other two species bear racemiform inflorescences and have very short pedicels. 27. Celastrus 1941. MO Scandent shrubs; branches terete, glabrous, smooth, shining, reddish-brown, the lenticels small, obscure, elliptic; axillary buds conoid, acute, about 1 mm. long. Leaves ovate, the apex obtuse, the base rotund, the margins shallowly crenate, 9-14 cm. long, 5.0-7.5 cm. wide, membranous, glabrous, the primary lateral veins 7-8 pairs, elevated below, plane and distinct above, the veinlets slightly elevated mm 11—13 mm. lone. Inflo usually twice compound, the primary peduncles glabrous, about 1.2-2.5 cm. long, associated with a vegetative bud in the axil; flowers bisexual, white, the pedicels about 1 mm. long, the articulation toward the base of the stalk. Calyx lobes im- bricate, obtuse, minutely erose, about 1.2 mm. long, brownish-punctate; petals oblong, rotund, more or less entire, about 2 mm. long and 1.2 mm. wide; disc fleshy, flat, about 2 mm. in diameter, the lobes subreniform; stamens attached slightly beneath the margin of the disc, about 2 mm. long, the filaments filiform, glabrous, the anthers ovoid, slightly apiculate; pistil short, conoid, about 1.5 mm. long, the style columnar and blunt. Fruit unknown. At altitudes 1,400-2,300 m.; Panama; flowering in April. Panama: chiriqui: Allen 31Q (MO, type; P). The floriferous branch of this species is very peduncle axillary vegetative bud. I have seen only the type collection. It is a young flowering branch. 41 In Lilloa 4:3 80. 1939. 1955] HOU REVISION OF THE GENUS CELASTRUS 281 Fig. 14. Celastrus Pringlei Rose 28. Celastrus pringlei Rose, in Contr. U. S. Nat. Herb. 5:195. 1899. (T.: Pringle 6842, MOl). Celastrus longipes Lundell, in Lilloa 4:381. 1939. (T.: Palmer 106, MO!). Scandent shrubs up to 5 m. high; branches terete, glabrous, brown or reddish- brown, densely lenticelled, the lenticels elliptic or ovate, slightly elevated, white; axillary buds suborbicular, about 1 mm. long. Leaves narrowly elliptic, the apex acute to acuminate, the base attenuate, the margins serrulate, 6-10 cm. long, 2-4 cm. wide, membranous, glabrous, the primary lateral veins 7 pairs, elevated below, plane and distinct above, the veinlets sliehtlv elevated beneath, visible above: bout 1 mm 12 mm. long. Inflorescences long, the primary axillary, 1- to 4-branched, racemiform, usually 2.0-3.5 cm. long, the peduncles glabrous, obscure to about 5 mm. long, the secondary peduncles about 3-8 mm. long, with 2-3 prophylls; flowers bisexual, white, the pedicels about 1.2 long, the articulation usually at the upper half of the stalk. Calyx lobes mm imbricate, ovate, marginate, scarious and slightly ciliate, about 1.5 mm. long; petals obovate, rotund, slightly erose, 2.5 mm. long and 1.2 mm. wide; disc fleshy, flat, the lobes depressed-subquadrate; stamens attached just beneath the disc margin, about 2.5 mm. long, the filaments linear, glabrous, the anthers suborbicular about 2 mm blunt. Fruits mm [Vol. 42 282 ANNALS OF THE MISSOURI BOTANICAL GARDEN septa 1.0—2.5 mm. wide, 1 -seeded; seeds broadly ellipsoid, 10—12 mm. long and 7 mm. wide, pinkish-brown, shining, the areolae distinct. In forests, at altitudes 790-2,300 m.; Mexico; flowering from March to June. Mexico: canyons of mountains above Cuernavaca, Pringle 6842 (MO, type of C. pringlei; GH, L, S, UC) ; Temascaltepec, Hinton 290 (F), 3506 (GH, US), 3574 (NY, US), 3717, 6076, 7203 (GH, US), 7380 (US), 9020 (GH, US), durango: San Ramon, Palmer 106 (MO, type of C. longipes; US), jalisco: McVaugb 10246, 10308, 13906 (MICH), michoacan: Tancitaro, Leavenworth 60 Hoogstraal 1023 (F, GH, MICH, MO, NY), morelos: Sierra de Tepoxtlan, Pringle 6998 (GH, MICH, MO, NY, S, UC) . The type specimen of Celastrus longipes seems to have larger leaves and longer peduncles than the type of C. pringlei. These characters are variable in all the specimens I have examined, and I think that they might be due to the differences in altitude. I have found some floriferous branches which are subtended by foliage leaves. This subtending foliage leaf may be from the previous year's growth. For example, a specimen (McVaugh 10308), collected at altitudes 2,400-2,600 m. in Jalisco, Mexico, on April 14, has a main branch bearing three floriferous branches. Each of the branches is subtended by a foliage leaf. When I made a cross-section of the main branch two porous rings of spring wood showed clearly. From this I assumed it might be an evergreen species, but, Dr. McVaugh told me he has col- lected another specimen in fruit with all of the leaves fallen off. 29. Celastrus vulcanicolus Donn. Smith, in Bot. Gaz. 61:373. 1916. (T.: Donn. Smith 2549, US! ) . Celastrus chiapensis Lundell, in Lilloa 4:380. 1939. (T.: Matuda 2080, MICH!). Celastrus siltepecanus Lundell, in Wrightia 1:155. 1946. (T.: Matuda 5192, MO!). Maytenus williamsii A. Molina R., in Ceiba 1:258. 1951. (T.: Merrill et al 15640, F!). Scandent shrubs up to 7 m. tall; branches terete, glabrous, gray to reddish- brown, the lenticels obscure on the gray branches while distinct on the reddish- brown ones, orbicular or ovate; axillary buds globose, about 1 mm. in diameter. Leaves elliptic, the apex acuminate or acute, the base cuneate or obtuse, the margins entire, repand, or slightly serrulate, 6-12 cm. long, 2.5-5.0 cm. wide, firmly mem- branous, glabrous, the primary lateral veins 7-9 pairs, curved toward the apex, below pet ary flowered, the primary peduncles glabrous, obsolete to about 5 mm. long, the secondary peduncles about 3 mm. long, usually with 2 prophylls; flowers bisexual, pale green, the pedicels about 2 mm. long, the articulation usually at the middle or lower half of the stalk. Calyx lobes ovate or deltoid, thick, obtuse, marginate, scarious, slightly erose, about 1 mm. long; petals oblong-elliptic, obtuse, minutely erose, about 2.0-2.5 mm. long and 1 mm. wide; disc flesh v, flat, about 2 mm. in mar gin, about 2 mm. long, the filaments filiform, glabrous, the anthers subglobose, slightly apiculate; pistil 1.5 mm. long, the style slender and blunt. Fruits ellipsoid, the valves broadly elliptic, about 14-18 mm. long and 7-10 mm. wide, 1 -seeded; 1955] HOU REVISION OF THE GENUS CELASTRUS 283 seeds ellipsoid, 12-14 mm. long and 6-9 mm. wide, reddish-brown, shining, the areolae obscure. In forests, at altitudes 1,300-2,400 m.; Mexico, Guatemala, Honduras, and El Salvador; flowering from December to January. El Salvador: Santa Ana, north of Metapan, Carlson goo (F). Honduras: Dept. Morazan, Merrill et al. 1 564.0 (F, type of May ten us williamsii; US) ; San Juancito, Williams d Molina R. 17100 (F). Guatemala: Dept. Alta Verapaz, Standley 71349 (F); Dept. Chiquimula, Steyer- mark 31480 (F); Dept. Quezaltenango, Steyermark 33635, 33692, 33775 (F), Standley 65402 (F), 65411, 86972 (F, MICH); Dept. Sacatepequez, Standley 63667 (F), Donn. Smith 2549 (US, type of C. vulcanicolus) . Mexico: chiapas: Fraylesca, Siltepec, Matuda 5192 (MO, type of C. siltepecanus) ; Cascada, Siltepec, Matuda 5148 (F); Mt. Ovando, Matuda 2080 (MICH, holotype of C. chiapensis) , 3944 (F, MICH, MO, NY, US), 16396 (UC). hidalgo: Molango, Moore 2698 (GH). oaxaca: Cumbre de Talea, Reko 4019 (US). This species is closely related to Celastrus pringlei Rose, but can be distinguished from it by the entire or remotely serrulate leaf margins, the obscure connectives, the larger fruits with wider septa, and especially the geographical distribution. The type specimens of Celastrus siltepecanus Lundell is similar to the present species except for the distinct and slightly elevated veins on both surfaces of the leaves, and the small distinct lenticels borne on the branches. These characters might be caused by the environment and are within the range of variation of the species. The specimens assigned to Maytenus williamsii A. Molina R. are Celastrus and are congruent with the present species. The leaves and fruits are smaller, which might be due to the high altitude of the habitat. 30. Celastrus casearhfolius Lundell, in Lilloa 4:379. 1939. (T.: Lehmann s.n.,T?l). Branchlets terete, glabrous, reddish-brown, densely lenticellate, the lenticels llary Leaves oblanceolate- mm oblong or elliptic-oblong, the apex acute to shortly acuminate, the base obtuse, the margins remotely serrulate, 4.5-10.5 cm. long, 1.4-4.0 cm. wide, chartaceous, glabrous, the primary lateral veins 7-9 pairs, veins and veinlets slightly elevated below, obscure above; petioles 3—5 mm. long. Inflorescences axillary, racemiform, 1- or 2-branched, up to 2.5 cm. long, the primary peduncles obscure, the secondary long. Flowers (young) bisexual, the pedicels obsolete to about 0.8 mm. long, the articulation at the upper half of the stalk. Calyx lobes imbricate, ovate or ovate-deltoid, erose-ciliolate; petals ovate-oblong, obtuse, sub- entire; disc fleshy, flat, the lobes obscure, the stamens arising from the margin of the disc, the filaments glabrous, the anthers suborbicular, apiculate; pistil ovoid, the style short and columnar, the stigmata obscure. Fruits cylindric, the valves oblong, about 16 mm. long and 7 mm. wide, 1 -seeded; seeds cylindric, obtuse at both ends, 1 5 mm. lone and 6 mm. wide, black and smooth. Colomb June [Vol. 43 284 ANNALS OF THE MISSOURI BOTANICAL GARDEN Colombia: Dept. de Antioquia, Daniel 329$ (F); highlands of Popayan, Lehmann s.n., 1,600-2,000 m. elev., May 1889 (F, holotype), 399 (L, NY). Celastrus caseariifolius is characterized by the innumerable, elevated and crowded lenticels, very much resembling a crowded colony of plant lice (aphids). Further, its chartaceous and oblanceolate-oblong leaves are distinctive. 31. Celastrus meridensis Pittier, in Bol. Soc. Venez. Cienc. Nat. 3:423. 1927. (T.: Gehringer 298, US!). Maytenus meridensis (Pittier) Cuatr. in Fieldiana, Bot. 27 2 :82. 1951. Scandent shrubs; branches slightly striate, glabrous, brownish, the lenticels sparse, elliptic, slightly elevated, white, obscure on the young branchlets; axillary buds ovoid, about 2 mm. long. Leaves ovate, ovate-oblong, or obovate, the apex rotund to abruptly acute, the base cuneate to rotund, the margins remotely crenate- serrate, 4.5-7.0 cm. long, 2-5 cm. wide, firmly chartaceous, glabrous, the primary lateral veins 5-7 pairs, elevated below, plane and distinct above, the veinlets visible below, obscure above; stipules filiform, about 1 mm. long; petioles 3-7 mm. long. Inflorescences axillary, simple, racemiform, up to 5 cm. long, the primary peduncles obscure, glabrous, the secondary peduncles about 1 mm. long, with 2 small pro- phylls; flowers bisexual, white, the pedicels obscure, accrescent, up to 3 mm. long on the fruiting specimens. Calyx lobes imbricate, ovate, rotund, slightly erose, thick, about 1.2 mm. long; petals oblong, obtuse, slightly erose, about 2.6 mm. long and 1.3 mm. wide; disc fleshy, flat, about 1.6 mm. in diameter, the lobes depressed, subreniform; stamens attached just beneath the disc margin, about 2 mm. long, the filaments linear, glabrous, the anthers ovoid, obtuse; pistil pear- shaped, about 1.5 mm. long, the style short, columnar and blunt. Fruits ovoid, the valves broadly ovate to suborbicular, about 11 mm. long and 7.5-9.0 mm. wide, the septa 2-3 mm. wide, 1 -seeded; seeds ellipsoid, 7 mm. long and 5 mm. wide, reddish-brown, shining, the areolae obscure. In thickets, at altitudes 2,490-2,700 m.; Colombia and Venezuela; flowering in July. Colombia: Cordillera Oriental, Dept. Boyaca, Cuatrecasas 1813, 1831 (F, US). Venezuela: Mucuruba, Gehriger 298 (US, type; F, NY). This species distinctly belongs to Celastrus. The morphological characters match the generic characters very well. The type specimen is a scandent, flower- ing plant. In addition to the type, I have several fruiting specimens at hand, which confirm this view. Cuatrecasas 42 transferred this species to Maytenus based on his own collections 1 (1813 and 1831, F, US). He says, "My specimens were obtained from trees, justifying their inclusion in the genus Maytenus". On examining the specimens, Cuatrecasas scandent plant. specimens Since all species are scandent shrubs, I assume that Dr. Cuatrecasas* specimens m trees" were in realitv scandent. 42 In Fieldiana, Bot. 27 2 :82. 1951. 1955] HOU REVISION OF THE GENUS CELASTRUS 285 Doubtful Species The type specimens or representative specimens of the following species are not available, while their original descriptions alone are not sufficient to place them. Celastrus discolor Levi, in Bull Geogr. Bot. 24:142. 1914 (T.: Cavalerie 3919). China. Celastrus grenadensis Urb. Symb. Antill. 5:51. 1904 (T.: Eggers 6222). Ind. Occ. Celastrus microcarpus D. Don, Prod. Fl. Nep. 191. 1825 (T.: Kamroop s.n.). Reg. Himal. Celastrus racemosus var. trinitensis Urb. Symb. Antill. 5:52. 1904 (T.: Baptist e 58 57) . Trinidad, Celastrus repandus Bl. Bijdr. Fl. Ned. Ind. 1145. 1825 (T.: none). Java. Celastrus reticulatus Wang in Chin. Jour. Bot. 2:68. 1937 (T.: Leu 233). China. Excluded Species omitte of which are species of Gymnosporia or Maytenus and require special study of isposition Wall. Cat. no. 4342. 1831 = Chailletia gelonioides Hook. f. Fl. Brit. Ind. 1:570. 1875. astrus adenophylla Miq. Ann. Mus. Bi Thunb. Fl. Jap. 78. 1784. 1865=Ilex crenata Celastrus alatus Thunb. Fl. Jap. 98. 1784 = Euonymus thunbergianus Bl. Bijdr. Fl. Ned. Ind. 1147. 1825. Celastrus alpestris Bl. Bijdr. Fl. Ned. Ind. 1145. 1825 = Perrottetia alpestris Loes. in Engl. & PrantL Nat. Pflanzenfam. Ill, 5:220. 1892. l Linnaea 15:458. 1841 = Acanthothamnus U. S. Nat. Herb. 23:684. 1923. \phyll Contr aquif 1856, nom. nud. Wall. Cat. no. 4319. 1831 = Gymnosporia neglecta Hook. Fl. Brit. Ind. 1:619. 1875. Celastrus bilocularis F. Muell. in Trans. Phil. Inst. Vict. 3:31. 1859 = Maytenus bilocularis (F. Muell.) Loes. in Engl. & Prantl, Nat. Pflanzenfam. 2 Abt. 20b:135. 1942. Jack, in Malay. Misc. 1:19. 1820 Wall Cat. no. 4340. 1831. Celastrus boaria Baill. Hist. PL 6:26. 1877, in text, nom. nud. = Maytenus boaria Molina, Saggio Stor. Nat. Chile, ed. 1, 177. 1778. Celastrus bodinieri Levi, in Fedde, Rep. Sp. Nov. 13:263. 1914 = Ilex purpurea Hassk. Cat. PL Bogor. 230. 1844; Rehd. in Jour. Arnold Arb. 14:239. 1933. Celastrus buxifolia Wall, in Roxb. Hort. Beng. 18. 1814, nom. nud.; Fl. Ind. ed. 286 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol.. 42 Carey & Wall. 2:396. 1824, in syn.: Celastrus rigidus Wall. = Gymnosporia wallichiana Laws, in Hook. Fl. Brit. Ind. 1:621. 1875. Celastrus cavaleriei Levi, in Fedde, Rep. Sp. Nov. 13:262. 1914 = Myrsine semi- serrata Wall, in Roxb. Fl. Ind. ed. Carey & Wall. 2:293. 1824; Rehd. in Jour. Arnold Arb. 15:292. 1934. Celastrus chungii Merr. in Sunyatsenia 3:253. 1937 = Tripterygium Wilfordh Hook, in Benth. & Hook. Gen. Pi. 1:368. 1862-67. Celastrus circumcissus Pavon, mss.; Briq. in Ann. Conserv. & Jard. Bot. Geneve 20:253. 1919 = Maytenus orbicularis (Willd.) Loes. in Engl. Bot. Jahrb. 50 (Beibl. 111):10. 1913. Celastrus colombianus Cuatrec. in Fieldiana Bot. 27 2 :81. 1951 = Ilex scandens Cua tree, in Lloydia 11:207. 1949. Celastrus confertus Ruiz & Pavon, Fl. Peruv. 3:7. 1802 — Maytenus confertus (Ruiz & Pavon) Loes. in Engl. & Prantl, Nat. Pflanzenfam. 2 Abt. 20b: 146. 1942. Celastrus crenatus Forst. f. Prod. 19. 1786 = Gymnosporia crenata (Forst.) Seem. Fl. Vit. 1:40. 1865. Celastrus crenatus sensu F. Brown in Bull. Bishop Mus. Honolulu no. 130:158. 1935, non Forst. = Gymnosporia sp., ex char. Celastrus crenatus sensu Hook. & Arn. Bot. Beechey Voy. 61. 1841 = Gymno- sporia vitiensis (A. Gray) Seem. Fl. Vit. 1:41. 1865. Celastrus crenatus Roth, Nov. PL Sp. 156. 1821 = Gymnosporia Montana Benth. Fl. Austral. 1:400. 1863. Celastrus crenulatus Wall. Cat. no. 4323. 1831, nom. nud. Celastrus cuneifolius (Wr. ex A. Gray) Gomez de la Maza in Ann. Inst. Segunda Ensefianza 2:172. 1895 = Euonymus cuneifolius C. Wright ex A. Gray in Mem. Am. Acad. n. s. 8:171. 1861 — not Celastrus (capsules 2-valved). Celastrus cunninghamii F. Muell. in Trans. Phil. Inst. Vict. 3:30. 1859 = May- tenus cunninghamii (F. Muell.) Loes. in Engl. & Prantl, Nat. Pflanzenfam. 2 Abt. 20b:136. 1942. flora F. M. Bailey in Queensland Agr. J 29:178, pi. 22. 1912 = Maytenus cunninghamii (F. Muell.) Loes. in Engl. & Prantl, loc. cit. 1942. Celastrus dilatatus Thunb. in Trans. Linn. Soc. 2:332. 1794 = Orixa japonica Thunb. Nov. Gen. Pi. 3:57. 1783. disp 1859 = Maytenus dispermus (F. Muell.) Loes. in Engl. & Prantl, Nat. Pflanzenfam. Abt. 2. 20b:135. 1942. Celastrus diversi folia Hemsl. in Jour. Linn. Soc. 23:123. 1886 = Gymnosporia diversifolia Maxim, in Bull. Acad. Sci. St. Petersb. Ill, 27:459. 1881. Celastrus dubia Spreng. Syst. Veg. 1:774. 1825— not Celastrus (capsules 2-valved). Celastrus emarginatus Ruiz & Pavon, Fl. Peruv. 3:6, t. 229, f. a. 1802 = May- tenus 1919. emarginatus Willd. Sp. PI. 1 2 :1128. 1798 = Gymnosporia emarginata Enum. PI. Zeyl. 409. 1864. Celastrus esquirolianus Levi. Fl. Kouy-Tcheou, 69. 1914 Sieb. & Zucc. in Abh. Bayer. Akad. Muench. II, 4:146. (Fl. Jap 1:38). 1845; Rehd. in Jour. Arnold Arb. 15:13. 1934. CRENATUS 1955] HOU REVISION OF THE GENUS CELASTRUS 287 Celastrus esquirolii Levi, in Fedde, Rep. Sp. Nov. 13:262. 1914 = Sabia parvi- var. nitidissima Levi. Fl. Kouy-Tcheou, 379. 1915; Rehd. in Wall J 1934. Celastrus euonymoidea Levi. Fl. Kouy-Tcheou, 419. 1915 = Grewia feddei (Levi.) Burret in Notizbl. Bot. Gart. Berlin 9:678. 1926. / feddei pi. 280). 1896 = Gymnosporia sp., ex ill. 13:263. 1914, quoad specim. Esquirol 3189 = Grewia henryi Burret in Notizbl. Bot. Gart. Berlin 9:674. J 1937. / 1914, excl. Esquirol 31 89 Grewia feddei (Levi.) Burret, loc. cit. 678. 1926. us finlaysonianus Wall. Cat. no. 4324. 1831, nom. nud. floribundus fournieri P; 1841, nom. nud. f Panch & Sebert, Not. Bois Nouv. Caled. 234. 1874 = May- tenus fournieri (Panch. & Sebert) Loes. in Engl. & Prantl Nat. Pflanzenfam. 2, Abt. 20b:138. 1942. Vahl, Symb. Bot. 2:42. 1791 = Elaeodendron glaucum aucus Pers. Syn. PL 1:241. 1805. Celastrus haenkea Spreng. Syst. Veg. 4 2 :88. 1827 = Schoepfia flexuosa Roem. & Schult. Syst. Veg. 5:160. 1819. Celastrus hamelii Spreng. Syst. Veg. 1:774. 1825 = Rhamnus racemosus Duham. Traite Arb. et Arbust. ed. nov. 3:48. 1806. Celastrus heterophyllus Savi in Mem. Accad. Sci. Torino 38:163 (Mem. de C. G. Savi, p. 11, tab. II, /. 2.). 1835 — not Celastrus (fruit triangular). Celastrus heyneana Roth in Roem. & Schult. Syst. Veg. 5:421. 1819, p.p. = Gymnosporia heyneana Laws, in Hook. Fl. Brit. 1:620. 1875. Celastrus ilicifolius Schrad. in Goett. Gel. Anz. 1:716. 1821 = Maytenus truncatus Reiss. in Mart. Fl. Bras. II 1 : 5. 1861. Celastrus japonicus (Thunb.) Koch, Dendrol. 1:625. 1869 = Orlxa japonica Thunb. Nov. Gen. PL 3:57. 1783. Celastrus jodinii Steud. ex Gopp. in Gartenflora 3:312. 1854, in obs.: Ilex cunei- folia Hook. f. = Trichilia jodinii (Steud. ex Gopp.) Briq. in Candollea 6:21. 1935. Celastrus kouytchensis Levi, in Fedde, Rep. Sp. Nov. 13:264. 1914 = Rhamnus crenatus Sieb. & Zucc. in Abh. Bayer. Akad. Muench. II, 4:146. (Fl. Jap. Fam. Nat. 1:38.) 1845. Celastrus leptopus Drake in Grandidier, Hist. Madag. 35. (Hist. Nat. PI. Atlas 3, pi. 28o A ). 1896 = Gymnosporia sp., ex ill. Celastrus linearis var. madagascariensis Drake in Grandidier, loc. cit., pL 286 s . 1896 = Gymnosporia sp., ex ill. Celastrus lineatus (Wr.) Gomez de la Maza in Ann. Soc. Espan. Hist. Nat. 19:239. 1890 = Maytenus lineatus C. Wr. in Griseb. Cat. PL Cub. 54. 1866. 288 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 42 J' 1904 = Gymnosporia sp. Wall, in Roxb. Fl. Ind. ed. Carey & Wall. 2:400. 1824— not Celastrus (ovary many-celled). astrus lycioides Bross. ex Willd. not Celastrus (branches spiny). 1819 Celastrus lyi Levi, in Fedde, Rep. Sp. Nov. 13:264. 1914 = Rhamnus esquirolii Levi, in Fedde, Rep. Sp. Nov. 10:473. 1912. Celastrus macrocarpus Ruiz & Pavon, FL Peruv. 3:8, /. 230, f.b. 1802 = May- TENUS MACROCARPUS (Ruiz & PaV.) Geneve 20:361. 1919. Jard Celastrus ? magellanicus DC. Prod. 2:8. 1825 = Maytenus magellanicanus Hook. f. Fl. Antarct. 254. 1847. Celastrus mairei Levi, in Fedde, Rep. Spec. Nov. 13:264. 1934 = Sabia yunnan- ensis Franch. in Bull. Soc. Bot. Fr. 33:465. 1886; Loes. in Ber. Deut. Bot. Ges. 32:543. 1914. Willem 1798; Roem. & Schult. Syst. Veg. 5:428. 1819 — not Celastrus (leaves ternate). Celastrus maytenus Willd. Sp. PL 1 2 :1127. 1798 = Maytenus boaria Molina, Saggio Stor. Nat. Chile ed. 1. 177. 1778. Celastrus mexicanus Moc. & Sesse ex DC. Prod. 2:8. 1825 = Wimmeria mexi- cana (Moc. & Sesse) Lundell in Bull. Torr. Bot. Club 67:618. 1940. Celastrus micrantha Roxb. Hort. Beng. 86. 1814, nom. nud.; Fl. Ind. ed. Carey & Wall. 2:393. 1824 — not Celastrus (leaves pinnate). Celastrus mollis Decne. in Rev. Hort. II, 4:425. 1845-46 — not Celastrus (leaves opposite) montana PL Sp. 154. 1821== Gymnosporia Montana Benth 1819; Roth, Nov. astrus may a O. Kuntze, Rev. Gen. 3 2 :37. 1898 = Moya spinosa Griseb. PL Lorentz. 63, pi. I, f. 3. 1874. astrus muelleri Benth. Fl. Austral. 1:399. 1863 = Maytenus sp. astrus myrtifolius Linn. Sp. PL 196. 1753 = Prunus myrtifolia (Linn.) Urban, Syn. Ant. 5:93. 1904. astrus neglect a Wall. Cat. no. 4341. 1831 = Gymnosporia neglecta Laws. Hook. FL Brit. Ind. 1:619. 1875. nepalensis jm Wight 1840 = Pittosporum flori- k 1834. Bot. Bemerk. 34. 1844 = Simmondsia Jour. Bot. 3:400, t. 16. 1844. Celastrus obtusi folia Roxb. Hort. Beng. 86. 1814, nom. nud.; Fl. Ind. ed. Carey & Wall. 2:393. 1824 = Gymnosporia trigyna Baker, Fl. Maurit. 50. 1877. Celastrus octogonus L'Her. Sert. 7. 1788 = Maytenus orbicularis (Willd.) Loes. in Engl. Bot. Jahrb. 50 (Beibl. Ill) :10. 1913. 1955] HOU REVISION OF THE GENUS CELASTRUS 289 astrus op posit a Wall, in Roxb. Fl. Ind. ed. Carey & Wall. 2:398. 1824 Pleurostylia cochinchinensis Pierre, FL For. Cochinch. Fasc. 20, sub. /. S°5y in text. 1894. astrus orbicularis Willd. mss. ex HBK. Nov. Gen. et Sn. PL 7:65. 1825 (Willd.) Loes. in Engl. Bot. J 1913. Celastrus orixa Sieb. & Zucc. in Abh. Bayer. Akad. Muench. II, 4:150. 1845 Orixa japonica Thunb. Nov. Gen. PL 3:57. 1783. ilifolius Steud. Nom. Bot. ed. 2, 1:2 Hook. Fl. Brit. Ind. 1:619. 1875. 1840 = Gymnosporia ovata Celastrus ovatus Hill. Veg. Syst. 13:62, /. 12. 1824 = Colubrina ferruginosa Brongn. in Ann. Sci. Nat. I, 10:369. 1827. Celastrus ovatus Wall. Cat. no. 4308. 1831 = Maytenus ovata (Wall.) Loes. in Engl. & Prantl, Nat. Pflanzenfam. 2 Aufl. 20b: 140. 1942. Celastrus oxyphyllus Wall. Cat. no. 4312. 1831 = Gymnosporia acuminata Hook. Fl. Brit. Ind. 1:619. 1875. Celastrus pallidus Wall. Cat. no. 4307. 1831 = Gymnosporia Montana Benth. Fl. Austral. 1:400. 1863. Celastrus parviflorus Vahl, Symb. Bot. 1:21. 1790 = Gymnosporia sp. Celastrus parvifolius A. Rich. Ess. Fl. Cub. 349. 1845 — not Celastrus (fruits 2- valved and 1- to 2 -seeded). Celastrus pauciflora Wall, in Roxb. Fl. Ind. ed. Carey & Wall. 2:400. 1824— not Celastrus (ovary 1 -celled). Celastrus pentagyna Zipp. ex Span, in Linnaea 15:186. 1841, nom. nud. Celastrus quadrangulatus Schrad. in Goett. Gel. Anz. 1:716. 1821= Maytenus quadrangulatus (Schrad.) Loes. in Engl. & Prantl, Nat. Pflanzenfam. 2 Aufl. 20b: 142. 1942. Celastrus retusa Poir. in Lam. Encycl. Meth. Suppl. 2:146. 1811 = Maytenus retusa Briq. in Ann. Conserv. & Jard. Bot. Geneve 20:351. 1919. Celastrus rhombifolius Hook. & Arn. in Hook. Bot. Misc. 3:170. 1833 = Iodina rhombifolia Hook. & Arn. ex Reissek, in Mart. Fl. Bras. 1 1 1 :78. 1861. Celastrus richardi Gomez de la Maza, Dice. Bot. Nom. Vulg. 25. 1889 = May- tenus buxifolius Griseb. Cat. PL Cub. 53. 1866. Celastrus richardi y. latifolius Gomez de la Maza in Ann. Soc. Espafi. Hist. Nat. 19:239. 1890, nom. nud. Celastrus richardi 8. cochlearifolius Gomez de la Maza, loc. cit. 239. 1890 Maytenus cochlearifolius Griseb. Cat. PL Cub. 53. 1866. Celastrus richardi c. elaeodendroides Gomez de la Maza, loc. cit. 239. 1890 Maytenus elaeodendroides Griseb. loc. cit. 54. 1866. Celastrus rigida Wall, in Roxb. Fl. Ind. ed. Carey & Wall. 2:396. 1824 = Gymno- sporia wallichiana Laws, in Hook. Fl. Brit. Ind. 1:621. 1875. Celastrus robustus Roxb. Hort. Beng. 18. 1814, nom. nud.; Fl. Ind. ed. Carey & Wall. 2:395. 1824 = Kurrimia sp. [Vol. 42 290 ANNALS OF THE MISSOURI BOTANICAL GARDEN Wight 1834 = Gymno- sporia rothiana Laws, in Hook. Fl. Brit. Ind. 1:620. 1875. astrus royleanus Wall. Cat. no. 4317. 1831 loc. cit. 620. 1875. istrus rufa Wall, in Roxb. Fl. Ind. ed. Care sporia rufa Laws. loc. cit. 620. 1875. ROYLEANA 1824 = Gymno- licifol, Hook 14:242. 1933. 1914 = Ilex macro- . in Jour. Arnold Arb. Celastrus seguini Levi, in Fedde, Rep. Spec. Nov. 13:262. 1914 = Myrsine semi- serrata Wall, in Roxb. Fl. Ind. ed. Carey & Wall. 2:293. 1824. Celastrus semiarillata Turcz in Bull. Soc. Nat. Mosc. 36 : :599. 1863 — not Celastrus (branches spiny). Celastrus sepiarius Dennst. Schlus. Hort. Ind. Malab. 31. 1818, nom. nud. Celastrus serrulatus Roth, Nov. PL Sp. 155. 1821, p. p. = Gymnosporia rothiana Laws, in Hook. Fl. Brit. Ind. 1:620. 1875. Celastrus spicatus Veil. Fl. Flum. 92, /. 138. 1827 = Gouania corylifolia Raddi, in Mem. Soc. Ital. (Modena) 18, /. 394. 1820. Celastrus spinifolius Larrafiaga, Escritos D. A. Larrafiaga 2:96. 1923— not Celastrus (capsules bivalved). Celastrus spinosus Royle, 111. Bot. Himal. 157. 1835 = Gymnosporia royleana Laws, in Hook. Fl. Brit. Ind. 1:620. 1875. Celastrus striatus Thunb. Fl. Jap. 98. 1784 = Euonymus altus (Thunb.) Sieb. Syn. PI. Oecon. Jap. 49. 1830. Celastrus suaveolens Levi, in Fedde, Rep. Sp. Nov. 13:263. 1914 = Ilex suave- olens (Levi.) Loes. in Ber. Deut. Bot. Ges. 32:541. 1914. Celastrus tetramerus Standley in Contr. U. S. Nat. Herb. 23:679. 1923 = Phyllanthus sp. Celastrus tristis Levi, in Fedde, Rep. Sp. Nov. 13:263. 1914 = Rhamnus napal- (W boaria Molina, Saggio Stor. Nat. Chile, ed. 1, 177. 1778. 1875. f. a. 1802 = Maytenus J Jahrb. 29:446. 1900. 1886 = Gymnosporia Celastrus verticillata Roxb. Hort. Beng. 18. 1814, nom. nud.; Fl. Ind. ed. Carey Wall. 2:391. 1824 Fl. Ind. Orient. 154. 1834. Wight TEN US /. b. 1802 = May- 1825. 1955] HOU REVISION OF THE GENUS CELASTRUS 291 Celastrus vitiensis (A. Gray) [incorrectly ascribed to Benth. & Hook, by] Drake, Fl. Polyn. France, 30. 1893 = Catha vitiensis A. Gray, Bot. Phanerog., Wilkes U. S. Expl. Exped. 287, Gray) Seem. Fl. Vit. 40. 1865. 1854 Gymnosporia vitiensis (A. Celastrus wallichianus Spreng. Syst. Veg. 5, Index 150. 1828 = Gymnosporia wallichiana Laws, in Hook. FL Brit. Ind. 1:621. 1875. zstrus wallichianus Wall, in Roxb. Fl. Ind. ed. Carey & Wall. 2:4 Gymnosporia wallichiana Laws, in Hook. loc. cit. 621. 1875. 1824 Celastrus tvallichiiG. Don, Gen. Syst. 2:8. 1832 = Linn. — not Celastrus. Celastrus wightianus Wall. Cat. no. 4332. 1831 Wall Pleurostylia cochin- chinensis Pierre, Fl. For. Cochinch. Fasc. 20, sub. /. 305 in text. 1894. Celastrus yunnanensis Levi. Cat. PI. Yun-Nan, 32. 1915 = Prema parvilimba P'ei in Mem. Sci. Soc. China 1:62. 1932; Rehd. in Jour. Arnold Arb. 15:324. 1934. Celastrus zeylanica Roth ex Roem. & Schult. Syst. Veg. 5:427. 1819 = Scutia commersonii Brongn. in Ann. Sci. Nat. I, 10:363. 1827. Enumeration of the Species Subgenus I. Celastrus Series 1. Paniculati 1. paniculatus Willd. la. lb. ssp. paniculatus ssp. serratus (Blanco) Ding Hou lc. 2. 3. 4. 5. 6. 7. ssp. multiflorus (Roxb.) Ding Hou novoguineensis Merr. & Perry subspicatus Hook. f. richii A. Gray madagascariensis Loes. angulatus Maxim. scandens Linn. Series 2. Axillares 8. monospermus Roxb. 9. m&nospermoides Loes. 10. hindsii Benth. 11. glaucophyllus Rehd. & Wils 12. hooker i Prain 13. membranifolius Prain 14. vanioti (Levi.) Rehd. 15. hypoleucus (Oliver) Warburg 16. gemmatus Loes. 17. orbiculatus Thunb. 18. rosthornianus Loes. 19. punctatus Thunb. 20. kusanoi Hayata 21. hirsutus Comber 22. stylosus Wall. 22 a. ssp. stylosus 22b. ssp. glaber Ding Hou 23. aculeatus Merr. 24. flagellars Rupr. Subgenus II. Racemocelastrus 25. lenticellatus Lundell 26. racemosus (Reiss.) Loes. 27. panamensis Lundell 28. pringlei Rose 29. vulcanicolus Donn. Smith 30. caseariifolius Lundell 31. meridensis Pittier 292 [Vol. 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN Index to Exsiccatae Italicized numerals refer to collectors' numbers, s. n. (sine numero) to unnumbered collections; parenthetical numerals refer to the numerals of the species conserved in this revision. Adams, J. W. 1 326, 1656, 286 5, 5016 (7) . Adams, J. W. & M. L. T. 1426, 21 31 (7). Adams, J. W., M. L. T. Adams & E. T. Wherry 1 519 (7). Adams, J. W. & G. Hopkins 907 (7). Adams, J. W. & R. Tash 568 (7). Adams, J. W. & E. T. WTierry 2 300 (7). Ahern's collector 2889 (Bureau of Sci. no.) (lb). Alcasid, — , & G. E. Edaiio 4.699 (lb). Alims, — . 1605, 2103 (9). Allard, H. A. 473, 5654 (7). D'Alleizette, Ch. 1288 (5). Allen, P. H. 319 (27). Anderson, T. 104 (22a); 107, 108, IOO (8). Andrews, D. W. 39 (7). Backer, C. A. 37252 (lc). Bacon, E. S. s. n. (7). Baenitz, C. IIl8 (7). Balansa, M. 1870 (3). Ballard, C. A. s. «. (7). Barber, C. A. 6846, 7018 (la). Barnhart, J. H. 580, 1 194, 1394 (7). Bartlett, H. H. 6864 (9). Batchelder, C. F. s.n. (7). Bennett, W. M. 419, 1842, 2223, 5690, 5720, 6308, 6902, 7923, 8682 (7). Bergman, H. F. 2101, 2630, 2855, 3120 Bergman, S. 439 (17). Berkheimer, D. 4756 (7). Bicknell, E. P. 3836 (7). Biswas, K. 2033 (8). Bock, C. & A. v. Rosthorn 790 (6). Bohlin, B. 104 (17). Boswastore, — . Brandis, — • 73 (22a). 2032 (la). Brass, L. J. 12999 (9). Breitung, A. J. 6172 (7). Brooks, W. P. 44, 544 (17). Bush, B. F. 118, 732 (7). Bush, Y. 60 (7). Calder, C. C. & M. S. Ramaswami 6*d (la). ^ Carr, C. E. 11260 (3) ; 13036 (9) ; 15698 Carlson, M. C. goo (29). Carter, H. G. 466 (lc). Cavalerie, J. 406 (IB) ; 3976 (10). Chandler, H. P. 510 (7). Chase, V. H. 255, 853, 4057, 16652 (7). Chen, P. E. 2447, 2701 (19). Chen, Y. IOOO (18). Cheng, C. H. J 57 3 (23). Cheo, T. Y. & L. Yen 1 03 (17). Chiao, C. Y. 1720 (18); 1721 (11); 21 16, 2126, 2395, 2612, 3016 (17); 1 8641 (22b); 18768 (23). Chiao, C. Y. & C. S. Fan 458 (10); 767 (11). Chien, S. S. 5920 (18). Ching, R. C. 1 336, 1664 (23); 1677 (19); 2508 (23); 2638 (16); 2661 (19); 2700 (11); 2845 (22b); jodj (15); 316S, 546l (22b); 6057 (14); 7080 (22b); 7/20, 7525 (8); 7566 (16); A?05 (22b); 8433 (8); 20^p, 20<5jc? (11); 20682 (16); 2/67o > , 220^2 (11); 22II0 (18). Ching, — . & — . Tso 580 (17). Chou, K. H. 8215, 8460 (23). Chow, C. L. 5707 (18); 6448, 6468 (11). Chow, H. C. 315 (18); jdp (17); 405, 495 (6)1548 (10); 572 (18); 597 (6); 7 4777 (23); 47 ; 62 35> 7264 (23); 8436, 8446 (10). Chung, H. H. & S. C. Sun 592 (16). Chung, Z. S. 83605 (16). Clarke, C. B. 682c (12); 56/5 (10); /20*<5 (22a); 18678C (22a); 23<5o7B, 2^57, 26339 (la); 2670/B (22a); 27O40D (12); 27683 (8); 3575^ (12); 37949B (8); £rp//C (10); 45/52 (8); 46569 (12). Clemens, J. & M. S. Clemens 3523 (2); 458l, 5084 (9); 5/52, 5394 (2); 6253 <5<5cx5 2pp/p 1955J HOU REVISION OF THE GENUS CELASTRUS 293 31827, 32052, 32120, 32860, 33182, 50253, 50641, 51295 (9). Clemens, M. S. 11073 (2); 11680 (7); 43683 (3). Clokey, I. W. 2479 (7). Coert, J. H. 3758 (lb). Colani, E. 4049 (lc). Collect, H. 588, 672 (la). Corbett, — . & — . Williams 88 (7). Core, E. L. 3778 (7). Cory, V. L. 14152 (7). Cuatrecasas, J. 1813 (31). Cuiran, H. M. 7499 (lb). Cuming, H. 1209, 1 324 (lb). Curran, H. 3415 (lb). Curtis, C. 404, 2005 (9) . Daniel, H. 3295 (31). Darsteri, A. C. 1 7 (8). David, A. J 78 3 (17). Davis, John. 199, 716, 720, 3709, 6368, 7344 (7). Day, M. A. 43, 96 (7). Deam, Chas. C. 15456, 18042, 18240, 18331, 18810, 19320, 35238> 379H, 38 1 41 (7). Degener, Otto 1 4798, 14940, 14973, 14978, 15050, I53I5> J5330, 1*3*2, 15424, *5427 (4). Demaree, D. 6746, 6955, 11340, 11572 (7). Deplanch, — . 92 (3). Desiloa, F. s. n. (8). Dickason, F. G. 5845 (la). Dickey, S. S. 27 (7). Dorgelo, J. D. 2290 (lc). Dorsett, P. H. & W. J. Morse 619 (24); 1070, 7074, 7152, 7206 (17). Dougboo, — . 86225 02). Drummond, J. R. 21515, 21516, 21651, 21691 (la). Ducloux, F. 132 (16). Dudgeon, W. & L. A. Kenoyer 237 (la). Dunn, — . 44 (19); 2479, 2479B, 3420 (23). Duthie, L F. 821 1 (la). Eccles, R. G. 29 (7). Edafio, G. E. 999 (9); 77599 (lb). Edmondson, T. W. 3426, 3876 (7). Eggleston, W. W. & K. M. Wiegand 22503 (7). Elizabeth, E. E. 351 1, 3903 (7). Elmer, A. D. E. 11411 (9). Elwell, L. H. 202, 203 (7). Emig, W. H. 370 (7). Endert, F. H. 3846 (9). Enequist, J. 435 (7). Engelmann, G. 1107 (7). Esquirol, J. 4 O^). Evans, I. H. N. 1 3247 (lc). Evers, R. A. 143 (7). Evers, R. A., G. N. Jones & F. F. Jones 541 (7). Eyles, D. E. & M. S. Eyles Eyles, M. S. 8380 (7). 8505 (7). 190 (22b); 210, Faber, E. 227 (18). Fan, C. S. & Y. Y. Li 236 (11); 263 (22b). Fang, W. P. 571 (6); 595 0«); ^02* (11); JJ<52 (6); 2124 (16); 3089 (11); 3424, 3486 (18); J7J7, ^//d (11); 5529 (16); 55 13032, 14 6 78, 14835 (7). Forbes, H. O. £27, 961, 980, 980a (10); 3829 (22a). Forrest, G. 7812 (18); 8704 (16); 9396 ( 22b) ; 9948, IO069, IOI42, III52 (11); 7777J (16); 772 (6); /307 (16). Haneeflf, Md. & Md. Jan 2446 (9). Harbison, T. G. 1611 (7). Harsukh, — . 21 400 (la). Hayden, A. 9590, 10070 (7). Hay ward, H. E. 616 (7). Heiser, Charles B., Jr. 1801 (7). Heller, A. A. & E. G. Halbach 522 (7). Henrici, S. II 5 1 (7). Henry, A. 0986a (11); 1893 (20); 2084 (6)12837 (15); 31 1 5 (IS); 3241 (10); 54<>5 ( 6 ) ; J4P5, J495A, j/ (6); 7272 (17); 1288, 1314, 1712 (6); 1758 (17); J«?J/, 2^2, 2670 (6) ; 27p/, 27P5 ( 17) ; 2pj2, 3010, 3097 (6). Hieronymus, P. 443 (la). Hinckley, L. C. UOc? (7). Hinton, G. B. 2po, 3506,' 3574, 3717, 6076, 7230, 7380, 9020 (28). Hitchcock, A. S. 60 (7). Ho, H T. 60086 (8). Hooper, D. 38801, 39537 (la). Hopkinson, — . 297 (18). Home, J. 1 135 (4). Horr, W. H. & R. L. McGregor E493 (7). Horsey, R. E. 2po, 1283, 1488, 1545, 1581, 1748, 1754, 2258, 2380 (7). How, F. C. 70368 (10); 70573 (lc); 73564 (8). How, F. C. & N. K. Chun 70273 ( 8 ) . Hsiung, Y. K. 5516 (23); 5655, 570 1 (16); 57^ (23). Hu, H. H. 784, 889 (23); 1 1 19, 1585 (16); 1631 (24). Hu, W. K. 7380 (11). Huk, A. (52 (la). Hunnewell, F. W. 5<5 506 161 (19); 162 (17). Ome, Tsuro (?) (17) Otanes, F. 500 (19). 17985 (lb). Ownbey, M. 7570 (7). Painter, J. H. 7794 (7). Palmer, E. 706 (28). Palmer, E. J. 772, 772, 13993, 15237* 16145, 18923, 21238, 22124, 22814, 26057, 26960, 27583, 28521, 28785, 3 3JI34, 32933> 34170, 35954, 30053, 30900, 37i3i> 41926-A, 43587, 45308, 45570, 45645, 490i6 (7) Pammel, L. H. 43, 269 (7). Pammel, L. H., H. H. Hume & S. R. Fitz 1647 (7). Pammel, L. H., H. E. Pammel & P. S. Mo Nutt 522 (7). Paradhan, K. S. 34 (la). Parham, H. B. R. (Mrs.) 3 (4). Pease, A. S. 14, 3410, 3671, 11963 (7). Peattie, D. C. 48 A (7). Pennell, F. W. 1 1 48, 16382, 16487 (7). Petelot, A. 340, 863, 991 (10); 3184 (8);5&p (22b) 55937 (8) ; 5936 (21); 5941 (11); 5946 (22b); 6368 (8). Peters, J. E. 29 (7). Phelps, O. P. 647, 648 (7). Pierre, L. 895 (la); 2794 (lc). Poilane, — . 287 1 4 (10). Poole, S. F. 102 (7). Pradham, — . 34 (la). Dr. Prain's collector 70, 78 (la); 289 (12); 4jp (22a); 795, W02 (8). Pratt, — . (55 (7). Prazer, J. C. 54, 167 (la) ; 2?<5 (lc) ; 273 (la). Pretz, H. W. 77403 (7). Pringle, C. G. 6842, 6998 (28); £733 (26). Purdom, W. 944, 945 (17); 946, 946* (6). Purpus, C. A. 7094 (26); 7370 (25); 8080, 8926 (26). Ram, B. 309 (la). Ramaswami, M. S. 1 490 (la). Ramos, M. & Deroy, D. 22511 (lb). Ramos, M. & G. Edano 38961 (9). Ramos, R. 9, 346, 1813, 42711 (lb). Randolph, L. F. & F. R. Randolph 35 (7). Rao, M. R. 75 20596, 20825, 22955, 23533, 24382, 24489, 24563, 24679, 24874, 25065, 26348 (7); 57015 (26). Stocks, J. E. 231 (la). Stone, W. 3821, 3921, 10200, 10682, 10847, I 3319, 13375 (7). Strachey, R. & J. E. Winterbottom 2 (la). Strathdickie, L. 1 124 (3). Stresemann, E. 90 (9). Thwaites, C. P. 1232 (la). To, K. P. 1868 (10). To, K. P. et al. Wong 2784 (20). 131,442 631 (23); 889 (10). W 41 88 1 (7). Sulit, M. D. 9917, 10086 Sun, C. L. 781, 1247 (16). Sun, S. C. 1189 (22b); /o* (17); 1189 (11). Suzuki, S. 1 1 162 (10); 4440.20, 455009, Tsang, W. T. Toppin, S. M. 4292, 6154 (10). Toroes, R. S. 4009 (9). Tso, C. L. 21 1 52 (6). Travis, M. T. 2871 (7). Trelease, W. 62, 63 (7). True, R. H. 92, 11 48 (7). True, R. H. & J. K. Edwards 5396 (7). Tsai, H. T. 51340, 51343 (18); 51546, 5 I 591, 51730 (8); 51945 (21); 52375 (10); 52477 (8); 52545 (10); 52552 (8); 5290/ (18); 54026 (11); 54565, 545/tf (22b); 54799 (21); 54949 (22b); 55076, 55^39 ( 8 ) ! 5542/ (21); 55670 (8); 5dopd (18); 56542 (10); 569/7 ( 8 ) ; 57255, 5727/, 57545, 575' o, 57769 (18); 57*99 (11); 57974 (18); 5*555 (11); 595^5 (16); 59548 (18); 59709 (16); 59824 (11); 60/52 (8); 6o2<5/ (21); 60420 (22b); 60459 (8); 60499 (21); 607c?/ (18); 609/9 (22b); 60933 (18); d/004 (11); 67/56, 61629 (22b) 67(5*0 (8); 61753 (18); 62295 (8) d259<5 (21); 62841 (22b); 65059 (18) 631 16 (16). 425 47oo/?> 4^024, 493027 (17). Suzuki, T. 6/95 (20). Swartley, J. C. Taam, Y. W. 578 (10); 720 (8); 724 (23); 75P (10); /04/ (23); 1477 (10); '792 (8); 1798, 1951 (10); 20*7 (8); 2147 (10). Tai, F. C. & C. M. Feng 525* (22b). 77, 16566 (10); 20092 (23); 20209, 20604 (1°); 20<575 (6); 20*55 (16); 2// 2300 (10); 466/ (16); 4945, 4.970 (6); 50/5, 5360 (16) ; 5543, 5601, 5614(1%); 6450 (16) ; 6832 (22b); 7499(18); ^J (6) ; 9354, 95/3 (18); /OO47 (16); /0460 (23). C. L. 20788, 20796, 21026 (20); (10). Ward, F. K. 9362 (21); 9455 (12). Waterfall, U. T. 1715, 2830 (7). Watkins, I. 5 J (7). Watson, J. 176 (la). White, C. T. 1943, /2J/7 ( 3 )- Whitford, H. N. 81, 414 (lb). Whitney, E. G. 2/132 (6); 21640 (23). Wiegand, K. M. Tso, C. L. & Y. Tsiang 2040 (10). Tsoong, — . 82941 (22b). Tsui, T. M. /3#(23); 34? (8) ; 39/ (10) ; 5^2 (18); 763 (16). 231 1 (7). 2765. <5794> 10248 ( 7 ) . Williams, L. O. & A. Molina R. Wilson, E. H. Turner, 431 (7). Uno, K. 21823 ( 24 ) . Valeton, T. 102 (10). Versteegh, — . /<5^3 /O (29). /£ (11); 753, 753A (6); P'9 (16); 952, 952 A, 972 (11); /063 (15); I 106, II 48 1176 (14); U84 (22b); (ii); "75 72o > 9 (18); Vicary (?), Visher, S. S. 175 (9). (la) 2110 (7). Voogd, C. N. H. de 682, 2338 (lb) Wadmond, S. C. 425 (7). Wagner,—. 68 1 6 (7). Wahl, H. A. 4959 (7). Walker, E. H. 57 1 7 (17). Wallich, N. 431 1 (8). Walter, N. & C. M. Bangham Wang, C. 32148 ( 1 8 ) ; 33029 ( 1 c ) ; 35222, 35893, 36064, 36607, 40415 (8). Wang, C. W. 62839 (11); 62990, 63709 (16); 64147, 68293 (18); 68718 (la); 69361, 71910 (18); 72857 (8); 72996 ( la) ; 734/3, 73379, 73390, 73825, 73839, 870 (10). (15); 1302, I 519, 2305, 2 305 A (16); 2306 (15); 2307(10); 230o'(17); 2309 (18); 23/O (11); 231 1 (18); 23/2 ( 14) ; 2215a (16); 3J24 (10); 3325, 41 i 7, 4122, 4/37(H); 4187, 4195(H); 4317(H); 478l (18); #£2(11); 6121,6304(19); 6627, 6907, 7412(17); 8300(10); 8450 (24); 86ll, 8801, 9203 (17); 9981 (20); 10425 (24); 10461 (17); //09<5 (19); II214 (20). Winckel, W. F. (10). 92P, 98P, 1633P, 1682? Wood, C. E., Jr. 2710, 3383, 3784 (7). Wray, L., Jr. Wright, C. 175, 103 1 (9). , 9 (8); 54 (19); 91 (10); 92 (8); 6/7 (10). 74430 (8); 74570 (8); 74966 (la); 74630 (la); Yamada, K. (la); 74907 (8); 749OO (lc); 75024 75183 (8); 7555/ (la); 75846 (10); 73920 (8); 76355, 76644 (la); 77202* (8); 77541, 77833 (lc); 7 (10); 79560 (lc); 7977^ (8); 0*0/36, 8018 1 80194 (lc);