• CO C > a > •' ft H & I ■» ■ * * * I N W 1 ' ^ -- * * i i ■ - r H d I - % # * J '* v »# W 1 A MONOGRAPH OF THE AMERICAN SPECIES OF THE GENUS DYSCHORISTE 1 CLARENCE EMMEREN KOBUSKI Assistant, Arnold Arboretum of Harvard University Formerly Rufm J. Lackland Research Fellow in the Henry Shaw School of Botany of Washington University Introduction Any attempt to determine specifically herbarium specimens of the genus under present consideration formerly proved very unsatisfactory because of the inadequacy of many of the original descriptions and also because of the small representation of type or authentic material in American herbaria. These facts, along with an especial interest in the genus and its allies, led to the present study. At first it was hoped that a monographic treat- ment of the whole genus might be made. A general survey of the material deposited in American herbaria, however, showed the Old World species to be so poorly represented that it was deemed advisable to exclude them from the present discussion and to include only the American species. Later the writer plans to visit some of the larger European herbaria and to supplement this monograph by a critical study of the far-eastern species. This investigation was made possible only through the cooper- ation of the botanists connected with the various herbaria from which material was borrowed. Sincere appreciation is due Dr. B. L. Robinson of the Gray Herbarium, W. R. Maxon of the United States National Herbarium, and D. C. Davies, Director of the Field Museum, who so willingly loaned their entire col- lections of Dyschoriste for this study. It was found necessary also to borrow types, and to obtain fragments and photographs of type collections from several European herbaria. Dr. Santiago Ramon y Cajal, Instituto Cajal, Madrid, and Professor Eduardo Balguerias y Quesada, Jardin Botanico, Universidad de Madrid, Missouri Washi submitted Washi Issued April 30, 1928. Ann. Mo. Bot. Gard., Vol. 15, 1928 (9) [Vol. 15 10 ANNALS OF THE MISSOURI BOTANICAL GARDEN very kindly furnished an excellent photograph of a little-known species, the type of which is preserved in the Madrid Herbarium. Professor Boris Fedtschenko, Jardin Botanique Principal, Lenin- grad, U.S.S.R., obligingly supplied two types essential for the completion of this monograph. Dr. A. W. Hill and T. A. Sprague, Royal Botanic Gardens, Kew, Dr. L. Diels, Botanischer Garten und Museum zu Berlin-Dahlem, Dr. C. H. Ostenfeld and Dr. Carl Christensen, Botanisk Garten, K0benhavns Universitet, as well as Dr. Adele Lewis Grant, Huguenot College, South Africa, who so willingly made critical comparisons with types at the Kew Her- barium on her journey to Africa, have all contributed either directly or indirectly, in material loaned or in verification of specimens submitted for comparison. The writer takes this op- portunity to express his gratitude for their generosity and kindly assistance. This study was made at the Missouri Botanical Garden, and thanks are due to the Director, Dr. George T. Moore, for the use of the excellent library and herbarium facilities which this in- stitution affords. Also, especial thanks are extended to the Curator of the Herbarium, Dr. Jesse M. Greenman, under whose constant guidance a,nd supervision this work has been made possible. History The genus Dyschoriste was proposed by Nees in the third volume of Wallich's 'Plantae Asiaticae Rariores' l published in 1832. The genus was segregated from Ruellia on account of stamen, corolla, and fruit characters, and was based on Ruellia depressa Wallich, namely, Wallich's No. 2379 from East India, which, however, is not conspecific with Ruellia depressa L. Two other East Indian species, D. cernua and D. litoralis, were referred by Nees in the same work to his new genus. In 1833, only a year later, David Don in Sweet's 'British Flower Garden' 2 described the genus Calophanes. Don's new genus was also a segregate from Ruellia, and was founded on Ruellia oblongifolia Michaux, which in turn was based on speci- mens collected in the state of Georgia. 1 Wallich, N. PL As. Rar. 3: 81. 1832. 'Sweet, R. Brit. Fl. Gard. II, 2: 181, pi. 181. 1833. 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 11 These two generic names were current in botanical literature for many years, as representing two supposedly distinct genera indigenous to remote regions — Dyschoriste of the eastern hemi- sphere and Calophanes of the western hemisphere. Nees, the foremost student in his time of the Acanthaceae, in his treatment of this family for Martius' 'Flora Brasiliensis/ * in 1847, accepted Calophanes and described seven Brazilian species of this genus. The same author and in the same year elaborated the Acanthaceae for DeCandolle's ' Prodromus ' 2 and maintained both names as representing separate and distinct genera. In this work, which was the first to present a comprehensive treatment of the group, five species of Dyschoriste and twenty-seven species of Calo- phanes, as well as several varieties, were recognized. Bentham and Hooker in the ' Genera Plantarum,' 3 1876, treated these two previously supposed distinct generic elements as congeneric, but unfortunately they took up the later name Calophanes and relegated Dyschoriste to synonymy. Mr. C. B. Clarke, who contributed the treatment of the Acanthaceae for Hooker's 'Flora of British India/ 4 1885, fol- lowed Bentham and Hooker's generic interpretation of the group and recognized four East Indian species of Calophanes, namely, C. Nagchana Nees, C. littoralis T. Anders., C. vagans Wight, and C. Dalzellii T. Anders. Under C. Nagchana Nees the following species are cited as synonyms: C. depressa T. Anders., Ruellia Nagchana Ham., R. erecta Burm., R. depressa and R. cernua Nees, Dipteracanthus Nagchana Nees, Dyschoriste depressa, and D. cernua Nees. In 1891, Dr. O. Kuntze 6 revived the name Dyschoriste Nees and transferred thereto several species, including Ruellia erecta Burm. which was described and illustrated in 1768, being the oldest Lindau. in 1895 the Acanthaceae for Engler and Prantl's 'Die Natiirlichen Pflanzenfamilien,' 6 followed Kuntze in recognizing the genus Dyschoriste. 1 Martius, C. F. P. de. Fl. Bras. 9: 26. 1847. J DeCandolle, A. P. Prodr. 11: 106, 107. 1847. » Bentham, G. & Hooker, J. D. Gen. PI. 2: 1077. 1876. • Hooker, J. D. Fl. Brit. India 4: 410. 1885. 6 Kuntze, O. Rev. Gen. PI. 2: 486. 1891. 6 Engler, A. & Prantl, K. Nat. Pflanzenfam. 4 3b : 302. 1895. 12 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 Clarke, in working up the Acanthaceae for the 'Flora Capen- sis/ ! 1901, took up the name Dyschoriste, thus reversing the position taken by him in Hooker's 'Flora of British India/ mentioned above. He recognized D. depressa Nees as a valid species along with four other African species, one of which is D. erecta Clarke, thus apparently disregarding the D. erecta (Burm.) 0. Ktze. Several new species have been described from time to time and referred to either Dyschoriste or Calophanes, but no compre- hensive treatment of the group as a whole has been published since that of Nees in DeCandolle's 'Prodromus.' Assuming that there is absolute identity and thus complete synonymy of the several elements which Clarke referred to Calophanes Nagchana in the 'Flora of British India/ then, as pointed out by Dr. Kuntze, the name erecta, as the oldest specific name involved, must be retained and the binomial Dyschoriste erecta (Burm.) 0. Ktze. becomes the valid combination for the plant concerned and D. depressa (Wall.) Nees must be regarded as a synonym of it. Since this study has been confined to the American species we must admit the observations of Clarke and Kuntze and accept Dyschoriste erecta (Burm.) 0. Ktze. as the type species of the genus until the eastern snecies in Question be examined General Morphology Roots. — The root system in the genus Dyschoriste is not very extensive. All the species are perennial and the roots, in turn, are of the simple fibrous form. By the casual observer, however, some of the slender underground stems of the previous year's growth are sometimes mistaken for roots. Stems. — There is considerable variation in the stem and its habit of growth. In all cases, the stem or stems arise from a ligneous perennial base. However, the mode of ascent varies. Many species are prostrate and the stems spread over the ground in several directions. In these cases the leaves assume a secund position. Often when the stems are short a rosette appearance is attained for the whole plant. The species D. oaxacensis illus- 1 Flora Capensis 5 1 : 15. 1901. 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 13 trates this character. However, it is not a stable specific char- acteristic. A second habit of growth is the ascending type. It is in this category that the majority of species is placed. The stem not infrequently becomes more or less geniculate ; this mode of growth is very characteristic of D. Pringlei. The third habit of growth is the erect type. It is to this type that the sturdy D. hirsutissima, D. oblongifolia, D. ovata, and D. trichanthera belong. Some species may follow consistently a distinct habit of growth, while others may have the stems ascending or erect even on the same plant. One can usually associate stem growth exceeding a length of 4-5.5 dm. with the erect habit, and shorter stems with the ascending or prostrate habit. Along with this, the prostrate type will possess small leaves and the erect type will have leaves with a more extensive surface. Several species, among them D. oblongifolia, possess slender underground stems which are common in perennials. These stems have the appearance of roots but on close observation buds and modified leaves can be seen. After passing under the ground for some distance they come to the surface and then rise erect. The stem may be terete or quadrangular. The latter is the more common type in the genus. In the species D. quadrangularis the stem is not only angular but the angles are winged. This condition is probably brought about by the decurrent petiole of the leaf extending down the stem. Leaves.— The leaves of the species in the genus Dyschoriste present a great variety of differences. All species have leaves with entire margins, except D. bilabiata which is not distinctly dentate but has a decided tendency in that direction. Several species, such as D. crenulata and D. hirsutissima, show a tendency toward crenulate margins. Others combine the crenulate with the repand margin. Along with these characteristics the margin is usually ciliate. In shape the leaf varies from that of the narrow, linear D. angusta, D. Greenmanii, and D. Purpusii to that of the oblong-ovate D. quadrangularis which grows to a length of 10 centimeters. Some species have two types of leaves, the lower or cauline leaves often being larger and different in shape from the uDoer leaves in whose axils branches and flowers are crowded. [Vol. 15 14 ANNALS OF THE MISSOURI BOTANICAL GARDEN The surface of the leaf itself is usually pubescent. When the pubescence is sparse or absent, an abundance of cystoliths is usually seen. Often the cystoliths, since they lack an orderly arrangement, are mistaken for appressed hairs. This cystolithic character is not sufficiently definite for specific delimitation. Some plants have a more copious formation of cystoliths than others. This character exists on the stem and calyx as well as the leaves. The venation which is very pronounced on the lower surface varies very little within the genus. The usual form is the feather- veined type. Inflorescence. —In all cases the flowers are axillary and sub- tended by bracts and sometimes bracteoles. Only occasionally are the flowers solitary in the axils. There are usually several to many flowers at the node, giving the appearance of a cymose cluster as in D. quadrangularis or a capitulum as in D. capitata and D. pinetorum. D. Greenmanii is an excellent example of a species with a solitary flower at the node. The majority of species have flowers which an; pedicellate, often so short, however, that a subsessile effect is presented. Calyx. — In the calyx of Dyschoriste is found one of the most constant characters of the genus. It is usually five-parted and always persistent. The only deviation from the five condition is found in D. maranhonis where the calyx-lobes are occasionally only four. In all cases the lobes are subulate- setaceous and usually ciliate. The ciliation may vary from a long whitish, flaccid pubescence to a very short hirsuteness. When the calyx proper is pubescent, the pubescence is usually confined to the nerves. The tissue connecting the lobes of the calyx is usually very membranaceous and tears apart very easily, iking it difficult and quite unsatisfactory to use the ratio between tube-length and lobe-length as a character for specific differentiation. The lobes are usually quite equal in length. However, here again variation is found. Cystoliths, as in the leaves, are very abundant; but in the calyx they are frequently disposed in a more or less regular arrangement. Corolla.— There is very little differentiation to be found in inerous m the corolla of the bilabiate however, is not as distinct as in some other eenera of the 1928) KOBUSKI — MONOGRAPH OF DYSCHORISTE 15 Acanthaceae. The corolla is five-lobed, the two posterior lobes being coalescent to a greater extent than the three anterior lobes. The length of the corolla varies from 10 to 17 mm., as found in D. decumbens, D. hygrophiloides, D. saltuensis, D. quadrangularis, and D. angustata, to 25-28 mm., exemplified in D. xylopoda, D. humilis, and D. ovata. One species, D. Pringlei, has a corolla measuring 35-38 millimeters long. None, however, reach the length of 70-80 millimeters as found in some of the Ruellias. The proportion between tube and throat is variable in the genus. The ventricose throat is found quite often. The condition should exist in all species because of the contiguity of the adnate fila- ments in the posterior portion of the throat and tube. This ventricosity, hence, is more pronounced in the larger-flowered species. The narrow tube of the corolla is usually slightly flared at the base to make room for the disc and ovary. The ampliation from the tube to the throat is very variable and may be abrupt or gradual according to the species. In all cases, the external surface is quite pubescent. In the species D. trichanthera the pubescence is found on the interior as well as the exterior surface of the throat. Stamens. —The stamens are didynamous. The long and short filaments on each side are contiguous or united at the base by a membrane which extends from the point of adnation to the base of the corolla tube. A very distinctive feature of the anthers is the mucronate appendages at the base of the anther cells. These mucronate appendages are characteristic of the genus and are very easily seen with the hand-lens. Under the low power of the compound microscope they are found to be composed of several multicellular strands of cells closely compacted together. strands of and a dentate ragged appearance is given to the whole appendage. This is doubtless what Nees saw when he described the appendages of D. quitensis as "2-3-toothed." A similar example was found in D. Schiedeana. On microscopic study, however, the so-called dentations were found to be nothing more than shreds of tissue torn away slightly from the compacted mass. The anther cells are usually parallel and oblong in shape. In the case of D. sagit- tata and Z). maranhonis the cells are so disposed as to have a , a [Vol. 15 10 ANNALS OF THE MISSOURI BOTANICAL GARDEN sagittate appearance. In both the species mentioned the apex as well as the base is appendaged. As a rule the anther cells are glabrous but in the species D. trichanthera the anther cells are very pubescent. The mode of dehiscence is by a longitudinal slit on the side of the anther cell. The filaments are commonly pubescent. ^ Pistil— There is little variation in the parts of the pistil. A disc is present beneath the ovary in all species. The ovary itself is two-celled, glabrous, and oblong. Little or no variation is found in the filiform pubescent style. Only the anterior lobe of the stigma is developed, and this lobe is usually linear and oblique with a flattened stigmatic surface. However, in D. hygrophiloides the stigma is curved, while in D. sagittata it is basally lobed. In D. maranhonis the stigma is reflexed. Capsule.— The capsule of the genus is quite uniform. The constant linear, glabrous and four-seeded characters, combined with other diagnostic characters, help considerably in generic determination. Retinacula or hooked appendages on the median ridge of the valves hold the flat suborbicular seeds in place. When dry the seeds appear to have many soft, appressed hairs. These same hairs when wetted diverge, elongate, and become m Geographical Distribution The geographical distribution of the American species of the genus Dyschoriste offers very interesting problems. The accom- panying maps demonstrate very clearly that there are three distinct areas of distribution: (1) southeastern United States; (2) southwestern United States and Mexico; and (3) South America. Two species, namely, Dyschoriste oblongifolia and D. angusta, occur in the southeastern United States area. This area extends the width of the coastal plain from southern Virginia to southern Florida. The regions of distribution of the two species do not overlap. D. angusta is confined to the wet region of Dade and Palm Beach Counties in southern Florida, while D. oblongifolia extends northward through the remainder of the area, seeking the dry, sandy pine woods. 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 17 *> o a Q CO a t o o o •43 o 03 *-. o U) 60 a O •s 03 fcfi The second area extends from Texas and southern Arizona southward to the Isthmus of Tehuantepec and contains the greater number of species. Of the forty recognized species in the genus Dyschoriste, twenty-one, or more than half, are confined 18 ANNALS OF THE MISSOURI BOTANICAL GARDEN Vol. 15 Fig. 2. Map showing the geographical distribution of the South American species of Dyschoriste. to this area. It is an interesting fact that the distribution of some species is almost coincident with the geological formation of the country. D. decumbens, which occurs on the plateau 1928| KOBUSKI — MONOGRAPH OF DYSCHOBISTE 19 region between the Sierra Madre ranges, is an excellent example of this fact. D. hirsutissima extends from Sonora southward along the western slope of the Sierra Madre range to Oaxaca. Many species have a localized distribution only, D. Greenmanii, D. crenulata, D. saltuensis, and D. angustifolia being examples. Some of these localized areas are characterized by three or more species. An instance of this is a small region around Guadalajara in the state of Jalisco where four species are represented. An- other illustration of limited areal distribution occurs in the northern part of Oaxaca which harbors D. oaxacensis, D. angusti- folia, D. capitata, and D. hirsutissima. Many species, especially the localized ones, appear to be extremely edaphic since they inhabit only regions near volcanoes. The center of distribution in this second area falls within the region represented by the states of Puebla, Michoacan, and Mexico. The third and last area, namely that of South America, com- prises more territory than either of the areas indicated above and includes the seventeen remaining species of the genus. The material examined in all cases was not very copious, hence an accurate range of geographical distribution of these species could not definitely be ascertained. Nearly all species appear to occur in isolated and limited areas, but the relationship between some of these areas indicates that a greater overlapping of areas would occur were it not for the paucity of herbarium material. Three species, namely, D. quitensis, D. ciliata, and D. repens, are found in Peru and the Andes of Ecuador. D. Niederleinii and D. humilis are found in Argentina. The other thirteen species inhabit northern Paraguay and southern Brazil, and it is here that the South American center of distribution occurs. An unusual feature of the geographical distribution is the isolation of the areas defined. At present, there is no one species which connects up any two areas. There seems to be no satis- factory explanation to account for the absence of the genus between the Andes of Ecuador and the Isthmus of Tehuantepec in Mexico. Members of the genus may be found between these two remote regions, but until the entire Andean range has been explored more thoroughly from a botanical standpoint one would hardly venture an explanation of the marked discontinuous dis- tribution which the genus now presents. (Vol. 15 20 ANNALS OF THE MISSOURI BOTANICAL GARDEN The non-occurrence of the genus in the Mississippi Valley is equally surprising. Since the flora of this entire area is com- paratively well known, it is hardly possible that the members of the genus would be overlooked if they there existed. The only solution seems to be the possible age of the genus. Dyschoriste is probably a pre-glacial genus which, prior to the Oligocene period of the Cenozoic era, extended continuously across the southern United States. However, the Eocene and Oligocene seas encroached upon the United States in the present Mississippi Valley, thereby splitting the distribution areas of the genus into two parts. Phylogeny Because of the large number of closely allied species in the genus Dyschoriste it is quite necessary that the phylogenetic discussion of the group be made from a purely hypothetical standpoint. The fact that the discussion is confined to the American species alone seconds this consideration, since the eastern species of the genus exceed the American species in number. On account of the three distinct geographical distribution areas, which have been discussed before, a tree method of illus- trating probable phylogenetic sequence proved unsatisfactory; hence the method used in the accompanying chart was devised finally to illustrate the apparent relationship of the species of the western hemisphere. This chart if superimposed on a map of the regions inhabited by the genus would coincide with the specific; regional distribution. It was felt reasonably certain that all species of the genus have evolved from a common ancestor designated in the chart as x. From this ancestor, species and groups of species have evolved. One might ask why, since the species seem to be placed in definite groups, subgenera or sections have not been designated. This question was given much thought and consideration; it was felt finally, however, that on account of the relative uniformity of the essential morphological characters within the genus, except in the case of group II, no adequate basis exists for the designation of subgenera or sections. It may be observed that all the designated groups with the 1928] KOBUSKI — MONOGRAPH OP DYSCHORISTE 21 exception of group II originated from the common ancestor at approximately the same time. Group II, on account of the muticous character of its anther appendages, the extremely- small leaves, the very small flowers, and the fruit characters, has been >parated from the Dyschoriste and raised to the rank of a new genus which is considered as intermediate between the hypothetical type and Dyschoriste proper. In this genus, Apassalus, 1 containing three species, the species Apassalus diffusus Fig. 3. Phylogenetic chart of the species of Dyschoriste. has reached hi development the reduction of the number of ovules to two, one being borne in each valve of the capsule. The other two species contain the four seeds which of the majority of the genus Dys choriste. Apassalus is confined to the islands of Cuba and Haiti and to the southeastern United States. 1 See Kobuski, C. E. A new genus of the Acanthaceae. Ann. Mo. Bot. Gard. 15: 1-8, pi. 1-2. 1928. [Vol. 16 22 ANNALS OF THE MISSOURI BOTANICAL GARDEN Group I, involving ten species, is confined to the plateau regions of Mexico. In this instance D. decumbens, on account of its extended range and characteristic relation to all species con- cerned, is considered the base species. D. Lloydii and D. crenu- lata are species having vital characteristics similar to D. decumbens but differing sufficiently in minor characters to be considered direct descendants from the base species. A small group con- taining three species, D. Purpusii, D. Greenmanii, and D. Rosei, stands by itself. The highest development is reached in D. Greenmanii and D. Rosei in which cases the inflorescence has been reduced to a solitary flower at each node. All species of the last group have very slender linear leaves. Another branch from D. decumbens, as the chart illustrates, is the linearis-jaliscensis branch. Although separated somewhat in regional distribution these two species are closely allied through their flower and foliage structures and are undoubtedly derived from D. decumbens. The species D. oblongifolia and D. angusta seem to be closely related to D. linearis; in fact, D. linearis was once considered a variety of D. oblongifolia. However, the two species under dis- cussion are confined to the southeastern United States in their distribution and, as the distribution map of North America shows, are not connected definitely with the Mexican-southwestern United States species. This suggests the probability that there may be a relationship between the two species in Florida and the genus Apassalus also found there and in the West Indies. The migration may have been northward through the Antilles, and a connecting link between the southeastern United States species and the southwestern United States-Mexico group may never have existed. In group III, one finds an entirely different situation. Here we have four large, erect species, each showing a definite charac- teristic development toward advancement. Perhaps the highest development is found in D. hirsutissima which extends the length of the western Sierra Madre range and possesses a well-developed glandular pubescence. This is the only instance of this character in the whole genus. D. bilabiata and D. quadrangularis are close relatives but do not seem to have been derived from the D. hir- 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 23 sutissima line. Instead, they undoubtedly arose along parallel lines of development. In D. bilabiata we find a distinct dentation of the leaf. This group is also characterized by large, petiolate leaves, in some cases as long as ten centimeters, an unusual feature in the genus. Group IV for the most part occurs in southwest Mexico, that is, the states of Oaxaca, Jalisco, and Michoacan. These species are the ascending foliose type with rather small, ovate leaves. The inflorescence is usually subcapitate, and it is sometimes dif- ficult to distinguish off-hand the species here included. They all, with the exception of D. pinetorum, seem to have evolved along a parallel line of development. D. pinetorum, on account of its resemblance to D. Pringlei, undoubtedly evolved from it. It is in D. Pringlei that we find the largest flower of the genus Dyschoriste, and in this species a close resemblance is shown to the genus Ruellia. The highest point of development in the North American species just discussed is found, according to my opinion, in D. hir sutissima , D. Greenmanii, D. Rosei, and D. linearis. In the South American species of the genus, a similar situation is found. Here the species can be placed in five groups. Group V contains the three species on the western coast inhabiting Peru and Ecuador. Specialization in them is not particularly noticeable. The species D. ciliata possesses rather muticous anther appendages. This would ally the species to the genus Apassalus. It was necessary to accept the word of Nees in this instance, however, as only a photograph of the type of the species could be obtained. The species of Group VI are found in Argentina. Here D. humilis reaches the highest point of development in the reduced number of seeds. As in Apassalus diffusus, only two seeds are produced, a single seed occurring in each valve of the capsule. Group VII is not unusual in its development. In this group are found six closely related species — perhaps interrelated — but showing no special development. Here again the herbarium material at hand is very sparse, the study in a few instances being confined to photographs. In Group VIII are four species, two of which are described for [Vol. 15 24 ANNALS OF THE MISSOURI BOTANICAL GARDEN the first time. In the species D. trichanthera is found the spicate inflorescence along with pubescent anthers. The former char- acter links it up with D. Schottiana, while the latter character shows the relationship which exists between D. trichanthera and D. lavandulacea . D. paraguariensis is undoubtedly a branch from D. lavandulacea but possesses more highly developed floral characters. The last group, namely, group IX, includes only two species. It is here that the sagittate or divergent anther cells are found. The highest development of the group and probably the highest development in the genus is shown by D. maranhonis, in which both incomplete didynamy and reduction in corolla and calyx- lobes are found. Nees describes D. maranhonis as being glandu- lar-pubescent. A fragment of the type specimen was obtained and failed to demonstrate this character. Summary The conclusions drawn as to the probable phylogeny of the group under consideration are reached after a comparative study of the outstanding morphological characters which may be sum- marized as follows : (1) Muticous appendaged anthers are more primitive than those with apiculate appendages. (2) Divergent anther cells are more advanced than parallel anther cells. (3) Glabrous anthers are more primitive than pubescent anthers. (4) Four-seeded capsules are more primitive than two-seeded capsules. (5) Numerous flowers in an axil is a more primitive condition than the solitary-flowered axil because the presence of bracts in the solitary-flowered species shows reduction to have taken place in the telescoping of the inflorescence. (6) An unlobed stigma is more advanced than the lobed stigma. (7) Entire-margined leaves are primitive. The dentate margin is an advancement, and the crenulate margin type is intermediate. (8) Glandulosity is more advanced than pilosity. 1928] KOBTJSKI — MONOGRAPH OF DYSCHORISTE 25 (9) Procumbent plants are more primitive than erect plants which have evolved through ascending plants. (10) Complete didynamy is more primitive than incomplete didynamy. (11) Winged stems are more advanced than the un winged quadrangular stems. (12) Reduction of corolla- and calyx-lobes from five to four is a criterion of specialization and advancement. Abbreviations The abbreviations used to indicate the herbaria in which the specimens cited in the present paper occur are as follows: B = Botanischer Garten und Botanisches Museum, Berlin, Germany. C = Botanisk Garten, K^benhavns Universitet, Copen- hagen, Denmark. Ch = University of Chicago (deposited in the Field Museum) . FM = Field Museum of Natural History. G = Gray Herbarium of Harvard University. K = Royal Botanic Gardens, Kew, England. L = Jardin Botanique Principal, Leningrad, U.S.S.R. M = Missouri Botanical Garden. Ma = Jardin Botanico, Universidad de Madrid, Madrid, Spain. US = United States National Herbarium. Taxonomy Dyschoriste Nees in Wallich, PL As. Rar. 3: 75, 78. 1832; Nees in DeCandolle, Prodr. 11: 106. 1847; O. Kuntze, Rev. Gen. PL 2: 485. 1891 ; Lindau in Engl. & Prantl, Nat. Pflanzen- fam. 4 3b : 302. 1895; Gray, Manual, ed. 7, 743. 1908. Calophanes D. Don in Sweet, Brit. Fl. Gard. 2: pi. 181. 1833; Nees in Mart. Fl. Bras. 9: 25. 1847; Nees in DC. Prodr. 11: 107. 1847; Benth. & Hook. Gen. PL 2: 1077. 1873-76; C. B. Clarke in Hooker, Fl. Brit. India 4: 410. 1885; Gray, Syn. FL N. Am. 2 1 : 324. 1878, and ed. 2, 1886; Chapman, Fl. Southeastern U.S., ed. 3, 365. 1897; Small, Fl. Southeastern U.S. ed. 1, 1082. 1903, and ed. 2, 1913. Vol. 15 26 ANNALS OF THE MISSOURI BOTANICAL GARDEN Linostylis Fenzl. in Linnaea 23: 94. 1850. Herbaceous, caulescent perennials, prostrate, ascending or erect, glabrous or pubescent. Leaves opposite, sessile or petioled, usually entire. Inflorescence; cymose, capitate or spicate, terminal or axillary. Flowers subtended by foliaceous bracts and bracte- oles. Calyx deeply 5-lobecl, lobes usually subulate-setaceous, ciliate, lineolate. Corolla-tube usually erect, occasionally slightly ampliated at the base; limb spreading, oblique, obscurely or distinctly bilabiate, 5-lobed. Stamens 4, didynamous; filaments of a long and short stem united at the base and adnate to the base of corolla-tube, pubescent; anther 2-celled, cells oblong, sharply mucronate at the base, parallel or slightly divergent, glabrous or occasionally pubescent. Ovary 2-celled, glabrous, ovules 2 or occasionally 1 in each cell; style filiform, pubescent; posterior lobe of stigma rudimentary, anterior lobe oblique, slightly flattened. Capsule; included in the persistent calyx, oblong-linear, glabrous, 2-4-seeded, separating with difficulty at maturity into 2 valves, 1-2 seeds to each valve held in position by retinacula. Seeds flattened, suborbicular, mucilag- inous when wetted. Type species: Dyschoriste erecta (Burm.) O. Ktze. in Rev. Gen. PI. 2:485. 1891. Key to Species 1 . Plants glandular-pubescent 1. D. hirsidissiina Plants glabrous or pubescent but not glandular 2 2. Inflorescence spicate 3 Inflorescence not spicate 4 3. Anther cells pubescent; internal surface of corolla-throat pubescent. 2. D. trichanthera Anther cells glabrous; internal surface of corolla-throat glabrous.S. D. Schottiana 4. Leaves mostly linear 5 Leaves ovate, other than linear 14 5. Corolla approximately 10 mm. long 1+. D. angusta Corolla 15 mm. or more long 6 6. Corolla 15-20 mm. long 7 Corolla 25-30 mm. long 11 7. Leaves 2 mm. or more wide 8 Leaves 1 mm. or less wide 5. D. Purpusii 8. Anther cells pubescent 6. D. lavaridulacea Anther cells glabrous 9 9. Stem glabrous, except for distinct pubescence at node; flowers solitary at node 7. D. Greenmanii Stem evenly pubescent; flowers usually two or more at node 10 1928) KOBUSKI — MONOGRAPH OF DYSCHORISTE 27 10. Plants low-growing, ascending, 1 dm. or less high; style approximately 6 mm. long; S. Amer. sp 8. D. Niederleinii Plants strict, 3-6 dm. high; style approximately 12 mm. long; Mex. sp. 9. D. Schiedeana 11. Flowers characteristically solitary at the node 10. D. Rosei Flowers not solitary at the node 12 12. Leaves mostly linear, 2 mm. or less wide 11. D. jali iscensis Leaves linear to linear-lanceolate, more than 3 mm. wide 13 13. Stem villous-hirsute 12. D. angustifolia Stem hirsute with rigid hairs 13. D. linearis 14. Leaves distinctly dentate 14. D. bilabiata Leaves not distinctly dentate 15 15. Cinereous-pubescent throughout 16 Not cinereous-pubescent 17 16. Leaf margins entire, not crenulate 15. D. decumbens Leaf margins distinctly crenulate 16. D. crenulata 17. Inflorescence capitate or subcapitate 18 Inflorescence other than capitate 21 18. Anthers emarginate, elongate at the apex 17. D. capitata Anthers without an elongation at the apex 19 19. Corolla 35-38 mm. long 18. D. Pringlei Corolla 20-21 mm. long 20 20. Leaves glabrous, oblanceolate 19. D. oaxacensis Leaves pubescent, obovate-elliptic 20. D. pinetorum 21. Corolla 15 mm. or less in length 22 Corolla 20 mm. or more in length 33 22. Leaves sessile 23 Leaves petiolate 25 23. Plants with converging anther cells giving a sagittate appearance; stigma lobed 21. D. sagittata Plants with parallel anther celht; stigma unlobed 24 24. Plants with glabrous leaves; S. Amer. sp 22. D. Serpyllum Plants with pubescent leaves; Mex. sp 23. D. Lloydii 25. Leaves glabrous 26 Leaves pubescent 28 26. Leaves 6-7 cm. long 27 Leaves 1.5-3 cm. long 24. D. microphylla 27. Anthers distinctly calcarate at base; leaves lanceolate; Mexican species. 25. D. saltuensis Anthers only slightly calcarate at base; leaves ovate-elliptic; S. Amer. species 26. D. ciliata 28. Stems winged; leaves approximately 10 cm. long 27. D. quadrangularis Stems not winged; leaves 2-4 cm. long 29 29. Calyx-tube glabrous, lobes cilial e with softly hirsute hairs; leaves lanceo- late 28. D. quitensis Calyx tube pubescent, lobes ciliate with stiflish hairs; leaves ovate to rotund 30 30. Anther cells diverging, giving a sagittate appearance; corolla and calyx lobes frequently reduced to 4; incomplete didynamy common. 29. D. maranhonis [Vol. 15 28 ANNALS OF THE MISSOURI BOTANICAL GARDEN Anther cells parallel; corolla and calyx characters constant; complete didynamy 31 31. Style approximately ft mm. long; 32 Style 10 mm. or more long SO. D. hirsuta 32. Lower leaves obovate to subrotund, emarginate at the apex 31 . D. hygrophiloides Lower leaves ovate, obtuse but not emarginate at the apex SB. D. repens 33. Calyx 9-10 mm. long S3. D. Pulegium Calyx 15 mm. or more long 34 34. Flowers crowded in glomerules 35 Flowers usually in pairs at the nodes 36 35. Stem erect; corolla barely exceeding or equalling the calyx in length; Mex. sp 84. D. ovata Stem geniculate-ascending; corolla 5 mm. or more longer than calyx; S. Amer. sp 35. D. amoena 36. Leaves pubescent 37 Leaves glabrous 39 37. Villous-pubescent throughout 86. D. xylopoda Plants not villous-pubescent 38 38. Capsule 4-seeded; N. Amer. sp 37. D. oblongifolia Capsule 2-seeded; S. Amer. sp 38. D. humilis 39. Leaves distinctly acuminate; the two bracts at each axil nearly equalling the leaf in all characters 89. D. paraguariensis Leaves obtuse at apex; bracts foliate but not equalling the leaf in size 40 40. Leaves sessile; corolla 25-28 mm. long; Am. bor. sp. . 37a. D. oblongifolia f . glabra Leaves petiolate; corolla 20 mm. long; S. Amer. sp 40. D. Tiveediana 1. Dyschoriste hirsutissima (Nees) O. Ktze. Rev. Gen. PL 2: 486. 1891. Calophanes hirsutissimus Nees in DC. Prodr. 11: 109. 1847; Hemsl. in Biol. Cent-Am. Bot. 2: 502. 1882. Calophanes bilobatus Rose in Contr. U.S. Nat. Herb. 1: 109. 1891. Stems branching, ascending from a stout perennial base to a height of 10-12 dm., somewhat quadrangular, more or less pubescent with the pubescence, in some cases, restricted to the edges, occasionally glandular at the apex; leaves petioled, ovate to oblong-ovate, 3-8 cm. long, 1.5-3 cm. wide, acute at both ends, margin usually entire or slightly crenulate, sometimes slightly denticulate, pubescent on both surfaces, younger leaves often densely so and glandular; inflorescence axillary, subtended by glandular-pubescent, subulate bracts ; calyx 5-lobed, subulate- setaceous, extremely glandular-pubescent, approximately 11 mm. 1928] KOBUSKI — MONOGRAPH OP' DYSCHORISTE 29 long, lobes more than on3-half the total length of the calyx; corolla subbilabiate, puberulent on the outer surface, occasionally glandular, averaging 14 mm. in length, tube about the same length as the abruptly ampliated throat; stamens adnate to a little below the middle of the corolla-throat; stigma oblique; capsule 4-seeded; seeds ob ique. Distribution: slopes of the Sierra Madre of western and southern Mexico. Specimens examined: Southwestern Chihuahua, Aug.-Nov. 1885, Palmer 285 (G, US) ; Alamos, 180 miles s.e>. from Guaymas, Sonora, alt. 418 m., 26 March-8 April, 1890, Palmer 402 (G, US); Sierra de los Alamos Mt., 6 miles due south of town of Alamos, Sonora, 14 March, 1910, Rose, Standky & Russell 12838 (US) ; dry hillside, Acaponeta, Tepic, 10 Apri ., 1910, Rose, Standley & Russell 14298 (US); dry rocky slopes ne^r Guadalajara, Jalisco, 11 Dec. 1888, Pringle 2154 (G) ; hills near Guadalajara, Jalisco, 15 Nov. 1889, Pringle 2939 (FM, G); Ouernavaca, Morelos, 15 Nov. 1865, Bourgeau 1262 (G); Cuicatlan, Oaxaca, alt. 1000 m., 9 Dec. 1895, Gonzalez 48 (G) ; Monte Alban, Oaxaca, 1933 m., 23 Nov. 1894, Pringle 6053 (G, M, US); Monte Alban, Oaxaca, alt. 2833 m., 24 Nov. 1894, L. C. Smith 323 (G) ; Monte Alban, near Oaxaca City, Oaxaca, alt. 1900-2000 m., 23 Nov. 1894, L. C. Smith 729 (M, US); Monte Alban, Oaxaca, 29 Dec. 1895, Seler 1733 (G); Valle de Oaxaca, Oaxaca, alt. 1650 m., 18 Nov. 1906, Conzaiti 1521 (FM); Tehuantepec, June, 1906, Gandoger (M 120892); Hacienda de Guadalupe, date lacking, Ehrenberg 1223 (B type, M fragment and photograph). 2. Dyschoriste trichanth* ra l n. sp. PI. 4. Dyschoriste maranhonis Lindau in Bull. Herb. Boiss. II. 3: 628. 1903, as to Fiebrig 4856, Hassler 5908, 7780, not O. Ktze. Stems stout, branched, erect, 5-6 dm. high, glabrate, pubescent near the apex, basal porion densely covered with cystoliths, swollen at the nodes; leaves oblong-ovate to ovate, younger 1 Dyschoriste trichanthera Kob., sp. nov., caulibus suffruticosis, erectis, 5-6 dm. altis, glabrescentibus, apice pube?centibus, inferiore cystolitherissimo, tumidis ad nodos; foliis oblongo-ovatis vel ovutis, 5-7 cm. longis, 2-3 cm. latis, integerrimis vel crenulatis, petiolis 10-12 mm. long s; floribus axillaribus, aliquot ad singularos nodos, fere prope apicem spicatis; bracteis parvis, foliace.s, cibatis, pubescentibus, bracteolis acuminatis. 4-7 mm. loneis: calvte 13-14 mm. loneo, lobis subulatis, setaceis, 8-9 30 ANNALS OF THE MISSOURI BOTANICAL GARDEN Vol. 15 pubescent, older leaves glabrate, 5-7 cm. long not including the petiole, 2-3 cm. wide, entire to crenulate, petiole 10—12 mm. long; flowers axillary, crowded at the nodes near the apex giving a spicate appearance; bracts small, foliaceous, ciliate and pubes- cent, bracteoles acuminate, 4-7 mm. long; calyx 13-14 mm. long, lobes subulate-setaceous, 8-9 mm. long, often recurved at the tip, ciliate, with 2 kinds of multicellular hairs, both flaccid and delicate; corolla distinctly bilabiate, 10-20 mm. long, rose or violet, lobes obtuse, emarginate, puberulent on external surface, distinctly pubescent on the internal surface; stamens barely included, adnate to about opposite the labiation of the corolla, anthers pubescent, style filiform, 10-20 mm. long, stigma oblique, linear; capsule not seen. Distribution : along rivers, northern Paraguay. Specimens examined : In the region of the river Capivary, Paraguay, date lacking, Hassler 5908 (G); between the rivers Apa and Aquidaban, Paraguay, Jan. 1908-9, Fiebrig 4856 (G); in region along the river Apa, Paraguay, Nov. 1901, Hassler 7780 (G type, M photograph and fragment). 3. Dyschoriste Schottiana (Nees) Kobuski, n. comb. Hygrophila Schottiana Nees in Mart. Fl. Bras. 9: 22. 1847; Nees in DC. Prodr. 11: 87. 1847. Dyschoriste crinita (Nees) O. Ktze. Rev. Gen. PI. 2: 485. 1891; Lindau in Bull. Herb. Boiss. 7: 575. 1899. Calophanes crinitus Nees in Mart. Fl. Bras. 9: 26. 1847; Nees in DC. Prodr. 11: 107. 1847. Herbaceous perennial; stem erect, 5-6 dm. high, profusely branched, hirsute; leaves oblong-lanceolate, 4-5 cm. long, 1-1.5 cm. broad, tapering below into a short petiole, entire, glabrous, margin and midrib of under surface scabrous; inflorescence axil- lary, cymose, many-flowered, subtended by bracts; calyx deeply mm. longis, apice saepe recurvatis, ciliatis cum duobus generibus multicellularum 10-20 vel 10-20 the river Apa, Paraguay, Nov. 1901, E. Hassler 7780 (G). pubescentibus; stylo Type collected along 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 31 5-parted, densely hirsute, about 10 mm. long, lobes subulate- setaceous; corolla more o:* less bilabiate, approximately 18-20 mm. long, pubescent on the external surface; capsule 8-9 mm. long, 4-seeded, glabrous; seeds flattened, suborbicular. Distribution: southeastern Brazil. Specimens examined: Prov. Goyaz, Brazil, Feb. 1841-42, Gardner 8951 (K type, M photograph) . 4. Dyschoriste angusta (Gray) Small, Fl. Miami, 168. 1913; Small, Fl. Florida Keys, 135. 1913. Calophanes angusta Graj , Syn. Fl. N. Am. ed. 2, 2 1 : 456. 1886; Small, Fl. Southeastern U.S. ed. 1, 1083. 1903, and ed. 2, 1913. Calophanes oblongifolia var. angusta Gray, Syn. Fl. N. Am. ed. 324. 1878, and ed A dm. high; stem erect or ascending from a creeping base, slightly puberulent, occasionally branching ; leaves many, subsessile, 'inear to linear-lanceolate, 1.5-2 cm. long, lineolate, entire; bracts foliaceous, about one-half as long as the leaves; flowers axillary; calyx 8-9 mm. long, lobes subulate- setaceous, ciliate, distinct to near the base, hardly surpassing the capsule at maturity; corolla blue;, purple, or rarely, white, slightly bilabiate, approximately 1 cm. long, tube slightly shorter than the limb and a little ampliated at the base; stamens adnate to the base of the limb o:' the corolla; anthers ovate, filaments widening at the base; capsule glabrous, linear, 4-seeded; seeds somewhat oblique. Distribution: southern Florida. Specimens examined: Palm Beach County: Pdm Beach, 26 Dec. 1895-11 Jan. 1896, Hitchcock 1455 (FM). Dade County : pine lands, Grossmanns, 24 May, 1904, Britton 155 (FM); Cocoanut Grove, 26 Dec. 1895, 11 Jan. 1896, Hitch- cock 1456 (FM); Miami, near river, 21 Nov. 1903, Eaton 385 (FM) ; in pine lands between Cocoanut Grove and Cutler, 13-23 Nov. 1903, Small & Carter 776 (FM) ; cabbage field in pine woods, Grossmanns, 25 Feb. 1905, Eaton 1247 (G); BiscayneBay, 1874, Palmer 847 (G, M, US); Miami, May, 1877, Garber (G, M 32 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 120923); Biscayne Bay, June, 1880, Curtiss 1938 (G); Lemon City, 3 March, 1892, Simpson 528 (G, US); rocky, calcareous land, Miami, 6 April, 1897, Curtiss 6858 (M, FM, G, US); Black Point Bridge near Cutler, 27 Feb. 1920, Young 301 (US). 5. Dyschoriste Purpusii 1 n. sp. PL 5. Perennial, erect or ascending from a suffruticose base, branching quite profusely, more or less pubescent, with short, stout, white hairs; leaves sessile, linear to linear-lanceolate, 18-20 mm. long, 1 mm. or less broad, entire, pubescent; flowers slightly pedicellate, subtended by two foliaceous bracts; calyx 5-parted, 9-11 mm. long, lobes unequal, slightly longer than the tube, pubescent, ciliate; corolla 15-17 mm. long, pubescent on the external sur- face, tube very narrow, not ampliated at the base, approximately 7 mm. long, limb 4 mm. long; anthers of the shorter pair of stamens occasionally smaller; style 11-12 mm. long, stigma ob- lique; capsule linear, glabrous, approximately 10 mm. long; seeds 4, oblique. Distribution: south Mexico. Specimens examined: Puebla: rocky hills, Tehuacan, June, 1905, Purpus 2362 (M type, US, G, FM) ; vicinity of San Luis Tultitlanapa, July, 1908, Purpus 3347 in part (M). 6. Dyschoriste lavandulacea (Nees) 0. Ktze. Rev. Gen. PI. 2: 486. 1891. Calophanes lavandulaceus Nees in Mart. Fl. Bras. 9: 27. 1847; Nees in DC. Prodr. 11: 112. 1847. Stems erect from a perennial base, 1.5-2 dm. high, sparingly pubescent, quite angular; leaves sessile, linear-lanceolate, 40-50 1 Dyschoriste Purpusii Kob., sp. nov., caulibus perennis, erectis vel ascendentibus a suffruticosa basi, ramis profusia, plus minusve pubescentibus cum brevibus albidis capillibus; foliia sessilibus, linearibus vel linearo-oblanceolatis, 18-20 mm. longis, 1 mm. minusve latis, integerrimis, pubescentibus; floribus parum pedicellatis, sub- tendentibus bracteis; calyce 5-diviao, 9-11 mm. longo, lobis inaequalibus, paulo tubo longioribus, pubescentibus, ciliatis; corolla extus puberula, 15-17 mm. longa, tubo angustissimo, non ampliato ad basem, plus 7 mm. longo, limbis 4 mm. longis; staminibus postero-lateralibus minoribus; capsula lineari, glabra, 10 mm. longa, 4-sperma. — Type collected on rocky hills, Tehuacan, June, 1905, C. A. Purpus £362 (M). 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 33 mm. long, 4-5 mm. wide, tapering to an acute apex, entire, glabrous; inflorescence somewhat glomerulate, several-flowered; bracts minute, 3-5 mm. long, resembling the calyx in texture; calyx 5-lobed, 13-14 mm. long, tube about 5 mm. long, glabrous and covered with cystoliths except for the ciliate, subulate- setaceous lobes; corolla 5-lcbed, slightly bilabiate, 20 mm. long, pubescent on the external surface, tube one-half the total length of the corolla, lobes quite truncate; anther cells slightly puberu- lent; mature capsule not seen. Distribution: south-centril Brazil. Specimens examined: In dry fields, near Rio Pardo, Brazil, Sept. 1826, Riedel 501 (L type, M photograph). 7. Dyschoriste Greenmanii 1 n. sp. PI. 6. Plants about 2 dm. high, ascending from a perennial base; stems slender, branched, pubescent at the nodes, otherwise quite glabrous; leaves sessile, linear, 20-25 mm. long, approximately 2 mm. broad, entire, ciliate, sparingly pubescent; flowers few, solitary at the nodes, subtended by 2-foliaceous bracts; calyx deeply 5-parted, 15 mm. long, flaccid-pubescent on the main nerves; lobes subulate-setaceous, approximately 10 mm. long; corolla pubescent on the external surface, 17 mm. long, scarcely exceeding the calyx in length, tube 6.5-7 mm. long, throat more or less equalling the tube in length; style 12 mm. long, pubescent, stigma oblique; capsule linear, glabrous, 7-8 mm. long, 4-seeded. Distribution: northeastern Mexico. Specimens examined : Vicinity of Victoria, Tamaulipas, alt. 320 m., 1 May-13 June, 1907, Palmer 492 (US type, M photograph and fragment). 1 Dyschoriste Greenmanii Kob., sp. nov., planta prope 2 dm. alta, ascendens a perenne basi; caulibus gracilibus, ramosis, pubescentibus ad nodos, aliter glabris; foliis linearibus, sessilibus, 20-25 mm. longis, 2 mm. latis, integerrimis, ciliatis, parce pubescentibus; floribus paucis, solitiriis ad nodos, subtendentibus 2-foliaceis brac- teis; calyce profunde 5-diviso, 15 mm. longo, flacciio-pubescente nervis, lobis subu- lato-setaceis, prope 10 mm. longis; CDrolla 17 mm. longa, paulo calyce longiore, tubo 6.5-7 mm. longo, fauce plus minus\e aequante tubum; stylo 12 mm. longo, pubes- cente, stigmata obliqua; capsula lineari, glabra, 7-8 mm. longa, 4-sperma. — Type collected in the vicinity of Victoria, Tamaulipas, Mexico, 1 May-13 June, 1907, E. Palmer m (US). [Vol. 15 34 ANNALS OF THE MISSOURI BOTANICAL GARDEN 8. Dyschoriste Niederleiaii Lindau in Engl. Bot. Jahrb. 19 (Beibl. 48): 15. 1894. Low-growing perennial; stems ascending, about 1 dm. high, branches tetragonal, minutely puberulent; leaves linear, approx- imately 30 mm. long, 5 mm. broad, somewhat obtuse at the apex, entire, glabrous, sparsely pilose at the base, petiole 2 mm. long; flowers single, axillary, subtended by small bracts; calyx 5-parted, puberulent, 11 mm. long, tube and calyx lobes of equal length; corolla puberulent on the external surface, 15 mm. long, ventri- cose; style 6 mm. long, filiform, pubescent; mature capsule unknown. Distribution: Argentina. Specimens examined : "Ad Primer Misionero de Hernandez," Puck and Fernandez (Niederlein 42), Argentina, Feb. 1884, (B type, M photograph). 9. Dyschoriste Schiedeana (Nees) O. Ktze. Rev. Gen. PI. 2: 486. 1891. Calophanes Schiedeanus Nees in DC. Prodr. 11: 111. 1847; Mueller in Walpers, Ann. 5: 647. 1858, including var. multi- fiorus; Hemsl. in Biol. Cent.-Am. Bot. 2: 502. 1882. Perennial, ascending or erect from a suffruticose base, 3-6 dm. high; stems somewhat angular, hirsute, branched near the base; leaves usually linear-lanceolate, lower cauline leaves occasionally obovate, 20-25 mm. long, 3-5 mm. broad, acute at the apex, narrowed at the base into a very short petiole, entire, hirsute on both surfaces; flowers axillary, usually two in an axil, subtended by bracts which equal or nearly equal the calyx in length ; calyx 5-parted, 11-12 mm. long, lobes 7 mm. long, subulate-setaceous, hirsute; corolla 14-15 mm. long, pubescent on external surface, tube 4 mm. long; mature capsule 7-8 mm. long, linear, glabrous, acute at apex, 4-seeded; seeds typical. Distribution: eastern Mexico. Specimens examined: Nuevo Leon: near Monterey, alt. 550 m., Aug. 1911, Arsene 6411 (US). Vera Cruz: in fields near Jalapa, date lacking, Schiede 122 (M photograph of type, B type); Mirador, date lacking, Sartoriu* (US 55208). 19281 KOBUSKI — MONOGRAPH OP DYSCHORISTE 35 10. Dyschoriste Rosei 1 n sp. PL 7, fig. 1. Low-growing perennial; stem pubescent, branched, ascending or erect, 12-15 cm. high; leaves sessile, linear, entire, glabrous, 18-25 mm. long, 2 mm. broad; flowers few, solitary at the nodes, usually near the apex of the stem, slightly pedicellate, subtended by 2-foliaceous bracts; calyx 5-parted, glabrous except for the ciliate margin of the unequal, subulate-setaceous lobes, shorter posterior lobes 8-9 mm. "ong, anterior lobes 11-12 mm. long; corolla externally pubescent, 25 mm. long, tube 10 mm. long, diverging abruptly into a broadly ampliated throat which is approximately equal the tube in length; stamens occasionally incompletely didynamous: ovary 2-celled, each cell containing 2 ovules, style 17-18 mm. long, stigma oblique, 2 mm. long; mature capsule not seen. Distribution: western Mexico. Specimens examined: Durango: without definite locality, 13 Aug. 1897, Rose 2259 (US type, M fragment and photograph). Jalisco: on road between Mesquite and Monte Escolebo, 26 Aug. 1897, Rose 3581 (US). 11. Dyschoriste jaliscensis 2 n. sp. PL 8. Stems several, 3-4 dm. high, erect from a ligneous, perennial base, branching, pubescent; leaves linear to linear-oblanceolate, 2.5-3.5 cm. long, 2 mm. dt less broad, narrowed at the base, 1 Dyschoriste Rosei Kob., sp. :iov., humilis perennis; caule pubescente, ramis ascendentibus vel erectis, 12-15 cm. altis; foliis linearibus, sessilibus, 18-25 mm. longis, 2 mm. latis, integerrimis, glabris; floribus paucis, solitariis ad nodos, fere prope apicem, subpedicellatis, subi endentibus bracteis; calyce 5-diviso, glabro, lobis inaequalibus, subulato-setaceis, po* terioribus lobis 8-9 mm. longis, anterioribus lobis 11-12 mm. longis; corolla extus puberula, 25 mm. longa, tubo 10 mm. longo, diver- gente subito in late ampliato faice; staminibus didynamis, subinde imperfectis; ovario biloculo, stylo 17-18 mm. longo; stigma obliqua, 2 mm. longa; capsula ignota. •Type collected in the state of Dirango, 13 Aug. 1897, Rose 2259 (US). * Dyschoriste jaliscensis Kob., up. nov., planta 3-4 dm. alta; caulibus pluribus, erectis a lignoso basi, ramis pube3centibus; foliis linearibus, linearo-oblanceolatis, 2.5-3.5 cm. longis, 2 mm. minusve latis, basi attenuatis, integerrimis, pubescentibus; floribus maioribus, bracteis foliaceis, 10 mm. longis; calyce 17-18 mm. longo, pu- bescente, lobis 11-12 mm. longis, subulato-setaceis, ciliatis; corolla prope 30 mm. longa, tubo 12-13 mm. longo, pauce fauce longiore; antheris 2 mm. longis, basi bicalcaratis; stylo 20-21 mm. longo, stigmata obliqua, capsula ignota. — Type col- lected on rocky hills near Guadalaj ira, Jalisco, 27 June, 1893, Pringle 5481 (G). [Vol. 15 36 ANNALS OF THE MISSOURI BOTANICAL GARDEN entire, pubescent; flowers comparatively large, subtended by 2-foliaceous bracts which arc? about 10 mm. in length; calyx 17- 18 mm. long, pubescent, lobes 11-12 mm. long, subulate, seta- ceous, ciliate; corolla approximately 30 mm. long, tube 12-13 mm. long, slightly longer than the throat; anthers about 2 mm. long; style 20-21 mm. long, stigma oblique; mature capsule not seen. Distribution : western Mexico. Specimens examined : Jalisco: rocky hills near Guadalajara, 27 June, 1893, Pringle 6481 (US, G type, M photograph and fragment). 12. Dyschoriste angustifolia (Hemsl.) 0. Ktze. Rev. Gen. PI. 2: 485. 1891. Calophanes angustif olius Hemsl. in Biol. Cent.-Am. Bot. 2: 502. 1882. Stem erect, strict, 4-5 dm. tall, more or less villous-hirsute ; leaves subsessile, linear-lanceolate, 1.5-2.5 cm. long, 3-5 mm. broad, acute at the apex, attenuate at the base, entire, scabrous; flowers axillary, disposed in dense shortly pedunculate cymes in the axils of the upper leaves, subtended by narrow bracts which almost equal the calyx in length; calyx deeply 5-parted, 15 mm. long, scabrous, lobes subulate-setaceous, nearly equalling the tube of the corolla, ciliate; corolla bilabiate, puberulent on the external surface, approximately 25 mm. long; anther cells shortly mucronate at the base; ovary 2-celled, cells 2-ovulate, glabrous, stigma linear, oblique; mature capsule not seen. Distribution: southern Mexico. Specimens examined: Oaxaca: without definite locality, coll. of 1842, Ghiesbreghl (K, M photograph). 13. Dyschoriste linearis (Torr. & Gray) O. Ktze. Rev. Gen. PI. 2: 486. 1891; Lindau in Engl. & Prantl, Nat. Pflanzenfam. 4 3b : 302. 1895; Lindau in Bull. Herb. Boiss. II. 6: 844. 1906. Dipteracanthus linearis Torr. & Gray, Bost. Jour. Nat. Hist. 5: 50. 1845 (PI. Lindh. 1: 50). Calophanes linearis Gray, Syn. Fl. N. Am. 2 1 : 324. 1878, and 19281 KOBUSKI — MONOGRAPH OF DYSCHORISTE 37 ed. 2, 1886; Hemsl. in Bio.. Cent.-Am. Bot. 2: 503. 1882; Small, Fl. Southeastern U. S. 1083. 1903, and ed. 2, 1913. Calophanes oblongifoliux var. texensis Nees in DC. Prodr. 11: 108. 1847; Torr. in Emory's Rept. U.S. & Mex. Bound. Surv. 2 (Bot.) : 122. 1859. Calophanes ovatus Nee? in DC. Prodr. 11: 108. 1847, not Ruellia ovata Cav. Ruellia ovata Benth. Fl. Hartweg. 89. 1842, not Cav. i.e., as to plant of Drummond from Texas. Stem 18-42 cm. high, erect and strict, branched and diffuse, hirsute with both rigid and short hairs, sometimes sparsely- pubescent or nearly glabrous, not cinereous; leaves linear-oblan- ceolate to oblong-spathuUte, 1.8-6.5 cm. long, entire, lineolate, rather rigid, pubescent on midrib and veins, margin ciliate; bracts foliaceous, frequeotly in short-leafed specimens equalling the length of the leaf; calyx 5-cleft, densely lineolate, giving the appearance of appressed hairs, lobes 9-13 mm. long, subulate- setaceous, more or less hispid, ciliate, calyx tube 4.5-6 mm. long, in most cases one-half the length of the lobes; corolla somewhat bilabiate, 26-27 mm. long, pubescent on external surface, tube 5-7 mm. long and slightly shorter than the abruptly ampliated limb ; anther cells oblong ; capsule 4-seeded ; seeds flat. Distribution: Texas to New Mexico and northern Mexico. Specimens examined: Texas: rocky prairies, 12 July, 1903, Reverchon (M 120836); western Texas, 1890, Nealley (Ch 254803) ; 1846, Lindheimer 504 (US) ; Drummond 2 no. 1 78 (G type) ; Drummond 256 (G) ; dry prairies, Bay City, Matagorda Co., 6 May, 1916, Palmer 9667 (M) ; prairies, Ganado, Jackson Co., 20 March, 1916, Curtiss 9216 (M); dry open ground, Vanderbilt, Jackson Co., 10 May, 1916, Palmer 9708 (M) ; Calhoun Co., 10 Aug. 1920, Drushel 4136 (M) ; dry rocky prairies near Dallas, Dallas Co., June- July, Curtiss 194-1 (FM, M) ; dry rocky prairies, Dallas, Dallas Co., May-June, 1879, Reverchon (FM 88468); dry rocky prairies near Dallas, Dallas Co., date lacking, Reverchon 1941 (G, M, US); rocky limestone prairies, Dallas Co., 15 May, Reverchon 722 (M, US) ; Dallas Co., 23 May, 1903, Bebb 1848 (FM); rocky prairies, Dallas Co., 18 May, 1900, Reverchon 2114 (M); field and gardens, Fort Worth, 38 ANNALS OF THE MISSOURI BOTANICAL GARDEN (Vol. 15 Tarrant Co., 8 June, 1909, Ruth 104 (US) ; along roadsides near Fort Worth, Tarrant Co., 5 July, 1909, Ruth 30 (FM) ; dry grounds near Fort Worth, Tarrant Co., 1 June, 1910, Ruth 81 (FM); Austin, Travis Co., 25 June, 1920, Tharp 733 (US); Austin, Travis Co., 1897, Buckley (M 120803); dry hills, Austin, Travis Co., 13 May, 1872, Hall 431 (G) ; dry prairies, Austin, Travis Co., 1G May, 1872, Hall 428 (US) ; along Corpus Christi Bay, Nueces Co., alt. sea level, 9-12 April, 1894, Heller 1529 (G, M US, FM); dry open ground, Strawn, Palo Pinto Co., 27 June, 1918, Palmer 14252a (M); Dublin, Erath Co., 1893, Maxwell 49 (Ch); Round Top Mt., Comanche Co., 9 May, 1900, Eggert (M, 120809); Gillespie Co., date lacking, Jermy 472 (M); rich hillside, Boerne, Kendall Co., 19 May, 1916, Palmer 9811 (M) ; in dried river beds of mountain rivers north of Braunfels, Comal Co., 1846, Lind- heimer 325 (M) ; in pastures, Bracken, Comal Co., 3 Aug. 1903, Groth 230 (G); humid prairie and along margin of shrubs near New Braunfels, Comal Co., May, 1848, Lindheimer 677 (G, M, FM, US); in grass and on black prairie loam, New Braunfels, Comal Co., May, 1846, Lindheimer 111 (G, M); Comanche Springs, New Braunfels, Comal Co., May, 1851, Lindheimer 1063 (M, G, FM, US); New Braunfels, Comal Co., May, 1851, Lind- heimer 552 (M) ; in open pastures, 5 miles south of San Antonio, Bexar Co., 14 May, 1920, Schultz 146 (US); San Antonio, Bexar Co., 1918, Slater (US 891769); San Antonio, Bexar Co., 1884, Havard (Ch 252081) ; Bexar Co., date lacking, Jermy 62 (M, US) ; San Antonio, date lacking, Jermy 249 (G) ; San Antonio, Bexar Co., 27 April, 1911, Mr. & Mrs. Clemens 1069 (M); Bexar Co., date lacking, Jermy 31 (US); San Diego, Duval Co., 1885, Croft 6465 (M) ; San Diego, Duval Co., July, 1885, Croft 6660 (M) ; dry open ground, Baird, Callahan Co., 26 May, 1918, Palmer 13698 (M) ; Abilene, Taylor Co., 22 May, 1902, Tracy 8079 (G, M, FM, US); prairie north of Abilene, Taylor Co., 7 June, 1900, Eggert (M 120800); calcareous banks, Menard Co., 11 May, 1917, Palmer 11871 (M); dry alluvial soil along creek, Lacey's ranch, Kerr Co., 10 June, 1917, Palmer 12229 (M) ; rocky ground, Sweet- Nolan Co., 27 May (M); Kni bocker ranch, Dove Creek, Tom Green Co., May, 1880, Tweedy 180 (US); Fort Clark, Kinney Co., 10 May, 1893, Mearns 1432 1928) KOBUSKI — MONOGRAPH OF DYSCHORISTE 39 (US); Devils River, Valverde Co., May, 1913, Orcutt 6230 (M); prairie north of Stanton, Martin Co., 13 June, 1900, Eggert (M 120810) ; western Texas to El Paso, New Mexico, El Paso Co., May-Oct. 1849, Wright 432 (G, FM, US). New Mexico: Slaughter Canyon, Guadalupe Mts., 12-20 Aug. 1924, Standley 40624 (US). Mexico: Sierra Madre, 45 miles south of Saltilla on border of states of Coahuila and Nuevo Leon, July, 1880, Palmer 2083 (G) ; roadside, Piedras Nigras, Coahuila, May, 1883, Havard (Ch 267840, US 147426) ; near Huasemote, Durango, 15 Aug. 1897, Rose 8495 (US). 14. Dyschoriste bilabiata (Seemann) O. Ktze. Rev. Gen. PI. 2: 486. 1891; Lindau in Bull. Herb. Boiss. 7: 575. 1899. PL 9. Calophanes bilabiatus Seem. Voy. H. M. S. Herald, 324. pi 65. 1852-57; Hemsl. in Biol. Cent.-Am. Bot. 2: 502. 1882. Stems 6-7 dm. high, erect from a perennial base, branching, pubescent; leaves ovate-oblong, 4-5 cm. long, 1.5-2 cm. broad, acute at the apex, narrowed at the base into a petiole, repand- denticulate, densely pubescent on both surfaces; flowers axillary, cymose, cymes pedunculate, 3-5-flowered, subtended by subulate bracts; calyx 5-lobed, 12 mm. long, tube 5 mm. long, pubescent, lobes subulate-setaceous, ciliate; corolla subbilabiate, pale blue, 14 mm. long, tube 5-6 mm. long, subventricose, pubescent on the external surface; filaments hairy; ovary glabrous, style filiform, stigma linear, oblique; mature capsule linear, glabrous, 4-seeded. Distribution: western Mexico. Specimens examined : Cero de Pinal, Sinaloa, Dec. 1848, Seemann 1518 (K type, M photograph only). 15. Dyschoriste decumbons (Gray) O. Ktze. Rev. Gen. PI 486. 1891; Lindau in Engl. & Prantl, Nat. Pflanzenfam. 302. 1895. •s decumbens Gray, Syn. Fl. N. Am. ed. 1, 2 1 : 325. 1878, and ed.*2, 1886; Hemsl. in Biol. Cent.-Am. Bot. 2: 502. 1882. 40 [Vol. 15 [SSOUBI Calophanes oblongifolius Torr. Bot. Mex. Bound. Surv. 122. 1855, not Don. Cinereous-puberutent throughout; stems mostly spreading on the ground from a ligneous base, occasionally erect, branched; leaves spatulate to oblanceolate, 2-3 cm. long, 0.5-1.1 cm. broad, entire, apex usually obtuse, sometimes slightly mucronate, base attenuated, often ha,ving the; appearance of a petiole ; flowers few, in foliose, bracteate clusters; calyx 15-20 mm. long, at maturity exceeding the capsule by as much as 10 mm., 5-cleft, tube 5-7 mm. long, lobes subulate-setaceous, hardly twice the length of the tube; corolla purple, 18-20 mm. long, tube a little longer than the throat, slightly ampliated at the base; anther cells oblong, filaments united at the base of the corolla-throat; seeds 4, sub- orbicular and flattened. Distribution: dry soil, western Texas to Arizona, and the plateau region of northern Mexico. Specimens examined : Texas: Chenates region of western Texas, 1889, Nealley 580 (857) (US); infrequent on slopes between Marfa and Alpine, 15 April, 1919, Hanson 688 (US). New Mexico: Valley of the Rio Grande, 1851, Wright 1462, 1468 (M). Arizona: Sonoito Valley, Santa Cruz Co., alt. 1833 m., Aug. 1874, Rothrock 687 (US); Fort Huachuca, 1890, Patzky (US 721394); Fort Huachuca, Cochise Co., May, 1892, Wilcox (US 55273, M 120796); roadway, Chiricahua Mts., Cochise Co., alt. 1400 m., 9 Oct. 1907, Blumer 2228 (FM) ; Fort Huachuca, Cochise Co., 1894, Wilcox 150 (US); foothills of Santa Rita Mts., near Greaterville, Pima Co., alt. 1666 m., 16 Sept. 1916, Shreve 4978 (US); plains about Huachnca Mts., Aug. 1882, Lemmon (US 55278); locality lacking, 1875, Rothrock (US 55277); Fort Hua- chuca, Cochise Co., 26 April-21 May, 1890, Palmer 472 (US). Mexico: Sierra Mojado Mts., Coahuila, 19 April, 1892, Jones 874 (US, M); near the border of Coahuila and Nuevo Leon, Feb.-Oct. 1880, Palmer 1009 (US); Saltillo, Coahuila, alt. 1600 m., 1911, Arsene 6472 (US) ; Saltillo, Coahuila, July, 1880, Palmer 2082 (G); Saltillo, Coahuila, May, 1898, Palmer 125 (US, M); Lerios, 15 leagues east of Saltillo near the border of Coahuila and 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 41 Nuevo Leon, alt. 3000 m., 10-13 July, 1880, Palmer 1010 (15458) (US, M) ; on road near Colatlan, Zacatecas, 31 Aug. 1897, Rose 3615 (US) ; exact locality lacking, San Luis Potosi, 1897, Schaffner 354 (647) (US) ; San Luis Potosi, alt. 2000-2500 m., 1878, Parry & Palmer 699 (FM, M, G, US) ; near Queretaro, 20-23 Aug. 1909, Rose & Rose 111 48 (US); San Andres Mts., Chihuahua, 22 Aug. 1900, Trelease 852 (M); hills near Chihuahua, Chihuahua, 30 Sept. 1886, Palmer 1075 (M) ; vicinity of Chihuahua, Chihuahua, alt. 1300 m., 1-21 May, 1908, Palmer 208 (US); rocky hills near Chihuahua, Chihuahua, May, 1885, Pringle 66 (US, G, FM); Cosihuiriachic, west of the city of Chihuahua, Chihuahua, 20 Sept. 1846, Wislizenus 185 (M); City of Durango, Durango, 1 Aug. 1898, Nelson 4597 (US) ; in the vicinity of Durango, Duran- go, April-Nov. 1896, Palmer 809 (FM, US, M, G) ; Palmer 980 (US, M); Palmer 276 (US, G); Sonora, 8 Sept. 1851, Thurber 974 (G) . 16. Dyschoriste crenulata 1 n. sp. PI, 7, fig. 2. Stems several, 1-2 dm. high, erect or ascending from a peren- nial, ligneous base, pubescent; leaves more or less spathulate to obovate, 2-3 cm. long, 0.3-1 cm. broad, acute to obtuse at the apex, attenuate at the base, densely cinereous pubescent, margin crenulate; calyx 5-parted, 17-18 mm. long, nearly equalling the length of the corolla, tube and lobes of nearly equal length, lobes subulate-setaceous, cinereous, ciliate; corolla 18-19 mm. long, pubescent on the external surface, throat slightly longer than the tube; anthers occasionally unequally didynamous, style 11-12 mm. long, stigma oblique: mature capsule not seen. Distribution: south Texas, south into Tamaulipas. Specimens examined: i Dyschoriste crenulata Kob., 3p. nov., planta 1-2 dm. alta; caulibus pluribus, erectis vel ascendentibus a perenne lignoso basi, pubescentibus; foliis subsessilibus, plus minusve spathulatis vel ple-umque obovatis, 2-3 cm. longis, 0.6-1 cm. latis, apice acutis vel obtusis, crenulatis, basi attenuatis, cinereo-pubescentibus; calyce 5-diviso, 17-18 mm. longo, prope aequante corollam, tubo lobes aequante, lobis subulato-setaceis, cinereis, ciliatie; corolla 18-19 mm. longa, extus puberula, fauce tubo paulo longiore; staminibus clidynamis, subinde imperfectis; stylo pubescente, 11-12 mm. longo, stigmata obliqua; capsula ignota. — Type collected on road from "San Fernando to Jimeney," stato of Tamauupas, Mexico, 26-27 Feb., 1902, E. W. Nelson 6604- (G) . 42 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 Texas: Brazos Santiago, 1399, Nealley 124 ($57) (US). Mexico: "San Fernando to Jimeney," Tamaulipas, 26-27 Feb. 1902, Nelson 6604 (G type, US isotype, M photograph and fragment). 17. Dyschoriste capitata (Oerst.) O. Ktze. Rev. Gen. PI. 2:486. 1891. Calophanes capitatus Oerst. in Vidensk. Meddel. 121. 1854; Mueller in Walpers, Ann. 5: 647. 1858; Hemsl. in Biol. Cent.- Am. Bot. 2: 502. 1882. Stems several, frequently branching and ascending from a ligneous base, 4 dm. high, subtetragonal, often geniculate, pubes- cence becoming more pronounced and flaccid near the apex; leaves obovate, 15-27 mm. long, 8-12 mm. broad, obtuse at the apex, attenuate into a petiole varying from a subsessile condition to 5 mm. in length, entire, ciliate, upper surface hirsute, especially on basal portion of midrib and petiole, sparingly so along midrib and veins of lower surface; flowers congested in heads at the apex of the stem and branches, subtended by oblanceolate bracts, the basal portion invested with long whitish hairs; calyx 9-10 mm. long, 5-lobed, joined for one-third its total length, possessing the same pubescence as the bracts, together giving a distinctly whitish appearance to the inflorescence, lobes subulate-setaceous; corolla 15-16 mm. long, puberulent on the external surface; filaments terminated by an emarginate, muticous prolongation at the apex; staminodium sometimes present; ovary glabrous, stigma linear, oblique; mature capsule glabrous, 8-9 mm. long, 4-seeded. Distribution: mountains of southern Mexico. Specimens examined: Prov. of San Luis Potosi, 1851, Oersted 808 1 (C); Sierra de San Felipe, Oaxaca, alt. 2000 m., 15 June, 1897, Pringle 6718 (FM, G, M, US) ; Valley of Oaxaca, alt. 1550 m., 8 June, 1897, Con- zatti &< Gonzales 282 (G). 18. Dyschoriste Pringlei Greenm. in Proc. Am, Acad. 40: 32. 1905. 1 This citation refers to a photograph of the only Oersted specimen of D. capitata found in the Copenhagen Herbarium. 1028] KOBUSKI — MONOGRAPH OF DYSCHORISTE 43 Stems several, 1-2 dm. in length, erect or ascending from a ligneous perennial base, densely hirsute-pubescent or subtomen- tose; leaves lance-elliptic to slightly obovate, 1.5-4 cm. long, 0.5-1.6 cm. broad, obtuse or acute, entire, narrowed below to a subpetiolate base, sparingly hirsute-pubescent on both surfaces; flowers crowded in the axils of the upper leaves, forming a sub- capitate, leafy inflorescence; calyx 13-14 mm. long, densely pubescent with white flaccid hirsute hairs, divided to somewhat below the middle, divisions lance-attenuate; corolla tubular- campanulate, 3-4 cm. long, externally pubescent, more or less purplish-maculate, at least in the dried state; stamens adnate to the corolla for about one-half its length, anthers rather conspicu- ously calcarate; ovary glabrous, style pubescent; mature capsule not seen. Distribution: southwestern Mexico. Specimens examined : Barranca of Rio Blanco near Guadalajara, Jalisco, alt. 1500 m., 22 July, 1902, Pringle HS18 (G, FM, US) ; deep canyons near Guadalajara, Jalisco, 1 July, 1889, Pringle 2907 (G type, FM, M photograph). 19. Dyschoriste oaxacensis \ l n. sp. PI. 10. Stems several, procumbent, ascending from a woody base, 1-2 dm. high, pubescent with lineolations showing through pubes- cence; leaves sessile, oblanc€ olate, occasionally somewhat spathu- late, 15-20 mm. long, 3-5 mm. broad, obtuse at the apex, ciliate, sparsely pubescent or glabrous, appealing scabrous because of the irregular scattering of cystoliths; flowers axillary, congested at the apex of stem and branches, producing a capitate-like in- florescence, subtended by oblanceolate bracts, approximately 10 1 Dyschoriste oaxacensis Kob., sp. nov., caulibus pluribus, procumbentibus, lineolatis, 1-2 dm. altis; foliis sessilibus, oblanceolatis, rare subspathulatis, 15-20 mm. longis, 3-5 mm. latis, ciliatis, pauce pubescentibus; floribus axillaribus, in apice caulis ramorumque capitatim congeutis; bracteis plus minusve 10 mm. longis; calyce 12 mm. longo, glabro, cystolithero, lobis subulatis, setaceis, ciliatis, 7 mm. longis; corolla subbilabiata, extus puberal a, 20 mm. longa, tubo 7 mm. longo; antheris basi bicalcaratis ; stylo lineari, pubescente, 13-14 mm. longo, stigmata lineari, obliqua; capsula 10-11 mm. longa,, glabra, 4-sperma; seminibus subrotundatibus, planis, humectatis mucilaginosis. — Type collected on calcareous hills, Las Sedas, Oaxaca, Mexico, 9 July, 1891, Prinjle 6712 (G). 44 ANNALS OF THE MISSOURI BOTANICAL GARDEN (Vol. 15 mm. long, calyx about 12 mm. long, except for the lobes glabrous and covered with cystoliths, lobes subulate-setaceous, ciliate, 7 mm. long; corolla externally puberulent, 20 mm. long, tube 7 mm. long, somewhat bilabiate; stamens adnate below the middle of the corolla limb; ovary glabrous, style 13-14 mm. long, stigma linear, oblique; mature capsule 10-11 mm. long, glabrous, 4- seeded; seeds oblique, somewhat rounded, flattened. Distribution: southern Mexico. Specimens examined: Oaxaca: calcareous hills, Las Sedas, alt. 2000 m., 19 July, 1891, Pringle 6712 (M type, G, FM, US); Las Sedas, alt. 2000 m., 2 June, 1907, L. C. Smith /+19 (G); Nochixtlan, alt. 2000 m., 19 June, 1907, Conzatii 1858 (FM). 20. Dyschoriste pinetorum 1 n. sp. PI. 11. Stems erect or ascending from a woody, perennial base, branches often arising from nodes of prostrate or erect growth of previous year, subquadrangular, 20-30 cm. high, nodes frequently 5-6 cm. distant, pubescent especially at the apex; leaves obovate- elliptic, 25-35 mm. long, 10-18 mm. broad, acute to subrotund at the apex, subsessile, attenuate at the base into a very short petiole, entire, ciliate, hirsute on both surfaces, the pubescence confined to midrib and veins on the under surface, veins con- spicuous; flowers disposed in heads at the tips of the stems and branches and subtended by oblanceolate bracts; calyx 11-13 mm. long, divided two-thirds the distance to the base into five subu- late-setaceous, ciliate lobes, pubescence similar to that of the bracts, together giving a canescent appearance to the leafy capitate inflorescence; corolla puberulent on the external surface, 1 Dyschoriste pinetorum Kob., sp. nov., caulibus erectis vel ascendentibus alignoso perenne basi, subquadrangularis, 20-30 cm. altis, nodis saepe 5-6 cm. diversis, pubescentibus praesertim apice; ramis saepe crescendentibus ex nodis prostratorum vel erectorum caulorum antecedents anni; foliis subsessilibus, obovato-ellipticis, 25-35 mm. longis, 10-18 mm. latin, apice acutis vel subrotundis, basi in petiolum brevissimum attenuates, integris, ciliatis, utrinque hirsutis, praesertim ad costos nervosque subtorum; floribus apice caulis ramorumque capitatim congestis; calyce 11-13 mm. longo, diviso ad ^ a basi in quinque subulatis, setaceis, ciliatis lobis, pubescentibus, canescentibus; corolla extus puberula, 20-21 mm. longa, tubo limbum aequante; stylo hirsuta, stigmata lineari, obliqua; capsula ignota. — Type collected in sandy fields under pines, near Patzcuaro, Michoacan, 31 July, 1892, C. G. Pringle 4134 (G). 1928] KOBUSKI— MONOGRAPH OF DYSCHORISTE 45 20-21 mm. long, tube and ;hroat of approximately equal length stamens adnate to about :he middle of the corolla-tube; styL hirsute, stigma linear, oblique; mature capsule not seen. Distribution: southern Mexico. Specimens examined: Michoacan: sandy fields under pines near Patzcuaro, 31 July 1892, Pringle A1SA (G type, isotypes in M, Ch, FM, US). 21. Dyschoriste sagittata 1 n. sp. PI mm Low-growing perennial; stems ascending 1-2 dm. high from a ligneous base, glabrous or nearly so, densely covered with cystoliths, quadrangular, somewhat winged, branched; leaves sessile, elliptic-obovate, 15-25 mm. long, 9-12 mm. broad, usually obtuse at the apex and base, glabrous, entire; bracts slender, lanceolate, glabrous, 8 mm. long, bracteoles minute, acuminate, long; calyx 8-10 mm. long, minutely pubescent on the nerves, lobes about 5 mm. long, subulate-setaceous, sparsely and minutely ciliate ; corolla pu bescent on the external surface, barely exceeding the calyx in length, ventricose, slightly bilabiate, lobes rounded, margins crenate; stamens small, filaments adnate to a little below the middle of the corolla-throat, anther cells converg- ing towards the acute apex, slightly diverging at the calcarate base, giving a sagittate appearance, approximately 0.5 mm. long; style 4-5 mm. long, minutely pubescent or glabrous, stigma lobed; mature capsule not seen. Distribution: Paraguay. Specimens examined: Paraguay: in region along the Alt a Parana River, 1909-10, Fiebrig 6S8S (G type, M fragment and photograph). 1 Dyschoriste sagittata Kob. sp. nov., humilis perennis; caulibus ascendentibus, 1-2 dm. altis a lignoso basi, glatris, cystoiitheris, quadrangularis, alatis, ramosis; foliis sessilibus, elliptico-obovatis, 15-25 mm. longis, 9-12 mm. latis, glabris, apice basique obtusis; bracteis angustis, lanceolatis, glabris, 8 mm. longis, bracteolis minutis, acuminatis, 2-3 mm. lon;is; calyce 8-10 mm. longo, puberulente in nervis lnKnnim «*«l\«>iim lnhin nmno S mm. Innins. Mihulatis-setaceis. oarce et minute mi ata, lobis rotundis, marginibus c*enatis; antheris sagittatis, basi divergentibus et bicalcaratis, apice acutis; stylo 4-f. mm. longo, parce pubescente vel glabro, stigmata trilobata, medio lobo longissimo, circiter 0.5 mm. longo; capsula ignota. — Type collected in the region along the Alta Parana River, Paraguay, coll. of 1909-10, Fiebrig 6383 (G). 46 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 486. 1891 um (Nees) O. Ktze. Rev. Gen. PI. 2: Calophanes Scrpyllum Nees in Mart. Fl. Bras. 9: 26. 1847; Nees in DC. Prodr. 11: 110. 1847. Stems 1 dm. or less high, erect from a suffruticose base, pubes- cent ; lower leaves ovate, upper leaves ovate-lanceolate, narrowed at the base, 10-12 mm. long, 3-5 mm. wide, entire, glabrous, eubsessile; flowers few, usually toward the apex of the stem, subtended by leafy, glabrous bracts; calyx unequally 5-parted, submembranaceous , sparsely pubescent, 12 mm. long, lobes greatly attenuated, approximately twice as long as the tube, ciliated; corolla 12-13 mm. long, pubescent on the external surface, tube very short, not more than 3-4 mm. long, ampliation into throat apparently beginning at the base of the tube ; stamens abruptly yet obtusely appended at the apex, bicalcarate at the base; ovary 2-celled, each cell possessing 2 ovules, style sparsely pubescent, filiform, stigma oblique; mature capsule not seen. Distribution: southeastern Brazil. Specimens examined: dry fields near Rio Pardo, Sept. 1826, Riedel 45 M fragment and photograph) I 23. Dyschoriste Lloydii 1 n. sp. Stems branched near the base, erect, pubescent, 1-1.5 dm. high; leaves sessile, oblong-oblanceolate, 18-20 mm. long, 3-4 mm. broad, sparingly hirsute-pubescent on both surfaces, often con- fined to the midrib and veins, entire; bracts foliaceous, nearly equalling the calyx in length; calyx 10-10.5 mm. long, tube 4-5 mm. long, sparingly pubescent, lobes ciliate, subulate-setaceous; corolla 14 mm. long, tube 5 mm. long, approximately equalling the throat in length; ovary and stamens typical of the eenus: i Lloydii centibus, 1-1.5 dm. altis; foliis sessilibus, oblongo-oblanceolatis, 18-20 mm. longis, 3-4 mm. latis, integris, ulrinque parce hirsuto-pubescentibus, praesertim ad costos nervoBque; bracteis foliaceis; calyce prope 10-10.5 mm. longo, tubo 4-5 mm. longo. parum pubescente, lobis ciliatis, subulato-setaceis ; corolla 14 mm. longa, tubo 5 mm.' ' ' J v — — -- ******* *v^*_i ft **j vuuw %j UUUi longo, prope aequante ampliatum faucem; capsula lineari, glabra, 7-8 mm. longa, semmi Zacatecas, Mexico, 1908, F. E Lloyd 199 (US). Hacienda 1928] KOBUSKI — MONOGRAPH OF DYSCHOBJSTE 47 labrous. 7-8 mm. lone, 4-seeded; seeds flattened suborbicular, oblique. Distribution: central M Soecimens examined: flats, Hacienda de Cedros, summer 1908, Lloyd (US microph ylla (Cav.) 0. Ktze. in Rev. Gen. PI PL 486. 1891. Calophanes microphyllui (Cav.) Nees in DC. Prodr. 11: 113. 1847; Hemsl. in Biol. Cent.-Am. Bot. 2: 502. 1882. Ruellia microphylla Cav. Ic. 6: 63, pi 586, /. 2. 1801 ; Spreng. Syst. 2: 821. 1825. Dyschoriste Jasminum 0. Ktze. in Rev. Gen. PL 2: 486. 1891. Calophanes J asminum-mexicanum Nees in DC. Prodr. 11: 110-111. 1847; Hemsl. in Biol. Cent.-Am. Bot. 2: 502. 1882. Stem rising or ascending from a perennial base, pubescent; leaves distinctly ovate, obtuse-rotund at the apex, attenuate at the base into a petiole, 1.^-3 cm. long (including petiole), 0.9-1.2 cm. broad, glabrous except for slight pubescence on midrib and margin, entire; inflorescence terminal or on rather short lateral branches, subtended by fDliaceous bracts; calyx 5-parted, 12-13 mm. long, somewhat pubescent, especially on the lobes, lobes subulate-setaceous, one-half the total length of the calyx, ciliate; corolla puberulent on the external surface, 13-14 mm. long, tube 8-9 mm. long, ampliating abruptly into the short throat, lobes rounded; stamens adnate below the center of the corolla throat, filaments broadening toward the base; style filiform, pubescent, stigma ampliated, posterior lobe rudimentary; capsule glabrous, 4-seeded, about 8 mm. long. Distribution: southern Mexico. Specimens examined: Puebla: Chalmo y Sar. Miguel, 1789-1794, D. Luis Nee (Ma type, M 928687, photograph) ; vicinity of Puebla at the Rancho Losado, alt. 2194 m., 29 Aug. 1909, Bro. Nicolas 299 (US); vicinity of Puebla, date and number lacking, Bro. Arstne (US 1004058); Cerro Guadalupe, vicinity of Puebla, alt. 2200 m., June, 1908, ArsSne 1983 (M, G, US) ; entre les haciendas Santa [Vol. 15 48 ANNALS OF THE MISSOURI BOTANICAL GARDEN Barbara et Cristo, sur l'Alseseca, alt. 2150 m., 27 June, 1907, Arsene 1528 (US); Hanta Barbara, Puebla, 1 June, 1907, Arsene 1075 (US, M). Mexico: hills in the valley of Mexico, alt. 2500 m., 24 Aug. 1902, Pringle 11322 (G, US). Michoacan: vicinity of Morelia, Punguato, alt. 2100 m., 16 July, 1909, Arsene SOU (M, \JS); Arsene 52a (FM, US); Morelia, alt. 2000 m., 4 Aug. 1910, Arsene (US 1134412); vicinity of Morelia, north of Zapote, alt. 1950 m., 4 Aug. 1910, Arsene 5728 (M, G, US); vicinity of Morelia, Cuincho, alt. 1900 m., 1 July, 1909, Arsene 7803 (M, US). 25. Dyschoriste saltuensis Fernald, Proc. Am. Acad. 33: 92. 1898. A slender, erect, suffrutescent plant; stems branching, subtet- ragonal, densely covered with short appressed hairs, ciliate at the nodes; leaves lanceolate, obtuse at the tips, tapering below into a short petiole, the lower cauline, 6 cm. long, 1.5 cm. broad, the upper scarcely half as large, above covered with cystoliths, beneath strigilose-pubescent on the midrib; flowers axillary, solitary or in glomerules of 2 to 5, peduncles 3 or 4 mm. long; bracts minute; calyx hirsute, 8-10 mm. long, divided half way to the base into 5 lance-subulate lobes; corolla light purple, pubescent without, 15 mm. or less in length, the slender tube equalling the calyx and spreading into a campanulate throat; lobes oblong, truncate, 4 mm. long; filaments hirsute, style hirsute; mature capsule approximately 10 mm. long, glabrous; seeds 4, flat, oblique. Distribution: mouii tains of southwestern Mexico. Specimens examined: Guerrero: vicinity of Acapulco, Oct. 1894-March 1895, Palmer 506 (G type, M, FM, Ch, US). 26. Dyschoriste ciliata (Nees) O. Ktze. Rev. Gen. PI. 2: 486. 91; Lindau in Engl. Bot. Jahrb. 19 (Beibl. 48): 15. 1894. Calophanes ciliatus Nees in DC. Prodr. 11: 110. 1847. Ruellia ciliata Ruiz in DC. Prodr. 11: 110. 1847. Stem procumbent, glabrescent, with the apex and ascending 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 49 branches puberulent ; lower leaves more or less spathulate, 3 cm. long, 1 cm. broad, upper leaves ovate to elliptic, 6-7 cm. long, 2-3 cm. broad, obtuse to acute at the apex, entire, nearly glabrous, the base cuneate-decurrent into a petiole about 1.5 cm. long; flowers axillary, in glomerules, subsessile, subtended by oblong, ciliate bracts; calyx 5-parted, 11 mm. long, joined for more than one-half the total calyx-length, lobes subulate-setaceous, ciliate; corolla infundibuliform, a little longer than the calyx; anthers slightly bicalcarate at the base. Distribution: Peru. Specimens examined: Peru: near Huanuco, 1787, Ruiz (B type, M fragment and photograph, 927773). 27. Dyschoriste quadrangi laris (Oerst.) 0. Ktze. Rev. Gen. PI 2: 486. 1891 ; Greenm. in Proc. Am. Acad. 33: 487. 1898. Calophanes quadr angulariv Oerst. Vidensk. Meddel. 120. 1854 Mueller in Walpers, Ann. 5: 647. 1858; Hemsl. in Biol. Cent, Am. Bot. 2: 503. 1882. Stem erect, 8-10 dm. high, distinctly quadrangular, with cili ated wings; cystoliths especially at the swollen nodes which are quite distant; leaves ovate, oblong, 7-10 cm. long, 2-3 cm. broad acute at the apex, attenuata at the base into a petiole, repand crenulate; flowers verticillai;e, in cymose clusters at the nodes calyx 5-parted, 1 1 mm. long, subtended by short, subulate bracts tube glabrous, lineolate, than the lobes which are subulate-setaceous, canescent-pubescent along the nerves, ciliate ; corolla su bbilabiate, 1 1 mm. long, tube slightly shorter than the limb ; stameis adnate to the middle of the corolla- tube; anthers oblong with bisal appendages about 0.5 mm. long, filaments accrescent at point of attachment; capsule lanceolate, 8 mm. long, 4-seeded. Distribution: eastern Mexico. Specimens examined: San Luis Potosi: Los Canoas, 29 Aug. 1891, Pringle 5020 (G). Vera Cruz: Potrero de C msoquitla, Nov. 1841, Liebmann (G cotype) ; rocky soil, Zacuapan and vicinity, Oct. 1906, Purpus (M, G, FM). (Vol. 15 50 GARDEN 28. Dyschoriste quitensis (HBK.) 0. Ktze. Rev. Gen. PL 2: 486. 1891 ; Lindau in Bull. Herb. Boiss. 5 : 679. 1897. Calophanes quitensis (HBK.) Nees in DC. Prodr. 11: 110. 1847. Ruellia quitensis HBK. Nov. Gen. 2: 240. 1817; Kunth, Syn. 2: 37. 1837. Stem procumbent or ascending from a woody base, 3-4 dm. high, branched, somewhat quadrangular, puberulent; leaves oblong, elliptic-lanceolate, narrowed acutely at both ends, 25- 32 mm. long, 10-12 mm. broad, entire, puberulent, constricted below into a short petiole,: flowers axillary, subtended by lance- olate bracts which about equal the calyx in length; calyx 5- parted, 7-8 mm. long, united for about one-half the total calyx- length, nearly as long as the corolla, lobes subulate-setaceous, quite hirsute; corolla 8-9 mm. long, pubescent on the external surface; stamens and pistils typical of the genus; mature capsule 7-8 mm. long, acute at the apex, 4-seeded; seeds typical. Distribution : mountains of Andes, Ecuador. Specimens examined: Ecuador: near Quito, around Panecilli, alt. 3000 m., Humboldt (B type, M photograph); in Andes of Ecuador, 1857-59, Spruce 5989 (G). 29. Dyschoriste maranhonis (Nees) O. Ktze. Rev. Gen. PI. 2: 486. 1891; Lindau in Engl. & Prantl, Nat. Pflanzenfam. 4 sb : 302. 1895. Z ahlbrucknera maranhonis Pohl. in Mart. Fl. Bras. 9 : 26. 1847 ; DC. Prodr. 11: 108. 1847. Calophanes maranhonis Nees in Mart. Fl. Bras. 9: 26. 1847; Nees in DC. Prodr. 11: 108. 1847. Ruellia viscosa Pavon in Mart. Fl. Bras. 9: 26. 1847; DC. Prodr. 11: 108. 1847. Stem ascending or erect from a perennial base, 1-1.5 dm. high, pubescent, densely so at the apex, branched; leaves oblong- oblanceolate, 20-25 mm. long, 6-8 mm. broad, obtuse at the apex, tapering to a short-p etiolate base, crenulate-subrepand, lower leaves hirsute, densely so on veins of lower surface, upper leaves subtomentose ; flowers axillary, subtended by lanceolate bracts; calvx usuallv 5-narted. 7-8 mm. long, united for about one-half 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 51 the total calyx-length, pubescent, lobes subulate-setaceous, ciliate with long flaccid hairs; com. la approximately 14 mm. long, tube gradually ampliated into the throat, pubescent on the internal as well as the external surface, occasionally only 4-lobed; stamens occasionally incompletely didynamous, the anthers sagittate, appendaged at both ends, acutely appendaged at the base; ovary glabrous, style filiform, pubescent, stigma coiled; capsule linear, 9-10 mm. long; seeds 4. Distribution: southeastern Brazil. Specimens examined : Brazil: St. Ignacio, date hcking, Sellow (B type, M fragment and photograph). 30. Dyschoriste hirsuta (Oerst.) 0. Ktze. Rev. Gen. PL 2: 486. 1891. Calophanes hirsutus Oerst. in Kjoeb. Vidensk. Meddel. 71. 1877-78. Robust perennial, suffruticose at the base; stems erect, 4-6 dm. high, branched, at firs; pubescent, glabrate; leaves ovate, 15-20 mm. long, 9-12 mm. broad, subrepand, pubescent on both surfaces, petiolate, with the petiole 2-3 mm. long; flowers axillary, subtended by two oblanceolate bracts; calyx 5-parted, approxi- mately 10 mm. long, lobes united for one- third the total calyx- length, subulate-setaceous, pubescent on the nerves, ciliate; corolla 15-16 mm. long, paie violet, pubescent on the external surface, tube and throat approximately equal in length; stamens and ovary typical of the genus. Distribution: southeastern Brazil. Specimens examined: Brazil: on fields betAveen Serra da Piedade and Lagoa Santa, 2 May, 1864, Warming (C type, M fragment and photograph). 31. Dyschoriste hygrophilo ides (Nees) O. Ktze. Rev. Gen. PL 2: 486. 1891; Lindau in Ergl. & Prantl, Nat. Pflanzenfam. 4 3b : 302. 1895. Calophanes hygrophiloides Nees in Mart. Fl. Bras. 9: 26. 1847; Nees in DC. Prodr. 11: 109. 1847. Stems geniculate, ascending from a ligneous perennial base, [Vol. 15 52 ANNALS OF THE MISSOURI BOTANICAL GARDEN 3-4 dm. high, pubescent; leaves petiolate, 2-3.5 cm. long, 1-2 cm. wide, lower leaves obovate-subrotund, more or less emarginate at the apex, attenuate at the base into a rather short petiole, upper leaves elliptic-ovate, softly pubescent on both surfaces, margin somewhat sinuous ; inflorescence axillary, in glomerules of 2-5 flowers; bracts linear-oblanceolate, setaceous, pubescent, shorter than the calyx, resembling calyx-lobes, bracteoles present ; calyx 5-parted, total length approximately 13 mm., lobes 7-8 mm. long, extremely setaceous, villous-ciliate ; corolla somewhat bilabiate, puberulent on the external surface, 13-14 mm. long, tube and throat of about equal length; stamens adnate below the middle of the corolla limb, anther cells ovate; style filiform, quite pubescent, 4 mm. or less in length; stigma curved, mature capsule not seen. Distribution: southeastern Brazil. Specimens examined: Brazil: in grassy fields, Parana, 10 Oct. 1914, Dustn 15640 (G, M photograph). 32. Dyschoriste repens (Nees) 0. Ktze. Rev. Gen. PL 2: 486. 1891 ; Lindau in Engl. & Prantl, Nat. Pflanzenfam. 4 3b : 302. 1895. Calophanes repens Nees in DC. Prodr. 11: 109. 1847. Ruellia repens Ruiz ace. to Nees in DC. Prodr. 11: 109. 1847, in synonymy. Stem spreading, ascending, geniculate from a perennial base, densely pubescent, branching; branches rather short, ascending, densely foliate; leaves obovate (lower) to ovate (upper), 2.5-3.5 cm. long, approximately 1 cm. broad, obtuse to acute at the apex, tapering at the base to a distinct petiole, ciliate, entire, hirsute on the upper surface, pubescence of unequal hairs on the midrib of the under surface; flowers axillary, subtended by foliaceous bracts; calyx 5-parted, 10-11 mm. long, nearly equalling the corolla, lobes united for about one-half the total length of the calyx, subulate-setaceous, conspicuously ciliate; corolla 11-12 mm. long, pubescent on the external surface, tube comparatively broad, short; ovary 2-celled, style filiform, pubescent, 6 mm. long, stigma oblique; capsule not seen. Distribution: Peru. Specimens examined: 1928] KOBUSKI — MONC "GRAPH OF DYSCHOKISTE 53 Peru: "near Cheuchin," date lacking, Ruiz (B type, M frag- ■ ment and photograph). 33. Dyschoriste Pulegium <|Nees) O. Ktze. Rev. Gen. PL 2: 486. 1891. Calophanes Pulegium Nees in Mart. Fl. Bras. 9: 25. 1847; Nees in DC. Prodr. 11: 109. 1847. Stem erect from a suffruticose base, subveiutinous-pubescent ; leaves sessile, obovate, 2-2.5 cm. long, approximately 1 cm. broad, obtuse at the apex, tapering at the base, crenulate; flowers axillary in sessile glomerules; calyx 5-parted, 9-10 mm. long, hirsute, joined one-qiarter the distance from the base, lobes subulate-setaceous, ciliate; corolla about twice the total calyx length; stamens and pistils typical of the genus; mature capsule lanceolate, 4-seeded. Distribution: southeaster! i Brazil. Specimens examined: Brazil: date and exact locality lacking, Sellow 173 (B type, M photograph). 34. Dyschoriste ovata (Cav.) 0. Ktze. Rev. Gen. PI. 2: 486. 1891; Lindau in Engl. & Prantl, Nat. Pflanzenfam. 4 3b : 302. 1895; Lindau in Bull. Herb. Boiss. 5: 678. 1897. Calophanes ovatus Benth. in DC. Prodr. 11: 108. 1847, as to Cavanilles plant (not Hartweg plant) ; Hemsl. in Biol. Cent.-Am. Bot. 2: 502. 1882. Ruellia ovata Cav. Ic. 3: 28, pi 25^. 1794; Willd. Sp. PI. 3: 363. 1801. Erect perennial, 4-6 dm. high; stems quite quadrangular, pubescent, sometimes winged, occasionally branched; leaves ovate, obovate or occasionally elliptical, 3-4 cm. long, 1-1.7 cm. broad, obtuse at the apex, attenuated at the base into a very short petiole, quite glabrous with the exception of occasional stiff hairs on the midrib and veins, cystoliths abundant, giving the appearance of appressed hairs, margin entire and ciliated, some- times slightly repand-crenulate ; flowers crowded among the foliaceous bracts at the nod3s, giving a glomerulate appearance; calyx 5-parted, quite glabrDus, covered with cystoliths, 14-15 54 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 mm. long, joined for about one-third the total calyx-length, lobes subulate-setaceous, ciliate; corolla 20-25 mm. long, tube less than 10 mm. long, puberulent on the external surface; stamens, pistil, and fruit typical of the genus. Distribution: southern Mexico. Specimens examined: Vera Cruz: Nogales, Mt. Orizaba, alt. 1400 m., 16 Aug. 1891, Seaton 892 (Ch, G); Borrego, near Mt. Orizaba, 26 Aug. 1866, Bourgeau 2908 (US, G). Morelos: near Cuernavaca, 30 July, 1906, Pringle 18838 (US); near Cuernavaca, alt. 1500 m., 28 July, 1896, Pringle 7249 (US). Michoacan: Morelia, alt. 2100 m., 8 Aug. 1912, Arsene 9027 (US). 35. Dyschoriste amoena (Nees) O. Ktze. Rev. Gen. PL 2: 485. 1891; Lindau in Engl. & Prantl, Nat. Pflanzenfam. 4 3b : 302. 1895. Calophanes amoenus Nees in Mart. Fl. Bras. 9: 27. 1847; Nees in DC. Prodr. 11: 110. 1847. Stem ascending from a perennial base, approximately 3 dm. high, geniculate, branched, puberulent toward the apex, other- wise glabrous; leaves narrowly oblong-ovate to slightly oblong- oblanceolate, 4-5 cm. long, 1-1.5 cm. broad, obtuse to acute at the apex, entire, glabrous, attenuate at the base into a very short petiole; inflorescence axillary, bracteate, glomerulate toward the apex; calyx deeply 5-parted, 2 cm. long, quite robust, pubescent, lobes less attenuate than in the majority of species, ciliate; corolla 5 mm. or more longer than the calyx, pubescent on the external surface; stamens and ovary typical of the genus; mature capsule not seen. Distribution: southeastern Brazil. Specimens examined: Brazil: date and exact locality lacking, Sellow (B type, M photograph and fragment). 36. Dyschoriste xylopoda 1 n. sp. PL 15. 1 Dyschoriste xylopoda Kob., sp. nov., caulibus erectis vel ascendentibus, basi crasse lignose, 2-3 dm. altis, villoso-pubescentibus; foliis sessilibus, lanceolato- oblongis, inferioribus rare ovatis, 2.5-3.5 cm. longis, 1 cm. minusve latis, integerrimis; 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 55 Stems strict, rising from a thick woody base to a height of 2-3 dm., villous-pubescent throughout; leaves sessile or nearly so, lanceolate-oblong, or the lowermost occasionally ovate, 2.5-3.5 cm. long, approximately 1 cm. or less broad, entire; flowers 2-3 on a short peduncle in the axils of the leaves, subtended by linear- lanceolate bracts which are a little shorter than the calyx; calyx deeply 5-cleft, 17-18 mm. .ong, lobes 11-12 mm. long, subulate- setaceous, villous-ciliate ; CDrolla 25-27 mm. long, pubescent on the external surface, tube 10 mm. long; stamens adnate below the middle of the corolla limb, anthers oblong-ovate, 2-3 mm. long; ovary 2-celled, glabrous, style filiform, 20 mm. long, pubes- ;ma linear, oblique, nearly 2 mm. long; mature fruit seen. Distribution: southern Mexico. Specimens examined: Jalisco: hills near Guadalajara, 19 July, 1893, Pringle 4442 (M type, G, FM). 37. Dyschoriste oblongifolia (Michx.) O. Ktze. Rev. Gen. PI. 2: 486. 1891; Lindau in Engl. & Prantl, Nat. Pflanzenfam. 4 3b : 302. 1895. PI Ruellia oblongifolia Michx. Fl. Bor.-Am. 2: 23. 1805; Pursh, Fl. Am. Sept. 2: 420. 1814. ? Ruellia biflora L. Sp. PI. 2: 635. 1753 (a doubtful synonym — refer to D. Don in Sweet's Brit. Fl. Garden). Calophanes oblongifolia (Michx.) D. Don in Sweet, Brit. Fl. Gard. 2: pi 181. 1833; Gray, Syn. Fl. N. Am. ed 1, 2 1 : 324. 1878, and ed. 2, 1886; Chapman, Fl. Southeastern U.S. ed. 3, 365. 1897; Britt. & Brown, 111. Fl. 3: 202. 1898; Small, Fl. Southeastern U.S. ed. 1, 1083. 1903, and ed. 2, 1913. Dipter acanthus biflorus Nees in Linnaea 16: 294. 1842. Dipter acanthus oblongifotius Chapman, Fl. Southeastern U.S. ed. 2, 303. 1889. floribus 2-3, pedicellatis, axillaribi s; bracteis linearo-lanceolatis, calyce haud paulo breviore; calyce profunde 5-diviso, 17-18 mm. lcmgo, lobis 11-12 mm. longis; subu- lato-setaceis, villoso-ciliatis; corolla 25-27 mm. longa, extus puberulenta, tubo 10 mm. longo; filamentis basi connatis, antheris ovato-oblongis, 2-3 mm. longis; stylo 20 mm. longo, pubescente, stigmata lineari, obliqua, prope 2 cm. longa; capsula ignota. — Type collected on hills, near Guadalajara, Jalisco, Mexico, 19 July 1893, C. G. Pringle 444% (M). 56 ANNALS OF THE MISSOURI BOTANICAL GARDEN (Vol. 15 i Herbaceous perennial; stem quadrangular, branched at the base, erect, pubescent or softly hirsute, 4-8 dm. high; leaves short-petiolate, oblong 5 cm. broad, rounded or obtuse at the apex, narrowed at the base, entire or slightly crenulate, softly hirsute; flowers solitary, axillary on very short pedicels, subtended by narrowly oblong, leafy bracts; calyx 15-18 mm. long, deeply 5-parted, subulate-setace- ous, lobes ciliate; corolla blue, usually purple-maculate in the throat, approximately 25-27 mm. long, the tube shorter than the abruptly ampliated throat, puberulent on the external surface, lobes rounded ; filaments slightly pubescent at point of adnation ; anther cells oblong; mature capsule 10-12 mm. long, lanceolate; seeds 4, flattened, oblique. Distribution: sandy pine barrens, southern Virginia to Florida. Specimens examined: Virginia: date and locality lacking, probably southeastern portion of the state, Thurber (G). South Carolina: sandy ground, north of Graniteville, Aiken Co., 21 May, 1899, Eggert (M); locality lacking, May, 1867, Ravenel (G, US) ; barrens near Beaufort, Beaufort Co., 26 April, 1917, Churchill 743 (M). Georgia : sand hills between Grovetown and Forrest, Columbia Co., 10 June, 1902, Harper 1812 (M, G, US); low places north of Belair, Richmond Co., 22 May, 1899, Eggert (M); Ocmulgee River swamp below Macon, Laurens Co., Small (FM) ; Savannah, Chatham Co., 1842, Curtis (M); Darien Junction, Mcintosh Co., 31 May, 1909, H. H. Smith 2219 (FM); Darien Junction, Mcin- tosh Co., alt. sea level, 25-! Oglethorpe Co., 1836, Short (M). Florida: sandy pine lands, date and exact locality lacking, Mohr (US 721388); eastern Florida, 1895, Curtiss (US); pine barrens, Duval Co., 21 April, 1902, Fredholm 5101 (G, US); pine barrens, Duval Co., 28 April, 1902, Fredholm 5127 (G); near Jacksonville, Duval Co., 2 May, 1893, Curtiss 4428 (M, US) ; dry pine barrens, near Jacksonville, Duval Co., 8 May, 1894, Curtiss 4667 (FM, US); Jacksonville, Duval Co., April, 1869, Canby (M, G, US); dry pine barrens, near Jacksonville, Duval Co., Curtiss 1938 (M, G, US); south Jacksonville, Duval Co., 7 April, 7 June, 1895, Small (FM) 1928] KOBUSKI — MONOGRAPH OF DYSCHORISTE 57 1897, Churchill (G); dry sandy pine barrens, St. Augustine, St. Johns Co., May-Aug. 1875, Reynolds (M); sandy pine barrens, DeLand, Volusia Co., date lacking, Harkness (M); dry pine woods, DeLand, Volusia Co., 7 May, 1910, Hood (M); vicinity of Eustis, Lake Co., June-July, 1894, Hitchcock 1454 (FM, M); Eustis, Lake Co., 26 Apiil, 1896, Webber 520 (M); sandy soil, high pine lands, vicinity of Eustis, Lake Co., Nash 184 (M, G, US); Eustis, Lake Co., 28 May-15 June, 1895, Nash 1774 (US); Winter Park, Orange Co., March, 1900, Huger 19 (M); Clarcona, Orange Co., 18-22 Aug. 1399, Pieters 121 (US); dry sand, Okee- chobee region, Brevard Co., 2 June, 1903, Fredholm 5870 (G); Lake City, Jefferson Co., June- July, 1898, Hitchcock 1452, 1453 (FM) ; Rosewood, Levy Co., June, 1876, Garber (FM); dry sandy ground, Polk Co., 12 April, 1894, Ohlinger 1415 (FM); dry land, Polk Co., 11 June, 1894, Ohlinger 188, 1487 (FM); Lake Alfred, Polk Co., 11 June, 1922, Armstrong & Armstrong (M); Polk Co., March, 1890, Milligan (US) ; pine barrens, Tampa, Hillsborough Co., Aug. 1898, Ferguson M); in pine lands near St. Petersburg, Pinellas Co., 10 Nov. 1907, Deam2832 (G); pine woods, Manatee Co., 16 March, 1887, Roth'ock (FM 160054, 322461) ; sandy field, Bradentown, Manatee Co., 15 May, 1900, Tracy 6683 (M); Fort Myers, Lee Co., 1904, Weatgate 8607 (FM) ; Fort Myers, Lee Co., July-Aug. 1900, Hitchcoci: (FM); Aspalaga, Liberty Co., May, 1898, Chapman (M). Alabama: date and locality lacking, Buckley (M, US). 37a. Dyschoriste oblongifolia (Michx.) O. Ktze. forma glabra n. f. Stem and leaves glabrous; otherwise as the species. Distribution: Florida. Specimens examined: Florida: Tocoi, St. Johns Co., 1874, Palmer 846 (M, G); Lake City, Columbia Co., 4 May, 1893, Rolf 190 (M, FM); Gaines- ville, Alachua Co., 5 June, 1910, Hood (M); Fort Myers, Lee Co., July-Aug. 1900, Hitchcock (M); cypress swamp and low pine land, vicinity of Fort Myere, Lee Co., 8 May, 1916, J. P. Standley 179 (M, G, FM, US) ; in pine land, Mullock Creek District, about 8 miles southeast of Fort Myers, Lee Co., May-June, 1917, J. P. Standley 443 (M, G, FM, US); in pine land, vicinity of Fort [Vol. 16 58 ANNALS OF THE MISSOURI BOTANICAL GARDEN Myers, Lee Co., 12 May, 1910, /. P. Standby IS (M, G, FM, US) ; pine woods, vicinity of Fort Myers, Lee Co., 28 Feb. 1916, ./. P. Standley 12852 (US); vicinity of Fort Myers, Lee Co., 29 Feb. 1916, J. P. Standley 12917 (US); moist pine lands, vicinity of Fort Myers, Lee Co., 14 Dec. 1919, P. C. Standley 18894 (US); high pine land, Jessamine, 17-20 April, 1899, Barnhart 2680 (FM). 38. Dyschoriste humilis (Griseb.) Lindau in Engl. Bot. Jahrb. 19(Beibl. 48): 15. 1894; Lindau in Bull. Herb. Boiss. II. 3: 628. 1903. Ruellia geminiflora Kth. var. humilis Griseb. in PL Lor. 176. 1874, and Symb. Argent. 259. 1879. Stem slender, branched below, ascending from a ligneous, perennial base, pubescent; leaves oblong-elliptic, 2-3.5 cm. long, 0.4-0.5 cm. broad, tapering acutely both at the apex and at the short-petiolate base, puberulent, ciliate, entire or sinuous ; flowers in twos or threes, axillary, subtended by foliaceous bracts about 8-10 mm. long; calyx in an thesis approximately 14 mm. long, puberulent, lobes setaceous, 8 mm. long, at maturity the calyx sometimes attains a length of 20 mm.; corolla 21-22 mm. long, puberulent on the external surface, tube shorter than the broadly ampliated (8-9 mm. in diameter) throat; stamens adnate below the middle of the corolla throat, anther cells oblong, a little over 2 mm. long, cells of individual anthers often differing, one base distinctly acute and minutely apiculated, the other base blunt or slightly mucronate, apex of cells slightly acute; style sparsely pubescent, filiform, 19 mm. long, stigma about 2 mm. long, posterior lobe rudimentary; mature capsule exceeding the calyx lobes, 10 mm. long; retinaculum in center of each cell; seeds two, flat, oblique. Distribution: Argentina. Specimens examined: Argentina: Chaco Santafichna, Mocovi, 5 Nov. 1903, Venturi 65 (US); Cordoba, Dec. 1891, Kuntze (US 701502); Cordoba, 21 Dec. 1876, Hieronymus (FM 51116, US 282198); near the city of Cordoba, 1874-75, Hieronymus (FM 51115a, US 282197). 1928 KOBUSKT — MONOGRAPH OF DYSCHORISTE 59 39. Dyschoriste paraguariensis 1 n. sp. PL 16. Stems erect, 2-3 dm. high, strict, somewhat quadrangular, nearly glabrous, sparsely pubescent at the nodes; leaves sessile, lanceolate-elliptic, 2-2.5 cm. long, 0.5-0.8 cm. wide, acute at the apex, glabrous, covered w ith an irregular scattering of cystoliths, margins entire, not ciliated; flowers in twos, axillary, subtended by two foliaceous bracts which resemble the leaves in nearly every respect; calyx glabrous, except for the ciliation on the lobes, covered with a regular array of cystoliths, 14 mm. long at maturity, lobes lanceolate, setaceous, 10 mm. long; corolla approximately 18 mm. lcng, puberulent on the external surface, tube about 9 mm. long; style filiform, about 12 mm. long, pubes- cent, stigma oblique; miture capsule linear, 11-12 mm. long, glabrous, 4-seeded; seeds flattened. Distribution : Paraguay. Specimens examined: Paraguay: in region of the river "Tapiraguay," Aug., Hassler 4355 (G, type). 40. Dyschoriste Tweediana (Nees) O. Ktze. Rev. Gen. PI. 2: 486. 1891. Calophanes Tweedianut Nees in DC. Prodr. 11: 108. 1847. Stem ascending from a perennial base, 4-6 dm. high, pubescent; leaves ovate-elliptic, 3-3.5 cm. long, 0.9-1.5 cm. broad, acute to obtuse at the apex, taper: ng at the base into a very short petiole, repand-subcrenate, glabrous; flowers axillary, 1-3 aggregated on very short peduncles in the axils, subtended by oblong-lanceolate bracts which are shorter than the calyx; calyx deeply 5-parted, pubescent, 14 mm. long, lobes subulate-setaceous, ciliate; corolla infundibuliform, 5-lobed, 20 mm. long, tube 8 mm. long, pu- bescent on the externa] surface, lobes ovate, obtuse; anthers 1 Dyschoriste paraguariensis Kob. sp. nov., caulibus erectis, 2-3 dm. altis, strictis plus minusve quadrangularis, uubglabris, parum pubescentibus ad nodos; foliis sessilibus, lanceolato-ellipticis, 2-2.5 cm. longis, 0.5-0.8 cm. latis, apice acutis, glabris, lineolatis; floribus duobus in axillaribus, bracteis duobus, foliaceis; calyce glabro, cystolithero, 14 mm. lonj;o, lobis lanceolatis, setaceis, 10 mm. longis; corolla violacea, 18 mm. longa, extus puberula, tubo 9 mm. longo; filamentis basi connatis; stylo 12 mm. longo, piloso; capsula lineari, 11-12 mm. longa, glabra, 4-sperma; seminibus planis. — Type collecte i in the region of the river "Tapiraguay," Paraguay, Aug., Hassler 4355 (G). 60 ANNALS OF THE MISSOURI BOTANICAL GAHDEN [Vol. 15 appendaged at the base, appendages connivent (according to Nees); seeds suborbicular, flattened. Distribution: southeastern Brazil. Specimens examined : Brazil: in dry mountain forests in Prov. Bonar, at river Jacuhy in Rio Grande do Sul, date lacking, Tweedie 771 (K type, M photograph) . Excluded Species Calophanes calif ornica Rose ace. to Vasey & Rose in Contr. U.S. Nat. Herb. 1 : 85. 1890 = Ruellia calif ornica (Rose) Johnston in Proc. Cal. Acad. Sci. IV, 12: 1171. 1924. Calophanes cubensis A. Rich, in Sagra, Hist, de Cuba 11: 160. 1850 = Hygrophila brasiliensis (Spreng.) Lindau in Urb. Symb. Ant. 2: 183. 1900. Calophanes Palmen Gray ace. to Watson in Proc. Am. Acad. 22: 443. 1887 - Spigelia scabrella Benth. PI. Hartweg. 45. 1840. Calophanes peninsularis Rose ace. to Vasey & Rose in Contr. U. S. Nat. Herb. 1: 75. 1390 = Ruellia peninsularis (Rose) Johnston in Proc. Cal. Acad. Sci. IV, 12: 1172. 1924. Dyschoriste Candida Brandegee in Zoe 5: 242. 1908 = Ruellia Candida (Brandegee) Kobuski, n. comb. Dyschoriste cubensis Urb. Symb. Ant. 7: 381. 1912 = Apas- salus cubensis (Urb.) Kobuski in Ann. Mo. Bot. Gard. 15 : 2. 1928. Dyschoriste diffusa Urb. Symb. Ant. 7: 380. 1912 = Apas- salus diffusus (Urb.) Kobuski in Ann. Mo. Bot. Gard. 15: 1. 1928. Dyschoriste humistrata (Shuttl.) O. Ktze. Rev. Gen. PI. 2: 486. 1891 = Apassalus humistratus (Shuttl.) Kobuski in Ann. Mo. Bot. Gard. 15: 3. 1928. List of Exsiccatae The distribution numbers are printed in italics. Unnumbered collections are indicated by a dash. The number in parentheses is the species number used in this monograph. Armstrong, Dr. & Mrs. G. M. — (37). Blumer, J. C. 2223 (15). Arsene, Bro. 6411 (9); 6472 (15); Bourgeau, M. 1262 (1); 2903 (34). 35, 52a, 1075, 1528, 1933, SOU, 5728, Britton, N. L. 155 (4). 7308 (24); 9027 (34). Buckley, S. B. — (13); — (37). Barnhart, J. H. 2680 (37a). Canby, W. M. — (37). Bebb, R. 1848 (13). Chapman, G. W. — (37). 1928) KOBUSKI — MONOGRAPH OF DYSCHORISTE 61 Churchill, J. R. 743 Mohr, C. (37). Clemens, Mr. & Mrs. J. 1069 (1 J). Conzatti, C. 1581 (1); 1858 (19). Conzatti, C. & Gonzalez, V. 281 (17). Croft, Miss M. B. 6465, 6660 (13). Nash, G. V. 184, 177 A (37). Nealley, G. C. {857) (16). (13); 580 (15); 184 Nee, D. L. Curtis, M. (37). (24). W. 4597 Curtiss, A. H. 5858 (4); 1941, 9U6 (13); Nicolas, Bro. 299 (24). -, 19S8, 4488, 4667 (37). Deam, Mrs. C. C. 8832 (37). Drummond, T. 178, 856 (13). Drushel, J. A. 4136 (13). Duges, A. (15). Dus6n, P. 15640 (31). Eaton, A. A. 386, 1847 (4). Eggert, H. 1 (13); (37). Ehrenberg, K. 1823 (1). Ferguson, A. M. (37). Fiebrig, K. 4856 (2); 6383 (21). Fredholm, A. 5101, 5127, 5870 (J7). Gandoger, M. Garber, A. P. (1). (4); (37). Gardner, J. 3951 (3). Ghiesbreght, A. (12). Gonzalez, V. 43 (1). Groth, H. A. 230 (13). Hall, E. 488, 431 (13). Hanson, H. C. 638 (15). Harkness, W. E. — (37). Harper, R. M. 1312 (37). Hassler, E. 5908, 7780 (2); 4355 (39). Havard, V. (13). Heller, A. A. 1529 (13). Hieronymus, G. (38). Hitchcock, A. S. 1455, 1456 (4); (37u). (37a). 1452, 1453, 1454 (37) Hood, S. C. — (37); — Huger, A. M. 19 (37). Humboldt, A. — (28). Jermy, G. 31, 62, 249, 472 (13). Jones, M. E. 374 (15). (38). Kuntze, O. Lemmon, J. G. Liebmann, F. M. (15). (27). Niederlein, G. 48 (8). Oersted, A. S. 808 (17). 1437 Orcutt, C. R. 6230 (13). 492 (7); 2033, 9667 (13); 125, 208, 276, 809, 472, 980, 1009, 1010, 1075, 2082 (15); 606 (25); 346 (37a). Palmer, E. J. — , 9708, 9811, 11871, 12229, 13698, 13758, 14252a (13). Parry, C. C. & Palmer, Ed. 699 (15). Patzky (15). Pieters, A. J. 121 (37). Pringle, C. G. 2154, 2989, 6053 (1); 6481 (11); 66 (15); 6718 (17); 2907, 11318 (18); 6712 (19); 4134 (20); 11322 (24); 5020 (27); 7249, 13838 (34); 4448 (36). Purpus, C. A. 2362, 8347 (5); 2263 (27). W (37). Reverchon, J. — , 722, 1941, 2114 (13). Reynolds, Miss M. C. Riedel, L. 501 (6) ; 45 Rolf, P. H. 190 (37a). (37). 8495 3615 (15). Rose, J. N. & Rose, J. R. 11148 (15). Rose, J. N., Standley, P. C. & Russell, P. G. 12833, 14298 (1). Rothrock, J. T. — , 687 (15); (37). Ruiz, L. (26); (32). Ruth, A. SO, 81, 104 (13). Sartorius, C. (9). Schaffner, J. G. 854 (647) (15). Schiede, C. J. W. 122 (9). Schulz, Miss E. D. 146 (13). Lindheimer, F. Ill, 325, 504, 558, 677, Seaton, H. E. 398 (34). 1063 (13). Lloyd, F. E. 199 (23). Maxwell, C. F. 49 (13). Mearns, E. A. 1438 (13). Secmann, B. 1518 (14). Seler, E. 1733 (1). Sello (Sellow), F. — (29); — , 173 (33); (35). Milligan, Mrs. J. M. (37). W (37). 62 [Vol. 16 THE MISSOURI BOTANICAL GARDEN Shreve, F. 4978 (15). Simpson, J. H. 628 (4). Slater, Mrs. H. D. — (13). Small, J. K. (37). Small, J. K. & Carter, J. J. 776 (4). Smith, H. H. 2219 (37). Tracy, S. M. 507.9 (13); 6683 (37). Trelease, W. 852 (15). Tweedie, J. 771 (40). Tweedy, F. 180 (13). Venturi, S. 55 (38). Warming, E. (30). Smith, Rev. L. C. 823, 729 (1); 419 (19). Webber, Ft. J. 520 (37). Spruce, R. 6989 (28). Westgate, J. M. 3607 (37). Standley, Miss J. P. 18, 179, 448, 12852, Wilcox, T. E. — 160 (15). 12917 (37a). Wislizenus, A. 185 (15). Standley, P. C. 40624 (13); 18894 (37a). Wright, C. 482 (13); 1462, 1468 (15) Tharp, B. C. 788 (13). Thurber, G. 974 (15); (37). Young, J. P. 801 (4). Index to Species synonyms in italics, and previously published names in ordinary type. type Apassalus 21 cubensis 60 difTusus 60 humistratus 60 Calophanes 26 amoenus 54 angustxfolius 30 angusta 31 bilabiatus 3! I bilobatus 28 calif ornica 60 capitatus 42 ciliatus 48 crinitus 30 cubensis A. Rich 60 dccumbcns 39 hirsuhssimus 28 hirsutus 51 hygrophiloides 51 J asminum-mexicanum 47 lavandulaceus 32 linearis 36 maranhonis 50 microphyllus 47 oblong/folia D. Don 55 oblongifolia var. angusta 31 oblongifolia var. texensis 37 oblongifolius Torr 40 ovatus Benth 53 ovatus (Cav ) Nees 37 Palmeri 60 pi ninsularis 60 Pulcgium 53 quadrangular is 49 quitcnsis 50 .... 52 ... 34 . * . . «5tc repens Schiedeanus Schiedeanus var. multiflorus Serpyllum 46 Tweedianus 59 Dipler acanthus 55 bifiorus 55 linearis 36 oblongifolius 55 Dyschoriste 25 amoena 54 angusta 31 angustifolia 36 bilabiata 39 Candida 60 capitata 42 ciliata 48 crenulata 41 crinita 30 cubensis 60 decumbens 39 depressa f 12 diffusa 60 erecta 12, 26 Greenmanii 33 hirsuta 51 hirsutissima 28 19281 KOBUSKI — MONOGRAPH OF DYSCHORISTE 63 humilis 58 humistrata 60 hygrophiloides 51 jaliscensis 35 Ja8minum 47 lavandulacea 32 linearis • 36 Lloydii 46 maranhonis Lindau 30 maranhonis Nees 50 microphylla 47 Niederleinii 34 oaxacensis 43 oblongif olia 55 oblongifolia f . glabra 57 ovata 53 paraguariensis 59 pinetorum 44 Pringlei 42 Pulegium 53 Purpusii 32 quadrangularis 49 quitensis 50 repens 52 Rosei 35 sagittata 45 saltuensis 48 Schiedeana 34 Schottiana 30 Serpyllum 46 trichanthera 29 Tweediana 59 xylopoda 54 Hy grophila 60 brasiliensis 60 Schottiana 30 Linoslylis 26 Ruellia 10 biflora 55 caJifomica 60 cane ida 60 cilia !a . . . 48 depr 388a 10 gemmifl&ra var. humilis 58 microphylla 47 oblongifolia 10, 55 ovati \ Benth 37 ovatii Cav 53 peniasularis 60 quitensis 50 repens 52 viscosa 50 Spigelia 60 scabrella 60 Zohlbrucknera 50 maranhonis 50 [Vol. 15, 1928] 64 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 3 Fig. 1. Flower of D. oblongifolia (Michx.) O. Ktze. X 3. Fig. 2. Open corolla of D. oblongifolia (Michx.) O. Ktze. X 3. Showing character and position of stamens. Fig. 3. Open calyx of D. oblongifolia Open calyx of D. oblongifolia (Michx.) O. Ktze. X 3. Fig. 4. Pistil of D. oblongifolia (Michx.) O. Ktze. X 3. Fig. 5. Dehiscing capsule of position of retinacula and seeds. long X 3. Showing Fig. 6. X 3. alyx oblong Ann. Mo. Hot. Gahd., Vol. 15, 1928 Plate :} 4 KOBUSKT— MONOGRAPH OF DYSCHORISTE [Vol. 15, 1928] 66 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 4 Dyschoriste trichanthera Kobuski From the type specimen, Hasslrr 7780, in the Gray Herbarium of Harvard University. Ann. Mo. Bot. Gard., Vol. 15, 1928 Plate 4 ■*■ lit A) AJb&& XrU* -IL- L^J^ri.,^ thri- ft *£ * . // ' ' < *A *A* I /,-. KOBUSKI— MONOGRAPH OF DYSCHORISTE [Vol. 15, 1928J 68 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 5 Dyschoriste Purpusii Kobuski From the type specimen, Pur pus 2362, in the herbarium of the Missouri Botanical Garden. Ann. Mo. Hot. Gaud., Vol. 15, 1928 Plate 5 St H -j ■ *r rl«nv .. ■■ s- ***■* < ( - * . A » , ) ' .»<< » . I < J -.A* - - I 2. , 4 ..<.«, W *" v >" ; J ( ■ r . . * KOIU'SKI MONOCHAI'H OF DYSCIK >K ISTK [Vol. 15, 1928 70 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 6 Dyschorisie Greenmanii Kobuski From the type specimen, Palmer 49%, in the United States National Herbarium Ann. Mo. Bot. Gard., Vol. 15, 1928 Plate a f| %* \m MUM >/M v ; *t *H* b)t,f4fl*% V " * X - 4 t f *£ \ :<• t KOIU'SKI— MONCKJRAPH OF DYSCHORISTE [Vol. 15, 1928 72 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate plate 7 Kig. 1. Dyschoriste Rosei Kobuski From the type specimen, Rose 2259, in the United States National Herbarium Fig. 2. Dyschoriste crenvlata Kobuski From the type specimen, Nelson 6604, in the Gray Herbarium of Harvard Univer- sity. \x\. Mo. Hot. Gabd.. Vol. 15, 1928 Plate 7 c td '< c c 5 o 32 H [Vol. 15, 1928| 74 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 8 Dyschorisle jalisccnsis Kobuski From the type specimen, Pringle 5481, in the Gray Herbarium of Harvard Univer- sity Aw. Mo. Hot. Card.. Vol. 15, 1928 Plate s c HUKOLK. PLANTS MEXICANS 4 I tun :t ;al»» ■ + v t-i ' / 7 •a 5*¥*7 ]byvU+u*Xk,. i+M***^ fcU-, %^ KOBUSKI— MONOGRAPH OF DYSCHORISTE [Vol. 15, 1928] 76 ANNALS 01 THE MISSOURI BOTANICAL GARDEN ] Explanation of Plate plate 9 Di/schoristc bilabiata (Seem.) O. Ktze. From the type specimen, Seeman 1513, in the Royal Botanic Gardens at Kew, England. Ann. Mo. Bot. Gard., Vol. 15, 1928 Plate 9 ^ / jn^* \*» tf ;r*3 e 9293B3 I £Wti**** titJux+ %0+J c e4^ K ^y * t »r .. i * <>*v\» * i ft ' ■ *r > ^ /r/3 9 •tf. . KOHTSKl MOXOCRAPH OF DYSUK >K ISTE [Vol. 15, 1928] 78 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 10 Dyschoriste oaxacensis Kobuski From the type specimen, Pringle 6712, in the herbarium of Ihe Missouri Botanical Garden. Ann. Mo. Hot. Gard., Vol. 15, 1028 Plate 10 u « ■ a ^* I tl 1 ■■A * ft 6712 Dwchoriste Ji im-WexH um. ** |wm-nvnv HJj QWPw W^ > ( f KOHCSKI— MONOdUAPII OF DYSCIK MUST]] [Vol. 15, 1928 80 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 11 Dyschoriste pinetorum Kobuski From the type specimen, Pringh 4134, in the Gray Herbarium of Harvard Uni- Tersity. Aw. Mo. Hot. Gard., Vol. 15, 1928 Plate 1 1 KOBl'SKI— MONOGRAPH OF DYSCHORISTE [Vol. 15, 19281 82 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 12 Dyschoriste sagittata Kobuski versity. type specimen, Fiebrig Ann. Mo. Hot. Gard., Vol. 15, 1928 Platk 12 i ?d M0«i*\ff • ■ , '< KjauJLk i •&• A i+iytU* r t N .-. I KOBUSKI— MONOGRAPH OF DYSCHOHISTE [Vol. 15, 1928| 86 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 14 Dyschoriste microphylla (Cav.) O. Ktze. Madrid Botanical Ann. Mo. Bot. Gard., Vol. 15, 1928 Plate 14 r "<*%,., / %/» / c\ A m n uui h LAA~ ** ' - KO*-j«,r >. ^ /C /(L. \ KOBUSKI— MONOGRAPH OF DYSUIOKISTK [Vol. 15, 1928) 88 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 15 Dyschoriste xylopoda Kobuski Missouri Garden, Ann. Mo. Hot. Gard., Vol. 15, 1928 Plate 15 'AW* i h - £ %r t W-/*AVpVrti.M- $* r % - "i C 9 I KOUE, 4: t ItAfl J ft 1411 < i*l^i*l8isiM -, tttgiltaititt, «CrM ■ .. *,> I ^ . Si ,: * ■ ■ ■ ■ mwVAWrt.TOuVAr^.o>*^ KoM'SKl -MONOCHAIMI OK DYSCHORISTE [Vol. 15, 1928] 90 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 16 Dyschoriste paraguariensis Kobuski From the type specimen, Hassler 4355, in the Gray Herbarium of Harvard Uni- versity. 4 > Ann. Mo. Hot. Gard., Vol. 15, 1928 Plate lt> 1 V ...**«. rammi *. I v .V. v , , ■ v I r 1 I V ' A^M4*A-# | U. ■ * v» * * I y ROBI'SKI— MONOGRAPH OF DYSOIIORISTE STUDIES IN THE UMBELLIFERAE . I 1 MILDRED E. MATHIAS Jessie R. Barr Research Fellow in the Henry Shaw School of Botany of Washington University A critical study of the genus Cymopterus has necessitated a detailed investigation of about twenty allied genera including Glehnia Schmidt. Asa Gray in 1859 doubtfully referred a plant from the Cooper collection in the region of Puget Sound, Washington, to the genus Cymopterus, and in 1860 published the species as Cymopterus ? littoralis. Bentham ('67) in Bentham and Hooker's 'Genera Plantarum,' published in September, 1867, transferred this species to his new genus Phellopterus. F. Schmidt ('67), some time between January and July inclusive of the year 1867, in his 'Prolusio Florae Japonicae' published the new genus Glehnia 2 with one species, G. littoralis, basing it on a Maximowicz plant from Hakodate. In this work Schmidt also included Cymopterus ? littoralis Gray as a distinct species. In 1868 in his 'Flora Sachalinensis ' he recognized that his Glehnia littoralis was con- specific with Cymopterus ? littoralis Gray and adopted the generic name Phellopterus of Bentham. Upon critical examination of the Cooper and Maximowicz types Schmidt's view as to the congeneric nature of the two is confirmed. As the generic name Glehnia of Schmidt was published at least two months prior to the Phellopterus of Bentham, on the basis of priority, it must be retained as the correct name for the genus and the Maximowicz plant must be taken as the generic type. The historical type of the genus is then the plant collected by Maximowicz in Hakodate, Japan, in 1861 and must bear the specific name Glehnia littoralis Schmidt. Bentham, G. ( 9 67)). In Bentham, G. f and J. D. Hooker, Genera Plantarum 1: 905. September, 1867. Gray, Asa ('59). "Botany of Japan/' Mem. Am. Acad. N.S. 6 2 : 391, 428. 1859. , C60). Stevens' Report of U.S. Explorations & Surveys from the Mis- River 1860. 1 Issued April 30, 1928. 2 The genus Glehnia was so named in honor of Peter von Glehn who collected with Schmidt on the Island of Sachalin. Ann. Mo. Bot. Gard., Vol. 15, 192* (91) (Vol. 15 92 ANNALS OF THE MISSOURI BOTANICAL GARDEN Schmidt, F. ('67). Prolusio Florae Japonicae in Miq. Ann. Mus. Bot. Lugd. Bat. 3: 61. January- July, 1867; Prolusio Florae Japonicae, 249. 1867. , ('68). Flora Sachalinensis. Mem. Acad. Imp. Sci. St. Petersbourg, VII, 12*: 138-140. 1868. Glehnia Schmidt, Prol. Fl. Jap. in Miq. Ann. Mus. Bot. Lugd. Bat. 3: 61. Jan.-July, 1867; Prol. Fl. Jap. 249. 1867; Baillon, Hist. Plant. 7: 215. 1880; Coult. & Rose, Contr. U. S. Nat. Herb. 7: 165. 1900; Piper, Contr. U. S. Nat. Herb. 11: 429. 1906; Henry, Fl. S. Brit. Col. 223. 1915; Piper & Beattie, Fl. N. W. Coast, 267. 1915; Carter & Newcombe, Prel. Cat. Fl. Vane. 61. 1921. Phellopterus Benth. in Benth. & Hook. Gen. PI. 1: 905. Sep- tember, 1867, not Phellopterus Nutt. (section under Cymopterus in Torr. & Gray, Fl. N. Am. 1: 623. 1840) in Coult. & Rose, Contr. U. S. Nat. Herb. 7: 166. 1900; Schmidt, Mem. Acad. Imp. Sci. St. Petersbourg, VII, 12 2 : 138. 1868; Franchet & Savatier, Enum. Plant. Jap. 1: 185. 1875; Wats. Bibl. Ind. 1: 430. 1878; Franchet, Cat. Plantes, in Mem. Soc. Nat. Sci. Cherbourg 24: 221. 1884; Coult. & Rose, Rev. N. Am. Umbell. 21, 81. 1888; Macoun, Check List Can. Plants, 25. 1889; Cat. Can. Plants 5: 329. 1890; Howell, Fl. N. W. Am. 1: 259. 1898; Engl. & Prantl, Nat. Pflanzenfam. 3 8 : 221. 1898; Ito & Matsumura, Tent. Fl. Lutch. in Jour. Coll. Sci. Imp. Univ. Tokyo 12: 529. 1899; Yabe, Rev. Umb. Jap. in Ibid. 16 2 : 92. 1902; Boiss. Omb. Cor. in Bull. Herb. Boiss. II, 3: 955. 1903; Nakai, Fl. Kor. 1 in Jour. Coll. Sci. Imp. Univ. Tokyo 26': 272. 1909. Herbaceous, subacaulescent, glabrous or pubescent perennials. Leaves coriaceous, petiolecl, bipinnatisect, broadly ovate in general outline. Inflorescence pedunculate, villous, peduncles shorter than or equalling the leaves; involucre usually absent, sometimes present in the form of a few linear bracts; involucel of conspicuous linear-lanceolate bracts. Calyx teeth inconspicuous. Stylopodium lacking. Fruit ovate-oblong to globose, glabrous or pubescent, flattened dorsally ; lateral and dorsal wings present ; wings broadened at the base; oil-tubes large, numerous, 2-6 on the commissural side; strengthening cells absent. Type species: Glehnia littoralis Schmidt, Prol. Fl. Jap. in Miq. Ann. Mus. Bot. Lugd. Bat. 3: 61. 1867; Prol. Fl. Jap. 249. 1867. 1928) MATHIAS — STUDIES IN THE UMBELLIFERAE. I 93 Abbreviations The following abbreviations have been used in citations to indicate the different herbaria from which material has been obtained for study : M = Missouri Botanical Garden Herbarium; G = Gray Her- barium of Harvard University; NY = New York Botanical Garden Herbarium; US = United States National Herbarium; W = Herbarium of the University of Washington deposited in the Washington State Museum; = Herbarium of the Univer- sity of Oregon; OAC = Herbarium of the Oregon Agricultural College; C = Herbarium of the University of California; P Herbarium of Pomona College. Key to the Species Fruit pubescent, species of the eastern hemisphere 1 . G. littorolis Fruit essentially glabrous, species of the western hemisphere 2. G. leiocarpa 1. Glehnia littoralis 1 Schmidt, Prol. PL Jap. in Miq. Ann. Mus. Bot. Lugd. Bat. 3: 61. 1867; Prol. FL Jap. 249. 1867. PL 17, fig. 2, 3, 5; PL 18; PL 19, fig. 1. " Archangelica officinalis, Hoffm.?" in Gray, "Account of the Botanical Specimens" from Narrative of the Perry Expedition 2: 312. 1856. "Cymopterus (?) littoralis, glaber" Gray, "Botany of Japan," in Mem. Am. Acad. N. S. 6 2 : 428. 1859, nomen nudum. Cymopterus ? littoralis Gray, "Botany of Japan" in Mem. Am. Acad. N. S. 6 2 : 391, 428. 1859, as to specimens from eastern hemisphere, nomen nudum. "Cymopteris glaber (A. Gray)" Black, "Catalogue of Japan 1 Glehnia littoralis Schmidt, em,— Planta humila, subacaula; foliis, petiolis excludentis, 5-13 cm. longis latisque, supra hirtellis in rachides nervosque, subtus glabris vel crebre tomentosis, ultimis segmentis foliorum oblongo-obovatis vel segmentis terminalibus cuneatis, 0.5-5 cm. longis, 0.4-4 cm. latis, apice rotundatis vel acutis, plus minusve cartilagineo-dentatis; petiolis 3-12 cm. longis, subdilatatis, hirtellis vel glabris; inflorescentiis umbellatis pedunculatis, crebre villosis; pedunculis subcrassis, subinde ramosis, foliis brevioribus vel aequantibus; umbellis patulis, 6-30-radiatis, radiis 1-3.5 cm. longis; involucro 1-3-bracteato; umbellulis capitatis, bracteis involucellorum pluribus, lanceolato-attenuatis; fructibus ovato-oblongis vel subglobosis, 0.4-1.5 cm. longis, villoso-pubescentibus, pilis multicellulatis; alis lateralibus saepe dorsalibus latioribus; vittis multis, 2-6 in commissurem. — Col- W.tad in Hakodate. Janan. 1861. Maximowicz (Gray Herb.), co-type. 94 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vou 15 Plants" in Hodgson, "A residence at Nagasaki and Hakodate in 1859-1860," 335. 1861, nomen nudum; Bonplandia 10: 92. 1862, nomen nudum. Phellopterus littoralis (Gray) Benth. in Benth. & Hook. Gen. PI. 1: 905. 1867, as to plants of eastern hemisphere; Hanee, Spic. Fl. Sin. in Jour. Bot. 16: 12. 1878; Forbes & Hemsley, Jour. Linn. Soc. Bot. 23: 331. 1888; Engl. & Prantl, Nat. Pflanzen- fam. 3 8 : 221. 1898, as to plants of eastern hemisphere; Ito & Matsumura, Tent. Fl. Lutch. in Jour. Coll. Sci. Imp. Univ. Tokyo 12: 529. 1899; Yabe, Rev. Umb. Jap. in Ibid, 16 2 : 93. 1902; Boiss. Omb. Cor. in Bull. Herb. Boiss. II, 3: 955. 1903; Nakai, Fl. Kor. 1. in Jour. Coll. Sci. Imp. Univ. Tokyo 26 1 : 272. 1909; Fl. Kor. 2. in Ibid. 31: 492. 1911. "C glaber" Gray ace. to Schmidt, Mem. Acad. Imp. Sci. St. Petersbourg, VII, 12 2 : 139. 1868, nomen nudum. "Phellopterus littoralis ace. to Schmidt, Mem. Acad. Imp. Sci. St. Petersbourg, VII, 12 2 : 138. 1868. "Phellopterus littoralis Schmidt" ace. to Franchet & Savatier, Enum. Plant. Jap. 1: 185. 1875; Franchet, Cat. Plantes, in Mem. Soc. Nat. Sci. Cherbourg 24: 221. 1884. "Glehnia littoralis (Gray) Schmidt" ace. to Coult. & Rose, Contr. U. S. Nat. Herb. 7: 165. 1900, as to plants of eastern hemisphere. Low subacaulescent plants; leaves, excluding petiole, 5-13 cm. long, about as broad, hirtellous on the rachises and nerves of the upper surface, glabrous to densely tomentose beneath, the ul- timate leaf-segments oblong-obovate or the terminal segments cuneate, 0.5-5 cm. long, 0.4-4 cm. broad, rounded to acute at the apex, somewhat unequally cartilaginously dentate; petioles 3-12 cm. long, somewhat inflated, hirtellous to glabrous; in- florescence pedunculate, densely villous; peduncles stoutish, sometimes branched, shorter than or equalling the leaves; umbels spreading, 6-30-rayed, rays 1-3.5 cm. long ; involucre 1-3-bracted; umbellets capitate, bracts of the involucel several, lance-attenu- ate; fruit ovate-oblong to subglobose, 0.4-1.5 cm. long, villous- pubescent with multicellular hairs, lateral wings usually broader than the dorsal wings; oil-tubes numerous, 2-6 on the com- missural surface. 1928] MATHIAS — STUDIES IN THE UMBELLIFERAE. I 95 Type specimen: Maximowicz, "Glehnia littoralis F. Schmidt. Fl. Sachalin ined." Iter secundum. Japonia. Hakodate. 1861. (type probably in Herb. Leningrad; co-type in the Gray Her- barium of Harvard University). Distribution: eastern hemisphere, along sandy sea-shores, from southern China northward, and in Japan. This plant is commonly known in Japan as " Hama-bofu" in relation to its maritime habitat, hama meaning sea-coast, and bofu, a medicinal plant. Specimens examined: Japan: Insula Sachalin, 1860, Schmidt (G); Kamiiso, Prov. Oshima, Hokkaida, 12 July, 1890, Miyabe & Tokubuchi (G); Hakodate, Iter secundum, 1861, Maximowicz (G co-type); Insula Jesso, circa Hakodate, 1861, Albrecht (G); Yezo, Ishikari, 10 Sept. 1903, Arimoto (G, M); Nambu, Nippon, 1865, Maximo- wicz, coll. Tschonoski (NY) ; Isoya, Shiribeshi, July, 1883, Take- nobu (G); seashore, Prov. Rikuzen, 9 July, 1913, Yasuda (W); Isl. Futami, 24 June, 1910, Flora Japonica, collector unknown (US 1155343); Loo-Choo Islands, 1853-56, Wright 98 (G, US); Corea, 1859, Wilford (NY). Siberia: Vladivostok and vicinity, May-Oct. 1919, Topping 2236 (G). China: Tsingtao, 1911, Zimmermann (G, US 795348); "Putoo Island— Clekiane," Henry (M); "Pootoo Isle, Chekiang," Faber M 6 (US); Delatache and Amoy, Henry (NY). 2. Glehnia leiocarpa 1 Mathias, nom. nov. PI. 17, fig. 1, 4; PI. 19, fig. 2. Cymopterus ? littoralis Gray, Mem. Am. Acad. N. S. 6 2 : 391, 428. 1859, as to American specimens, nomen nudum; Stevens' Rept. U. S. Expl. & Surv. from Miss, to Pacific Ocean 12 2 : 62. 1860; Jeps. Man. Fl. Plants Calif. 731. 1925. Phellopterus littoralis (Gray) Benth. in Benth. & Hook. Gen. •Planta humila, subacaula; foliis, petiolis excludentis, 2.5-15 cm. longis latisque, supra hirtellis in rachides nervosque, subtus crebre tomentosis, ultimis segmentis foliorum oblongo-obovatis vel segmentis ter- minalibus cuneatis, 0.5-5 cm. longis, 0.4-3 cm. latis, apice rotundatis vel acutis, plus minusve dentatis, marginibus subinde cartilaginibus; petiolis 2.5-14 cm. longis, subdilatatis, hirtellis; inflorescentiis umbellatis pedunculitis, crebre villosis; pedun- culis subcrassis, subinde ramosis, saepe foliis brevioribus, rare aequantibus; umbellis 1 Glehnia 96 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 PL 1: 905. 1867, as to American plants; Engl. & Prantl, Nat. Pflanzenfam. 3 8 : 221. 1898, as to American plants. "Glehnia littoralis (Gray) Schmidt" ace. to Coult. & Rose, Contr. U. S. Nat. Herb. 7: 165. 1900; Piper, Contr. U. S. Nat. Herb. 11: 429. 1906; Piper & Beattie, Fl. N. W. Coast, 267. 1915; Carter & Newcombe, Prel. Cat. Fl. Vane. 61. 1921. ' ' P hellopterus littoralis Schmidt" ace. to Wats. Bibl. Ind. 1: 430. 1878; Coult. & Rose, Rev. N. Am. Umbell. 81. 1888; Macoun, Check List Can. Plants, 25. 1889; Cat. Can. Plants 5: 329. 1890; Howell, Fl. N. W. Am. 1: 259. 1898. Glehnia littoralis Schmidt ace. to Henry, Fl. S. Brit. Col. 223. 1915. Low subacaulescent plants; leaves, excluding petiole, 2.5-15 cm. long, about as broad, hirtellous on the rachises and nerves of the upper surface, mostly densely tomentose beneath, the ultimate leaf-segments oblong-obovate or the terminal segments cuneate, 0.5-5 cm. long, 0.4-3 cm. broad, rounded to acute at the apex, unequally dentate, margins sometimes cartilaginous; petioles 2.5-14 cm. long, somewhat inflated, hirtellous; inflores- cence pedunculate, densely villous ; peduncles stoutish, sometimes branched, usually shorter than the leaves, rarely equalling them; umbel globose to spreading, 5-13-rayed, rays 0.5-4.5 cm. long; involucre 1-3-bracted; umbellets capitate, bracts of the involucel several, lance-attenuate; fruit ovate-oblong to subglobose, 0.4-1.2 cm. long, essentially glabrous (sometimes with a few scattered multicellular hairs), lateral wings sometimes broader than dorsal wings; oil-tubes numerous, 2-6 on commissural surface. Type specimen: J. G. Cooper, "sandy shores, Washington Terr. (Shoal Water Bay)." 1854. (The type is in the Gray Herbarium of Harvard University and is labeled "Cymopterus ? littoralis, n. sp." in Gray's handwriting; co-types are in the Herbarium of the New York Botanical Garden and in the United States National Herbarium.) globosis vel patulis, 5-13-radiatis, radiis 0.5-4.5 cm. longis; involucro 1-3-bracteato; umbellnlis capita tis, bracteis involucellorum pluribus, lanceolato-attenuatis; fructibus ovato-oblongis vel subglobosis, 0.4-1.2 cm. longis, fere glabris vel subinde sparse pubescentibus, pilis multicellulatis; alis lateralibus dorsalibus latioribus; vittis mul- tis, 2-6 in commissurem. — Collected on sandy shores, Shoal Water Bay, Washington Territory (State of Washington), 1854, J. G. Cooper (Gray Herb.), type. 1928] MATHIAS — STUDIES IN THE UMBELLIFERAE. I 97 Distribution: North America along sandy sea-coasts from San Francisco, California northward. Specimens examined: Alaska : along the Ankow River, near Ocean Cape, vicinity of Yakutat Bay, 1 July, 1892, Funston 51 (NY, M, C). British Columbia: vicinity of Ucleulet, Long Beach, Van- couver Island, 25 June, 1909, Macoun 78600 (US); sand, Oak Bay, Vancouver Island, 31 May, 1887, Macoun (G). Washington: Lopez, San Juan Islands, 25 June-1 Aug. 1917, S. M. & E. B. Zeller 963 (NY, M, G, US) ; Puget Sound, Wilkes Expedition (NY, US 44092); Port Angeles, 26 June, 1908, Webster (W); sand dunes, Ocean Park, April, 1908, Rigg (W); Ilwaco, 21 June, 1904, Piper 5002 (US); Oyhut, Chehalis County, 7 June, 1897, Lamb 1249 (NY, M); drifting sand, common along the ocean beach, Westport, Chehalis Co., 26 June, 1892, Hender- son 885 (US) ; ocean beach, Westport, Chehalis County, 26 June, 1892, Henderson (W); sand dunes, Westport, June, 1917, Grant (NY); sand spit, Sequim, June, 1915, Grant (NY, M 788926); Seattle, July, 1915, Freiberg (M 813695); sandy dunes, mouth of "Joe Creek," near Moclips, 28 June, 1908, Foster 824 (US); sandy sea-shores, Port Angeles, 26 June, 1908, Flett 8375 (US) ; Olympic Mts., Clallam Co., July, 1900, Elmer 2768 (NY, M, US); M. Beach, Westport, 10 July, 1907, Cowles 512 (M); "sandy shores, Washington Terr. (Shoal Water Bay)." 1854, Cooper (G type, NY, US); beach sand, Copalis, June-July, 1902, Conrad 892 (US); Copalis, 30 May, 1912, Bardell (M 813656). Oregon: Clatsop Beach, Clatsop Co., 21 Aug. 1902, Sheldon 11252 (NY, M, G, P, US); Gearhart, 19 June, 1904, Piper 6241 (US); Gearhart, 19 June, 1904, Piper 6131 (US); sandy sea- beach, Newport, 3 July, 1918, /. Nelson 2292 (G); Nestart's Bay, Tillamook Co., 29 June, 1894, Lloyd (NY); on strand, Nestucca, July-Aug. 1901, Kirkwood 149 (NY); sandy sea-shore, mouth of the Umpqua River, 18 June, 1885, Howell 1 (OAC, US, M, 1151); on sand dunes, mouth of Tillamook Bay, 16 July, 1882, T. Howell (NY) ; on shifting sand, Tillamook Bay, 14 July, 1 his herbarium labels. appearing [Vol. 15 98 ANNALS 1882, T. J. Howell (M, US 33339); Clatsop Beach, 20 July, 1891, J. Howell (M 863104); on shifting sands of sea-shore, Coos Bay, 19 Aug. 1911, House 4705 (US, NY); drifting sand, ocean beach, Tillamook Bay, 14 July, 1882, Howell & Henderson (O); drifting sand, ocean beach, Clatsop, 30 July, 1887, Henderson 385 (M, OAC) ; beach, below Florence, 20 May, 1925, Henderson (0) ; sand of the ocean above high tide, Rockaway, 16 Sept. 1925, Henderson (O) ; sea-shore, Fort Stevens, 7 July, 1886, Henderson (O); Bayocean, Garibaldi, 28 Aug. 1914, Hitchcock 12370 (US); sands of the Oregon coast between Umpqua and Coos Bay, 12 Aug. 1880, G. Engelmann (M); on sand dunes of the ocean, Gearhart, Clatsop County, 1 Sept. 1898, Coville 861 (US). California : in drifting sand, Humboldt County, sand hills of ocean beach at Samoa, opp. Eureka, 7 Aug. 1901, Tracy 1261 (C) ; Samoa Beach, Humboldt Co., 17 June, 1911, Smith 3854 (NY); Trinidad, Humboldt Co., 7 June, 1911, Smith 3806 (NY); Trini- dad, Humboldt County, 6 July, 1911, Smith 3806 (US); Pebble Beach, Crescent, Del Norte Co., 17-20 June, 1925, Parks 8257 (C 279023); sandy dunes at Humboldt County camp, 7 miles south of Trinidad, 24 July, 1924, A. A. Heller 13882 (NY); Crescent City, Del Norte Co., 30 June, 1899, Davy & Blasdale 5960 (C); Point Arena, Mendocino Co., 24 July, 1900, Davy 6050 (C); peninsula, Eureka, 23 Aug. 1904, Congdon (C 140694); sea-shore peninsula, Eureka, Humboldt Co., 23 July, 1904, Congdon (M); Humboldt Bay, May, 1901, Chandler 1145 (C); Trinidad, Humboldt Co., 18 July, 1916, Abrams 6140 (NY, O). The genus Glehnia is characterized by its maritime habitat, broad leaf divisions, thick coriaceous texture of the leaves, and prominent wing development of the fruit. The two species are separated largely on fruit characters. Glehnia littoralis Schmidt, the species of the eastern hemisphere, always has a pubescent fruit. The pubescence is villous with multicellular hairs. The mature fruit may be only slightly pubescent due to the falling off of the hairs but in such cases it has a tuberculate appearance showing the previous attachment of these hairs. In the young fruit the pubescence is densely villous. As a rule the oil-tubes of the fruit are smaller and more numerous than in the other 1928] MATHIAS — STUDIES IN THE UMBELLIFERAE. I 99 species. The characters of inflorescence and foliage are similar in both species. There is quite a range of variation in foliage pubescence of Glehnia littoralis. The type of the species, the Maximowicz plant from Hakodate, represents an intermediate condition, and upon an examination of additional material may- prove to be a hybrid between the glabrous and pubescent forms (pi. 18, fig. 2). The leaves are hirtellous on the lower surface and on the veins and rachises of the upper surface. The one extreme of variation in pubescence is typified by the plant col- lected by Wright in the Loo Choo Islands and labeled by Gray "Cymopterus littoralis ?? Gray, var. glabra, vel sp. am" (pi. 18, fig. 1). The leaf is essentially glabrous, the hirtellous condition being limited entirely to the veins and rachises. The margins of the leaves are more frequently cartilaginous than in other forms. The other extreme of variation is typified by the plant collected by Schmidt in Sachalin in 1860 (pi. 19, fig. 1). This plant super- ficially more closely approaches the species of North America. The lower surface of the leaf has the same dense tomentose pubescence that occurs in Glehnia leiocarpa. However, an ex- amination of a large amount of material from the eastern hemi- sphere shows a great number of intergrading forms; a gradual variation exists from the extreme glabrous form to the densely tomentose one. The fruit in all forms is similar, and the pubes- cence characters of the foliage are of no value in separating Glehnia littoralis into varieties or forms. Glehnia leiocarpa, on the other hand, shows a very constant pubescence character. The leaves in every case are densely tomentose beneath. The fruit is glabrous with the exception of occasional multicellular hairs on the margins of the wings. In no case was a tuberculate appearance observed which would point to the previous attachment of hairs in younger conditions. The young fruit in most cases is essentially glabrous. Moreover, a cross-section of the fruit shows the oil-tubes to be larger and generally fewer in number than in G. littoralis. An interesting geographical distribution is shown in connection with this genus. The two closely related species occur along the coast on both sides of the Pacific Ocean (fig. 1). Glehnia leiocarpa extends from Alaska to northern California and G. littoralis from [Vol. 15 100 ANNALS OF THE MISSOURI BOTANICAL GARDEN a o CD o o 3 00 bfi a o CO bC 1928] MATHIAS — STUDIES IN THE TJMBELLIFERAE. I 101 Siberia to southern China and through Japan. It is also inter- esting to note the distribution of the different pubescence types of G. littoralis. The more pubescent plants and those most nearly approaching G. leiocarpa occur in the northern region of the distribution area of the species, while the more glabrous plants are found in the southern range of distribution. Such a dis- tribution seems to indicate that the ancestors of this species occurred in the intermediate area and in the land bridge con- necting North America and Asia somewhere in the Bering Sea region . A similar distribution for other genera has been pointed out by various workers in this field. One of the earliest important works was Dr. Gray's 1 article on the "Botany of Japan" in which he showed the similarity of the flora of northwest as well as eastern America to that of Japan. In this work he also mentions the distribution of the genus Glehnia. Butters, 2 more recently, has pointed out a similar distribution for the genus Athyrium. Berry 3 has shown this distribution for Castanopsis, Pasania, Corylus. Juglans, and other genera. The writer is indebted to Dr. George T. Moore, Director of the Missouri Botanical Garden, for the use of the library and her- barium of that institution. Sincere appreciation is due Dr. N. L. Britton and Dr. J. K. Small of the New York Botanical Garden, Dr. B. L. Robinson and Dr. Ivan M. Johnston of the Gray Herbarium, Dr. Wm. R. Maxon of the United States National Herbarium, Prof. L. F. Henderson of the University of Oregon, Dr. Helen M. Gilkey of Oregon Agricultural College, Prof. T. C. Frye and Miss Martha R. Flahaut of the University of Washing- ton, Dr. N. L. Gardner of the University of California, and Dr. Philip A. Munz of Pomona College for the privilege of examining material in the herbaria of the above-mentioned institutions or for the loan of material necessary for this study. Thanks are also due Dr. John H. Barnhart of the New York Botanical 1 Gray, A. "Botany of Japan." 376-449. 1859. 1 Butters, F. K. Taxonomic and geographic studies in North American ferns I. The genus Athyrium and the North American ferns allied to Athyrium Filix femina. Rhodora 19: 169-207. 1917. • Berry, E. W. Tree ancestors. 270 pp. 1923. 102 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN Garden, Dr. J. N. Rose of the United States National Herbarium, and Dr. F. A. F. C. Went of the Botanical Laboratory of Utrecht for their assistance in bibliographical details. Especial thanks are due Dr. J. M. Greenman, Curator of the Herbarium of the Missouri Botanical Garden, for his advice and assistance. List of Exsiccatae The distribution numbers are printed in italics. The number in parenthesis is the species number used in this revision. Abrams, L. 61 40 (2). Albrecht, N. Arimoto, S. Bardell, E. M. (1). (1). Howell, J. Howell, T. Howell, T. J. (2). (2). (2). (2). Chandler, H. P. 1146 (2). Kirkwood, J. E. 149 (2). Lamb, F. H. 1249 (2). Congdon, J. W. (2). Lloyd, F. E. (2). Conrad, H. S. 892 (2). Cooper, J. G. Macoun, J. — , 78600 (2). (2). Maximowicz, C. J. (i). Coville, F. V. 861 (2). Cowles, H. C. 612 (2). Davy, J. B. 6050 (2). Maximowicz, C. J. (coll. Tschonoski) (1). Miyabe, K. and Tokubuchi, E. (1). Davy, J. B. and Blasdale, W. C. 6960 Nelson, J. C. 2292 (2). (2). Elmer, A. D. E. 2768 (2). Engelmann, G. — (2). Faber, E. M» (1). Flett, J. B. 3375 (2). Parks, H. E. 8257 (2). Piper, C. V. 6002, 6181, 6241 (2). Rigg, G. B. Schmidt, F. (2). (1). Sheldon, E. P. 11262 (2). Fl. Japonica (collector unknown), — (1). Smith, H. H. 8806, 8854 (2). Foster, A. S. 824 (2). Takenobu, S. (1). Freiberg, G. W. — Funston, F. 51 (2). (2) Grant, J. M. (2). Topping, L. 2236 (1). Tracy, J. P. 1261 (2). (2). Heller, A. A. 18882 (2). Henderson, L. F. — 885 (2). (1). (2). Henry, A. a). Hitchcock, A. S. 12370 (2). House, H. D. 4705 (2). Webster, E. B. Wilford, C. Wilkes Expedition, Wright, C. 98 (1). (1). Zeller, S. M. and E. B. 963 (2). Yasuda, A. Howell T. and Henderson, L. F. — (2). Zimmermann, R. (1). 1928 MATHIAS- — STUDIES IN THE UMBELLIFERAE. I 103 Glehnia leiocarpa Mathias Index of Species New species and combinations are printed in bold face type; synonyms in italics; and previously published names in ordinary type. "Arch-angelica officinalis Hoffm.?" 93 Athyrium 101 1 ' C. glaber > ' Gray 94 Castanopsis 101 Corylus 101 "Cymopteris glaber (A. Gray)" 93, 95, 99. 101 Glehnia littoralis Schmidt 91, 92, 93, 95, 96, 98, 99, 101 Black 9 Cymopterus Raf 91, 92 Cymopterus sect. Phellopterus Nutt. 92 Cymopterus? littoralis Gray. . .91, 93, 95 "Cymopterus littoralis?? Gray var. glabra" 99 "Cymopterus (?) littoralis glaber" Gray 93 Glehnia Schmidt 91, 98, 101 "Glehnia littoralis (Gray) Schmidt" 94,96 Juglans 101 Pasania 101 Phellopterus Benth 91, 92 Phellopterus Nutt 92 Phellopterus littoralis (Gray) Benth. 94,95 " Phellopterus littoralis Schmidt" 94, 96 "Phellopterus littoralis" ace. to Schmidt 94 [Vol. 15, 19281 104 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 17 Fig. 1. Mature fruit of Glehnia leiocarpa Mathias, collected on "sandy shores, Washington Terr. (Shoal Water Bay)," Cooper, 1854 (Gray Herb.), type. X 6. Fig. 2. Mature fruit of Glehnia littora 1 is Schmidt, collected on the Island of Sachalin, Schmidt, 1860 (Gray Herb.). X 6. Fig. 3. Mature fruit of Glehnia littorolis Schmidt, collected in Yezo, Ishikari, Arimoto, 10 Sept. 1903 (Mo. Bot. Gard. Herb.). X 6. Fig. 4. Cross-section in median plane of immature fruit of Glehnia leiocarpa Mathias, collected on "sandy shores, Washington Terr. (Shoal Water Bay)/' Cooper, 1854 (Gray Herb.), TTPB. X 10. Fig. 5. Cross-section in median plane of mature fruit of Glehnia littorolis Schmidt, collected on the Island of Sachalin, Schmidt, 1860 (Gray Herb.). X 10. Ann. Mo. Bot. Gaud., Vol. 15, 1928 Plate 17 1 2 3 4 5 MATH I AS— STUDIES IN TMBELLIFERAE [Vol. 15, 1928] 106 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 18 Glehnia litioralis Schmidt Fig. 1. From specimens in the Gray Herbarium of Harvard University, namely W rig hi and Albrecht. Fig. 2. From the co-type specimen, Maximowicz, in the Gray Herbarium of Harvard University. Ann. Mo. Bot. Gakd.. Vol. 15, 1928 Plate 18 > CO CO H CO 3 > td It 1 > IO * 1 ** - ■ r | ■ ■ ■ > I u [Vol. 15, 1928 108 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 19 Fig. 1. Glehnia littoralis Schmidt, from a specimen collected by Schmidt, in the Gray Herbarium of Harvard University. Fig. 2. Glehnia leiocarpa Mathiaa, from the type specimen, Cooper, in the Gray Herbarium of Harvard University. Ann. Mo. Bot. Gard., Vol. 15, 1928 Plate 19 2 > X > i 3 X r 1 ►- ~ IC > CONCERNING THE STATUS OF THE GENUS LATERNEA DAVID H. LINDER Mycologist to the Missouri Botanical Garden Instructor in the Henry Shaw School of Botany of Washington University While in Cuba during the summer of 1 924 , the writer collected a member of the Clathreae which was subsequently determined in Saccardo ('88) as Clathrus triscapus (Turpin) Fr. In going over the literature concerning the simple columnar species of Clathrus, it was observed that there were few statements as to the manner in which the gleba is borne. Examination of the figures accompanying the original descriptions led the writer to the conclusion that the majority of these simple species carry the gleba in the same fashion as do the more complex ones in which the columns anastomose to form a latticed sphere. In these latter forms, exemplified by Clathrus cancellatus and C. crispus, the gleba is closely applied to the inside of the columns or receptacles. However, in the genus Laternea, of which L. triscapa is the original species, the columns, strictly speaking, are stipes united above, and these subtend an angular body, subovate in outline, the "lanterne" of Turpin (1822). A comparison of C. columnatus , C. crispus, and C. cancellatus brings out the fact that except for the gross morphological dif- ferences, the structure of the simple columnar and the more complex latticed species is essentially the same; that is, the columns may be relatively rough or even smooth on the outside, but on the inside they are always rough and pitted (pi. 20, figs. 3-6). It is to this pitted inside surface of the columns that the gleba is applied. Studies of preserved young material of Clathrus columnatus and observations of the other two species in the field at all stages of development amply support this view. There is certainly no evidence that, at the time of rupture of the volva, any definite receptacle other than the column is present. Aside from the fact that the gleba of Laternea triscapa is strictly confined to the angled, subovate, specialized receptacle pendant from the i unction of the apices of the columns, it differs from Ann. Mo. Bot. Gabd., Vol. 15, 1928 (109) [Vol.15 11U ANNALS OF THE MISSOURI BOTANICAL GARDEN C. columnatus and other similar members of that genus by being proportionately taller and more slender. In addition, the col- umns are less angular and the surfaces are smooth, both on the inside and outside (pi. 20, fig. 1, 2). With the above distinction in mind, it becomes quite evident that Turpin (1822) was thoroughly justified in creating the genus Laternea for that form which bears the gleba in the manner men- tioned. Accepting this view , then Laternea triscapa becomes the only representative of the genus and Laternea columnata, L. pusilla, L. rhacoides, L. Spegazzini, L. angolensis, and possibly L. bicolumnata, considered as belonging to the genus by Lloyd ('09), should be excluded and treated as members of the genus Clathrus, following the treatment by Fischer ( '86) . Certainly this is a more natural grouping, especially since Clathrus colmnnatus tends towards the more complex type represented by C. cancel- latus. The latter species may at times be columnar below and only show anastomosis of the receptacle above, while in the former, as is shown in pi. 20, fig. 6, there is a tendency for the columns to divide to produce four or even five. If, however, it is deemed more convenient to separate the simple columnar members from the genus Clathrus, then the genus Colonnaria Rafinesque (1808), on the basis of priority, should be restored, and Clathrus of Michelius (1729) should be reserved for those forms with anastomosed receptacles. In view of the former uncertainty concerning Laternea, it seems advisable, while restoring it to its original status, to re- describe the genus, and also the species as follows : Laternea Turpin: Columns slender, smooth, usually three, subtending from the junction of the apices an angular, subovate receptacle to which the gleba is restricted. Laternea triscapa Turpin: columns 3, "capucine buff" * at base, becoming "cadmium orange" above; smooth on inne;r and outer surfaces, 5-6.2 cm. long, 4-5 X 6 mm. in diameter, united above; receptacle pendant, "nopal red," angled, subovate in outline, 10 X 13 mm.; gleba deep olive; volva white, 15 X 20 mm. In sugar cane field at edge of woods, Soledad, Cuba. Sep- 1 Ridgway, R. Color standards and nomenclature. Washington, D. C, 1912. s 1928] LINDEP — STATUS OF THE GENUS LATERNEA 111 tember, 1924, hinder (in Farlow Herb, at Harvard Univ. and writer's herbarium). In conclusion, the writer wishes to express his indebtedness to Prof. William H. Weston, Jr. for the loan of the preserved material of Clathrus columnatus Bosc. Bibliography Fischer, E. ('86) . Versuch einer systematischen Uebersicht liber die bisher bekannten Phalloideen. Konigl. Bot. Gart. u. Mus. Berlin, Jahrb. 4: 1-92. pi. 1. 1886. Lloyd, C. G. ('09). Synopsis of the known Phalloids. pp. 48-65. /. 59-64' Cincinnati, 1909. Michelius, P. A. (1729). Nova plantarum genera iuxta Tournefortii methodum dis- posita. p. 214. pi 93. Florence, 1729. Rafinesque, C. S. (1808). Prospectus of Mr. Rafinesque Schmaltz's two intended works on North American botany; the first on the new genera and species of plants discovered by himself and the second on the natural history of the funguses, or mushroom-tribe of America. N. Y. Med. Repository, 2nd hexade 5 (11): 355. 1808. Saccardo, P. A. ('88). Sylloge Fungorum 7: 18-21. 1888. Turpin, P. J. F. (1822). Diet. d. Sci. Nat. 25: 248. 1822. [Vol. 15, 19281 112 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 20 Fig. 1. Latemea triscapa, showing specimen. Natural size. Burfaces of the columns. Photograph of freshly collected Fig. 2. Laternea triscapa. Receptacle and upper part of columns enlarged three times to show the definite development of tissue to form the receptacle. Fig. 3. surface rough, pitted inner surface. The gleba may be seen still adhering to the inner surface at the junction of the columns or receptacles. From preserved material collected in Porto Rico by P. V. Siggers. Natural size. Fig. 4-6. Clathrus columnatus. Fig. 4 shows the upper portion of the recep- tacle enlarged three times. There is no evidence of any tissue that may be con- sidered comparable to that found in the receptacle of L. triscapa, the gleba being applied to the inner pitted surface. Figures 5 and 6 show the receptacles spread out and viewed from below; the latter figure illustrates the division of one of the columns to form a 5-colunmar receptacle. 4 /s natural size. Aw. Mo. Hot. Gabd., Vol. 15, 1928 Plate 20 2 4 3 « LIXDER— THE GENUS LATERNEA Annals of the Missouri Botanical Garden Vol. 15 APRIL, 1928 No. 2 SOURCES OF ENERGY FOR AZOTOBACTER. WITH SPECIAL REFERENCE TO FATTY ACIDS P. L. GAINEY Soil Biologist, Kansas Agricultural Experiment Station l hingt Introduction The thermo-chemical phenomena involved in the fixation of free nitrogen by various micro-organisms are not well under- stood. It has been assumed that the fixation process is endo- thermic in nature and that the necessary energy is, in the case of the Azotobacter group of organisms, derived from the oxidation of organic compounds, principally of a carbohydrate, acid, or alcohol nature. Regardless of whether the initial process through which nitrogen is brought into combination is exo- or endo-thermic, no one has been able to establish definitely a measurable fixation of nitrogen by Azotobacter, or any other nitrogen-fixing group of organisms, in the complete absence of some form of oreranic matter. Furthem found to run more and nitrogen fixation have been of the organic material available, provided the material is non-nitrogenous in nature. It may be assumed safely, therefore, that an organic food material of some kind is essential in the metabolism of this group of organisms. This being true, it would seem highly desirable, both from a theoretical and practical standpoint, to secure as much information as possible relative 1 An investigation carried out in part at the Missouri Botanical Garden in the Graduate Laboratory of the Henry Shaw School of Botany of Washington University and in part in the Research Laboratory of Soil Biology of the Kansas Agricultural Experiment Station, and submitted as a thesis in partial fulfilment of the require- ments for the degree of doctor of philosophy in the Henry Shaw School of Botany of Washington University. Ann. Mo. Bot. Gabd., Vol. 15, 1928 (113) [Vol. 15 114 ANNALS OF THE MISSOURI BOTANICAL GARDEN to the different organic food substances suitable for these organ- isms and also the relative efficiency with which different com- pounds can be used. As an aid in the solution of some of the more complicated thermo-chemical questions involved it seemed desirable to ascertain, if possible, whether any quantitative relationship existed between the potential energy content of the organic material utilized, on the one hand, and the quantity of growth and nitrogen fixed, on the other. It was with the hope of securing information along these lines that this work was under- taken. More specifically this investigation has been concerned with seeking an answer to the following questions: (1) Is there any difference in the relative availability of the lower fatty acids as a source of carbon, or organic food substance, for Azotobader? (2) If Azotobader exhibits differences in ability to utilize different fatty acids, can such differences be associated with the structure, size, or energy content of the molecule? The following criteria were used in judging the ability of Azotobader to utilize the various acids: (a) the variation in visible growth; (b) the disappearance of the acid; (c) the fixation of nitrogen; (d) changes in the hydrogen-ion concentration of the medium fatty acids were selected for study because compose a series of compounds exhibiting many characteristics in common, yet in the series the molecule increases in definite increments from fairly simple to fairly complex. Furthermore, the heat of combustion of these compounds increases directly as the molecular weight increases. By including the iso com- pounds, two molecules varying in structure but identical in composition and energy content could be compared. An addi- tional desirable characteristic possessed by this series of com- pounds is that they are found as free acids or as constituents of fats in nature, and some of them are already known to serve as organic food for Azotobader. One other characteristic essential in a series of compounds suitable for a study of this nature is susceptibility to fairly easy quantitative analysis. Not many series of organic compounds of which the members possess the 19281 GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 115 desirable characteristics enumerated above can be found, and possibly there is no other series in which as many members may be used so advantageously. Unfortunately, lack of solubility prevents any quantitative use of members of this series above six carbon atoms, but even then there would seem to be enough members of the series that can be used to give valuable informa- tion if carefully studied. Methods General procedure — The general procedure has been to prepare, in one batch, as carefully as possible, all the medium neces- sary for a single experiment. Measured quantities of this were placed in the culture flasks which were then stoppered with cot- ton and sterilized in the autoclave. After sterilization the flasks were inoculated as uniformly as possible with a heavy suspen- sion of organisms washed from the surface of mannitol-soil-ex- tract agar upon which active vigorous growth was taking place. No old or apparently non-vigorous growing culture was used as an inoculum. After varying periods of incubation the cultures were removed from the incubator and the qualitative and quan- titative analyses were made as indicated. Where supplementary aeration was resorted to, an inlet tube of glass, containing five to seven small openings near the end, was inserted in a rubber stopper in such a way that when the stopper was tight in the culture flask the end of the tube almost reached the bottom of the flask. The stopper was also provided with an outlet tube to be connected to a vacuum system. The stopper and connecting tubes were sterilized separately and inserted after inoculation. Cotton was forced into the ends of the connecting tubes as a precaution against possible contamina- tion. In the aerated experiments 300-cc. Pyrex Erlenmeyer flasks were used as culture containers, and all those in any one experiment were connected in series so that the same quantity of aeration was provided for all. While incubating, a vigorous bubbling of air through the media was continuously maintained. Before entering the first culture flask the air was washed through flasks arranged as mentioned above, of acid, alkali, and water. Despite special precautions to prevent contamination there was one type of foreign organism difficult to keep out. I Vol. 15 116 ANNALS OF THE MISSOURI BOTANICAL GARDEN Where no special aeration was provided the cultures consisted of 50 re. of media in 300-cc. flasks; 100 cc. media in 750-cc. flasks; or 200 cc. media in 1000-cc. Pyrex Erlenineyer flasks. These quantities of media in the flasks indicated always exhibited a large surface area compared to the depth, and while aeration was certainly not as vigorous as where air was drawn through the culture, nevertheless it was ample for very rapid growth. Growth at the bottom of the culture was frequently observed before it made its appearance on the surface, indicating aeration through- out the culture. These cultures were left stationary except when being handled for examination, and even then care was taken not to shake so vigorously as to break up any film that might be forming on the surface. Medium — Unless otherwise stated the medium used in the various experiments had the following composition, and its suitability is evidenced by the very rapid growth that took place when the organic material was assimilable: K*HPO< 2.50 gms. MgSO< 20j.iq. NaCl 20 gm. CaCl 2 06 gm. FeCl.i (10 per cent, sol.) 1 . drop Organic material I per cent Distilled water 1000 CC. In some of the preliminary experiments only 0.50 gm. of K 2 HP0 4 was used, but it was observed that when such a small quantity of phosphate was added the hydrogen-ion concentration sometimes changed so rapidly and markedly that growth was very soon inhibited by the increase of hydroxyl-ions. A rapid increase in hydroxyl-ion concentration was always observed when large quantities of a metallic salt of an organic acid were metab- olized, and probably arose from the formation of an hydroxide by the metallic-ions set free when the acid radicle was assimilated by the organisms. Even with 2.5 gms. phosphate and an excess of CaC0 3 the buffering effect was barely sufficient to permit of complete* oxidation of 1.0 per cent acid. In fact, in some instances then 1 is evidence to indicate that the high all alinity accompanying vigorous oxidation not only prevented further activity, as is 1928! GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 117 evidenced by the failure of the organisms in certain cultures to oxidize all the acid even where abundant growth occurred, but actually resulted in the death of most or all of the organisms present. In case of some of the less soluble acids the quantity added was only 0.5 per cent. When the organic material was fatty acid it was, in all except a few preliminary experiments, added to the entire volume of distilled water. An excess of CaC0 3 was then added, and the material boiled until the reaction became neutral to brom-thymol- blue, indicating complete transformation into the calcium salt, after which it was filtered. This procedure was resorted to in order to hasten the completion of the reaction between the weakly dissociated acid and calcium carbonate. The other salts were then added, and if any change in the reaction took place it was again adjusted to neutrality by the addition of sodium hydroxide or sulphuric acid as needed. The medium was then measured into the culture flasks, care being taken to keep it well agitated in order to secure an equal distribution of the precipitate, a small quantity of CaC0 3 added, the flask plugged with cotton and sterilized in the autoclave at ten pounds pres- sure. The final reaction of medium prepared as indicated was never far from P H 7.0. Obviously, one could not depend upon the original weight of any organic material subjected to the manipulations described in the preceding paragraph as indicating the final concentration. It was necessary, therefore, to prepare controls and make quanti- tative analyses of the final concentration of the organic materials in all cases. The agar used for maintaining stock cultures, for testing the purity of cultures, and for the preparation of the inoculum was a soil-extract-mannitol agar prepared as follows: One thousand gms. fertile garden soil were added to 1000 cc. distilled water and subjected to fifteen pounds pressure in the autoclave for thirty minutes, after which CaC0 3 was added and the mixture filtered. The clear filtrate was made up to 1000 cc. with dis- tilled water. To 900 cc. distilled water was added 100 cc. soil extract, 0.5 gm. K 2 HP0 4 , 10 gms. mannitol, and 15 gms. agar agar. After heating in the autoclave to bring the agar agar into [Vol. 15 118 GARDEN solution, and while still hot, phenolphthalein and sufficient sodium hydroxide were added to give a distinct pink color. Cultures — The following include all cultures employed in any experiment, together with their origin. They were selected from among more than one hundred available cultures of Azoto- bacter. Those strains that were used to any appreciable extent were selected primarily because of their vigorous growing and nitrogen-fixing ability. Cultures Nos. 3a and 3b were strains of Azotobacter chroococcum secured from S. A. Waksman, of New Brunswick, N. J. Cultures Nos. 4, 5a, 5b, 6, 7, 8, 57, 58, 59, 60, 62, and 66 were isolated from different Colorado soils in the laboratory of W. G. Sackett, Fort Collins, Colo. Culture No. 94 was a strain of Azotobacter vinelandii from the Bureau of Plant Industry, U. S. Dept. Agr., Washington, D. C. Culture No. II was received from W. Omeliansky, Leningrad, Russia. Culture No. 218, a strain of Azotobacter chroococcum marked "K," was received from Chr. Barthel, Stockholm, Sweden. Cultures U C" and "R" were received from the Rothamsted Experiment Station, England. "C" came originally from a single cell strain of H. R. Christensen's and "R" was isolated from soil. Cultures Nos. 178, 187, 188, 194, and 165 were all isolated in this laboratory from the following soils, respectively: No. 178, Gloucester loam from Minnesota; No. 187, field soil from New York; No. 188, cotton and sugar cane soil, Virgin Islands; No. 194, soil from V. L. Winogradsky, Paris, France; No. 165, irrigated potato field soil from Wyoming. The only unidentified strain used to any appreciable extent was No. 62. This culture possessed the characteristics of Azotobacter chroococ- cum in that it grew abundantly as grayish-white opaque, distinct colonies, soon turning brown and eventually black with a more or less wrinkled dry surface. Before any culture was used in any experiment it w r as carefully tested for purity by repeated streaking and re-isolation from individual, microscopically examined colonies, until assured of the presence of only one type of organism. Furthermore, after incubation most cultures were again examined for purity and if evidence of contamination was present it has been so recorded. Inoculum. — The inoculum was prepared by streaking the 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 119 entire surface of a Kolle flask of soil-extract-mannitol agar with the desired culture, incubating 48-96 hours, or until the entire surface was covered with a uniform thick growth, and suspending this growth in 25-50 cc. of sterile water. This gave a suspension of such density as to be practically opaque in a depth of only half an inch. One or two per cent of this was used as the inoculum, thus insuring a very heavy inoculation. Incubation. — All cultures were incubated either at summer room temperature or in an incubator at 28-32° C. Room tem- perature was quite favorable in the summer but during the winter the temperature dropped too low at night for active growth. Most of the aerated experiments were run at room temperature, while all non-aerated experiments were incubated at 28-32° C. CHEMICAL METHODS Dextrose. — Quantitative dextrose determinations were made by the Shaffer-Hartmann ('21) iodometric method. Where a heavy growth of Azotobacter had taken place the slime-like material present was precipitated by adding 1.0 per cent of a mixture of 2.5 gms. phosphotungstic acid and 5.0 gms. H 2 S0 4 before sugar determinations were made. Preliminary experi- ments proved that sugar could be recovered quantitatively when added to a culture and treated by this method. Total nitrogen. — The Gunning modification of the Kjeldahl method was employed for total nitrogen. If the culture con- sisted of only 50 cc. of medium the entire volume was utilized, otherwise after making the culture up to the original volume 25-cc, or more frequently 50-cc, duplicated samples were run. A small piece of copper wire, 7 gms. of anhydrous sodium sulphate, and 35 cc. H 2 S0 4 were added and digestion continued for one hour after the solution became clear (see Latshaw, '16). Table i indicates the degree of accuracy with which duplicate determina- tions checked. In some of the experiments where aeration was employed the data for nitrogen determinations did not seem conclusive, and it has been thought best to leave them out entirely. No significance is attached to an increase in nitrogen of less than 0.5 mgs. per 100 cc. of medium, and all the data recorded in the tables are based upon 100 cc. I Vol. ir> 120 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE I ACCURACY WITH WHICH TOTAL NITROGEN DETERMINATIONS CHECKED Sample No. Mrs. nitrogen recovered from Peptone solution Peptone solution Azotobacter culture 1 2 3 4 5 6 Average 9.58 9.52 9.61 9.58 9.52 9.52 9 . 55 9.61 9 . 52 9.52 9.58 9 . 52 9.58 9 . 66 2.21 2 . 34 2.14 2.21 2.08 2.27 2.25 Volatile acid. — Practically all the acids used were Eastman Kodak Co, products. Quantitative determinations have been made by distilling 100 cc. from a total volume of 110 cc. Pyrex Erlenmeyer flasks of 300 cc. capacity connected to Liebig con- densers and surrounded an asbestos shield were used as distillation flasks. These were heated by an electric hot plate, and it required 30-45 minutes, depending upon the acid, to dis- til over 100 cc. Titrations were made in increments of 20 cc. unless it had previously been noted that practically all acid had disappeared, in which case the entire 100 cc. were titrated at one time. This fractional titration was employed in order to enable the plotting of the titration curves to detect the trans- formation of a higher into a lower acid. Phenolphthalein was employed as an indicator, and care was exercised that all vessels and wash water were neutralized before being used. Figures 1 and 2 are given to show the relative titration curves of the different normal acids and also to show that there was no indication of an acid with a higher molecular weight being trans- formed into one of lower molecular weight. The curves for the standard acid solution and for the cultures in which abundant growth had taken place coincide as well as would curves from two different batches of acid. The culture distillation curves are from cultures in which approximately half the original acid had disappeared. Curves for iso compounds would show the same thing. These curves are plotted on a basis of the per cent of the total recovered that came over in each 20-cc. fraction, when 100 cc. were distilled from a total volume of 110 cc. 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 121 Calculations of the quantity of acid present were based upon quantitative distillations of carefully standardized acids distilled from pure water to which a small quantity of sulphuric acid had also been added. The data in table n show the per cent t o u • c! o Fig. 2. Distillation curves for volatile acids from standard solutions and from cultures in which approximately half of the acid had been metabolized. recorded. When distilled from the culture medium the volatile acid was freed from calcium by adding an excess of sulphuric acid. The C0 2 thereby liberated was removed by aerating vigorously for thirty minutes. In the aerated cultures, run for the longer periods of time, 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 123 a < o o o < O o w hH* CO Q w w CO Q •J g CO O P8 < O « c o n3 CD CD ► O o (-1 C ftH CD Ph I o o 3 a c3 CD o Ht3 o3 p4 T3 CD o3 o Hi O o3 o CO +3 5 03.2 d o o Ph o CD O s o oo*ocooo I>i-hNC0^ ON ONO »ON00OiOi iO CM CM CM iO CM CO CO CD T3 0> OOOOO "tf r* N oo oo WOSiOCO WO5 00ON CO^ONOiOi P3 < NiOOiN^ *OCM CO »OGO CM Tt< *O00 COOCMOCM CM^cO CM^ CO GO w Q o CO Q O o CO O H - H CO a a O Q CD o o T3 ol O C3 OP CD Ph cd • I— « cd C3 > o t". p< O I"* )- o £h > (NW^NO ONhOOJ iC i> a a o c3CO h« C5kO 00^ CO 00 lONOiOOS 03 C^ E ooooo Tt< i—l O C5 l> O 00 t-i CO GO tQ t"- Ci C5 C5 02 S o OO o C^ t^ GO CO o »oo CO OtNWNN CN Tt< CO N 05 ^ CO fe ooooo CO »0 CO cO (N i-H GO CO (M rf CO r- Oi CO o3 (N s coc^ "^ -^ i-H 00 CO CO o3 E GO cO^ CN < N O O !>• CO (Nre no doubt principally due to evaporation but probably degree to the mechanical removal of moisture due to the bursting of so many bubbles. Owing to such discrepan has been difficult to evaluate accurately the utilization c by the growing cultures in certain instances. We have fore recorded as questionable such losses in the aerated c acid Q possibl exceeded such losses where observ been thought best to check ime would not permit accurate determinati ion of the hydrogen order to determine if the made of brom table for growth. For this purpose use has been brom-thymol-blue red red, phenolphthalein, and thvmol-bl Since the medium was only tested roughly as to whether it was acid, alkaline, or neutral to those indicators within whose range its reaction lay, the figures recorded in the various tables are merely approxi- mate. Only in those instances where the medium was found to be alkaline to thymol-blue and is recorded as 9.0 + has there been any indication that the reaction was unfavorable. In all 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 125 probability a medium alkaline to thymol-blue has an unfavorable, if not toxic, effect upon Azotobacter (see Johnson and Lipman, '22). Preliminary Experiments A large number of experiments of a preliminary nature have been performed. In fact, before any method, or step in a method, or culture was adopted for experimental use it was carefully tried to see that it would work satisfactorily. Among these preliminary experiments there are, aside from those already mentioned, a number that seem to be of sufficient significance to record. Experiments reported by Hunter ('23) indicated that by draw- ing a current of air through the medium, growth of Azotobacter and fixation of nitrogen could be greatly stimulated. Since the use of any method that would hasten growth seemed desirable, in view of the slow assimilation of certain of the fatty acids reported by Mockeridge ('15), it was thought that Hunter's method might be used advantageously. Therefore, an experi- ment was designed not only to confirm Hunter's results but at the same time to determine whether varying the rate of aeration would influence quantitatively the consumption of the organic food and the fixation of nitrogen. The results of such an experi- ment are reported in table v. No method of measuring quantitatively the rate of flow of air through the medium was available; however, in the samples subjected to "slow" aeration a slow continuous flow of bubbles, perhaps one a second, was maintained. In the "medium aerated cultures the air was drawn through at least ten times as rapidly, while the cultures subjected to "rapid" aeration probably received ten times as much air as the "medium" aerated cultures. The data show very definitely that increasing the rate of aeration increases both the rate of dextrose consumption and nitrogen fixation. There is some indication that the dextrose may possibly be utilized somewhat more efficiently with limited aeration, the average nitrogen-dextrose ratio being 1 : 79 for the "slow" aerated samples, whereas the corresponding ratio for the other samples was 1 : 115 and 1:113 respectively. >> [Vol. 15 126 ANNALS OF THE MISSOURI BOTANICAL GARDEN > CQ < n o W X P o O o o J n p o < o - w n3 d c a; < B cd E a t-. CD o d c x G 8 CD • tG ^ag QO CO iO 4C CO CO o a bf>0 g 00 tJI (N - go t» r>» a r>- d P 3 bcU? 5?2x GO "^ CM GO ^ *C CO CM U0 CO CM CO CD co O s -a CD a X 5 CD T3 d o CD rt< 00 00 CO CM iO 00005 M fa o CMCO^iO CD bC oj CD < 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 127 In this connection it might be well to comment of the term " dextrose Certain investigators in speaking of this relationship have made use of the term " carbon-nitrogen ratio." This term, it is believed, does not accurately express the relationship. If it is a question of the organism securing or setting free a certain quantity of energy per unit of carbon, as is usually considered, obviously the expression C : N ratio is incorrect in that the energy freed per unit of carbon depends upon the other elements combined with the carbon as well as upon the carbon. For exam caproic acid contains approximately twice as much energy per gram of material and one and one-fourth times as much energy per gram of carbon as does dextrose, both being six carbon atom compounds ould seem, therefore, that it would be much more logical to express this relationship upon an energy or molecular basis. In an effort to secure a desirable culture with which to carry out the more extensive investigations recorded in the next part of this paper, the experiments reported in tables vi and vn were designed. All the cultures available at that time were included in these tests. TABLE VI VARIATION IN UTILIZATION OF DEXTROSE AND FIXATION OF NITROGEN BY DIFFERENT CULTURES OF AZOTOBACTER Culture No. 3a 3b 4 5a 5b 6 7 8 Incubation 2 days Mgs. dextrose consumed Mgs. nitrogen fixed 169 197 171 179 169 247 157 148 2.20 2.48 1.38 2.06 2.34 2.34 .82 1.38 Mgs. dextrose per mg. nitrogen fixed 77 79 124 82 72 106 192 107 Incubation 5 days Mgs. dextrose consumed 620 674 522 679 729 692 250 368 Mgs. nitrogen fixed Mgs. dextrose per mg. nitrogen fixed 4.96 125 6.06 111 1.10 474 3.58 189 3.58 203 3.04 227 .82 610 1 66 222 These data indicate that culture No. 3a was approximately as efficient in fixing nitrogen as any other. At the same time 128 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE VII UTILIZATION OF DEXTROSE AND NORMAL BUTYRIC ACID* AND FIXATION OF NITROGEN BY DIFFERENT CULTURES OF A/OToRACTER Culture No. 3a 3b 4 5a 5b 6 7 8 Mgs. dextrose utilized Mgs. nitrogen fixed 057 5.40 057 4.55 84 .50 057 5.40 057 6 20 620 3.02 84 .28 057 3.44 Mgs. dextrose used per mg. nitrogen fixed 177 210 168 177 154 208 300 284 Mgs. butyric acid utilized Mgs. nitrogen fixed 00 98 3 06 00 99 99 1.38 1.76 .00 1.65 1.21 1.38 .00 .82 Mgs. butyric acid used per mg. nitrogen fixed 72 55 58 80 72 121 * Butyric acid neutralized with sodium hydroxide. it grew abundantly, thus enabling ready detection of growth. It was also known to be a strain of Azotobader chroococcum. For these reasons it was temporarily selected for further study. These data also indicate that certain cultures (Nos. 4 and 7), to all appearance Azotoba de oped very poorly in the dextrose and not tit all in the buty d med \\ by the quantity of nitrogen fixed per unit of than material used, the actively growing cultures were apparently capable of utilizing butyric acid much more effectively dextrose. In this particular experiment the increased effe: ness with which butyric acid was used might have been d the relatively low per cent present, It has been frequently observed that pared with dextrose of eriment place per unit of organic material consumed. In this ex] only 0.1 per cent butyric acid was present. Having selected culture No. 3a the next step was U in a preliminary nay with several acids. According, experiment recorded in table vm was arranged. For some the unknown reason this culture was tested ble to utilize acetic vere secured in other experiments. Other cultures having been added to our collection, further qualitative tests were carried out, among which was the protocol 1928J GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 129 TABLE VIII UTILIZATION OF VARIOUS ACIDS BY CULTURE NO. 3A Mgs. recovered per 100 cc. culture solution Acid Controls Incubation period in days Unaerated Aerated 2 5 9 13 Formic Acetic Propionic Butyric Valeric Dextrose 644 1026 1051 1018 920 934 696 977 1001 896 954 716 982 998 994 896 690 659 431 996 984 884 423 635 394 1002 997 875 000 673 264 996 1001 884 000 arranged in table ix. This experiment was also designed to gain some information relative to the effect of the cation as well as the acid radicle. The salts included were the only ones available of those particular acids at that time. The results presented in table ix indicate very strongly that the cation is probably of as much significance in determining the availability of an acid as is the anion. None of the formates permitted growth. This has been characteristic of formic acid in all tests conducted with it and is probably due to the formation of formaldehyde. Aluminum acetate appears to be the most readily available salt of acetic acid tested, all the strains being able to assimilate it readily. Uranium acetate, on the other hand, was not assimilated by any of the cultures. Between aluminum and uranium lay calcium, ammonium, magnesium, potassium, and sodium salts, their availability being approximately in the order given. Ammonium acetate, while apparently assimilated by all ten cultures, supported vigorous growth only in two instances. It is possible that when supplied with nitrogen the very small quantities of organic impurities finding their way into the cultures enabled the various cultures to make perceptible growth. This, though, does not seem probable. Only half the cultures were capable of making visible growth when the sodium salt was the only source of organic matter supplied, and only one out of the ten assimilated it readily. Culture No. 60 could not even metab- olize dextrose readily. This experiment, as well as others reported in this paper, [Vol. 15 130 ANNALS OF THE MISSOURI BOTANICAL GARDEN X! ►J PQ O O < O CO & O fa fa Q w co co>* o o fa fa o CO CO CO O < > fa O O J 8 o a 3« s» CJ 08 3 oa T3 cj O o a 03 ol cu CJ 03 O 8 of 6 3 3 o3 0? CJ O 03 c 1 5 2 «3 0) cj 0$ a> s o E 03 03 £ a> .5 c3 ++++T++ + + + + + I + + + I +++•-+#- + 2 0) +** + + +="* +^ + O- + + + + + + + + + + + + + ++++++++++ ++++++++++ + + + + as 3 c3 n3 S ° n I I I I HZ 03-* oJQC-J l © C<3 IQ «0 iO "3 O (Q CO +J o3 cj -3 3 ,g * 00 * *\ fl J3 +^ GO & o *A X-. i bJD a -2 «+* o 3 ^^■^ o3 C o fc-l o 6 00 3 CD 3 a C 4 a> 3 * 2 83 03 3 o o3 03 o o a 8- CD F OO o o3 a -3 '55 oo S s o3 3 03 o3 ^5 b£ o3 03 '53 cp CJ 00 T5 CJ c3 oo 00 a o o3 03 s o cj a O 0) o a 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 131 indicates a very wide variability in the metabolism of organisms belonging to the genus Azotobacter. The data also emphasize the great need for more specific physiological studies of this very interesting group of organisms. It would appear utterly futile to attempt to apply the findings from the study of one strain or species to any other strain or species. Just as Lohnis and Smith ('16) have pointed out the futility of attempting to apply the morphological findings in any particular medium or at any particular time to the group as a whole, it is well to emphasize the same with regard to physiological studies. Since culture No. 62, apparently a strain of Azotobacter chroococ- cum and hence very closely related to culture No. 3a, seemed from the above-reported experiment, as well as from a number of unrecorded tests, to possess the ability to utilize a wider variety of salts of fatty acids than any other available culture, it was selected for the more intense studies reported in the next part of this paper. Also, even though aluminum acetate was undoubt- edly more readily available to some strains of Azotobacter than the calcium salt, the latter served equally as well for culture No. 62; and since calcium salts have found a much wider use in biological studies than aluminum it seemed desirable to use the calcium salt, thus making any results that might be secured more comparable with those reported by others. In addition, calcium salts are somewhat more easily prepared than aluminum. Calcium was therefore used as the basic element in succeeding studies. The question of the influence of the cation should certainly receive more study, and it is hoped that such studies may be continued in the near future. The data presented in table ix are only indicative of what may be expected. Experiments with Culture No. 62 As previously mentioned, culture No. 62 probably belonged to the species Azotobacter chroococcum. Preliminary experiments indicated that it was a vigorously growing and strong nitrogen- fixing strain when supplied with a suitable form of organic material such as dextrose and certain of the lower fatty acids. The experiments conducted with this culture were all aerated. [Vol. 15 1.32 ANNALS OF THE MISSOURI BOTANICAL GARDEN Aeration was employed because the work of Mockeridge ('15) indicated that the rate at which certain organic materials were assimilated was extremely slow, long periods of incubation being necessary to insure appreciable utilization. Previous work had shown that the rate of growth, consumption of certain sugars, and fixation of nitrogen could be materially facilitated by draw- ing a current of air through the medium. It was hoped, by employing a similar method in these experiments, to shorten the time of incubation necessary to secure quantitative results of a definite character. Increased aeration unquestionably stimulated growth in many instances, but occasionally some difficulty was experienced in obtaining entirely satisfactory checking in quantitative nitrogen and volatile acid determina- tions following prolonged aeration, and in addition it was more difficult to maintain pure cultures. For these reasons the quan- titative experiments in which Azotobacter vinelandii was em- ployed were not aerated. Utilization of formic acid. — Experiments were carried out in which formic acid was used as the sole organic constituent of the medium, but there was no indication of either growth, utilization of the acid, or fixation of nitrogen, and therefore the data are not recorded. Utilization of acetic acid. — The data with regard to the utiliza- tion of acetic acid, recorded in table x, are quite conclusive in showing that the calcium salt of this acid is readily available to culture No. 62. Within seven days practically all the original 1.0 per cent of acid had disappeared, accompanied by abundant growth and a marked change in the reaction of the medium. In fact, it is probable that the hydroxyl-ion concentration was such as to inhibit further growth. Quite marked fixation of nitrogen was evident, bub owing to an error in the method employed in the total nitrogen determinations in this experi- ment, the data are not recorded. Utilization of propionic acid. — The ability of culture No. 62 to utilize readily the calcium salt of propionic acid is quite evident from the data presented in table xi. Within nine days practically all the acid had disappeared, and a marked increase in the hydroxyl-ion concentration and nitrogen content of the cultures had occurred. 19281 GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 133 TABLE X UTILIZATION OF ACETIC ACID BY CULTURE NO 62 Approximate Mgs. acid recovered Flask No. Days incubated Purity reaction expressed as Ph Control 1 Control Sterile 7.0-7.4 1059 2 Sterile 7.0-7.4 1062 3 2 Pure 7.0-7.4 831 4 2 Pure 7.0-7.4 906 5 4 (a)* 7.0-7.4 668 6 4 (a) 7.0-7.4 654 7 7 (a) 8.6-9.0 36 8 7 (a) 8.6-9.0 85 9 9 Pure 9.0 + 39 10 16 Contaminated 9.0 + 48 11 16 Pure 9.0 + 48 12 22 Contaminated 9.0 + 124 13 I ! 22 Pure 9.0 + 91 Mgs. acid utilized 229 154 392 406 1024 975 1021 1012 1012 936 969 Mgs. nitrogen fixed 0,0 o S o O o3 m c3 m a 3 (d o X a> cd o S3 CO CD CQ 55 3 £1 Not tested. Flask No. 1 2 3 4 5 6 7 8 9 10 TABLE XI UTILIZATION OF PROPIONIC ACID BY CULTURE NO. 62 Days incu- bated Control Control Control 18 Control 18 3 . 5 7 9 12 18 Purity Approximate reaction expressed Sterile Sterile Contaminated Sterile Pure Pure Pure Pure Contaminated as P H 7.0-7.4 7.0-7.4 8.2-8.6 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 9.0 + 9.0 + 9.0 + Mgs. acid recovered 958 952 17 974 829 705 626 21 37 17 Mgs. acid utili- zed 132 256 335 940 924 944 Mgs. nitro- gen fixed Mgs. acid used per mg. nitro- gen fixed 1.72 4.74 3.64 4.74 76 54 92 196 In this experiment control flasks Nos. 1 and 2 were not aerated, while Nos. 3 and 4 were aerated under the same conditions and for as lone: a time as any of the inoculated flask The culture adjacent to control culture No. 3 foamed badly, resulting in the contamination of No. 3 with Azotobacter; therefore it is not considered in the quantitative calculations. It is evident, [Vol. 15 134 ANNALS OF THE MISSOURI BOTANICAL GARDEN though, from a comparison of control flasks No. 1 and No. 2 with No that no volatilization of the propionic acid Therefore the deer of acid that ccurred in the presence of pure cultures must have been due o its assimilation by the organisms. Utilization of butyric acid. — The data presented in days table xn show that culture No. 62 is also capable of uti normal butyric acid in its metabolism. Again only seven were required for almost complete assimilation of the acid present, with corresponding decreases in the hydrogen-ion con- centration. The quantities of nitrogen fixed were also marked, as was the visible growth of the organisms. TABLE XII UTILIZATION OF NORMAL BUTYRIC ACID AND FIXATION OF NITROGEN BY CULTURE NO. 62 Flask No. Days incu- bated Control Control 2 2 5 5 7 12 17 17 Purity Approximate reaction expressed as P H Mrs. acid recovered Mgs. acid utili- zed Mgs. nitro- gen fixed Mgs. acid used per mg. nitro- gen fixed 1 2 3 4 5 6 7 8 9 10 Sterile Sterile Pure Pure Pure Pure Pure Pure Contaminated Contaminated 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 9.0 + 9.0 + 9.0 + 978 995 956 878 706 669 439 78 61 111 30 108 280 317 547 908 925 875 .09 1.25 2.73 4.15 2.59 6.77 4.95 6.29 334 86 102 76 212 134 186 140 Utilization of isc ►ns to which the i Under the experimental condi c u icid.— recorded in table xm were subjected No. 62 could iso-butyric acid in its metabolism. It is true that decrease in the quantity of volatile acid present cultures but the quantities were small. Besides, lated sterile aerated controls, No. 2 and No. 3, showed some the the same decrease as same length of time. ulated flask No. 10, incubated the I addition there was no eptible and the Quantities of nitrogen fixed did not exceed the experim Visible growth 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 135 questionable. It seems safe, therefore, to conclude that culture No. 62 could not utilize the calcium salt of iso-butyric acid under the conditions obtaining in these experiments. TABLE XIII UTILIZATION OF ISO-BUTYRIC ACID BY CULTURE NO. 62 Flask Days No. incubated Control 1 Control 2 17 Control 3 17 4 2 5 2 6 5 7 5 8 7 9 12 10 17 Purity Sterile Sterile Sterile Pure Pure Contaminated Pure Contaminated Pure Contaminated Approximate reaction expressed as Ph 7.0-7.4 7.0-7.4 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 Mgs. acid Mgs. acid recovered utilized 831 None 774 None 773 None 886 None 800 None 779 None 778 None 799 None 746 None 758 None Mgs. nitro- gen fixed None None None None None None None None None None Utilization normal valeric acid. — Not only was normal valeric acid not available to culture No. 62 but it actually was sufficiently toxic to kill all the introduced organisms within three days. This is the only instance in which any acid studied, other than formic, has actually killed the culture except when marked change in reaction occurred. There was of course no utilization of the acid or fixation of nitrogen. The slight loss of volatile acid previously referred to is evident in the data presented in table xiv. Utilization of monohydrated valeric acid. 1 — The data presented in table xv with regard to this acid are inconclusive. There is a distinct loss of acid, evidently not due to its removal through aeration, because uninoculated aeration controls No. 3 and No. 4 incubated one day longer than the longest incubated inoculated flask showed no loss in volatile acid. On the other hand, there was no perceptible change in reaction, and the quantities of nitrogen fixed, if any, were too small to detect, there being no 1 Mono- and trihydrated valeric acids were iso compounds made by Merck and Co. [Vol. 15 136 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE XIV UTILIZATION OF NORMAL VALERIC ACID BY CULTURE NO. 62 Flask No. Days incubated Control 1 Control 2 3 3 4 5 11 6 11 7 18 8 18 9 32 10 32 Purity Sterile Sterile Sterile Contaminated Sterile Sterile Sterile Sterile Sterile Sterile Approximate reaction expressed as V H 6.6-7.0 6.6-7.0 6.6-7.0 6.&-7.0 6.6-7.0 6.6-7.0 6.6-7.0 6.6-7.0 6.6-7.0 6.6-7.0 Mrs. acid Mgs. acid recovered utilized 992 None 992 None 988 None 890 None 990 None 913 None 973 None 979 None 935 None 951 None 1 Mgs. nitro gen fixed None None None None None None None None None None TABLE XV UTILIZATION OF MONOH YDRATED VALERIC ACID BY CULTURE NO. 62 Flask No. 1 2 3 4 5 6 7 s 9 10 11 12 Days incubated Control Control Control 33 Control 33 3 6 11 11 18 18 32 32 Purity Sterile Sterile Sterile Sterile Pure Pure Pure Pure Pure Pure Pure Pure Approximate reaction Mgs. acid Mgs. acid expressed recovered utilized as Ph 7.0-7.4 1055 , 7.0-7.4 1058 7.0-7.4 1064 7.0-7.4 1066 . 7.0-7.4 1051 10 7.0-7.4 944 117 7.0-7.4 911 150 7.0-7.4 972 89 7.0-7.4 899 162 7.0-7.4 907 154 7.0-7.4 871 190 7.0-7.4 933 1 28 Mgs. nitro gen fixed None None None None None None None None None None None None difference in the nitrogen content of flasks No. 11 and No, 12 and sterile controls No- 3 and No. 4. Utilization of trihydrated valeric acid. — Here again the quanti- ties of volatile acid not recovered and increases in total nitrogen were not sufficient to be regarded as significant. If the quantities of acid disappearing are based upon aerated control No. 3 no losses are evident. Unfortunately, this sample became con- 1028| GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 137 taminated, and conclusions based upon it would not be entirely- valid. On the other hand, if the quantities of acid recovered from the various flasks are compared with those recovered from controls No. 1 and No. 2, analyzed at the beginning of the experi- ment, small losses of acid are evident. What has been said with regard to losses of acid is equally true of increases in nitrogen. It is preferred, therefore, to regard it merely as a questionable possibility that culture No. 62 assimilates this acid qualitatively. The quantitative utilization of this acid, as well as of the mono- hydrated sample, is certainly small, if it occurs at all, compared to that of some of the other acid studied. TABLE XVI UTILIZATION OF TRIHYDRATED VALERIC ACID BY CULTURE NO. 62 Approximate | Flask Days Purity reaction Mgs. acid Mgs. acid Mgs. nitro- No. incubated expressed recovered utilized gen fixed as P H Control 1 Sterile 7.0-7.4 784 J5 ated Con tro 1 Control 19 Control 19 3 3 7 7 13 13 19 19 Purity Sterile Contaminated Contaminated Pure Pure Pure Contaminated Contaminated Contaminated Pure Pure Approximate reaction ex pressed as Ph 7.0-7.4 9.0+ 7 0-7.4 6 6-7.0 7 0-7.4 7.6-8.0 7 8-8.2 9 + 9 + 9 + 9.0 + Mgs. acid recovered 408 137 359 351 374 337 260 87 75 102 109 Mgs. acid utili- zed 57 34 71 148 321 333 306 21 19 Mgs. nitro- gen fixed 1.02 .98 .50 2.16 2.96 Lost 4.80 Mgs. used acid per mg. nitro- gen fixed 56 34 142 69 104 64 Utilizat >/ acid. — While the data presented table xvin may be regarded as inconclusive, they nevertheless indicate that iso-caproic acid is not available as an organic food for culture No such )sses were equally as marked from the aerated controls as from anv inoculated cu of acid, but l-inoculated, is. Further- Flask No. 1 2 3 4 5 6 7 8 9 10 11 12 TABLE XVIII UTILIZATION OF ISO-CAPROIC ACID BY CULTURE NO. 62 Days incubated Control Control Control 19 Control 19 3 3 7 7 13 13 19 19 Purity Sterile Sterile Sterile Sterile Pure Pure Pure Pure Pure Contaminated ? Sterile Approximate reaction expressed as Ph 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 Mgs. acid recovered 560 533 520 497 526 522 505 517 493 513 518 Mgs. acid utilized c eS a o 00 c cr a o t^ Mgs. nitro- gen fixed 3 o cr .3 1928| GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 139 more, there was no appreciable changes in reaction or detectable increases in the nitrogen content. Summary of experiments with culture No. 62. — Under the experimental conditions to which culture No. 62 was subjected in the experiments herein reported the calcium salts of formic and normal valeric acids were strongly germicidal. Similar salts of acetic, propionic, normal butyric, and normal caproic acids served as readily available sources of organic food. The other salts tested, namely, iso-butyric, mono- and trihydrated valeric and iso-caproic, were either not available or utilized very slowly and in small quantities. Experiments with Azotobacter vinelandii The culture of Azotobacter vinelandii used in these experiments (Culture No. 94) was obtained from N. R. Smith, of the Bureau of Plant Industry, U. S. Department of Agriculture. It grew vigorously upon soil-extract-mannitol agar, less vigorously upon beef-extract agar, and produced the typical green fluorescence upon the former medium. It also grew very abundantly in the liquid medium employed when dextrose, mannitol, and calcium salts of several of the fatty acids were supplied as the organic material . Utilization of formic acid. — There was no evidence either in the qualitative or quantitative experiments of the ability of this organism to utilize calcium formate and therefore the quantita- tive data are omitted. Utilization of acetic acid. — An examination of the data pre- sented in table xix should convince any one of the ability of this culture to utilize readily the calcium salt of acetic acid. A very marked change in the reaction, an almost complete disappearance of the acid accompanied by definite fixation of nitrogen, together with an abundance of visible growth, sub- stantiate the above conclusions. In this experiment the hydroxy 1- ion concentration evidently reached a point that could not be tolerated by the organisms, most of them being dead when the flasks incubated for the longer period of time were examined. Utilization of propionic acid. — Azotobacter vinelandii can readily assimilate propionic acid under the conditions obtaining in the 140 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE XIX UTILIZATION OF ACETIC ACID AND FIXATION OF NITROGEN BY AZOTO BACTER VINELANDII Flask No. Days incu- 1 iated 1 2 3 4 Control Control 4 4 5 9 6 9 7 14 8 14 9 24 10 24 11 33 12 33 Puritv Sterile Sterile Pure Pure Pure Pure Pure Pure Pure* Pure* Sterile Pure* Approximate reaction expressed as P H Mgs. acid recovered 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 9.0 + 9.0 + 9.0 + 9.0 + Mgs. acid utili- zed 876 S59 783 790 667 575 423 389 3 3 5 11 84 77 200 292 444 478 864 864 862 856 Mgs. 1 1 i tro- gen fixed Mgs. acid used per rag. nitro- gen fixed -.06 .20 -.20 .12 50 .96 1.60 3 . 06 2.36 2.22 384 — ■ ' 440 S88 498 540 282 366 386 thin three weeks the The * Very few living organii-ims. experiment reported in table original 1.0 per cent of acid had practically disappeared abundance of visible growth, the marked increase in hydroxyl- ion concentration, and definite increases in the nitrogen content of the cultures are additional proof of the ability of the organism to assimilate this particular acid. TABLE XX UTILIZATION OF PROPIONIC ACID AND FIXATION OF NITROGEN BY AZOTO BACTER VINELANDII Flask No. 1 2 3 4 5 6 7 8 9 10 11 12 Days incu- bated Control Control 7 14 23 23 30 30 39 39 46 Hi Purity Sterile Sterile Pure Pure Pure Pure Pure Sterile Sterile Sterile Sterile Sterile Approximate reaction exj Tessed as P H 7 . 0-7 7 . 0-7 7 . 0-7 7.1+7 9.0 + 9.0 + 9.0 + 9.0 + 9.0 + 9.0 + 9.0 + 9.0 + 4 4 4 4 Mgs. acid recovered Mgs. acid utili- zed Mgs. nitro- gen fixed 1009 999 790 575 16 45 7 10 7 7 4 6 214 429 988 959 997 994 997 997 1000 90S .32 1.46 5.92 6.06 4.46 4.02 3.12 2.36 2 . 22 3.12 Mgs. acid used per mg. nitro- gen fixed 668 194 168 158 224 247 319 422 450 319 Utilization of normal butyric acid.— Some irregularities were exhibited in the growth in different culture flasks containing 1928) GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 141 normal butyric acid and inoculated with Azotobacter vinelandii despite all efforts to make the duplicate flasks as nearly identical as possible. These irregularities are reflected in the quantity of acid unrecoverable, the changes in reaction, and in the increases in nitrogen content as recorded in table xxi. However, these data unquestionably show a ready utilization of this acid by the culture in question. The rapidity of disappearance of the acid is not as great as with acetic and propionic acids, but the increases in nitrogen per unit of acid assimilated are greater. TABLE XXI UTILIZATION OF NORMAL BUTYRIC ACID AND FIXATION OF NITROGEN BY AZOTOBACTER VINELANDII Days t Approximate Mgs. acid recovered Mp. Mgs. Mgs. acid Flask Purity reaction acid nitro- used per No. incu- bated expressed as Ph utili- zed gen fixed mg. nitro- gen fixed 1 Control Sterile 7.0-7.4 806 2 Control Sterile 7.0-7.4 819 ■ 3 9 Pure 7.0-7.4 761 51 .12 424 4 9 Pure 7.0-7.4 750 62 .38 162 5 16 Pure 7.0-7.4 700 112 .76 148 6 16 Pure 7.0-7.4 700 112 .64 176 7 25 Pure 7.0-7.4 499 313 2.62 118 8 25 Pure 7.0-7.4 314 498 5.80 90 9 39 Sterile 9.0 + 46 766 6.12 126 10 39 Pure 8.4-8.8 321 491 3.82 128 11 53 Pure 9.0 + 51 761 6.12 124 12 53 Pure 8.8-9.2 411 401 3.56 112 Utilization of iso-butyric acid. — The data presented in table xxn indicate rather strongly that the calcium salt of iso-butyric acid is not nearly so readily available as is the corresponding salt of the normal acid. Even after forty-nine days over half of the original acid was still present, the reaction had changed only slightly, and the quantity of nitrogen fixed was small compared with that fixed where the normal acid was present. Besides, the visible growth (unmistakably present) was also small compared to that where acetic, propionic, or normal butyric acid was the source of organic food. Utilization of normal valeric acid. — This acid apparently was not assimilated by Azotobacter vinelandii as readily as some of the lower members of the series. However, the figures presented in table xxm show practically complete disappearance in flasks 142 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE XXII UTILIZATION OP ISO-BUTYRIC ACID AND FIXATION OF NITROGEN BY AZOTOBACTER VINELANDII Flask No. 1 2 3 4 5 6 7 8 9 10 11 12 Days incu- bated Control Control 7 7 14 14 17 17 26 26 38 49 Purity Sterile Sterile Pure Pure Pure Pure Pure Pure Pure Pure Pure Pure Approximate reaction ex pressed as Ph 7.0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 7 0-7.4 Mgs. acid recovered 835 846 795 790 753 Lost 700 728 662 634 495 467 Mgs. acid utili- zed 45 50 87 140 112 178 206 343 373 Mgs. nitro- gen fixed Lost Lost .26 .32 .12 .64 .64 .64 1.22 1.14 Mgs. acid used per mg. nitro- gen fixed 334 174 278 322 282 338 Nos. 9, 11, and 12. The more rapid assimilation in these in- stances, though, might have been partially due to the contamina- ting organisms. That there was, however, unmistakable utiliza- tion in the uncontaminated flasks, Nos. 7, 8, and 10, is proved by the decreased acid content, change in reaction, and increase in nitrogen content. Both the total quantity of nitrogen fixed and the relative quantity fixed per unit of acid assimilated were large, the latter being greater than for any lower member of the fatty acid series. TABLE XXI II UTILIZATION OF NORMAL VALERIC ACID AND FIXATION OF NITROGEN BY AZOTOBACTER VINELANDII Flask No. 1 2 3 4 5 7 8 9 10 11 12 Days incu- 1 >ated Control Control 7 7 14 14 26 26 38 38 49 49 Purity Sterile Sterile Pure Pure Contaminated Contaminated Pure Pure Contaminated Pure Contaminated Contaminated Approximate reaction expressed as P H 7 . 0-7 7.0-7 7.0-7 7.0-7 7.0-7 7.0-7 8.8-9 8.8-9 9.0 + 7.4-7 9.0 + 9.0 + 4 4 4 4 4 4 2 2 -7.8 Mgs. acid recovered 892 883 810 835 770 750 197 621 54 414 62 60 Mgs. acid utili- zed 77 52 117 137 690 266 S33 473 825 827 Mgs. ni t ro- gen fixed .70 .70 .96 1.34 6.68 2.80 6.76 6.56 7 65 11.98 Mgs. acid used per mg. nitro- gen fixed 110 74 122 102 102 96 122 72 108 70 mm 19281 GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 143 Utilization of monohydrated valeric acid. — Growth in the pres- ence of this acid was slow, and the total amount appeared to be much less than with the normal valeric or with the acids of smaller molecular weight. This was reflected in the rate and total disappearance of the acid as well as in the quantity of nitrogen i fixed. It is also evident from the data recorded in table xxiv that slight, if any, change took place in the hydrogen-ion con- centration. However, the amount of nitrogen fixed and acid consumed, coupled with the presence of visible growth, show conclusively that this acid can be assimilated by the culture in question, but probably not as readily as the other low molecular weight straight-chain, fatty acids, at least not under the condi- tions of these experiments. TABLE XXIV UTILIZATION OF MONOHYDRATED VALERIC ACID AND FIXATION OF NI TROGEN BY AZOTOBACTER VINELANDII Flask Days • No. incu- bated 1 Control 2 Control 3 9 4 9 5 21 6 21 7 28 8 28 9 38 10 38 11 49 12 49 Purity Sterile Sterile Pure Pure Pure Pure Pure Pure Pure Pure Pure Pure Approximate reaction expressed as P H 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 7.0-7.4 Mgs. acid recovered Mgs. acid utili- zed Mgs. nitro- gen fixed 989 1021 831 830 816 841 685 714 637 712 631 663 192 175 189 164 320 291 368 293 374 342 .64 .76 1.40 1.52 2.62 1.52 2.04 1.66 1.90 2.16 Mgs. acid used per mg. nitro- gen fixed 300 230 136 108 122 192 180 176 196 158 Utilization of trihydrated valeric acid. — W was said with regard to the monohydrated valeric acid also applies to the trihydrated, except that the quantities of acid assimilated and the quantities of nitrogen fixed were smaller. No perceptible in hyd and one-fifth of the acid was not recoverable after seven weeks o incubation. These facts would indicate that this acid is assimi lable by Azotobacter vinelandii with somewhat more difficulty thai either the normal or monohydrated valeric acids. The data ar< presented in table xxv. 144 [Vou 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN TABLE XXV UTILIZATION OF THI II YDRATED VALERIC ACID AND FIXATION TROGKN BY AZOTOBACTER V1NELANDII OF NI- Flask No. 1 2 3 4 5 6 7 8 9 10 It 12 Days incu- bated Control Control 7 7 14 14 24 24 35 35 49 49 Purity Sterile Sterile Pure Pure Pure Contaminated Pure Pure Pure Pure Contaminated Pure Approximate reaction expressed as P H 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.0 7.4 •7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 Mgs. acid recovered 784 788 730 753 729 714 752 708 656 668 632 628 Mgs. acid utili- zed 66 33 57 72 34 78 130 118 151 158 Mgs. nitro- gen fixed .38 .26 .26 .32 .50 .88 1.28 1.22 .76 .64 Mgs. used acid per mg. nitro- gen fixed 148 122 218 224 68 88 116 96 202 246 Utilization of normal caproic acid. — It would appear from the information recorded in table xxvi that normal caproic acid can be metabolized by Azotobacier vinclandii the most readily of any fatty acid tested. Within four days a very heavy growth was evident, half the acid added to the medium had disappeared, the reaction had become alkaline to thymol-blue, and marked fixation of nitrogen had taken place. Within nine days the volatile acid had reached the minimum recorded for any incuba- tion period. Furthermore, the quantity of nitrogen fixed per TABLE XXVI UTILIZATION OF NORMAL CAPROIC ACID AND FIXATION OF NITROGEN BY AZOTOBACTER VINELANDII Flask Days • I ^ - 1 Ap] r No. incu- bated runty e> 1 Control Sterile 7 2 Control Sterile 7 3 4 Pure 9 4 4 Pure 9 5 9 ? 9 6 9 Pure 9 7 16 Pure 9 8 16 Pure 9 9 28 Contaminated 9 10 28 Contaminated 9 11 37 Contaminated 9 12 37 Pure 9 Approximate reaction expressed Ph 0-7.4 0-7 . 4 + + + + + + + + + + Mgs. acid recovered Mgs. acid utili- zed Mgs. nitro- gen fixed 408 381 267 180 39 41 32 28 27 31 35 35 128 215 356 354 363 367 368 364 360 360 3 . 92 4.66 5.86 5.48 6.06 6.62 5.28 6.50 8.34 5.10 Mgs. acid used per mg. nitro- gen fixed 32 46 60 64 60 56 70 56 82 70 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 145 unit of acid assimilated was greater than for other acid tested This would seem to indicate ing the size of the molecule does not necessarily decrease its availability. Utilization of iso-caproic acid. — This acid was, according to the data presented in table xxvn, readily assimilated, though the rate of growth, acid utilization, and nitrogen fixation were not equal to the corresponding rates where normal caproic acid was the sole organic food supplied. Similarly, the ratio between acid consumed and nitrogen fixed was twice as wide as for the normal acid. The iso compound then, it would seem, is not only less readily metabolized but can also not be used as economically as the straight-chain molecule of this acid. UTILIZATION OF TABLE XXVII ISO-CAPROIC ACID AND FIXATION OF NITROGEN AZOTOBACTER VINELANDII BY Flask No. Days incu- bated Purity 1 Control Sterile 2 Control Sterile 3 7 Pure 4 7 Pure 5 16 Pure 6 16 Pure 7 28 Pure 8 28 Pure 9 37 Pure 10 37 Pure 11 51 Sterile 12 51 Pure* Approximate reaction expressed as Ph Mgs. acid recovered 7.0-7.4 417 7.0-7.4 409 7.8-8.2 323 7.8-8.2 272 8.2-8.6 90 8.2-8.6 75 9.0 + 23 9.0 + 22 9.0 + 77 9.0 + 23 9.0 + 17 9.0 + 15 Mgs. acid utili- zed 90 131 323 358 390 391 336 390 396 398 Mgs. nitro- gen fixed Mgs. acid used per mg. nitro- gen fixed .50 180 .90 144 2.28 142 2.16 156 2.54 154 3.18 122 3.70 90 2.80 140 2.54 156 2.68 148 * Only four colonies developed on two plates. Summary of experiments with Azotobacter vinelandii (culture No 94) — Azotobacter vinelandii is capable of growing in a medium containing as the only organic material the calcium salt of the following fatty acids: acetic, propionic, normal butyric, iso- buty normal iso-valeric (monohydrated valeric and trihydrated valeric), normal caproic, and iso-caproic. organism failed to grow under similar conditions in the p of calcium formate. This The rate of growth varies with the different calcium salt added (Vol. 15 146 ANNALS OF THE MISSOURI BOTANICAL GARDEN to the medium, being very rapid with normal caproic and very slow with all of the iso acids tested. When growth takes place there is an increase in the nitrogen content of the medium and a decrease in the quantity of volatile acid. The quantity of nitrogen fixed in the presence of any given acid corresponds more or less closely with the quantity of acid disappearing. The increase in nitrogen per unit of acid con- sumed by the organisms varies with different acids. If only normal acids are considered the quantity of nitrogen fixed per unit of acid decomposed increases as the molecular weight of the acid increases. Growth, disappearance of acid, and increase in nitrogen are not as rapid where iso acids are added to the medium as when the acid is a normal compound. Experiments with Other Cultures Since there were a number of instances in which the results secured with cultures No. 62 and No. 94 did not agree, it seemed desirable to extend somewhat similar tests to other cultures. In order to secure very active nitrogen-fixing strains for these tests the experiment recorded in table xxvni was arranged. It was also hoped that this experiment would show to what extent vigorous growth in liquid media, utilization of dextrose, and fixation of nitrogen could be correlated. Erlenmeyer flasks of 300 cc. capacity containing 1.0 per cent dextrose medium were prepared and inoculated heavily with the cultures indicated. One of the triplicate flasks was used qualitative sugar tests. Growth observations were recorded n the remaining two flasks and after fifteen days' incubation for the quantity of nitrogen fixed was determined. Since the dextrose had completely disappeared in all but and figures represent approximately quantity of nitrogen fixed per 1000 mgs. dextrose consumed. It will be noted that with few exceptions the quantities of nitro- gen iixed do not vary very widely. The smallest quantity fixed in any instance where complete disappearance of sugar had taken place was 2.64 mgs. while the largest was 11.46 mgs. Most of the cultures showed a fixation of about 8 to 10 mgs. per 1000 1928) GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 147 THE TABLE XXVIII ABILITY OP VARIOUS AZOTOBACTER CULTURES TO GROW IN PRESENCE OP, AND TO ASSIMILATE, DEXTROSE, AND THEIR RELATIVE ABILITY TO FIX NITROGEN THE Culture No. 219 187(a) 55 209 C 94 204 220 15 19 209 27 215(4) II 216 144 16 103(4) 226(B) 188 86 97 95 B 7 48 11(2) 56 S-2 194 185 165 . 25 44 I 218 III 215(2) 178 198(1) 15(1) Control Control Control Quantity of growth after varying periods of incubation 2 days ? + + + ++ 4-4- + ? ? ? + + + + ? + + + + + + + ? ? ? + + + + + + + + + + + + + ? ? ? + + ? + ? ? ? + + + 4 days ? + + + + + + + + + + + + + + + + + + + ? 4-4-4-4- 4-4-4- ? 4-4-4- 4-4-4- 4-4-4- ? ? 4-4-4-4- + 4-4-4-4- 4-4-4-4- 4-4-4- 4-4-4- ? + 4- 4-4-4- 4- ? 4- 4- + 4-4-4- 4- 4-4-4- 4-4-4- 4-4-4- ? 4-4-4- 8 days 4- 4-4-4-4- 4-4- 4-4-4-4- 4-4-4- 4-4-4-4- 4-4-4-4- 4-4- ? 4-4-4-4- 4-4-4-4- 4- 4-4-4- 4-4-4- 4-4-4-4- 4-4- 4-4- 4-4-4-4- 4-4-4-4- 4-4-4-4- 4-4-4-4- 4-4-4-4- 4-4-4- 4- 4-4- 4-4-4- 4-4-4- 4-4- 4- 4-4- 4-4-4- 4-4-4-4- 4-4-4- 4- 4-4-4- 4-4-4- 4-4-4- 4-4-H- 4-4- 4-4-4- 4- Presence of dextrose ^ after varying per- iods of incubation 8 days Abundant Abundant Abundant Abundant Abundant :0 Abundant Abundant Abundant Abundant Abundant Abundant Abundant Abundant 14 days Abundant Abundant Abundant Abundant Abundant Nitrogen fixed per 100 cc. medium Mgs. 8.14 10.14 9.16 9.92 10.62 8.00 9.16 9.10 2.64 10.00 9.54 10.06 8.66 10.06 9.04 10.38 2.74 8.92 7.04 8.28 11.46 9.30 7.06 10.70 7.96 10.32 8.92 9.16 6.88 9.80 8.08 10.06 8.94 8.46 8.66 8.84 6.56 6.24 10.62 8.78 7.64 mgs. of dextrose or 100 to 120 parts of sugar consumed for each part fixed. This would indicate that the [Vol. 15 148 ANNALS OF THE MISSOURI BOTANICAL GARDEN food available, provided the particular culture in question is capable of utilizing it, is possibly of more significance in determin- ing the quantity of nitrogen fixed than is the culture. There was, however, a rather marked variation in the rate at which various cultures consumed the dextrose. Another very evident fact is that the quantity of visible growth cannot be taken as necessarily indicating the relative fixation of nitrogen. Cultures Nos. 219, 55, 220, 27, 144, B, 44, and 178 all showed a relatively small visible growth, yet they were among the most active nitrogen-fixing strains. Cultures Nos. 15 and 16 were evidently unable to utilize dextrose very readily. It is possible, though, that they may have fixed as much nitrogen per unit of dextrose consumed as any other culture. From the cultures tested in the experiment just described nine, Nos. 7, C, 165, II, 188, 194, 187, 178, 218, and in addition R, were chosen for a comparative study of their ability to utilize the various acids studied in previous experiments. These particular cultures were selected both because of their active nitrogen-fixing ability and because they were secured from such widely varying conditions. Furthermore, they exhibited rather marked variations in cultural characteristics, indicating that they represented different strains or possibly species. In this experi- ment 100 cc. of the medium were placed in 750-cc. or 1000-cc. flasks, thus giving excellent aeration. The rate of visible growth is indicated in table xxix, while the quantities of acid consumed and nitrogen fixed are recorded in table xxx. The data presented in tables xxix and xxx tend to confirm the previous results secured from cultures No. 62 and 94 in that rather marked variations are exhibited in the ability of various cultures to grow in the presence of calcium salts of different organic acids as the only organic matter present. All ten of the cultures readily assimilated dextrose, though culture No. II probably with somewhat more difficulty than the others. All grew in the presence of the calcium salts of acetic, propionic, normal valeric, normal butyric and normal caproic acids, the latter apparently being more easily metabolized than 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 149 ffl < O CO s WD MO o. o o go Hen me tfta O fe h3 < a i 43 43 O a 0) 72 o I c CO o3 S O 55 o-d ^"3 43 o3 > ^ o3 o3 o CP o3 s C 03 o3 > i O go 2 o3 X2 o3.2 O ^ c3 o ooooo ooooo ooooo Q 03 CD' < C 3 6 + + + + + + + + + + + + + + + +++++ + + + oooo + 5^- e- •» ^- + + + + + + + + + + + + + + 5^. c^- + + + + + + + + + + ++_)_++ + + ++ + + + + c-. c*-» c— O ^-' 0^- •*■ "f" C^* C— O- O- r- 00+ + + o^>*-f -f ♦** oo + + + ++ + +++ +++++ i-r-r-r, J + + + I ++ ++ + + + +++ +++ +++ + + + + as £i oT3 0) 03-O o'd O c3 J2 00 C^ ojxi cd a) 150 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN H < a ^3 pC o o oc.~ o 0*0 8-3 .IS -C OS ago > 6 o a8 oft o & o3 a3 c O 3 05 £ vtf «§ « «< +++ ++++ +++ +++++ +++++ +++++ +++++ + + + + + + o o o o o 3 6 O © © © © © ooo©© +++++ + + ++++ ++++ ++++ + + + + + + + ++++ ++++ ++++ ©ooo© +++ +++++ ++++ + + +++ +++ +++ +++ ++ ++ ++ ++++ ++++ +++++ +tt + + + + © © © © «- ++ ©c— *-*- -j- +++++ 000*- + +++ + ++ + fb« fh» fu» -J_ + + + + +++++ 0©e^ + + + + + + + 4 + + + ++++ + + + + + + + ++++ ++++ ++++ + + + + + + ++ a) Xi OT3 Qj o3 £2 uT3 a> 00 00 oojQ t>T3 > oo 35 OO 13 a? ooo © o • O ©Ooo coco o -3* 9 ° O O fl a a — 03 pD >» >> >» 03 03 o3t3 - co iOOO ^ fe ►* *4-i ^^ V< o3 f3 «3 0J 0) o c3 «! 03 saa 03 CO 00 G G a ooo oo t c8X^ oTJ'c; 00 os te 1 OOO 1928) GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 151 .J < c O S3 < fa O PS fa fa Q p fa Eh CO Q fa O fa O O J b o s I i o CO 03 e o 55 -C 03 > 03 p£3 o3 u 03 CP 03 a O ^ cj 03 03 ► i 8£'3 3 c3 U3 c3.2 O w T3 0> E 5 6 2- o 03 pS o a GQ 03 *o CO is o s '53 T4-4-TT+ oooooooooo ++*. + + + + + •-+-K-t++t + + +^•+4- -t* ♦ti + + + + + + ++ + + + ++ ++ + + + + + + + 4-^4-^ CM OOOiOOh. ac oo c© oo oooooooooo CO 0000000000*000000000 IN-^O 00 CM iO CM i-i^ CM Tt<05i> COCOCMCOCOi-HCMCMCOCO C0^H^CN^C0»O CM CM^O^CDOiCOCOI^-COl^ iocDoor>-Tt<^oot^^>o OO^^CM-^CO^O^C^O NNCDXN^NNNCO WC0COCOC0CMCOC0COC0 i>-o (NCNOiNONO co^rs.o5^t- CM CM CDQiCXN^CDOJOiOC t^ot^^coioooooooo CMCOCOCMCOCMCOCMr-HCO ^•<*r^iOT*oocMTt cC^OCOCOcOcMcOkOCOOO CO-^CM CO^GOCMkOkCO OOOOOi^COiOOpOOOOI^ 000000 X>"^CO»0 » ^-4 ►2 •35 c3 !a a *OT3 Ui » ^* O cap c3 | T? 1 T3 CD aci Tri- a. >> lenc -C c3 > c3 s O P 1 O 3 (?3 JQ o3.^ o s ce §2 Ph o XJ or << 33 0? L- 3 t 91 fc a CD ca CD b£ O b/0 ^ CO^OCOCDiOOOOWX ooocoooo ooocoooo ) cO(N00cN^C000CMtM»O OC0O0C Tt C^COCO*-iC^-QO*OiOiOI>CO*Or^ CO 'N(NC^'-ht^- Tt< Or-HCO^O^-^OOOcD OIhhhiNhin 05 »-h 1^ O CO ^ CO 00OO 10 (N CO CO C^ O O Oi p- CO t^ N GO rf t^ *tf -1* 00 CD CCWCOiOOON^CDOO t>.iOTt<^H^t^GOcot^CNr^ r^ooco»o^HCor^GOiOGOco t^r^coGOaocccDTtHr^^^ CN^OCOiOCO'^CO'^ClCNCO • co 00 CD 10 CO 00 I CO CO tf^U^ig - bC 8 In 2£ CO 1^ 00 050^0- H ^ CO -t CO CO CD 10.42 1 1.45 CO OS CO 00 1^ a bfi 1 a> 01 S a. wig O 2 bca 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 153 any of the others, with valeric ranking next. It is quite evident, then, that increasing the size of the molecule does not decrease the availability. Not one of the cultures seemed capable of growing when the calcium salt of iso-caproic acid was the sole organic compound present. Growth with the calcium salts of iso-butyric or with the two hydrated valeric acids (the only organic materials present) was usually either very slow or absent. Culture No. 188 seemed to utilize a wider variety of acids with more ease than any other, its growth being recorded as -f- after three weeks with all the different acids except iso-caproic. As previously pointed out, the quantity of visible growth is not necessarily associated with the quantity of nitrogen fixed or the quantity of organic material consumed. This is again evident if the visible growth as recorded in table xxix is compared with the quantity of acid utilized and nitrogen fixed as recorded in table xxx. 9 The quantitative data presented in table xxx show very definitely that all ten of the strains of Azotobacter used in this experiment can utilize, as a source of organic food for nitrogen- fixing purposes, all the acids tested with the exception of tri- hydrated valeric and iso-caproic. Cultures C and 218 apparently utilized the trihydrated valeric acid. The iso compounds, though, cannot be assimilated as readily as the straight-chain molecules. Since this experiment could not be repeated no effort will be made to analyze critically the results secured with the individual cultures. It is quite evident that some of the cultures can assimilate certain acids very much more readily than can other cultures. Furthermore, some were capable of fixing very much more nitrogen per unit of acid with a given acid than were others. The milligrams of acid consumed per milligram nitrogen fixed, the nitrogen fixed per gram acid, and the quantity of nitrogen fixed per calory of contained energy for the various acids were averaged and recorded at the end of table xxx. In a general way this summary agrees with the results secured with Azotobacter vinelandii. The quantity of nitrogen fixed per gram of acid metabolized increases as the molecular weight increases. In fact the increase in quantity of nitrogen fixed is more rapid [Vol. 15 154 ANNALS OF THE MISSOURI BOTANICAL GARDEN than the increase in heat of combustion, indicating a more efficient use of the energy contained in the larger molecules. The conclusion seems justified, then, that the calcium salts of acetic, propionic, normal butyric, normal valeric, and normal caproic acids can be very readily assimilated by numerous strains of Azotobacter, while the ability to assimilate iso-caproic acid is rather limited among these organisms. The other acids tested, namely, iso-butyric and iso-valeric acids, while capable of being assimilated by most, of the cultures tested, are certainly not as readily available as are acetic, propionic, normal valeric, and normal caproic. As a further check on the relative availability of the various acids the protocol arranged in table xxxi was carried out experi- mentally. The acids used in this experiment were all from new lots and, because of limited quantities available, only approxi- mately 0.5 per cent concentrations were used. Even with 0.5 per cent, large quantities of heptylic and caprylic acids remained undissolved in the culture medium. Erlenmeyer flasks of 300 cc. capacity containing 50 cc. of the medium and inoculated in duplicate served as cultures. In addition to the fatty acids a number of polyhydric alcohols of varying molecular weights and some of identical molecular weights but varying configuration were used in order to see if similar variability in assimilability, as observed for acids, existed for alcohols. Culture No. 94 again demonstrated its superiority over either No. 178 or No. 218 to utilize a variety of fatty acids, and also to metabolize more readily the alcohols. Both the latter cultures again failed to utilize any of the iso compounds readily and some not at all. These data, secured from entirely different batches of acids, tend to substantiate the observations noted from experi- ments recorded earlier in this paper. One new acid, di-methyl- ethyl-acetic, was added but apparently was not even assimilated by culture No. 94. The same variability as regards the utilization of different acids by different strains of Azotobacter is also evident when the alcohols are considered. Of the eight polyhydric alcohols tested, only sorbitol and mannitol were readily metabolized by 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 155 -3 O o CO & O M O ao Q CO o Si C O c i-9 n O 0!) W 00 • O 3 03 CO* T+I+^+Too+o+o + + o +oo + + ao CO >> o3 + O^+OO + 3 o ao n3 + OCOOOOOO , OO + CO + +_ + o-f ++^* o oooooooooooooo + + w + 00 o 5 00 C0_c3 + ++ + + e- +p- -f 4"0 O © O +© O O +0 © + + + + + T3 3 + + + + +o ! *- ' c- © © © O + + + + + + ©oo+o© + T3 + © Tc- +*- O O O o+ooo T oo + +e<-c^c-000 00 000 +0 • c 3 0+0+++00 ©+©+++ s a o TJ ?TJ c3 £ #3 o '■0 <5 32 ■l-J O « u 03 !2 >►> o3 3 '8 o3 2 CT3 *0 a c o o 03 I* o «3 i O CO *2 i g 2 ° « g «3 c3 o3 it .2 2 < •* a g v § g a^Sj! o o o o.iS-^n o o 156 ANNALS OF THE [Vol. 15 all three cultures, the former more rapidly than the latter. Dulcitol, an isomer of sorbitol and mannitol, apparently cannot be utilized by any of the three cultures. Inositol, a six carbon hexahydric ring alcohol, was readily utilized by culture No. 94, with more difficulty by culture No. 218, and not at all by culture No. 178. The same was true of glycerol. The four carbon alcohol of this series, erythritol, was utilized only by culture No. 94. The two and five carbon numbers, ethylene glycol and adonitol, were apparently not available to any of the three cultures. This fact is rather interesting since Mockeridge obtained the maximum fixation in her experiments where ethylene glycol was the sole source of energy. Apparently the configura- tion of the molecule plays an important role in determining the ability of a given strain of Azotobacter to utilize a compound. Discussion Attention has already been called to the salient facts brought out in the various experiments in the summaries accompanying the individual tables. In the limited discussion to follow an effort will be made merely to correlate these various points with the special object of toying to see if satisfactory answers to the original questions propounded in the introduction can be made. In the first place, it is frankly admitted that certain procedures followed in some of the earlier experiments did not prove as satisfactory as had been hoped. This was true of the aeration in that the quantitative acid and nitrogen data did not, for some unknown reason, check as well as did later experiments. Because of these irregularities as much significance is not attached to the data secured in connection with those experiments as to those in which culture No. 94 was employed. Secondly, such variation in the ability of different cultures to utilize different organic food substances was not anticipated or the experiments would have been confined entirely to identified cultures, thereby making it possible to compare results here reported with results obtained by other investigators. This desirability was realized in time to make use of known cultures in the major portion of these experiments, and data obtained with known organisms are regarded as much more significant. 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 157 In the third place the inadequacy of the data, in many or possibly all instances, from a quantitative point of view, is realized. It is believed, however, that in this respect they are equal or superior to any thus far reported. Where quantitative determinations, such as were employed in these experiments, are necessary the accumulation of mass data is a slow process. The limited data available are presented not as definite proof but rather as indicating certain tendencies, and it is hoped that others may see fit to conduct experiments along similar lines, thus bringing about the accumulation of sufficient data to justify definite conclusions. It is desired to call attention again to, and to emphasize, the danger of applying the findings from a study of one strain of Azotobacter to other strains. If there is any one thing indicated by the experiments reported in this paper it is the marked variability in the metabolism of organic compounds of different strains or species in this group of organisms. This fact makes it rather difficult to compare results that have been reported from one laboratory with those from another, because frequently no indication whatsoever has been given as to the origin or identification of the culture studied. In the future much greater emphasis should be placed upon the identity of the culture being studied. For the reason just set forth it is believed that very little would be gained by a comprehensive review of investigations dealing with the utilization of various organic substances by this group of organisms. A cursory perusal of the literature dealing with Azotobacter will impress one with the great variety of organic compounds that may be assimilated by these organisms. Also it brings out the wide variation in efficiency with which such compounds can be used when the quantity of nitrogen fixed in their presence is the criterion by which such efficiency is judged. To illustrate the points suggested in the preceding paragraph the reader is referred to table xxxn. This table is an adaptation of one recently used by Bonazzi ('26) enlarged to include the work of Mockeridge and a few examples from the data previously presented in this paper. Over fifty separate and distinct organic compounds are here 158 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN CO O < O Q Q Oh o o o H h i o3 a c o3 b£ «-. O E o3 »-. bfi i-i c bC O co <", g bfi o co ,fl T3 s si 11 i3 o3 0> 6 * fl 0) In eg Oh a; H a CO o o CO o o 03 o + E T3 s o o *s bfi o COCN CO o cot>- • • fl *OcOO> CO CO CO OJ ooo -^ CN 00 O ^ io O *0 CO *o r^ or^oo^ r^o ^ T* *o o *o CO CO Tt< Tt< 00 co »o CN O t^ CO CO »-« CO iNOO^ CO CO O CO ooo ^ coo O 00 00 00 t^ CO Tt^ CO *oi^ooooi^Ttia> ^ CO oo o CO Tf^ r-H lO 00 CO *0 CO ooi^oaot^^o r- ^ 00 O *ct< 0) CO o o cp ^ co t • O o3 w » C^3 O o o 3§ 1 CO o a co CO o a.- co O H O o O o o3 XL bC cu fl cu ^^■fl >* >> cu 43 >>!? Swow^wi I o CO 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 159 2 as o3 s O a S3 os & as d b£ O <$ ffl OS a OS p OO CD co r^ o oo t^ ^CJ *OCO(N *-*^ *0 t d '3 M as Ph CO o CO c as a as as 5 S a as d a d :0 «3 o o o3 o CO as co 00 0«MCO(N CN CM 00 WOM CO o o u + CN CO OS > 0Q o3 ^^ as ^ S OS W c3 c3 GQ OS D o CO 03 0j Q) O O O o o3 c3 c3 ^^ • .—4 ^* ^"* • 03 o3 o3 o3 o3 £0000 as .2 i 03 o as as o as as as 03 > I 03 O 03 03 O O o3 03 i i 03 03 • p— « as o3 as '3 o3 5 • fa as as o3 d o3 £_£ as as as as a> as a o

h o3 03 0) 03 a i i d o3 as c3 a a c5 o 'SB i i o3 o3 OO as II d as s* 03 s ^ o « as o3 ^ o a> i o3 O 0£ o as o3 o3 o a O as o3 8 O 03 t o C3 "S CO o as 2 s g 'S c3 as as as as as co o Fh c3 8 o ^ c3 bp o3 L3 as d [Vol. 15 160 ANNALS OF THE MISSOURI BOTANICAL GARDEN recorded as being assimilable by Azotobacter. Among these are representatives of classes of organic compounds possessing, in many instances, very few characteristics in common. Further- more, the compounds here listed are only those reported by five of the many investigators in this field and are by no means intended to represent all compounds that have been tested and found capable of supplying the organic needs of members of the Azotobacter group of organisms. Among the carbohydrates are examples of polysaccharides, gums, and sugars. Among the latter are tri-, di-, and mono-saccharoses as well as hexose and pentose sugars. Alcohols are represented by mono-, di-, tri-, tetra-, and hexa-hydrox compounds, also by straight-chain and iso arrangements of the carbons, with the additional variations in polarity of inactive, dextro-, and levo-rotary molecules. There are also twenty-five salts of organic acids, including representa- tives of a number of dissimilar groups. There would seem to be no question, then, but that among the species of Azotobacter there are members capable of utilizing a very wide variety of organic compounds as the only organic requirement for the fixation of nitrogen. Not only is this true but the same strain of organisms can function as a nitrogen fixer when supplied with a wide variety of compounds. When it comes to the comparative efficiency with which these various compounds can be used, measured by quantitative gains in nitrogen, the data are too inadequate to permit of drawing any very definite conclusions. The work with impure cultures would have to be eliminated from consideration. This leaves only a few compounds that have been tested by two or more investigators. Of these ten may be selected that were studied in common by Hoffmann and Hammer ('10), Krainsky ('08), and Mockeridge ('15). The quantity of nitrogen fixed per gram of material as reported by these investigators is indicated in fig. 3. The data plotted may not be very accurate, since there is no indication of quantitative determinations of the residual organic material having been made, except that Mockeridge states that qualitative tests showed complete absence of the organic com- pound. Furthermore, one is forced to assume that the original 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 161 1 .0 per cent of material added remained quantitatively unaltered during the process of sterilization, a condition that certainly mieht not obtain this connection it is believed that the policy followed in experiments herein reported, of making quantitative determinations on controls submitted to the same treatment as the cultures, is a much safer procedure for ascertaining the quantity of the organic compound available to the organisms. Fig. 3. different fixation numerous fig. 3 show a marked qualitative similarity. The absolute quantities of nitrogen fixed per gram of material vary very widely. This might be taken to indicate a variation in the efficiency of the cultures, but, as mentioned in the preceding paragraph, it may merely mean that the organic material was not used up quantitatively in, for example, Krainsky's experiments. If it is assumed that in all instances approximately the same quantity of organic material was available, and that it was used up quantitatively, then we would be justified in concluding that there are wide variations in the efficiency with which different [Vol. 15 162 ANNALS OF THE MISSOURI BOTANICAL GARDEN cultures utilize the same compounds and also a marked contrast in the efficiency with which the same culture used different organic materials in the fixation of nitrogen. There are strong indications, however, that different cultures may utilize many of the same compounds with approximately the same relative efficiency. Attention was called to this in connection with the experiment reported in table xxm, in which out of forty cultures tested the quantity of nitrogen fixed, when dextrose served as the organic food, did not vary very widely in most instances. The work here reported has been confined almost entirely to the salts of fatty acids, principally calcium salts. And attention may be called again to the possible marked effect that the cation may have upon the availability of the fatty acids, as indicated in table ix. It is necessary of course to use the salt of any acid to be tested in order to avoid the inhibiting effect of the hydrogen-ion concentration produced by the free acid. The only previously reported work with which these data can be compared directly is that of Mockeridge, recorded in table xxxii. That portion of Mockeridge's data dealing with fatty acids has been recorded in table xxxm parallel with a summary of the data presented in this paper. In this table are recorded in parallel columns the molecular weights, heat of combustion, nitrogen fixed per gram of acid consumed, and nitrogen fixed per calory of heat energy contained in the material consumed. Qualitatively, the only point of difference between the data presented here and those reported by Mockeridge is in the utilization of formic acid. In no instance have either quantita- tive or qualitative evidence of the growth of any culture of Azotobacter in the presence of a salt of formic acid been observed in this work. Quantitatively, the data agree in showing that as the molecular weight or heat of combustion increases, the quantity of nitrogen fixed per unit of acid consumed increases. This is brought out clearly in fig. 4. There is, however, a difference, possibly significant, in the type of curves plotted from the two sets of data as shown in fig. 4 where the nitrogen fixed is plotted against molecular weight, or heat of combustion, since both increase arithmetically in the compounds as arranged in the figure. 1928] GAINEY — SOURCES OF ENERGY FOR AZOTOBACTER 163 H n < c o < Ph Q <4 few CO HO HHfcj 63 a Kg KO w O a o 2 a 03 bfi b£ H o 03 OO^IN 10 0) bfi s 0) bfl si b£ 3 03 o t^l>COCO CO*Q CD CO bfl Is CO bfl o 1-4 03 •a Tt^T^C000OOO(NOC0 CN (M I bfl °_- S x bX) s t-4 bfl 03 2* COCDiONX^)O^»O00 OC^iOXONCO»ONNX d o 0'43 * g o Cj o O 03 s c3 ;- b£ 0) TfOI>CDCCCDCD0000cDcD C0^iOC0 O t^rfTtiT^CDC^'^COcO(NC^ COCO(NXCOCSOOOO^)N 1 03 03 1 ^v>4 a o ... o * p— * H o 2 *-. 03 O 03 M ^ o3 03 O c3 > I c3^ a in l O

< 03 bO o3 %-> c bC § o3 E ++ 000 * 00 03 2 o3 O a c o3 C3 03 03 CO ^3 03 bJD •s 03 03 o o- 03 0> o 03 a o w ? s O 03 o CO 03 O) O T3 OJ ^c3 13 03 164 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN This difference is shown nitrogen fixed is plotted the compound. trikingly » in fig. 5 in which the of enerev contained in ^ IE v. .010!) 0091 .0107 H .074 .014 .0107 or (-)t .009 (b) Cucumis Con. G . 1 247 .045 .0135 . OOfiS Series I F . 1226 (c) Phaseolus Con. G .045 (-) 0126 093S . 03S4 Series ] F (d) Ipomoea Con. G 0110 009S .0814 .1565 . 0296 (-)or .0114 .0102 040s 0243 . 0239 Series I F' 012 Oil H .117 .117 034 .029 (-) .0145 Leaf Pali- sade Upper epid. Lower epid. Nicotiana (0 Bryophyllum (g) Raphanus Con. Series I Con. Series I Con. Series I G H G H G F' .143 . 0525 .0218 .011 .1326 . 0529 (-) .0112 .460 (-) .0153 .0121 .4338 (-) .0145 .1604 .0495 .0142 .0151 .1612 . 0635 (-) .0123 (h) Raphanus Series 1 1 Series III A B C D .1293 1582 .158 .149 0195 049 047 .0411 014S .019 015 .0134 01176 .0103 009 1 .0117 Con. E .163 . 038 .013 .007 1928] ELTINGE — EFFECT OF ULTRA-VIOLET RADIATION 201 (i) Lactuca (j) Bryophyllum * Series II Series III Control Series II Series III Control A B C D E A B C .4464 .0202 .0126 D E Leaf Palisade Upper epid. Lower epid. .0884 .0182 .0128 .0078 .087 .0207 .0154 .0109 .0896 .0204 .012 .0092 .0882 .0176 .0134 .0089 .1013 .0263 .0154 .0126 .360 .0153 .0121 .522 .0189 .0099 .8075 .0153 .0144 .5103 .0207 .0121 * F', plants rayed 8 weeks at 100 iches without a screen. t (— ), lacking. Leaf Palisade Upper epid. Lower epid. (k) Cucumis Series II A .1307 .0448 .0134 .007 B .1397 .0459 .0096 .0117 Series III C .1363 .0454 .0117 .007 D .127 .0453 .0133 .0071 Control E 1257 0484 0126 0072 (!) Coleus Series II Series III A .1181 .0364 .021 .0086 B 1056 0299 0184 0078 C 1232 037 0159 0103 D .1374 .0371 .0156 .007 Control E .138 .0375 .0168 .0112 (m) Phaseolus Series II Leaf Palisade Upper epid. Lower epid. A .1369 .0518 .018 .012 B Series III C .1016 .038 .0111 .0069 D Control E 0915 036 010 0067 (n) Zea Mays Series II A 1408 0263 0204 B .1402 .0260 . 0304 Series III C .1467 .020 .026 D .1366 .0296 .0196 Control E .1195 . 0274 .019 (o) Nicotiana (p) Ipomoea Series II Series III Control Series II Series III Control A B C D E 1 A B c D E Leaf Palisade Upper epid. Lower epid. .143 .0525 .0218 Oil .143 .0498 .019 .0112 .1444 .0362 .0148 .0142 .1584 .0487 .0162 .0086 .145 .0414 .0207 .0103 .1206 .0355 .0179 .0168 .1184 . 0375 .0193 .017 .1316 .0467 .0188 .0176 .168 .0604 .0243 .0212 .133 .0476 .0226 .017 Stems. — Cross-sections from the base of non-rayed Cucumis stems seven weeks old were found to have smaller diameters than any of those from rayed stems. Similar sections from the base of Cucumis stems in Series II A and III D were found to be a little broader, while those in Series II B and III C had the [Vol. 15 202 ANNALS OF THE MISSOURI BOTANICAL GARDEN greatest diameter (table vm). Series II B had the thickest leaves and the greatest growth in height. All rayed Cucumis stems in Series II and III also showed larger bundles than control stems, though there was little difference between the different rayed groups. The average width of the rayed bundles from the outside of the stem toward the center was 0.54 mm. and that of similar control bundles was 0.36 mm. The amount of bast tissue was about the same for control as for rayed stems. Cross-sections from the base of stems of control Nicotiana plants were also found to be smaller in diameter than most of the rayed ones. Series II A developed stems a little smaller in diameter than those of the control plants. Stems in Series II B and III C were larger in diameter, and those in Series III D were still larger. It will be remembered that this was the group of tobacco plants that showed the thickest leaves and the healthiest appearance (table vm). As to development of vascular tissue the control plants again had the thinnest vascular cylinders which were 0.36 mm. in thickness. Next in order of thickness came Series II A (0.378 mm.) followed by Series III C (0.505 mm.) and III I) (0.54 mm.). The tracheae were also smallest in the control plants and largest in Series III D with Series II A, B, and III C intermediate and equal. In all cases the walls of the tracheae were thicker in rayed plants than in non-rayed ones. When sections from the bases of rayed Zea Mays stems of the different series were measured, it was found that those in Series II A were a little smaller in diameter than those from the control stems. Series II B had stems a little larger and III D still larger than those in Series II B. Stems in Series III C were the largest of all. This was also the group of plants that showed the thickest leaves and the greatest growth. There was a noticeable range of variation present in the vascular bundles of the different groups of Zea Mays plants. Rayed stems in Series II A showed bundles smaller than those of the control stems, both as to entire bundle and as to size of vessels, but the phloem was better developed than in control stems. Series II B showed bundles of about the same size as control stems, but here the phloem was as well developed as in Series II A and the mechanical tissue much better developed than in 1928| ELTINGE — EFFECT OF ULTRA-VIOLET RADIATION 203 either the control stems or those in Series II A. The best- developed bundles were found in Series III C. Here the phloem and mechanical tissue were very well developed. The walls of the vessels were much heavier here than in any other group of Zea Mays plants. The pith cells in the control plants and those in Series II A were angled, while those in Series II B and III C and D were oval in shape, showing more air-spaces. The oval pith cells were also larger than corresponding angled ones. Series III C was also found to have many more layers of cells making up the cortex than any of the other groups of Zea Mays plants (pi. 33, figs. 1-8, and table viii). When sections of Coleus stem from the different groups of plants were measured, it was found that the control plants again showed smaller stems than any of the rayed plants. The stems largest in diameter were found in Series II A and B. It was these two groups also that showed the greatest growth in height (table viii). The radial diameter of the vascular bundles of the control stems was 0.495 mm., while that of Series III D was 0.612 mm., that of III C, 0.62 mm., and of Series II B, 0.675 mm. This corresponds well with the fact that greatest growth in height was found in this group. Bast tissue was present about equally in all rayed and control stems of Coleus. Sections of Phaseolus also showed the control plants to have stems smaller in diameter than any of the rayed plants. Series II A has stems having the greatest diameter. It will be remembered that this group of Phaseolus plants also showed the thickest leaves. Series III C had stems just a little smaller in diameter than those in Series II A (table vni). When the TABLE VIII SHOWING THICKNESS IN MILLIMETERS OF RAYED AND NON-RAYED STEMS Plant Series II Series III Control A B C D E Cucumis Nicotiana Zea Mays Coleus Phaseolus 5.0 8.8 8.0 6.0 3.8 5.5 9.5 8.8 6.0 5.5 9.5 10.7 5.5 3.5 5.0 10.0 9.5 5.5 4.0 8.5 8.5 4.8 3.0 [Vou 15 204 ANNALS OF THE MISSOURI BOTANICAL GARDEN vascular cylinder in the different rayed groups was compared it was found to be by far the thickest in Series 1L A and III C, averaging 0.54 mm. in diameter, while that of the control stems averaged 0.49 mm. PHYSIOLOGICAL Chlorophyll decomposition. — A medium green 80 per cent alcoholic chlorophyll solution was made from Nicotiana leaves and put in test-tubes of pure fused silica. These were placed horizontally in white dishes and exposed to the different con- ditions, with results given in table ix. TABLE IX SHOWING THE AMOUNT OF TIME NEEDED TO DECOLORIZE CHLOROPHYLL SOLUTION Sunlight in greenhouse Sunlight outside Diffused light in greenhouse At 30 inches from an unscreened lamp in diffused light Ultra-violet lamp screened with vita glass plus diffused light Sunlight outside under a screen of vita glass Sunlight outside under a screen of quartz-lite glass 12 4 35 39 40 6 5 min. min. min. min. min. min. min. 12 m. () min. 2 min. 18 min. 20 min. 22 min. 3 min. 2J/> min. These results show plainly that ultra-violet rays do not hasten the decomposition of chlorophyll. Vita glass is thicker than quartz-lite, and hence used at close range the difference in thickness would account for the longer time required for the decomposition under vita glass in sunlight. Starch storage. — Sections of Coleus stem at the end of seven weeks showed more starch in control plants than in any of the plants rayed as in Series II and III. It was impossible, however, to distinguish between the different rayed stems. In sections of Phaseolus stem starch was present in the cortex of plants rayed as in Series II A, while similar control plants showed very little if any starch in the cortex. Determination of P n . — Lactuca plants under experiment for eight weeks as in Series I F were used for this work, the leaves and stems being determined separately. The P H of the leaves and stems of both rayed and control plants was found to be 6.0. The P H of leaves and roots of Raphanus was determined sepa- 19281 ELTINGE — EFFECT OF ULTRA-VIOLET RADIATION 205 rately, and here again the rayed and control plants responded alike, that of the leaves being 6.2 and the roots 6.0. Thus raying with ultra-violet rays seems to have no effect on the P H of plants. Dry weights.— In the experiments carried on in 1926 to 1927 as described in Series I, which consisted of plants rayed with an unscreened lamp, dry weight determinations were made of the entire tops of Lactuca and both tops and roots of Raphanus. In all cases greater dry weight was was found in the This can be well seen in the results in table : (a) and (b) SHOWING TABLE X IN GRAMS THE DRY WEIGHT OF PLANTS IN (50 INCHES), F (100 INCHES), AND G (CONTROL) SERIES I H (a) Raphanus 4 weeks H Tops Roots 36 41 G 69 79 4 weeks F .326 .126 G .49 .174 8 weeks (b) Tops of Lactuca plants 2 leaves H G 4 weeks 8 weeks 018 003 .244 .770 9 leaves 2 leaves H G 805 825 2.22 3.85 F 116 200 G .24 1.15 9 leaves In the experiments carried on in 1927 to 1928 the dry weight was determined for fifty grams of wet weight of leaves. Rayed Zea Mays plants of Series II and III showed greater dry weight than corresponding control plants. Series II A showed the smallest dry weight of the rayed plants which was where the poorest growth in rayed plants of Series II and III was found. Lactuca plants in Series III D had the greatest dry weight. Those in Series II A and III C showed smaller dry weight than the control plants. Ipomoea plants in Series III D had the greatest dry weight. It was also in this group that greatest growth and thickest leaves were found. Nicotiana plants showed greatest dry weight in the control plants and the smallest in Series II A. [Vol. 15 206 ANNALS Cucumis plants had the smallest dry weight in Series II B, and in this group were the greatest growth and thickest leaves with the largest air-spaces. Phaseolus plants in Series II A showed the greatest dry weight and also the thickest leaves. Plants of Raphanus showed the greatest weight in Series III C and D and the next greatest in the control plants. Bryophyllum plants also had the greatest dry weight in Series III C and D, though all rayed plants in Series II and III had weights than similar A comparison of dry weights of the various plants will be found in table TABLE XI SHOWING THE DRY WEIGHT IN GRAMS PER FIFTY OKA MS OF WET WEIGHT OK PLANTS IN SERIES I (RAYED WITHOUT A SCREEN), SERIES II (SCREEN OF VITA GLASS) AND SERIES III (SCREEN OF QUARTZ-LITE GLASS) l'lant Zea Mays Lactuca Ipomoea Nicotiana Cucumis Phaseolus Raphanus Bryophyllum Series I H 6.0331 3.1950 2 . 8295 Series II A 5.462 2 . 2307 7 . 4775 5.3150 4 . 8072 6 . 9050 3 . 8845 3.5100 B 6 . 1 496 2 . 8396 7 . 0642 5.5130 4 . 5494 3 . 9775 3 . 4055 Series III C 5 . 8290 2 . 3322 7 . 6735 5 . 9240 4.7149 5 . 3994 4 . 3200 4 . 2975 D 5 9232 3 . 2320 7.7290 5 . 9675 4 . 7544 4.3410 3 . 7585 Control E 5 . 2030 2 . 7959 7 . 3554 6 . 3225 4 . 8420 6 . 3695 4.1361 3 . 52S6 Ash determination. — The results were not conclusive, but they >int toward an increase in ash in plants rayed with an unscreened lamp In plants rayed with a screened lamp the ash was, in the majority of cases, less than in the control plants. In Cucumis showing best growth there was less ash and also smaller explained dry weight than in the control plants. This can be however, by the presence of many large air-spaces in those leaves while in the control leaves the air-spaces were smaller. In Phaseolus the amount of ash again corresponded very well with the dry weights, there being the greatest dry weight where there was the greatest amount of ash. This also corresponded with the thickness of the leaves. For comparison of the results see table xn. 1928) ELTINGE — EFFECT OF ULTRA-VIOLET RADIATION 207 TABLE XII SHOWING THE WEIGHT IN GRAMS OF ASH FOR 3 GRAMS OF DRY LEAF MATERIAL Plant A B C D E F Zea Mays .310 .272 .269 .295 .325 .349 Lactuca .600 .610 .611 .581 .595 .565 Ipomoea .410 .302 .395 .392 .505 Nicotiana .527 .533 .505 .515 .570 Cucumis .562 .521 .505 .530 .599 Phaseolus .464 .435 .455 Raphanus .638 .630 .643 .558 • .789 Bryophyllum .564 .5193 .482 .430 .540 .610 The effect of ultra-violet radiation transpiration. — When leaves of Phaseolus, Cucumis, Lactuca and Coleus were placed in bottles of water, sealed with paraffin, and rayed at 50 inches from the unscreened lamp, it was found through weighings taken every 30 minutes of bottles and leaves combined that at first the rayed leaves lost as much as the controls. Then there was a time when less weight and sometimes no weight was lost by rayed * leaves. After that there was a loss equalling that of the control leaves kept in darkness or in some cases surpassing it. When the stomata were examined at the end of three hours those in the rayed leaves were found to be closed, while those in the leaves kept in darkness were partly open. When the rayed and were weighed at the beginning and end of the experiment, found that all the rayed leaves had lost weight while the had remained It will be noted Coleus behaved a little differently than did the other leaves used This might be explained by the fact that Coleus has stomata or the under surface only, times with the different experiment was repeated The petioles of leaves of Coleus Blumei var. V erschaffeltii were paraffined and the leaves placed in a horizontal position, some being rayed on the upper side, some on both sides, and others placed in darkness, Those in darkness and those rayed on the upper surface were partly wilted at the end of twelve hours while those rayed on both sides were still turgid. When weighed, however, the leaves rayed on both sides and those upon one side only were found to have lost much more weight than the leaves kept in darkness (table xm, and pi. 26, fig. 3) . [Vol. 15 208 BOTANICAL seo\us a: Cuc 2 A B C 2> 3 4 ELTINGE— EFFECT OF ULTRA-VIOLET RADIATION [Vol. 15, 1928 222 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 24 (SERIES II AND III) Fig. 1. Cucumis sativus. A. Plant not rayed. B. Plant rayed for seven weeks at 100 inches, using a screen of quartz-lite glass. C. Plant rayed for seven weeks at 50 inches, using a screen of quartz-lite glass. D. Plant rayed for seven weeks at 100 inches, using a screen of vita glass. E. Plant rayed for seven weeks at 50 inches, using a screen of vita glass. Fig. 2. I pomoca Batatas. A. Plant not rayed. B. Plant rayed for seven weeks at 100 inches, using a screen of quartz-lite glass. C. Plant rayed for seven weeks at 50 inches, using a screen of quartz-lite glass. D. Plant rayed for seven weeks at 100 inches, using a screen of vita glass. E. Plant rayed for seven weeks at 50 inches, using a screen of vita glass. Fig. 3. Zea Mays. A. Plant not rayed. B. Plant rayed for seven weeks at 100 inches, using a screen of quartz-lite glass. C. Plant rayed for seven weeks at 50 inches, using a screen of quartz-lite glass. D. Plant rayed for seven weeks at 100 inches, using a screen of vita glass. . E. Plant rayed for seven weeks at 50 inches, using a screen of vita glass. Fig. 4. Lactuca sativa. A. Plant not rayed. B. Plant rayed for seven weeks at 100 inches, using a screen of quartz-lite glass. C. Plant rayed for seven weeks at 50 inches, using a screen of quartz-lite glass. I). Plant rayed for seven weeks at 100 inches, using a screen of vita glass. E. Plant rayed for seven weeks at 50 inches, using a screen of vita gl Ann. Mo. Bot. Gakd.. Vol. 15, 192S Plate 21 ■•* W. * * •** A 8 »*., ** J> C ^W****™*™*" I ^t^ ■***— ™W 2 3 4 ELTINGE— EFFECT OF ULTRA-VIOLET RADIATION [Vol. 15, 1928) 224 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 25 (SERIES IT AND III) Fig. 1. Raphanns sativus. A. Plant not rayed. B. Plant rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite glass. C. Plant rayed for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. D. Plant rayed for seven weeks at 100 inches from the light, using a screen of vita glass. E. Plant rayed for seven weeks at 50 inches from the light, using a screen of vita glass. Fig. 2. Coleus Blumei var. "Spotted Gem." A. Plant rayed for seven weeks at 50 inches from the light, using a screen of vita glass. H. Plant rayed for seven weeks at 100 inches from the light, using a screen of vita glass. C. Plant rayed for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. D. Plant rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite glass. E. Plant not rayed. Fig. 3. Cole \is Blumei var. Yerschaffeltii. A. Plant rayed for seven weeks at 50 inches from the light, using a screen of vita glass. B. Plant rayed for seven weeks at 100 inches from the light, using a screen of vita glass. C. Plant rayed for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. I). Plant rayed for seven weeks at 100 inches from the light, using a screen of quarts-lite glass. E. Plant not rayed. Fig. 4. BryophyUum pinnatum. A. Plant rayed for seven weeks at 50 inches from the light, using a screen of vita glass. B. Plant rayed for seven weeks at 100 inches from the light, using a screen of vita glass. C. Plant rayed for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. I). Plant rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite glass. E. Plant not rayed. Aw Mo. Hot. Gard., Vol. 15, L928 ] >LATE 25 ^ e -• -» .m V D E I 2 3 \A ELTINGE— EFFECT OK I'LTRA-YK >U;T RADIATION [Vol. 15, 19281 226 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 26 Fig. 1. Lactuca saliva (fifteen leaves). A. Plant rayed for eight weeks at 100 inches from an unscreened lamp. B. Plant not rayed. Fig. 2. Coir us Blumei var. Versi ha ffeltii. A. Plant rayed for thirty seconds each day at 100 inches from an unscreened lamp. B. Plant not rayed. Fig. A. Leaves of Colrus Blumei var. Verschaffeltii, with petioles paraffined. A. Leaf twelve hours after it had been rayed for 13^ hours on each surface at thirty inches from an unscreened lamp. B. Leaf twelve hours after it had been rayed for 1J/2 hours upon the upper surface at thirty inches from an unscreened Lamp. C. Unrayed leaf after twelve hours. Fig. 4. Zea Mays. A. Plant not rayed. B. Plant rayed for six weeks at 100 inches from an unscreened lamp. Fig. 5. Bryophyllum pinnatum. A. Plant not rayed. B. Plant rayed for six weeks at 100 inches from an unscreened lamp. Fig. 6. Phaseolus vulgaris. A. Plant rayed for four weeks at 100 inches from an unscreened lamp. B. Plant not rayed. Ann. Mo. Bot. Card., Vot.. 15, 1928 Plate 26 A A «*« A e 4 1 5 8 3 G 6 ELTINGE— EFFECT OF ULTRA-VIOLET RADlATInN 228 ANNALS OF THE MISSOURI BOTANICAL [Vol. 15, 19231 Explanation of Plate PLATE 27 Camcra-lucida drawings of equal magnification, using 4-inm. objective and 10 X eyepiece. Fig. 1. Leaf of Lactuca saliva not rayed. Fig. 2. Leaf of Lactuca saliva rayed for four weeks at 100 inches from the light without a screen. Fig. i*. Leaf of Lactuca saliva rayed for eight weeks at 100 inches from the light without a screen. Fig. 4. Leaf of Lactuca saliva rayed for four weeks at 50 inches from the light without a screen. Fig. 5. Leaf of Nicotiana Tabacum rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite gass. Fig. 6. Leaf of Nicotiana Tabacum not rayed. Fig. 7. Leaf of Nicotiana Tabacum rayed for four weeks at 50 inches from the light without a screen. Ann. Mo. Hot. Gaud., Vol. 15, 1928 Plate '27 1 5 2 3 4 7 ELTINGE— EFFECT OF ULTRA-VIOLET KADIATIOX [Vol. 15, 1928] 230 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 28 Camera-lucida drawings of equal magnification, using 4-mm. objective and 10 X eyepiece. Fig. 1. Leaf of Phaseolus vulgaris not rayed. Fig. 2. Leaf of Phaseolus vulgaris rayed for four weeks at 100 inches from the light without a screen. Fig. 3. Leaf of Phaseolus vulgaris rayed for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. Fig. 4. Leaf of Cucumis sativus not rayed. Fig. 5. Leaf of Cucumis sativus rayed for four weeks at 100 inches from the light without a screen. Fig. 6. Leaf of Phaseolus vulgaris rayed for seven weeks at 50 inches from the light, using a screen of vita glass. Ann. Mo. Box. Gaud., Vol. 15, 1928 Plate 28 1 <4 2 5 6 ELTIXGE— EFFECT OF ULTRA-VIOLET RADIATION [Vol. 15, 1928] 232 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 29 Camera-lucida drawings of equal magnification, using 4-mm. objective and 10 X eyepiece. Fig. 1. Leaf of Cucumis sativus unrayed. Fig. 2. Leaf of Cucumis sativus rayed for seven weeks at 100 inches from the light, using a screen of vita glass. Fig. 3. Leaf of Cucumis sativus rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite glass. Fig. 4. Leaf of Ipomoea Batatas unrayed. Fig. 5. Leaf of Ipomoea Batatas rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite glass. Fig. 6. Leaf of Ipomoea Batatas rayed at 50 inches from the light without a screen. Ann. Mo. Bot. Gaud., Vol. 15, 1928 Plate 29 1 Ci 2 3 S ELTTNC.E— EFFECT OF ULTRA-VIOLET RADIATION [Vol. 15, 1928] 234 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation op Plate PLATE 30 Camera-lucida drawings of equal magnification, using 4-mm. objective and 10 X eyepiece. Fig. 1. Leaf of C oleics Blumei var. Verschaffeltii rayed for seven weeks at 50 inches from the light, using a screen of vita glass. Fig. 2. Leaf of Coleus Blumei var. Verschaffeltii rayed for seven weeks at 100 inches from the light, using a screen of vita glass. Fig. 3. Leaf of Coleus Blumei var. Verschaffeltii not rayed. Fig. 4. Leaf of Coleus Blumei var. Verschaffeltii rayed for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. Fig. 5. Leaf of Coleus Blumei var. Verschaffeltii rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite glass. Aw. Mo. Bot. Gabd., Vol. 15, 1928 Plate 30 1 2 <4 3 5 ELT I NC.E— EFFECT OF CLTRA-VIOLET RADIATION [Vol. 15, 1928] 236 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 31 Fig. 1. Diagram of the cross-section of a stem of Coleus Blumei, the stippling indicating the normal distribution of red color. Fig. 2. Diagram of the cross-section of a stem of Coleus Blumei, showing the distribution of red color at the end of the fifth raying at 50 inches from the light without a screen. Fig. 3. Diagram of the cross-section of a stem of Coleus Blumei, showing the distribution of red color at the end of the tenth raying at 50 inches from the light without a screen. Fig. 4. Cross-section of a leaf of Coleus Blumei, showing normal distribution of red color. Fig. 5. Cross-section of a leaf of Coleus Blumei, showing the distribution of red color after ten rayings at 50 inches from the light without a screen. Drawn to the same scale as fig. 4. Ann. Mo. Bot. Gaud., Vol. 15, 1928 Plate 31 I 4 2 5 3 i <: i ;r i n ( ; e— e ffect o f u ltr a- v i < ) let r a d i at i o n i 238 [Vol. 15, 1928] ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 32 Camera-lucida drawings of equal magnification, using a 4-mm. objective and 10 X eyepiece. Fig. 1. Leaf of Zea Mays unrayed. Fig. 2. Leaf of Zea Mays rayed for seven weeks at 50 inches from the light, using a screen of vita glass. Fig. 3. Leaf of Zea Mays rayed for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. Fig. 4. Leaf of Zea Mays rayed for seven weeks at 100 inches from the light, using a screen of quartz-lite glass. Fig. 5. Leaf of Zea Mays rayed at 100 inches from the light for seven weeks, using a screen of vita glass. Mays using Ann t . Mo. Bot. Gaud., Vol. 15, 1928 Plate 32 1 2 5 e ELTINGE— EFFECT OF ULTRA-VIOLET RADIATION 240 [Vol. 15, 1928] ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 33 Camera-lucida drawings of equal magnification, using 16-mm. objective and 10 X eyepiece. Corresponding bundles were used in all cases (sixth bundle from the epidermis). Fig. 1 . May, stem. Fig. 2. Mays for seven weeks at 50 inches from the light, using a screen of quartz-lite glass. Fig. 3. Mays for seven weeks at 50 inches from the light, using a screen of vita glass. Fig. 4. Mays for seven weeks at 100 inches from the light, using a screen of vita glass. Fig. 5. Mays weeks at 50 inches from the light, using a screen of quartz-lite glass. Fig. 6. The amount of cortex present in a stem of Zea Mays rayed for seven weeks at 50 inches from the light, using a screen of vita glass. Fig. 7. The amount of cortex present in a stem of Zea Mays rayed for seven weeks at 100 inches from the light, using a screen of vita glass. Fig. 8. The amount of cortex present in an unrayed stem of Zea Mays. Ann. Mo. Bot. Gaud., Vol. 15, 192S Plate '.Y-\ i 3 2 4 6 5 7 8 ELT INGE— EFFECT OF ULTRA-VIOLET RADIATION Annate of the Missouri Botanical Garden Vol. 15 SEPTEMBER, 1928 No. 3 THE PROBLEM OF SPECIES IN THE NORTHERN BLUE FLAGS, IRIS VERSICOLOR L. AND IRIS VIRGINICA L. EDGAR ANDERSON Geneticist to the Missouri Botanical Garden Assistant Professor of Botany in the Henry Shaw School of Botany of Washington University Table of Contents I. Introduction 241 II. Taxonomy and morphology 247 III. Intra-specific variation in Iris versicolor and Iris virginica 258 IV. Hybridization of Iris versicolor and Iris virginica 299 V. Summary 304 VI. Discussion 306 VII. Bibliography 312 I. Introduction The problem of species is a dual one. It asks two questions : What are species? How have they originated? In studying the nature of species we are first of all concerned with the extent to which the species of the orthodox taxonomists reflect actual divisions among plants and animals. Do such species, in the main, rest upon discrete natural groups of individuals or do they represent merely an artificial cataloguing device, ordering up, as best they may, the myriad forms occurring in nature. It might be supposed that upon so fundamental a problem biologists would have come to some working agreement. Even the most casual survey of recent literature will show that this is far from being the case. While several of those who have committed themselves to print on the subject may use such phrases as "modern biologists recognize this to be the case," or, all geneticists agree, etc., etc.," it is a noteworthy fact that no two of them are of the same shade of opinion. At one extreme are those who believe with Lotsv ('16) that, "The species is merely a u Ann. Mo. Bot Gard., Vol. 15, 1928 (241) 242 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 conception of the human mind and a very primitive anthropocen- tric conception in the bargain. * * * The species of the taxonomist is the comparative insignificant rest of large swarms of individuals which arose from the cross of two parents. * * * Species in the present taxonomic sense do not exist." On the other hand, we have such statements as Harper's ('23) "Linnean species do, by and large, constitute recognizable groups of more or less freely inter-breeding individuals. Only extremists deny the possibility of segregating and recognizing such units." The disagreement as to species is just as extreme in regard to their size as in regard to their nature. If groups corresponding to our species do occur in nature, do they fit most readily into the species of the orthodox taxonomists (what we may call Linnean species for want of a better name) or would some other unit better express the relationships which we actually find among indivi- duals? What are we to do with the vexed question of true-breed- ing sub-groups (variously termed micro-species, Jordanons, iso- reagents) whose existence has been demonstrated for numerous Linnean species? Are we to retain them as significant sub-groups within the Linnean species, are we to treat them as of even greater importance, or are we to cast them out altogether? It should be a relatively simple matter to answer these questions for any one species. A comprehensive survey of variation within the species over its entire range would show whether it were an independent unit or whether it merged into other forms ; whether there were recognizable sub-groups within it; and what inter- relationships obtained between the individuals which go to make up these sub-groups and the species itself. However, such an intensive study should be able to contribute to other questions besides the nature of species. It should offer valuable evidence for the more popular question of their origin. During the last quarter of a century an attempt has been made to study evolution experimentally. In the course of these ex- periments two types of mutations have been observed in our laboratories and breeding plots. They have been thought to be the same sort of changes which, operating in the past, have brought about the evolution of our present-day forms. The first type of mutation has been shown to be due to changes which take 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 243 place at a particular point in the germ-plasm. Since only a single gene is affected, such changes have been called gene- mutations. The second type of mutation has been shown to be due to re-alignments of the germinal material, to duplications of chro- mosomes or whole sets of chromosomes. The morphological results of these two types of mutation are quite dissimilar. An intensive and extensive survey of variation within a single species should therefore be able to demonstrate which of these two processes has been most active in causing progressive changes within that species. Such a study should, in other words, enable us to evaluate the evolutionary importance of gene-mutations and chromosomal re-alignments. The present study is just such an intensive and extensive survey of two closely related species. It is an attempt to present a fairly complete picture of the variation within two natural groups of individuals over their entire range. It has as yet been almost purely morphological in scope. Though the morphologi- cal differences between individuals and groups of individuals undoubtedly rest upon basic physico-chemical ones, our knowledge of these physiological differences is as yet too incomplete, among the flowering plants, to possess much phylogenetical significance. So far as is known, no such complete survey of variation within a species or group of species has ever before been made with plant material. It is an ambitious attempt for a single individual unless the problem be made as simple as possible. If we are to learn anything about the ultimate nature of species we must first of all reduce the problem to the simplest possible terms and study a few easily recognized, well-differentiated species. The group to be studied should therefore possess few sub-groups and inter- grading forms. Unfortunately those species which have been selected for intensive study in the past have been chosen by reason of their very complexity. The northern blue flags (Iris versicolor of the seventh edition of Gray's 'Manual') were accordingly chosen for the study, prima- rily, because they were a comparatively simple, stable, and well- marked group. They possessed certain other features which have materially reduced the labor involved in locating and study- [Vol. 15 244 ANNALS OP THE MISSOURI BOTANICAL GARDEN ing a large number of individuals. They are common, conspicu- ous, colonial, and perennial. Though the merits of these charac- ters are practically self-evident, it may not be out of place to call attention to the colonial nature of Iris versicolor L. and its close relatives. Plants of these species are seldom found as isolated individuals but usually occur in colonies of a few to several hun- dred plants. Nor are these colonies merely the vegetative off- spring of a single plant as is the case with so many colony-forming species. The individuals vary so strikingly in the size, shape, and color of their flowers, that at blooming time the limits of a single clone can readily be determined. These seldom exceed a few square feet, though exceptionally large clones do sometimes occur, as will be described below. It was the truly colonial nature of the blue flags which was particularly useful in the pres- ent study, making possible the examination of large numbers of plants in each locality with a minimum of effort. An attempt was made to visit as much as possible of the range of the northern blue flags during their flowering season. Numer- ous colonies were studied in detail and measurements made on twenty to fifty individuals of the characters which had been selected for study. Representative plants were sent back from each of the colonies and established in an experimental plot at the Missouri Botanical Garden, where they were made the subject of genetical, morphological, and cytological studies. In 1923 collections were made in central Michigan. In the spring of 1924 a trip was taken through northern Missouri, eastern Iowa, southern Wisconsin, central Michigan, western New York, and central New Hampshire. In 1925 collections were made in central Missouri, northwestern Arkansas, northern Ohio, including the Bass Islands, western and central New York, and southern Vermont. Fruiting material was collected that fall from central and northern New Hampshire and eastern Michigan. In the spring of 1926 a trip was taken through southern Illinois, southern Missouri, eastern Arkansas, western Tennessee, and central Kentucky. In June, collections were made in Missouri, Michigan, northern Vermont, Ottawa, Canada, and at Lake Timagami, in northern Ontario, Canada. In the early spring of 1927 another southern trip was taken, principally 1928 ANDERSON — PROBLEM OF SPECIES IN IRIS 245 to collect material from Mississippi and Alabama. In June of that year collections were made in Ohio, Maryland, Pennsylvania, New York, Ontario, northern Michigan, and Wisconsin. In 1928 Fig. 1. Localities at which I. versicolor and I. virginica have been studied. North and South Carolina and eastern Tennessee were visited in early May. With the exception of the extreme north the terri- tory has now been quite thoroughly covered as shown by the map in fig. 1. 246 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 Extensive collections of flowering material preserved in alcohol and of ripe seed capsules were made at several points by Mr. R. E. Woodson, Jr. and by the writer. The work on the fresh and preserved material has been supplemented by taxonomic studies in the herbarium of the Missouri Botanical Garden and at various other herbaria. ACKNOWLEDGEMENTS It is scarcely possible to carry on a study involving geographical distribution without enlisting the help of a very large number of people. For aid in assembling information and material for study it is a pleasure to acknowledge the cooperation of those whose names are listed below, and of several others who have been inadvertently overlooked : Dr. J. R. McLeland, R. W. Shreve, A. N. Rood, R. S. Sturte- vant, W. M. Boswell, Dr. A. C. Kinsey, Dr. P. C. Mangelsdorf, Prof. J. G. Jack, Dr. J. N. Couch, E. A. Mosley, Mr. Holtiwanger, Dr. 0. E. Jennings, John MacDougal, A. F. Satterthwaite, M. E. Mains, Miss G. V. Butcher, Mrs. Jasper Blackburn, A. C. Anderson, Herbert Stoddard. Miss Katherine Perkins, Mrs. Will Spear, Miss Edith Mason, J. B. Stevenson, E. J. Layton, Henry Edmonds, J. E. Gilmour, J. Steyermarck, John H. Rush, Miss Martha Beardsley, Mrs. Dorothy M. Anderson, Phil Rau, Miss Mildred Mathias, Miss Katherine H. Litch. Sincere appreciation is due the curators of the United States National Herbarium, the Gray Herbarium of Harvard University, the Dillenian Herbarium, the herbaria of the Field Museum, of the University of Wisconsin, of the Michigan State College, of the Philadelphia Academy of Natural Sciences, of the Linnean Society, and the British Museum, for the opportunity of studying material from these institutions. Thanks are also due Professor J. Paul Goode, of the University of Chicago, for permission to use his Homolosine Equal Area Projection Map (fig. 1.) For much advice and helpful information in connection with the taxonomic aspects of the problem, the author is deeply indebted to his colleague, Dr. J. M. Greenman. For watchful care of the experimental garden during his absence he is pleased to thank Mr. P. A. Kohl. The study would have been less complete in many ways but for the assistance of Mr. R. E. Woodson, Jr. in 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 247 making the collections and tabulating the data. It is a pleasure to acknowledge his helpful companionship and material aid. II. Taxonomy and Morphology HISTORY OF THE SPECIES It has been found that what commonly passes for Iris versi- color L. is made up of two species. One, inhabiting New England, New York, Pennsylvania, northern Ontario, and northern Mich- igan, has lanceolate petals much shorter than the sepals, a short ovary, and shiny D-shaped seeds. The other extends in the middle-western states from the Great Lakes to the Gulf of Mexico and up the Atlantic seaboard to southern Virginia. It has obovate-spatulate petals which are nearly as long as the sepals, a long ovary, and dull, round or D-shaped seeds. As will be brought out in Parts III and IV of this paper, the species are wholly distinct and crosses between them are partially sterile. Their distribution is shown in more detail in fig. 1 . It now becomes necessary to determine their relation to the species named by Linnaeus. In the first edition of the 'Species Plantarum' he named, as numbers 10 and 11, two similar Amer- ican irises. To the first, which he named Iris versicolor, he referred numbers 187 and 188 of Dillenius and the Iris latifolia Virginiana etc. of Ehret. The second he named 7m virginica, basing it upon Gronovius ' Iris corollis imberbibus, etc. The two species of Iris studied by the writer are very similar to each other and, like many other species of the genus, are difficult of recogni- tion in ordinary herbarium specimens. Were it not for the fact that we have such ample material it might be difficult, if not impossible, to determine the relation between the names given by Linnaeus and these two common species of eastern North America. Fortunately there are such good illustrations and such excellent herbarium specimens that there can be no doubt that the common blue flag of the East and North is identical with Iris versicolor of Linnaeus and that the species of the South and Middle West is identical with his Iris virginica. For Iris versicolor I have studied figs. 187 and 188 of Dillenius, Tab. VI and VIII of Ehret (pi. 35), and a photograph (pi. 36, fig. 1) of Dillenius ' specimen in his herbarium at Oxford. Ehret 's [Vol. 15 248 ANNALS OF THE MISSOURI BOTANICAL GARDEN excellent figures leave no doubt about the identification. Par- ticularly to be noticed are the short ovary, the lack of auricles on the dissected style, the broad-bladed sepal, and the lanceolate petals. Dillenius ' figure and his specimen show very small petals ; the herbarium specimen in addition has a short ovary. In the case of 7ns virginica, I have examined Clayton's No. 259 at the British Museum, and Linnaeus ' own specimen of Iris virginica at the Linnaean Society. Both are in an excellent state of preservation and each shows heavy pubescence on the blade of the sepal. Jacquin's plate of Iris virginica is an excellent figure of Iris versicolor. It shows a plant with small petals, sepals with no trace of a yellow eye, and long pedicels. The plate (pi. 36, fig. 2) accompanying Radius ' description of Iris Carolina shows it to be identical with Iris virginica. The sepal and petal are both waved, a character often assumed by Iris virginica, never by 7ns versicolor. The petals are nearly as long as the sepals, and according to the description the plants have few flowers. Furthermore, the species is distinguished from 7ns virginica of Linnaeus, according to Jacquin, which as we have seen, is 7ns versicolor of Linnaeus. Plants from the same general region were described as 7ns caroliniana by Watson in 1898. In his description he speaks of "the elliptical blade lilac, veined with purple and with a yellow spot reaching to the center. ' ' The petals are described as oblong- spatulate. The type specimen, in the Gray Herbarium, shows the typical roundish seeds with a spongy surface, which character- ize 7ns virginica. Material from practically the same locality was studied in the spring of 1928 by the writer. When large numbers of plants were examined there was found to be no consistent difference of any sort between the irises of the Carolina Coastal Plain and those of the Mississippi Valley. There is therefore no ground for maintaining 7ns Shrevei of Small. He distinguishes it from 7ns versicolor but makes no mention of its relationship to 7ns Carolina (Iris virginica L.), with which it is identical. The illustration accompanying his description is evidently made from a badly stunted plant. Specimens collected from the same locality visited by Small are much larger than those 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 249 in his illustration . The locality was visited when the plants were in full bloom and nearly all bore the bright yellow-pubescent spot on their sepals which shows in his illustration of Iris Carolina but is lacking from his illustration of Iris Shrevei. Furthermore, plants transplanted from that locality to the experimental garden have borne as many round as D-shaped seeds, though his illustra- tion shows only a D-shaped seed- COMPARATIVE MORPHOLOGY Distinguishing differences in italics. See also pis. 37-39, 41-43 IRIS VIRGINICA L. ROOTSTALK A creeping rhizome 2.0-4-0 cm. in di- ameter. Stele, in fresh specimens, yellowish white to dirty yellow. Cortex pale pinkish white. IRIS VERSICOLOR L. 0-2 ameter. Stele, in fresh specimens, yellowish white to dirty yellow. Cortex pale pinkish white. ROOTS Wrinkled, 2-4 mm - in diam., white. Wrinkled, 1-3 mm. in diam., white LEAVES Narrowly ensiform, 1-5 cm. broad, gray-green, often flushed with purple below. Narrowly ensiform, 0.5-3.0 cm. broad, gray-green, often flushed with purple below. STEM Sparingly branched above, green or greenish purple. Freely branched above, green or greenish purple. SPATHE VALVES 4-12 cm. long, occasionally becoming foliaceou8. 3-4-5 cm. long, never becoming foli- aceous. PEDICELS 2.5-8.0 cm. long in the flower, rarely equalling the longest spathe valve. OVARY 1.8-3.8 cm. long in the flower; ovary wall relatively thick; ovules nearly filling locules at anthesis. 1.0-8.0 cm. long in the flower, often equalling or surpassing the longest spathe valve. 0.8-2.0 cm. long in the flower; ovary wall relatively thin; ovules filling scarcely half the locules at anthesis. TUBE Wall of tube thick, conspicuously glandular, closely appressed to style base within. Wall of tube thin, not conspicuously glandular, style base distinctly separate from tube. 250 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 Fig. 2. Camera-lucida drawing of cross-sections of ovaries of /. versicolor (right) and /. virginica (left), resin ducts in black, vascular bundles outlined; X 10. SEPALS 4.0-84 cm. long, I.6-4.O cm. broad; blade of sepal oblong-ovate to ovate, usually some shade of bright blue, al- though white and lavender specimens have been found. cm ovate usually some shade of purplish-blue, although white and lavender speci- mens have been found. Blade bearing a bright yellow conspicu- Blade often ivithout a conspicuous spot ously pubescent spot at the base. at the base; when present, greenish or greenish yellow. Hairs of the pubescence inconspicuous and macroscopic, longer than the thick- and microscopic, shorter than the ness of the sepal. conspicuous thickness of the sepal. PETALS Oblong-lanceolate to oblong-spatulate , 3.0-7.0 cm. long, 1.0-3.0 cm. broad, often (in about 50 per cent of the cases) notched at the apex, of the same gen- eral color as the sepals, delicate in texture, easily bruised, wilting readily. STYLE Style branches inwardly auriculate at their convergences. STYLE CRESTS Reflexed, variously toothed, 0.7-2.0 cm. in height. Stigma a triangular tongue. STAMENS Anthers 0.9-2.1 cm. long. Pollen-grains oblong-ovate in outline. Lanceolate to oblong-lanceolate, 2.2- 54 cm. long, 0.5-2.2 cm. broad, very rarely notched at the apex, of the same general color as the sepals ; firm in texture, not readily wilting. Style branches not auriculate at their convergences, or at most but weakly so. Reflexed, variously toothed, 0.7-1.6 cm. in height. Stigma a triangular tongue. Anthers 0.7-1.6 cm. long. Pollen-grains oblong-ovate in outline. 1928] ANDERSON — PEOBLEM OF SPECIES IN IRIS 251 CAPSULE Capsular lining irregularly minute- striate; inner surface mainly dull. SEEDS Round to D-shaped in outline. Seed-coat corky and dull 3 the surface appearing irregular, with occasional broad depressions under a hand-lens. Capsular lining regularly minute- striate, giving the inner surface a uniformly shiny appearance. Always D-shaped in outline. Seed-coat relatively thin, hard, and shiny, the surface appearing regularly pitted, the pits in definite rows under a hand-lens, Owcn Sound, 0«t. Situat* C«i.«x,N.Y. Co*ew*m*o,N.Y OftfoRO, N K. Strrtforo U.K. L« Cr©s« ,W How ^ovi Rocft.,TeNW A NX ft. In-. • seed-coat: uooer row. Iris versi- color, lower row, J. virginica; X 20. GEOGRAPHICAL DISTRIBUTION It is planned to give a detailed discussion of the distribution two sDecies in a later section. Until then certain outstand- of the may be briefly summarized and virginica and shown by the map in fig. 1, Iris versicolor occurs in New England, New York, Pennsylvania, northern Michigan, Wisconsin, and Minnesota, and eastern Canada, Iris virginica extends from the Gulf states to central Michigan and Wisconsin and west to the eastern edge of Nebraska and Kansas. On the Atlantic coast it is common in the Carolina Coastal plain, though it is rare in the mountains and was not seen in the Piedmont. At their zone of contact in Michigan, Wisconsin, and Minnesota, the range of Iris versicolor coincides very nearly with the northern 252 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN A B C Fig. 4. Camera-lucida drawings of seed capsules showing outline and cross-sec- tion with seeds, X V%\ A, I. virginica, from La Crosse, Wis.; B, I. versicolor, from Conewango, N. Y.; C, I. virginica, from Bonnieville, Ky. woods forest, that of Iris virginica with the southern hard 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 253 Both species, while distinctly moisture-loving, belong to the border zone between marsh and dry land. They are not able to compete successfully if the situation is wet enough for Typha or dry enough for grasses. In those regions which have been com- pletely drained and brought under cultivation they have very nearly disappeared. None were found in the drained areas of southeastern Missouri, though the species was formerly common there. Several local naturalists have reported that Iris virginica was much commoner in Tennessee and North Carolina before the lowlands in these states were brought under cultivation. In those sections of the country, on the other hand, where such areas have been left as pasture lands, both species have increased. Livestock eat the grass in preference to the Iris, and the iris colonies spread out vigorously, often covering several acres. The seeds of both species float easily, and seedlings can often be observed coming up on wet sandy beaches of the Great Lakes. It would not be surprising to find Iris virginica establishing itself locally along the St. Lawrence River system. Two such colonies have actually been found : one on a deserted lake beach at Saint Ignace, and one on the shores of the Niagara River a few miles above the Falls on the American side. ABBREVIATIONS The abbreviations used to indicate the herbaria in which the specimens occur are as follows : ASP = Academy of Natural Sciences of Philadelphia. BM = British Museum. MBG = Missouri Botanical Garden. US = United States National Herbarium. TAXONOMY Iris versicolor L. Sp. PI. ed. 1, 39. 1753; Curtis, Bot. Mag. 1 : pi 21. 1790; Redoute, LiliacSes 6: pi. 839. 1812; Edwards, Bot. Gard. 2 : pi. 80, fig. 2. 1812 ; Baker in Gard. Chron. N. S. 6 : 614- 615. 1876; Meehan, Native Flowers & Ferns 1: pi. 86. 1878; Millspaugh, Med. Plants 2: pi. 173. 1892; Britton & Brown, 111. Flora, ed. 1, 1: 448, pi. 1069. 1896, in part; Small, Fl. Southeast. U. S. 306. 1903, and ed. 2,305. 1913, in part; [Vol. 15 254 ANNALS OF THE MISSOURI BOTANICAL GARDEN Gray's Manual, ed. 7, 299. 1908, in part; Britton & Brown, 111. Flora, ed. 2, 1: 537, pi. 1328. 1913, in part; Dykes, Genus Iris, 79-81. 1913, in part; Small, Addisonia 9: 55-56, pi 316. 1924. Pis. 38, 40, 41 (left fig.); 42 fig. 2; 43 figs. 6-10. [Iris americana versicolor stylo non crenato Dillenius, Hort. El- tham. 1. 155, pi. 187. 1732.] [Iris americana versicolor stylo crenato Dillenius, Hort. Eltham. t. 155, pi. 188. 1732.] [Iris latifolia Virginiana Ehret, Plantae Depictae pi. 6 & 8. 1748.] 7m virginica L. ace. to Jacquin, Icones Plant. Rariorum 2: pi. 223. 1786-1793. Iris virginica L. Sp. PI. ed. 1, 39. 1753 ; Michx. Fl. Bor. Amer. 22. 1803; Sims in Curtis' Bot. Mag. 19: pi. 703. 1804; Baker in Gard. Chron. N. S. 6; 615. 1876. Pis. 34, 37, 39, 41 (right fig.); 42 figs. 1, 3, 4; 43 figs. 1-5; 44. [Iris corollis imberbibus Gronovius, Flora Virginica, 11. 1739.] Iris Carolina Radius, Naturforsch. Ges. Leipzig Schrift. 1 : 158, pi. 3. 1822 ; Small, Addisonia 9 : 49-50, pi. 313. 1924. Iris versicolor L. in part; Britton & Brown, 111. Flora, ed. 1, 1: 448, pi. 1069. 1896; Small, Fl. Southeast. U. S. 306. 1903, and ed. 2, 305. 1913; Gray's Manual, ed 7, 299. 1908; Britton & Brown, 111. Flora, ed. 2, 1: 537, pi. 1328. 1913; Dykes, Genus Iris, 79-81. 1913. Iris caroliniana Wats, in Gray's Manual, ed. 6, 514. 1890; Baker, Handbook Irid. 12. 1892 ; Sargent, Gard. & Forest 6 : 334- 335. 1893; Britton & Brown, 111. Flora, ed. 1, 1: 449, pi 1071. 1896; Wats, in Proc. Am. Acad. 25: 134. 1898; Small, Fl. Southeast. U. S. ed. 1, 306. 1903, and ed. 2, 305. 1913; Gray's Manual, ed. 7, 300. 1908; Stapf in Curtis' Bot. Mag. IV, 8: 94, pi. 8465. 1912. Iris versicolor L. var. virgimea L., ex. Baker (err. typ.), Hand- book Irid. 12. 1892. Iris georgiana Britton in Britton & Brown, 111. Flora, ed. 2, 1 : 537, pi. 1330. 1913. Iris Shrevei Small, Addisonia 12: 13-14, pi 391. 1927. 19281 ANDERSON — PBOBLEM OF SPECIES IN IRIS 255 IRIS VERSICOLOR L. I Specimens examined : Canada : Nova Scotia: Canso, July 24, 1901, Ward (US); Sable Is., Aug. 16, 1913, SL John 1184 (US). New Brunswick: Campobello Is., July 8, 1880, Smith (US). Ontario: Lincoln Co., June 10, 1897, McCalla (US); Paris, vicinity, July 4, 1926, Mathias 636 (MBG). United States : Maine: St. John River, July 22, 1917, St. John & Nichols 2280 (US) ; Westbrook, June 8, 1906, Richer 50 (US) ; Fort Kent, river flats, July 10, 1908, Mackenzie 8422 (MBG) ; Seal Harbor, swamp, June 28, 1887, Redfield 7117 (MBG) ; St. Francis, gravelly shore of St. John River, Aug. 5, 1893, Fernald 147 (MBG). New Hampshire: Peterborough, Aug. 20, 1913, Batchelder (ASP); Shelburne, June 17, 1915, Deane (US); Shelburne, Oct. 28, 1915, Deane (US) ; Connecticut Lakes, Oct. 20, 1895, Stevens (US). Vermont: Brandon, swale, June 1921, Button (MBG); So. Burlington, July 8, 1914, Knowlton (ASP). Massachusetts: Dedham, June 13, 1897, Greenman (MBG); Brookline, June 26, 1896, Greenman 2321 (MBG) | Granville, meadow brook, June 20, 1914, Seymour 174 (MBG) So. Framingham, May 31, 1890, Sturtevant (MBG); Cambridge May 29, 1901, Floyd (MBG) ; Nantucket, pond margin, July, 1923 Mason (MBG) ; Ashland, wet grounds, Morong (MBG) ; Nonquit June 5, 1888, Sturtevant (MBG); Woods Hole, July 26, 1911 Pennell8164 (ASP) ; Wilbraham, June 12, 1876, Pillsbury (ASP) Boston, 1816, Boott (US); Marthas Vineyard, June 29, 1916 Seymour 6158 (US); Nantucket, June 2, 1900, Day 64 (US) Marthas Vineyard, Aug. 1888, Harrison (US). Connecticut: Middletown, May, 1836, Bigelow (MBG) exact locality and date lacking, Wright (MBG) ; Stratford, June 7 1893, Eames (US). New York: Otis Summit, July 27, 1903, Williamson (ASP) St. Regis Falls, June 28, 1903, Hudson 81 (US) ; Ithaca, June 23 1885, Coville (US) ; Van Courtlandt, June, 1893, Pollard (US) ; N Harpersfield, June 21, 1906, Topping (US) ; Elizabethtown, swamps 256 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 July 2, 1875, Redfield 7911 (MBG) ; Ithaca, June 4, 1877, Trelease (MBG) ; east of Cherry Valley, July 2, 1926, Mathias 652 (MBG) ; banks of Lake George, Aug. 24, 1856, Engelmann (MBG) ; locality lacking, 1840, Short (MBG) ; Bedminster, meadows, date lacking, Perry (MBG) . New Jersey: Vineland, Sept. 12, 1923, Bassett & Long (ASP) ; Camden, May 22, 1922, Bassett (ASP); Quaker Bridge, Sept. 11, 1923, Dreisbach 1884 (ASP); Folsom, June 10, 1911, Long 5940 (ASP); Cor's, Ocean Co., Oct. 11, 1910, Long 5582 (ASP); Cold Spring, Cape May Co., June 1, 1911, Brown (ASP) ; Atlantic City, margin of meadows, June 3, 1875, Redfield 7910 (MBG). Pennsylvania: Lester, Delaware Co., May 28, 1912, Findlay (ASP); Chester Co., June 4, 1905, Weston 164 (US); Harris- burg, Aug. 13, 1888, Small (US). Maryland : Spesutic Is., May 25, 1879, Smith (US) ; Chester- town, Aug. 18, 1900, Vanata (ASP). District of Columbia: Washington, May and June, 1878, Ward (US). Michigan : Cheboygan Co., sedge pool, July 17, 1917, Gates & Gates 10597 (MBG); Douglas Lake, Cheboygan Co., July 17, 1916, Ehlers (MBG) ; Burt Lake, Cheboygan Co., July 4, 1917, Gates & Gates 10492 (MBG) ; L ' Anse Co., July, 1892, Eby (MBG) ; Keweenaw Peninsula, Aug. 1878, Engelmann (MBG). Wisconsin: La Pointe, Sept. 3, 1878, Engelmann (MBG). IRIS VIRGINICA l. Specimens examined : United States: District of Columbia: Alexander Is., May 15, 1915, Van Eseltine 866 (US) . Virginia: Norfolk Co., May 11, 1898, Kearney 1079 (US); exact locality and date lacking, Clayton 259 (BM). North Carolina: Hendersonville, wet meadows, May 26, 1898, Biltmore Herb. 542b (MBG, US) ; Roandale Farm, June 8, 1895, Wetherby 1067 (US). South Carolina : Cat Island, Georgetown Co., Aug. 16, 1915, Alexander 111 (US, NY). Ohio: Middleburg, Cuyahoga Co., June 20, 1897, Watson 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 257 (MBG) ; Clare, Hamilton Co., spring-fed swamp, May 30, 1906, Braun (MBG) . Kentucky: Elizabethtown, May 23, 1926, Anderson & Woodson 88 (MBG) ; exact locality and date lacking, Shor f(MBG) ; Bowling Green, swamp, May 21, 1897, Price (MBG); Drake's Creek, Warren Co., May, 1891, Price (MBG); Livermore, Mc- Lean Co., June 2, 1920, E. J. Palmer 17707 (MBG); Ohio City, swamp, June 5, 1901, Price (MBG); Clay City, May 23, 1903, Biltmore Herb. 542 (US). Tennessee: Stanton, May 20, 1926, Anderson & Woodson 60 (MBG) ; Knoxville, wet ground, May 1896, Ruth (MBG). Georgia: McQueen Is., Chatham Co., brackish marshes, April 30, 1904, Harper 2180 (MBG). Florida : Lake City, May 23, 1897, MacKenzie (MBG) . Michigan: Adrian, June 24, 1901, Dewey 508 (US); Ann Arbor, June 10, 1884, Sudworth 50 (US) ; Flint, swamp, June 18, 1923, Anderson (MBG). Indiana : Terre Haute, May 19, 1889, Evermann (US) ; Roby, June 18, 1910, Lansing 2789 (US) ; Miller's, May 30, 1902, Chase 1 799 (US) . Alabama: White River marshes, April 14, 1898, Mohr (US); Hollywood, May 15, 1902, Biltmore Herb. 5J$g (US) ; Pearlington, April 8, 1880, Mohr (US) ; Mobile, May 16, 1893, Mohr (US). Illinois: Spoon River, Stark Co., June 9, 1907, Chase 1425 (US) ; Macon Co., June 12, 1915, Clokey 2438 (US) ; Morgan Park Ridge, May 27, 1907, Dixon & Gage 702 (US); Kankakee, May 27, 1913, Crampton 124 (US) ; Decatur Co., May 29, year lacking, Clokey (MBG); French Village, St. Clair Co., May 27, 1903, Eggert (MBG); Olney, April 14, 1921, Ridgway 1381 (MBG); Fish Lake, St. Clair Co., July 16, 1858, Norton (MBG) ; Wood- lawn, Jefferson Co., May 16, 1898, Eggert (MBG); Decatur, date lacking, Clokey 2789 (MBG) ; Shawneetown, swampy borders of lake, June 19, 1919, E. J. Palmer 15554 (MBG); Reevesville, Johnson Co., low swampy open ground, June 3, 1919, E. J. Palmer 15350 (MBG); Mounds, Pulaski Co., May 7, 1919, E. J. Palmer 15074 (MBG). Mississippi: Beauvoir, March 28, 1898, Tracy 4468 (MBG). Minnesota : Nicollet, June, 1892, Ballard (US) ; Swan Lake, [Vol. 15 258 ANNALS OF THE MISSOURI BOTANICAL GARDEN July 16, 1917, Metcaljlfi (US); Fort Snelling, June wood (US). Iowa: Tiffin, June 5, 1909, Somes 8132 (US); Fayette Co June 5, 1894, Fink 37 (US); Hamilton Co., July 1891, Rolfs (MBG); Missouri Valley, June 21, 1897, Pammel 587 (MBG); Iowa City, date lacking, Hitchcock (MBG) ; Armstrong, Emmet Co., June 18, 1897, Cratty (MBG); Ames, June 21, 1897, Ball & Meeker 524 (MBG). Missouri : St. Louis Co., June 1912, Craig (MBG) ; Oakwood, Ralls Co., open woods, June 19, 1917, Dams 7553 (MBG); Randolph, May 18, 1895, Mackenzie 544 (MBG) ; Jackson Co., low wet ground, common, June 17, 1892, Bush 1475 (MBG); Jefferson City, June 1870, Krause (MBG) ; Carthage, swamps, May 27, 1906, E. J. Palmer 883 (MBG) ; Kimmswick, May 23, 1885, Wislizenus 407 (MBG) ; Winfield, Lincoln Co., June 7, 1916, Davis 1 456 (MBG) . Arkansas: Nettleton, May 7, 1893, Eggert (MBG); Craig- head, lakes, May 7, 1893, Eggert (MBG). Louisiana: Hammond, April 4, 1889, Gallup 16 (US); Minden, April 14, 1901, Trelease (MBG); Madisonville, May 1888, Joor (MBG) ; Covington, April 1920, Arsene 12400 (US). . Nebraska: Nebraska City, June, 1889, Webber (MBG). Kansas : Wyandotte, moist places, Aug. 5, 1897, Clothier 1067 (MBG, US). Texas: Troupe, May 9, 1902, Reverchon 2793 (MBG, US); Orange, April 19, 1899, Bray 61 (US); Keechi, Leon Co., April 20, 1918, E. J. Palmer 18420 (MBG) ; Beaumont, Jefferson Co., April 22, 1916, E. J. Palmer 9522 (MBG). III. Intra-specific Variation in Iris virginica Wherever either species was studied, the individual plants which went to make up a colony were found to vary strikingly among themselves. They varied in every conceivable character- both vegetative and floral, though the most and form, in color and color pattern, in number and arrangement, in texture, and in odor. Plate 34 shows such of these differences as can be Ann. Mo. Hot. Gaud., Vol. 15, 1028 Plate 34 H o GO p pc ■a p C 2- 2. O 3 -a o CD CO 5* 5 19281 ANDERSON — PROBLEM OF SPECIES IN IRIS 259 recorded by the camera for six: individuals from a typical colony, that of Iris virginica at Portage des Sioux, Missouri. They are not selected extremes but are the first six individuals which were located with at least three flowers apiece. The pictures were taken under uniform conditions as regards lighting, exposures, development, etc., and such differences as occur are due to differences in the color and texture of the flowers themselves. It may not be out of place to call attention to some of the differences which obtain between these six plants, since they are similar to the differences which were met with everywhere throughout the study. Plant No. 1. Petals narrow and ruffled, ovary short. Plant No. 2. Petals flat and wide, flower pale blue in color. Plant No. 3. Entire plant large and robust, of great vegeta- tive vigor, clone covering half an acre, flowers marked with dark blue veins on a light back- ground, spot on sepal large and brilliant. Plant No. 4. Flowers small and light colored, sepals broadly spatulate. Plant No. 5. Flowers pale gray-blue, sepals long and narrow, stems and leaves very short, plant floriferous. Plant No. 6. Flowers dark reddish blue, ovary short, stems slender, few-flowered. In order to summarize and average such differences as these it becomes necessary to select a few for concentrated study. The characters chosen should fulfill three conditions ; they should be easily measurable; they should show some variability; and they should not be easily affected by environmental influences. The first qualification eliminates a number of very conspicuous differences, such as color and color pattern, which are not adapted to quick and accurate measurement. The second bars those characters, such as the length of the anther, which are practically indentical for a large number of species of Iris. All of the characteristics of the plant are, of course, open to the objection that they are affected by the environment but some of them are much more stable than others. The leaf characters are particu- larly poor. The largest leaf on a plant may be two or three times [Vol. 15 260 ANNALS OF THE MISSOURI BOTANICAL GARDEN as wide as the smallest. The size and shape of the seed capsules are likewise extremely variable on the same plant in different seasons and under different conditions. In general the dimen- sions of the floral elements are less easily affected, and since they are of predominant importance from a taxonomic standpoint they were very largely used in the present work. After a year 's preliminary study, seven characters were chosen for measurement and have been used throughout the work. Five others were subsequently tried extensively, and three of these are being used at the present time. The characters chosen and the exact manner in which measurements were taken are shown in fig. 5. These characters are measured as follows : Sepal length. This is the maximum length of the sepal from the base of the stamen to the tip. Sepal width. The maximum width at right angles to the length. Sepal taper. Distance from the tip to the point of maximum width. Petal length. Unfortunately there was no exact lower boundary as in the case of the sepal, and some of the variation in the earlier measurements is due to the fact that different points were used as a base. It was eventually defined as the line between the bases of the lateral flanges at which it breaks off naturally when pulled over backward. Petal width. The maximum width at right angles to the length. Petal taper. The distance from the tip to the point of maximum width . Crest. The maximum length of the appendages of the stigma, measured from their tips to a line connecting the ends of the stigmatic lip. Ovary. The junction of the ovary and pedicel is distinct in both species ; that between the ovary and the tube is much less so, being marked by a vague ridge. The summit of this ridge is taken as the line of demarcation and the ovary is measured from it to the base. Tube. The tube is measured from the same point to the base of the stamen, one sepal having been removed. 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 261 Stamen. The length of the anther from the attachment of the filament to the tip. Care was taken to measure only fully opened, uninjured flowers. Not more than one flower of a clone was measured (except in those few cases where intra-clonal variation was being studied). It was found upon experiment that one single set of measurements gave almost as consistent results as when two sepals and two petals were measured on each plant. An experiment was made on the Fig. 5. Diagram showing how floral elements of Iris were measured. colony at Portage des Sioux, Missouri, to determine if the measure- ments were being affected by the flowers not being fully expanded. Twenty flowers were measured at four o'clock in the afternoon and remeasured at seven the next morning with practically identical results. The first five year's measurements have been summarized in tables i-iv. The figures in italics are the class containing the median or mid value. The median has been used as an average rather than the mathematical mean, since it is less influenced by occasional extreme values. The extra-small size of a frost-bitten or insect-mutilated petal, for instance, is not really significant, yet a single such extreme observation might seriously influence the position of the mathematical mean. It would have no more effect on the position of the median than would any petal of less than average size. One fact is immediately apparent from an inspection of the summary. No single measurement will serve as a criterion for 262 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN £ > H o CO o feC O o o3 cd o3 o o o3 CD CO WOT MhC^h lOCO oscoN.*o^-. -lOC^HNWON I ^ CO CO 30 • CO CO o CO Tt< t^co ^ IO ^ -«COfHCOOOcOwoOocO'^cO(N a CD cd jd a cd c3 CD pu, ^ IQ iO iocmiococncogococo^io^^co^ Gor- CM CO COCO cooc Co CM CD CO OS OS IQ t>- -^jNhoO Oo^Cot^t^COCMCMO CS Tt< &) CO I CO

G i £ 03 Q a^d > CO o c d d o ® s ^ 03 o3 o3 d d o3 o d o3 C» P P o3 |8 »- o o3 X P o o o c3 O O O O s3 x - >iOQ OQ CO CQ'^H ^ a> «P CD CD CD. 2 A o JZ3 T3 T3 73 73 £ ^ co c3 . . . . O •*"3 o3 CQ CD > O :^^a** H iJ^ ct3( - ) ^^L^: t:i3 ^' 73 o3r^Kro3 . . . .0.^03 JB0 A I>t^-l^-COCDCDCDCDCDCDCDiOCD"^iOCD»OC040CDr-r^^GOGO Osososososososos OS osososososososososososososososos d o CO CD ^3 CD CQ CD o3 CO CD ca CD -d 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 263 "8 P a o PQ "8 .s o < £ « 0) 55 O 5 o <3 a3 c-> d o3 o 03 P iONiOHCOQOOOCOCOCOOOO^COiOWNOOOOiOiOiO CN HCO(N(NOiO'HCO(NrHrH(NCO(NCOCOWW^ CQ 0) s I Tt< 00 00 I> o r^ l 00 CO CD - C^Tji ^^CO^iCoCO r>- co »h co co ^ O«D CN HCO(NC^COlOrHCO(NrHrH(NCO(NCOCOCC^ -^ ^ oo *o a* tjh o a> oo oo > 44 K pra *H -ft I 5 CI o b£) PQ ^^ c3 03^ 72 ifl O o*^ w o a o a> 1 cu & s s § flj a 03 o3 sEMgcgafiflarffe'ggafljfigaflaa JB8^ OOOOOOOOtJH (N(N • CD^COIn- cDfNC^iO^ iO cD t^OCO^O^OOO CO CD • 1 • CO -h- t^ CD t^ CO CD • I • CD CD cd CD CD CD rd bfi 3 cd *0 00 *0 coco Tt* iO *C COCOio»o (N»0O^ CO CD oo oo 00 o> o\ CO Tt< OiiOcDr^ C^C^OC^i-hOOCDCD iO CO 00 CO CD -tf CO CO >«^ CD *h CD CD bC g § (D +3 «S P •5-5 ^SS • c3 ^ ^^ g? rt 5J .9 "o CD CD CD j^a^ TticDrtcDcDt^T^'^'^ , ^t^"^fcOt^ C^(N(NCNCNCN(NC>1CN(N(NC^(N(N(N(N 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 265 CD p o < o o GO tf 0) S3 bfi C o o 0> o GO 3 C^CNCOi-h CM ^ht-KNCO(NCOCOCNCO CM ■«^ CO T*Hrt< -i 00 io CO lO tJ< CO CO io io GOCO^-CO • 00 CM CO O ^t- t^OCO»-iCMC>OCOiO^coT^icow5'-HcoTt- IO GO • I • CO CO ObCOOt^^COrt^OJ^r^COi005l>05CO>^^.t>l^cOcOCOCNl I "tf CO CO CD^OW^H^^ •^OOiloCNiOrHCOCOCOCO • IO • CM CM CO CO cocoes -^ CM CM CO CO CO CO CM a> cm H J>H w ggggoqPQ a a a §P h QhPhPl ( ^Z^^^>>>>>^SoOO^^^^ a o *d o o i o °° o s CD a « l o o o . bD bfi_»H OJ3r^ o3 o 03 P O IP OP CP o -d a o3 T5 a o3 w CU a 03 03 _T! CO O o3 c3 a c3 bfi 03 ie -9 a o3 a o3 J« 9 A tN-t^ioioiOT^ioiot^-ioioioiococoTjit^.t^.cocot^r^-i^-t*- C^CNC^CMC^CM(NC^C^CNC^CMC^CN(N(NCS|CNCNICNICN1C^C^C^ 266 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 T3 CD g o fc ► X o c « CO CO ,-: l • CO CO Oi CM »o CO no 00 ^ *C to ^ IO CO CO CO CO CM COrt* *0 CO CO • I • ^ <^ CO CM CO O CM CO ^ iO 00 • I CO CO CO -^ CO CO t^ o • I • CM CO CO CO CM CO CM O CM CM CO O^ CO CO ^o CO 3 o3 w CD O O CD O SSSSSo a o >> o3 o3 „_ o3 CD O o3 a b/OCD T3 o3 o o3>HT3 a^- 5 J^8 A r^ i^ t>- 1^. xji t>. CM CM CM CM CM CM 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 267 H ffl £ CO « 53 o I o o c3 CD "55 CO "3 x? o c3 CD O P > jequinu TOOT *OCN CN hC^ o io lOCO r^oot^»Oi-HOOQOcocoiOooooot^ CN ^HCOCN i-HCN CO ^i(M "tfCNCNCN CO cd CD a p CD "3 CD CM TfH T^ CO T^ t^ GO CO CO »o CO • I • CO CO CO ^ CO CO CO CO CN lO CN CN LO T^ CO a 05hc^ CN kO T* CN • CN CN^CN CO»-HCNi-HCN -CN^t^CN CN *C CO io^-co C& o • I • CN CO CN -CO • tF ^ ^ CN CN i-H CO *C CO CO O CO -^ CO -iOCoOJN GO CN CN ^^>CO • CN CD i-h CO CO O iO t^ CN tJh ^h C5 Oo CN Go *o f-t e^ iO CO • I • CN CN t^i-H^CO^HCOCO'-H^- •N.Co^cOCO CO TtH • I CN CN COCN »OCN •© CN I CN CN CN i-hCN O O CN t^ CO iO CO CO t^ ■ I • CN CN iO ^ Tti CO ^ CO CN *• <^ CN CO *0 CN COCN CN OlNHlCTf CO r^CN t^co CO iO T*i CN jaquinu CO — CD CD a a CN i CO t^ CO CN CN iO CO ■ I • CN CN CN CO CN CN CN CN CN CN CN CO Ci^CN CN CN CN CN CNCO CN CN OJ CO p CD O 13 CD Oh 0) CN *OCN t* GO CN Oc ^ ^h CN ^t-CO t-* CO 00 6- rt< 00 Tt< • CO ^ CO CN rHCN CN CO ^ ^Tt< COCO -^ »^ t^t^» iO CO CO • rt^ CO CO CO CO CO CO CO *o *o Q* CN ^h C^ CO CN 0^ CO tF CO CO CN HTJ^ U-) iQ -t^CO CO i-h CN CO CO CN CO O5COC0 »0 G> © CN CN t^ GO iO CO CD 03 CN CO ©0 CN CO »OClr-iNNOi0HHNC0»O>O(N000CC0C0iClO00NC0'HOiO CN --h CN ^h tO CO i-H CN rH CO CN rH CM CO tF CN "^ CN CN CN CN CN oo 00 CN m cq P P P r-3L^L3 CD CD CD.^.^V. 9 ^3 O o Si a o CO o O W) P CD CD ° ^P CD CD £ C2 P O CD .2 P P o 3 -O o3 P P O rf^ 1 PLd m X X X X 75 p p p p 030000 >iW CQ OQ OJ CD CD CD CD CD 'P HD T3 H3 ti M^ fl •% 5 P cd ;i^ d ^5 °J OP5," fl 3 3 • Jqgfl • • • .O--i03a3 Q CD 2 A ^^A t^t>-t^cOcOcOcOcOcOCOc040CO'^iOCOfcOCOiOcOi>»COTj O CO I T* GO CO CO iO CD • I CO CO CM CO CO I CO CM cm CO CO CO o • I CM CO COT^ t>- 00 • I o i cm r- oo *C CD • I • CO Tji CO -(N COCO CDiOOCDOo CiCO(N ^-00^00 Co lOCD CDO^OCO lO^T^O^r^^Co^^^Co^COT^T^iOC^iOt^^^OiU^cD CO •^-r-HCO CM CO O CO CM CO CM CO iC CD ^h OJ CM t^ CM (N CM CM •^ CO O CO CM *0 ^ CO *C CO CM CM jaqranu »OHOOOOOCDCOON05^0iOCOC^iOOOO»OcO , ^(NrH(N CM CO CM CM CD *C »-h CO CM r-< ,-h CM CO CM CO CO CO CM "^ "^ CM 2 a XI OS I CM CO t^ I 00 CM CM *C CD I • I CM CM CO ^ ^^^^t^<^CMCOt^T-HT^COCMCM'^ai^OCMiO CO CD 00 Co CO *C 00 CO O iO CO CO CN CMCN CO^h CO CO c5 o <33 o QQ £/2 .s >os n as o CU •> SO c3 CO i3« a3 CO a> CO o3 ^S« 111 S S?§ g| I oT^^^I^ 2 §-3 g^5 nil ss ^S^faCQO^SaCQO^^^^ScQ^OOGoa^^OOfa &23S o 2^ ^ cu bfi 2 00 03 Jv X88I OOOOQO^T^^l^iOiO^OiOiO^O^OTtiT^TtHCDCDcOcOcD^^^CD^ C<1C^CMC^C^C^C^C^C^C^C^C^CS|(NC^C^C^(NC^CNC^CM 19281 ANDERSON — PROBLEM OF SPECIES IN IRIS 269 1 d g "8 | fc tf 2 •53 5 W) O ■jj cd 3 g O CO 3 jaquina iOO?DO»OCOt} 00 • I ~ CO CO iO CD ■ I J CO CO CO , TjH • I CO CO o (N CO COt»^ K 00 CM GO C^ cm iO CO GO 00 00 CO CD CO CN CM Co COCO *OCM ^ CM CO o> COt^^CM CO^» CO cscoco 03 a, 02 CO t* • I CM CM GMCO(N(NOh^hcOO CM CM CM CM CM^^nCMt^ O O . I COCM I 00 fcO CO CO T^ T^ ^ CM jaquinu iOOOO»0^^05WNCO^ COCOi-iCOCN^ iOCMCO 2 0> a d 13 • I CM CO GO CM CM lO , CO CM CM CO ^t* CM CM Tf CO CM (N CM CM CO *C CM CM Q> O I CM OCO *^ OiiO^ CM ^ Tt* *^ CO CO CM CM 00 ^*0^ OQ 00 CO C^ *o CO T^oo^-icooococococoiN.a> CO T* CM CM CM CO o , o 1^ 00 • I iO CO • I • co cm a> CM T^l> l^ CM CM J3 -d ^3 *d M M « d OCPOOOOddaQOQOQ-M+a 33333323^^66 d o d d fc a s ffJFII g g . d § ^ £ > c3 O 0) d °3 bOHH QQ Vh bC d d 'o jto* COt>-COCOtN.'^<' , tfl'*t»T^t > -^t< l Ot^' CMCM> 'o -a 03 O "3 •g jaqranu Nd^MC005»OOCO»ONNN«:^icOCO(NcOCOTt ■t-h CM •coa 1 • i-h CO CO CO CO CM • • * t-h CM • 1 • CO CO r-HCOr-H •r-HCO •»-HCNCN ) • CO CO H O CM CO •CM os o • 1 • CM CO co-^-co • • COt^CM o o •T-HCOOoOOi-HCO^CoiOCO'^lOiO N 00 • 1 • CM CM OCD^rHCO»HCOCOCOCOf-ikOOO^iOfHQoNiCCOOifHCO(N iO CO * 1 CM CM "^ t-h O rt< CM r-H Co Qq rt< Oo CO CO 00 CO "tf t-H CO r-i CO CM ^f Tt< ^t " CO Tt* • I ■ CM CM *0 • CO CO CM -CNCOCMTtH00>-iCMCOCM • '• CN CO :- •CM ^h CM CM c^ r-i _ ^ oo ^ ^ ,_, ; 1 lON^T^HHHH I OS O • 1 • - : CM -h i : — ' : r-H ON r-H '> r : t^ 00 • I :-— : CMCO t-h ! *C CO • 1 • ^0 CO ^ tinue jaquinu F»°JL COOCO-^'-HCiCOGOcOrHcOXCOlO-^lOO^O'-^COCOCOtOCM CM CN CO t-h r-H CM t-h pH CN CO CM CO CO CN CO CM CM CO CM CM iH t-h h Con Petal width in centimeters os o • 1 CM CO l> 00 • 1 • CM CM ABLE : iO CO • 1 • CM CM CO Tt< • 1 CN CM • • • - : H t-h CM 1 • . . . CN CN 1 • CM • - - ': - os o • 1 t-h CM * : OS • CM • CMCM - CM r-H i CO • ^-H N 00 • 1 •*- : CO • a* • - : T^CO • • OSCO r-H -r-lTt CO • • kO CO r-I ' r-i CM COCO • CO t-H CO Oi r-H • •tHIOQO • • • HQoNCNiO^iNCO CO "^ • 1 10^)00 • rt< TH CM CO *0 CM N CO O OS • N»CN050«OrH(N • 1 CO CO §* t-H r-H CN ^t •*0 -OCN>-iCO"^iOCMOO^HOCo^CMI>.rH OS o • 1 XCO^r^N • • • • • • iO iO • US *« OQ IQ *• CO CM t-h '• CO rt< t-h *• '. i-i t- GO • 1 • COHCOOCO tO CO • 1 ^ ..::::: : :^ : : : : : : : : : : : : : c c a. 1 2 • t3 » » • • ^h o3 c3 c3 c3 Conewango N. Y. Conewango N. Y. Cherry Cr. N. Y. ViUenova N. Y. Hubbardsville N. Y. Pownal Vt. Pownal Cent. Vt. Clarendon Vt. Middlebury Vt. New Haven Jet. Vt. Holderness N. H. CC • • ■ aa * d d d • • • • c B 4 Hood Liverpool Harmonsburg Meadville Greenville Duxbury Alberton Ottawa L. Timagami Truro Antigonish Baddeck McAdam j« »A CM OS CM OS CM OS 4C CN CM OS CM OS 4C CM to CM CM CM OS CM CM CM Oi CO CM CD CO CM CM OS CM o> CM Oi CO CM OS CO CM CM CN OS CN OS CN 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 271 73 a o n < z k X o o 02 « o J-H o 03 a; CO O a c3 o CO 'oS jaqinnu »o^ iO CO CO(NH(N CO cm ^ ^ • CO ■*< t^ 00 CO CO *o CO CO CO 13 CO Tt< • I CO CO cm CO CO a> o • I CM CO *C CO TF CM ^O Oq *cO CO CO 00 CO i-h t* t^ 00 CM CM CM CM ^ CO CM ^O 1Q CO • I • CM CM ^COCOt-h ^-r^ co T* CM o I cm r- oo *o CO • I CO rt< CO CM CO CO CO jaquinu >ON^ CO CO CM i-HCM co CD s CM I CO 00 CM CM CM I CO CM CO CM I CM CM C^ CM Oi O CM COtJ^CO 00 03 CMCM^iO Oi lO CO CO 00 00 i-HCO CO Tt< CM CM CO 00 *o CO *«H *OCO ^^ CO *o 5 ATAXIA o o « o o SSSSSo >> if* GO O 03 cp • o • oj &g o |h<» IT'S xmjL r^ r>- •** t^ t^ i> CM CM CM CM CM CM O Oi O* G5 OS OS 272 [Vol. 15 ANNALS OP THE MISSOURI BOTANICAL GARDEN 3 < < 'A 00 tf CP Gq o bC a o 08 0> CM CO • I I joquinu co a 0) o I -3 0) 00 O • I • CM CM O |> CM (M ^ *0 CM CM CM CM CO CM CM CM 00 a> co I CM CO I 00 05 • I CO t> CM CM NN CO I>o0 CM i-i CO iO 00 | w CM ON ^CM CM CO CM -CM rH 00O CM tF ^ CM CM CM Cb CM CM CO CO GO '^OOCOCM -^ ^•CM ^iQh lOlO CO Tt< CO tQ CO i— I IQ 1*3 CO *C CM CMCM^^OCM COrt< Tji CO CM ^ CM CM CM CM CM CO £^ T^ CO COt^ 00 io 00 Co COCO lOCM ioc^hnno»o CM phCM^-h lOCO CM HCO IO CO CM CM 00 t^CM I CM CO CO CM CM h l> 00 CO^cOCO^HCi^CO CO ^O CM lo ^CM A © CM CM CM CO Tt^^iO CO LQ i-h CM CO 00 CO CO CM CM CO COO CM CO CO CO CM CM ia -* *o io • CM rt< COCO QO T^CO -*.^M IO CM 03 ai co co a a a I -^ooooooo Assesses c3 o3 o o o 43 bC O Fh O 0? CO a H bC d o co O bfifl > ■g ' d 0) 0> 3^-2 -a d o A i.HJlii OP 03 a d o3 o3 o3 a a3 CO 3 d o M vs X X X O i$ d d d c3 O O O CO ^ CQ CO CO ^ fl-> CD QJ -tvt^cOcOcOcOcOcOcOcOiOcOTt»iOcOiOcO»OCOt>-Tt ft 03 • I CM CM CO CM I In. CM cm cm (N CM CO CM 03 OQ CM CM 00 OS CO CM CO I jaquinu OQ a 00 CM I OS CM CO t^ • I • CM CM CM CM CM CO • I CM CM CM CM 00 OS & 3 Ph CD l> • I • CM CO • I 00 OS • I o3 DQ O H CO CD C» CO OS CN CO CO 00 aa CD »OOOOiO CO CM CO CO CO CD CM CM t^CM OSCN ^) lOt^ ocm ^ ^ CD ^ IO 00 O CO iO CO CD *0 *00©0*0 CO CM CO CO CO CM CM CM 00 • *0 CO »C O TtCO t^OCM ^HOO t*CM CM • • • *OOCO CM I COCO CM CM ^ lOCD^ oo^*o <^ cD*C l>T^ CM ^OOOOOOO O -^ *G J3 *G ^ M -C Xi A A AAA A A ,a JT pjG _C pU ~G -C £j£^ ,H,H,H,H ' H ' H,H ' N,,,<,N,-, H'H.H'H.5 OQ QQ OQ a a> p» pn 2 S OQ OQ c3 o3 mm ^3^3 bo T3 a> a P. o 0> as ^-c:m^^ 03 5 §2 3 2**Z 3 5 s 2^§|§ MM 5?| |2 a o 2^ ^ 9 J 0) a p p #s a CS.J5 q a Pi Q3 ^> >H ^b*ipi0fc0i0iC40i0T^TjH^OcDCDcDc0Tt • I CN CN CD a CD O 0) a 03 • I • CN CN T*1 CN CM CN CO • I CN CN CNCO CO 00 OS *C C^ CO l> COO Tt< I iO *OCN CN CO • I • ^ I jdqumtz CN CN CQ CD CD a a CD CD 00 CN I CN CO CN CN CN CN CN CN CO CN CN CN 00 C5 co r- CD cj CD Ph t^CO ^ IO o^ CN CO ooo I CO 00 O • I co b- • I • CD $ QQ I I CN CO CO t^ o CO CN CO Tt< 0^ CO CD O CD ss <5 g g . . . C m » x (« ^H CNCNCNCNCNCNCNCNCNCN O^GiG5Gi05CiOiC7i05Ci 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 275 d d O O PS O o P N •53 8 3 1 cd a o ,d a o3 J3 c3 d "P a jaqainu toot, CD 00 o> CO t* • I • *0 C^ CO (NOi §3 03 o3 (N ^ *o cd r> • I (Ml Tji lO O CN * • m • ■ CNCO OcD(NCDCOTt<^C c3 > o j=3 a> 03 n a d o o PhPh d o d OJ c3 3 T3 T5 d 55 d OS b« 33 o3 c3 o3 I a o o H rd g o cJ p O t^|>iOiOiO^^OiOt— iOkO^OOCDCDTtiI>-t— CDCDt— t-t— t- C^C^C^(NC^C^C^C^C^C^IC.L •HHCOIN CM 00 G5 • * \ \ \ \ CO t^ a m • • ■ • • CO B * 1 • 1 # # • • • • | ■ •(NrH '• '• CD a hH >> ~: 1 ^.* g o cm « 3 HH H i "3 o 03 go f • 1 • # * HH j CD 03 *C3 r-i ' ' »-tCO^ Co -co CM CO HH ^^F ■ 1 ■ 1-HCO00CO '00 pa cd ►h HH 2 '2 COCOCOCO -i-H 00 O 1 • 1 • • • • • • • 1 • »-H • * • • • • • * * • CD jlArGjZrrl . 5 CD o o o o -^> siissiso as cd o Eh 2 S +iT3 o3 o 03 «s! fl s! o 8 1 t^» l> l> l> TfH t^ CM CM CM CM CM CM O) O* O* O) O) o* KH ■ i-H y- 1 , " H1 r-ir-ir-4 1928 ANDERSON — PROBLEM OF SPECIES IN IRIS 277 > < Q £ PS GO « CD O 3 o cd CD '55 O o cd o o3 CD O TO 1 laqranu FM>X (M Od GO t 00 CD CD a co IC ^ d cd 6 d S-H O CO CM os 00 laquinu t^ © CO *OCO ONNNiO(NOOOO(NWOOi (NhWWhiNcO Tji c^ CO CO OlON tH CO CO CO «* rH-*OCM^CO^CN ^*-CO CO i-H ^ ^-i— I CD <* CO ^O .cOiO(N Tt<(N OCO^CM •cOt^ CM CO CO CM CM ^ CM CO CM CM^CM •00 CM CM CO CO coco CM CM CO • 't^. • CM • O • • • • O *C CO CM ■ -CO "^ -i-H • • • -CM CO Tji CM ^ 00 -tf »o T* ^ *o to CO -CM ^ *C CO CM cm -b- CO ** CO Tt< 00 cm CO CD > a o o) a bfl 3 o 72 OQ fl G co. 2 a a a a o •d bfl o d d o a 9 5P d 8M-8 s a 0) 03 OQ ■■a CD Q a « o fl a o o a3 o3 d a g d c3 Ul GO «•= JB3^ -2'Sbs JUS .9 cj So .05 -TS O XJ p o o3 >fh ^ >% CQ QQ CQ 0) CU «X> Q) ^3 "O T^T? 14 > o P o bO a o3 513 i*sj t^- 1^- 1^. CO CO CO C^ C^ CN (N CM -T^00000000 CN(N(N> d rd o 03 *o d d H-4 joqranu IOHCD00O5 *O0 'HNNHCOiO(N^»OOOOiCiO|2gai rH (N -CO • CO *C ^h CO ^h pH h (N CO (N CO CO CO • • • rH t * * * 1.4 CO »W ft • • ft • • • 1 rH • • rH • H CO ' ■ # CO T* Tj< CO >H "^ CM • | rH '-00 •CO -rH • • (^ .HCOH . rH CO CO rH CO ^ CO rH C5 "^ ^ • • 1.2 rH ^Or H '• ^ "• • • * t> CO •OM^(NO^^0N^»0O»0^ • •HTt *h oo • • CO -* : : « • i b. Oi 00 Oi CO l> c^ • • 1.1 1.2 .* I '. CO C<1 t- .lOU (NCMC^CMCNCNCNC O) O) O O) 9 9 9 c ^cOcOcOcOcO^t^^cO N ■d .3 o £ GO « jaqainu wot CO »o CO s 6 8 S Pi o c^ a> x r- CO iO Tt< CN iC CO T*< 00 t^ CO IC ^ b© c <5 CO C^l o o 00 COO CO c^ ** CO CNt^ tQ O CN JO - ^5 cO CQ M M c8 o O St ^ 08 o5 . d a P pS S ^ S c3 O a> W* bC a o a; a; »1 ■^cot^-cocoi^-^-^^T^r^-^ict^ CN « o o a QC » (V 2 '53 o bfl g O a Q c3 T3 >> d o3 O •> G M [ J9qranu 1 • co o co co io a • »c t>- go r^- Tji Tt< tj . ^ _ ,-< CM ,_h ^H . CO CM CO CO CM CO i •COCM • • - • Crest in centimeters 2.0 <3> 00 ft • CO ft * ft ft * ft * ft !- »09 • • • • ft •CO - • *co •CO^H ' ^^HTji "• '• '• '• CO ft ■i> « •^00 - !*^rt< ■ '• '• '• CM • ■ ^ ' ■ H -OJHHCOOOOirH ■ ^00-^ '. '. '. '. • ■COt-hCMiOCMt^ • COCOOOCM CO ^ ^h ! if-HlO • • • « o ft • CM CM CO CM 00 00 • O Q) Q> ^ CM O »-h ■ ■ co^ • • • • ft • ^ CO CO i-H CO *0 • CO CO Oi CO ' •CM i-l « :^^ : : : : GO ft .^^^h^hc^^h *. CMiot^iO • :~ : :- ': ': ': ': • 1 'CO -*-h CM CO CM ' » • • • « ■ | , J9qumu CM CM CO • CO C5 • *o r^ CO t^- « Tji _ pH • -H ^ 'COCMCOCO < CO « 2.0 Anther length in centimeters • 00 • ft CO ■ CO • :- 1 1.5 CO • - : : CM -CM « ^CM « • ^ CM CM • CM CO -iOCMr-H >H* >H* >H >H • a +£.*».*. ^ c c fl • • . . o Hood Liverpool Harmonsburg Meadville Greenville Conewango Conewango Cherry Creek Villenova Hubbardsville Pownal Pownal Cent. Clarendon Middlebury ft/ New Haven Jc. Holderness Duxbury Alberton Ottawa L. Timagami Truro Antigonish Baddeck McAdam 1V3£ 1 1927 1927 1925 1925 1925 ^CMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCM 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 281 T3 CD a a © > < V, > ffl < X o o □Q GO K bfi o eg o c3 o joqinnu *ooo"^ o CO t^ CN CN CD CO o> go lO^H lOTf rt< CM CO jaqinna GO l> CD iC ^ CO CN O GO t^ CD 03 O CD CD CD O SSSSSo J*9 A 282 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN separating the two species. It is thus apparent at the outset that no biometric method of distinguishing the two species can be a simple matter. While it is certainly true, as Hall and Clements ('23) and McLeod ('26) have suggested, that taxonomy needs the development of more exact methods, there are serious limitations to a wholesale inclusion of biometry in every-day taxonomic procedure. These limitations may be easily demonstrated by a simple example. Figure 6 shows five petals of 7m virginica and five A B Fig. 6. A, outlines of five petals each of /. versicolor; B, outlines of five petals each of I. virginica, of Iris versicolor. For the purpose of taxonomic description they may be conveniently and accurately separated by describing those of Iris versicolor as ovate-lanceolate and those of 7ns virginica as 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 283 obovate them by biometric methods by no means so easy, as table v shows TABLE V Petal length Iris virginica No. 1 2 3 4 5 Petal width Petal taper Taper width cm. 1.4 1.7 1.7 1.3 1.6 cm. 1.5 0.8 0.9 1.1 1.2 1.1 0.5 0.6 0.9 0.8 Iris versicolor No. 1 2 3 4 5 4.7 4.8 4.1 3.0 4.5 1.3 1.5 1.2 1.2 1.4 2.0 1.7 1.3 1.2 1.7 1.5 1.1 1.1 1.0 1.2 Neither the length nor the width will suffice The taper (i. e. the length between the point of maximum width and the tip) somewhat better The ratio between this latter measurement and the width is still better though it fails to separate the entire lot. None of the measurements is as good for purposes of distinc tion as the single terms "ovate-lanceolate" and "obovate spatulate." Only by combining all three measurements into i complex ratio would it be possible to demonstrate, mathemat ically, the discontinuity between the two sets of Minot (quoted by Thompson, '17) has said of genius ha\ As n The fact that men evolved wonderful methods of dealing with numerical relations should not blind us to another fact, namely, that the observational basis of mathematics is, psychologically speaking, very minute compared with the observational basis of even a single minor branch of biology. * * * While therefore here and there the mathematical methods may aid us, we need a kind and degree of accuracy of which mathematics is lapable absolutely The above example illustrates the two chief reasons why biom- etry must necessarily be limited in its application to taxonomy. In the first place, mathematics is swift and efficient only in re- cording differences in number ; it becomes cumbersome in record- 284 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 differences in form, however useful it may be for a deeper ysis of the forces which produced the form. Yet it is just such differences in form which are most commonly met with in taxonomic work This point is shown in a survey of the which specific distinction is based in two representative families, (one from the Monocotyledons and one from the Dicotyledons) in the seventh edition of Gray 's ' Manual \ The differences were classified as being based on shape, on absolute size and number, and on comparative size and number. In all doubtful cases the preference was given to differences in number. Iridaceae Boraginaceae Differences in shape 21 31 Differences in absolute size or number Differences in comparative size and number 13 11 12 9 There is an even more fundamental reason why the customary methods of mathematics are not well adapted to taxonomic work. Such methods are comparatively simple when the variations in one characteristic are being traced; they become involved and cumbersome when the simultaneous changes number of variables are studied. m a large Yet this last is essentially the When he distinguishes between a method of the taxonomist. group of sugar maples and a group of silver maples, for instan he is summarizing a large number of differences— differences form, size, color, and Mathematics can deal with such problems through the study of correlation, but it is slow and laborious work and though it may be useful in the analysis of some particular problem it is not adapted to general taxonomic use. An attempt has therefore been made to develop a new method presenting biometric data which would combine the good methods of mathematics and comparative morphol- Like This of ogy. Like mathematics, it is accurate and object morphology, it leaves something to the trained eye has been accomplished by diagramming the data in a series of ideographs. Figures 7 and 8 show how the four measurements the petal and sepal are combined Essen- 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 285 Fig. 7. Diagram showing typical flower of /. virginica and resulting ideograph Fig. 8. Diagram showing typical flower of /. versicolor and resulting ideograph. tially the ideograph consists of a diagrammatic white petal, superimposed upon a diagrammatic black sepal. By construct- ing an ideograph of this sort for each plant measured, it is possible [Vol. 15 286 ANNALS OF THE MISSOURI BOTANICAL GARDEN to show simultaneously the variation in the four variables con- sidered, for an entire population of plants, or to compare the variation in one population with that in another as in figs. 10 to 13. Such ideographs would seem to be of general usefulness in taxonomic work. They are capable of demonstrating slight differences in proportion which are not revealed by figures alone. In the genus Iris, for instance, though Irisfulva, Iris foetidissima, and 7ns prismatica each belong to a different section of the genus, the three species have flowers of nearly the same size. The dimensions of the petals and sepals are so nearly the same that the species hardly appear distinct when the figures are compared, as in table vi. When these are arranged as ideographs, however, as in fig. 9, the essential differences in proportion between the three species are clearly demonstrated and they are seen to be morphologically distinct. TABLE VI COMPARISON OF FLORAL DIMENSIONS OF THREE SPECIES OF IRIS Sepal length Sepal width Petal length Petal width • in cm. in cm. in cm. in cm. 4.4 2.0 4.0 1.1 4.9 1.8 4.3 1.2 Iris prismatica 4.7 2.1 4.3 1.4 4.8 1.8 4.6 1.1 J 4.5 1.8 4.2 1.1 4.9 2.2 3.7 1.4 4.9 2.4 4.0 1.2 Iris fulva 5.6 2.6 4.5 1.6 5.3 2.4 4.1 1.3 5.0 2.2 3.9 1.1 1 4.3 1.8 3.9 0.9 3.9 1.6 3.6 0.8 Iris foetidissima 4.0 1.8 3.6 0.9 4.4 2.0 3.9 0.8 1 4.5 2.1 4.0 0.9 Ideographs for twenty individuals each of sixteen repre- sentative colonies are uped in figs- 10 fig. 13 (I. versicolor). They inff variation in size and or< L2 (/. virginica) and proof of the strik- which has been found studied. In marked contrast to the variation be tween individuals of the same specie the general resemblance between cc While several different colonies have 19281 ANDERSON — PROBLEM OP SPECIES IN IRIS 287 A B C Fig. 9. A, ideographs of five plants of /. fulva; B, ideographs of five plants of /. foetidissima; C, ideographs of five plants of J. prismatica. individual tendencies, and while in the case of those colonies which have been measured in successive years (fig. 10) the peculiarities nersist from vear to vear, there is practically no differentiation between one region and The onlv generalization that can be made is that Iris versicolor becomes on the average a little smaller as one goes from north to south and that 7m virginica becomes a little larger. Thus, although the colony at Stanton, Kentucky (fig. 12), seems slightly unusual by reason of its com- Darativelv large petals the other colonies studied in Kentucky and ities. ille and Camden) do not show these peculi much variation in proportion and almost 288 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 Lin wood, Mich. 1926 Linwood, Mich. 1927 Frankenmuth, Mich. 1926 Frankenmuth, Mich. 1927 Fig. 10. Ideographs of twenty individuals each of four colonies of /. virginica 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 289 1 Sunbury, Ohio Portage des Sioux, Mo. '26 Anna, 111. Eastover, S. C. Fig. 11. Ideographs of twenty individuals each of four colonies of /. virginica [Vol. 15 290 ANNALS OF THE MISSOURI BOTANICAL GARDEN Stanton, Ky. Bonnieville, Ky Jackson, Miss. Kimborough, Ala. Fig. 12. Ideographs of twenty individuals each of four colonies of /. virginica 1928] ANDERSON — PROBLEM OP SPECIES IN IRIS 291 Timagami, Ont Truro, N. S Ottawa, Ont. Hubbardsville, N. Y. Fig. 13. Ideographs of twenty individuals each of four colonies of /. versicolor 292 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN much in size in the colonies from southern Michigan as there is within the whole group of colonies of Iris virginica from the Great Lakes to the Gulf of Mexico. Above all, when the ideographs are considered as a whole, the two species remain completely and absolutely distinct. In spite of a wide range of variation in separate characteristics, when the combination as a whole is studied it is found to be strikingly TABLE VII MEDIAN MEASUREMENTS OF IRIS FLOWERS FROM DIFFERENT LOCALITIES IRIS VIRGINICA Number of individ- uals 25 21 50 35 21 20 15 38 27 15 38 43 23 40 30 28 27 26 30 28 23 63 33 30 30 30 35 25 45 43 19 35 30 30 25 25 52 27 33 Locality Kimborough, Ala. Jackson, Miss. Camden, Tenn. Bonnieville, Ky. Stanton, Ky. Eastover, S. C. Maysville, N. C. Anna, 111. Vulcan, 111. Vulcan, 111. Farmington, Ark. Wicks, Mo. Valley Park, Mo. Portage des Sioux, Mo. Portage des Sioux, Mo. Portage des Sioux, Mo. Louisiana, Mo'.. Fort Madison, la. Sunbury, O. Mill Creek, O. Huron, O. Bay Bridge, O. Middle Bass Is., O. Lawrence, Mich. Schoolcraft, Mich. Centerville, Mich. Hartland, Mich. Armada, Mich. Yale, Mich. Otisville, Mich. Otisville, Mich. Frankenmuth, Mich. Frankenmuth, Mich. Linwood, Mich. Linwood, Mich. Muskegon, Mich. La Crosse, Wis. Pardeeville, Wis. Slinger, Wis. Year 1927 1927 1926 1926 1926 1928 1928 1926 1925 1926 1925 1925 1926 1926 1927 1928 1924 1924 1927 1925 1925 1925 1925 1926 1926 1926 1926 1924 1924 1924 1926 1926 1927 1926 1927 1924 1924 1927 1924 Median Median sepal sepal length ■ width in cm. in cm. 6.3 2.8 5.9 2.6 6.7 3.0 6.7 3.0 5.5 2.6 6.7 2.8 7.5 3.0 6.3 2.8 6.3 2.6 6.3 2.6 6.3 3.0 5.9 2.6 5.9 2.8 5.9 2.8 6.3 3.0 6.8 3.2 6.3 2.8 5.9 2.8 6.7 3.0 5.1 2.1 5.5 2.4 5.1 2.4 5.1 2.6 5.9 2.8 5.9 2.6 6.3 2.8 5.9 2.8 5.5 2.2 5.5 2.4 55 2.4 55 2.4 5.9 2.6 5.9 2.8 5.9 2.9 5.9 3.0 5.5 2.4 5.9 2.6 5.5 2.8 5.9 2.4 Median petal length in cm. 5.7 4.9 5.7 5.7 4.9 5.7 6.1 4.9 5.3 5.3 5.3 4.5 5.3 4.9 5.7 5.8 5.3 4.5 5.3 4.1 4.1 4.1 4.5 4.9 4.9 5.3 4.9 4.1 4.5 4.5 4.5 4.9 4.9 4.9 4.9 4.5 4.9 4.9 4.9 Median petal width in cm. 1.8 1.8 2.0 1.8 1.6 1.9 2.0 1.8 1.8 1.6 2.2 1.6 1.8 1.6 2.0 2.0 1.6 1.6 19 1.4 1.4 1.4 1.6 2.0 1.8 1.8 1.8 1.4 1.7 1.6 1.8 1.8 1.8 2.0 1.8 1.6 1.6 1.8 1.8 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 293 IRIS VERSICOLOR Number of individ- uals 28 22 35 29 28 35 27 38 37 34 25 20 26 32 26 22 Locality Year Hood, Md. Liverpool, Pa. Harmonsburg, Pa. Conewango, N. Y. Villenova, N. Y. Hubbardsville, N. Y. Pownal, Vt. Pownal Center, Vt. Clarendon, Vt. New Haven Jet., Vt. Middlebury, Vt. Duxbury, Mass. Alberton, Ont. Ottawa, Ont. Timagami, Ont. |Truro, N. S. 1927 1927 1925 1924 1927 1925 1925 1925 1925 1926 1926 1927 1927 1926 1926 1927 Median sepal length in cm. 5.9 5.9 5.5 5.9 5.9 5.5 5.1 5.1 5.1 5.9 5.9 5.5 5.5 5.9 5.9 6.3 Median sepal width in cm. 2.8 3.0 2.8 3.2 3.0 2.6 2.2 2.4 2.6 3.0 3.0 2.8 3.0 3.0 3.2 2.8 Median petal length in cm. 3.7 4.1 3.3 4.1 4.1 3.3 3.3 3.3 3.3 4.1 4.1 3.7 4.1 4.1 4.1 4.1 Median petal width in cm. 1.2 1.4 1.2 1.8 1.6 1.0 1.0 1.1 1.0 1.5 1.4 1.4 1.6 1.4 1.2 1.4 constant. Iris versicolor remains always and unmistakably Iris color, and Iris virginica remains always and unmistakably Iris There is not the slightest tendency merge into the By employing another mathematical concept, that of the Dnstruct a different sort of ideograph and pro- duce something like a composite picture of each If for each colony we take the average length of sepal, average width of sepal, average length of petal, and average width of petal, and construct an ideograph we will obtain a figure which will present graphically the averages of all four measurements for the colony in question. be a purely hypothetical figure; it i kind of proportions most commo met with in the particular colony but it will serve for convenient comparison between separate colonies and between successive measurements of the same colony. As in the earlier presentation of averages, the median is used rather than the mathematical mean because it is less influenced by occasional extreme values. The median values and also the number of individuals measured are presented in table vn. The resulting ideographs are shown in fig. 14. The examination of these composite ideographs strengthens the conclusions already arrived at. The differences between colonies 294 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN ft********* *.*# Ciu«* U»««« ClMT*A>.ll, Nooft.Aa LivcBroob.fik H«*f«»ttA C»«f imj*», *t ViUfjiOv*. MX Hl*Mfttt«U.to fW««K^ &*«£«*¥*. CuA£***.». A.wJtMr,.* ^»tXBuat.tt 0«i»DR\>Utt. Ai.*r«r«MtOtC £rr«.rt.C.r. Tm*Q*m.Qm. T«.M» « *. Composite ideographs of 16 colonies of /. versicolor. Ovw*uA««'n AuMDNmriU H—M— t l U..*~4j L+wmmttt Composite ideographs of 39 colonies of /. virginica. Fig. 14 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 295 are very slight. In spite of the fact that the colonies measured extend from the Gulf Coast to the Great Lakes in the case of Iris virginica, and from Maryland to the north woods in the case of Iris versicolor, there is practically no evidence* of regional differentia- tion; that is, of the formation of morphologically distinct geo- graphical subspecies within either 7m versicolor or Iris virginica. As before, Iris versicolor is seen to become slightly dwarfed to- wards the South, whereas Iris virginica h smaller in the North. The general proportions of each species, however, remain strik- ingly similar throughout their entire ranges. There is no out- standing difference in proportion which characterizes any one region, though individual colonies may show slight peculiarities. Thus while the colony at Wicks, Missouri, is distinctive by reason of its proportionately small petals, that at Valley Park, Missouri, only a few miles away, is characterized by proportionately large ones, while the near-by colonies of Portage des Sioux, Missouri, and Vulcan, Illinois, have petals of about average proportions. In addition to differences in size and proportion, another sort of difference has been noted in the plants grown in the experi- mental garden. There is a general tendency, in both species, for the plants from northern colonies to come into bloom sooner than those from southern colonies. There is a good deal of variation in blooming time within each locality, of course, and this will mask the general tendency to a certain extent but for three springs it has been observed that the earliest plants to come into bloom are from northern colonies and the latest £,re those from southern ones. Aside from differences in size and in blooming periods no consistent regional differences have been found within either species. It is, of course, possible that such differences existed but were not revealed by the methods used. However, these methods did distinguish successfully between the two species. If regional differences do exist, they must be of an entirely different order of value from the differences between species. CHARACTERISTIC TENEENCIES Quite as remarkable as any clear-cut differences between the ro snecies were certain characteristic tendencies of one or the 296 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN other species to depart from the normal. The spathe valves of the two species are usually very similar, the longest ones of Iris versicolor and the shortest of Iris virginica being quite indistin- A B Fig. 15. range Variation in spathe valves of /. virginica, X X A- guishable morphol Those of Iris virginica, how become long and foliaceous like those s versicolor were never seen to do sc Iris often (in nearly 50 per cent of the cases examined) had notch in the petal. Only twice have such notches bee versicolor. due to mutilation, since they could be found in buds several days before opening; they were probably of the vascular system of More or less associated ruffled, crimped, or even crenate petals. Flowers characteristics are shown Only tendency observed in Iris versicolor and 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 297 then it was not highly developed. These are only a few of a host of variable characters which were occasionally expressed in one species and rarely or never in the other. There is nothing new in this experience. It has been the common lot of all who have attempted to distinguish groups of organisms, be they species or families or phyla. The quotation from Bergson which Mrs. Arber ('25) uses in distinguishing between the Monocotyledons and the Dicotyledons is quite as appropriately used here as there. "Tie group must not be defined by the possession of certain characters but by its tendency to emphasize them. " The point in calling attention to this commonplace phenomenon is to suggest that if most of the differences between species are of this sort, taxonomists might well undertake surveys of variation within the species they are treating, on a scale much larger than is customary at the present time. Such studies should be particu- larly important in any attempt at a phylogenetic treatment of related species. 7ns virginica is closely related to 7ns hexagona, for instance, and one of the morphological indications of this relationship is that in about one case out of fifteen or twenty the spathe valves of 7ns virginica are foliaceous like those of 7. hexa- gona. PECULIAR FORMS A few of the individuals observed have been so unusual as to deserve special mention. Albinos. — Complete albinos with no trace of color other than yellow have been found only in 7. virginica. These were found at two places, at Lawrence, in western Michigan, and near Lake Saint Clair, in Ontario. Partial albinos are fairly common in both species. They are white, shaded or faintly lined with blue. They have been observed at the following localities : 7. virginica — Lawrence, Michigan (in same colony with albino) ; Farmington, Arkansas; Maysville, S. C; Dennison, Mich. 7. versicolor — Timagami, Ontario; Central Connecticut (M. E. Mains). Occasional individuals were found in each species in which the flower color is lavender or red-purple instead of blue. One such 298 ANNALS OF THE MISSOURI BOTANICAL GARDEN Vol. 15 form has been offered in flower catalogues for some years under the name of i". versicolor kermisina. Similar color forms have been found at the following localities : /. virginica — Otisville, Mich.; Burke, Iowa; Camden, Tenn. J. versicolor — Stratford, N. H.; Hubbardsville, N. Y. Occasional individuals were found with sepals strongly re- flexed. Three flowers from such an individual are illustrated in pi. 39, figs. 7-9. THE GENETIC RELATIONSHIPS OF INDIVIDUAL DIFFERENCES IN IRIS VERSICOLOR An experiment is under way to determine, for each species, the extent to which the peculiarities which characterize individual plants are passed on to their offspring, under the conditions obtaining in nature; in other words, to determine, from an inspection of the variation between sister seedlings, the amount of interbreeding which takes place under natural conditions. If the species is naturally self-fertilized ninety-nine times out of a hundred, all seedlings grown from a single parent plant should be practically alike and similar to the parent. If, however, there has been continuous out-crossing with other individuals, the sister seedlings should differ from each other and from the parents; as much, for instance, as human brothers and sisters differ from each other. Several such series of sister seedlings from seed-pods collected on plants of Iris versicolor at Connecticut Lakes, N. H., bloomed in 1928. There were not enough flowers per plant to make careful qualitative studies but there were sufficient to allow certain preliminary conclusions to be drawn. One of these series is illustrated in pi. 40, and shows one flower from each of five sister plants. It will be noted that while the flowers are in general very similar, they are not as much alike as flowers from the same plant (see pis. 37, 33 and 39, where series of flowers from the same plants are illustrated). They are, however, much the general run of individual ( The most conspicuous differences are in regard to the Those of figs. 9, 5, and 6 (pi. 40) are unusually broad and flat, those 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 299 of number 7 are somewhat smaller, while number 8 's are narrow and folded. There were accompanying differences in color which are only suggested by the photographs. Number 8, for instance, was distinctly lilac, while the others were blue. Similar results have been obtained in all the series of Iris versicolor which have blossomed to date. It is therefore concluded that 7. versicolor, though usually self-fertilized, is frequently cross-pollinated under natural conditions. The relation of this conclusion to the ques- tion of Jordanons or micro-species, is deferred to the Discussion. Hybridization of Iris versicolor and Iris virginica a preliminary report It has been found upon experiment that 7m versicolor and Iris virginica are partially fertile in crosses with each other. There has not yet been enough material in the breeding plot to make exact quantitative studies, but whereas nearly all the crosses which have been tried within either species have set seed, only Fig. 16. Ideographs of three hybrid plants; left, ratural hybrid from Engadine, Mich.; center, artificial hybrid produced at the Missouri Botanical Garden; right, natural hvbrid from London. Ontario. slightly more than half of those between the two species have done so. The resulting hybrid seed has a low percentage of germina- and the seedlings show hybrid This noticeable when the young seedlings were grown in the cold-frame side by side with seedlings of the The seedlings [Vol.. 15 300 ANNALS OF THE MISSOURI BOTANICAL GARDEN of Iris versicolor and of 7ns virginica varied greatly in vigor, whereas the hybrid seedlings were all uniformly vigorous. Of the various crosses which have been made, two have flowered. They are reciprocal crosses between plants I-AAA-3 and IAAG. I-AAA-3 is a plant of Iris virginica collected at Flint, Michigan, in 1923, and is illustrated in pi. 41, right figures. It is large- flowered and the flower has bright blue veining on a light back ground. The petals are unusually broad. IAAG is a plant received from Dr. J. R. McLeland of Pleasanton, Kansas, though it came originally from Medina, N. Y. It is a typical Iris versi- color and is outstanding in the dark color of its flowers and its stems. 1 1 is illustrated in pi. 41 , left figures. The seedlings of cross IXB (I-AAA-3 X IAAG) were similar in size, leaf number, leaf height, etc., to the most vigorous plants of either parent species. The seedlings of the reciprocal cross, IXA (IAAG X I-AAA-3) were even more vigorous. Plants one and two years old grown from seed showed leaves one fourth again as high and about half again as many leaves per plant as the parents. IXA and IXB flowered for the first time in 1927. Ideographs and photographs of their flowers and of those of the parent species are shown in fig. 16 and pi. 41. It should be emphasized that the measurements and pictures of IXA and IXB were made from plants growing in the cold-frame. They would have been much larger if grown in the water-side plots with the other plants. Thirteen plants of IXA and three plants of IXB were brought to flowering age. The two sets of hybrids were very similar aside from the greater vigor of IXA. While the seedlings varied slightly among themselves they were remarkably uniform in general aspect. They were intermediate in all the differences which characterize the two parents, though they were easily distinguished by their vegetative vigor. In general appearance and in practically all measurable characters they much more closely resembled Iris versicolor than Iris virginica. Their most outstanding differences from Iris versicolor were their much larger and laxer petals and the peculiar spots at the base of the sepal. The spots were due to the combination of the bright yellow pubescent spot of Ms virginica and the dark blue cross- 1928 ANDERSON — PROBLEM OF SPECIES IN IRIS 301 veining and brown stippling of Iris versicolor. The resulting "eye" is unlike anything in either species. Fig. 17. Cross-sections of sepals showing pubescence: A, I. virginica; B, I. X robxusta; C, I. versicolor; X 25. A microscopical examination of the pubescence of the sepal showed that of the hybrid to be intermediate in form between the two species though, due to the larger cells of the hybrid, it is almost as long as that of Iris virginica. Camera-lucida drawings of all three are shown in fig. 17. The hybrid was likewise inter- mediate in the glandular development at the base of the stamens. Figure 18 shows cross-sections of the perianth tubes. Fig. 18. Cross-sections of perianth tubes showing, a, style, and b, glandular development of inner wall of tube; left, Iris versicolor, right, /. virginica, center, /. X robusta. The of crosses IXA and IXB were partially sterile. Counts showed about 50 per cent abnormal pollen and many of the ovules aborted. Ovaries with one exposed locule are shown in pi. 41. slants from which the capsi same cold-frame and had had for pollination. Al small and shrunken of the seed capsules on the hybrid umerous other hybrids between the two species have been 302 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 made, and while they have not flowered they are similar in their vegetative characteristics to those already described. Dr. J. R. McCleland, an iris amateur of Pleasanton, Kansas, has grown several crosses between the two species, some of which I have seen, and they are similar to IXB and IXA. Natural hybrids resembling those produced in the experimental plot have been found at several points where the ranges of the two species overlap. They would be commoner were it not for the barriers which exist between the two species; geographical barriers such as the extensive limestone areas west of the Alle- ghenies in which neither species is common; physiological barriers which prevent them from being wholly fertile with each other. The location of the hybrids and their relation to the distribution of the parent species is shown in fig. 19. Most of the hybrids found were very similar to those produced in the experimental plot. Those observed in Ontario resembled the garden-made hybrids very closely. There was little variation between them although they were found in several places. They were nowhere particularly abundant except in one creek bed just west of London, Ontario, where there was a large colony, appar- ently all belonging to one clone. Iris virginica was fairly common throughout this district. Iris versicolor was not found there though it was fairly common a short distance east of London. At Yale, Michigan, a single plant of Iris versicolor and a few hybrids were found. They were in a meadow near a small graveyard from which /. versicolor may have spread. There were no plants of Iris virginica in the immediate neighborhood, though it is common in eastern Michigan and there were large colonies within two miles. An extensive search of the neighborhood failed to locate any other plants of Iris versicolor or of the hybrids. At Tawas City, Michigan, a single hybrid plant was found growing along the roadside among a group of plants of Iris versi- color. A colony of Iris virginica was found less than half a mile away. At only two points were the hybrids found making up large colonies. These were both in the northern peninsula of Michigan, and though only a few miles apart they were so different in their composition that a separate description will be given of each. 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 303 2* if? CD d 5 2. cr c o P O P 5 fie E I p S 9 S B X 3 s- ST ST. o r o g* 304 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 WEST ST. IGNACE Along the main highway three miles west of Saint Ignace a very remarkable colony was located. It had apparently originated by the intermingling of the two species, followed by numerous crosses and back crosses. The ideographs in fig. 20 give a slight idea of the bizarre mixture of types which was encountered; there was everything from typical versicolor to typical virginica with numer- ous intermediates and a number of very curious forms unlike any- thing seen before either in wild colonies or in the experimental plot. Most of the plants exhibited a high degree of vegetative vigor, though the environment did not seem to be unusually favorable. ENGADINE At Engadine, a few miles west of West St. Ignace another peculiar colony was found. It was most remarkable by reason of the great vegetative vigor of the plants. There was a degree of variation between the plants quite comparable to that found in colonies of the parent species, but by no means so extreme as that found in the colony at West St. Ignace. The plants were all very similar in their general aspect to the hybrids produced in the experimental plot. Ideographs of the colony show its lesser variability, and show the general similarity to the experimentally produced hybrid in fig. 16. Though admittedly difficult, if not impossible, to distinguish by means of ordinary herbarium material, the hybrid appears of sufficiently common occurrence to warrant a special name. X Iris robusta E. Anderson, 1 new hybrid (Iris versicolor X I. virginica) Intermediate between Ins versicolor and /. virginica, but more robust; partially sterile. V. Summary The common blue flags of eastern North America are found to be made up of two separate and distinct species; a northern and 1 Iris robusta E. Anderson (Iris versicolor X /. virginica) Virginicam et versicolorem intermedia sed robustior, partialiter sterilis. A Iris virginica B Iris versicolor C West St. Ignace, Mich. D Engadine, Mich. Fig. 20. Ideographs of twenty individuals each of the two hybrid colonies studied: C, West St. Ignace, Mich.; D, Engadine, Mich.; above, average colonies of /. virginica and /. versicolor for comparison. 306 ANNALS OF THE [Vol. 15 eastern one, Iris versicolor L., and a southern and a middle- western one, Iris virginica L. Statistical studies have been made of 10 colonies of Iris versi- color mid 39 of Iris virginica. Each species is found to be a natural group. In spite of wide variation within each species there is no tendency whatsoever for one to merge into the other, and the general average of each species is practically the same wherever it is studied. Although variation between individuals is met with in every colony examined it has not resulted in any appreciable regional differentiation, since the species in question moved into their present home at the close of the glacial period. Iris versicolor and Iris virginica are kept separate in part by natural geographical barriers and in part by physiological differ- ences which prevent them from being perfectly fertile with each other. .Along the narrow zone where the two species come in contact hybridization occasionally takes place. Natural hybrids, closely resembling those produced experimentally, have been found at five points. At two of these localities hybrid populations of considerable size had arisen and have been studied in greater detail. There is some evidence that constant new forms might originate in this manner. VI. Discussion There are several objections which must be met before the above conclusions can be accepted as generally applicable. In the first place it might be argued that there were regional differ- ences present within the two species which were not revealed be- cause of the methods used, the characters chosen for study, etc. It is quite possible that such unrecognized differences existed, but if so they must have been of an entirely different order from the differences between species. The methods used did distinguish effectively between Iris versicolor and Iris virginica and did demonstrate that no such differences existed within either of the two species. For the material studied the conclusion therefore seems unavoidable that the differences within species are of an entirely different nature from the differences between them. A more valid objection to the general application of the con- clusions drawn in this work is that the two species belonged to an 1928) ANDERSON — PROBLEM OF SPECIES IN IRIS 307 old and well-defined genus and that different relations between species might be found to obtain in such a group as the genus Aster, for instance. From purely a priori reasoning it would seem quite possible that distinctions between species might be of a different nature in different parts of the vegetable kingdom. Conclusions drawn from a study of two species belonging to the Monocotyledons would certainly have to be confirmed with more closely related material before they could be accepted as valid for such a distantly related group as a genus of the Compositae. Another hindrance to the general application of the conclusions reached in the study of these two species is the geophysical nature of the region in which they occur. It is without great barriers of any sort. With a greater degree of isolation, such as would occur in a region cut by great mountain chains, the differences within species might perhaps be very different. It would be particu- larly interesting to compare the regional differentiation of the closely related western species, Iris missouriensis . LINNAEAN VS. JORDANIAN SPECIES The main conclusion drawn from this Investigation is that the Linnaean species is a natural and permanent group. It should therefore be the most effective one for purposes of classification. In regard to Jordanian species I am in complete agreement with Clausen ('27) when he says, " The old botanists worked according to their own common sense and delicate biological feeling, but they had a conception of species far more nearly coinciding with what we now arrive at by careful statistical observation of var- iation in the field and by cytological investigations and crossing experiments than our modern small-species taxonomists. 'Eine Art ist eine Art, ganz gleichgultig ob sie Diapensia lapponica, Viola tricolor, oder Hieracium margineUiceps heisst', says du Rietz. I quite agree that it is a matter of supreme indifference to Nature what we decide shall be the definition of a species. But it is by no means a matter of indifference to ourselves whether we accept or reject a form of nomenclature which obscures Nature's own chief system of division. " In my opinion those who have considered the Jordanon to be of prime importance, taxonomically and phylogenetically, have [Vol. 15 308 ANNALS OF THE MISSOURI BOTANICAL GARDEN ascribed undue significance to the fact that it comes true from seed. With more genetical or horticultural experience they would have realized that coming true from seed (homozygosity) is a mere corollary of the amount of inbreeding which has taken place and that it is of minor taxonomic and phylogenetic signiiicance. We can best demonstrate the actual significance of the Jord- anon's true-breeding quality by examining the relationships between individuals in two different sorts of species; in one continuously cross-pollinated and in one continuously self-pol- linated. These relations are diagrammed in fig. 21 at a and b. Let us take four individuals of a continuously cross-polli- nated species, such a one as Aster anomalus, for instance, to name one that has actually been investigated (unpublished data). We may diagram these four individuals as A, B, C, and D (b, fig. 21) and if we examine them they will be found to possess certain inherent differences in such characteristics as the number, length, and shape of the ray flowers, the number and arrangement of the branches of the inflorescence, etc. Though two individuals may sometimes agree in a single characteristic, no two will possess the same combination of characters. In the next generation, however, none of these particular combinations will reappear. Plant A will be pollinated by pollen from other plants, as, for instance, by B as in the diagram, and the resulting offspring will show new combinations of characters unlike anything in the previous generation. In the diagram one of them is taken as an example and named L. Plants M, N, and O are similar new combinations arising from cross-pollinations between the other plants of the previous generation. There will be similar new recombinations of characteristics for each new generation. In other words, there will be just such a reshuffling of characteristics as occurs from generation to generation in human families; man being another species in which similar out-crossing prevails. If we now turn to a self-pollinated Linnaean species and study the differences between individuals we will find that they are very similar to those studied in Aster anomalus. We may therefore diagram them similarly (a, fig. 21) as A, B, C, and D. In such a self-pollinated species, however, a single individual acts as father and mother for the next generation. There will ordinarily 1928] ANDERSON — PROBLEM OF SPECIES IN IRIS 309 be no crosses between A and B or B and C as there were in Aster anomalus. Furthermore, since such close: inbreeding has obtained in the past, each individual will be practically homozygous (pure- breeding), and its progeny will resemble their parent and each other very closely. The seedlings of A will be so similar to their parent that they can be diagrammed as A, those of B will show the same combination cf characters which dis- tinguished B from A, C, and D, and may be diagrammed as B. This resemblance will continue from generation to generation as long as no cross-pollination occurs. Such a species is therefore divisible into a number of pure lines or Jo::danons. 4 B C V A MM kHn ? 9 ? £? CO C 3 o o 73 2| 33 x to a • ,•« v^ *v ♦ • ■ i ■ * 318 [Vol. 15, 19281 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 37 Flowers of Iris virginica, X Fig. 1. Plant ABC — collected at Orchard Farm, Mo Fig. 2. Plant ABC — collected at Orchard Farm, Mo Fig. 3. Plant ABC — collected at Orchard Farm, Mo Fig. 4. Plant ABC — collected at Orchard Farm, Mo Fig. 5. Plant ABC — collected at Orchard Farm, Mo Fig. 6. Plant ABE — collected at Fort Madison, la. Fig. 7. Plant ABE — collected at Fort Madison, la. Fig. 8. Plant ABE — collected at Fort Madison, la. Fig. 9. Plant ABQ— collected at Gilbertville, la. Ann. Md. But. Gabo.. Vol. 15, 1U2S Plati; 37 AXDEHSoX— PROBLEM OF SPECIES IX I HIS [Vol. 15, 1928] 320 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 38 Flowers of Iris versicolor, X ^ Fig. 1. Plant ABY— collected at Cohasset, Mass. Fig. 2. Plant ACF— collected at New Haven, Conn. Fig. 3. Plant ACF— collected at New Haven, Conn. Fig. 4. Plant BBG — collected at Harmonsburg, Pa. Fig. 5. Plant collected at Cedar Lake, Nova Scotia. Fig. 6. Partial albino, var. " Stella Main"— collected in Connecticut. Figs. 7-9. Seedlings grown at the Missouri Botanical Garden fron at Connecticut Lakes, N. H. seed Ann. Mo. Bot. Gard.. Vol. 15, 1928 Plate 38 ANDERSON— PROBLEM OF SPECIES IN IRIS 322 (Vou 15, 1928) ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 39 Flowers of Iris virginica, X Figs. 1-2. Plant ABB— collected at Valley Park, Mo Fig. 3. Plant ACZ— collected at Fish Creek, Wis. Fig. 4. Plant ABT— collected at Yale, Mich. Fig. 5. Plant ABT— collected at Yale, Mich. Fig. 6. Plant ABT— collected at Yale, Mich. Fig. 7. Plant ABP— collected at Otisville, Mich. Fig. 8. Plant ABP— collected at Otisville, Mich. Fig. 9. Plant ABP— collected at Otisville, Mich. Aw. Mo. Hot. Gard., Vol. 15. 1928 Plate 3<» ANDERSON— PROBI EM OF SPECIES IX IRIS 324 [Vol. 15, 1928) ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 40 One flower each of five sister plarts of Iris versicolor grown at the Missouri Botani- cal Garden from a seed capsule collected at Connecticut Lakes, N. H. Ann. Mo. Bot. Gabd., Vol. 15, 1928 Plate 40 JZBDB IZBDB IZBDB IZBDB I2LBDB WG^.wtv ■M ANDERSON— PROBLEM OF SPECIES IN IRIS 326 [Vol. 15, 1928] ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 41 Iris between them. Top row, immature capsules with one locule exposed. Middle row, immature capsules. Bottom row, flowers. rts ns (Iris X robusta) Ann. Mo. Bot. Gard., Vol. 15, 192S Plate 41 ANDERSON— PROBLEM OF SPECIES IN IRIS [Vol. 15, 1928 328 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 42 Fig. 1. Typical flower stalk of Iris virginica. Fig. 2. Typical flower stalk of Iris versicolor. Fig. 3. Iris virginica photographed at Huntingdon, Tenn. Fig. 4. Iris virginica photographed at Wilmington, N. C. Ann. Mo. Hot. Gard., Vol.. 15, 1928 Plate 42 AXI)KUS(.)N— PROBLEM OF SPECIES IN IRIS [Vol. 15, 1928 330 ANNALS OF THE MISSOURI BOTANICAL GARDEN Explanation of Plate PLATE 43 Figs. 1-5. Iris virginica. Fig. 1. Collected seeds from nine localities, X H- Fig. 2. Seed surface (St. l^ouis, Mo.), X 30. Fig. 3. Seed surface (Catawba, Ohio), X 30. Fig. 4. Seed, X 7. Fig. 5. Lining of seed capsule, X 30. Figs. 6-10. Iris versicolor. Fig. 6. Collected seeds from nine localities, X V% Fig. 7. Seed surface (Sagamore, Mass.), X 30. Fig. 8. Seed surface (Conewango, N. Y.), X 30. Fig. 9. Seed, X 7. Fig. 10. Lining of seed capsule, X 30. Ann. Mo. Hot. Gabd., Vol. 15, 1928 Plate 43 4 f •• • • # V * % *► <# ^J % ) «•* f • 4 i # 4 Albert Ltittim m w % • Frw/terfrtirtdfliih. p 0T f /f yf orf A/.Scrft.'*- 0^««Sa»^l3*t Cwnecticui Ukex.WH 4 % > ; M ^ * <* * ■""* % 7a*a^ £ifV ^f/c/>. Fostoxix flith * * >> La. CroW, \V7i c #♦ «% • Stt*ffrrJ, #■ H A | CjlU**A, 0ll/o /?**a, I //.'*>. 3 Be^ievJft, fy. N/ An tu.c.Hct ; Mali I 2 3 ->- *, ; Ik ^*_ ft 4 K 5 .->: 6 7 8 9 10 A \ % ^ 9 % > 4 4 « 4 f 4*4 Or W, A/. A * % # * * * 4 O * 4 4 & 4 4 SfetuJbftM Cy, WY Cone*an90.N.V ■A- X * ■A :*» ->, "»*4>. -*■- *** 4» * «* <* ^ * ** - *♦ * *»* *•* * *■ ■: *4x * -* Hi I *MP ANDERSON— PROBLEM OF SPECIES IN IRIS [Vol. 15, 1928] 332 ANNALS OF THE MISSOURI BOTANICAL GARDEN Fig. I. Explanation of Plate PLATE 44 o Dlants of Iris virainica from V cinders along railroad track; lower row, two plants from the same colony but growing in rich swamp. Fig. 2. Large single clone of Iris virginica> Maysville, N. C. Fig. 3. Colony of Iris virginica at Camden, Tenn. Ann. Mo. Hot. Gard., Vol. 15, 1928 Platk 44 > O !Z5 C C ^ fd x Annals of the Missouri Botanical Garden Vol. 15 NOVEMBER, 1928 No. 4 A NEW VARIETY OF VERNONIA LINDHEIMERP ESTHER L. LARSEN Instructor in Botany in the University of Montana Formerly Jessie R. Barr Fellow in Botany, Washington University A very interesting Vernonia was collected by Roxana S. Ferris and Carl D. Duncan along the Sanderson-Sheffield Road, twelve miles from Sanderson, Terrell County, Texas, July 19, 1921. The plant was distributed as Vernonia Lindheimeri Gray & Engelmann; but upon critical study and comparison with a relatively large suite of specimens representing this species in the Missouri Botanical Garden Herbarium, the Ferris and Dun- can plant shows such marked variation from the type of the species that it seems worthy of recognition as an outstanding variety. A description is recorded as follows: Vernonia Lindheimeri Gray . 1924. Leaves glabrous or glabrate above, beneath barbate in the axils of the midvein, or glabrate; calyx-lobes 3-5 mm. long, greenish. Distribution: waste-land, and hedgerows, general over central and southern Mexico. Specimens examined: Mexico: Chihuahua: Tierras Verdes, May, 1891, Hartmann 534 (G); southwestern Chihuahua, Aug.-Nov. 1885, E. Palmer M (G, US); Batopillas, April, 1892, Hartmann 1032 (G). Sinaloa: Ymala, Aug. 16-25, 1891, E. Palmer 1470 (US type); Mazotlan, April 5, 1910, Rose, Standley & Russell 14064 (US); vicinity of Rosario, April 14, 1910, Rose, Standley & Russell 14544 (US); San Ignacio, June 19, 1918, Monks & Salazar 405 (US); Guadalcupe, April 18, 1910, Rose, Standley & Russell 14675 (US); La Cruz, 1921, Ortega 4175 (US); between Rosario and Concepcion, July 27, 1897, Rose 3260 (US); San 1928 WOODSON — STUDIES IN APOCYNACBAE. II 355 Ignacio, March 12, 1918, Montes & Salazar 268 (US); Rosario, July 8, 1897, Rose 1578 (US); Colomas, July 16, 1897, Rose 1688 (US). Jalisco: Baranca, near Guadalajara, June, 1886, E. Palmer 132 (G, US); Chiquilistlan, May 15, 1892, Jones 335 (MBG, US) ; Tequila, July 5-6, 1899, Rose & Hough 4777 (US) ; Baranca, near Guadalajara, May 28, 1891, Pringle 5151 (G); vicinity of Colima, April 5, 1897, Seler 3436 (G); Baranca of Guadala- jara, alt. 4000 ft., June 10, 1898, Pringle 6872 (G, NY, F, ANSP, US, MBG); Guadalajara, June 25, 1892, Pringle 5363 (G); San Sebastien, Jan. 15, 1927, Mexia 1490 (US); Bolanos Aug. 10-19, 1897, Rose 2888 (US). Durango: Chocala, March 7, 1899, Goldman 858 (US). Morelos: Cuernavaca, May 11, 1898, Pringle 6847 (US). Nayarit: Ojos de Agua, near Ixtlan, Sept. 23, 1926, Mexia 733 (US). Popular names of this variety are "Berrarco," "Berraco 6 Tapaco," and the gum of the fruit is said to be used like chicle {Monies & Salozar 405, US). The embarrassment of monographers who find themselves forced to regard as " typical" an anomalous form of a species because of priority in publication over a more common variety is illustrated in a peculiar fashion by Stemmadenia tomentosa Greenm. and its var. Palmeri. As early as 1891 the herbarium name " Stemmadenia Palmeri Rose" was distributed with speci- mens of the glabrcscent or barbate variety of the former species. The name did not appear in publication, however, until 1893, when Urbina, in compiling his 'Catalogue of Mexican Plants,' happened upon specimens of the genus bearing the inscription of S. Palmeri Rose in a rather poor script, and erroneously published the name for the first time as a nomen nudum. Ur- bina mistook the name of Dr. Rcse for an abbreviation, and gave the author as "Kosc." It is indeed fortunate that a de- scription was not included under that authorship. In 1900 Dr. Greenman published Stemmadenia tomentosa, and in so doing spoke of the characteristics of £. Palmeri Rose, which he evidently assumed to be a correctly published name. It was not until 1924 that Stemmadenia Palmeri was published (Vol. 15 356 ANNALS OF THE MISSOURI BOTANICAL GARDEN by Rose in Standley's 'Trees and Shrubs of Mexico.' The legal place of publication of the species must evidently be regarded as ex Greenman, Proc. Am. Acad. 35: 311. 1900. 2. Stemmadenia sinaloana Woodson, n. sp. 1 PI. 48, fig. 1. Shrubs or small trees; leaves 8-12 cm. long, 5-6 cm. broad, glabrous, or very slightly puberulent upon the lower surface, petiolate, the petioles 7-10 mm. long; inflorescence 1-4-flowered; corolla yellow, the proper-tube 1 .5-2 cm. long, the proper-throat conical, about 1.5 cm. long, about 1.5 cm. broad at the orifice, the limb 1.5-2 cm. long; calyx about one-sixteenth the length of the proper tube, the segments ovate-reniform, 1.2 mm. long, about 4 mm. broad, obtuse at the apex, or completely rounded, scarcely imbricated, unequal, greenish; follicles unknown. Distribution: known only from the type locality in Sinaloa. Specimens examined : Mexico: Sinaloa: Rosario, Jan. 1895, Lamb 467 (G type). S. sinaloana is especially noteworthy in the genus Stemmadenia by reason of its peculiarly reduced calyx. In that respect it is closest related to S. tomentosa Greenm., from which it differs having a calyx less than one-half as large (1 mm and in having the calyx-lobes subreniform and rounded at the apex instead of oblong and ovate with acute or acuminate apex as in the latter sDeeies. 3. Stemmadenia glabra Benth. Bot. Voy. Sulph. 124. L 44. 1844; Hemsl. Biol. Cent.-Am. Bot. 2: 310. 1881; Miers, Apoc. S. Am. 74. 1878; K. Sch. in Engl. & Prantl, Nat. Pflanzen- fam. 4 2 : 149. 1895; Standi. Contr. U. S. Nat. Herb. 23: 1156. 1925; Standi. & Calder6n, Lista Pr61im. PI. Sal. 174. 1925. PI. 47, fig. 1. Shrubs or small trees, 2-10 m. tall; leaves 14-20 cm. long, 7-8 cm. broad, glabrous, petiolate, petioles 5-10 mm. long; 1 Stemmadenia sinaloana sp. no v., arborea glabra vel subpuberulenta; foliis ob- longo-lanceolatis 8-12 cm. longis 5-6 cm. latis; petiolis 7-10 mm. longis; corollae tubo conico-infundibuliformo 3-3.4 cm. longo, lobis ca. 1.5 cm. longis; calycis lobis parvis ovato-reniformibus inaequalibus ca. 2 mm. longis ca. 4 mm. latis obtusis, viridibus— Sinaloa, Rosario, Jan. 1895, F. H. Lamb 467 (Gray Herb., type). 1928 WOODSON — STUDIES IN APOCYNACEAE. II 357 inflorescence 1-4-flowered ; corolla deep yellow, the proper-tube 2-2.5 cm. long, the proper-throat conical, about 2 cm. long, 2-2.5 cm. broad at the orifice, the limb 2.5-3 cm. long; calyx about equalling the length of the proper tube, the segments 1.5- 2.5 cm. long, .8-1.0 cm. broad, strongly imbricate in two unequal series, the larger yellow, the smaller greenish yellow; follicles about 5 cm. long, 3-3.5 cm. broad. Distribution: tropical forests and thickets, Central America. Reported also from Mexico. Specimens examined : Costa Rica: between San Pedro de Montes de Oca and Cur- ridabat, Dept. San Jose, Feb. 2, 1924, Standley 32798 (US); Cartago, Feb. 1924, Standley 36469 (US). Honduras: Amapala, Isla de Tigre, Feb. 14, 1922, Standley 20713 (US). El Salvador: vicinity of La Uni6n, Dept. La Uni6n, alt. 150 m., Feb. 13-21, 1922, Standley 20686 (G, NY, US); Laguna de Magugue, Dept. La Uni6n, alt. 60 m., Feb. 18, 1922, Stand- ley 20948 (G, NY, US); La Uni6n, Sept. 21, 1860, Sutton-Hayes (G) . Nicaragua: southwestern slopes of Santiago Volcano, near Masaya, alt. 300-480 m., July 5, 1923, Maxon 7647 (G, US); Ometepe Island, Jan. 1893, C. L. Smith (G); Managua, shores of Lake Managua, June 24, 1923, Maxon, Harvey & Valentine 7270 (US) ; Managua, vicinity, June 30, 1923, Maxon, Harvey & Valentine 7589 (US); Laguna de Masaya, July 6, 1923, Maxon 7727 (US). Dr. Sutton Hayes remarks {Sutton-Hayes, G) that the popular name of this species in El Salvador is "Cajon del Mico." Ac- cording to Standley {Standley 32793, US), the popular name in Costa Rica is "huevos de Caballo," or "Girijarro," and the sap is used for corns and tooth-ache. 4. Stemmadenia obovata (Hook. & Am.) K. Sch. in Engl. & Prantl, Nat. Pflanzenfam. 4 2 : 149. 1895. Bignonia {?) obovata Hook. & Arn. Bot. Beech. Voy. 439. 1841. Stemmadenia pubescens Benth. Bot. Voy. Sulph. 125. 1844; 358 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 1 S Miers, Apoc. S. Am. 74. 1878; Hemsl. Biol. Cent.-Am. Bot. 2:310. 1881. Shrubs or small trees, 2-15 m. tall; leaves 10-20 cm. long, 7-10 cm. broad, pubescent, or glabrate above, petiolate, peti- oles 5-8 mm. long; inflorescence 1-6-flowered; corolla deep yel- low, the proper-tube 1.5-2.5 cm. long, the proper-throat 1.5-3 cm. long, 2-2.5 cm. broad at the orifice, the limb 1.5-2.5 cm. long; calyx much surpassed by the length of the proper-tube, the segments 1.5-2 cm. long, .8-1.0 cm. broad, strongly imbri- cated in two unequal series, both series yellowish; follicles 4- 4.5 cm. long, 3-3.5 cm. broad, acute at the apex. Distribution: tropical forests and thickets, southern Mexico and Central America. Specimens examined: Mexico: Guerrero: El Correjo, alt. 900 m., May 18, 1899, Langlassi 1029 (G). Costa Rica: Salinas, July, 1890, Pittier 1177 (US). Nicaragua: Managua, vicinity, June 30, 1923, Maxon, Har- vey & Valentine 751$ (US); La Paz, Dept. Leon, Jan. 31, 1903, Baker 2270 (G, US); Managua, June 30, 1926, Chaves 215 (US). El Salvador: near La Cebadilla, 1922, Calderon 1230 (G); Laguna de Olomega, Dept. San Miguel, alt. 75 m., Feb. 20, 1922, Standley 21034 (G). 4a. Var. mollis (Benth.) Woodson, n. comb. Stemmadenia mollis Benth. Bot. Voy. Sulph. 125. 1844; Hemsl. Biol. Cent.-Am. Bot. 2: 310. 1881; Miers, Apoc. S. Am. 75. 1878; K. Sch. in Engl. & Prantl, Nat. Pflanzenfam. 4«: 149. 1895; Urbina, PI. Mex. 214. 1897; Donn.-Sm. Enum. PI. Guat. 4: 105. 1895; Areschoug, PI. ca. Guayaquil Coll. 127. 1910; Standi. Contr. U. S. Nat. Herb. 23: 1156. 1924; Standi. & Calder6n, Lista Prelim. PI. Sal. 174. 1925. Stemmadenia calycina Brandg. Univ. Cal. Publ. Bot. 10: 188. 1922. Upper surface of leaves persistently tomentose. Distribution: tropical forests and hedgerows, southern Mex- ico, northern Central America, and northwest-central South America. 1928) WOODSON — STUDIES IN APOCYNACEAE. II 359 Specimens examined : Mexico : Vera Cruz: Bafios del Carrizal, Aug. 1912, Purpus 6230 (G, NY, US, MBG); San Francisco, May, 1894, C. L. Smith 1339-1374 (G); Remulatero, April, 1922, Purpus 8771 (G, NY, US, MBG). Guerrero: Iguala, Aug. 1905, Rose, Painter 31 / x ■■d c X - > K IC STUDIES IN THE APOCYNACEAE. Ill A Monograph of the Genus Amsonia ROBERT E. WOODSON, JR. ifus J. Lackland of Washington University Historical Discussion When the second edition of Linnaeus 's 'Species Plantarum' appeared in 1762, one of the many additions to the species pre- sented in the first edition (1753) was Tabernaemontana Am- sonia, 1 a plant the exact genus of which Linnaeus himself was not precisely sure, qualifying it to the genus Tabernaemontana with the remark "Affinis Camerariae et Tabernaemontanae." The attitude with which Linnaeus treated his Tabernaemontana Amsonia is essayed by Sir J. E. Smith, 2 and throws much illu- mination upon problems concerning the genus which will receive subsequent treatment in this monograph: "Tabernaemontana . The herbaceous plants, supposed by Linnaeus to belong to this genus, constitute, as we have already said, and as Lin- naeus himself originally thought, a very distinct one, of which we shall now treat by the name of Amsonia. We can give no positive account of the meaning or origin of this word except that its author, according to Miller, [3) was Clayton. Linnaeus in his own copy of Gronovius' Flora Virginica, ed. 1. p. 26, has written Amsonia as a generic name, to what Clayton took for a species of Nerium, and has subjoined also in manuscript the characters of the follicles and seeds. This plant, in the second edition of the Species Plantarum, is the T. Amsonia; and so it remained until Mr. Walter restored it to rank as a genus ; but without throwing any light upon the name." The name Amsonia has indeed been an enigma, and Rafin- l L. Sp. PI. ed. 2, 2:301. 1762. 2 Sm. in Rees, Cycl. 35. 1819. 3 Miller, Gard. Diet. ed. 5, 2*: art. "Tabernaemontana." 1807 Issued December 22, 1928. Ann. Mo. Bot. Gard., Vol. 15, 1928 (379) [Vol. 15 380 ANNALS OF THE MISSOURI BOTANICAL GARDEN esque 1 even went so far as to change the name to "Ansonia," referring to the passage above quoted from Smith, and naively remarking that he had been acquainted with several Ansons, but never an Amson, and so the name must be misspelled. With the fresh stimulus of Rahnesque's contention a special search was made for the origin of the name Amsonia, and for a time it appeared that Rahnesque's intuition had been well guided, for, although all of the published floras and manuals dealing with the genus spoke readily of "Dr. Amson, a colonial physician," or "Charles Amson, a physician of South Caro- lina," no authentic trace of that gentleman could be found, either in published encyclopedias or standard reference works. Historical societies in Virginia and the Carolinas were invoked to no avail. An Amson, any Amson whatever, was not forth- coming. However, Ansons were frequent, including a certain Lord George ison, a royal governor fond of explorative exped of which he had discovered and brought to ci tion a new esculent pea. Rafinesque was about to be vindi- cated, when a letter from Clayton to John Bartram appeared which seems to solve the problem, although not completely. The letter, which was written from Gloucester County, Vir- ginia, Sept. 1, 1760, follows: " Dear Friend: "I have sent you, enclosed, some seed of a new plant, which 1 presume is a stranger in your northern part of the world. Indeed it grows here only in the southern parts of the colony. I have it in my garden, but have quite? forgotten whether I showed it to you, when I had the favor of your company. If I did, I believe I told you it was to be called A?nsonia, after a doctor, here; but I think the name inscribed upon the inclosed more proper, as it answers to the particular form of its seed. "I intend to send you some of the seed of our thorny Sensitive Plant by the first opportunity that offers, after it is ripe; "And remain, dear sir, your sincere friend "And most most humble servant, ^____ "John Clayton. " 1 Raf. New Fl. N. Am. 4: 58. 1838. 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 381 Allowing for orthographical errors, then, Amsonia was defi- nitely named for a certain Dr. Amson, a physician of Gloucester Co., Virginia; but regarding his complete name, or the positive form of spelling of his family name, doubt still remains. Lord Anson, however, can undoubtedly be discarded as a possibility. The "name inscribed upon the inclosed," which Clayton thought to be "more proper," was evidently Tabernaemontana Amsonia. Tabernaemontana Amsonia was immediately conspicuous among the other Tabernaemontanas both in habit and in hab- itat, since it was the only temperate herbaceous member of the genus, and in 1788 attracted the attention of Thomas Walter, 1 who described from it a new genus, naming the type species, in transposition of the Linnaean combination, Amsonia Tabernae- montana. At the same time, Walter 2 also described a new plant from the Carolinas which he assigned to the same new genus, calling it A. ciliata. As a result of his explorations in the southeastern United States, Andre Michaux 3 was able to expand the little genus with the addition, in 1803, of two new species, A. latifolia and A. angustifolia . The latter was a transfer to Amsonia of a species placed in the genus Tabernaemontana by Aiton 4 in 1789. Pursh, 6 in 1814, also published a new species which he named A. salicifolia. Probably the most interesting addition which has ever been made to the species of Amsonia was in 1819, when Roemer and Schultes 8 transferred to that genus a plant which had been assigned to Tabernaemontana by Thunberg 7 in 1784. The spe- cies, A. elliptica, was a native of Japan, and is yet the only known member of the genus not native to North America. Thunberg was slightly hesitant, as was Linnaeus, in committing his plant to the genus Tabernaemontana, but relied upon the precedence of the earlier author, remarking, much as did Lin- i Walt. Fl. Carol. 98. 1788. * Walt. I. c. 1788. » Michx. Fl. Bor. Am. 1: 121. 1803. * Ait. Hort. Kew. 1 : 300. 1789. * Pursh, Fl. Am. Sept. ed. 1, 1: 184. 1814. * Roem. & Schult. Syst. Veg. 4: 432. 1819. 'Thunb. Fl. Jap. 111. 1784. [Vol. 15 382 ANNALS OF THE MISSOURI BOTANICAL GARDEN naeus, "Valde affinis Amsoniae," a significant statement in the light of the later disposition of the species. The same year there appeared in the Rees 'Cyclopaedia' the description by Sir J. E. Smith 1 of a species of Amsonia which he called A. tristis. The plant was reported grown in an Eng- lish garden from seed collected in North America by Lyon, who contributed the plant from which Pursh published his A. salicifolia. Smith gave to his plant the common name "brown- ish-flowered Amsonia," since, as he wrote, "The flowers . . . are of a dingy brown hue, the segments of their limb strongly reflexed, at least in fading." The species has not been reported since its publication, and because of the rather suspicious- sounding description of the flowers, the name has generally been referred to as a synonym of A. Tabernaemontana Walt., or of A . salicifolia Pursh. Rafinesque's 'New Flora of North America' appeared in 1838, with a comparative sketch of the "Ansonias" then recognizable, and added one new species, A. tenuifolia* the originality of which the author took unusual pains to indicate. In the 'Prodromus' of De Candolle the genus Amsonia re- ceived its first collective treatment. Besides recognizing the «pecies which had previously been published with the exception of A. knuifolia Raf., Alphonse De Candolle* described a new plant which he termed A. salicifolia Pursh var. ciliolata, from Alabama and Louisiana. The following year the genus gave evidence of the growing botanical knowledge of the southwestern United States by the publication by Torrey and Fremont 4 of A. tomentosa from "west of the Rocky Mountains." Fourteen years later, as a result of the activities of the Mexican Bound- ary Survey, Torrey 5 published a second species, A. longiflora, a very distinct plant of the region about El Paso, Texas. Some time later the genus came to the attention of Asa Gray during the course of the preparation of the 'Synoptical Flora/ 1 Sm. in Rees, Cycl. 35: end of art. "Tabernaemontana." 1819. * Raf. New Fl. N. Am. 4: 58. 1838. * A. DC. in DC. Prodr. 8:384. 1844. 4 Torr. & Frem. in Fr6m. Rept. 1843-1844, 316. 1845. * Torr. in Rept. Mex. Bound. Surv. 2<: 159. 1859. 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 383 and in 1877 1 he published two new species from the Southwest, A. brevifolia and A. Palmeri. The following year when the 'Synoptical Flora' 2 was issued, there appeared also a new variety of the genus which Gray termed A. angustifolia Michx. var. Texana. Besides being noteworthy for the contribution of a new variety, Gray's treatment of Amsonia in the 'Synoptical Flora' constitutes a scientific and comprehensive treatment of the group in its phylogenetic aspects. In 1894, in accordance with the Rochester Code of Nomen- clature, a double name was made by Britton 3 for the type spe- cies of the genus. This name, Amsonia Amsonia, is still current among some botanists. K. Schumann, 4 in Engler and Prantl's 'Naturlichen Pflanzen- familien,' elaborated upon Gray's treatment in the 'Synoptical Flora' and divided the genus into two sections which he called Euamsonia and Sphinctosiphon. In the first section, three spe- cies were recognized, namely, A. Tabemaemontana Walt., A. ciliata Walt., and A. elliptica (Thunb.) Roem. & Schult., and in the second section, four species, A. Palmeri Gray, A. longi- fiora Torr., A. brevifolia Gray, and A. tomenlosa Torr. & Frem. In the twentieth century numerous additions have been made to the genus Amsonia. In 1900 A. A. Heller 8 elevated Gray's A. angustifolia Michx. var. Texana to specifie rank. In Small's 'Flora of the Southeastern United States' two new species are contained, A. ludoviciana Vail 6 and A. rigida Shuttleworth. 7 Other new specific contributions have been A. latifolia M. E. Jones, 8 1908, A. Eastwoodiana Rydberg, 9 A.arenaria Standley, 10 and A. hirlella Standley, 11 in 1913. Jepson, 12 in 1925, reduced A. lomentosa Torr. to a variety of A. brevifolia Gray. 1 Gray, Proc. Am. Acad. 12: 64. 1877. » Gray, Syn. PI. N. Am. 2 l : 81. 1878. » Britton, Mem. Torr. Bot. Club 5: 262. 1894. • K. Sch. in Engl. & Prantl, Nat. Pflanzenfam. 4*: 143. 1895. 1 Heller, Muhlenbergia 1 : 2. 1900. • Vail, in Small, Fl. Southeast. U. S. 935. 1903. 7 Shuttlew. in Small, I. c. 1903. • Jones, Contr. West. Bot. 12: 50. 1908. • Rydb. Bull. Torr. Bot. Club 40: 465. 1913. 10 Standi. Proc. Biol. Soc. Wash. 26: 117. 1913. " Standi. I. c. 1913. u Jepson, Man. Fl. PI. Cal. 768. 1925. (Vol. l. r > 384 ANNALS OF THE MISSOURI BOTANICAL GARDEN Evidently the first printed illustration of an Amsonia was one presumably of A, Tabernaemontana in Plukenet, t. 115, fig. 3, 1769, where it appears as "Apocynum Virginianum Asclepia- disfolio erectum floribus pallide caeruleis radici crassa." Because of the peculiar distribution of the genus as well as because of its growing need for a taxonomical revision, it was thought appropriate that a rather broad study be made of the genus Amsonia. Such a study was begun at the Gray Her- barium of Harvard University under the oversight of Dr. B. L. Robinson and Prof. M. L. Fernald, and completed at the her- barium of the Missouri Botanical Garden under Dr. J. M. Greenman. To Professors Robinson, Greenman, and Fernald, the author wishes to express his obligations most heartily for their kindly criticism and their ready suggestions. To Mr. T. H. Kearney the author is also indebted for much valuable aid with regard to the difficult species of the southwestern United States. Various herbaria have been visited, also, or specimens have been borrowed, and to the curators of each the author would express his gratitude. Gross Morphology The genus Amsonia is one of the few members of the Ado- cynaceae which are temperate or subtemperate, and contains within its several species only erect perennial semi-woody herbs. Roots. — The root system of the group is characteristically fibrous. The crowns usually become woody with increasing age, and produce numerous clustered stems. Latex tubes occur in abundance, as in all the members of the family. Stems. — The stem system is typically that of an erect peren- varies relatively little. A mature stem is usually di- and vided into several branches. The species inhabiting the arid regions of the southwestern United States and northern Mexico frequently have stems which are branched to a much greater extent, and much lower upon the stem than the more temperate species of the states of the Southeast and Middle West. Leaves. — The leaves of the genus are alternate to subverti- cillate, and vary greatly in size and outline. Through the suc- cession of species, extremes are found in the leaves of A. Tab- ernaemontana, which are broadly ovate-elliptic, usually measur- 19281 WOODSON — STUDIES IN APOCYNACBAE. Ill 385 ing 3-5 cm. long and 1.5-2.5 cm. broad, petiolate and opposite, to the subverticillate leaves of A. salpignantha, which are linear- lanceolate to linear-filiform, measuring 2-5 cm. long and .5-4 mm. broad, and decidedly sessile. The leaves may also be gla- brous to glaucous, as they are in A. salicifolia, or densely to- mentose, as they occur in A. tomentosa. The leaves are always entire, and are never cordate. In only one species, A. Tabernae- montana, are the bases of the leaf-blades other than acute when a petiole exists. Inflorescence. — The inflorescence is a thyreoid or corymbose cyme. The amount and shape of the inflorescence, however, is varied. The largest inflorescence of the genus is found in A. Tahernaemoniana var. Gattingeri, which frequently contains over fifty blossoms, and the smallest in A. Palmeri, which usually has only five or six. The inflorescence may have very incon- spicuous bracteoles, as in the subgenus Euamsonia, or quite conspicuous bracteoles, giving the whole inflorescence a chaffy appearance, as in the subgenera Sphinctosiphon and Ariicularia. The inflorescence may also be surrounded by the foliage, as in A. arenaria, or held high above the foliage by a long, nearly leafless stalk, as in A. ciliata var. tenuifolia. Pedicels may be relatively long, as in A. Tahernaemoniana var. salicifolia, or frequently lacking altogether, as in A. longiflora. Calyx. — The variation in the calyx is marked. In A. Tahernae- moniana var. salicifolia the calyx is 1 mm. long or less in en- tirety, the lobes being minutely triangular-ovate. In A. tomen- tosa the calyx is as shallow as in the former species, but the lobes are fully 3-5 mm. long and are subulate-aristate. The calyx may be glabrous or pubescent, occasionally becoming sparsely hirsute. Corolla. — The corolla is regularly five-lobed. The tube dilates upward, and may be unconstricted, as in the subgenus Euam- sonia, or markedly constricted at the mouth, as in the subgenera Sphinctosiphon and Articularia. Variation in the length of the tube is great, ranging from 6-8 mm. in A. ciliata var. tenuifolia and allied species, to 3-3.5 cm. in A. longiflora and A. salpignantha, which have the most conspicuous flowers of the genus. The color varies from a clear cerulean blue, tinged to tawny-white in the tube, [Vol. 15 386 ANNALS OF THE MISSOURI BOTANICAL GARDEN in most of the eastern species, to white or a faint livid greenish blue in some of the western species. The tube is always villous within, and may be pubescent or glabrous without. The lobes of the cor- olla are spreading, and may be ovate to narrowly lanceolate in out- line. The length of the lobes varies from one-half the length of the tube to an equal length, save in the large-flowered species of the section longiflorae of the subgenus Sphinctosiphon, where the ratio of the length of the tube to that of the lobes may be from about 3:1 in A. longifiora to 5 : 1 in A. salpignantha . Stamens. — The stamens number five, and are adnate to the corolla-tube well above the middle. The anthers are ovate- lanceolate, acute above, obtuse below, unappendaged, and fertile throughout their entire length. The stamens are free from the stigmatic-cap. Pistil. — The two carpels of the gynoecium, which are unilocu- late and contain many two-seriate anatropous ovules, are united by a common filiform style, which is about the length of the corolla-tube, to a position immediately below the stamens, where it is surmounted by a stigmatic-cap bearing the stigma. The stigmatic-cap is constructed in three elements, the lower of which is a reflexed membranaceous appendage, originating from the summit of the stylar shaft, the central, a tangled mass of short papillae, and the upper, the stigma itself, which may be depressed-capitate or truncate, as in the subgenus Euamsonia, or apiculate by two distinct obtuse lobes, as in the subgenera Sphinctosiphon and Articularia. Fruit. — The fruit of Amsonia is a pair of follicles which are cylindrical and acuminate, and may be slender and continuous, as in the subgenera Euamsonia and Sphinctosiphon or torose and definitely articulated into thickish constricted segments, as in the subgenus Articularia. In either case the seeds are one- seriate, cylindrical, and unappendaged, but in Articularia the endosperm is conspicuously thicker and more corky than in the endosperm of the other subgenera. Systematic Position The genus Amsonia is placed in the tribe Plumeroideae of the Apocynaceae because of its free unappendaged stamens. 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 387 The characters of an ovary containing six to many ovules, an eglandular calyx, coriaceous fruit, a hypercraterform corolla, and included stamens moreover place the genus in the subtribe Euplumeroideae. The closest related genus to Amsonia appears to be for various reasons Haplophyton. The two genera are found in common territory from southern California to southwestern Texas. Mor- phologically the greatest dissimilarity lies in the seeds which are appendaged in Haplophyton. The leaves, which serve to aid the differentiation of the two genera in the 'Synoptical Flora,' 1 are not as widely separated as is ordinarily to be sup- posed, since they are not absolutely opposite in Haplophyton and alternate in Amsonia, but are more nearly approximate in the former and frequently subverticillate in the latter. The stamens are nearly alike in both genera, but are somewhat larger in Haplophyton. The character of the stigmatic head in that genus is also much like that of the stigmatic-head in the subgenus Euamsonia of Amsonia, although more elongate, but lacks a membranaceous reflexed appendage. However, a distinct swollen region occurs upon the stylar shaft of Haplo- phyton just below the papillose cap, which might be regarded as a primitive stage in the development of the more elaborate appendage of Amsonia. Rhazya is also a genus closely related to Amsonia, but possesses a disc and a jointed clavuncle among other dissimilarities. Relationship and Distribution op the Subgenera ■ The genus Amsonia , although relatively a small group, is readily separable into three subgenera, which, while interlocking closely, are distinct and well differentiated entities in the whole. The series of subgenera range in geographical and evolutionary succession from east to west and south in North America, upon which continent the bulk of the species occur, only one species being found in eastern Asia. The subgenus Euamsonia is the largest of the divisions in number of species and varieties and the most widely spread, embracing five species and four varieties in the southeastern United States, and one species in Japan. The second largest 1 Gray, Syn. Fl. N. Am. 2>: 81. 1878. (Vou 15 M88 ANNALS OF THE MISSOURI BOTANICAL GARDEN subgenus with regard to number of species and extent of dis- tribution is Sphinctosiphon, which occurs with eight species in the central-southwestern United States and adjacent Mexico, having for its center of distribution southern New Mexico and northern Chihuahua. Ariicularia is the smallest of the sub- genera, and contains four species limited to southern California, southern Nevada, southwestern Utah, and western Arizona; while one species, A. arenaria, is isolated from the general dis- tribution of the subgenus to which it belongs, in extreme south- western New Mexico and adjoining Chihuahua. The situation of Euamsonia in having species of the south- eastern United States and Japan is by this time of more or less frequent knowledge, and no speculations will be devoted to it, since similar instances have been reported. 1 ' 2 - etc . The occur- rence of the three subgenera in the southern United States and northern Mexico is, however, of general interest. A study of the genus in North America suggests forcibly that it is a genus of mesophytic origin which exhibits an increasing adaptation to an arid habitat. Euamsonia is the one subgenus of a mesophytic habit, and since it is represented by the relict species in Japan to which reference has already been made, it is taken as the most primitive. The subgenera Sphincto- siphon and Articularia are plants of distinctly arid habitat, and the morphological differentiation which those groups ex- hibit are interpreted as divergences from the primitive condition represented by Euamsonia. The genus Amsonia is relatively advanced among the Plumer- oideae because of its highly differentiated stigmatic-cap, among other characters, and it is upon the basis of further differentia- tion in that respect that the first subgeneric division is made. In the subgenus Euamsonia the stigma proper is depressed- capitate or truncate, and appears merely as the freer summit of the papillose central region of the stigmatic-cap to which reference was made in detail in the previous section concerning Gross Morphology. The mouth of the corolla-tube, moreover, is relatively open, continuing the dilation of the tube. In the 1 Gray, A. Mem. Am. Acad. N. S. 6: 377-449. 1859. * Fernald, M. L. Quart. Rov. Biol. 1: 227. 1926. 1928) WOODSON — STUDIES IN APOCYNACEAE. Ill 389 subgenus Sphinctosiphon an evolutionary advance is detected in the elevation of the stigma to the position of two distinct apiculate lobes, and the constriction of the mouth of the corolla- tube. Such differentiations are obviously of use to the plant for insect pollination in an arid region. The subgenus Articularia demonstrates a further advance in the articulation of the follicles, which are quite slender and continuous in Euamsonia and Sphinctosiphon, into thickish con- stricted segments in much the same manner as the legumes of certain desert Leguminosae, beside having the apiculate char- acters of the stigma. The seeds of the follicles of Articularia, moreover, are larger and ovoid, and the endosperm is thick- ened, but of a light and corky texture, an evident construction to facilitate easy dissemination in an arid habitat. The seeds of Euamsonia and Sphinctosiphon, on the other hand, are roughly cylindrical with a relatively thin, hard endosperm. The subgenera of Amsonia are remarkable for their inter- rupted distribution, a factor which lends even sharper distinc- tion to the morphological differences which they display, and suggests certain hypotheses for their origin. Euamsonia, with the greatest number of species and varieties, has been found naturally in all of the southeastern United States, with a gener- ally characteristic habitat of moist woods, ravines, or stream- sides, save in its extreme western limits, where A. ciliata var. texana is found on rocky hillsides and prairies. Sphinctosiphon, with the next largest number of species, is confined to southwestern Colorado, southeastern Utah, New Mexico, and adjacent portions of Arizona, Chihuahua, and Texas. Thus Sphinctosiphon and Euamsonia are entirely sepa- rate in range, except in south-central Texas, which contains two species of Euamsonia and a limited colony of A. salpig- naniha of Sphinctosiphon. Besides the anomalous occurrence of A. salpignantha within the southwestern limits of Euamsonia the nearest that the subgenera approach each other is evi- dently in western Texas, Euamsonia being found in the Wichita Mountains of north-central Texas, and Sphinctosiphon in the Guadaloupe Mountains about two hundred miles to the south- west. The species of Sphinctosiphon, although in an arid region, [Vol. 15 390 ANNALS OF THE MISSOURI BOTANICAL GAHDEN partake of the nature of Euamsonia in frequenting the borders of ponds, streams, and branches. Articularia, with the fewest number of species, is confined to southern California, southern Nevada, southwestern Utah, and northwestern Arizona, save for the species A. arenaria, which has a distribution analogous to the anomalous distribution of A. salpignantha of Sphinctosiphon, occurring separate from the other species with which it has its affinities, in extreme south- western New Mexico (Grant County), and adjacent Chihuahua, within the distributional area of Sphinctosiphon. The species of Articularia demonstrate the most extreme endurance for aridity, being found most frequently in the sand of the open desert, whence, it is rather safely supposed, occurs the striking mor- phological adaptations which they exhibit. Thus Sphinctosiphon and Articularia possess rather distinct areas of distribution, save for A. arenaria of Articularia which occurs fully three hundred miles, to present knowledge, from the known range of its kindred species, a perplexing situation. It is also possible that A. Eastwoodiana and A. Jonesii, species of Articularia and Sphinctosiphon respectively, meet in southern Utah and northern Arizona. At any rate, A. Kearneyana, oc- curring in regions midway between the territories of Sphincto- siphon and Articularia, for reasons which will be advanced later, appears in all probability an hybrid between A. Palmeri of the former subgenus and A. brevifolia of the latter. Thus it is seen that the subgenera of Amsonia occupy essentially distinct and isolated ranges, Articularia and Sphinctosiphon, the most nearly related of the groups morphologically and ecologically, being also the most neighborly distributed, and both distinctly removed from Euamsonia, the supposed primitive subgenus, morpho logically and geographically. If we are to believe that by the Cretaceous the modern angio- spermous type of vegetation had become fully established throughout the world, 1 we may assume that the genus Amsonia was by that time in a flourishing condition with a wide distri- bution over the southern half of what is now the United States 1 Grabau, A. W. Textbook of geology 2: 687, 1922. 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 391 Fiz. 1. Land mass of North America in Comanchean time. and adjacent Mexico. 1 The position of the land masses in 1 Stopes, M. C. Ancient plants, p. 85. 1910. Dr. Stopes wrote, in support of such an assumption: "Specimens of Cretaceous plants from various parts of the world seem to indicate that there was a striking uniformity in the flora of that period all over the globe." 392 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN Fig. 2. Land mass of North America in late Comanchean time Comanchean, or lower Cretaceous, time would lend support to the speculation that the genus was allowed at that time practically an uninterrupted range (fig. 1), and because of that reason was very likely of a more or less uniform character. The 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 393 fact that the genus is now found as a relict in Japan is reason enough for assuming a wide range for its species. During the late Comanchean time, however, the continuous range supposed for the genus was broken by the inundations of the Colorado Trough (fig. 2), which occurred over nearly the whole of northeastern Mexico, Texas, and parts of Oklahoma, Colorado, Kansas, and New Mexico. This invading sea could scarcely be regarded as less than a most effective opportunity for generic variation through isolation, especially since the vege- tation in the isolated mass was by that time bearing evidence of an adaptation to aridity. 1 The Cretaceous time, proper, is well known for the extensive inundations which then occurred widely in North America, and the break in the hypothetical distribution of Amsonia was heightened by an increase of the seas of the Colorado Trough, which cut completely through the continent from what is now the coast of the Territory of Mackenzie to the Gulf of Mexico. Troughs of the western coast also caused intrusions during the early periods of the Cenozoic, reaching a climax during Mio- cene time, when large tracts of southern California and adja- cent Lower California and Arizona were separated as islands (fig. 3). By the Pliocene time, North America had largely assumed the shape with which we are now familiar. With isolated land masses corresponding roughly to the lo- calities of probable origin of the subgenera Sphinctosiphon and Arlicularia, a fair degree of credence might be allowed the assumption of their differentiation upon those lands, the first instance of isolation, the intrusion of the Colorado Trough, possibly giving rise to the development of the type of Sphincto- siphon from the primitive condition represented at present by Euamsonia, and the second, the production of islands by the inundations of the west coast troughs during the Miocene, pro- viding an opportunity for the divergence of the type of Arlicu- laria from the group now represented by Sphinctosiphon. It is 1 Schuchert, C. Outlines of historical geology. 1924. "In general we may say that after the early upper Cretaceous time ... the climate the world over was . . . warm temperate in character [p. 612.] With the Miocene, however . . . more or less of desert climates developed in the Cordilleran areas of North America and have prevailed there ever since [p. 626]." 394 [Vol. 15 ANNALS OF THE MISSOURI BOTANICAL GARDEN Fig. 3. Land mass of North America in upper Cretaceous time. believed that the discontinuous areas of the subgenera sui the hypothesis offered for their The fact that the dis tribution of the three subgenera still bears evidence of disrup- tion thousands of years after the Mesozoic and Cenozoic inun- dations appears quite striking, and bears additional evidence of 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 395 the old age 1 of the genus in its inability to reunite its former distribution. In this respect, the genus Amsonia offers an in- teresting parallel to the endemics of the unglaciated regions of boreal America, whose ranges were broken by the Pleistocene glacial phenomena: "These older species in North America have long since passed their period of aggressiveness. 'Left undis- turbed they persist in their old habitats, but they fail to move into new and immediately neighboring territory.' " ? Abbreviations Abbreviations indicating the herbaria where specimens cited in this monograph are deposited are as follows : Baker = C. F. Baker Herbarium of Pomona College. F = Field Museum of Natural History Herbarium. G = Gray Herbarium of Harvard University. MBG = Missouri Botanical Garden Herbarium. NE = New England Botanical Club Herbarium. NY = New York Botanical Garden Herbarium. P = Pomona College Herbarium. ANSP = Academy of Natural Sciences of Philadelphia Her- barium. PBC = Philadelphia Botanical Club Herbarium. TAXONOMY Amsonia Walt. Fl. Carol. 98. 1788; Michx. Fl. Bor. Am. 1: 121. 1803; Pursh, Fl. Am. Sept. ed. 1, 1: 184. 1814; Roem. & Schult. Syst. Veg. 4: 432. 1819; Smith in Rees, Cycl. 35: end of art. "Tabernaemontana." 1819; Elliott, Sketch Bot. S. C. & Ga. 316. 1821; Endl. Gen. PI. 582. 1838; A. DC. in DC. Prodr. 8: 384. 1844; Pfeiffer, Norn. Bot. I 1 : 156. 1873; Benth. & Hook. Gen. PI. 2: 703. 1876; Gray, Syn. Fl. N. Am. 2 l : 81. 1878; Durand, Index Gen. Phan. 262. 1888; Baill. Hist. PI. 10: 180. 1891; Coulter, Contr. U. S. Nat. Herb. 2: 262. 1892; Coville, Contr. U. S. Nat. Herb. 4: 142. 1893; 1 The presence of Amsonia in the Cretaceous period would give to the genus an age, based upon the most capable of present calculations (Schuchert, C. I. c. 485. 1924), of at least 45,000,000 years. ' Fernald, M. L. Persistence of plants in unglaciated areas of boreal America. Mem. Am. Acad. 15: 244. 1925; Antiquity of vascular plants. Quart. Rev. Biol. 1:227. 1926. 396 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 K. Schumann in Engl. & Prantl, Nat. Pflanzenfam. 4 2 : 143. 1895; Mohr, Contr. U. S. Nat. Herb. 6: 674. 1901; Small, Fl. Southeast. U. S. 934. 1903; Dalla Torre & Harms, Gen. Siph. 406. 1904; Harper, Ann. N. Y. Acad. Sci. 17 1 : 175. 1906; Robinson & Fernald in Gray, New Man. ed. 7, 661. 1908; Nelson in Coulter & Nelson, New Man. Rocky Mt. Bot. 385. 1909; Matsumura, Index PL Jap. 2: 505. 1912; Wooton & Standley, Contr. U. S. Nat. Herb. 19: 504. 1915; Rydb. Fl. Rocky Mts. 668. 1917; Davidson & Moxley, Fl. South. Cal. 278. 1923; Tidestrom, Contr. U. S. Nat. Herb. 25: 418. 1925; Jepson, Man. FL PL Cal. 768. 1925. Ansonia Raf. New Fl. N. Am. 4: 58. 1838. Lactescent herbaceous caulescent perennials, glabrous or pu- bescent. Leaves alternate or subverticillate, sessile or petio- late, membranaceous or somewhat thickened and fleshy, entire. Inflorescence a terminal thyrsoid or corymbose cyme. Calyx five-parted, the lobes acuminate or subulate. Corolla salver- form, villous within, tube cylindrical, dilating, open or con- stricted; lobes ovate to lanceolate, spreading or nearly erect. Stamens five, adnate to the corolla-tube above the middle, included; anthers ovate to ovate-lanceolate, obtuse, unappen- daged. Disk wanting. Carpels two, united by the filiform style surmounted by a truncate stigmatic-cap. Stigma de- pressed-capitate, or apiculate by two distinct lobes, surrounded by a spherical papillose mass, and appendaged by a reflexed membrane. Ovules in each carpel many, two-seriate. Follicles two, cylindrical, continuous or articulated. Seeds many, one- seriate, cylindrical, unappendaged. Embryo straight. Type species: A. Tabernaemontana Walt. FL Carol. 98. 1788. Synopsis of the Subgenera and Sections KEY TO THE SUBGENERA A. Stigma depressed-capitate or truncate Subgenus I. Euamsonia B. Stigma apiculate by two distinct lobes. a. Follicles continuous, not articulated Subgenus I [. Sphinctobiphon b. Follicles torose, articulated into thickish constricted segments Subgenus III. Articularia Subgenus I. Euamsonia (K. Schumann) Woodson Subgenus I. Euamsonia (K. Schumann) Woodson, n. comb. ^Euamsonia K. Schumann in Engl. & Prantl, Nat. Pflanzen- ■ 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 397 fam. 4 2 : 143. 1895; Dalla Torre & Harms, Gen. Siph. 406. 1904. Bracteoles inconspicuous; orifice of the corolla-tube not con- stricted in anthesis; stigma depressed-capitate or truncate; fol- licles slender and continuous, not articulate, fibrous, not horny in texture; seeds irregularly oblong in outline, truncate at either end, variously pitted and wrinkled; plants of the southeastern United States and Japan. Spp. 1-5. Key to the Species and Varieties a. Corolla glabrous without. b. Leaf -blades elliptic, distinctly petiolate throughout 1. A. rigida bb. Leaf-blades oblong-lanceolate to linear, sessile or subsessile above. c. Corolla-tube 6-8 mm. long. d. Stem-leaves 4-15 times as long as broad; inflorescence barely held above the foliage £. A. ciliata dd. Stem-leaves 15-30 times as long as broad; inflorescence held high above the foliage 2a. A. ciliata var. tenuifolium cc. Corolla-tube 9-12 mm. long. d. Corolla-lobes about half as long as the tube; pedicels 3-5 mm. long; species of North America. .2b. A. ciliata var. texana dd. Corolla-lobes about equalling the tube; pedicels 5-10 mm. long; species of Asia 8. A. elliptica aa. Corolla pubescent without. b. Follicles glabrous. c. Leaf-blades ovate to oblong-lanceolate, the bases of the lower obtuse to broadly acute 4- -4- Tabernaemontana cc. Leaf-blades lanceolate to linear-lanceolate, the bases of the lower acute to acuminate. d. Inflorescence loose, few-flowered; foliage glabrous, glaucous beneath 4a. A. Tabernaemontana var. salicifolia dd. Inflorescence dense, many-flowered; foliage pubescent, gla- brate in age 4b. A. Tabernaemontana var. Gattingeri bb. Follicles pubescent, at least upon the upper portion 5. A. ludoviciana 1. Amsonia rigida Shuttlew. in Small, Fl. Southeast. U. S. 935. 1903; Harper, Ann. N. Y. Acad. Sci. 17 1 : 175. 1906. PI. 51, figs. 4-6. Herbaceous perennial from a thickened somewhat woody root; stems 8-15 dm. tall, regularly branched above, glabrous; leaves alternate, numerous, the blades almost exactly elliptic, isophyl- lous, i. e., the lower and the upper leaves of nearly like outline, green above, glaucous or glaucescent beneath, 2.5-6 cm. long, .5-1.5 cm. broad, distinctly petiolate throughout; flowers rela- [Voi,. 15 398 ANNALS OF THE MISSOURI BOTANICAL GARDEN tively numerous in fairly loose cymes; pedicels 5 mm. long or slightly less; calyx 1-1.5 mm. long, glabrous, the lobes triangu- lar-ovate; corolla salverform, the tube 6-8 mm. long, gradually dilating upwards, glabrous without, the lobes lanceolate, 7-10 mm. long, widely spreading; stigmatic-cap about as tall as broad, stigma depressed-capitate; follicles slender, continuous, gradu- ally attenuate, 7-11 cm. long, sessile, glabrous, 7-10-seeded; seeds 5-11 mm. long, oblong in outline, truncate at either end, variously wrinkled and pitted, dark brown. Distribution: swampy or moist pine forests, northern Florida and southern Georgia. Specimens examined : Georgia: Alapaha, swampy pine woods, June 25, 1901, Cur- tiss 6820 (G, MBG, NY, US); Sumter Co., moist pine barrens, Aug. 21, 1900, Harper 448 (G, MBG, NY, US, F) ; same locality, Sept. 6, 1900, Harper 606 (NY, US); same locality, Aug. 21, 1900, Harper 440 (NY, US). Florida: Chattahoochee, May, 1882, Curtiss (G); data lack- ing, Chapman (G, ANSP, MBG); St. Marks, June, 1843, Rugel (MBG); Chattahoochee, 1891, Chapman (MBG). 2. Amsonia ciliata Walt. Fl. Carol. 98. 1788; A. DC. in DC. Prodr. 8: 385. 1844; Wood, Classbook Bot. 589. 1860 Chapm. Fl. South. U. S. 343. 1897; Mohr, Contr. U. S. Nat Herb. 6: 674. 1901; Small, Fl. Southeast. U. S. 935. 1903 Harper, Ann. N. Y. Acad. Sci. 17 1 : 175. 1906. PL 51, figs. 7-8 Tabernaemontana angusiifolia Ait. Hort. Kew. 1: 300. 1789 Willd. Sp. PI. I 2 : 1247. 1798. Amsonia angustifolia (Ait.) Michx. Fl. Bor. Am. 1: 121. 1803 Pursh, Fl. Am. Sept. ed. 1, 1: 184. 1814; Roem. & Schult. Syst Veg. 4: 432. 1819; Ell. Sketch Bot. S. C. & Ga. 317. 1821 Darby, Bot. South. States, 434. 1860; Gray, Syn. Fl. N. Am. 2 l :81. 1878. Ansonia ciliata (Walt.) Raf. New Fl. N. Am. 4: 58. 1838. Ansonia angusiifolia (Ait.) Raf. I. c. 1838. Herbaceous perennial from a thickened woody root; stems 7-15 dm. tall, clustered from the base, erect or slightly ascend- ing, sparsely branched above, the branches ascending, pubes- 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 399 cent, glabrous or glabrate in age; leaves numerous, crowded, subverticillate above, slightly heterophyllous, i. e., the lower leaves broader and of a slightly different outline than the upper, linear-lanceolate, or the lower oblong-lanceolate, pubescent, or glabrate in age; inflorescence dense, barely held above the foli- age; pedicels 3-5 mm. long, sparsely pubescent; calyx 1-1.5 mm. long, glabrous, or with a few scattered hairs, the lobes triangular-ovate; corolla salverform, the tube 6-8 mm. long, glabrous without, the lobes 7-8 mm. long, oblong-lanceolate, erect or spreading; stigmatic-cap slightly broader than tall, stigma depressed-capitate or truncate; follicles slender, con- tinuous, 9-11 cm. long, gradually attenuate, sessile, glabrous, 7-11-seeded; seeds 5-11 mm. long, oblong in outline, truncate at either end, variously pitted or wrinkled, dark brown. Distribution: pine forests, occasionally entering fields; North Carolina, South Carolina, southern Georgia, northern Florida, southern Alabama, and northeastern Texas. Specimens examined: North Carolina: data lacking, Curtis (G). South Carolina: Aiken, April, 1882, Velden (MBG); Aiken, sand hills near Granite ville, May 7, 1899, Eggerl (MBG) ; Aiken, May, 1869, Canby 68 (MBG, G, NY) ; Columbia, woods, May 9, 1899, Sargent (G); data lacking, Ravenel (G, NY, US); Colum- bia, dry sandy pine woods, May, 1890, Taylor (F). Georgia: Richmond Co., slopes of sand hills about 8 miles west of Augusta, June 10, 1902, Harper 1319 (US, NY, F); Augusta, date lacking, Olney & Metcalfe 76 (G); data lacking, Wilkins (G). Florida : Tallahassee, date lacking, Berg (NY) ; Aspalaga, dry pine woods, April, year lacking, Curtiss 2269 (G, MBG, ANSP, US, F); River Junction, fields and open woods, April 22 and May 16, 1898, Curtiss 6376 (G, NY, US, MBG) ; data lacking, Chapman (G); Aspalaga, May, 1898, Chapman (MBG); Coffee Co., rocky open ground, flood plains of Pea River, May 15, 1925, E. J. Palmer 27233 (MBG); Chehaw, June 24, 1915, Drushel 4572 (MBG). Alabama: data lacking, Durand (ANSP). Texas: San Marcos, June 6, 1897, Stanfield (NY); Mid- [Vol. 15 400 ANNALS OF THE MISSOURI BOTANICAL lothian, April 30, 1895, Plank (NY); Turtle Creek, Kerr Co., date lacking, Bray 239 (US); Orange, April 17, 1899, Bray 60 (US). Although recognizing that A. ciliata Walt, antedates A. an- gustifolia (Ait.) Michx., Gray placed Walter's species in sy- nonymy with the latter species, remarking that ciliata was an inappropriate name. The specimens with Gray's labels in the Gray Herbarium truly are glabrate or glabrous, being overly matured specimens, hence Dr. Gray's impression. In any event, Walter's name can scarcely be discarded. 2 a. Var. tenuifolia (Raf.) Woodson, n. comb. Ansonia tenuifolia Raf. New Fl. N. Am. 4: 58. 1838. Amsonia salicifolia Pursh var. ciliolata A. DC. in DC. Prodr. 8:384. 1844. Amsonia ciliata Walt. var. filifolia Wood, Classbook Bot. 589. 1860. Amsonia tenuifolia (Raf.) Harper, Ann. N. Y. Acad. Sci. 17 1 : 175. 1906. Herbaceous perennial from a fibrous root; stems 3-10 dm. tall, single or sparingly clustered from the base, erect or slightly ascending, sparingly branched above, the branches ascending, pubescent or glabrate in age; leaves numerous, crowded, sub- verticillate, scarcely heterophyllous, i. e., the lower leaves barely broader and of about the same outline as the upper, linear- lanceolate to filiform, pubescent or glabrate; inflorescence dense, held high above the foliage by a slender, usually leafless stalk; pedicels 3-5 mm. long, barely strigose or glabrous; calyx 1-2 mm. long, glabrous, or with a few short hairs, the lobes triangu- lar-attenuate; corolla salverform, the tube 6-8 mm. long, gla- brous or slightly canescent without, the lobes 4-6 mm. long, ovate to oblong-lanceolate, erect or spreading; stigmatic-cap about as tall as broad, stigma depressed-capitate or truncate; follicles slender, continuous, 8-14 cm. long; seeds 7-12 mm. long, oblong in outline, truncate at either end, variously pitted or wrinkled, brown. Distribution: sand-hills and barrens, also rocky margins of streams; North Carolina, South Carolina, Georgia, Florida, Ala- 1928) WOODSON — STUDIES IN APOCYNACEAE. Ill 401 bama, southern Arkansas, Missouri, Texas, and central Mexico. Specimens examined : United States: North Carolina: data lacking, Curtis (G); White Hall, May 13, 1896, Biltmore 1400 (US). South Carolina: Aiken, May 21, 1899, Eggeri (MBG). Georgia: Altamaha, sand-hills, date lacking, Chapman (G); Augusta, sand-hills, June 10, 1902, Harper 1319 (G, MBG, NY) ; Bainbridge, low woods bordering Flint River, July 13, 1899, Curliss 6476 (G, MBG, NY, US) ; Bulloch Co., sand-hills along Big Lott's Creek, June 17, 1901, Harper 915 (G, MBG, NY, US, F) ; Camilla, Mitchell Co., dry sand barrens, Aug. 7, Harper 1166 (G, NY, US); Dublin, Laurens Co., sand-hills of Oconee River, April 20, 1904, Harper 2188 (G, MBG, NY, US, F); Jasper City, 1846-48, Porter (G); Dooly Co., dry pine barrens near Gum Creek, Sept. 3, 1900, Harper 577 (US); Burke Co., Aug. 15, 1897, Hopkins 83 (NY); Thomson, McDuffie Co., sand-hills, Sept. 9, year lacking Bartlett 1493 (P); Vidalia, April, 1914, Huger (MBG); Macon, date lacking, Green (ANSP). Florida: Bellair, Sept. 3, 1895, Nash 2546 (G, MBG, US, F); Clarcona, Orange Co., date lacking, Meislahn 210 (US); Gotha, March 28, 1919, Nehrling 12 (US); pine woods west of Jacksonville, April, 1848, Rugel 21 (US, MBG, F, NY); data lacking, 1873, Fell (ANSP); Cocoanut Grove, 1899, Rodman (G); Sumter Co., grassy pine-barrens, March 11, 1883, Donnell- Smith (G); "East Florida," date lacking, Buckley (G); "Middle Florida," date lacking, Eaton (G); data lacking, Buckley (G); Alachua Co., June-July, 1898, Hitchcock (MBG); Lake Brant- ley, Aug. 1, 1895, Williamson (ANSP). Alabama: data lacking, Buckley (MBG); Bon Secour (near Mobile), June 29, 1893, Mohr (US). Missouri: Ozark Co., rocky open ground, bald knobs, near Tecumseh, Oct. 9, 1927, E. J. Palmer 38081 (MBG) ; summit of bald knob across river from Tecumseh, Ozark Co., Nov. 11, 1928, Anderson & Woodson 4000 (MBG). Arkansas: Logan Co., rocky margins of small streams, Oct. 18, 1923, E. J. Palmer 24203 (G); Hot Springs, Aug. 5, 1879, Letterman (MBG); Arkadelphia, May 10, 1884, Letterman (MBG). [Vol. 15 402 ANNALS OF THE MISSOURI BOTANICAL GARDEN Texas: data lacking, Wright (G); Medina Lake, Bandera Co., limestone ledges, creek banks, June 14, 1917, E. J. Palmer (MBG); Johnson Co., rocky prairies, April, 1882, Re- verchon 84 (MBG). Mexico: Michoacan: Morelia, June, 1901, Arsene (F). 2b. Var. texana (Gray) Coulter, Contr. U. S. Nat. Herb. 2: 262. 1892. Amsonia anguslifolia Michx. var. Texana Gray, Syn. Fl. N. Am. 2 1 :81. 1878. Amsonia texana (Gray) Heller, Muhlenbergia 1: 2. 1900; Small, Fl. Southeast. U. S. 935. 1903; Rydb. Fl. Colo. 269. 1906; Nelson in Coulter & Nelson, New Man. Rocky Mt. Bot. 385. 1909; Clem. & Clem. Rocky Mt. Fl. 100. 1914. Herbaceous perennial from a slightly woody root; stems 2-5 dm. tall, usually clustered from the base, erect or slightly as- cending or spreading, occasionally pubescent when young, mostly glabrous ; leaves alternate, numerous, quite heterophyllous, i. e., the lower leaves broader and of a different outline than the upper, ovate to oblong-lanceolate below, lanceolate to linear- lanceolate above, occasionally with short scattered hairs; in- florescence compact, barely held above the foliage; pedicels 3-5 mm. long; calyx 1.5-2.5 mm. long, glabrous or glabrate, the lobes triangular-lanceolate to subulate; corolla salverform, the tube 9-11 mm. long, glabrous without, the lobes 4-6 mm. long, ovate to ovate-lanceolate, spreading; stigmatic-cap broader than tall, stigma truncate; follicles slender, continuous, 6-10 cm. long, rather abruptly acuminate, sessile, glabrous. 5-15-seeded; seeds 5-11 mm. long, oblong in outline, truncate at the ends, variously pitted or wrinkled, brown. Distribution: dry and rocky hillsides and prairies; Oklahoma, and Texas. Specimens examined: Oklahoma: Crusher Spur, Murray Co., rocky mountain-side, April 12, 1913, Stevens 29 (G, MBG, US); Fort Sill, May 20, 1892, Sydone (NY); Tishomingo, on hillsides, common, April 8, 1910, Houghton 3606 (G); vicinity of Fort Sill, April 12, 1916, 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 403 Clemens 11727 (MBG); Cache, Comanche Co., dry hillsides, decomposed granite, July 19, 1917, E. J. Palmer 12597 (MBG); Davis, Arbuckle Mts., April 1, 1916, Emig 399 (MBG). Texas: Comanche Springs, March, 1849, Lindheimer (G, MBG); Dallas, rocky prairies, April, 1875, Reverchon (G, NY, MBG); Dallas, dry uplands, March-June, year lacking, Rever- chon (G, MBG); "Upper Colorado," rocky places, 1847, Lind- heimer 660 (G type, MBG, US, F); Fort Worth, rocky hill- sides, May 7, 1911, Ruth 241 (G, NY, ANSP, F); "Witicha Mtns.," July, 1852, Torrey (G, NY) ; data lacking, Lindheimer (ANSP, MBG); Dallas, dry soil, April-June, 1877, Reverchon 598 (US, MBG) ; Dallas, common in woods, May 7, 1900, Bush 646 (US, NY, MBG); Dallas, rocky prairies, June 30, 1877, Hall 515 (US, MBG, NY); Forks, May 27, year lacking, Rever- chon (MBG); Boerne, Kendall Co., low rocky creek banks, April 6, 1917, E. J. Palmer 11471 (MBG); Dallas, rocky hills, West Dallas, June 22, 1899, Eggert (MBG) ; Hood Co., prairies, May 4, 1900, Eggert (MBG); Dallas, cement works, April 12, 1902, Reverchon (MBG); data lacking, Lindheimer 4 (MBG); Gillespie Co., date lacking, Jermy 145 (MBG); Dallas, open limestone hills, May 4, 1918, E. J. Palmer 13496 (MBG); Lacey's Ranch, Kerr Co., moist rocky creek banks, June 11, 1917, E. J. Palmer 12233 (MBG); Bull Creek, near Austin, April 11, 1914, Young (MBG); Boerne, Kendall Co., moist rocky creek banks, April 20, 1917, E. J. Palmer 11616 (MBG); Dallas, high prairies, April 12, 1902, Reverchon 3122 (MBG). 3. Amsonia elliptica (Thunb.) Roem. & Schult. Syst. Veg. 4: 432. 1819; A. DC. in DC. Prodr. 8: 384. 1844; Franch. & Savatier, Enum. PI. Jap. 1: 315. 1874; K. Sch. in Engl. & Prantl, Nat. Pflanzenfam. 4 2 : 143. 1895; Matsumura, Index PI. Jap. 2: 505. 1912. PI. 51, figs. 9-10. Tabernaemontana elliptica Thunb. Fl. Jap. 111. 1784. Ansonia elliptica (Thunb.) Raf. New Fl. N. Am. 4: 58. 1838. "Amsonia elliptica Sieb. & Zucc.'' in Gray, Mem. Am. Acad. 6:403. 1857. Herbaceous perennial from a slightly thickened root; stems 4-7 dm. tall, single or clustered from the base, erect or slightly 404 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 ascending, glabrous, or very slightly pubescent when young, branched above, the branches ascending or somewhat spread- ing; leaves alternate, relatively distant, the blades relatively narrow, lanceolate to linear-lanceolate, the lower 5-10 times as long as broad, both the bases and the apices narrowly acute to acuminate, glabrous above, glaucescent beneath, becoming green in age; inflorescence loose, relatively few-flowered, pedicels 5-10 mm. long; calyx 1-2 mm. long, the lobes triangular-lanceo- late, glabrous; corolla salverform, the tube relatively broad, 10-12 mm. long, glabrous without; the lobes of about equal length, oblong-lanceolate, spreading; stigmatic-cap about as broad as tall, stigma depressed-capitate or truncate; follicles relatively stout, continuous, or very slightly torose, 4-6 cm. long, sessile, glabrous, 5-10-seeded; seeds 5-10 mm. long, oblong in outline, truncate at either end, variously pitted or wrinkled, brown. *■ Distribution: northern Japanese Archipelago. Specimens examined: Japan: Hakodate, 1861, Maximomcz (G); Jesso, near Hako- date, 1861, Albrecht (G); Todahara, Musashi, May 24, 1891, Watanabe (G) ; Tokio, May 7, 1879, Matsumura (US) ; Musaski, Toda, May 27, 1911, collector lacking (US). f 4. Amsonia Tabernaemontana Walt. Fl. Carol. 98. 1788; Pers. Syn. 1: 269. 1801; A. DC. in DC. Prodr. 8: 384. 1844; Rept. Torr. Bot. Mex. Bound. Surv. 159. 1859; Gray, Syn. Fl. N. Am. 2 l : 81. 1878; Wood, Classbook Bot. 589. 1860; Gat- tinger, Tenn. Fl. 63. 1887; Coulter, Contr. U. S. Nat. Herb. 2: 262. 1892; K. Schumann in Engl. & Prantl, Nat. Pflanzen- fam. 4 2 : 143. 1895; Chapm. Fl. South. U. S. 343. 1897; Rob- inson & Fernald in Gray, New Man. ed. 7, 661. 1908. PI. 51, figs. 11-13. Anonymus suffrutex Gronov. Fl. Virg. ed. 2, 35. 1762. Tabernaemontana Amsonia L. Sp. PI. ed. 2, 2: 301. 1762; Willd. Sp. PI. I 2 : 1246. 1798. Tabernaemontana humilis Salisb. Prodr. 148. 1796. Amsonia latifolia Michx. Fl. Bor. Am. 1: 121. 1803; Pursh, Fl. Am. Sept. ed. 1, 1: 184. 1814; Roem. & Schult. Syst. Veg. 4: 1928| WOODSON — STUDIES IN APOCYNACEAE. Ill 405 432. 1819; Elliott, Sketch Bot. S. C. & Ga. 316. 1821; Darby, Bot. South. States, 434. 1860. Amsonia tristis Sm. in Rees, Cycl. 35: end of art. "Tabernae- montana." 1819; A. DC. in DC. Prodr. 8: 384. 1844. Ansonia latifolia (Michx.) Raf. New Fl. N. Am. 4: 58. 1838. Amsonia Amsonia (L.) Britton, Mem. Torr. Bot. Club 5: 262. 1894; Britton & Brown, 111. Fl. 3: 1. 1898; S. Coulter, Rept. Dept. Geol. Ind. 24: 880. 1899; Hitchcock, Fl. Kans. 13. 1899; Gattinger, Fl. Tenn. 137. 1901; Mohr, Contr. U. S. Nat. Herb. 6: 674. 1901; Small, Fl. Southeast. U. S. 935. 1903; Lowe, Miss. State Geol. Surv. Bull. 17: 227. 1921. Herbaceous perennial from a thickened, slightly woody root; stems 3-10 dm. tall, usually clustered from the base, erect or slightly ascending, branched above, the branches ascending or spreading, occasionally somewhat pubescent when young; leaves alternate, relatively distant, ovate to oblong-elliptic, the bases of the lower obtuse to broadly acute, occasionally sparsely pubescent upon the lower surface when very young; inflores- cence relatively small and dense, barely held above the foliage, pedicels 3-5 mm. long; calyx 1-1.5 mm. long, glabrous, the lobes triangular-ovate; corolla salverform, the tube 6-8 mm. long, pubescent without, the lobes 4-6 mm. long, oblong to oblong-lanceolate, spreading; stigmatic-cap about as tall as broad, stigma depressed-capitate; follicles continuous, 8-10 cm. long, rather abruptly acuminate, sessile, glabrous, 5-15-seeded; seeds 5-11 mm. long, oblong in outline, truncate at either end, variously pitted and wrinkled, dark brown. Distribution: moist woods and waste-lands, river-banks, etc.; South Carolina, Tennessee, Illinois, eastern Missouri, eastern Oklahoma, eastern Kansas, southeastern Arkansas, escaped from cultivation in Massachusetts, New Jersey, Pennsylvania, and Delaware. Specimens examined : Massachusetts: Boston, Back Bay waste-lands, Aug. 12, 1903, Williams (G); Hampden, June, 1911, Knowllon (NE). New Jersey: South New England Road, introduced in field, Cold Spring, Cape May Co., July 7, 1918, Brown (PBC). 406 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol 15. Delaware: Wilmington, waste places, June 3-July 18, 1896, Commons (PBC). Pennsylvania: Oakdale, near Philadelphia, June, 1863, Mar- tindale (PBC) ; Philadelphia, Broad Street & Germantown R. R., 1865, Martindale 4864 (MBG) ; Gradyville, Delaware Co., June 9, 1898, Painter (PBC) ; near Philadelphia, May, 1889, Leeds (PBC) ; Gradyville, June 3, 1904, Vail 546 (US). South Carolina : Greenville Co., ravines near Caesar's Head, Aug. 5, 1881, /. D. Smith (G). Tennessee: Knoxville, thicket on Tennessee River bank, April and July, 1890, Ruth 174 (G); Knoxville, April, 1894, Ruth 466 (P, US); Knoxville, June, 1898, Ruth 480 (MBG). Illinois: Chandlersville, Aug. 19, 1886, Seymour 1584 (G, P). Missouri: St. Louis, July 2, 1895, Glatfelter (G, MBG); St. Louis, date lacking, Engelmann (G, MBG); Eagle Rock, un- common in barrens, June 22, Bush 11 (MBG, US); uncommon in rich woods, 4 miles e. of Carthage, May 27, 1906, E. J. Palmer 921 (MBG); Newton Co., cherty barrens, July 15, 1906, E. J. Palmer 12 (MBG); Carthage, rich woods, May 27, 1906, E. J. Palmer 818 (MBG); Noel, low ground, May 10, 1915, Bush 7513 (MBG); Noel, McDonald Co., thickets, hillsides, Sept. 12, 1913, E. J. Palmer 4305 (MBG); Swan, common in woods, Oct. 4, 1899, Bush 753 (MBG). Arkansas: Fort Huron, date lacking, Edward (G); Fayette- ville, May, year lacking, Harvey 38 (ANSP) ; Fulton, low ground, April 17, 1905, Bush 2378 (MBG); Fayetteville, May 10, 1919, Wells (US). Louisiana: Hammond, April 10, 1889, Gallup 4 (US). Oklahoma: Leflore Co., Page, on bank of mountain creek near Rich Mountain, Sept. 8, 1913, Stevens 2670 (G, US) ; Page, Leflore Co., on rocky mountain-side, April 25, 1915, Blakely 3425 (G); Poteau, Leflore Co., July 13, 1915, E. J. Palmer 8286 (MBG). Kansas: Cherokee Co., rocky woods, May 7, 1897, Hitchcock 76a (MBG). 4a. Var. sallcifolia (Pursh) Woodson, n. comb. Amsonia salicifolia Pursh, Fl. Am. Sept. ed. 1, 1: 184. 1814 t 1928] WOODSON— STUDIES IN APOCYNACEAE. Ill 407 Roem. & Schult. Syst. Veg. 4: 432. 1819; Elliott, Sketch Bot. S. C. & Ga. 316. 1821; A. DC. in DC. Prodr. 8: 384. 1844; Darby, Bot. South. States, 434. 1860; Wood, Classbook Bot. 589. 1860; Small, Fl. Southeast. U. S. 935. 1903; Britton, Man. Fl. 737. 1907. Ansonia salicifolia (Pursh) Raf. New Fl. N. Am. 4: 58. 1838. Herbaceous perennial from a slightly thickened root; stems 3-5 dm. tall, usually clustered from the base, erect or slightly ascending, glabrous or very slightly pubescent when young, branched above, the branches ascending or somewhat spread- ing; leaves alternate, the blades relatively narrow, lanceolate to linear-lanceolate, the lower 5-10 times as long as broad, both the base and the apex narrowly acute to acuminate, gla- brous above, glaucous or glaucescent beneath, becoming green in age; inflorescence loose, relatively few-flowered, pedicels 3-7 mm. long; calyx about 1 mm. long, the lobes minutely triangu- lar, glabrous; corolla salverform, the tube relatively narrow, 6-10 mm. long, scatteringly pubescent without, the lobes 5-7 mm. long, lanceolate, spreading; stigmatic-cap somewhat broader than tall, stigma depressed-capitate or truncate; follicles rela- tively slender, continuous or very slightly torose, 8-10-seeded; seeds 5-10 mm. long, oblong in outline, truncate at either end, variously pitted or wrinkled, brown. Distribution: river-banks and moist thickets generally; Vir- , North Carolina, South Carolina, Georgia, Alabama Kentucky, Tennessee, Indiana, Illinois, Missouri, A and mens examined Virginia: Petersburg, date lacking, Tuomey (ANSP). North Carolina: Biltmore, river-banks, May 11, 1897, Bilt- more 81b (G, MBG, US, NY); same locality, May 2, 1896, Biltmore 81 (MBG) ; Weldon, April, 1897, Williamson (ANSP) ; same locality, April 19, 1908, Williamson (ANSP); Hot Springs, May 5, 1884, Smith (ANSP); Warm Springs, May 6, 1887, Smith (ANSP); Columbus, 1897, Townsend (US); Statesville, May, 1878, Hyams (US); Granville Co., May, 1873, Faxon (G) ; Mt. Tryon, Polk Co., moist rich soil on a level spot along a mountain rill, May, 1918, Millspaugh 4030 (F); Weldon, April 19, 1908, Bartram (G); data lacking, Curtis (G). [Vol. 15 408 ANNALS OF THE MISSOURI BOTANICAL GARDEN South Carolina: Oconee Co., Clemson College, low woods, April 16, 1906, House 1851 (US, NY). Georgia: Thompson's Mills and vicinity, Gwinnett Co., April 8, 1908, Allard 197 (US); Macon, date lacking, Green (ANSP); Stone Mountain, Yellow River, May 3, 1899, Sargent 69 (G). Alabama: Albertsville, April 22, 1899, Hosdy (US); Chicka- saw, rich places in the barrens, April, 1919, Graves 566 (US); Tuskaloosa, April, 1892, Ward (US); Auburn, Lee Co., April 22, 1900, Earle (G). Mississippi: Pass Christian, May 16, 1924, Cooper (MBG); Rolling Fork, April, 1895, Boyce (US). Louisiana: Feliciana [East or West?] Parish, Carpenter (G); east of Baton Rouge, April 20, 1874, Joor (F). Indiana: banks of the Wabash River, June, 1868, Allen (F). Kentucky: Bowling Green, 1903, Price (MBG); barrens of Kentucky River near Hopkinsville, date lacking, Buckley (MBG). Tennessee: Franklin Co., May 5, 1898, Eggert (MBG); Nash- ville, 1880, Hubbard 2268 (G). Illinois : exact locality lacking, 1845, Mead (G) ; Grantsburg, April 28, 1900, collector lacking (P); Conologue, April, 1924, Woodson (MBG). Missouri: bank of Meramec River near Windsor Springs, April 19, 1891, Douglass (US); Cave Spring, 1887, Blankinship (US); Cave Spring, Greene Co., June 18, 1905, Standley (US); Tyson, St. Louis Co., May 19, 1918, Drushel 8757 (MBG); banks of the Meramec River, Minke, St. Louis Co., May 17, 1919, Greenman 39U (MBG); Chadwick, May 15, 1907, Bush 4461 (MBG). Arkansas: Baker Springs, Howard Co., April 10, 1909, Kel- logg (MBG); data lacking, Pitcher (ANSP). Texas: Beaumont, low wet woods, March 16, 1918, E. J. Palmer 18090 (MBG). 4b. Var. Gattingeri Woodson, n. var. 1 Herbaceous perennial from a thickened slightly woody root; 1 Var. Gattingeri var. nov., plus-minusve pilosa varietatem genu i nam simulans differt foliis longioribus basi acutis; corollae tubo lanoso. — Tennessee, Nashville, June, year lacking, A. Gattinger (Gray Herb. type). 1028] WOODSON— STUDIES IN APOCYNACEAE. Ill 409 stems 3-10 dm. tall, pubescent, becoming glabrous, somewhat clustered at the base, erect or ascending, branched above, the branches ascending or spreading; leaves relatively distant, alter- nate, the blades lanceolate to linear-lanceolate above, the lower 5-10 times as long as broad, both base and apex narrowly acute to acuminate, green, pubescent, frequently densely so, becoming glabrous in age; inflorescence compact, many-flowered, pedicels 2^4 mm. long; calyx 2-4 mm. long, the lobes narrowly triangu- lar, glabrous, or with a few scattered hairs; corolla salverform, the tube 7-10 mm. long, densely pubescent or villous without, especially in the sinuses of the lobes, the lobes 5-8 mm. long, lanceolate, spreading; stigmatic-cap much broader than tall, stigma truncate; follicles slender, continuous or very slightly torose, 9-14 cm. long, acuminate, sessile, glabrous, 7-11-seeded; seeds 5-12 mm. long, oblong in outline, truncate or slightly tapered at the ends, variously pitted or wrinkled, brown. Distribution : woods and ravines, northern Georgia, Tennessee, Illinois, Missouri, southeastern Kansas, eastern Oklahoma, and northeastern Texas. Specimens examined: Georgia: Jasper City, 1847, Porter (G). Tennessee : Nashville, June, year lacking, Gattinger (G type, MBG); Nashville, islands in Cumberland River, September, 1878, Gattinger (MBG, NY, ANSP, F). Kentucky: barrens of the Kentucky River, exact locality lacking, 1860, Short (MBG). Illinois: Athens, 1861, Hall (G, P, MBG); Olney, Richmond Co., Turkey Creek bottoms, May 19, 1914, Ridgway 104 (G, MBG) ; Grantsburg, April 28, 1900, Baker (P) ; East Hannibal, June 6, 1913, Davis 898 (MBG, US); along road west of Fish Lake, St. Clair Co., July 16, 1898, Norton (MBG) ; damp shady thickets, American Bottoms, opposite St. Louis, May, 1845, Engelmann (MBG); Queens Lake, Clinton Co., May 20, 1917, Ledman (MBG) ; Venedy, May 18, 1926, Anderson & Woodson 5 (MBG); Conologue, May 16, 1926, Woodson & Stevenson U (MBG). Missouri: Alba, rich bluff woods, April 29, 1909, E.J. Palm- er 1819 (G, MBG); St. Louis, July 28, 1910, Sherff 801 (G, F, 410 ANNALS OF THE MISSOURI BOTANICAL GARDEN [Vol. 15 MBG); Webb City, gravelly branches, Sept. 2, 1909, E. J, Palmer 2620 (G, MBG); Winfield, Lincoln Co., June 7, 1916, Davis 1403 (MBG); Bower's Mill, Lawrence Co., rich hill-side woods, April 22, 1908, E. J. Palmer (MBG) ; Allenton, June 10, 1884, Kellogg (MBG); Elmont, May 23, 1914, Emig (MBG); Gascondy, July 21, 1914, Emig 221 (MBG); Gray's Summit, May 15, 1926, Greenman 4493 (MBG); Allenton, June, 1880, Letterman (MBG, US). Arkansas: Benton Co., date lacking, Plank (MBG); Eureka Springs, April 27, 1899, Trelease (MBG); Little Rock, May, 1886, Hasse (F). Oklahoma: Miami, on dry bank of draw, Aug. 26, 1913, Stevens 2337 (G, MBG, US); Page, on rocky mountain-side, April 25, 1915, Buckley 3425 (G, MBG); rocky hills, Wichita Mts. not common, July, 1891, Sheldon 224 (MBG). 5. Amsonia ludoviciana Vail in Small, Fl. Southeast. U. S. ed. 2, 935. 1913. Herbaceous perennial from a slightly thickened woody root; stems 5-11 dm. tall, pubescent, at least when young, sparingly branched, erect or ascending, the branches erect or ascending; leaves relatively distant, alternate, the blades elliptic, both base and apex acute to acuminate, 5-8 cm. long, essentially glabrous above, densely white-lanose beneath, pedicels 2-4 mm. long; inflorescence relatively dense, several- flowered, pedicels 2-4 mm. long; calyx 2-3 mm. long, the lobes triangular, 1-1.5 mm. long, pubescent; corolla salverform, the tube 5-9 mm. long, densely pubescent or villous without, the lobes about equalling, or slightly exceeding, the tube, lanceolate, spreading; stigmatic-cap about as broad as tall, stigma truncate; follicles slender, continuous, 8-10 cm. long, acuminate, sessile, manifestly pubescent, 6-10- seeded; seeds 5-12 mm. long, oblong-ovoid in outline, truncate at the ends, variously pitted and wrinkled, dark brown. Distribution: known only from southern Louisiana. Specimens examined: Louisiana: New Orleans, date lacking, Ingalls (NY); Shacky- nody, April, year lacking, Hale (NY). 1928] WOODSON — STUDIES IN APOCYNACEAE. Ill 411 Subgenus II. Sphinctosiphon (K. Schumann) Woodson Subgenus II. Sphinctosiphon (K. Schumann) Woodson, n. comb. ^Sphinctosiphon K. Schumann in Engl. & Prantl, Nat. Pflan- zenfam. 4 2 : 143. 1895; Dalla Torre & Harms, Gen. Siph. 406. 1904. Bracteoles conspicuous, giving the inflorescence a chaffy ap- pearance; mouth of the corolla-tube markedly constricted at an thesis; stigma apiculate by two distinct obtuse lobes; follicles continuous, not articulated, fibrous, not horny in texture; seeds j in outline, truncate at either end, variously pitted and wrinkled; plants of the southwestern United States and north- ern Mexico. Spp. 6-13. Section I. Micranthae Woodson. Corolla-tube 1-1.5 cm. long; calyx 1-4 mm. long; follicles 4-7 cm. long; seeds 4-8 mm. long. oblon Key to the Species a. Follicles slender; seeds fertile. b. Corolla-lobes 3-6 mm. long, ovate or oblong. c. Corolla-lobes 3-4 mm. long; plant entirely glabrous 6. A. Palmeri cc. Corolla-lobes 5-6 mm. long; plant pubescent, at least the calyx- lobes. d. Stem and leaves glabrous; pedicels 3-5 mm. long; inflores- cence loose 7. A. pogonosepala dd. Stem and leaves pubescent; pedicels 1-2 mm. long, or prac- tically sessile; inflorescence dense. e. Calyx glabrous; inflorescence many-flowered. . .8. A. hirtella ee. Calyx pubescent; inflorescence few-flowered. . .9. A. Standleyi bb. Corolla-lobes 6-8 mm. long, lanceolate 10. A. latifolia aa. Follicles short; seeds sterile 11. A. Kearneyana 6. Amsonia Palmeri Gray, Proc. Am. Acad. 12: 64. 1877; Gray, Syn. Fl. N. Am. 2 1 : 82. 1878. PI. 52, figs. 14-15. Amsonia Fremontii Rydb. Bull. Torr. Bot. Club 40: 465. 1913. nomen. Herbaceous perennial from a somewhat thickened root, gla- brous; stems 3-5 dm. tall, usually clustered from the base, erect or slightly ascending, sparingly branched above, the branches ascending; leaves alternate, relatively numerous, oblong-lanceo- late to linear-lanceolate above, the blades 2.5-7 cm. long, 4-8 mm. broad; inflorescence relatively few-flowered and loose, held well above the foliage; pedicels 1-3 mm. long, or practically 412 [Vol. 16 GARDEN lacking; calyx 3-4 mm. long, sparsely hairy, the lobes subulate; corolla salverform, the tube constricted at the mouth, 1-1.8 cm. long, the lobes ovate to ovate-oblong, 3-4 mm. long, erect or spreading; stigma apiculate by two distinct obtuse lobes; follicles 4-6 cm. long, acuminate, sessile, glabrous, continuous, 5-10-seeded; seeds 4-8 mm. long, oblong in outline, truncate at either end, variously pitted or wrinkled, chocolate-brown. Distribution : Arizona and New Mexico. Specimens examined : New Mexico: exact locality lacking, 1851-52, Wright 1669 (G type, MBG). Arizona: exact locality lacking, 1884, Lemmon 3248 (G); 50 miles s. of Lee's Ferry, June 12, 1890, M. E. Jones (P, US); Hillside, May 1, 1903, alt. 3700 ft., Jones (MBG) ; Beale's Spring, date lacking, Lemmon & Lemmon (US); exact locality lacking, 1887, Mearns 152 (NY). i 7. Amsonia pogonosepala Woodson, n. sp. Herbaceous perennial from a thickened woody root; stems 5-8 dm. tall, glabrous, clustered from the base, erect or slightly ascending, freely branched above, the branches ascending or spreading; leaves alternate, relatively numerous, glabrous, lance- olate to oblong-lanceolate, the blades 1-1.5 cm. broad, 5-7 cm. long, acute to acuminate at both base and apex, petiolate, the petioles 1-3 mm. long; inflorescence loose, relatively many- flowered; pedicels 2-A mm. long; calyx 3-6 mm. long, the lobes subulate, conspicuously cilia te, 2-5 mm. long; corolla salverform, the tube constricted at the orifice, 12-15 mm. long, glabrous without, the lobes 5-6 mm. long, ovate to oblong, spreading; stigma apiculate by two distinct obtuse lobes; follicles 1.2-8 cm. long, acuminate, sessile, glabrous, continuous, 4-15-seeded; seeds 7-10 mm. long, oblong-truncate in outline, variously pitted or wrinkled, reddish brown. 1 Amsonia pogonosepala sp. nov. 7 humila saepe basaliter ramosa 5-8 dm. alta; ramis erectis vel laxe ascendentibus glabris; foliis lanceolatis oblongo-lanceolatis petiolatis 5-7 cm. longis 1-1.5 cm. latis glabris; lobis calycis piloso-ciliatis subulatis 2-5 mm. longis; corollae lobis ovatis vel ovato-oblongis 5-6 mm. longis tubo sub- clavato dimidio brevioribus distendatis; stigmate subtrochleari apice bilobato; fol- liculis teretibus gracilibus continuis glabris sessilibus 2-8 cm. longis. — Arizona, dry rocky hills, San Francisco Mts., April, 1881, H. H. Rusby 256 (MBG type). 19281 WOODSON — STUDIES IN APOCYNACEAE. Ill 413 Distribution: southern Arizona. Specimens examined : Arizona: dry rocky hills, San Francisco Mts., April, 1881, Rusby 256 (MBG type, ANSP, NY) ; small sandy wash between Apache Junction and Canyon Lake, June 21, 1928, Harrison pi ign subscriptions: Wheldon & \\ 2, 3 and 4 Arthur St., N London, W. C. 2, England^M VFI OF THE MISSOURI BOTANICAL ( 1RDI n t Director GEORGE T. MOORE Assistant to the Director. Katherine H. Leigh. Herma von Schrenk, Pathologist. Jesse M. Greenman, C itor of the Herbarium Edgar Anderson, I ST VOLI Phys logist. David H> Linder, Mycologist. Roland a L, LaGardb, icist. ch nt. Nell C Horner, Librarian and Editor of Publication BOARD OF TRl \Tl OF THE MISSOURI BOTANICA] GARDEN President, Vice-President, iEORGE C. HITCHCOCK. Second Vice-President, SAMUEL C DAVI Daniel K. Catlin. Thomas S. Maffftt V. C. F. Meyer. rt T. Pej Phi li C. Scan John I Shepley. EX-OFFICIO MEMBERS Chancellor of Washington Universit\ Fr: aiCK F* Johnson, Bishop of the Diocese of Missouri. Victor J Miller, Mayor of the Git of St. Louit Georg T. Moore, President of Th V of St ivoui Ben Weidle, President of the Board of Educati t c t Lot m Daniel Brece Secretary