ESSAYS IN GEOBOTANY IN HONOR OF WILLIAM ALBERT SETCHELL .{j-^i/A^-ou..- ^ /it^tr'cXiU^ AYS IN E GEOBOTANY In Honor of WILLIAM ALBERT SETCHELL EDITED BY T. H. GOODSPEED . 0 WOOC MOU€ UNIVERSITY OF CALIFORNIA PRESS BERKELEY, CALIFORNIA 1936 UNIVERSITY OF CALIFORNIA PRESS BERKELEY, CALIFORNIA CAMBRIDGE UNIVERSITY PRESS LONDON, ENGLAND COPYRIGHT, 1936, BY THE REGENTS OF THE UNIVERSITY OF CALIFORNIA PRINTED IN THE UNITED STATES OF AMERICA BY SAMUEL T. FARQUHAR, UNIVERSITY PRINTER Editorial Preface In 1934, the Editorial Committee of the Uni- versity of California approved a proposal to issue a volume in honor of Professor Setchell, who, in 1 93^, after forty years of active par- ticipation in the life and wor\ of the Univer- sity, would become seventy years of age and Professor of Botany Emeritus. The editor was authorized to request articles dealing with the subject of the proposed boo\ from distin- guished botanists in this country and abroad. The magnitude of the editorial tas\ which proved to be involved, together with certain unforeseen circumstances causing delays, have made it necessary that this anniversary volume appear post festum. The editor is under great obligation to the various authors for their in- terest and cooperation, to Dr. Helen-Mar Wheeler, and to the editorial sta§ of the Uni- versity of California Press for much assistance and advice. Since the material in this volume brings to- gether authoritative statements concerning the status of the various problems of geobotany, its usefulness as a text for advanced students and as a boo\ of reference should be consid- erable. T. H. GOODSPEED, EDITOR. [VII] ^ ^ Contents PAGE Portrait of William Albert Setchell Frontispiece By PETER VAN VALKENBURGH William Albert Setchell: A Biographical Sketch .... xi By T. H. GOODSPEED The Rate of Plant Migration i By O. W. ARRHENIUS The Origin of Crepis and Related Genera, with particular reference to Distribution and Chromosome Relation- ships 9 By E. B. BABCOCK The Succession and Distribution of Cenozoic Floras around the Northern Pacific Basin 55 By R. W. CHANEY The Origin of the Desert Climax and Climate 87 By E E. CLEMENTS The Strand and Dune Flora of the Pacific Coast of North America: a Geographic Study 141 By W. S. COOPER The Genetic Phy to geography of the Southwestern Pacific Area, with particular reference to Australia .... 189 By LUDWIG DIELS The Role of the Terrestrial Alga in Nature i95 By E E. FRITSCH The Plant as a Metabolic Unit in the Soil-Plant System . . 219 By D. R. HOAGLAND 5894 ;5 CONTENTS PAGE Malaysian Phytogeography in relation to the Polynesian Flora 247 By E. D. MERRILL Plant Communities of the World 263 By EDUARD RUBEL Antarctic Plants in Polynesia 291 By CARL SKOTTSBERG The Published Writings of William Albert Setchell ... 313 WILLIAM ALBERT SETCHELL A Biographical Sketch By T. H. GooDSPEED WILLIAM ALBERT SETCHELL was bom on April 15, 1864, in Norwich, Connecticut. Remotely o£ Danish an- cestry,' the Setchell family came to America from the North Coast of England in late Colonial or early Revolution- ary times. They intermarried with the old Standish, Case, and Jans families, whose American beginnings were still older. His mother was born in England of English-Welsh parents, and came to America as a small child. He is, therefore, of a decidedly American descent which involved English, Dutch, and Welsh ancestry. From his earliest years, Profesor Setchell exhibited a decided predilection for "natural history!' As a small child, among other pursuits in this domain, he sought attractive and interesting plants from which he would pull portions and plant them, to be replaced, when faded, with others. This and other evidence of interest in plants ultimately attracted the attention of the family physician, who, when the boy was nine or ten years old, gave him a copy of Mrs. Lincoln's Botany for Young Ladies, which contained a species of manual arranged on the Linnean system. Therewith he puzzled out the names of many plants and made a beginning in systematics which cannot have failed to leave its mark, particularly since the labor involved in these first steps stimulated rather than dampened his enthusiasm for plant study. It was further aroused (in 1876) by a journey with his parents and his grandmother to Philadelphia and the Cen- tennial Exposition, with stops in New York City and elsewhere; [xi] XII T. H. Goodspeed he saw to it that visits were made to botanical gardens and con- servatories, and to zoological gardens and aquaria besides. When he v\^as sixteen and a student in Norwich Free Acad- emy, the curriculum there included a course in Gray's Lessons in Botany in which some attempt was made at determining species. The young man showed so lively an interest that he was asked to assist in collecting material for the class. He was also given special privileges; he could leave the classroom to follow his own desires in plant collecting, or pursue independent study in the botanical section of the school library. There is perhaps an indication here that the "project method" in secondary edu- cation had its champions even before the present generation of schoolmasters! His collecting enthusiasm was paralleled by that of George R. Case, Deputy Collector of Internal Revenue in the Norwich District, with whom Professor Setchell became acquainted about this time. Together they proposed the prepa- ration of a local flora to include all plants found within a radius of ten miles from the center of the city of Norwich, the species to be determined by Gray's Manual and to be arranged in order of flowering. In this last proposal appears the background of a research interest which was to engage Professor Setchell's efforts many years later. Finally, in 1883, the list of Norwich plants was published. Of further significance in this early period of Professor Setchell's botanical experience was his discovery of the fern Aspleneum montanum some hundreds of miles west of its previously known range, an event which was communicated to Professor Daniel Cady Eaton of Yale University. The appear- ance of his first scientific contribution, together with the enthusi- asm engendered by a botanical discovery, led the young man to consider himself somewhat established in a botanical career. When, in 1883, he entered Yale University, there was almost A Biographical SJ^etch xiii no instruction offered in botany, at least not to members of the classical course in Yale College. However, Professor Eaton, hav- ing some acquaintance with him as a result of the fern discov- ery and appreciating his botanical ardor, opened the way. He admitted him to his home, where, as Professor Setchell writes, "I occupied a table in a corner of his combined library, her- barium, and study room, a large and magnificent room with mantelpieces carved to represent certain North American fern species. A good library, a good herbarium, and a very sympa- thetic instructor!" Various factors in this stimulating and satis- fying environment went far in determining Professor Setchell's future botanical inclinations as well as broadening his immedi- ate botanical horizon. For a year or two he had been collecting seaweeds, and the presence of W. E. Sailord in New Haven en- couraged this pursuit. Saflord, then an ensign in the United States Navy, had obtained leave to study the marine algae under Eaton. Setchell and he immediately combined forces and soon began to make progress in their joint phycological studies. At the same time, he became acquainted with a fellow student, J. B. Hatcher, who was then studying the Hepaticae. Hatcher, who was graduated soon after Setchell's freshman year, obtained an appointment in the United States Geological Survey and was sent to Kansas to engage in paleontological field work. For two or three years he collected representatives of the plains flora, sending his specimens to Setchell, and with Eaton's aid, Setchell determined them. His work on this material, containing, as it did, genera and species almost all of which were new to him, made the first considerable contribution to Setchell's knowledge of the morphology and taxonomy of vascular plants and of the factors which influence their distribution. His earliest contact with what was to become his major re- XIV T. H. Goodspeed search interest came about 1880, when he made a series of visits to the seashore in Eastern Connecticut and collected marine algae. The algae were named according to the information con- tained in Hervey's Sea Mosses, and the phycological experience thus derived was supplemented by the later association with Eaton and Saflord, already referred to. An important influence which during this early period further stimulated Professor Setchell's growing enthusiasm for algology was an acquaintance- ship, made through Eaton, with Mr. Isaac Holden, of Bridge- port, Connecticut. He was Vice-President and Business Manager of the Wheeler and Wilson Sewing Machine Company, and spent his week ends and holidays in exploring the seashore and studying his collections. While Setchell was at Yale, and later when he was at Harvard, Holden and he collected and studied together and ultimately formulated a project which was to be realized some years afterward when they associated with them- selves another amateur phycologist, Mr. Frank S. Collins, of Maiden, Massachusetts. In the absence of any manual of North American algae, and because they were not then in a position to prepare one, they proposed to issue series of mounted and named specimens which would make available to the inter- ested public the products of their collections and studies. The ultimate result was the preparation of 51 fascicles of the Phyco- theca-Boreali-Americana (1895-1919), which distributed to cor- respondents throughout the world some 200,000 specimens and which represent a collection that will always be fundamental in investigations of the algae of North America. To the triumvirate who gained for the Phycotheca-Boreali- Americana its initial success there was added, during Professor Setchell's early years in Berkeley, N. L. Gardner, another phy- cologist who at the beginning had amateur status only. In 1897, A Biographical Sketch xv Gardner, then a school teacher in Coupeville, Whidbey Island, Washington, submitted to Professor Setchell his collections of seaweeds from the Puget Sound region, an area which had not been explored for marine algae in more than fifty years, and where Gardner found unusually large numbers of new or little-known species. The scientific and later personal and professional relation thus established between Professors Setch- ell and Gardner continued during the latter's undergraduate years at the University of Washington and preparation for the doctorate at Berkeley. For many years as colleagues at the Uni- versity of California they have collaborated in the preparation of a series of monographs, intended ultimately to cover the algal flora of the Pacific Coast. In his senior year at Yale, Professor Setchell had decided to continue his studies in cryptogamic botany, as an avocation at least. The question of a vocation, which had necessarily to be answered soon, resolved itself in his mind into a choice between becoming a teacher of the classics in preparatory schools, taking up the profession of medicine, or attempting to carry on his botanical studies to the point where he might hope to obtain a university appointment. This third alternative, although attrac- tive, seemed to him least likely of success until, on the recom- mendation of Eaton and with the aid of W G. Farlow, he ob- tained an appointment as a Morgan Fellow in Harvard Univer- sity to pursue studies in both botany and zoology. Professor Setchell's period of graduate study at Cambridge (1887-1890) was replete with those broadening and refining influences, scientific and cultural, which today, because of early specialization and the administrative and curricular pressure upon faculty and students, are so rarely available to the graduate student in most American universities. In Farlow's laboratory XVI T. H. Goodspeed he was associated particularly with Kingo Miyabe and W. C. Sturgis. Both of them were more advanced in their studies than he, but they worked with him sympathetically. Among his fel- low students in Mark's laboratory were G. H. Parker, H. H. Field, C. B. Davenport, H. B. Ward, C. H. Eigenmann, and C. W. Woodworth. These friends, together with advanced students in other fields — including John Manley and E. B. Delabarre — formed a most stimulating group. Besides their attachment to the natural sciences, they shared an enthusiasm for literature, painting, and music. They attended concerts and lectures, and visited the Boston Museum of Fine Arts; and their discussions in the laboratory were not restricted to scientific topics. The re- finement of Professor Setchell's taste, the breadth and quality of his literary background, and his interest in and knowledge of the fine arts are referable, in part at least, to these Harvard experi- ences superimposed upon his early training in the classics. During the two years following, at Cambridge, Professor Setchell served as an assistant in biology and held a proctorship. As proctor, he had to do with matters of discipline in the college dormitories and with the conduct of examinations. At the be- ginning of the second year he was granted the degree of Master of Arts, the only requirements for which were to request its conferment and to submit evidence of a proper record in course work. Since the assistantship permitted him only half-time for his graduate studies, he found it necessary to give up the re- search in zoology which he had begun under Mark and to give himself exclusively to the thesis topic under Farlow. The prob- lem decided upon for the thesis concerned the morphology and development of Saccoriza dermatodea, a characteristic kelp of the northern New England coasts. In the meantime, a certain small group of fungi, the Doassansia group, had aroused in him A Biographical S\etch xvii a new and keen interest, stimulated by his having obtained the germination of curious spore masses or sori. Farlow frowned upon this new enthusiasm as tending to occupy time better spent upon research on the thesis topic. Nevertheless Professor Setchell succeeded in continuing the Doassansia work by care- fully apportioning the morning hours in the laboratory: until 9 130 he concerned himself with Doassansia, thereafter with Sac- coriza, and thus was always engaged upon the latter when Far- low made his daily round. Together with fellow students in the laboratory, Professor Setchell was encouraged to call occasionally in the evening at Farlow's bachelor apartments. On these occasions the professor regaled the group with anecdotes having to do with the person- alities and work of many botanists, living and dead. By this means Professor Setchell acquired an extensive though some- what unassorted mass of information on botanical history and, more, an enduring interest in the subject. Many years later, a result of this manifested itself in the preparation of a semester of lectures entided "The History of Botany'.' This was the first organized course in the undergraduate curriculum at the Uni- versity of California which dealt exclusively with the historical development of a biological field. For many years it was deserv- edly popular with students of other disciplines besides the bio- logical. In 1890, Professor Setchell obtained the doctorate, and a year later accepted appointment as an assistant in botany in Yale University. Soon thereafter he began to consider, among other problems in the marine algae, the distribution of the kelps, and presented his preliminary conclusions before the Connecticut Academy of Science in a short paper entided, "Geographical Distribution of the Laminariaceae'.' In the discussion which fol- XVIII T. H. Goodspeed lowed, questions were raised concerning the factors which had brought about the distribution of certain marine algae, and in particular, the possible importance of the temperature factor was discussed. Here, then, was the origin of a point of view which Setchell has effectively developed. Indeed, one of his most significant research contributions has had to do with the estab- lishment of critical temperature intervals in relation to the dis- tribution of the algae and, more recently, of other aquatic plants and of terrestrial plants. With respect to the algae, for example, he has shown that an interval of 5° C. in amplitude determines the zone of distribution of a species, and the seasonal invasion of these zones by members of other zones. Five years later. Professor Setchell was called to the Univer- sity of California to become Professor of Botany and Chairman of the Department, an appointment which he held continuously until his retirement as Professor Emeritus in 1934. He was one of those younger men whom Benjamin Ide Wheeler brought to Berkeley to assist him in the building of the greater university— an institution which will never cease to reflect the foresight and wisdom of one who was perhaps the greatest of the older school of American university presidents. Among Professor Setchell's older associates on the faculty were Joseph LeConte, Hilgard, Rising, and Hesse, and among the younger, Ritter, Leuschner, Louderback, and Blasdale. Under E. L. Greene, his predecessor, the Department of Botany included Marshall A. Howe as an instructor, and as assistants W L. Jepson, Ivar Tidestrom, and J. Burtt Davy. Later, W. J. V. Osterhout and H. M. Hall became members of the staff. For many years, by virtue of his position as head of the Department of Botany, Professor Setchell acted as Botanist of the California Agricultural Experiment Station. In 1898, there was erected a small building provided with labora- A Biographical Sketch xix tories and certain herbarium space, and this was occupied by the department for more than thirty years — indeed, until all bio- logical units were housed in the Life Sciences Building. Over a period of forty years. Professor Setchell has signifi- candy contributed to the life of the University of California. At the beginning he had much to do with the organization of the greater university. He served on administrative and faculty committees, and was one of those who interpreted the academic atmosphere and traditions of Harvard and Yale in terms of Western ideals and aspirations. Of special importance to the University was the experience he had acquired in his years as a proctor at Cambridge, in dealing effectively and sympatheti- cally with undergraduates. For many years he was both official and unofficial adviser to members of the student body, and their appreciation was expressed by his early election to membership in their honor societies. In his official capacity he assisted in the formulation of the then unique system of student self-govern- ment, suggested and fostered by President Wheeler and per- petuated by later administrations, which has been adopted by so many institutions of collegiate and lesser grade as one of the surest means of developing individual and group loyalty and responsibility among undergraduates. In his unofficial capacity he still continues to act in loco parentis to many young men and particularly to those in whom he sees promise, and in many parts of die world his former students and foster children ac- knowledge their indebtedness to him for the guidance, encour- agement, and assistance — often material assistance — which he so wholeheartedly offered them in their college days. Setchell's scientific contributions reflect a remarkably wide range of interests, some having their origin in the influences and environment which were part of his biological training and XX T. H. Goodspeed which have already been mentioned, others being referable to influences which have operated more recently, and still others being somewhat incidental products of circumstance. Certain of his diverse research concerns represent fields which he plowed and later passed on for cultivation to colleagues and students. Soon after coming to Berkeley, Setchell met Dr. H. W. Hark- ness, President of the California Academy of Sciences, who had long studied the subterranean or hypogaeous fungi, and who desired assistance in preparing his observations for publication. This incidental meeting revived Setchell's interest in the fungi which had had its beginning in Farlow's laboratory, and as a result he has been engaged for many years with the aid of N. L. Gardner, Helen M. Gilkey, Lee Bonar, and H. E. Parks, in assembling representative collections of California species. Though the fungi herbarium and its study have been turned over to Professor Bonar, these collections still enlist his interest. Later on, in the Botanical Garden, of which he was Director for many years, the possibility of growing for instruction and display a collection of the commercial varieties of the tobacco plant and other species of Nicotiana led to the obtaining of seed from foreign as well as domestic sources. The early trials of this seed indicated so much confusion with respect to nomenclature that he began to give serious thought to the situation. Rather extensive garden cultures soon produced a series of stable and correcdy named species and varieties and, in addition, evidence that the races of commercial tobacco represented hybrid com- binations. For some years Setchell gave up his summers to an analysis of varietal hybrids in Nicotiana Tabacum and to the re- sults of hybridization among members of the collection of other species which he now had available. From these early studies is derived the extended research program dealing with the cyto- A Biographical Sketch xxi genetics of the genus Nicotiana which for many years has been carried on by R. E. Clausen, the writer, and their students. In these illustrations of Setchell's more or less incidental research interests are revealed a scientific curiosity and an enthusiasm which were so balanced by a sense of values that profitable ven- tures have been assured to inheritors of these projects. In 1920 came what proved to be a major influence upon his career, in an invitation from the Carnegie Institution of Wash- ington to carry on certain investigations in the Samoan Islands. From this beginning there was derived a sustained interest in a variety of Polynesian problems — the building up of coral reefs, the origin of insular floras, the biological as well as the geological history of the crustaceous corallines, ethnobotany — and from it, too, have come his important contributions concerning the paleo- geography of the Pacific area. In the following ten years he made a series of journeys in this area, and brought back exten- sive collections from Hawaii, Tahiti, Tonga, Fiji, Australia, New Zealand, Japan, and South Africa. His undergraduate training in the systematics of vascular plants formed the background for an accumulating knowledge and experience in this field which enabled him to collect with discrimination the significant ele- ments of the Pacific floras with which he came into contact. From these expeditions he acquired first-hand knowledge of the geographic and geologic as well as of the floristic features of the Pacific area in their relation primarily to the fundamental concepts of plant distribution involved. A large place in his thinking today is occupied by these matters, and by his con- tinued investigation of the factors which have brought about the distribution of the marine algae. It is therefore not inappro- priate that this volume which honors his scientific achievements should be concerned with geobotany. XXII T. H. Goodspeed Soon after his coming to California in 1895, Setchell under- took an exploration of the West Coast of North America, look- ing toward a complete treatment of its algal flora. This project has continued to be his major research concern for more than forty years. His phycological contributions, and particularly his monographic treatment, in collaboration with N. L. Gardner, of numerous groups of Pacific Coast algae, have to do with a field in which he was a pioneer and which he has cultivated assiduously and successfully. Early in this work he visited on more than one occasion those European herbaria in which he could familiarize himself with types in the algae. At the same time, he collected extensively from California to Alaska and, on his first journey around the world, in 1903-1904, collected and studied new or litde-known algal floras. The broad character of the background thus acquired, together with research enthu- siasm matched by industry, a rare sense of morphological values, and a mature judgment, are reflected in the fullness of his phy- cological contributions and their critical quality. Relieved of teaching and administrative duties as Professor Emeritus, Setch- ell is increasing his productivity. At present he is engaged, among other things, upon the monographing of certain large genera of the algae, such as Sargassum and C odium, and of various genera of the Balanophoraceae; consideration of clima- tological relations of the Wallacean area; and studies of the morphology, taxonomy and, particularly, the distribution of the genus Zostera. In 1920 he married Mrs. Clara B. Caldwell, of Providence, Rhode Island. Mrs. Setchell immediately endeared herself to his colleagues and other friends by her tact, dignity, and sym- pathetic reaction to her new environment. In a quiet and un- pretentious fashion she entered into the Hfe of the laboratories, A Biographical Stretch xxiii coming each day to assist her husband, first in the organization and classification o£ his Hbrary, notes, and collections, and later, having perfected herself in microtechnique, in the researches which he had in progress. During the next fifteen years, Mr. and Mrs. Setchell undertook together a series of extended scien- tific expeditions, principally in the Pacific area. They both en- joyed travel, and in this as in every other aspect of their life together, they found nothing but experiences of mutual interest. Mrs. Setchell's unselfishness and serenity of spirit were never more apparent than during the more than two years of disabil- ity which preceded her final illness, and her obvious desire that her condition should as little as possible interfere with the prog- ress of her husband's work was realized. In this brief biographical statement no effort has been made to recite the total of Professor Setchell's scientific concerns or achievements. Nevertheless, by illustration and comment the breadth of his biological horizon and the importance of his research contributions have, hopefully, been exhibited. An obvi- ous effort has been made throughout to indicate the character of the associations and influences which have operated in init- iating and directing his scientific pursuits. Given those qualities of mind which he possesses, certain, at least, of the diverse influ- ences which are sure to approach the biologist may be produc- tive of the broader, more significant, and more useful type of scientific career and should, therefore, not be thrust aside with- out due consideration. If a somewhat personal tone appears as characteristic of some parts of this effort to present a sketch of a distinguished man, it must be referred to the writer's intimate acquaintance with him for more than a quarter-century. Among the thirty scientfic societies and other organizations of which Professor Setchell is at present a member, the National XXIV T. H. Goodspeed Academy of Sciences, American Pliilosophical Society, Ameri- can Academy of Arts and Sciences (Fellow), California Acad- emy of Sciences (Fellow), Washington Academy of Sciences, Torrey Botanical Club (Fellow), Linnean Society of London, Societe de Biographic, Societe Linneenne de Lyon, Societe pour I'Etudes Oceaniennes (Fellow), American Anthropological Association, American Geographical Society, American Asso- ciation for the Advancement of Science (Fellow, and Past Vice- President, Section G — Botany), Phi Beta Kappa, Sigma Xi, Alpha Zeta, Phi Sigma (Past National President), Alpha Epsi- lon Sigma, Author's Club (London), Bohemian Club (San Francisco), represent some of those which have honored his scientific achievements and include certain others which indi- cate the range of his contacts, scientific and otherwise. The Principal Journeys and Expeditions of W. A. Setchell 1896. Expedition during the summer by wagon from Berkeley to the Santa Cruz Mountains, California, through the San Joaquin Valley to Yosemite Na- tional Park, and back to Berkeley (with W. L. Jepson) — collections of vascular plants, fungi, and fresh-water algae. 1898. Yellowstone National Park — investigation of thermal algae. 1 899. Expedition to Alaska and the Bering Sea (with L. E. Hunt, W. L. Jep- son, and A. A. Lawson) — studies of marine algae and general cryptogamic col- lecting. 1900. Summer in Hawaii (with L. E. Hunt) — studies of marine algae and ethnobotany. 1 903-1 904. First sabbatical leave from the University of California. Around the world eastward from Berkeley by way of Philadelphia, Europe, Egypt, India, New Zealand — herbarium studies, investigation of marine algae and thermal algae, and general collections. 1908. Summer in European herbaria — study of types in the algae. 191 1. European herbaria — study of types in the algae. 1920. Summer in Tutuila, Samoan Islands — investigation of the origin and development of coral reefs, studies in ethnobotany, general collecting. 1922. Summer in Tahiti (with H. E. Parks and Mrs. Clara B. Setchell) — further investigation of the origin and development of coral reefs and problems of plant distribution in the Pacific area, general collecting. 1924. Summer in Hawaii (with Mrs. Clara B. Setchell) — further investigation of coral reefs and distributional problems. 1926-1927. Second journey around the world, westward from Berkeley (with Mrs. Clara B. Setchell). Tonga (with H. E. Parks, J. E. HoflFmeister, and J. M. Ostergaard), Fiji, Samoa, Hawaii, Japan, Australia, New Zealand, South Africa — continuation of earlier studies of problems of the Pacific area; collections of marine algae, general collections. 1929. To Java and the Pan-Pacific Congress, Great Barrier Reef of Australia, and Rarotonga (with Mrs. Clara B. Setchell) — collections of tropical marine algae, general collections. 1931. Summer in Alaska (with Mrs. Clara B. Setchell) — taxonomic and dis- tributional studies of the genus Salix. 1932. Summer in Alaska and journeys on the Yukon (with Mrs. Clara B. Setchell) — continuation of Salix studies. [ XXV ] I The Rate of Plant Migration By O. W Arrhenius IN THE SKERRIES (small islands) along the east coast of Sweden the migration of plants plays a great role. Every year the land rises from about 0.3 to 1.5 cm. New areas are constantly wrested from the sea, and thus the water level in the soil changes from year to year. Consequently the changes in the plant associa- tions are very great, and the rate of migration is important in the formation of the new associations. Usually these skerries are small — one acre or less. Their high- est parts are rocks washed clean by the waves. Between these rocks, in sheltered places, we find the soil apt for vegetation. The most common profile of a place covered with vegetation in the younger skerries is shown in figure i. Rocks Main direction of the wind Fig. I. In order to examine these and related questions, some "bare areas" were laid out in the early summer of 1920. The areas were prepared as follows. In a typical plant association, a frame of I square meter divided with iron strings into 100 small squares was laid out. The plants were counted separately in each small area. After this the humus layer was dug away. The upper part of the mineral soil was then removed, put through a fine sieve, [I] 2 O. W. Arrhenius and replaced. The humus also was sieved and replaced on top of the mineral soil. The corners of the area were marked with poles. This bare area, of course, had not the same properties as the old soil; many of the factors were changed in one direction or the other. These areas were visited in 1922, 1924, and 1932. The district of these investigations belongs to the most arid region of Sweden. The annual rainfall is low — from 300 to 400 mm. The moisture of the air is rather high, but the evapora- tion is strong because of the hard and everlasting winds. The water is the dominating factor. Tables i to 6 (below) present the results of the investigations. The figures give the real numbers of the plants in each square meter. As a plant the section between the soil and a stem has been counted. When it is possible and of interest, the location of the different plants is given on a map. This shows the number of plants of different species in every small square. The expla- nation of the signatures is found in the tables; for example, V5 in table i marks 5 Vaccinium vitis idaea. From the tables, we see that the plant migration is going on very slowly, especially where the soil is dry. The most extreme condition we find in table 3, where the plants have not succeeded in establishing themselves after twelve years. Most interesting to see, however, is how the migration goes on. The migration has always taken place from the border. It is those squares which are situated next to the old vegetation which show new vegetation. And where the areas are subject to a prevailing wind — from the shore — we see that the migration chiefly has taken place from the windward side. The long-distance migration must be of very small importance in these associations. The Rate of Plant Migration 3 We may summarize the results as follows. The migration of plants in this district is very slow. The rate of migration is chiefly dependant upon the moisture content of the soil. The chief migration takes place within a very small distance— 10-20 cm. — and by the aid of the wind. Q o OS u CO & w I < ess id o K H o PQ < o Q o o en i-i > o o < > I < ^9 •J •J < u I M iz; > <>< > *-4 > > > 1-H > > ^-4 > > > 1— < > > o I— < > < > 1—4 > > >S > > > > > > > > > > 1— t 1— 1 o OS C4 ON o o\ > > > > 1-H > > 1— t > CM > > > > >u > > «— 1 > C<1>-H > > > r-4 U CM > >u CM > u U > 1— ( > u CM > > > >u > > r4 > > CM > CM > CM > > > > 1—1 > > 1-H > CM > CM > 1— ( > U CM > 1— I > u CM > > > > cs > Cv» > CM > CM > > > > > ^-4 > CM > CM > CM > > > CM ON CM CM ON ; w^ u^ o CM ON 3 bfl o o CO CO O 1 S a X 3 i-i (U 3 k> s 3 c j3 cS rt 3 TABLE 2 — Herbrich Pinus Wood on Munko Thick humus on fine sand — Level 1922 HI Rl Fl Bl Rl Rl VI VI Rl Al Gl CI VI Yl Nl Rl Rl 11 PI PI Al LI Gl Al Gl CI Bl Rl Fl PI Rl VI PI Rl V — Vaccinium vitis idaea G — Geranium sanguineum F — Fragaria vesca L — Luzula pilosa Melampyrum nemorosum H — Hypochaeris maculata Y — Lathyrus pratensis G — Gnaphalium dioicum B — Galium verum Vicia silvatica Salium boreale R — Rubus saxatilis Chrysanthemum leucanthemum Filipendula hexapetala Solidago virgaurea Festuca ovina Clinopodium vulgare Briza media Sorbus aucuparia Anemone hepatica Linum catharticum I — Viola canina Hieracium pilosella Achillea millefolium Veronica officinalis P — Pimpinella saxifraga Trifolium sp A — Anthoxanthum odoratum N — Angelica silvestris C — Luzula multiflora 1920 398 54 18 20 2 1 10 50 21 1 37 4 11 9 38 46 2 3 1 1 1 5 33 5 42 13 4 1922 4 3 2 1 1 1 3 1 2 1924 The area covered with a dense vegetation of the same type as the surrounding association TABLE 3 Arctostaphylus- Pinus Wood on Korsholmen Raw humus on coarse gravel, very dry Plant 1920 1922 1932 Arctostanhvlus uva ursi 180 107 1 .... Aira flexuosa Hieracium sd TABLE 4 Empetrum-heath on Maskaren Mild humus on rather moist sand. The island crowded with seagulls and other birds — Level Plant Empetrum nigrum... Aira flexuosa Viola tricolor Rumex acetosa Veronica spicata Tanacetum vulgare. Stellaria graminea... Rubus chamaemorus.. Sagina nodosa Anthoxanthum odoratum 1920 1922 Plant 1920 1922 160 14 6 1 10 3 3 1 8 74 12 28 1 13 180 70 Festuca ovina 51 Sedum acre 18 Veronica chamaedrys Galium verum 11 17 Myosotis sp 8 Potentilla anserina Rubus saxatilis 9 6 Valeriana officinalis.... Draba verna 5 4 Viola sp 2 Allium schoenoprasum 1 TABLE 5 Myrtillus Picea- Pinus Wood on Skaten Thick raw humus on sand and gravel Plant 1920 1922 Plant 1920 1922 Vaccinium myrtillus.. Aira flexuosa 216 112 74 6 5 4 Melampyrum pra- tense 12 3 1 Vaccinium vitis idaea Calluna vulearis Empetrum nigrum Sorbus aucuparia .... TABLE 6 o Dry Herbrich Association on Asmansskaret Mild humus on sand on the top of the slope. Very dry 1922 X G2 G3 IX G3 Bl Y2 mi G2 HI G2 VII Y2 Gl B2 G2 VI G2 Tl Y3 V Yl HI GS IV Tl Bl G4 III Gl Tl Y3 II Bl G2 I G4 G4 CI GS T5 Fl GS CIO AS G3 C4 A2 T2 G2 C4 T3 AS Tl Bl C6 T3 Gl Shore A Aira flexuosa G Galium verum C Clinopodium vulgare.... T Tanacetum vulgare Festuca ovina F Fragaria vesca Pimpinella saxifraga Rumex acetosa B Arabis hirsuta Y Hypericum perforatum Filipendula hexapetala. Scutellaria hastifolia Veronica spicata Plantago lanceolata H Hieracium pilosella Valeriana officinalis Draba verna 1920 919 232 147 57 19 16 S 4 3 1 2 1 1 1922 12 53 25 16 1 6 11 1924 1* 92 30 IS 16 II III IV 39 19 5 4 1 10 1 44 13 4 1 14 37 26 35 14 1 1 VI VII VIII IX X 40 36 1 1 • The Roman numerals show the rows of small squares as indicated in the map I The Origin of Crepis and Related Genera, with Particular Reference to Distribution and Chromosome Relationships By Ernest Brown Babcock THE PHYLETic RELATIONS of the gcnus Crepis cannot be adequately discussed without some consideration of nearly related genera. Although none of these other genera has been as thoroughly studied as Crepis, sufficient in- formation is available to indicate rather definitely that the posi- tion of Crepis with reference to its closest allies is unique. In order to make this clear, it is proposed, first, to review the evi- dence on comparative morphology, chromosome number, and geographic distribution of these closely related genera so as to present, as clearly as possible in brief and rather general form, a review of this small portion of the Compo sitae; and, second, to summarize the much more detailed information now available concerning Crepis, in order to see what further light this may throw on the problem of origin. But, before discussing this group of genera, a few remarks on criteria of relationship may be in order. This subject has been discussed by the present writer in an earlier publication^* in which it was approached from the cyto- genetic side. The same general principles that underlie the classi- fication of individual organisms into species must apply mutatis mutandis to the classification of species into genera. In other words, just as species are natural groups of individuals, so genera are, or should be, natural groups of species. The criteria of re- • Superior figures refer to items in the bibliography at the end of this essay. [9] 10 E. B. Babcoc\ lationship which have been used in the present study are com- parative morphology, geographic distribution, and chromosome number, size, and morphology, and with respect to certain species additional evidence has been derived from genetic data, especially on interspecific hybrids. The morphological features of greatest usefulness in this group of plants are mentioned be- low. Like most of the Compositae, these species are often diffi- cult to recognize and highly variable or polymorphic. Many of the species which have hitherto been published have had to be revised because they were based on inadequate material or too superficial study of the morphology. The species with which this essay deals have all been carefully reviewed, if described by others, and a number of new species which are included have been so recognized only after critical investigation. The importance of geographic distribution to the student of plant taxonomy has long been recognized. Studied in connec- tion with the evidence from comparative morphology, the facts of distribution may throw light on the phyletic relations of species as well as of larger groups, and may indicate centers of origin and distribution. Field studies on the variation and ecology of the living plants in nature has also come to be widely appreciated as an invaluable aid in taxonomy. But the individual investigator is definitely restricted in the amount of such work that is practicable, and, in attempting to classify the species in any large, cosmopolitan genus, finds it impossible to apply this method very generally. In recent years, however, the comparative study of the chromo- somes, especially of their number and morphology, has come to be recognized as a useful method of supplementing the evidence from gross morphology and geographic distribution. At the same time, the accumulation of a living collection makes it The Origin of Crepis and Related Genera 1 1 possible to compare these cultivated species and forms when grown under fairly uniform conditions. Although this can never take the place of thoroughgoing field study, yet it does compen- sate to a degree for the inability to see the plants growing in the wild. For there is no doubt that the inherent morphological and physiological peculiarities which distinguish species are faith- fully reproduced under cultivation, often with an obviousness that is not found in the study of herbarium material alone or even in limited field observations. In the two decades during which the writer has applied part of his time to the study of Crepis, fully half of all the known species have been introduced into cultivation and many of them have been maintained for use in genetical and cytological re- search extending over a long period. Some of these species have been represented by accessions from many localities throughout the range of distribution, and of others only one or a few acces- sions have been obtained. Every species in the collection has been examined cytologically, and several reports have been pub- lished on the chromosomes in relation to taxonomy and phy- logeny in this genus.^' ^' *' "' '^^ Meanwhile, the taxonomic litera- ture on the genus has been reviewed and most of the species have been studied critically by the use of herbarium material, at least. In the second part of this paper, the data on geographic distribu- tion in this genus will be summarized and compared with the evidence from chromosome relations with reference to phylo- geny. Let us first consider the general aspects of the genus and its position with reference to other most nearly related genera. Crepis and Some Related Genera The genus Crepis is a natural group of more than two hundred species, being the second largest genus of the sub tribe Crepidinae 12 E. B. Babcoc\ of the Cichorieae, the tribe which includes all species of the Com- positae having only ligulate florets in the flowerhead. The largest genus is Hieracium, with hundreds of species and thousands of subspecies, varieties, and forms, the constancy of which (at least of many) is made possible by some type of apomictic reproduc- tion. In marked contrast with this condition in Hieracium is the nearly complete absence of asexual reproduction in Crepis. A few species are stoloniferous and certain others spread from root sprouts, but apomixis is apparendy at a minimum and actually has been demonstrated in only one species, C. tenui folia, which is a peculiar species in other respects. (The occurrence of apo- mixis is also suspected in certain American species with odd chromosome numbers.) Next to Crepis in size is Lactuca, which in its present vaguely defined state contains between seventy- five and one hundred species. These largest genera, Hieracium, Crepis, and Lactuca, are well differentiated, although there are borderline species or sections as well as certain small inter- mediate genera which are discussed below. The other larger genera in the Crepidinae are Chondrilla, Taraxacum, Launea, Sonchus, and Prenanthes. All these but the last are sharply dis- tinguished from the three largest genera. Prenanthes, like Lac- tuca, is at present rather vaguely defined. At any rate, both are in need of thorough revision. Nevertheless, for purposes of the present discussion it is possible to deal with them as well-known groups. These four larger genera, Crepis, Hieracium, Lactuca, and Prenanthes, include annual, biennial, and perennial herbs, a few of the perennials being somewhat suffrutescent. The earlier leaves usually form a rosette. The flower stems may be simple, scapiform, and merely bracteate, or more or less branched with few or many heads and with the cauline leaves more or less re- The Origin of Crepis and Related Genera 13 duced, the uppermost being bractlike. In certain very low or tufted alpine species the flower heads may be few and on very short scapes, or numerous and congested. The heads are large, medium, or small and often many-flowered. The involucre is generally composed of distinct outer and inner series of bracts. The receptacle is naked or shortly ciliate (in two small sections of Crepis it is paleaceous) . The corolla is tubular below, ligulate above, and 5-dentate. The anther tube and style branches are generally similar throughout except in color and such details as the appendages, and the pappus consists of simple, barbellulate brisdes. These are the outstanding resemblances between the five genera. There are, however, numerous differences in the vegetative parts, including the indumentum, also in details of the inflorescence; and differences in the achenes and pappus, in habit, and in color of the flowers are particularly useful for pur- poses of classification. The differentiation in respect to some of these features has been summarized from the diagnoses of Hoff- man,^* Bentham and Hooker,^ Robinson and Fernald,'^ and others, and is presented in very general, condensed form in table I below. There are several small genera in the Crepidinae which re- semble Crepis more or less, and no doubt they all deserve care- ful consideration in any future revision of this subtribe. For purposes of the present discussion, five have been selected which have been merged previously with other genera and all of which, the present writer believes, must be reinstated. It is not the pur- pose of this discussion, however, to present the complete evi- dence and formally to propose their reinstatement. They will be used merely as illustrations of the usefulness of the cytological criterion. Ixeris Cass, has been included in Lactuca by Hoff- mann and it is certainly closer to Lactuca than to Crepis, but in PQ < •< a: 0. U ss =^ t: w < h a CO CO « •-' O W r-r. CO «^ ^ CO t ^ ^, o zi Hiera- cium Prenan- thes The Origin of Crepis and Related Genera 15 the strict sense of Cassini's definition' it cannot belong in Lac- tuca, because the achenes are not strongly compressed and they are lo-ribbed, the ribs being equal and often salient or rounded. Hence the present author agrees with Gray" and Nakai^ in their reinstatement of this genus. The genus Youngia Cass, is included in Crepis by Hoffmann, and it resembles Crepis more than it does Lactuca; but unfortunately Youngia has been made a re- pository for a number of species which cannot be classified here in the sense of Cassini's definition.* Because of the strongly ob- compressed marginal achenes and certain other details, Youngia Cass., sensu stricto, must also stand as a genus intermediate be- tween Crepis and Lactuca. The genus Aetheorrhiza Cass, is monotypic and the species has long been known as Crepis bulbosa (L.) Tausch. Because of the imbricate involucre and the peculiar achenes, which are superficially like those of Crepis but are actually unique, this genus should be reinstated. Another species to be excluded from Crepis is Paleya oligocephala Sch. Bip. The multiseriate, imbricate involucre is alone sufficient to bar it from Crepis, Lactuca, Hieracium, and Prenanthes. As the type species of this genus was Crepis albida VilL, the genus has become synonymous with Crepis. The only other species that has been published under Paleya (P. billotioides Sch. Bip.) is also received under Crepis. Hence P. oligocephala may constitute a monotypic genus. However, its relationships to certain other dubious species await investigation. In addition to the foregoing, two small genera have been erected by Nakai^ for the disposition of certain species which he considers neither Crepis nor Lactuca. They both resemble Ixeris in some degree, but one of them, Crepidiastrum, is strongly suflrutescent or shrublike and differs in habit, leaf arrange- ment, and especially in the Crepis-likc achenes, which are terete 1 6 E. B. Babcoc\ and beakless, with lo fine, rounded ribs. In fact, both involucre and fruits are more Hke Crepis than Ixeris. The other genus, Paraixeris, also resembles Ixeris somewhat, but differs in having the stem much branched, the fruiting heads nodding, and the achenes subcompressed with 14 or 15 narrow, rounded ribs and a short coarse beak. Nakai's contention that these two genera are intermediate between Crepis and Lactuca may be well founded, but in my opinion they are both nearer to Crepis than to Ixeris and, as will be shown below, their chromosome num- ber supports this view. At the same time the relations of these two small genera to Prenanthes should be investigated, particu- larly since in both there is a strong habital resemblance to this genus. We turn now to the cytological evidence. Although the data on chromosome numbers are far from complete, yet something is known about most of these related genera. At present, the basic haploid numbers as known in the four larger genera are: Hieracium, 9; Lactuca, 8, 9; Taraxacum, 8; Prenanthes, 9, 8; and in the smaller genera under discussion the numbers are: Crepidiastrum, 5; Paraixeris, 5; Chondrilla, 5; Ixeris, %\ Youngia, 8; Paleya, 8; Aetheorrhiza, 9. In Crepis, after the exclusion of those species which on morphological grounds belong in other genera, the numbers 8 and 9 do not occur as basic haploid num- bers, whereas 5 is a basic number. This indicates a closer relation between Crepidiastrum, Paraixeris, and Crepis than between Crepis and the other excluded genera. Furthermore, all the species of Ixeris, Youngia, and Aetheorrhiza thus far examined have chromosomes which are definitely smaller than those of most species of Crepis. Thus the chromosomes provide important evidence in sup- port of the delimitation of Crepis as oudined above. Cytology The Origin of Crepis and Related Genera 17 has been similarly helpful in determining the proper classifica- tion of certain dubious sections of Hieracium. The section Crepidisperma, in particular, is characterized by having the achenes more or less attenuate toward the summit, as in Crepis, and the pappus white as in most species of Crepis. Because of these features of fruit and pappus, certain species were originally described as Crepis, for example, C ambigua A. Gray. Cyto- logical examination of this plant showed the haploid number to be 9, which is characteristic of Hieracium and not of Crepis, and it is probable that the other species of this section have the same base number. Thus cytology supports the inclusion of these species in Hieracium. There exists a very puzzling group of high alpine Himalayan plants which have come to be known as the "Glomeratae" from the type species, Prenanthes glomcrata Dene., although this species and four or five similar ones have long been known as Crepis. The evidence from comparative morphology, however, strongly supports Decaisne's original classification in Prenan- thes, and now it is known that the haploid number of chromo- somes in P. glomerata is 8, a number not found in Crepis except in certain polyploids. For the present, however, this group may be considered as intermediate between Prenanthes and Crepis. The occurrence of 5 as a base number in several species of ChondrillcT raises a special problem concerning the relationship of this genus to Crepis. It has long been considered as closest to Taraxacum because of the similarly specialized achenes in the two genera. Although the base number for Taraxacum is 8, both genera might have sprung from primitive Crepis stock. This is only one of many similar problems awaiting investigation. Without going into further detail concerning the evidence from comparative morphology and cytology, and as a first step i8 E. B. Babcoc\ merely toward a better comprehension of the natural relation- ships of the genera under discussion, we may represent their phylogeny somewhat as in figure i. This arrangement is in CHROMOSOME NUMBER AMD PHYLOGEMY IW THE CREPIDIIVIAE HICRACIUM 9 CRCPiS 5.4 LACTUCA 8.9 Fig. I . Chromosome number and phylogeny in the Crepidinae. ^nly one species reported with respect to chromosome number. general agreement with the natural system worked out by Cassini" on the basis of comparative morphology. On the basis of chromosome numbers alone, Crepis may not appear to have had a common origin with Hieracium, Lactuca, Prenanthes, and most of the other genera which have been examined cyto- The Origin of Crepis and Related Genera 19 logically. Nevertheless, it does not seem probable that any other genera in the Crepidinae will be found to be more closely re- lated to Crepis than these. (Two small genera, Pterotheca and Rodigia, both having 5 as the haploid number, are merged with Crepis by the author.) Reference again being made to the evi- dence from chromosome numbers, the marked divergence of Crepis, Paraixeris, Crepidiastrum, and Chondrilla from the other genera represented in figure i, may point to certain evolutionary processes underlying the differentiation of these and perhaps other genera in the Cichorieae. As will be shown below, our present knowledge of Crepis makes it possible to advance a tentative hypothesis concerning the nature of these processes. The evidence from geographical distribution supports the con- ception of a common origin for Crepis and its allies. If the dis- tributional areas of the three largest genera are compared, it is found that they all include most of Eurasia and at least part of Africa (fig. 2). All the other genera which have been studied cytologically occur in this same geographical unit. This suggests a common origin for these genera, and the distribution of the more primitive types of Crepis and its allies indicates that the region of common origin is south-central Asia. Additional evi- dence will be presented in the second and third parts of this paper. The fact that Hieracium is so widely distributed might at first seem to indicate that it is therefore the most ancient. But this is not necessarily true, because certain unique features of this great genus are sufficient to explain its present large spread over the earth's surface. According to Zahn,*' there are numerous species of Hieracium which are intermediate between other principal species; these "Zwischenarten" often produce numberless prog- eny and sometimes considerably transgress the distributional The Origin of Crepis and Related Genera 21 range of the presumably parental species. In fact, the distri- bution of Zwischenarten far beyond the range of one or both parents has been frequently observed. Although many of these intermediate species are doubtless hybrids, their distributional dominance is easily explained by the high frequency of apomic- tic reproduction in this genus. Predominance of interspecific hybrids has been observed in many other genera, but in Hiera- cium it has been an important factor in bringing about the wide distribution of the genus. Undoubtedly another impor- tant factor is the usually small size of the achenes, which makes them especially well adapted for wind distribution. In Crepis, on the contrary, the fruits of many species, especially the more primitive ones, are larger and heavier and apomixis is almost unknown. Yet the distribution of Crepis is nearly as extensive as that of Hier actum, the chief diflference being that Hieracium occupies South America whereas Crepis is not known to occur south of northern Mexico. But Crepis, even after certain species are excluded, includes some of the most primitive types in the Crepidinae (e.g., C. pontana and C. sibirica). Furthermore, Crepis is more widely distributed in Africa than is Hieracium. Hence it is probable that Crepis represents a progenial stock from which both these genera were derived. The wide distribution of Hieracium in South America and the complete absence of Crepis from that continent present a distributional problem calling for more detailed study. A sug- gestion worth considering is that distributional relations such as these will eventually be found to have a bearing on the con- cept of continental displacement. Good" states that, although there is no satisfactory hypothesis of continental displacement, it is believed that one will be forthcoming and that the geog- raphy of the flowering plants will afiFord important evidence in 22 E. B. Babcoc\ its support. However, if both genera originated in south-central Asia and migrated into North America by way of a land bridge, even though Crepis is the more primitive, a difference in time of migration and the special adaptations of Hieracium species for rapid dissemination would explain the present distributional relations of the two genera. Furthermore, the South American species of Hieracium having attenuate achenes and fine white pappus must be among the most advanced members of the genus. It is well known that various species of flowering plants in South America are closely allied to North American species and appear to have migrated from north to south. Lactuca (fig. 2) has a distributional range about as extensive as that of Crepis but, in certain morphological features, particu- larly in specialization of fruits, it appears to be a more advanced group of species. Its extensive geographic range does not neces- sarily indicate antiquity greater than that of Crepis. Prenanthes is less extensive in distributional range than Lac- tuca (fig. 2), but the evidence from comparative morphology shows that Prenanthes is the more primitive. If the distribution of Prenanthes be examined in some detail, it will be found to be consistent with the conception that it was formerly more wide- spread. However, the small areas indicated on the map (fig. 2) in the Canary Islands, Java, South America, and probably Soco- tra Island may be disregarded. Herbarium studies made since this map was printed indicate that none of these species are of this genus. Nevertheless, both comparative morphology and dis- tribution indicate that Prenanthes may be more primitive than Crepis. The Genus Crepis The genus comprises three subgenera, Catonia, Eua-epis, and Bar/{hausia, which are differentiated primarily according to de- The Origin of Crepis and Related Genera 23 gree of specialization of the involucre and secondarily by the amount of specialization of the achenes. In Catonia, as the fruits mature, the involucre merely increases in size and the bracts may become somewhat thicker and indurate but with litde, if any, change in shape. In the other two subgenera, the bracts of fruit- ing heads have a median dorsal keel and become more or less swollen at the base by development of soft spongy tissue. The achenes in Catonia and Eucrepis are not as a rule provided with a true beak, the slender stipe bearing the pappus, but in Barl^^- hausia the beak is always present. There are naturally certain exceptions to these generally reliable distinguishing features, but such exceptional species usually exhibit some marks of re- semblance to other species whose subgeneric affinity is clearly marked. In general it is obvious that there is increasing special- ization of the involucre and the fruits as one passes from Catonia through Eucrepis to Bar\hausia. Along with this increasing specialization there is a definite trend toward reduction in length of life. All the species of Catonia are perennials, while about one-fourth of the Eucrepis species and three-fourths of the Bar\' hausia species are annuals. A somewhat similar situation exists in the three subgenera with reference to the proportion and nature of endemic species. Although some species are so rare or poorly represented in herbaria that the extent of their distribution is wholly conjec- tural, yet the distribution of the majority is sufficiently known to classify them as either restricted to a single geographical unit or of wider extent. In this sense, the proportion of endemics in Catonia is about one-half; in Eucrepis one-third, and in Bar\' hausia less than one-third of the species are thus restricted. Furthermore, on the basis of relative degree of specialization through reduction in size of plant and its parts, increase in num- The Origin of Crepis and Related Genera 25 ber of flower heads, and reduction in length of the hfe cycle, the endemics can be classified roughly as "old" or "young!' The former category obviously includes the more primitive species or relics, and the latter contains the comparatively recent ones. In all three subgenera about one-third of the endemics appear to be old species. Thus the proportions of endemics and of relics in the subgenera is in harmony with the conception that Catonia is the most primitive, Bar^hausia the most recent, and Eucrepis intermediate. The geographical distribution of the subgenera is in general agreement with this conception. As is shown in figure 3, Catonia occupies most of Europe, Asia, and Africa, but does not occur in the New World; Eucrepis covers about the same area in the Old World except that in Africa it is represented only in the north and northeast, but it is also found in western North America and in a few localities in Labrador and Newfoundland; Bar\' hausia is restricted to the Mediterranean region and adjacent areas. It is significant that Eucrepis, which has the widest dis- tribution, contains twice as many species as either of the other subgenera and is more complex in its difiFerentiation into mor- phologically distinct subgroups. It will be shown below that these evolutionary features are associated with greater variability in chromosome numbers and chromosome morphology in Eu- crepis. The phyletic relations between the three subgenera are summarized in figure 4. The somatic chromosome numbers found in the Old World species of Crepis are 6, 8, 10, 12, 14 (15-24), 16, and 40, while the North American species have 22, 33, 44, 55 ( ?), and 88 {}), besides 14 in two representatives of an Old World group. In figure 5 the distribution of these numbers in the subgenera is shown by indicating as exponents the numbers of species known 26 E. B. Babcoc\ to have certain chromosome numbers. From this representation it is obvious that Eucrepis is much more variable in chromosome numbers than the other subgenera. .\lthough 8 is the most prevalent diploid number in the genus, 10 must be considered more primitive, because the most primi- BARKHAUSIA 45 SPECIES PEDITERRANEWM MOSTLY ANNUALS 13 ENDEWIC OR LOCAL EUCREPIS 120 SPECIES 30 ANNUALS 90 PERENNIALS 105 OLD WORLD 15 NO. AHER. 40 ENDEMIC OR LOCAL CATOIMIA 43 SPECIES OLD WORLD PERENNIALS 20 ENDEMIC OR LOCAL Fig. 4. Phyledc relations of the subgenera of Crepis. tive species in the genus have lO chromosomes in their somatic cells, and because no species having lo chromosomes are as gready reduced or specialized as some of the species having 8. It follows, therefore, that all the 8-chromosome species have been derived from lo-chromosome progenitors; and similarly, the two 6-chromosome species were probably derived from 8-chro- mosome ancestors. The evolutionary processes involved in such changes have been discussed elsewhere.** **" All numbers higher than 10 are secondary, and have been derived by various pro- cesses of which interspecific hybridization and doubling of the whole chromosome complement are of most general importance. The Origin of Crepis and Related Genera 27 Throughout the genus there is close correspondence between chromosome numbers and external morphology of the plants. Although the most primitive species have 10 chromosomes, there CREPIS EVOLUTIONARY POSITIONS OF THE SUBGENERA CHROMOSOME NUMBER AND PHYLOGENY BARKHAUSIA 16^ 8 (10-18) 8^' EUCREPIS 44' 55? 88 22^ (15-24) CATOIMIA 12^ 10^ 8® PROGEMIAL\STOCK, 10 Fig, 5. Distribution of chromosome numbers in the subgenera of Crepis. are some fairly primitive 8-chromosome types. In both 10- chromosome and 8-chromosome series there is abundant evi- dence of progressive development from the woody-based peren- nials vsdth large lyrate or entire leaves, few large heads, large florets, and large unspecialized fruits to the short-lived annual types with small leaves, numerous small heads, small florets, and 28 E. B. Babcoc\ very small or highly specialized fruits. The basic, primitive dip- loid number is lo and there are three 5-paired phylogenetic lines, one in each subgenus. Furthermore, the subgenera are not sepa- rated by fixed limits. From the peculiarities of certain species, it is clear that Catonia tends to merge into Eucrepis, and Eucrepis into Bar\hausia. Hence it seems highly probable that the three lo-chromosome lines, one in each subgenus, had their origin in a common nexus. At any rate, the genus must be accepted as a natural unit. The chromosomes of Crepis species are of three distinct types, namely, those with a subterminal spindle-fiber constriction, those with a subterminal constriction and bearing a trabant or satellite attached to the proximal arm, and those with an approximately median constriction. By comparing total length and relative length of the proximal and distal arms, chromosomes of the first general type are subdivided into classes designated A, B, and C, the satellite-bearing chromosome is called D, and the small median-constricted chromosome, E.^ One exception to this gen- eral scheme occurs in a few species of Eucrepis in which the largest chromosome has a median constriction, but there are good reasons for considering this chromosome a modified A type. Every simple diploid species of Crepis has only one pair of satellite-bearing chromosomes. This fact alone often makes it possible to recognize secondary or derived species from the regular presence of more than two D chromosomes in the somatic cells, even without a complete analysis of the whole complement. All the lo-chromosome species of Crepis have one pair each of chromosome types A, B, C, D, E. Just as 10 is the most primi- tive number in the genus, so this combination of five types must be considered the basic genom (haploid complement of a diploid The Origin of Crepis and Related Genera 29 species). All the 8-chromosome species have one pair each o£ types A, B, C, D. The two 6-chromosome species have one pair each of A, B or C, and D. All species thus far analyzed with numbers higher than 10, have been found to have genoms re- sulting either from duplication of certain of the types present in the basic genom or from complete doubling of the whole comple- ment. Thus, in the 12-chromosome species, there may be dupli- cation of A, B, C, D, or E and sometimes one type is absent; whereas the 14-chromosome species always have the basic genom plus extra chromosomes of C, D, or E types. This indicates hybrid origin for both the 12- and the 14-chromosome species. Again, all the i6-chromosome species have only A, B, C, and D types and are polyploids derived from 4-paired ancestors. Analy- sis of distribution of chromosome types in the species with higher numbers has not been completed, but in some of the American species having 22 chromosomes in their somatic cells, the two pairs of D chromosomes are unlike. This strongly indi- cates that these species originated through interspecific hybrid- ization followed by doubling of both specific complements (amphidiploidy) . Similar evidence has been found in two Old World species having 40 chromosomes. Figure 6 illustrates the basic genom of three primitive species, one from each subgenus; and derived genoms with smaller chromosomes and with less and more than 5 are shown. Chromosome number and morphology are taxonomic criteria of great value in the genus Crepis, especially when used in con- nection with other available criteria. It is certain that morpho- logically similar species have similar chromosomes and, con- versely, close similarity in the chromosomes has often proved to be a helpful clue to a closer degree of relationship than was at first surmised from superficial gross morphology. Thus it has 30 E, B. Babcoc\ crocea taraxacoldes a ■ M ■ W ^ vesicaria I M ■ ■ roontana nana fontiana ZOr 011 cirj hypochaerldea argollca senecloides conyzaefolla pannonlca aculeata nv ni)< rfiK aurea leontodontoldes thomsonll siblrlca raulinll alblda Fig. 6. Primitive and derived genoms in Crepis: left, Catonia; middle, Eucrepis; right, Barl^hausia. The Origin of Crepis and Related Genera 31 been possible to work out with greater accuracy the classifica- tion of species into groups within each subgenus, and to ascertain phylogenetic relations. Figure 7 shows the genoms of ten Eu- crepis species which provide a fine illustration of marked speci- fic differentiation without any gross chromosomal alterations. The five species represented in the left column are perennials, and the others are annuals. Among the perennials, gymnopus of Japan, praemorsa of Russia, and incarnata of southern Europe are more similar to one another morphologically than to Eigiana and Reuteriana of Syria. The two latter species are closely re- lated, and very different from palaestina, pulchra, and pterothe- coides, which occur in the same region. In fact, the three species last named have in the past been referred to three different genera, Cymboseris, Phaecasium, and Crepis. Yet comparative morphology reveals the close relationship of palaestina and pul- chra, and establishes that pterothecoides, although most preco- cious and short-lived and very distinct from the other annuals, also belongs in this group. Of the two remaining annuals, granatensis is a very close relative of pulchra and occurs locally within the range of this widely distributed species, and Sto- janovii, an endemic of Bulgaria, is quite as distinct as pterothe- coides, although obviously related to pulchra. This group there- fore includes widely distributed and highly diversified species, yet the chromosomes throughout are very similar in appearance. Other such groups exist not only in Eucrepis but also in both the other subgenera. These groups of related but well-differentiated species, with very similar chromosomes among the species in each group, are significant in several ways. For one thing, they prove that close similarity in the chromosomes is direcdy correlated with close taxonomic relationship in Crepis. They also indicate that the 52 E. B. Babcoc\ reuterlana pterothecoldea elglana granatensia 1)0 ''Ol incamata pulchra praemorsa palaestina -IK* 'm gymnopua stojanovll Fig. 7. Haploid complements of a group of closely related but very distinct species of Eucrepis. The Origin of Crepis and Related Genera 33 genetic evolutionary processes involved in speciation within these groups are processes v^hich are not accompanied by many visible changes in the chromosomes. Factor or gene mutation (genovariation) is one such evolutionary process of general occurrence. Experimental genetic evidence is thus far very lim- ited from this group of plants. It can be stated, however, that Fi hybrids from C. pulchra x C. palaestina are highly fertile and produce F2 progeny exhibiting numerous combinations of char- acters from the original parents. This is just the result that would be expected if the two species differ in numerous genes and yet are homologous with respect to many others. On the contrary, Fi hybrids from C. pulchra x C. Stojanouii are nearly if not quite sterile, which suggests an almost complete lack of homology or some combination of genes in the hybrid which prevents sexual reproduction. Similar evidence from various groups of related species leads to the definite conclusion that gene mutation is an evolutionary process of general occurrence throughout this genus. But it must be viewed as superimposed upon or acting concurrently with those more profound genetic processes involving changes in chromosome number and chromosome morphology. Changes in size of the chromosomes have occurred in Crepis independently of changes in shape. Chromosomes have both increased and decreased in size in the evolution of the genus, but there is a general tendency toward reduction in size of the chromosomes concurrendy with reduction in size of the plant and reduction or specialization of parts. In figure 8 the genoms of another group of Eucrepis are shown. The differences in size of the chromosomes are striking and they illustrate the general tendency toward reduction in the more recent and highly special- ized species. C. oreades and C. Rohertioides are perennials but 34 E. B. Babcocl^ suffrenlana TfV "iV^ cretlca neglecta l)i fullglnosa (nl nicaeensis Ipi tectorum ID caplllarls lln parvlfiora 710 "irii robertioides oreades Fig. 8. Haploid complements of a group of Eucrepis species illustrating reduc- tion in size of the chromosomes concurrent with reduction and specialization in the plant. The Origin of Crepis and Related Genera 35 not of the most primitive category. C. tectorum is one of the most widespread annuals, whereas C. nicaeensis is a biennial (sometimes annual) of rather restricted distribution and prob- ably a more recent species. C. parviflora and C. capillaris are closely related species with similar distributional areas in differ- ent geographic regions. Thus far, C. capillaris and C. fuliginosa are the only 3-paired species found in the genus, but they are not very closely related. It is believed that they were derived from different 4-paired ancestors. The four species at the top of the chart are very closely related. C neglecta is of comparatively wide distribution, whereas cretica and fuliginosa are endemics and Suffieniana is very restricted. The last three are more re- duced or highly specialized than neglecta, and undoubtedly they are comparatively recent in origin. These four species are of special interest because they illustrate the tendency of some of the recent Eucrepis species to merge into Bar]{hausia. This is indicated by their more or less definitely beaked achenes, espe- cially in the three more recent species. Finally, this whole group of species is of unusual interest because of the probability that they have been derived from a 5-paired progenial line (see fig. 10). By application of this method of comparing chromosome rela- tions with the evidence from external morphology, the phyletic relations of all the Crepis species thus far studied cytologically have been v/orked out. These are represented in figures 9-1 1. The distribution of these various groups of related species may now be considered. In Catonia (fig. 9) only three lo-chromosome species have thus far been obtained, but two of them, sibirica and pontana, are the most primitive types in the whole genus. The first is dis- tributed from central Europe to Mongolia and southward to 36 E. B. Babcoc\ Transcaucasia, and the second occurs at higher altitudes from France to Rumania. The third, aurea, is much less primitive and occurs in various forms in the mountains of southern Europe CATOIMIA PHYLOGENY AND CHROMOSOME NUMBER AUREA 10 HOOKERIANA 8 PONTANA 10 SIBIRICA to VISCIDULA 12 PALUDOSA 12 CROCEA 16 BUNGEI 8 POLYTRICHA 16 CHRYSANTHA 8 6UREJENSIS 8 C0NY2AEF0LIA 8 KASHMIRICA 12 '^^ PRE-CATONIA Fig. 9. Phylogeny and chromosome number in Catonia. and northern Asia Minor. The 8-chromosome species fall into tv^o groups, one with four, the other with six species. Of the first group, blattarioides and alpestris are montane species of southern Europe; the latter also occurs in Asia Minor, which supplements the evidence from gross and chromosome morphol- ogy that it had a common origin with C. hypochaeridea of The Origin of Crepis and Related Genera 37 South Africa. Furthermore, there are central African species which are certainly related to hypochaeridea, and one of these, C. Newii, has just been found to have a genom very similar to that of C. hypochaeridea. The fourth species, Hookeriana, a Moroccan plant of the Grand Adas Mountains, is less closely related to the other three. Of the second group, conyzaefolia is the most primitive species and is distributed from southern Europe to central Asia; the other five, all of eastern Asia, are certainly less primitive. Although the 12-chromosome Catonia species must be regarded as derived from species with some more primitive number or numbers, yet they are fairly primitive types. Their distributional relations are interesting, for paludosa is second only to sibirica in distributional range (western Europe to western Siberia), while its close relative, viscidula, is restricted to the northern Balkan countries. C. J^cishmirica is not closely related to the other two and it must have sprung from a different ancestor or ancestors; as its name indicates, it occurs in the western Himalaya Mountains. From this survey of Catonia, it would appear that the center of origin and distribution of these species must have been in south-central or southwestern Asia. In Eticrepis (fig. 10) the lo-chromosome species comprise three groups, of which the first contains two alpine perennials, C. Raulinii, an endemic of Crete, and C. bithynica, which occurs in northern Asia Minor and the western Balkan Peninsula. The second group comprises C. multicaulis, a much less primitive perennial which is distributed from the western Himalayas to northern Scandinavia, together with sancta and bifida, two an- nuals which are closely related to each other and whose com- bined range is from southwestern Europe to southwestern Asia. The third group contains tingitana, a somewhat primitive perennial of northern Morocco and southwestern Spain which 38 E. B. Babcoc\ EUCREPIS-PHYLCX^EAJY AI\JD CHROMOSOME (\JUMBER ATRISAR8A 88? MOWTICOLA 559 \ BARBICERA 88? SCOPULO«U« 44?\ LaCUHIWATA 33.44.55? r OCCIDENTALIS 22.44 ^ .GRACILIS 22.55? ANOCRSONII 22 CLAUCA 22 RUNCINATA 22 PTDWTHECOIDK 8^ PULCHRA 8, GRANATENSIS 8. PALAESTINA 8 WLTIFLORA 8 0I09C0RI0IS 8^ STOJANCVIl 8 TUBAEFORMS 8^ REUTERIANA8 EICIANA8 AMPHIDIPLOID HYBRIDS 22 LEONTOOOWTODES 10 PRE-EUCREPIS Fig. 10. Phvlogeny and chromosome number in UrPRENlANA 8 NEGLECTA 6 8 PARVIFLORA 8 CAPILLARIS 6 TECTORUM 8 NICAEENSIS 8 ROeERTlOlDES 8 OREAOES 8 CILIATA 40 BIENNIS 40 HIEROSO- LYnrTANA 12 WLLEftn- OIDES 12 nUNOIERI 12 P^ONTANA 12 PYGHAEA 12 LYRATA 12 MOLLIS 12 FLEXUOSA 14 .ELEGANS 14 NANA 14 Eucrepis. The Origin of Crepis and Related Genera 39 has some afiinity with certain African species of Catonia; sube- rostris, an endemic annual of Algeria; and leontodontoides, a less primitive perennial of Italy and southern France. The 8-chromosome Eucrepis species comprise several distina groups, x^vo of which have already been mentioned in connec- tion with chromosome morphology. One of these, the oreades- Suffieniana series, may have originated from a lo-chromosome ancestral stock because there is a certain amount of affinity for the 40-chromosome species, biennis and ciliata, which are oao- ploids with 5 as base number. The evidence on distribution is in harmony with this idea. C. biennis is distributed from western Asia to western Europe and southward to southern Russia and the middle Balkans; C. ciliata occurs in the Caucasus region. C. oreades and Robertioides 2J^ woody-based alpine perennials, the former from middle Asia, the latter from Syria. C. nicaeensis occurs sporadically from the Balkan Peninsula to the Pyrenees; lector iim is distributed throughout most of Europe and the greater part of Siberia; capillaris is found in southern and mid- dle Europe and eastward to southern Ukraine and Crimea; parviflora occurs from the eastern Balkans throughout Asia Minor to Egypt and the Caucasus; neglecta and its close allies are distributed from southern France to Asia Minor, Greece, and Crete. Likewise, the gymnopus-pterothecoides group, taken as a whole, ranges from Japan to Spain and from middle Russia to Palestine, with five of the ten species occurring in Syria. The other two groups of 8-chromosome Eucrepis species are foimd in certain parts of the same general region. C. pannomca, lacera, and chondrilloides are very closely related species, and their distribution is in agreement ^^-ith the morphological evi- dence of relative age. C. pannonica is certainly the most primi- tive and is distributed from Hungary to southern Russia, lacera 40 E. B. Babcoc\ is intermediate and occurs in the southern half of Italy, and chondrilloides is most recent and is found only in a limited area extending from Trieste into southern Austria and northern Dal- matia. (Associated with the foregoing are the two polyploid species, C. incana and C. taygetica, of southern Greece.) In the argolica-multiflora group, however, the relation between distri- bution and the morphological evidence of relative age is re- versed. Both argolica and tubaeformis are the least specialized, particularly in achenial features, and each is known only from a single small area in Greece, whereas dioscoridis is abundant in parts of southern Greece, and multi flora, the most reduced species, occurs on the islands and shores of the Aegean Sea in eastern Greece and western Asia Minor. Hence these two groups of species illustrate very well the difficulty which was recognized by Willis"" in applying the "age and area" hypothesis to very small groups of species — the pannonica group is in agreement with the hypothesis, but the argolica group is not. This, how- ever, is merely an incidental observation. The main point is that both groups form part of the Eucrepis assemblage and occur in the same distributional area. Definitely associated with the foregoing species, although not closely related to any of them nor to each other, are two remark- able species, C. patula, an endemic of northern Algeria, and C. tenuifolia, which is distributed from the western Himalayas to Mongolia. The former is apparendy a relic which combines a number of primitive features with extreme reduction of the pappus, a combination of characters that is unique in the genus. The latter is believed to have originated as an interspecific hybrid between two 8-chromosome species followed by amphi- diploidy and subsequent alterations in chromosome number.* Turning now to other Eucrepis species with derived chromo- The Origin of Crepis and Related Genera 41 some numbers, we find three groups, namely, those with 12 chromosomes, those with 14 chromosomes, and those having 22 chromosomes as the base number. As in the groups discussed above, the position of each species in the series (fig. 10) is deter- mined on the basis of comparative morphology, considered with reference to geographic distribution which, in the 12-chromo- some group, is as follows: mollis extends from western Europe to middle Asia; lyrata is found in western Siberia; pygmaea occurs only in the European Alps; montana, only in Greece; mungieri, only in Crete; Willemetioides, in northeastern Persia; and hierosolymitana, in Palestine, Syria and Cyprus. At present only three species with 14 chromosomes are known, yet they comprise an even more widespread group than the pre- ceding one. C nana, in fact, is the most widespread species in the genus, being distributed from central Asia to eastern Labrador, at high elevations in the Rocky Mountains, and in the higher ranges of Utah, Nevada, and the Pacific Coast states. C. elegans occurs only in Alaska and the Rocky Mountains, and C. flexuosa is distributed from Outer Mongolia and Semipalatinsk to the Altai and Himalaya mountains. The remaining Eucrepis species occur only in North America. As is indicated in the chart (fig. 10), there is good evidence to support the hypothesis that all these species with 22 as the base number originated through hybridization between 14-chromo- some and 8-chromosome species followed by amphidiploidy. The region of origin is uncertain, but it seems most probable that it was in southern Alaska or an adjacent area to the south, and that the progenial species did not survive the subsequent glacial epoch in which their more vigorous offspring were able to mi- grate southward. At any rate, the distribution of these exclu- sively American species indicates a radiation from southwestern 42 E. B. BabcocJ{^ Canada and northwestern United States to the east, southeast, and south. Furthermore, the existence of 14-chromosome species in North America previous to the last or Wisconsin glacial epoch, is proved by Fernald's discovery" of C. nana in an un- glaciated island on the Gaspe Peninsula in eastern Canada. Although no 8-chromosome species are know^n to occur in Alaska, one such species (C. burejensis, which from its morphol- ogy might have been an ancestor of the occidentalis group of American species) is found in the Kurile Islands, and it might easily have been distributed farther east at an earlier period. It seems very probable, therefore, that these American species are descendants of 8- and 14-chromosome species which migrated into northeastern North America from Asia. Reviewing the evidence on distribution of these various groups of related species in the subgenus Eucrepis, we find it consistently pointing to south-central Asia as the region of common origin. In BarJ^hausia (fig. 11) the lo-chromosome species are cer- tainly more primitive as a group than any of the 8-chromosome species. They may be considered in two subgroups, namely, those more or less resembling C. albida, and those clustering around C. foetida. C. albida is a polymorphous sufFrutescent perennial species of southwestern Europe and the Grand Atlas Mountains in Morocco. C. alpina is a primitive type of annual which is distributed from the Caucasus to western Asia Minor, and C. syriaca is a closely related species occurring more to the south. C. rubra is a less primitive congener of the southern Bal- kans and Italy. It is worth noting, also, that there are two obvi- ously related species in Abyssinia which have not yet been studied cytologically. There is thus sufficient evidence of connection between albida and the region of common origin for Eucrepis and Catonia. The species in the foetida group are more closely The Origin of Crepis and Related Genera 43 connected in distribution. C. foetida (including several sub- species) is distributed from southwestern Persia and the Caspian region through Asia Minor to western Europe; commutata BARKHAUSIA PHYLOGENY AND CHROMOSOME NUMBER MYRIOCEPHALA 8 SENECIOIOES 8 NIGRICANS 8 THOMSONII 10 HACKELIl 16 TARAXACirOLIA 8y ' TARAXACOIDES 16 LYBICA 8^ \ ? .VESICARIA 8.16 / BURENIANA 8^ | ERITREENSIS 10 HYEMALIS 8,^ \ ^ ' ' ." FOETIDA 10 clausonis 8^^ \ /^^bursifolia 8 'commutata 10 TRIASII 8. \ SETOSA 8. / ^BELLIDIFOLIA 8 HIERACIOIDES 8 DIVARICATA 8 CANARIENSIS 8 FONT I AN A 8 RUBRA 10 ALPINA 10 SYRIACA 10-18 ALBIDA 10 PRE-BARKHAUSIA Fig. 1 1 . Phylogeny and chromosome number in Bar\hausia. occurs in Asia Minor and the southern Balkans; eritreensis is known only from Eritrea; Thomsonii is indigenous in north- western India; and the definitely affiliated 8-chromosome species, C. Bureniana, is a native of Persia, Baluchistan, and Afghanistan. All the 8-chromosome Bar\hausia species (fig. ii) are of the Mediterranean littoral, several being island endemics, including 44 E. B. Babcoc\ two from Madeira and one from the Canary Islands. Several are of rather wide distribution, while the tetraploid species con- nected with them are more restricted and undoubtedly more recent. It seems hardly necessary to go into further detail con- cerning their distribution, as they are certainly connected with Asia Minor and the Caucasus region. In general, the evidence on distribution of related groups of Crepis species is wholly consistent with the assumption that the center of origin and distribution for the genus is in south-central Asia. It follows, then, that there was subsequent migration to the west, with the development of secondary regions of specia- tion in Asia Minor, the Balkan Peninsula, the European Alps, Abyssinia, and the Mediterranean littoral, and to the northeast, with secondary centers in the Altai region, northeastern Asia, and northwestern America. Although there are still many in- teresting problems to be solved with reference to specific rela- tions and the nature of endemics,— concerning whether they are epibiotic or of recent origin,— the most fundamental problems may be considered as solved with the establishment of the center of origin in southwestern Asia and with the conclusion that 5 and 4 are the basic haploid chromosome numbers in Crepis and that 5 is the more primitive number. Origin of the Crepidinae What bearing have these conclusions on the problem of origin in the Crepidinae? It is impracticable to give a detailed study of distribution in Hieracium, Lactuca, and Prenanthes, but it can- not be without significance that in each of these genera south- central Asia is centrally situated with reference to generic limits of distribution (fig. i). Furthermore, the species of that unique smaller group, the Glomeratae, which show affinity for both The Origin of Crefns and Related Genera 45 Crepis and Prenanthes, are confined to the Himalayan region. In addition, the most primitive species which has been excluded from Crepis, namely, Paleya oligocephala, occurs in this region only. The existence of these species in the Himalayas, however, need not be stressed, inasmuch as the origin of the Crepidinae may have antedated the upheaval of that great mountain system. This evidence on distribution, however, is consistent with the concept that Crepis and its closest allies originated in south- central Asia. This leads naturally to the question of whether any genetical evolutionary processes can be suggested to account for speciation and differentiation within these genera. It has been intimated (p. 19) that our present knowledge of phylogeny in Crepis makes possible a tentative hypothesis of this kind. Before pre- senting this hypothesis, however, it is necessary to consider briefly two fundamental genetical processes which appear to have been operating in the evolution of Crepis. The first of these is reduction of chromosome number from 10 to 8. If monophyle- sis for the genus be assumed, it is necessary also to assume that such a process actually operated during the early development of the genus. Although there is no experimental evidence that new species can originate in this way, yet the combined evidence from comparative morphology, geographic distribution, and cytology in Crepis both necessitates and supports this assump- tion. It will be recalled that, although 10 is the most primitive number in the genus, yet 8 is found in the larger number of species and there are some rather primitive types among the 8-chromosome species. Furthermore, in certain groups of closely related species two different chromosome numbers occur. Thus the two 6-chromosome Eucrepis species are obviously related to species with 8 chromosomes, and in BarJ^hausia a lo-chromo- 46 E. B. Babcoc\ some group contains one 8-chromosome species. It has been pointed out by Babcock and Cameron'' that the most likely process by which 8-chromosome species might arise from 10- chromosome ancestors would be by reciprocal translocations between nonhomologous chromosomes, followed by meiotic ir- regularities leading to disappearance of one pair of chromo- somes. The disappearance of a pair of chromosomes, however, does not necessarily imply actual loss of any essential genes. Reduplications of small sections of chromosomes are known to occur and these might compensate for the loss of fragments lack- ing a spindle-fiber attachment. There has come to light in recent years abundant evidence that reciprocal translocations between nonhomologous chromosomes occur in nature and that they may have a profound effect on genetic relationships. A mecha- nism is known to exist, therefore, which makes possible the ori- gin of 8-chromosome species from lo-chromosome progenitors. The other fundamental process involved is that of interspe- cific hybridization and amphidiploidy. That this phenomenon may be of great importance in the evolution of the higher plants has been amply demonstrated by the production, through arti- ficial hybridization, of highly fertile and constant new forms which might be expected to maintain themselves under natural conditions. The classical example to be investigated cytologically is that of Nicotiana digulta of Clausen and Goodspeed,"' " who obtained it by crossing N. tahacum (n = 24) with A/", glutinosa (n = 12) . N, digltita has 36 pairs of chromosomes, is fairly fertile, and reproduces itself constantly. But when it was crossed with the parental species, the investigators found that the hybrid progeny were more or less fertile and displayed a tendency to revert to the types of the parental species. Thus, although N. digluta The Origin of Crepis and Related Genera 47 behaves like a distinct species, it probably could not survive under natural conditions unless isolated from the parental species. This was the first experimental verification of the hypo- thesis proposed by Winge,^'' that interspecific hybridization fol- lowed by doubling of the chromosome number might account for the arithmetical series of chromosome numbers which have been shown to exist in various genera of plants. Since then, numerous amphidiploid hybrids between distinct species have been produced experimentally and many of them behave like "good" species although many others are more or less sterile and irregular in reproduction. One of the most convincing examples is that of Raphanobrassica, which was produced by Karpe- chenko^" by crossing Rap h anus sativus with Brassica oleracea. The amphidiploid race obtained is constant and fertile under experimental conditions, and can be crossed with the parental species only with great difficulty; hence it must be considered a potential new species. Furthermore, numerous constant inter- specific hybrids have been discovered in nature, and these pre- sumably arose through amphidiploidy.^'' ^' *° In Crepis, cytological evidence has been found to indicate the important role played by interspecific hybridization and amphi- diploidy in the evolution of this genus (pp. 29, 38, 41). It may be objected that the experimental evidence from Crepis does not support this idea. It happens, however, that the three amphidi- ploids thus far obtained in Crepis resulted from crosses between nearly related species, so that there was sufficient homology between the chromosomes of the parents to cause much irregu- larity in distribution of the chromosomes to the gametes, with resulting sterility."^ The fact that amphidiploids were produced from hybrids between such species may be looked upon as sup- porting the hypothesis here proposed. 48 E. B. Babcoc\ If it be admitted that two such genetical processes as reduction in chromosome number from 10 to 8, and interspecific hybridiza- tion followed by amphidiploidy, may have occurred in the early evolution of this group of plants, a working hypothesis of the origin of the Crepidinae is readily forthcoming. This hypothesis is illustrated diagrammatically in figure 12, and may be stated as follows : (i) The original progenial stock consisted of one or more 10- chromosome species. (2) New species with 8 chromosomes originated through some such process as reciprocal translocation with subsequent elimina- tion of parts of chromosomes, leading eventually to the disap- pearance of one pair of chromosomes. (3) The genus Crepis arose from the complex of species thus created. (4) The 8- and 9-paired species of closely related genera origi- nated through interspecific hybridization between 5-paired and 4-paired species (or between 4-paired species) followed by am- phidiploidy. (5) Speciation and evolution, within Crepis at least, have in- volved the same fundamental processes together with autopoly- ploidy, and, concurrendy with these processes, genie mutation has played a continual role. The propositions just stated are submitted as the most logical conception of the early course of evolution in the Crepidinae on the basis of the evidence now available. Future research on Crepis and especially on the genera closest to Crepis may result in modification or extension of the hypothesis, but it appears that the general nature of the processes involved in the evolution of this group of plants has been disclosed. If this working hy- pothesis ultimately becomes an established theory, it will necessi- The Origin of Crepis and Related Genera 49 tate the recognition that gross chromosomal alterations have played a more basic role, in the evolution of some o£ the higher plants, at least, than has genie mutation. CREPIS 10.8 HIERACIUM 18 LACTUCA 18.16 PRENANTHES 18 INTERSPECIFIC HYBRIDIZATION AND AMPHIDIPLOIDY (5+4X2=18) 5-PAIRED AND 4-PAIRED SPECIES ORIGIN OF 8-CHROMOSOME SPECIES 10 -CHROMOSOME PROGENIAL STOCK Fig. 12. Origin of the Crepidinae. The diagram illustrates the hypothesis that the whole subtribe originated from a 10 -chromosome stock through: (i) origin of 8-chromosome species by loss of one pair of chromosomes; (2) development of Crepis from this group of 5 -paired and 4 -paired species; (3) origin of the other genera through interspecific hybridization between some of those species followed by amphidiploidy. A detailed discussion of various pertinent questions which w^ill arise in the minds of critical readers must be postponed until a future contribution. Let it suffice for the present to state 5© E. B. Babcocl^ that, in the mind of the present writer, all such queries revolve around problems which are of minor importance as compared with the genetic evolutionary processes on which the hypothesis is based. One such question may be mentioned briefly by way of illustration. An interesting issue is raised because, in those few species in the genera related to Crepis on which data are available, there appears to be only one pair of satellite-bearing chromosomes. This may at first appear inconsistent with the fact that certain species of Crepis are believed to have arisen through interspecific hybridization and amphidiploidy because they have two or more pairs of unlike D chromosomes. This criticism can be met by reference to the phenomenon known as amphiplasty.^" It has been found that, in hybrids between certain species, the satellite contributed by one of the parental species is permanendy lost, whereas in other interspecific hybrids the satellites from both parents are preserved. These well-established facts are sufficient to account for the apparent inconsistency. The genus Crepis is closely related to the other genera of the Crepidinae, if the evidence from comparative morphology can be trusted, and probably it can be, especially when it leads to the same conclusion as other criteria. The genus Crepis, in spite of the great variability in chromosome numbers which it exhibits, is a natural group with a common center of origin and a single, most primitive chromosome number. The center of origin of Crepis coincides with the center of origin for its closest allies. Only on cytological grounds is there an apparent separation of Crepis from other genera in this group. A hypothesis has been proposed which seems to account satisfactorily for these chro- mosome relations and to be consistent with other evidence of natural relationship in the Crepidinae. BIBLIOGRAPHY ^ Babcock, E. B. 193 1. Cytogenetics and the species-concept. Am. Nat., 45: 1-18. ' Babcock, E.B., and Cameron, D. R. 1934. Chromosomes and phylogeny in Crepis. II. The relationships of one hundred eight species. Univ. Calif. Publ. Agr. Sci., 6: 287-324. ' Babcock, E. B., and Lesley, M. M. 1926. Chromosome number and individuality in the genus Crepis. II. The chromosomes and taxonomic relationships. Univ. Calif. Publ. Agr. Sci., 2:315-341. * Babcock, E. B., and Navashin, M. 1930. The genus Crepis. Bibliog. Genet., 6: 1-90. ^ Bentham, G., and Hooker, J. D. 1873-76. Compositae, in Genera Plantarum, 2(1): 163-533. " Cassini, a. H. G. \%i'] . Aetheorrhiza, in Diet. Sci. Nat., 48: 425. ' Cassini, A. H. G. 1822. Ixeris, in Diet. Sci. Nat., 24: 49. ® Cassini, A. H. G. 1 83 1. Youngia, in Ann. Sci. Nat., 23: 88. * Cassini, A. H. G. 1826. Sur I'ordre des Synantherees, in Opuscules Phytologiques, 1 : 378-426. ^° Clausen, R. E. 1928. Interspecific hybridization in Nicotiana. VII. The cytology of hybrids of the synthetic species, digluta, with its parents, glutinosa and taba- cum. Univ. Calif. Publ. Bot., 11: 177-211. " Clausen, R. E., and Goodspeed, T. H. 1925. Interspecific hybridization in Nicotiana. II. A tetraploid glutinosa- tabacum hybrid, an experimental verification of Winge's hypothesis. Genetics, 10: 278-284. ^ Fern.\ld, M. L. 1929. Some relationships of the floras of the northern hemisphere. Proc. Internat. Cong. Plant Sci., 2: 1 447-1 507. " Good, R. D'O. 1930. The geography of the genus Coriaria. New Phjtologist, 29: 170-198. " Gray, A. 1857. Diagnostic characters of new species of phaenogamous plants, etc. Mem. Am. Acad. Arts and Sci., 6: 396-398. [51] 52 E. B. Babcoc\ " Heitz, E. 1926. Der Nachweis der Chromosomen. Verglcichcndc Studien iibcr ihrc Zahl, Grosse und Form im Pflanzenreich I. Zeitschr. f. Bot., 18: 625- 681. " Hoffmann, O. 1891. Compositae, in Engler and Prand, Die natiirlichen Pflanzenfamilien, IV(5): 87-402. " HoLLiNGSHEAD, L., and Babcock, E. B. 1930. Chromosomes and phylogeny in Crepis. Univ. Calif. Publ. Agr. Sci., 6: 1-53. " ISHIKAWA, M. 191 6. A list of the number of chromosomes. Bot. Mag. Tokyo, 30: 404-448. " ISHIKAWA, M. 1 92 1. On the chromosomes of Lactuca (a preliminary note). Bot. Mag. Tokyo, 35: 153-158. ^ Karpechenko, G. D. 1927. Polyploid hybrids of Raphanus sativus L. X Brassica oleracea L. Zeitsch. ind. Abs. Vererb., 48: 1-85. " Mann, M. C. 1925. Chromosome number and individuality in the genus Crepis. I. A com- parative study of the chromosomes of nineteen species. Univ. Calif. Publ. Agr. Sci., 2: 297-314. ^ Marchal, E. 1920. Recherches sur les variations numeriques des chromosomes dans la serie vegetale. Mem. de I'Acad. Roy. de Belgique, ser. 2, 4: 1-108. ^ Nakai. T. 1920. Notulae ad Plantas Japoniae et Koreae, xxiii. Bot. Mag., 34: 148-158. ^* Navashin, M. 1925. Morphologische Kernstudien der Crepis-Arten in Bezug auf die Artbildung. Zeitsch. Zellforsch. mikr. Anat., 2: 98-1 11. ^ Navashin, M. 1928. "Amphiplastie" — eine neue karyologische Erschemung. Zeits. indukt. Abst. Vererb. Suppl., 2: 1148-1152. ^ OSAWA, J. 1 913. Studies on the cytology of some species of Taraxacum. Arch. f. Zell- forsch., 10: 450-469. ^ OSTENFELD, C. H. 1 92 1. Some experiments on the origin of new forms in the genus Hieracium, subgenus Archieracium. Jour. Genetics, 11: 1 17-122. The Origin of Crepis and Related Genera 53 ^ Poddubnaja-Arnoldi, W. a. 1933. Geschlechte und ungeschlechte Fortpflanzung bei einigcn Chondrilla- Arten, Archiv wissensch. Bot., 19: 46-86. *° Poole, C. E 1932. The interspecific hybrid, Crepis rubra XC. foetida, and some of its derivatives, II. Two selfed generations from an amphidiploid hybrid. Univ. Calif. Publ. Agr. Sci., 6: 231-255. ^ Poole, C. E 1933. Constant species hybrids. Amer. Nat., 67: 188-190. '^ Robinson, B. L., and Fernald, M. L. 1908. Compositae, in Gray's New Manual of Flowering Plants. '^ Rosenberg, O. 1909^. Zur Kenntniss von der Tetrandenteilungen der Compositen. Svensk. Bot. Tidskr,, 3 : 64-77. ^ Rosenberg, O. 1909-?'. Ueber die Chromosomenzahlen bei Taraxacum und Rosa. Svensk. Bot. Tidskr., 3: 150-162. " Rosenberg, O. i9i2.Ueber die Apogamie bei Chrondrilla juncea. Svensk. Bot. Tidskr., 6: 915-919. "^ Rosenberg, O. 1917. Die Reductionsteilung und ihre Degeneration in Hieracium. Svensk. Bot. Tidskr., 11: 145-206. *" Tahara, M. 1910. Ueber die Zahl der Chromosomen von Crepis japonica. Bot. Mag. Tokyo, 24: 23-27. "^ Tahara, M., and Ishikawa, M. 191 1. The number of chromosomes of Crepis lanceolata var. platyphyllum. Bot. Mag. Tokyo, 25: 119-121. ^ Willis, J. C. 1922. Age and Area. A study in geographic distribution and origin of species. Cambridge Univ. Press. »" Winge, O. 1 91 7. The Chromosomes. Their numbers and general significance. Comp. Rend. Lab. Carlsberg, 13: 131-275. *° Winge, O. 1932. On the origin of constant species hybrids. Svensk. Bot. Tidskr., 26: 107-122. " Zahn, K. H. 1921-22. Hieracium. in Engler, A. Das Pflanzenreich, IV, 280; Heft 75, 79. The Succession and Distribution of Cenozoic Floras Around the Northern Pacific Basin By Ralph W. Chaney THE VEGETATION o£ wcstem North America in the Ceno- zoic era has been made known by the studies of Lesque- reux, Newberry, and Dawson, followed by a more modern group, among whom Knowlton was the outstanding contributor. Until recently, however, the floras of this most re- cent major unit of geologic time, occupying perhaps the latest hundred million years of the earth's history, have not been criti- cally studied on the continent of Asia. With the investigations of Kryshtofovich and others, it becomes possible to compare Tertiary floras on both sides of the Pacific Basin, and to deter- mine not only their sequence in geologic time, but also their probable routes of migration and their modern geographic situa- tion. The purpose of this paper is to summarize the evidence of the succession and distribution of the Cenozoic floras around the northern part of the Pacific Basin, and to oudine the general principles which guide the interpretation of this evidence. The underlying basis for the interpretation of the distribution of Cenozoic floras, both in time and in space, is the assumption of a trend from mild, humid climate at the early part of the era to cool, semiarid climate toward its close. This does not imply that there were not minor changes in reverse direction in the Cenozoic, nor that the change toward cooling and drying was of uniform rate. A series of fossil floras resulting from such a climatic trend is an example of what Clements'* has termed a * Superior figures refer to items In the bibliography at the end of this essay. [55] 56 R. W. Chaney clisere, that is, a sequence of climaxes with the corresponding cHmates, such as may be seen today from the foot to the summit of high peaks or from tropical to polar lands. Not only does the fossil plant record of the Cenozoic indicate such a cliseral sequence, but also the evidence of the animals, both invertebrate and vertebrate, and of the sedimentary rocks in which they are enclosed, is in accord with the assumption of this climatic trend through the Cenozoic. It has been discussed on the basis of the plant record in several publications by the writer and his asso- ciates,^''*'*' °' "■'■ ^ and will not be reviewed here in detail. Floras of older Tertiary age, including those referred to the Eocene epoch and some of those referred to the Oligocene,* have been critically studied in western North America from the west flank of the Sierra Nevada in central California, from the western border of the Cascade Range in central Oregon, from the John Day Basin of eastern Oregon, from the Puget Sound area and from the eastern slope of the Cascades in Washington, and from the Kenai Peninsula and adjacent Alaska.f These floras as represented in California and Oregon are made up dominandy of types of plants which have their modern range in the low latitudes of both hemispheres. Silicified stems make up a conspicuous part of the fossil record in the Auriferous * Confusion concerning the boundary between these two series of the Tertiary in western North America often makes impossible an exact age designation. The terms "Eocene" and "older Tertiary" as applied to American floras will be used synonymously in this paper, since most of the floras discussed are considered by the writer to be of Eocene age, tin addition to the papers listed in the bibliography, the following studies have advanced to the point of completed manuscript or of matured conclusions: C. A. HoUick, on the Tertiary floras of Alaska; R. S. LaMotte and others, on the floras of Washington; R. W. Chaney, on the Clarno flora of eastern Oregon; H. D. MacGinitie, on the older Auriferous Gravels floras of California. Cenozoic Floras Around the Northern Pacific 57 Gravels of the Sierra foothills, but they are rare or absent at the other localities, as are seeds and flowers. Leaves are the struc- tures most commonly preserved, occurring in the tuffaceous shales as impressions which show all the details of margin, nervation, and texture. The propriety of referring these leaf impressions to modern genera has been questioned on many occasions by botanists who feel that generic characters are neither constant nor recognizable in foliar structures. The paleobotanist, however, usually dependent on the evidence of leaves, has been forced to find in them criteria for recognition which are un- known to most botanists; he sees in the arrangement of the nerves, particularly near the margin, diagnostic characters by which a relationship to modern genera can often be estabUshed; with certain specimens he may be forced to indicate only a family relationship by the use of such names as Laurophyllum, Legu- minosites, or Menispermites. Familiarity with the foliage of modern plants in forest and herbarium makes possible his refer- ence of almost all leaf fossils to genera which indicate their sys- tematic position, leaving only a small residue in the form genus Phyllites. It is significant to note that botanists of wide experi- ence in field and herbarium, such as Abrams, Jepson, and Mason for temperate floras, and Gleason, Greenman, Merrill, Standley, and Wilson for those of lower latitudes, are able to recognize readily the leaves of modern woody plants, and to suggest the generic or family affinities of fossil leaves as well. Acknowledg- ment is due to these and many other botanists for the valuable assistance which they have given to the writer and his associates. As compared with the species of later Tertiary floras, Eocene leaves are large and thick, with the margins more commonly entire and with a consequendy greater development of campto- drome nervation. The families which are most characteristic, 58 R. W. Chancy together with the genera included in them, are given in the following list : Palmaceae Sabalites Liliaceae Smilax Juglandaceae Engelhardtia Fagaceae Quercus Moraceae Ficus Pourouma Aristolochiaceae Aristolochia Trochodendraceae Trochodendroides (Tetra- centron) Menispermaceae Abuta Cissampelos Menispermites Magnoliaceae Drimys Magnolia Anonaceae Anona Monimiaceae Siparuna Lauraceae Acrodiclidium Cinnamomum Cryptocarya Lindera Nectandra Ocotea Persea Hamamelidaceae Liquidambar Platanaceae Platanus Leguminosae Inga Lonchocarpus Mimosites Euphorbiaceae Acalypha Alchornea Aleurites Aporosa Euphorbiophyllum Mallotus Sapium Anacardiaceae Astronium Sapindaceae Ailophylus Cupania Sabiaceae Meliosma Rhamnaceae Rhamnidium Sterculiaceae Pterospermum Sterculia Dilleniaceae Davilla Tetracera Flacourtiaceae Xylosma Myrtaceae Calyptranthes Cenozoic Floras Around the Northern Pacific 59 Araliaceae Boraginaceae Aralia Cordia Sapotaceae Ehretia Chrysophyllum Verbenaceae Lucuma Petrea Ebenaceae Bignoniaceae Diospyros Callichlamys Styracaceae Caprifoliaceae Styrax Viburnum Apocynaceae Apocynophyllum Ferns and gymnosperms have not been included because o£ their comparative rarity in older Tertiary floras of western America. The plants listed indicate a climate characterized by an annual rainfall of not less than 80 inches and by a uniform temperature free from frost — a climate (to mention only those regions v^ith v^^hich the writer is familiar) essentially like that on the lower slopes of Vera Cruz, Mexico, up to an elevation of 5500 feet on the Pacific slope of Guatemala, in the savannas near sea level on the Pacific slope of Panama, and at an elevation of about 5000 feet in the Andes of Venezuela near Merida. There is evidence that the Goshen flora advanced northward from low latitudes to its Eocene or Oligocene position in west-central Oregon." A similar movement has been described for the somewhat older Wilcox flora of the southeastern United States.^" Knowledge of the Cretaceous floras of western America is so scant as to give little information about the pre-Eocene distribution of the fami- lies and genera involved, although it has been pointed out that some of the genera and most of the families were widely dis- tributed." If this movement from low latitudes can be demon- strated, it will represent the most extensive northward migration 6o R' W. Chaney recorded by plants in the Cenozoic era, and the only migration of importance in that direction until the Pleistocene or Recent, when the coniferous forest was shifted from southern to central and northern California."'" The study of the Eocene floras of Washington has not yet ad- vanced to a point where a list of the families and genera repre- sented can be given. Although numerous species represented in the states to the south are included, indicating that the north- ward movement above mentioned extended into this state, the leaves are for the most part characterized by smaller size, thinner texture, and more commonly serrate margins. It seems clear that they represent a less tropical assemblage than do those of Oregon and California and that their modern equivalents will be found in temperate rain forests of higher latitudes. The same conclu- sion may be reached concerning the Eocene floras from the Rocky Mountain province to the east. As now described, these floras include a large list of temperate genera such as Acer, Alnus, Betula, Cornus, Fraxinus, Populus, and Ulmus, along with subtropical types. Although many of the generic references must be revised before the lists of names can be depended upon as an indication of the nature of these floras, a survey of the average size, margins, and texture of the leaves as figured indi- cates that they represent a more temperate assemblage than those along the Pacific Coast at the same latitudes. This may be ex- plained as resulting from the more continental type of climate in the Rocky Mountain province as compared to that of the Pacific border states. The Kenai flora of Alaska, considered by HoUick and others to be of Eocene age, includes such genera as Diospyros, Ficus, and Magnolia; but the most conspicuous and widespread ele- ment of this flora comprises such genera as Acer, Alnus, Betula, Cenozoic Floras Around the Northern Pacific 6i Carpinus, Castanea, Cercidiphyllum, Corylus, Fagus, Ginkgo, Juglans, Liquidambar, Myrica (Comptonia), Platanus, Popidus, Prtinus, Quercus, Salix, Sequoia, Taxodium, Tilia, Trapa, and Ulmtis. The abundance of Sequoia emphasizes the temperate aspect of this flora, which with sUght modifications has been recorded from several other Eocene locaHties in the Arctic in- cluding Greenland and Spitzbergen. A flora of older Tertiary age, collected by Mason and others" on the shore of the Bering Sea south of Nome and on St. Lawrence Island, is made up of Alnus, Platanus, Sequoia, and other temperate genera. The significance of older Tertiary floras of Siberia, together with those of Manchuria, Korea, and adjacent parts of north- eastern Asia, has been admirably summarized by Kryshtofo- vich.^^ He concludes that nowhere in this region during the Tertiary were there tropical or subtropical conditions, but that as late as the Lower Miocene it was occupied by a deciduous forest "composed mosdy of temperate types such as Fagus, Ulmus, Alnus, Betula, Corylus, Populus, Juglans, Comptonia, and Trapa'! He points out the similarity of this flora to that from the Eocene of Alaska, and the marked difference between it and the subtropical floras of older Tertiary age in western Europe. His explanation of the cause of this difference is perhaps open to question, involving as it does the supposed migration of the North Pole to the northern part of the Pacific. It seems apparent that the southwestern part of Eurasia, extensively submerged in early Tertiary time, must have been characterized by a much warmer climate than the northeastern part of this land mass, far removed from the ameliorating influences of the Atlantic. In any event, with the pole in its present position, the climatic differences between these two parts of Eurasia are almost as great today as those indicated for the Eocene. In addition to 62 R' W. Chancy latitude the major factor in determining these differences seems to be the position of the two areas with relation to prevailing circulation of air and water currents. The resemblance of these early Tertiary floras of Alaska and northeastern Asia to the Miocene floras of Oregon and adjacent states will be seen by reference to the list of Miocene genera on page 68. It seems possible to conclude that in the early part of the Tertiary period there was no such uniformity of climate over a wide range of latitude as has been pictured by earlier writers, some of whom are responsible for the myth of the occurrence of tropical floras in the Arctics. The existence of such floras has been so frequendy disproved by Berry^' and others in recent years that their sup- posed climatic implications can be wholly disregarded. The fossil plant record indicates that there were differences in tem- perature and humidity along the Pacific Coast of North America during the Eocene which, though perhaps not as marked as those of today, indicate that latitude was a controlling factor, with subtropical vegetation toward the south and progressively more temperate floras northward into Alaska. The diiferentia- tion of Tertiary vegetation into a series of climaxes, or a cHsere, in response to differences in latitude and altitude was suggested by Clements almost twenty years ago." This change in the character of the vegetation because of the influence of latitude makes possible certain important inferences concerning the nature of the upland forests during the Eocene and subsequent epochs. The fossil record is of necessity in- complete, since it is usually preserved at sites of deposition situ- ated on the lowlands. Studies of the dispersal of modern leaves and fruits from the trees on which they grew indicate that under normal conditions of transportation they are rarely deposited except in the immediate vicinity of the parent trees. The assump- Cenozoic Floras Around the Northern Pacific 63 tion may readily be made that all the more common plant re- mains in a fossil deposit are of lowland types, from trees adjacent to the site of deposition. The actual record is, therefore, of these plants which grew in lowlands, where the accumulation of their leaves and other structures is made possible by their inclusion in sedimentary deposits. In the Tertiary of North America there are only occasional remains, such as wind-borne seeds and leaves of exceptional durability, which give material evidence of the nature of upland vegetation. It seems obvious that in any Tertiary situation the trees of the upper slopes and ridges had a more xerophytic aspect than that of the flood plain, lake borders, and coastal lagoons, the vegeta- tion of which has been preserved in the fossil record. Just as the redwood forest of today has a border phase dominated by Doug- las fir, live oak, and madrone, and as the rain forest of the tropics grades up into an association characterized by oaks and other temperate trees, so the redwood forest of the Miocene and the rain forest of the Eocene may be assumed to have had border phases conditioned by less humid climate the temperate range of which was more extreme. The nature of these border phases can rarely be direcdy determined from an adjacent fossil record, since only exceptionally has the upland vegetation been pre- served in sedimentary deposits. An indirect but effective method of determining the nature of the vegetation of the uplands may be followed by examining the contemporary floras of higher latitudes. In the same manner as the high altitude border phase of modern cUmax forests shows a parallel development in higher latitudes near sea level, the upland vegetation bordering a Ter- tiary climax in middle latitudes may be indicated by the nature of the fossil record to the north, at lowland situations suited to its deposition and preservation. 64 R' W< Chaney The presence of a typically temperate forest in Alaska during the Eocene therefore makes possible the assumption that a simi- lar assemblage occupied the uplands in Oregon and California at the time that the Goshen, Comstock, and other subtropical forests were living in the lowlands and shedding their leaves into adjacent basins of deposition. Unquestionably, there were differences in detail between these contemporary forests of high altitudes and high latitude; in the upland forest, Pseudotsuga or Abies may have taken the place of northern Sequoia as a dominant in the same way that Pinus and Abies supplant the dominant Picea of high latitudes in the subalpine forests of middle latitudes today. But the general aspect of the Eocene forest of the uplands may be reconstructed on the basis of the contemporary floras at high latitudes. Of further significance is the fact, already mentioned, that this Eocene vegetation of high latitudes, presumably deposited near sea level, and the Eocene vegetation assumed to have been present in the uplands of middle latitudes, have a definitely Miocene aspect. The up- land vegetation of Oregon and California during the Eocene epoch did not become part of the fossil record because it was destroyed in the course of transportation toward areas of deposi- tion in the lowlands. With the lowered humidity and tempera- ture of Miocene times, it was shifted downward toward sea level from the uplands, until it came to occupy the sites of deposition in which the earlier floras had accumulated in the Eocene. At the same time the southern limit of the high-latitude forest was shifted down from Alaska into Oregon and California. There- after this temperate flora entered the fossil record as the well- known redwood forest of the Miocene. Certain differences may be recognized between the middle-latitude Miocene flora and its high-latitude Eocene equivalent. But these are no greater than Cenozoic Floras Around the Northern Pacific 65 might be expected in view of the length of time involved and the distance covered. It is significant to note that this forest, charac- teristic of the lowlands of Oregon and adjacent states during the Miocene, was present in only slighdy modified form at higher latitudes on both sides of the Pacific during early Tertiary time, and that its high-altitude equivalent may be assumed to have occupied the upper slopes and ridges as far south as California. To a tradition-bound stratigraphic paleontologist, the occurrence of a "Miocene" flora in the uplands of Oregon and California during the Eocene epoch may prove disquieting, and may result in the conclusion, not infrequently expressed, that plant fossils are of litde value as time indicators.* With no denial that homo- taxiality, of both fossil floras and faunas, may be used as a rough basis for synchroneity in the dating of sedimentary formations, the conclusion may be expressed that for any detailed knowl- edge of the age of a fossil occurrence, data concerning latitude, altitude, and position with relation to ocean basins are essential. For the equally important considerations relating to the distri- bution and physical indications of the life of the past, such data are obviously required. In this connection it seems desirable to quote a statement written in 1879 by J. S. Gardner." This statement was not read by the present writer until after the preceding discussion in this paper had been completed. It indicates the early recognition by paleobotanists of the need for knowing something about the geographic occurrence of a fossil flora before its age can be de- termined : The nearly unbroken sequence seen in the Eocene floras extends • It should be borne in mind, however, that because of its remoteness from sites of deposition, this upland flora with Miocene characteristics did not enter the record during the Eocene epoch. The critical feature in stratigraphic dating is not the time of existence of a flora or fauna, but the time of its entrance into the sedimentary record. 66 R' W. Chancy into the Miocene. There is no great break in passing from one to the other when we compare them over many latitudes, and but litde change beyond that brought about by altered temperature or migra- tion. But if Tertiary floras of different ages are met with in one area, great changes on the contrary are seen, and these are mainly due to progressive modifications in climate, and to altered distribution of land. From Middle Eocene to Miocene the heat imperceptibly dimin- ished. Imperceptibly, too, the tropical members of the flora dis- appeared; that is to say, they migrated, for most of their types, I think, actually survive at the present day, many but very slighdy altered. Then the sub-tropical members decreased, and the temperate forms, never quite absent even in the Middle Eocene, preponde- rated. As decreasing temperature drove the tropical forms south, the more northern must have pressed closely upon them. The Northern Eocene, or the temperate floras of that period, must have pushed, from their home in the far north, more and more south as climates chilled, and at least, in the Miocene time, occupied our latitudes. The relative preponderance of these elements, I believe, will assist in determining the age of Tertiary deposits in Europe, more than any minute comparisons of species. Thus it is useless to seek in the Arctic Regions for Eocene floras, as we know them in our latitudes, for during the Tertiary period the climatic conditions of the earth did not permit their growth there. Arctic fossil floras of temperate, and therefore Miocene aspect, are in all probability of Eocene age, and what has been recognized in them as a newer or Miocene f acies is due to their having been first studied in Europe in latitudes which only became fitted for them in Miocene times. When stratigraphical evidence is absent or inconclusive, this un- expected persistence of plant types or species throughout the Terti- aries should be remembered, and the degrees of latitude in which they are found should be well considered before conclusions are published respecting their relative age. In response to physical changes controlled apparently by emer- gence with its resultant modification of ocean currents, and by mountain-making with its resultant modification of atmos- pheric circulation, a temperate forest came to occupy the lower Cenozoic Floras Around the Northern Pacific 67 altitudes in western North America in the Miocene epoch. Since it Uved in situations suited to the preservation o£ leaves and other parts of the plants, this forest has left a plentiful and wide- spread record from CaUfornia north to southern British Colum- bia, and eastward across Idaho and Nevada into Montana and Colorado. By the end of the Miocene, reduction in rainfall and temperature, with associated orogeny, had greatly restricted its distribution and limited its opportunities for entering the fossil record. This forest has survived, with important modifications, in many limited areas in western North America. The presence of its altered modern equivalent in the eastern United States at the present time suggests that in the Miocene it may have occupied this area as well, but the fossil record of its occurrence there is too scant to establish definitely such a distribution. The dominant species of the Miocene floras of western Amer- ica is Sequoia langsdorfii, the Tertiary equivalent of the living S. sempervirens, the coast redwood. Present in the older Tertiary floras of Alaska and northeastern Asia, this species is completely absent from the Eocene deposits of California and Oregon.* It is much reduced in distribution in the Pliocene epoch, and may therefore be used as an accurate index fossil for the Miocene of middle latitudes. The climatic and other physical requirements of the modern coast redwood are so restricted as to make its Tertiary equivalent a valuable indicator of past environments. Particularly is this true where remains of Sequoia langsdorfii are found with those of Alnus, Quercus (Lithocarpus), Umbellu- laria, and other genera which are its most regular associates along the northern coast of California at the present time. Ex- • S. langsdorfii is recorded from the upper Eocene of northwestern Washing- ton, in a flora intermediate between the typical subtropical type from the Eocene of more southern latitudes and the typical temperate type from the Eocene of Alaska. 68 R. W. Chancy tensive quantitative studies involving a comparison of the fre- quency of occurrence of leaves and fruits in Tertiary deposits and in contemporary stream deposits have been carried on by the writer in the past fifteen years. From numerous localities, of both Miocene and Recent deposition, a large amount of data has been compiled which shows that Alnus, Quercus (Lithocarpus), Sequoia, and Umbellularia made up the same dominant ele- ment in the redwood forest of the Miocene as these genera do today. A count of 20,611 fossil specimens from the type locality of the Lower Miocene Bridge Creek flora, in the John Day Basin of eastern Oregon, has shown that four fossil species of the above-named genera make up 86.44 per cent of the total; a count of 8422 modern specimens from a valley in Muir Woods, near the coast of central California, has shown that the four modern equivalents of these genera make up 85.44 per cent of the total." The general aspect of the Miocene forest, particularly in the early part of the epoch, must have been strikingly similar to that of the redwood belt on the California coast. A survey of the fol- lowing list of the common families and genera of the Miocene floras of western North America will emphasize, however, the presence of two important elements which are no longer domi- nant or present there. Ginkgoaceae Tsuga Ginkgo Taxodiaceae Taxaceae Glyptostrobus Taxus Sequoia Torreya Taxodium Pinaceae Cupressaceae Abies Libocedrus Picea Thuyites Pinus Liliaceae Pseudotsuga Smilax Cenozoic Floras Around the Northern Pacific 69 Salicaceae Rosaceae Populus Amelanchier Salix Cercocarpus Myricaceae Crataegus Myrica (Comptonia) Prunus Juglandaceae Ribes Juglans Rosa Hicoria Sorbus Betulaceae Leguminosae Alnus Cercis Betula Leguminosites Carpinus Anacardiaceae Corylus Rhus Ostrya Aceraceae Fagaceae Acer Castanea Hippocastanaceae Castanopsis Aesculus Fagus Sapindaceae Quercus (including Sapindus Lithocarpus) Rhamnaceae Ulmaceae Ceanothus Celtis Rhamnus Ulmus Tiliaceae Trochodendraceae Tilia Cercidiphyllum Onagraceae Berberidaceae Trapa Odostemon Cornaceae Lauraceae Cornus Sassafras Nyssa Umbellularia Ericaceae Saxifragaceae Arbutus Hydrangea Oleaceae Philadelplius Fraxinus Hamamelidaceae Caprifoliaceae Liquidambar Viburnum Platanaceae Platanus 70 R- W. Chaney The more important o£ the two elements of the Miocene flora which are absent or sparsely represented in the living forests of western North America is the deciduous element comprising such genera as Carpinus, Castanea, Fagus, Hicoria, Liquid- amhar, Nyssa, Ostrya, Sassafras, Tilia, and Ulmus. Whereas such genera as Aca-, Aesculus, Alnus, Amelanchier, Betula, Cornus, Corylus, Fraxinus, Juglans, Philadelphus, Platanus, Populus, Prunus, and Salix have survived on the Pacific Coast, retaining their deciduous habit, Ceanothus, Quercus, Rhamnus, and Rhus are predominandy evergreen; and a group of genera including many of the most numerous and wide-ranging forms in the modern forest — Arbutus, Castanopsis, Lithocarpus (Quercus), Myrica, Odostemon, and Umbellularia — are evergreen. A ten- tative suggestion may be made that the source of this element was in the south. The extensive representation of modern spe- cies of Lithocarpus in the low-latitude forests of Asia, and of evergreen members of the Lauraceae in the tropics of both hemispheres, suggests that their ancestors, shifting northward in the Eocene, may have survived in California and Oregon, or have continued their migration in the Miocene. There is some indication from the paleontologic history of the Ericaceae that they also had their origin in the south. In any event, the present abundance of broad-leafed evergreens in low latitudes and their known occurrence there in Cretaceous and Eocene time makes plausible the suggestion of their southern origin. The elimination of a large number of deciduous genera from the flora of western America, with an increased number of broad-leafed evergreens, appears to be related to fundamental changes in climate since the Miocene, involving the concentra- tion of rainfall in the winter months. This Lower Miocene flora as a whole indicates a temperate climate over much of western Cenozoic Floras Around the Northern Pacific 71 North America which was characterized by annual precipita- , tion of at least 40 inches, probably more evenly distributed throughout the year than at present. The mild, uniform tempera- ture in which frost played an inconspicuous part was controlled by the adjacent Pacific, whose influence extended farther inland than it does today, extensive highlands not being present until a later epoch of the Tertiary. The second Miocene element, now missing from the forests of North America, comprises a group of genera and species which have survived in Asia. Such genera as Cercidiphyllum, Gink^go, Glyptostrobus, and Trapa no longer occur outside of Asia. Other genera, still represented in North America, have species living in Asia which are more similar to the Miocene members of these genera; examples, with the fossil species in parentheses, are Acer pictiim (A. scottiae), Castanea henryi (C. orientalis), Cra- taegus pinnatifida (C. newberryi), Querciis myrsinae folia (0. consimilis, in part), Rhus sylvestris (R. rnerrillii), and Ulmus par vi folia (U. brotvnellii) . In this connection, it is of interest to note that the genus Sequoia, now extinct in Eurasia, was widely distributed over that continent in the Tertiary. If the specific characters of various dicotyledonous species from the Miocene of Asia were better known to the writer, it seems probable that he would be able to include in addition a list of modern species from North America which are as closely related to them as are the living Asiatic species of their genera, or even more nearly than these. In any event, although the Miocene floras of Asia are at present incom- pletely studied, it is apparent that they include numerous genera such as Acer, Alnus, Carpinus, Fagus, Populus, Sequoia, and Tilia, which are characteristic of this epoch in North America. There appears to be much less difference between the Miocene y 72 ^' W. Chaney and Eocene floras of Asia than between those of the same series in western North America. This may be accounted for by the comparative remoteness of the Asiatic floras from the Atlantic Ocean and its cUmatic effects; whereas changes in temperature of the Pacific were readily translated to the Tertiary floras of western America, similar changes in the Atlantic from the Eocene to Miocene epochs had less effect upon the Miocene floras of central and eastern Asia because of its distance from the region they occupied. In like manner many Miocene genera, par- ticularly of deciduous types, have survived to the Recent in the temperate parts of eastern Asia. The living forests of northern Japan, Korea, Manchuria, and northern China, as observed by the writer, have a definitely Miocene aspect. The resemblance of these forests to those of eastern North America was long ago recognized by Asa Gray,"'' who suggested a past interchange of vegetation between the two areas across high latitudes. In the light of modern studies of later Tertiary floras and their distribu- tion, it seems possible to conclude that this deciduous element, comprising genera and species from both sides of the Pacific, has survived on the eastern sides of both continents under climatic conditions involving summer rain and winter cold. The gradual elimination of deciduous types from the forest of the Pacific Coast, with the resultant dominance of broad-leafed evergreens, has resulted from the development there of a climate character- ized by mild, rainy winters and cool, dry summers. Comparison of the list of older Tertiary families and genera of Oregon and California (pp. 58-59), and of the list of Miocene families and genera (pp. 68-69), emphasizes that there is a great difference in their composition. With unimportant exceptions, the families and genera of the Eocene have their most extensive modern development in the warmer parts of the world, while Cenozoic Floras Around the Northern Pacific 73 those of the Miocene are in large part restricted to temperate areas. Where the same famiHes fall in both lists, as do the Liliaceae, Juglandaceae, Fagaceae, Lauraceae, Leguminoseae, and Anacardiaceae, they are for the most part represented in the two floras by different genera, which are tropical in the Eocene and temperate in the Miocene. In the lists as here composed, the only genera common to both are Liqtddambar, Platanus, Qiiercus, Stnilax, and Viburnum, all of which are ancient genera of wide distribution since the Cretaceous, ranging at the present time from temperate lowlands to middle altitudes in the tropics. It must be admitted that if all the genera reported from the Mio- cene, especially by the older authors, were included in the list (pp. 68-69), ^^^ resemblance between this list and the Eocene flora would be more marked. Such genera as Anona, Aralia, Cinnamomum, Diospyros, Ficus, Persea, and Sterculia have been regularly recognized in Miocene floras, but usually on evi- dence which is not accepted by the writer. While it is possible that an occasional representative of a tropical genus may have survived as a relict in the Miocene floras of western America, this in no way affects the general conclusion that the members of the older Tertiary forests had in large measure disappeared from the middle latitudes before the beginning of the Miocene. With altered climatic conditions, the northern limits of the ranges of the typical Eocene species shifted southward, and the temperate types of higher altitudes and latitudes moved down to take their places as the forest dominants of Oregon and California. Many of these* have continued to live in western America down to the present time, but subsequent changes in topography and climate have restricted their distribution to the most favored situations. * To almost every Tertiary species, a name different from that of its modern equivalent has been applied, but the actual differences as based on leaf characters may not always warrant these specific designations. 74 ' R. W. Chancy The forest dominated by the coast redwood is confined to a narrow belt along the coast from central California to southern Oregon. Its more hardy equivalents, among the conifers, extend northward along the coast to Alaska, and occupy the middle levels of the Cordillera, associated with most of the same dicotyle- donous genera. As above indicated, a similar forest occupies an extensive area in northeastern Asia. One of the great physical events of later Tertiary time in west- ern America was the outpouring of extensive and numerous sheets of basaltic and other lavas. The wide areal extent of these flows is itself an indication of relatively low relief over much of Washington, Oregon, and Idaho at the beginning of the Miocene. This eruptive activity has continued down into the Recent, but it reached its climax in the middle part of the Mio- cene epoch; at that time there were built up the thick series of flows which have formed the Columbia Plateau and which are so well exposed in the gorge of the Columbia River and in its tribu- tary streams of Oregon and Washington. The Lower Miocene sediments underlying these flows have been exposed by the cut- ting of the John Day and Crooked rivers in eastern Oregon,^' and the record of the Bridge Creek flora, with its dominant red- wood character as above described, has been exposed. During intervals of quiescence between eruptions, there developed a somewhat altered type of forest in Oregon and Washington. This forest, which has been described as the Eagle Creek^^ and the Latah'''"'* floras, differed from the Bridge Creek in the less dominant position of Sequoia, and in the greater abundance of maples, poplars, and such broad-leafed deciduous oaks as Quer- cus pseudolyrata. It is clearly of a more upland type than the flood-plain forest of the Lower Miocene, and appears to have occupied a region of greater relief. Even more definite is the Cenozoic Floras Around the Northern Pacific 75 suggestion of the Mascall and other related floras which occur in tufFaceous deposits overlying the Columbia lava, and which are referred to the Upper Miocene. In these later floras, Sequoia holds a relatively unimportant position; the dominant types are Abies, Acer, Aesculus, Arbutus, Betula, Platanus, Populus, and Quercus (both broad-leafed and evergreen types). In such florules as that from Trout Creek in southeastern Oregon" and from the Blue Mountains of eastern Oregon,"^ there is definite indication, in the occurrence of Abies, Amelanchier, Picea, and Sorbus, of an altitude of several thousand feet at the site of deposition. The general character of the Upper Miocene floras has been shown to resemble closely that of the forest now living on the borders of the redwood belt'^ under more exposed and less humid conditions than those of the Bridge Creek flora. The stratigraphic position of these Upper Miocene floras is commonly so well marked, separated as they are from the sediments bear- ing the Lower Miocene Bridge Creek flora by the thick series of basaltic flows of Columbia lava, that there is litde possibility of error in establishing their relative age; and the Upper Miocene vegetation is always of a less mesophytic type than that of the Lower Miocene. The habitat, involving in all probability a more exposed topographic setting, was also characterized by less rain- fall and greater extremes of temperature. These physical changes appear to have been caused by the incipience of the orogeny which has subsequendy produced the Cascade Range and re- lated structural units in the northern part of the Great Basin, together with renewed uplift of the Sierra Nevada. The effect of these mountains on the climate and vegetation of the areas to the east is too well known to merit description in this paper. It need only be mentioned that they cut off from the interior of western North America the atmospheric circulation which had \ 76 R' W. Chancy given it its mild, moist climate; as a result, the redwood and related coniferous forests became confined to the coastal slopes. From eastern Washington and Oregon the more mesophytic elements of the vegetation have gradually been eliminated until they survive only to a limited degree at higher altitudes in pro- tected situations, with Artemisia and Juniper us occupying the broad expanse of the Great Basin. The gradual steps by which the humid forest was restricted, not only in North America but also in Asia, are being made known by studies of Pliocene floras on both sides of the Pacific. With the breaking up of the extensive areal units by mountain- making, with resultant diversity in topography and climate, the situations in which plant remains might be preserved became fewer. As a result, the floral record of this latest epoch of the Tertiary is less completely known than the record of those pre- viously discussed. In western America, Pliocene floras are known only from California and Nevada.'"' ""' '' Except in the coastal region, where a modified redwood forest has persisted down to the present, they are made up, in large part, of trees such as Alnus, Platanus, Populus, Quercus (evergreen-oak types), and Salix, which occupy stream borders in regions of low rainfall. Another element, com^nsmg Arctostaphylos, Ceanothus, Cerco- carpus, Garry a, and Heteromeles, makes up the first unit of true chaparral that has been recognized in the fossil record. A third element, apparendy confined to higher altitudes, includes Cupressus, Libocedrus, Pinus (nut-pine type), Quercus, and Robinia, with the palm, Washingtonia, on the adjacent sand washes. Most of these Pliocene species are not only similar in leaf character to modern species of their genera, but they also appear to have occupied essentially the same habitats over the same range. The exotic element, unlike that of older Tertiary floras, Cenozoic Floras Around the Northern Pacific 77 is numerically unimportant. A significant example of this ele- ment is Ulmus brotvnellii, a species with small leaves having simple-serrate margins, first reported from the lower Miocene of Oregon, and ranging as far east as Colorado before the end of that epoch. Elm leaves of the same characteristic appearance have recently been recorded from the lake beds of Pliocene age in Shansi Province, China,^ and U. parvifolia, the modern equivalent of both fossil species is today widespread in the cooler and drier forested areas of northern China. Here is an example of a tree the leaf characters of which permitted its survival in western America long after most deciduous types, including sev- eral species of its own genus, had become extinct, and which has survived down to the present in the somewhat more favorable habitat afforded in northeastern Asia. As is true of the Pliocene floras of western North America, the known vegetation of this epoch in Asia is essentially like that now occupying the same areas. Acer, Amelanchicr, Picea, and Ribes are among the other genera recorded in Shansi. To the west, in Sinkiang and adjacent provinces of northwestern China, a small but critically important flora has recendy been described,^* ^^^ all of whose species have closely related living equivalents in central Asia at the present time. The two common trees are a poplar whose leaves are indistinguishable from those of the Populus euphratica now occupying river borders and sand washes in the interior of Asia Minor, and an elm that is essen- tially like Ulmus pumila, which occupied the arid portions of central and northern China, including the borders of the Gobi Desert of Mongolia. Associated with the leaf impressions of these and other species of the sand washes and canyons, there have been recorded the remains of an aquatic element which occupied the lake basins in which the sediments were accumulated. Leaves 78 R. W. Chancy and stems of Nuphar and Typha, together with grasses and sedges, give a picture not unlike that in many arid and semiarid parts of Asia and North America today, where low precipitation and the consequent reduction in importance of valley develop- ment result in the presence of numerous shallow basins at the foot of mountain ranges. Although the Pliocene floras of both continents show a marked resemblance to the modern vegetation of the same regions, there is apparent in most of them a somewhat less arid aspect than exists at present. The physical conditions character- izing this epoch may be summarized as having involved a topography essentially as diverse as that of today, together with a climate in which high range of temperature and annual rain- fall varying from lo to 20 inches resulted in the development of locally differentiated plant formations with relatively few generic and specific representatives. The Pleistocene floras of western America have been so re- cently and adequately described by Mason and others'^' that the present discussion warrants only a brief summary of the plant record for the epoch immediately preceding the Recent. Most of these floras represent coastal vegetation in California that is of essentially the same type as that now living, including a large number of herbaceous species represented mosdy by fruiting structures. The most significant difference in the present discussion is that certain of the Pleistocene florules, notably those from Carpinteria and Santa Cruz Island, are made up of species which occupied regions several hundred miles beyond the pres- ent southern limits of their range. The extension into southern California of the range of such species as Cupressus goveniana, Pinus radiata, Pseudotsuga taxifolia, and Sequoia sempervirens may be interpreted as an indication of increased precipitation Cenozoic Floras Around the Northern Pacific 79 and lowered temperature quite as definite as the advance of ice sheets from the north. The subsequent withdrawal of these species northward to and beyond the Monterey Peninsula is an equally definite record of a return to southern California of the dry and warm climate which now characterizes it. The relation of the Pleistocene floras of California to the modern coastal closed-cone pine forest, with its present differentiation into floris- tic units, has been so fully discussed by Mason"" that it need only be mentioned here. Pleistocene floras in Oregon, Washington, and British Colum- bia have not been extensively collected and studied, but those now available indicate the probable existence of differences from those of California which were consistent with the differences in latitude. A small flora from near Fairbanks, Alaska,^ is made up of Betula, Salix, and other boreal types. It may be suggested that in the Pleistocene there was essentially as great diversity in the forests from south to north along the Pacific Coast as now. The Pleistocene floras of Asia have only recendy become an object of study, especially in connection with the occurrence of early man in China."' *^ On the basis of our present incomplete knowledge, it may be supposed that they differed only in a small degree from the modern forests of the regions where they have been collected. In discussing the history of the Cenozoic plants bordering the northern basin of the Pacific, particularly with reference to their stratigraphic sequence and geographic distribution, it has been necessary at various points in this paper to describe and interpret the physical conditions under which they lived. This has been essential because their distribution both in time and in space has been conditioned by physical changes which have resulted in their migrations. The major movements have been southward So R. W. Chancy as a result of a progressive trend toward cooling and aridity since the early Tertiary, but secondary movements in the opposite direction have been detected and have played their part in estab- lishing the character of modern vegetation. The gradual shifting southward of the northern limits of most of the Eocene species has resulted in their restriction to the low latitudes of Asia and America, in the forests of which there is a remarkable corre- spondence of genera on both sides of the Pacific. Following closely behind them, as successful competitors under changing conditions of climate, the temperate Miocene flora with its dominant Sequoia element left the northern regions and became widely established in middle latitudes of both continents. Further modifications have served to eliminate the redwood from most of its former range and have localized the deciduous element on the eastern sides of the continents where summer rains and \ winter cold are prevalent, or to the higher altitudes of western America where this general climate is duplicated. With the late Tertiary uplift of the Cascade Range and the Sierra Nevada to essentially their present height, and with the consequent devel- opment of diversified topography of high relief, the relatively uniform floras of middle Tertiary time have also become diversi- fied, with major differences on either side of the mountain ranges, and with the various minor differences which are appar- ent in the modern vegetation. A modified type of Miocene forest has migrated as far south as southern Guatemala, where such genera as Alnus, Arbutus, Crataegus, Cupressus, Myrica, Ostrya, Pinus, Prunus, Quercus, and Salix occupy the highlands above the subtropical vegetation of the lower slopes, thus duplicating the association of these two climatic types which characterized the older Tertiary in Oregon and California some 2500 miles to the northwest. \ Cenozoic Floras Around the Northern Pacific 8i Profound changes in physical conditions around the northern basin of the Pacific in late geologic time, as shown by the lithol- ogy, structure, and areal extent of Cenozoic rocks, have altered the vegetation of this region, especially with respect to its distri- bution. It seems clear that a consideration of the paleontologic record, as well as of the living plants, forms the basis for any complete understanding of modern vegetation. BIBLIOGRAPHY ^ Clements, Frederic E. 1916. Plant succession, Carnegie Inst. Wash. Pub. 242, p. 347. ^ Chanev, Ralph W,, and Sanborn, Ethel I. 1933. The Goshen Flora of west central Oregon, Carnegie Inst. Wash. Pub. 439, PP- 1-103. ' Chaney, Ralph W. 1925. A comparative study of the Bridge Creek Flora and the modern red- wood forest, Carnegie Inst. Wash. Pub. 349, pp. 1-22. 1925. The Mascall Flora — Its distribution and climatic relation, Carnegie Inst. Wash. Pub. 349, pp. 23-48. 1925. Notes on two fossil hackberries from the Tertiary of the western United States, Carnegie Inst. Wash. Pub. 349, pp. 49-56. 1925. A record of the presence of Umbellularia from the Tertiary of the western United States, Carnegie Inst. Wash. Pub. 349, pp. 57-62. * Chaney, Ralph W. 1927. Geology and paleontology of the Crooked River Basin with special reference to the Bridge Creek Flora, Carnegie Inst. Wash. Pub. 346, pp. 45-138. ^ DoRF, Erling 1933. Pliocene floras of California, Carnegie Inst. Wash. Pub. 412, pp. 1-112. ° Chaney, Ralph W., Mason, Herbert L., and Potbury, Susan S. 1934. Pleistocene palaeontology of California, Carnegie Inst. Wash. Pub. 415, pp. I-I79- '' Sanborn, Ethel I. 1935. The Comstock Flora of west central Oregon, Carnegie Inst. Wash. Pub. 465, 1. ^ Potbury, Susan S. 1935. The La Porte Flora of Plumas County, California, Carnegie Inst. Wash. Pub. 465, II. * Chaney, Ralph W, and Sanborn, Ethel I. 1933. The Goshen Flora of west central Oregon, Carnegie Inst. Wash. Pub. 439, PP- 28-42. " Berry, Edward W. 1 91 6. The Lower Eocene floras of southeastern North America, U. S. Geol. Surv., Prof. Paper 91. " Chaney, Ralph W, and Sanborn, Ethel I. 1933. The Goshen Flora of west central Oregon, Carnegie Inst. Wash. Pub. 439, pp. 28-42. [82] Cenozoic Floras Around the Northern Pacific 83 " Chaney, Ralph W., and Mason, Herbert L. 1934. A Pleistocene flora from Santa Cruz Island, California, Carnegie Inst. Wash. Pub. 415, p. 21. " Chaney, Ralph W., and Mason, Herbert L. 1934. A Pleistocene flora from the asphalt deposits at Carpinteria, California, Carnegie Inst. Wash. Pub. 4i5> ?• 79- " Chaney, Ralph W. 1930. A sequoia forest of Tertiary age on St. Lawrence Island, Science, 72, pp. 653-654. " Kryshtofovich, a. N. 1929. Evolution of the Tertiary flora in Asia, New Phytologist, 28, pp. 303- 312. " Berry, Edw.\rd W. 1930. The past climate of the north polar region, Smithsonian Misc. Coll., 82, pp. 1-29. " Clements, Frederic E. 1916. Plant succession, Carnegie Inst. Wash. Pub. 242, pp. 356, 362. " Gardner, J. S. 1879-82. Mon. Brit. Eocene Flora, i, pp. 7-8. " Chaney, Ralph W. 1924. Quantitative studies of the Bridge Creek Flora, Am. Jour. Sci., 8, pp. 127-144. " Gray, Asa 1859. Observations upon the relations of the Japanese flora to that of North America^ and of other parts of the northern temperate zone, Am. Acad. Arts and Sci., n.s. Mem. 6, pp. 437-449. ^ Merrl\m, John C. 1 90 1. A contribution to the geology of the John Day Basin, Univ. Calif. Publ. Bull, Dept. Geol. Sci., 2, pp. 269-314. ^ Chaney, Ralph W. 1925. A comparative study of the Bridge Creek Flora and the modern red- wood forest, Carnegie Inst. Wash. Pub. 349, pp. 1-22. 1925. The Mascall Flora — Its distribution and climatic relation, Carnegie Inst. Wash. Pub. 349, pp. 23-48. 1925. Notes on two fossil hackberries from the Tertiary of the western United States, Carnegie Inst. Wash. Pub. 349, pp. 49-56. 1925. A record of the presence of Umbellularia from the Tertiary of the western United States, Carnegie Inst. Wash. Pub. 349, pp. 57-62. ^ Chaney, Ralph W. 1927. Geology and paleontology of the Crooked River Basin with special reference to the Bridge Creek Flora, Carnegie Inst. Wash. Pub. 346, pp. 45-138. 84 R. W. Chaney " Chaney, Ralph W. 1920. The flora of the Eagle Creek Formation, Contr. Walker Museum, 2, pp. 115-181. ^ Knowlton, Frank H. 1926. Flora of the Latah Formation of Spokane, Wash., and Coeur d'Alene, Idaho, U. S. Geol. Surv., Prof. Paper 140-A, pp. 17-81. " Berry, Edward W. 1929. A revision of the flora of the Latah Formation, U. S. Geol. Surv., Prof. Paper 154-H, pp. 225-265. " MacGinitie, Harry D. 1933. The Trout Creek Flora of southeastern Oregon, Carnegie Inst. Wash. Pub. 416, pp. 21-68. ^ Oliver, Elizabeth 1934. A Miocene flora from the Blue Mountains, Oregon, Carnegie Inst. Wash. Pub. 455, pp. 1-27. ^ Chaney, Ralph W. 1925. The Mascall Flora — Its distribution and climatic relation, Carnegie Inst. Wash. Pub. 349, pp. 23-48. ^^ DoRF, Erling 1933. Pliocene Floras of California, Carnegie Inst. Wash. Pub. 412, pp. 1-112. ^^ Webber, Irma E. 1933. Woods from the Ricardo Pliocene of Last Chance Gulch, California, Carnegie Inst. Wash. Pub. 412, pp. 113-134. ^^ AxELROD, Daniel I. 1934. A Pliocene flora from the Eden Beds, Am. Mus. Novitates, no. 729, pp. 1-4. ^ Chaney, Ralph W. 1933. A Pliocene flora from Shansi Province, Geol. Soc. China Bull., 12, pp. 129-142. ®* Chaney, Ralph W. 1935. The Kucha Flora in relation to the physical conditions in central Asia during the late Tertiary, Svenska Sallskapet for Antropologi och Geog- rafi, Geografiska Annaler, Sven Hedin, pp. 75-105, 1935. ^ Chaney, Ralph W, and Mason, Herbert L. 1934. A Pleistocene flora from Santa Cruz Island, California, Carnegie Inst. Wash. Pub. 415, pp. 1-24. 1934. A Pleistocene flora from the asphalt deposits at Carpinteria, California, Carnegie Inst. Wash. Pub. 415, pp. 45-79. ^ Potbury, Susan S. 1932. A Pleistocene flora from San Bruno, San Mateo County, California, Carnegie Inst. Wash. Pub. 415, pp. 25-44. Cenozoic Floras Around the Northern Pacific 85 "^ Mason, Herbert L. 1934. Pleistocene flora of Tomales Formation, Carnegie Inst. Wash. Pub. 415, pp. 81-179- ^ Mason, Herbert L., ibid. '* Frick, Childs 1930. Alaska's frozen fauna, Natural History, 30, pp. 71-80. " Chaney, Ralph W. 1927. Hackberry seeds from the Pleistocene loess of northern China, Am. Mus. Novitates, no. 283, pp. 1-2. *^ Chaney, Ralph W., and Daugherty, Lyman H. 1933. The occurrence of Cercis associated with the remains of Sinanthropus, Geol. Soc. China Bull., 12, pp. 323-326. The Origin of the Desert CUmax and CUmate By Frederic E. Clements IN THE POPULAR MIND, the word "desert"calls to memory a vast sandy waste, typified by a picture o£ the Sahara in the first book of geography. The idea has been further confused by the earher misconceptions concerning the "Great American Des- ert" of the Missouri Valley, which was actually a grassland cUmax (where an unrivaled agricultural development has taken place), and the Great Salt Lake Desert, an alkaline expanse left by the retreat of the lake after glacial times. It is evident that these regions have little in common, apart from the quality of vastness and desolation in human terms. The scientist's usage has natur- ally been more definite, basing itself in part at least upon the physical factors of rainfall and temperature, but even he has sometimes been misled by climatic relicts that have assumed a new role as indicators of disturbance. Cacti in particular have contributed to this result, because they bear the impress of the true desert, but they are often found far beyond its present cli- matic limits. Extensive areas of grassland badly denuded by overgrazing have also been called desert, such as the Red Desert of Wyoming and the southern San Joaquin Valley. However, the latter does have the warrant of a cover of flowery annuals, found likewise in the neighboring Mohave Desert. The popular usage finds some justification in the original meaning of the word as "abandoned" and hence uninhabited or uninhabitable. In geographic parlance a desert is a region which has little or no vegetation because of a scarcity of rainfall, and thus is relatively unpopulated; while in ecological terms it is an area marked by a peculiar vegetation or, in places or at times, j [87] 88 F- E. Clements by an absence of it, with low rainfall and high temperature as causal factors. Though the Arctic tundra is sometimes regarded as a desert, this is clearly a return to the primitive meaning of the word and is warranted neither by vegetation nor by climatic conditions. As a consequence, an accurate and practicable defi- nition of the term apparendy must be based primarily upon vegetation and rainfall, with some consideration of the animals and of high temperature during the growing season at least. In brief, a desert climax is one marked by the absence of forest or grassland, a critical deficit in rainfall, and a high potential evaporation caused by excessive heat and often by high winds also. In applying this test to the vegetation of North America, it becomes clear that the term desert is often loosely employed by ecologist and geographer alike. Thus, the so-called sagebrush desert does not meet the requirements over all or almost all of its extent, and even the more xeric Larrea community is desert in only a part of the area concerned. The most effective test is afforded by tracing the boundary of the grassland climax in the Southwest and, when this is correlated with rainfall or, better still, rainfall/evaporation, where this is possible, it provides the best delimitation of desert available today. In general, the isohyet of 5 inches marks the disappearance of grass dominants on the climax level, and thus is the readiest means of setting off the desert climax. Much attention has been given to following this boundary on the ground, but it will suffice here to state that the desert, as characterized by Larrea, Franseria, and their typi- cal associates, is confined to the Death Valley, Mohave, and Colorado regions and to a larger but less known area in Mexico. Though small districts of similarly low rainfall occur to the north, there temperature and evaporation are less telling, and The Origin of the Desert Climax and Climate 89 the Larrea climax is replaced by sagebrush and the latter then by grassland. It should be clearly recognized that the limits of the desert or Larrea climax cannot be drawn with exactness at present, and that no more than a fair approximation will ever be possible. Quite aside from the difficulties arising from the slow fluctua- tions caused by cycles of varying length and intensity, are those produced by overgrazing and its consequences in a climate where this process is the most destructive to grassland. Hence, the limits of the desert plains and mixed prairie associations must be determined by scattered relicts, which are sought most diligently and with greatest success in the wet phase of the cycle, when they are most in evidence. Moreover, these are the very regions deficient in weather records, and the isohyets are based upon relatively few and recent observations so that they serve only as a general index. However, if it be borne in mind that all climaxes are characterized by broad ecotones, it will be seen that these criteria serve as a practicable basis for circum- scribing the desert region. From the essential nature of climates and climaxes, it follows that deserts usually lie in direct contact with grassland rather than woodland or forest. This is particularly true of the Colo- rado complex and its wide extension in Mexico, though the detailed relations are there much less known. Since the domi- nants are typically bushes or shrubs, they may come into touch with another scrub formation, such as the sagebrush on the north or along the middle slopes of bordering ranges, but this is partly at least a consequence of grazing disturbance. The desert is largely encircled by three associations of the grassland forma- tion, namely, the mixed prairie on the northeast, desert plains along the extended eastern boundary, and bunchgrass prairie to go RE. Clements the west (Clements, 1920; Weaver and Clements, 1929). The last two are to be considered as climatic differentiations of the mixed prairie, arising from the same causes that led to the development of desert in an originally more or less uniform vegetation. This is indicated by the fact that the desert plains of southern Arizona and northern Mexico today swing north as a border along the Colorado Desert. Hence, the origin of the desert vegetation is inseparably bound up with the disappearance of the grass domi- nants in the region of desiccation and their realignment in the areas round about. The typical dominant of the desert climax, Larrea tridentata, resembles the sagebrush, Artemisia tridentata, in its distribution over an area much wider than that in which it is climax. In the United States it is found from western Texas to the eastern edge of the Great Valley of California but, of all this vast stretch, it is climax only in western Arizona and southeastern California, as well as in Mexico. In the wake of overgrazing, it has spread extensively to replace the vanishing grasses, and today its proper climatic region can only be determined by the presence of reUct grasses on the one hand and of its peculiar desert associates on the other. As in various other climaxes, disturbance pro- cesses have produced a misleading picture, the true significance of which can be discovered only by turning to the dynamics involved. This demands not merely a proper understanding of the effects of disturbance by man, catde, and rodents, but like- wise an insight into the direction and influence of minor and major climatic shifts. Both bio-ecology and paleo-ecology must be called upon to provide a complete synthesis and to permit a reconstruction of past biomes, as well as to trace the develop- mental processes active in producing the present vegetation. These are the objects of the present paper. The Origin of the Desert Climax and Climate 91 An effort has been made to marshal all the facts that have a bearing, direct or otherwise, upon the modification of grassland into desert under compulsion of a major climatic change. The primary cause of the latter is to be sought in the elevation of neighboring mountain systems, just as its effects are disclosed in the fossil record of the plants and animals that passed across the changing scene. In order to permit a sharp focus, the exist- ence of a grassland climax as widespread as that of today must be taken for granted. This assumption is well justified by the transmontane occurrence of the great genera of grasses, such as Stipa,Bout€loua,Sporobolus,Poa,Andropogon,ttc, and attested by an ungulate fauna far more varied than that of the present epoch. While the origin of the prairie climax has a direct bearing upon the present problem, this is a matter of such scope and complexity as to demand separate treatment in the future. In assembling the evidence from all sources, it is necessary to employ the essential method of paleo-ecology by proceeding from the existing biome to the past, and in this task the relict provides the missing key. These witnesses to past events are of several kinds. Chief among them are the relicts still present in the desert, which are few in the climax itself, but more frequent in the protection of mountains in or bordering upon the region. Much more numerous and also of signal value are what may be termed "transads" plants and animals either present or absent today, but at home in the desert when it was grassland as shown by their presence on both sides of the climatic barrier. Next in significance are the species that occur in the desert and on one side of it, either to the east or to the west. Of similar meaning are the changes that have taken place farther away in the grass- land without transforming it into desert. A correlated body of evidence is to be derived from the migration of the desert climax Q2 E E. Clements from the Mexican homeland, and the evolution of endemics at this or a later time. The indispensable background for this is the evaluation of fossil horizons of plants from Miocene to Pleistocene, with which must be associated the fossil animals and especially the vertebrates. Finally, behind all these lies the record of great orogenic movements expressed in the slow rise of mountain blocks, the reduced access of rain-laden winds, and the "spiral" desiccation of the regions within the lee of the rising chains. The word "spiral" is employed to characterize the resultant of progressive desiccation and of recurring climatic cycles of varying intensity, of which those of the Pleistocene were among the most pronounced. Relicts Within the Desert Borders Of this group the perennial grasses, as dominants of the original climax, are obviously the most significant, while the annuals are of little importance apart from their serai relations. These climax dominants have persisted only as a result of local pro- tection, maintaining themselves in some degree in shallow washes, in sand, or on rocky or northerly slopes, wherever edaphic compensation for climatic drouth was to be found. A number have taken refuge in mountain ranges by virtue of higher rainfall and reduced evaporation. Naturally, the grass genera of southern origin have supplied most of these relicts, though the circumpolar Stipa, Elymus, and Poa are numbered among them. In the following list, the species are arranged under the associations in which they play major roles today, accompanied by those of subclimax or serai character. As is to be expected from their nature and proximity, the desert plains exhibit the largest number. The Origin of the Desert Climax and Climate 93 Oryzopsis hymenoides Poa nevadensis Elymus sitanion Aristida purpurea Aristida p-wrighti Aristida divaricata Aristida d-parishi Aristida californica Aristida glauca Mixed prairie Hilaria jamesi Bouteloua gracilis Bouteloua racemosa Desert plains Hilaria rigida Triodia pulchella Sporobolus cryptandrus flexuosus Sporobolus wrighti Muhlenbergia porteri Muhlenbergia iemmoni California prairie Poa scabrella Stipa speciosa Stipa pulchra Coastal prairie Andropogon saccharoides Andropogon glomeratus Serai dominants Distichlis spicata Hordeum jubatum Sporobolus airoides Phragmites communis Sporobolus asperifolius Imperata hookeri Few of these constitute extensive communities today. Hilaria rigida and Triodia pulchella are the most important in this re- spect, and next in order are Oryzopsis hymenoides, Aristida p-turighti, Stipa speciosa, Poa scabrella, and Distichlis spicata. The most complete community relicts so far discovered occur in the Avawatz Mountains on the south rim of Death Valley at an altitude of 4300 feet. Here are associated Hilaria jamesi, H. rigida, Aristida p-wrighti, Stipa speciosa, Poa scabrella, Oryzop- sis hymenoides, and Sporobolus c-flexuosus, at their best on rocky north exposures. In sandy plains about the playas in the 94 E E. Clements western part of the Mohave, Poa scabrella, Oryzopsis, Stipa spe- ciosa, Elymus sitanion, and E. cinereus grow together, and in sandy areas at GoflFs and across the Colorado River at Yucca, Hilaria rigid a, Aristida p-wrighti, Muhlenbergia porteri, and Sporoboltis c-flexHosus occur. As would be expected, almost all these are southern in origin; they belong to the grass associations that border the desert today and have most recently withdrawn from it. By contrast, such major dominants of the mixed prairie as the two Boutelouas and Hilaria jatnesi persist only in the mountains, while the boreal Stipa cofnata and Agropyrum smithi have moved well to the north. The presence of two or more dominants from each of four associations out of the six that form the prairie climax denotes that the original grassland of the desert area belonged to a biome much less differentiated than at present. Most wide-ranging of the species concerned are Bonteloua racemosa and Sporobolus cryptandrtis, which extend from the Atlantic to California and from Canada into Mexico; B. gracilis and hirsuta, with a similar extension from the Middle West; and Andropogon glo7neratus, which stretches in a broken line from ocean to ocean in the south.* Transads These are dominants or subdominants that still exist on both sides of the desert and consequently must have extended through * Note: A steady effort has been made to verify and extend recorded ranges both in field and herbarium during the past decade or more, but it is not to be hoped that final accuracy can be attained in this respect. Some early records by competent botanists are no longer verifiable, especially where disturbance has obliterated rare extralimital stations as it seems to have done with Btichloe in Canada and Arizona. For the Southwest in particular, ranges have been drawn chiefly from Jepson's Manual for California, Tidestrom's Flora of Utah and Nevada, and Shreve's MS catalogue for Arizona. Sargent's Manual and Sud- worth's bulletins have been utilized for the drawing of tree ranges. The Origin of the Desert Climax and Climate 95 it during occupation by the grassland climax, but may or may not occur within its borders today. Here again the grasses are of the first importance, though perennial forbs and woody plants also possess much significance. Strictly speaking, transads are to be regarded as single species, but groups of closely related lin- neons have similar if somewhat less meaning, and this is also true of certain paired genera. The annuals have value, but it is often difficult to determine, and they receive small considera- tion, except for the group of desert endemics. Furthermore, hydrophytes and halophytes not only possess peculiar ability to migrate but they also evade climatic control in varying degree. Their presence, therefore, is not of great significance. Finally, a large number of species occur in the higher altitudes of moun- tains on both sides of the desert, but these have moved from the north along parallel systems and probably have never stretched across the desert region. In consequence, the species listed below comprise only those of which there is clear evidence that they once extended across some part of the cUmax desert, as represented by the region from Death Valley southward. There was desiccation to the east all along the Sierra-Cascade axis, but this was not sufficient north of Owens Lake to maintain the Larrea desert. Grass transads still present in the desert Oryzopsis hymenoides Phragmites communis Stipa speciosa Bouteloua gracilis Aristida divaricata Bouteloua racemosa Aristida d-parishi Andropogon saccharoides Aristida purpurea Andropogon glomeratus Poa scabrella Sporobolus airoides Poa s-nevadensis Sporobolus asperifolius Elymus sitanion Hordeum jubatum Distichlis spicata ly^ a6 F. E. Clements Grass transads no longer found in the desert Stipa comata Elymus triticoides Stipa minor Koeleria cristata Stipa eminens Epicampes rigens Stipa coronata ' Bromus marginatus Bouteloua hirsuta Andropogon scoparius cirratus Bouteloua rothrocki Aristida d-schiedeana Bouteloua radicosa Eragrostis secundiflorus Aristida p-fendleriana Panicum huachucae Several o£ the grasses named above are dominants of the true prairie, notably Andropogon scoparius, Bouteloua racemosa, and Koeleria cristata. The last-named retains this role through the mixed prairie and into the bunchgrass association of Cali- fornia. In reduced rank, though still as a dominant, B. racemosa appears in both mixed prairie and desert plains; Andropogon becomes postclimax in both. Bouteloua rothroc\i, B. radicosa in several forms, and Aristida d-schiedeana are dominants of the desert plains, while Elymus triticoides is the valley consociation throughout the California prairie. FoRB Transads Of this group, the perennial species are of chief importance as being more characteristic of climax areas, while the annuals are in large part and often wholly serai. Like the grasses, this group is divisible into those still persisting in the desert and those that have disappeared from it but that survive today on both sides. The latter are generally more mesic than the former, and a number of them, namely. Erysimum asperum, Glycyrhiza lepidota, Lotus americanus, Verbena hastata, and Urtica gracilis, extend east into the true prairie or farther. There are approxi- mately 70 desert transads, of which 41 are annuals. The transads The Origin of the Desert Climax and Climate 97 no longer present in the desert comprise 44, 20 of them annuals. The greater number of annuals in the first group accords with the general rule for the desert. The majority of the perennials extend at least to the western edge of the Great Plains, but the typical societies of the true and mixed prairies drop out long be- fore reaching the borders of the desert, as would be expected. As may readily be inferred, aquatic forbs afford an exception- ally high percentage of transads; in fact, there are but one or two of this group restricted to the desert and only a few that extend to either side. The number of transads still found in the desert barely exceeds a dozen, and these are almost wholly con- fined to the sedge family, to Juncus, Typha, and the duckweeds. The aquatic transads that have disappeared from the desert are twice as many and their relationships are much wider, the cos- mopolitan genera of the figwort family being best represented. Although the inference is plausible that many of these transads have persisted since the remote days of extensive swamp and meadow, the relative ease of migration among aquatics does not exclude more recent colonization. However, they not only aid in completing the details of the reconstructed picture, but they also prove that desert climates do not exclude perennial ponds and swamps with their characteristic water plants and animals. Thus, in desert as in grassland generally, such edaphic com- pensation for climatic dryness is more or less adequate, if occasional, and the organisms concerned cannot be direcdy em- ployed as climatic indicators. Shrub Transads These are likewise divided into desert transads and transads proper, the former numbering three times the latter. However, for the present purpose it is more convenient to arrange the o8 RE. Clements members of both groups in accordance with the cUmax in which they find their chief part. As to origin, by far the greater num- ber are of southern derivation, amounting to 75 per cent. The exceptions are chiefly dominants of the mountain chaparral that are common to the Petran and Sierran associations, such as Rhus trilobata, Cercocarptis parvifolius and ledifolius, Amelanchier alnifolia, Holodiscus discolor, and Jamesia americana. A few such as Artemisia tridentata, Atriplex canescens, Chrysotham- ntis nauseosus and viscidiflorus, Eurotia lanata, Gutierrezia saro- thrae, Sarcobatus vei-miculatus, and Ribes cereum range as far north as Canada. Desert scrub climax Larrea tridentata Prosopis juHflora EnceHa farinosa Acacia greggi Acamptopappus sphaerocephalus Baccharis emoryi Haplopappus HnearifoHus Baccharis glutinosa Dalea californica Hymenoclea salsola Prunus fasciculata Hymenoclea monogyra Nolina parryi Pluchea sericea Sagebrush climax Artemisia tridentata Sarcobatus vermiculatus Atriplex confertifolia Brickellia californica Atriplex canescens Viguiera deltoidea Chrysothamnus nauseosus Zauschneria californica Chrysothamnus viscidiflorus Pentstemon antirhinoides Grayia spinosa Eriodictyum californicum Eurotia lanata angustifolium Salvia carnosa Chaparral climax Rhus trilobata Holodiscus discolor Cercocarpus parvifolius Jamesia americana Cercocarpus ledifolius Arctostaphylus pungens Ceanothus cuneatus greggi Arctostaphylus pringlei Amelanchier alnifolia Purshia glandulosa The Origin of the Desert Climax and Climate 99 Prosopis and Acacia are postclimax, being found typically in washes or valleys along with other treelike species, while Bac- charis, Hymenoclea, and Pluchea are essentially serai. The ma- jor dominants of the Basin sagebrush are those listed in the first column; Sarcobatus constitutes the subclimax of the halo- sere, and the remaining species belong to the Coastal sagebrush. As noted earlier, most of the chaparral dominants are north- eastern, but Arctostaphylus and Ceanothus are derived from the Pacific Coast. Tree Transads Because of their higher water requirements and more direct exposure to the influence of a changing climate, trees present few simple transads. But three species of this group occur in the desert at this time, namely, Salix lasiolepis, exigtia, and Populus fre?nonti; the first two are usually shrubs, and all are confined to valleys or washes with higher water-content than that of the desert proper. For similar reasons, the same genera furnish most of the transads no longer present in the desert, such as Salix lev- igata, nigra, fttwiatilis, and Populus trichocarpa. The other member of this group is Celtis mississippiensis in the variety reticulata. Phylads and Genera TREES In addition to single species that cross the desert, there is a much larger number of phylads, that is, phyletic lines of closely related species, that stretch from the Atlantic or the Mississippi Valley to the Pacific. In some of these the relationship is so intimate that the ecologist would regard the forms concerned as sub- species or varieties of a single linneon. This is true of the phylad Quercus minor-gambeli-garryana, the last two members of 100 E E. Clements which appear to diiiFer in no important particular. Quercus minor and gam belt are in close proximity in western Texas; the latter occurs in the mountains o£ western Arizona and Nevada, and garryana extends into the southern Sierra Nevada. Q. alba and lobata are likewise nearly related, though the present gap be- tween them reaches from central Texas to the edge of the Great Valley in California. A similar lacuna exists between Q. velutina and \elloggi, which might well be regarded as varieties of the same species. An even more striking example is afforded by the virginiana phylad, of which the primary species ranges from Virginia through the South to Mexico and Lower California, where it meets Q. dumosa of central and southern California. 0. emoryi of the same group stretches from the mountains of western Texas through those of southern Arizona, and Q. engelmanni lies in contact with dumosa in southern California and in close proximity to virginiana. The phylad Q. breviloba-undulata- douglasi is morphologically continuous, the respective ranges being from Alabama to central Texas, western Texas to the east- ern border of the desert, and from northern California to the rim of the Mohave. Similarly, in the phyletic line Q. myrti folia- hypoleuca-wislizeni-agrifolia, the first extends from South Caro- lina to Mississippi, the second from western Texas to Arizona and Sonora, and the third from desert slopes to Lower California and northward, often meeting agrifolia in southern California. Phylads of similar significance in respect to evolution and migration are exhibited by Juglans, Fraxinus, Platanus, Populus, Pinus, and Juniperus, as well as by the more shrublike Arbutus, Aesculus, and Prunus. Juglans nigra-rupestris-californica is essentially a structural continuum, and from Massachusetts to California is interrupted only by the desert. The continuity in The Origin of the Desert Climax and Climate loi both respects is even greater, if possible, in Fraxinus ameri- cana-texensis-veltttina-coriacea-oregona; the first two overlap in Texas; velutina reaches from Texas to Inyo County, California; coriacea occurs in the desert region from southern Utah to south- eastern California, and oregona from the mountains of southern California to Washington. Platanus occidentalis-tvrighti-race- mosa extends from New England to southern and Lower Cali- fornia, touching the borders of the desert on both sides. Populus deltoidea and fremonti meet in southern Colorado and western Texas to span the continent, and the closely related wislizeni overlaps them both in Texas and New Mexico. Though best regarded as one polymorphic species consisting of monophylla, edtdis, cembroides, and quadrifolia, with com- plete overlapping in the needle number, the pinyon is usually treated as four. Pinas edulis ranges from Utah and Colorado to western Texas, Arizona, and northern Mexico; cembroides through the southern half of Arizona and adjacent Mexico to Lower CaHfornia; quadrifolia from Lower California to the arid Santa Rosa Mountains along the Colorado Desert, and monophylla from northeastern Utah to Nevada, northern Ari- zona, and the southern Sierras into Lower California. Jumperus utahensis and californica meet today in southeastern California, while the general areas of monosperjna and occidentalis confront each other over a long line. Arbutus menziesi makes its home from British Columbia to the mountains of San Diego County, California; it is separated by the desert from A. arizonica in southern Arizona and Chi- huahua, where A. arizonica is in proximity to A. xalapensis of Mexico and Texas. Aesculus californica reaches the margin of the Mohave in Antelope Valley; its cognates are glabra and octandra, which range from Pennsylvania to Oklahoma and to 102 RE. Clements western Texas, respectively. The wide interval is occupied in part by the related monotypic Ungnadia speciosa of Texas, New Mexico, and northern Mexico. Prunus nana-melanocarpa-de- missa is found from Newfoundland to California, the wide- spread melanocarpa overlapping the limits of the other two. Less definite though still suggestive evidence of former con- tinuity across the desert is presented by a number of paired genera, such as Castanea-Castanopsis, Catalpa-Chilopsis, Persea- Umbellularia, Sabal-Washingtonia, Taxodium-Sequoia, and Aes- culus-Ungnadia, as already mentioned. In each pair the eastern genus finds its western limits in Texas, except Catalpa, which stops in Mississippi; while the western genera are confined to California or the Pacific Coast, except Chilopsis, which reaches the lower Rio Grande, and Ungnadia, which ranges from east- ern Texas to New Mexico and Mexico. SHRUBS Several phylads of this group cross the continent almost without interruption, though it is a pertinent question whether each of these is not a single species. Sambucus canadensis ranges from New Brunswick to Colorado and Texas; it gives way in New Mexico to mexicana, which reaches California and there meets the closely related glauca. Cercis canadensis-reniformis-occiden- talis is unique in that all three members occur in Texas, reni- formis being the intermediate both morphologically and geo- graphically. Vitis exhibits a phylad of even greater continuous area : girdiana extends from coastal southern California into the deserts, californica grows throughout the state, arizonica occurs from Arizona to west Texas, and aestivalis from Texas along the Gulf and Adantic Coasts. Staphylea and Styrax are likewise transcontinental, but with a wide gap in the mountain region. The Origin of the Desert Climax and Climate 103 Staphylea trifolia is found from Quebec to Kansas, while bo- landeri is restricted to central and northern California; Styrax californica is more southern, and the limits of pulverulenta are from Virginia to Texas. Both forms of Amorpha occur over the wide area from New Mexico to California, jruticosa reach- ing southern Pennsylvania and californica the San Francisco Bay region. The remaining phylads are western, Philadelphus letfisi oc- curing from southern California to British Columbia, and tnicrophylla from Arizona to Colorado. Zauschneria californica ranges from California to Nevada, and arizonica from Califor- nia through Arizona to southwestern Colorado. Eriodictyum californicum extends from southern Oregon through central California, crassi folium southward to the border, trichocalyx along the Cross Ranges to the Santa Rosa Mountains at the west side of the desert, and californicum angusti folium from the east side through southern Nevada and Arizona to New Mexico. Ephedra californica of the southern part of the state meets tri- furca in the Mohave and Colorado deserts, and trifurca stretches east to Texas and Colorado, sharing the area east of the desert with torreyana.T\it more restricted Crossosoma bigelovi occurs from central Arizona across the desert to Lower California, while californica is confined to islands off the southern coast. Species of the Desert and Eastward or Westward FORBS The number of species found both in the desert and eastward into the desert plains or mixed prairie is 89; of these, 49 are an- nuals, 39 of which extend as far east as New Mexico, Colorado, or Texas. Almost all the perennials likewise stretch far to the lOA E E. Clements east, and the majority constitute societies in the mixed prairie o£ the Great Plains. It appears probable that most of the species of the desert-eastward group originated in the two prairie asso- ciations just mentioned and found their western limit at the desert's edge, though some may well have extended into it or even have arisen during its development. The species that occur in the desert and westward through more or less of cismontane California number 91 ; 67 of these are annuals, 60 of which are found more widely outside the desert. Nearly all the desert- westward perennials appear to be endemic, while the desert- eastward ones are chiefly migrants from the east, as already sug- gested; 7 likewise occur in Mexico. SHRUBS The total of desert-eastward shrubs is 43, 21 of which extend well to the east into Colorado or Texas. Of the westward group of 20, the large majority are endemic, only 4 being found also in Mex- ico; while of the eastern, 15 occur in Mexico and may well have originated there. DESERT ENDEMICS The word endemic is used in a general sense to apply to the desert region as a whole, with full recognition of the fact that the limits of the desert are most uncertain in Mexico and Lower California and none too definite in any direction other than west. On this basis, the group comprises 89 species of annuals, 78 perennials, and 69 shrubs, the total amounting to 236, in con- trast to the 247 species found in the desert but also extending into the climaxes on either side. Of endemic annuals, the largest numbers per genus are found in Phacelia, Cryptanthe, Gilia, Euphorbia, Langloisia, and Mentzelia; among perennials, Astra- The Origin of the Desert Climax and Climate 105 gains exceeds all others, with Aster and Sphaeralcea next. Opnntia, Dalea, and Salvia lead in the number of shrubby species, followed by Eriogoniim, Brickellia, Atriplex, and Echi- nocactus. Differentiation of the Colorado and Mohave Floras Though both deserts belong to the Larrea-Franseria climax, they exhibit a noteworthy difference in genera and species. This is evidently caused by the San Bernardino Range and its continua- tion to the Colorado River. This serves as a definite physical barrier, though it is not insuperable, and its eastern flank is readily turned. It is also a climatic obstacle to some degree, as was especially true during wet phases, and is in part a causa- tive factor in the not inconsiderable climatic differences between the two regions. Of endemics as defined above, the Mohave con- tains 90, the Colorado 57, and 89 are common to both ; most of the common element may antedate the separation of the deserts, but part of it probably has been caused by climatic pulsations and the resulting migrations. In contrast. Death Valley with two wide entrances has a mere half-dozen strict endemics and only a few more that it shares with the regions round about. With respect to the total number of species according to life form, 5 grasses are found in the Mohave that do not occur in the Colorado, 4 show the reverse relation, and 17 are present in both. There are 138 forbs in the Mohave, 69 in the Colorado, and 157 are common to both; in the same order, the shrubs number 51, 43, and 59. As would be expected, neither the few grasses nor the many forbs produce much visible difference. This is like- wise true of the shrubs and low trees over the general climax area, but these do afford conspicuous contrasts through valleys and washes and along mountain slopes and alluvial fans. In spite io6 F' E' Clements of the smaller number of woody species, these often give a dis- tinctive character to the Colorado Desert. The most notable are Washingtonia filifera, Cereus giganteus, Fouquiera splendens, Par\insonia torreyana, Olneya tesota, Dalea spinosa, Beloperone californica, and Simmondsia californica. It is understandable that these are all of southern derivation, while the Mohave has obtained most of its important shrubs from the west, north, and especially the east. Even the Joshua tree (Yucca brevi folia) y though characteristic, is not a strict endemic. Fossil Plants As is well known, grassland and desert do not aflEord ready op- portunity for the production of plant fossils, partly because leaves and stems are not easily detached, but chiefly for the reason that water bodies are relatively few in those regions. The paucity of fossils is naturally more pronounced in desert than in grassland and, so far as the climax desert is concerned, amounts to almost complete absence from the Pleistocene up to the pres- ent. Conditions were more favorable on both counts during the Pliocene and still more so in the Miocene, with the result that the fossil floras are increasingly better known for these periods of greater rainfall. Hence, in scrutinizing the evidence of the course of development, it is desirable as usual to proceed from the present and thus to consider the Pleistocene horizons first. The value of these depends in large measure upon proxim- ity, apart from number of species; those in southern California come first, none being found in the true desert so far; those farther north demand more interpretation in terms of differ- ence of climate and vegetation. Although the general practice of paleo-botany is to provide a separate name for a fossil species, even though it is regarded as The Origin of the Desert Climax and Climate 107 probably identical with a modern one, it facilitates comparison and interpretation to employ the name of the existing repre- sentative, and this plan is accordingly pursued. The stations from which Pleistocene fossils have been re- corded in California are as follows: (i) Rancho La Brea, in the suburbs of Los Angeles; (2) Carpinteria, near Santa Barbara; (3) Santa Cruz Island, one of the Santa Barbara group; (4) San Bruno, on the west shore of San Francisco Bay; (5) Tomales Bay, not far north of the last-named. As would be expected, the poorest collection is from the first locality and the richest from the last, though edaphic conditions largely determine this for the intermediate stations. The grouping below is approximately into trees, shrubs, and herbs, though the limits of the first two are far from precise. PLEISTOCENE Trees — Pinus tuberculata Cupressus macrocarpa Junipenis californica Sambucus glauca Qucrcus agrifolia Pinus muricata Pinus radiata Pinus remorata Pinus sabiniana Pseudotsuga taxifolia Sequoia sempcrvircns Cupressus goveniana Juniperus californica Quercus agrifolia Umbellularia californica Sambucus glauca Pinus remorata Pseudotsuga taxifolia Cupressus goveniana Myrica californica Rancho La Brea Shrubs — Celtis mississippiensis Celtis reticulata Carpinteria Pirus rivularis ? Arctostaphylus glauca Ceanothus thyrsiflonis Garrya elliptica Rhus diversiloba Eriodictyum californicum Myrica californica Herbs— Xanthium calvum Pteris aquilina Chorizanthe pungens Cymopterus litoralis Corethrogyne filagini- folia ? Xanthium calvum Arceuthobium campy- lopodum Santa Cruz Island Arctostaphylus sp. Ceanothus thyrsiflorus Cornus pubescens Arceuthobium campy- lopodum io8 E E. Clements Trees — Pseudotsuga taxifolia Sambucus glauca Pseudotsuga taxifolia Picea sitchensis Torreya californica Pinus muricata Pinus radiata Cupressus goveniana Acer macrophyllum Ainus rubra Arbutus menziesi Quercus agrifolia Umbellularia californica Myrica californica Sambucus glauca PLEISTOCENE {Continued) San Bruno Shrubs — Ceanothus thyrsiflorus Rhus diversiloba Symphoricapus albus Rubus parviflorus Baccharis pilularis Prunus emarginata TOMALES Arctostaphylus Colum- biana ? Ceanothus thyrsiflorus Ceanothus rigidus Garrya elliptica Baccharis pilularis Adenostoma fasciculatum Rhus diversiloba Amelanchier alnifolia Cornus californica Corylus rostrata cali- fornica Prunus emarginata Prunus subcordata Rubus parviflorus Rubus spectabilis Rubus vitifolius Symphoricarpus albus Vaccinium ovatum Arctostaphylus uva-ursi Herbs— Micromeria chamissonis Galium sp. Silenc sp. Montia perfoliata Hemizonia virgata Rumex salicifolius Rumex persicarioidcs Equisetum sp. Carex sp. Sparganium eurycarpum Pteris aquilina Fragaria californica Montia perfoliata Galium californicum Eriophyllum s. artemisi- folium Calandrinia caulesccns Atriplex hastata Daucus pusillus Camassia leichtlini Rumex salicifolius Rumex occidentalis Rumex maritimus Carex spp. Scirpus spp. Montia howelli Montia sibirica Montia fontana Oenanthe sarmentosa Ruppia maritima Arceuthobium campy- lopodum ? Datisca glomerata A detailed consideration o£ the significance o£ the species in each flora and of the cHmatic impHcations of the flora has been given by the several authors, namely, Frost (1927), Chaney and Mason (1933, 1930), Potbury (1932), and Mason (1934). For The Origin of the Desert Climax and Climate 109 the present purpose, the most important fact is their essential agreement that the entire stretch was more humid and corre- spondingly cooler during the Pleistocene than it is at present, though with the logical inference that this applies to the wet rather than the dry phases of the several cycles. The second value of the lists is to be derived from the general agreement in respect to important dominants, together with their sharp reduction inland at the south. The reference of these dominants to the present climaxes of California furnishes the basis for testing the reconstruction of the successive climaxes that have moved over the desert region from the Miocene to the present. PLIOCENE RiCARDo: Last Chance Gulch Woodland — Chaparral — Pinus cembroides monophylla Robinia neo-mexicana Cupressus arizonica ? Washingtonia r filifera Quercus agrifolia Alturus-Pico Coniferous: Coast-montane — Chaparral — Sequoia sempervirens Cercocarpus parvifolius Pseudotsuga taxifolia Amelanchier alnifolia Pinus lambertiana Prunus demissa Libocedrus decurrens Garrya fremonti Pinus muricata Heteromeles arbutifolia Ribes nevadense Rhus laurina Berberis aquifolium Ceanothus integerrimus Smilax californica Pteris aquilina Wooodwardia radicans Woodland — Serai — Quercus agrifolia Populus trichocarpa Quercus douglasi Populus fremonti Quercus chrysolepis Salix lasiolepis Quercus tomentella Salix exigua Castanopsis chrysophylla Platanus racemosa Umbellularia californica Alnus rhombifolia Aesculus californica Fraxinus oregona Arbutus menziesi Pinus sabiniana no RE. Clements In evaluating the Ricardo group, Webber (1933:134) states that all the species are possibly identical with those living now in the general region, but that difiFerencees in range justify their assignment to the Pliocene. Today, the closest approach to the grouping and the related climate is to be found in the San Jacinto Range bordering the Colorado Desert, where palms occur on the desert level, with pinyons and live oaks on the slopes toward the desert. The conclusion is reached that the climate was less extreme than that of the present desert. The evidence from the other Pliocene horizons of California is indirect because these are much more northerly, except the very scanty remains from the Pico of Ventura (Dorf, 1933:66). The several floras show a close relationship to the living forests of the region, modified somewhat by the effects of a major cli- matic cycle. It is assumed that the climate of the Lower Pliocene was rather drier and warmer than the present one, but that con- ditions became cooler and more humid in the Upper Pliocene than they are in the same districts today. Since the- Ricardo is regarded as Lower Pliocene, it is clear that the conclusions are not in accord, though this discrepancy may be explained by topographic control. However, there is distinct agreement that the climate of California was drier throughout the Pliocene than during the Miocene. MIOCENE Genera common to the Mascall and Florissant Deciduous climax — Coniferous climax — Serai or marginal Acer Sequoia Ulmus Vitis Castanea-Castanopsis Quercus Libocedrus Pinus Juglans Betula Cclastrus Ptelea Fraxinus Populus Hydrangea Liquidambar Salix Odostemon Magnolia Carpinus Sapindus Sorbus Crataegus Amelanchier Rosa Apocynum Typha The Origin of the Desert Climax and Climate iii These two floras have but two species in common, or six if those in the related horizon of Yellowstone Park are taken into account. However, this fact is of little significance, chiefly be- cause the Mascall of the John Day and Trout Creek have been critically reworked by Chaney (1925) and MacGintie (1933) on the basis of the newer ecological technique, while the Rocky Mountain floras yet await such scrutiny. Furthermore, whatever element of doubt may reside in species based upon leaf im- pressions, this is largely resolved in respect to genera. Conse- quendy, it is logical to assume that both regions exhibited essen- tially the same deciduous and coniferous climaxes, and that these stretched more or less continuously between the two regions in their respective habitats, while the plateaus and plains were occupied chiefly by grassland and scrub. The continuity of forest was doubtless greater in the Eocene because Fagus, Castanea, and Tilia were present both in the Rocky Mountains and the Pacific Slope, the intervening shrinkage of forest and extension of grassland being promoted by conditions during the Oligocene of the interior. Fossil Animals The carnivores and insectivores are omitted from the following lists for the various formations, primarily because they bear no direct relation to vegetation as food supply. This is likewise true of the raptores among birds. As dwellers in serai habitats, water birds possess little or no value as climatic or climax indicators for grassland or desert and, in this respect, the value of such ground forms as quail, meadowlark, and turtle dove is none too definite (see Grinnell, 1914, 1923, for the birds of Death Valley and the Lower Colorado) . Naturally, the major interest inheres in the ungulates because of the direct response of teeth to type of food and of feet to 112 F. E. Clements surface and cover (Kowalevsky, 1873; Osborn, 1910:238). The proboscideans are less satisfactory indicators, and the rodents are significant chiefly in connection with the food coactions and soil reactions that the same species or genera exhibit in the mod- ern biome. PLEISTOCENE The selection of the following formations during the three periods of the Pleistocene, Pliocene, and Miocene has been de- termined in large measure by position in the desert region, or proximity to it, or location in a climate similarly the result of mountain-making processes. The number of genera represented has also been a consideration of importance, though not suffi- cient to warrant including the rich faunas of the Potter Creek and Samwel Caves of northern California, which belong in a cUmate of much higher rainfall. For the reasons already given, the water birds are omitted except for the McKittrick area, where they are included for the purpose of suggesting the animal members of the water community in the Pleistocene (see Grin- nell, 1923:54-60). Lake Manix MAMMALS Ungulates — Equus, large Camel, small Equus, small Antilocapra Camelops ? large Proboscideans — Elephas or Mastodon Bautista Creek MAMMALS Ungulates — Equus Antilocapra Tapirus Odocoileus Camelid Capromeryx The Origin of the Desert Climax and Climate 113 Ungulates — Equus occidentalis Camelops Camel Proboscideans — Mastodon sp. Ground forms — Lophortyx californica Zenaidura macroura Oxyechus vociferus Water forms — Podilymbus podiceps Anas platyrhynchos Mareca americana Nettion carolinense Querquedula cyanoptera Spatula clypeata Dafila acuta Marila americana Erismatura jamaicensis Chen hyperboreus McKlTTRICK MAMMALS Bison sp. Antilocapra ? BIRDS Geococcyx californianus Corvus corax Branta dickeyi Jabiru mycteria Ardea herodias Grus canadensis Rallus virginianus Recurvirostra americana Macrorhamphus griseus Totanus melanoleucus Pelidna alpina Ungulates — Equus occidentalis Camelops hesternus Bison antiquus Bison latifrons Proboscideans — Elephas imperator Elephas columbi Rodents — Lepus californicus bennetti Sylvilagus bachmani cinerascens Sylvilagus audoboni sanctidiegi Citellus beecheyi fisheri Dipodomys agilis agilis Thomomys bottae pallescens Neotoma intermedia intermedia Neotoma fuscipes macrotis Rancho La Brea MAMMALS Antilocapra americana ? Odocoileus hemionus ? Capromeryx minor Platygonus sp. Mastodon americanus Perognathus panamintinus brevinasus Perognathus californicus californicus Peromyscus maniculatus gambeli Peromyscus californicus insignis Peromyscus eremicus fraterculus Microtus californicus neglectus Reithrodontomys megalotis longicaudus Onychomys torridus ramona 114 E E. Clements Ground forms — Lophortyx californica ? Parapavo californicus Columba fasciata Zenaidura macroura Otocoris alpestris Agelaius gubernator Ungulates — Equus near occidentalis Camelops ? c£. hesternus Rodents — Lepus near californicus Sylvilagus bachmani Dipodomys sp. Thomomys near bottae Neotoma sp. Microtus californicus Ground forms — Lophortyx californica Parapavo californicus Columba fasciata Turdus migratorius RANCHO LA BREA (Continued) BIRDS Xanthocephalus xanthocephalus Geococcyx near californianus Sturnella neglecta Corvus corax Corvus brachyrhynchus Carpinteria MAMMALS Odocoileus sp. Bison ? sp, Perognathus sp. Peromyscus sp. Onychomys ? sp. Eutamias sp. Sciurus sp. BIRDS Geococcyx californianus Corvus corax Corvus caurinus Sturnella neglecta Ungulates — Equus pacificus Camelops kansanus ? Camelops vitakerianus Proboscideans — Elephas imperator Rodents — Thomomys sp. Geomys sp. Microtus sp. Lepus townsendi Ground forms — Tympanuchus pallidicinctus Palaeotetrix gilli Fossil Lake MAMMALS Antilocapra sp. Camelus maximus Eschatius conidens Elephas columbi Ondatra oregona Castor sp. Castoroides sp. BIRDS Pedioecetes phasianellus Pedioecetes nanus The Origin of the Desert Climax and Climate 115 Of the six formations considered, only the Manix falls within the desert proper, being more or less central in the Mohave. The Bautista is scarcely beyond the western end of the Colorado. In summing up the evidence for the Manix, Buwalda concludes that the climate in the region and to the west was less arid than at present (1914:457). As to the Bautista, at present this region is grassland under a rainfall thrice as great and hence the difference between Pleistocene and modern conditions was probably less. The mammals of the McKittrick, which is in the "desert" of the upper San Joaquin Valley with a rainfall of 6 inches, are nearly identical with those of the Manix; the genera are all found in the Rancho La Brea and all but one at Fossil Lake, Oregon, indi- cating a general similarity of climate over this wide range. The ungulates of Rancho La Brea suggest conditions not much unlike those of the present; the antelope and deer are still found in the general region, and within the historical period both the bison and the reintroduced horse thrived in grassland of less rainfall. Dice (1925:123) believes that the similarity in the rodents indicates that there could have been no considerable or widespread difference in climate between the Pleistocene and the present. On the contrary. Miller thinks that the evidence from the birds, though inconclusive, points to a warm moist climate during the glacial period (1925:70). Hay reaches the conclusion that the Brea deposits were laid down in the Af tonian interglacial stage of the early Pleistocene, when the climate was warm (1927), though this term may well be relative to the Nebraskan and Kansan glacial stages that preceded and fol- lowed. The indications at Carpinteria are essentially the same as at Rancho La Brea, and at both places the rodents and ground birds are nearly all those of the present (Wilson, 1934; L. H. Miller, 193 1 ; A. H. Miller, 1932). ii6 E E. Clements Though noting the differences between the faunas of Rancho La Brea and Fossil Lake deposits, Hay concludes that those of the latter also belong to the Aftonian (1927:191), in spite of the indications that it was more desert-like and colder. It seems probable that these and other discrepancies already noted are to be explained by smaller differences in the climates of the Pleisto- cene of the Far West and less clearly marked limits between glacial and interglacial stages. When allowance is made for the wet and dry phases of the climatic cycles, it appears that the latter half of the Pleistocene differed litde from the present over the vast stretch of the grassland climax. Ungulates — Hipparion mohavense Hipparion spp. 2 Pliohippus tantalus Pliohippus fairbanksi Pliohippus sp. Proboscideans — Tetrabelodon ? sp. Rodents — Lepus ? sp. PLIOCENE RiCARDO MAMMALS Procamelus spp. 2 Pliauchenia sp. Alticamelus ? sp. Merycodus near necatus Merycochoerus ? californicus Thousand Creek MAMMALS Ungulates — Pliohippus ? sp. Equus ? sp. Camel, cf. Camelus americanus Pliauchenia ? sp. Sphenophalis nevadensis Proboscideans — Tetrabelodon ? sp. Ilingoceros alcxandrae Ilingoceros schizoceras Teleoceras ? sp. Prosthennops ? sp. The Origin of the Desert Climax and Climate 117 Thousand Creek Rodents — Lcpus vetus Citellus sp. Peromyscus antiquus Peromyscus sp. Arctomys (Marmota) nevadensis Arctomys (Marmota) minor Aplodonda alexandrae MAMMALS {Continued) Diprionomys parvus Diprionomys magnus Entoptychus minimus Mylagaulus monodon Eucastor lecontei Dipoides sp. Ungulates — Pliohippus spectans Hipparion sinclairi Hipparion occidentals Hipparion leptode ? Hipparion anthonyi Proboscideans: remains — Rodents — Otospermophilus gidleyi BIRDS Branta sp. Rattlesnake MAMMALS Pliauchenia ? or Alticamelus Sphenophalos sp. Ilingoccros or Sphenophalos } sp. Teleoceras ? near fossiger Prosthennops ? sp. ? Platygonus rex Lepus (Hypolagus) vetus The three PHocene formations considered comprise a wide extent of territory in the lee of the rising mountains, from the Mohave Desert at the south to northwestern Nevada and eastern Oregon. In spite of a difference of several degrees of latitude, the ungulate faunas have much in common, the significant fact being the predominance of advanced types. Merriam has char- acterized the Ricardo region as follows (1919:526) : The Ricardo fauna suggests climatic conditions permitting the de- velopment of vegetation suitable for grazing animals. This indicates a somewhat heavier growth of grass than is found in this region at the present time. There is nothing in the constitution of the fauna to suggest conditions radically different from those obtaining today, but the presumption is in favor of less extreme aridity than is now known on the western border of the Mohave Desert. The conditions Ii8 RE. Clements prevailing in this region in Ricardo time were probably not widely different from those now obtaining in the southern portion of the Great Valley of California. The same authority states that the mammals of Thousand Creek suggest plains with meadows or occasional lakes; the rainfall may have been somewhat higher and the temperature rather lower than at present (1911:219). Merriam, Stock, and Moody (1925:59) make no statement concerning the vegetation and climate of the Ratdesnake formation, but the comparison with the Ricardo and Thousand Creeek clearly indicates that these three must have been much the same throughout, the more advanced character of the Rattlesnake being offset by the south- ern position of the Ricardo. In support of the assumption of a widespread prairie climax during the Pliocene may be adduced the occurrence of the major grazing genera in the Middle West. Three species of Hipparion are recorded for the Snake Creeek of northwestern Nebraska, two for Florida, and one for the Blanco of Texas. The Blanco also exhibits one species of Pliohippus; the Clarendon of the same state, two species; the Snake Creek, three; and South Da- kota horizons, two. The Snake Creek and neighboring Valen- tine contain two species of Procamelus, and the former has one species of Pliauchenia. Of Merycodus, the Snake Creek has one form, the Valentine two, and the Santa Fe of New Mexico two. MIOCENE Barstow Ungulates — Hypohippus near affinis Procamelus spp. 2 Parahippus ? mourningi Pliauchenia sp. Merychippus (Protohippus) Merycodus ? coronatus Merychippus intermontanus Merycodus necatus ? Merychippus calamarius stylodontus Dromomeryx or Cervus ? sp. Merychippus sumani Prosthennops ? sp. Protohippus ? or Pliohippus ? Alticamelus ? sp. The Origin of the Desert Climax and Climate 119 MIOCENE BARSTOV7 {Continued) Proboscideans — Tetrabclodon ? sp. Rodents — Lepus ? sp. Virgin Valley Ungulates — Hypohippus near osborni Parahippus cf. avus Merychippus mesonesus Camel, near Procamelus Aphelops ? sp. Moropus ? sp. Merycodus near furcatus Merycodus near nevadensis Dromomeryx ? sp. near borealis Dromomeryx ? sp. Blastomeryx mollis Merychyus ? sp. Thinohyus ? sp. Proboscideans: remains — Rodents — Lepus vetus Palaeolagus nevadensis Aplodontia alexandrae Mylagaulus monodon Mylagaulus pristinus The faunistic data for the transmontane area during the Mio- cene are derived from the same locaHties as are those of the Phocene; they are represented by the Barstow, Virgin Valley, and Mascall. These are reenforced by the Cedar Mountain de- posits near Walker Lake in western Nevada, which closely resemble the first two, particularly in the ungulates. According to Merriam (1919:450), the Barstow fauna denotes an open country with much grass and herbage, better watered than the present desert. The abundance of Merychippus, merycodonts, camels, and related forms, indicates that grasses were well rep- resented and hence that the climax must have been grassland. The climate was probably semiarid, like that of parts of the Great Valley today. Likewise, the horses, camels, and meryco- donts of Cedar Mountain suggest a semiarid climate not gready different from that of the region at present, but the other forms 120 E E. Clements are regarded as indicating a slightly more humid climate with savannah and chaparral such as now occur around the margin of the Great Valley (Merriam, 1916:168). In discussing the Virgin Valley fauna, the same author (1911 : 206) concludes that the closest relationships are with the Mascall of Oregon and the Pawnee Creek of Colorado, though the simi- larity with the Snake Creek beds of northwestern Nebraska is ac- tually greater. Here again, the ungulates provide the significant evidence, and leave little question that the Great Basin and Middle West were peopled in large part by grazing types living in a climax of mixed prairie, but probably without the short- grasses of today. Deformation and Climatic Cycles in the General Region A comprehensive account of deformational and cUmatic cycles, with particular relation to changes of climax and the significance of the climax sere or clisere, has been given in Plant Succession (1916: 304, 321, 364). The geological materials for this were drawn chiefly from Chamberlin and Salisbury (1906) and from Schuchert (1910, 1914). In scrutinizing the earlier conclusions in respect to the cHmate of the desert region, further data have been taken from Schuchert (1924), Jones, Antevs, and Hunting- ton (1925), and Antevs (1929). For the present purpose, the deformation cycles of direct im- port are three, namely, Paleocene-Lower Oligocene, Upper Oligocene-Miocene, and Pliocene-Pleistocene, forming the long , interval from the disappearance of the Cretaceous Mediterran- \/ ean to the present. The immediate consequence of each cycle was a primary climatic pulsation, in which were involved smaller ones caused in all probability by solar fluctuations. The best known and most pronounced of these were the glacial-inter- The Origin of the Desert Climax and Climate 121 glacial alternations more recently regarded as five in number, the glacial phases known as Nebraskan, Kansan, lUinoian, lowan, and Wisconsin, with the four interglacial ones. There is also increasing evidence of smaller cycles, such as can be discerned in varves, tree rings, and other phenomena, but none of these was sufficiently long to lead to the evolution or mass migration of a climax. The Paleocene deformation appears to have been greater in the region of the Rocky Mountains than along the Pacific, with its efFect to be seen in a warm-temperate deciduous forest, while the Pacific area was clothed with a subtropical forest (Chaney, 1933). In the long quiescent period of the Eocene and Lower Oligocene, swamp sedges and grasses seem to have spread widely over the interior of the continent, as indicated by the presence of such genera as Carex and Phragmites in the Creta- ceous. It is probable that the evolution of grassland proceeded more rapidly in the period of mountain-making in the Upper Oligocene to produce the forerunner of the modern prairie in the Miocene, where the typical genus Stipa is recorded, along with horses of the grazing type, Merychippus and Protohippus. It is a plausible assumption that, by the time of the Miocene, grassland had become the distinctive climax in at least the central and southern parts of the Great Plains and had spread westward through the present desert, as attested by the ruling abundance of Merychippus in the Barstow and the appearance of a still more advanced grazing type, Protohippus or Pliohippus (Mer- riam, 1919:479). With the rise of the Cascadian axis in the late Miocene, this dominance of prairie must have been gradually extended, probably with the general replacement of tall sub- tropical grasses by the northern mid-grasses and western short- grasses. 122 RE. Clements The general trend toward desiccation in climate and climax may well have found its chief expression at the time of greatest elevation of the Sierra Nevada and Cross Ranges in the Pleisto- cene, to be succeeded by the four or five glacial cycles already mentioned. If the interglacial phases were long, it is plausible that desert conditions similar to those of today intervened more than once, though the secular trend of rainfall as well as the grass relicts still found in the desert, particularly in Death Valley, sug- gest that today constitutes at least one maximum. This view is supported by the presence of two horses, two camels, a large antelope, and a probable elephant from Lake Manix, indicative of at least a fairly continuous grass-cover as late as this deposit. As to the particular times of mountain-making that produced the desiccation of the grassland climate which finally terminated in the desert climate and climax, Hershey places the structural differentiation of the Sierra Madre-San Bernardino chain at the opening of the Quaternary, concluding that the great bajadas were finished by the Illinoian or lowan stage. Baker (1911) thought it probable that deformation immediately preceded or coincided with the formation of the Rosamond (Barstow) de- posits of the Upper Miocene, and that an arid climate like that of today prevailed during this time. An interval of diastrophism then intervened, followed by a long period of erosion which was succeeded by a time of mountain-making that produced the pres- ent ranges of the desert. Knopf (1916) concludes that the latest major uplift of the Sierra Nevada occurred early in the Pleisto- cene, but Blake (1914) regards the end of the Pliocene as the time of upUft in the Colorado region. Merriam (1919) considers the Ricardo beds to be derived from a Sierra range that, in late Tertiary, rose several thousand feet above the floor of the Mohave Desert. Recendy, Blackwelder (1931) has studied the eastern The Origin of the Desert Climax and Climate 123 slope of the Sierra Nevada to determine the times of Pleistocene glaciation and has constructed a tentative table of glacial stages in the western United States, in which the four recognized in the Sierran region, namely, McGee, Sherwin, Tahoe, and Tioga, are correlated respectively with the Nebraskan, Kansan, lowan, and Wisconsin of the Middle West. Reconstruction of the Clisere or Series of Climaxes Methods. — The concepts and methods of paleo-ecology were first developed in Plant Succession (1916:344-362) and were re- capitulated in Scope and Significance of Paleo-ecology (1918). They have been consistently applied by Chaney and his asso- ciates with striking results in a noteworthy series of monographs on western fossil floras, particularly the Bridge Creek, Mascall, and Goshen (Chaney, 1925, 1933). As for the methods employed, the foremost is that of causal sequence, which involves the basic relation of habitat, plant, and animal, and which may be illus- trated by the presence of Stipa in the Florissant Miocene. This suggests not only the existence of prairie but likewise of a grass- land climate and a grazing population. A similar but even more fundamental sequence begins with deformation and finds suc- cessive expression in gradation, climate, and the biome of plants and animals, the latter regularly exhibiting the final effect. Both sequences are concerned in the method of community bonds in which the reactions and coactions of the component species in the complex organism are intimate enough to serve as the essen- tial basis for reconstruction and interpretation. An important corollary of this is seen in the method of succession, especially in the form of the clisere or succession of climaxes in which climatic changes are at work. The essential feature of this is mass migra- tion, with evolution, which yields the method of phylogeny. l^ 124 E E. Clements An invaluable corollary of these is found in inference from distribution in space and in time. The latter enables us to close many a gap in the fossil record, and the former permits us to fill in the fragmented area of a dominant or community. This at- tains its most striking success in the form of the transad, and especially in that of the relict, where the community relation finds greater expression (Clements, 1934). Finally, an epitome of all other methods is that of cycles, in which it is recognized that all forces and effects exhibit cycles, not merely of deforma- tion, erosion, and climate, but also of mass migration and evolu- tion in the biome. Thus, evidence of elevation carries with it the necessary assumption of climatic shift as well as accelerated erosion, with inescapable effects in the climax series. Conversely, the displacement of a climax presupposes the chain of physical processes of which it is the final consequence. Out of all these ^ . principles springs the axiom that the present provides the surest route into the past and that reconstruction proceeds most re- liably in that direction. The former grassland climax. — The preliminary endeavor to outline the earlier grassland of the desert region was made a decade ago upon the basis of the relict grasses in the Colorado, Mohave, and Death Valley (Clements, 1922, 1923). As indicated previously, almost no perennial grasses survive today on the typical climax level and soil, except transient individuals ascrib- able to wet phases. Even though a product of the desert climate, the shrubby Hilaria rigida becomes dominant only in sand- plains or washes, as is true likewise of the other most xeric species, Triodia pulchella, Stipa speciosa, and Oryzopsis hyme- noides. In the driest part of the desert, where there is a rainfall of about 2 inches, such habitats supply compensation sufficient at least to double the rainfall in effect and to convert the climax of The Origin of the Desert Climax and Climate 125 Larrea-Franseria into a savannah of these with the grasses men- tioned. Toward the margins o£ the desert, where there is a rainfall of 5 inches, as in the vicinity of Goffs and Yucca, sand similarly increases the availability of the water content and supports a grass cover of Hilaria, Sporobolus cryptandrus flexuosus, Muh- lenbergia porteri, and Aristida purpurea. With a corresponding rainfall in che desert generally, these constituted the chief vege- tation, mixed or alternating with Stipa and Oryzopsis and dotted with taller Larrea and similar shrubs. Such a grassland was doubdess able to maintain at least a sparse population of camels, antelope, and horses, but bison were probably absent. Xeric forms of nearly all the species of rodents at present known for the region must have been present in abundance, and this pre- supposes a corresponding balance in carnivores. Still proceeding backward, a further rise in precipitation to from 10 to 12 inches was marked by a prairie denser in texture and richer in species. In terms of the grazing animals, this was even more significant because of the presence of the grama- grasses of the genus BotUeloua. The evidence for this community is drawn from transads found in marginal areas as well as from mountain relicts, which are really transads in climatic terms. Thus, the relict prairie at 4300 feet in the Avawatz Mountains between the Mohave and Death Valley consists of Poa scabrella and Hilaria jamesi, in addition to the more xeric Triodia, Ory- zopsis, Stipa, and Aristida, while at the northern edge on the slopes of the Argus Range occur Hilaria jamesi, Poa scabrella, and Stipa comata. In the district about Cabezon at the outermost edge of the Colorado Desert, a fine example of bunchgrass prairie persisted until it was plowed a few years ago. It was composed of Stipa speciosa, coronata, setigera, and eminens, Poa scabrella, 126 F. E. Clements Elymus sitanion, and Koeleria cristata. Even more eloquent are Bouteloua gracilis and racemosa in the neighboring mountains, and the unique reUct east of San Diego at Jamacha, containing B. hirsuta, radicosa, and rothroc\i. On the plateau that forms the eastern rim of the desert at Kingman, the desert plains consist of Bouteloua eriopoda, B. gracilis, Aristida divaricata, A. pur- purea, Sporobolus cryptandrus, S. c. flexuosus, S. strictus, Hilaria mutica, and Muhlenbergia porteri. Still higher and to the east, at Seligman, where there is a rainfall of 15 inches, most of the species of the desert plains have yielded to the mixed prairie, comprising B. gracilis and racemosa, Stipa pennata and comata, Sporobolus, Hilaria jamesi, and others, w^ith Andropogon scopa- rius, furcatus and, more rarely, nutans and halli in postclimax areas, such as valleys and sandhills. Of the 25 species mentioned above, all but seven are transads and most, if not all, of those seven must have extended some dis- tance into the region concerned. The consequence was a mixed prairie surpassing that of the present in wealth of species and in nutritive value as attested by eight short-grasses (six species of Bouteloua, Hilaria cenchroides, and probably Buchloe dacty- loides) as well as by H. jamesi, which is outranked only by B. gracilis and Buchloe in food value. Naturally, not all the domi- nants grew intermingled, but the cover exhibited various group- ings in accordance with the requirements of the species and the diversity of terrain. In the wet phases of the major climatic cycles, the animals must have reached not only the maximum in num- bers, but also in size, as exemplified by the horses, camels, and elephants of the late Pliocene and early Pleistocene. Among the grass transads are several tall-grasses such as Andropogon saccharoides, A. glomeratus, and Imperata hoo\' eri that are subtropical in nature and derivation. A larger num- The Origin of the Desert Climax and Climate 127 ber of similar species occur westward into central Arizona, but stop short of the desert, namely, Andropogon contortus, nutans, furcatus, and halli, Elionurus barbiculmis, Trachypogon secun- dus, and Tripsaciim lemmoni. All these find their best develop- ment in the southern part of the country and under a rainfall of 35 inches or more, though the first three range widely through the river valleys of the Great Plains, where they persist as relicts of an earlier warm-moist period. In their climatic relations in the United States they are typically postclimax to true prairie and subclimax to deciduous forest. This fact, taken with their occurrence from the South and East to the Canadian border and westward to Arizona or California, leaves little doubt that they constituted the dominants of an earlier tall-grass prairie that confronted the deciduous forests of the Eocene and Lower Oligo- cene and gradually replaced them as they withdrew in the face of cooler drier conditions in the later Oligocene and the Miocene. Their presence in the Southwest, both as transads and as ad- jacent dominants, justifies the assumption that they constituted the grassland that preceded the mixed climax described above and perhaps was still in occupation of much of the region before the Upper Oligocene. The mid-grasses of the ancestral mixed prairie consisted of more xeric dominants, as the name suggests. These are all more northern in derivation and with lower water requirements than the tall-grasses. Some of them, such as Stipa, Oryzopsis, and Hilaria, were apparendy North American in origin, while others were Eurasian. Stipa is the genus paramount of prairie and steppe the world over, and has probably played the major role not only in the evolution but also the intercontinental migra- tion of grazing animals. Its indurated lemma, sharp callus, and long twice-bent hygroscopic awn stamp it as the product of 128 E E. Clements a drying climate and give it distinct advantages over all its competitors in the processes of migration and ecesis. The genus is the final expression of the phyletic line of Muhlenbergia-Ory- zopsis-Stipa, which is predominandy American and originally North American, the first genus having been derived apparently from Calamagrostis, primarily a meadow grass of high latitudes. The mid-grasses were evidently in process of dispossessing the tall-grasses after the deformation of the Upper Oligocene, as shown by fossil Stipa in the Miocene, and probably had reached the Southwest in Lower Miocene in time to play their part in the evolution of the Miocene grazing horses, Merychippus and Protohippus, Further direct evidence of the presence of Stipa and its asso- ciates has been obtained by Elias from deposits of the Ogallala in the Lower Pliocene of western Kansas and adjacent Colorado (1932, 1934). In addition to several forms of Stipa, two other grasses occur, Panicum and Berriochloa, and four forbs of the borage family belonging to the genera Biorbia and Krynitz\ia. It is an axiom that the evolution of a particular grassland must have preceded considerably the adaptation of the graminivor- ous animals to it, just as deformation and climatic change, or sometimes the latter alone, must have antedated the develop- ment or migration of a particular climax. The short-grasses are likewise of American origin, but their homeland was the mountain plateaus of Mexico and Central America. Such far-ranging species as Botiteloua gracilis and racemosa and Buchloe dacty hides have pushed north to the Ca- nadian border or beyond, but the majority are confined to the region west and south of Texas. Aristida, though of somewhat different form, is a regular associate of Bouteloua and of similar southern origin and xeric nature. The two genera constitute the The Origin of the Desert Climax and Climate 129 chief dominants of the desert plains of today and hence the short-grasses of the Great Plains must have found their way into the mid-grass prairies of the Tertiary in a later dry period, per- haps as a result of Pliocene deformation. The short-grasses profit by overgrazing in competition with the taller species, and hence have spread irresistibly to the east during the historical period. They must have behaved similarly in some degree under the pressure of the great ungulate populations of the later Tertiary, in the evolution of which they apparently played a part. They "cure on the ground" and provide forage during the resting sea- son that is far superior to that of the taller forms, their high nutritive value being associated with the unique starch-sheaths of the leaves. In contrast to the mid-grasses, they prospered under warmer drier phases of the climatic cycles, and at such times probably constituted the larger part of the forage. The presence of a well-defined sagebrush climax lying above the grassland on slopes and plateaus, especially toward the north, is attested by the list of transads on page 98, which are domi- nants in this climax today. The most important of these are members of the main association typical of the Great Basin, but the coastal sagebrush of central and southern California is rep- resented by several of the lesser dominants such as Eriodictyutn, Pentstemon, and Viguiera. Two of the most important species, Artemisia californica and Salvia apiana, extend into the western edge of the Colorado Desert, while Eriogonum fasciculatum with its variety polifolium ranges across the desert and well to the east. No fossil remains of this climax have been found in or about the region, a fact readily explained by the general nature of the leaves and the arid habitat. As with the grassland, this community must have moved north and south and up and down the lower mountain slopes in its proper zonal relation in 130 F. E. Clements response to climatic pulsations and, as at present, it must have been the chief resource of the ungulates during periods of drouth. The chaparral climax, representative of the next zone above with higher rainfall, naturally suffered more from the trend toward desiccation and the number of transads is correspond- ingly smaller. However, almost all the important genera except Adenostoma and Heteromeles are represented, mosdy by species transads or by phylads that cross the desert, as in the case of Querciis dumosa, Prtinus demissa, Aesculus californica, and Cer- cis occidentalis. Furthermore, an additional genus, Garrya, is ^ found in the Pleistocene at Rancho La Brea; another, Robinia, in the Pliocene of the Ricardo; and a large number of the actual species occur in one or both of these horizons elsewhere in Cali- fornia. On the basis of phylogeny and association, chaparral must have been in existence in its proper position in the clisere of the Miocene or earlier, but it undoubtedly received recruits from time to time as the climate became drier. Although some of its dominants were undoubtedly utilized by browsers and the nuts and berries by rodents and birds, this community must have played a smaller part in the general economy. In the existing vegetation, woodland takes a m^inor role as a climax, though in the reduced form of savannah it covers a wide area from the Edwards Plateau of Texas to the Coast Ranges and north into Oregon. Its major components are junipers, nut pines, and live oaks, though it may contain other pines and evergreens such as Pinus sabiniana, Umbellidaria, or Arbutus, or, more rarely, such deciduous trees as Quercus douglasi and Aesculus californica. Probably the Joshua tree. Yucca brevifolia, finds its most natural position here also. As species with varieties, both Pinus and Junipei'us were present, as was Cufn-essus; the evi- The Origin of the Desert Climax and Climate 131 dence drawn from transads is confirmed by the occurrence of juniper and cypress at La Brea and Carpinteria and of cypress and nut pine at Ricardo. Quercus agri folia is found in all three horizons, while 0. douglasi and chrysolepis, Pinus sabiniana, Umbellularia, Arbutus, and Aesculus occur in the Pliocene else- where in California, the last three extending at least as far back as the Miocene of the John Day region. In the form of phylads, all these except Pinus sabiniana occur across the desert today, and it is probable that in the Pliocene and wet phases of the Pleisto- cene the woodland was much the same as it now is about the Grand Canyon, in southern Arizona, and in a few relict areas in California. As might be expected, the problem of the extent and composi- tion of former forest in the desert region presents more difficulty. In the bordering ranges, especially, the frequent presence of Pinus ponder osa, P. p. jeffieyi, and Pseud otsuga taxi folia indi- cates that the montane forest was of wider occurrence before the Recent period, but its movement must have been chiefly if not entirely along mountain axes and not across the lowlands. The most plausible assumption is that climax mixed prairie had come into possession in the Lower Miocene and that this had been preceded by a tall-grass community in the beginning subclimax to hardwood forest, a condition visible today along the western border of the oak-hickory association. This border receives ap- proximately 40 inches of rainfall at the present time, but it appears probable that more was needed in the Southwest to compensate for the evaporation caused by higher temperature and a longer season. In view of the fact that Chaney (1925:43) estimates the rainfall of the Mascall at 30 inches during Middle Miocene, a time when the oak-madrono forest was the chief community, it seems certain that it must then have been drier 132 F. E. Clements in the region of the present desert. This assumption is confirmed by the prevalence of Merychippus, and it seems fair to surmise that the precipitation amounted to about 20 inches and supported a grassland similar to but richer than that of central Nebraska today. This region is characterized along the valleys by relict forest, composed almost wholly of flood-plain species, such as willow, Cottonwood, box elder, elm, ash, hackberry, and walnut, though bur oak and linden also reach it, and the more meso- phytic oaks and hickories are not far to the east. In attempting to explain the wide distribution of grassland and the associated ungulates in Miocene time, it is necessary to invoke land connections and to consider the great succession that followed the withdrawal of the Cretaceous Sea in the Laramie or Paleocene. The most satisfactory hypothesis is that the result- ing marshes of vast extent were long peopled with reed-grass and sedge communities, as are the great tule swamps of Cali- fornia today, and that these for the most part yielded to tall- grass meadows in the Oligocene. With the broad connection be- tween the two northern continents at this time, a wide pathway came into existence for the mutual exchange that took place in the Middle Miocene. Had this been limited to Holarctic forms, a mere change of climate might be sufficient explanation, but the movements of certain groups in opposite directions, such as those of the dogs, true cats, catde, and elephants from Asia to North America and that of camels and horses from the latter to the former, presupposes an adequate connection at the north (Osborn, 1909; Matthew, 1915, 1930). The assumption of such an intercontinental grassland highway gains support from the great migration and expansion of Stipa, its associates, and the accompanying horses, tapirs, camels, and other forms in South America in the Pliocene and Pleistocene. [cLEMENTS: ORIGIN OF DESERT CLIMAX AND CI.IMATe] ■■^"»^y.agai'titaf.->TH4- ■„■ -^ Fig. I. a. Desert climax of Lcirrea-Franseria, the leaves killed by the 1934 drouth phase of the climatic cycle. Fig. I. b. Larrea transad relict in the grassland of the upper San Joaquin Valley. [cLEMENTS: ORIGIN OF DESERT CLIMAX AND CLIMATeJ FiiT. 2. a. Postclimax Hilaria rigida in a sandy wash, central Mohave Desert. Fig. 2. b. ReHct Stipu and Aristida among the blocks of a lava ridge, Death Valley. [cLtMhNTS: OKU. IN UK DEStRT CLIMAX AND cLIMATtJ Fig. 3. a. Stipa speciosa, Oryzopsis, and Poa in an open sa\annah of Yiicca-Laneii in Antelope Valley, the western arm of the Mohave Desert. Fig. 3. h. Stipa prairie between Banning and Cabezon at the western edge of the Colorado Desert. [CLEMENTS: ORIGIN OF DESERT CLIMAX AND CLIMATE] Fig. 4. a. Reed-swamp associes of Scirptis, Jitnctis, and Phragmites, Death Valley. Fig. 4. b. Open woodland of Joshua tree {Yucca brevi folia), western Mohave Desert. [cLEMENTS: ORIGIN OF DESERT CLIMAX AND CLIMATE] Fig. 5. a. Washingtonia filifera in canyons of the foothills near Indio, Colorado Desert. Fig. 5. b. Par /{ill son id toncyaiia in valleys of the dune-cw,ii^.;^\, Algodones, Colorado Desert. [cLEMENTS: ORIGIN Ol- DhShRT CLIMAX AND CLIMATtJ Fig, 6. ci. Consociation of Elymiis triticoides in the McKittnclv region, upper San Joaquin Valley. Fig. 6. b. Postclimax tall-grass, Andropogon saccharoides, in a trackway depression, Gila Desert, Arizona. [cLEMENTS: ORIGIN Of DESERT CLIMAX AND CLIMATeJ ^^^4^^^' Fig. 7. ^7. Desert-plains association with Yucca and Ephedra in southwestern New Mexico. «v Fi 'o CO to 164 W. S. Cooper four species of Tanacetum native to North America — the two Pacific species, T. htironense Nutt., ranging from New Bruns- wick and Maine to the Great Lakes and Hudson Bay, and T. bipinnattim (L.) Schultz, occurring from the lower Mackenzie River westward through central Alaska into eastern Siberia, and, according to Fernald (1923), across northern Asia to Russian Lapland. All four species are closely similar. Fernald suggests that T. huronense may fairly be considered a variety of T. bipin- natum. All but T. bipinnatum are rather strictly coastal, except that T. huronense is also found on the shores of the Great Lakes. Species of Stable Dunes Dominants. — Eriogonum parvifolium Sm. (fig. 8^) is an im- portant dominant in the dune shrub community throughout its range, which extends from San Francisco to the Mexican bound- ary. It sometimes strays inland a short distance, and occasionally occurs in habitats other than dune sand, such as bluffs and ter- races. The large genus Eriogonum is confined to western North America, where its members are conspicuous features of the vegetation in semiarid regions. The present species belongs to a group composed mainly of shrubs. A close relative is E. fascicu- latum Benth., one of the most important dominants of the coastal sagebrush climax of southern California. * Lupinus arboreus Sims (fig. 8^) is not confined to dune sand. It occurs in all sorts of sandy places, and is occasionally found at considerable distances inland. It is hardly a typical member of the dune shrub community. It ranges from Humboldt Bay southward to Ensenada, Baja California. Lupinus chamissonis Eschsch. (fig. 8c) is one of the most characteristic dominants in the dune shrub community. It is common from Marin County to southern California. There is (A u to Strand and Dune Flora of the Pacific Coast 171 1925). It belongs to the subgenus (or genus) Hor\elia, which, according to Rydberg (1898), comprises about 50 species, all natives of western North America. The group Cuneatae has nine species, all inhabitants of California and Baja California. Croton californicus Muell. Arg. (fig. 9^) is a very abundant and typical member of the dune shrub community. It ranges along the coast from San Francisco Bay southward to the tip of Baja California. It also occurs on both shores of the Gulf of Cali- fornia. At the northern end of its range it is confined to the im- mediate vicinity of the sea, but farther south it spreads inland. It is frequent in the valleys of southern California, and is re- ported from scattered localities in southern Nevada, Utah, Ari- zona, New Mexico, and Texas. Interior stations east of California are not plotted on the map. In addition to the typical form, which is mainly coastal, Ferguson (1901) recognizes three varieties, mainly of the interior. The genus Croton, numbering more than 600 species, is chiefly tropical. Ferguson recognizes 24 species as occurring in the United States. These range across the southern part of the country from the Atlantic to the Pacific, several ex- tending into Mexico. C punctatus Jacq., a near relative of C. californicus, is a seashore plant upon the south Adantic and Gulf coasts. Monardella crispa Elmer (fig. ^e) is found with the dune shrubs and also on sand that is imperfecdy stabilized. It occurs in the Santa Maria dune region, and at Point Reyes. Monardella is a genus of moderate size centering in California, especially in the dry foothill regions, with a few species in the central Rocky Mountains. Castilleja latifolia H. and A. (fig 9/) is frequent within its range; it occurs also on blufTs, but always close to the sea. It ex- tends from Westport, Mendocino County, south to Point Sur, 172 W. S. Cooper Monterey County, with an outlying station at Casmalia, Santa Barbara County. Castilleja is an important genus in western North America, with 16 species in California (Jepson, 1925). It is represented also in eastern America, northern Asia, and South America. Another member of the genus, C. foliolosa H. and A., though not strictly coastal, is a frequent member of the dune shrub community. Corethrogyne filaginifolia (H. and A.) Nutt. (fig. 9^) is a common and typical member of the dune shrub community. Along the coast it ranges from Bandon, Oregon, southward to the Mexican boundary. It occurs inland also, in the Coast Ranges south of San Francisco Bay, the lower and middle alti- tudes of the Sierras from Amador County southward, and the valleys of southern California. The genus Corethrogyne is considered by Hall (1907) to be monotypic. The species is exceedingly variable. Hall recognizes eight varieties; twenty-five or more so-called species have been segregated by various authors. Certain of these occupy consistent geographic areas and others do not. The genus belongs to a group within the tribe Astereae, which includes such characteristic western American genera as Aster and Erigeron. Artemisia campestris L. subsp. pycnocephala (Less.) Hall (fig. 9^) is a frequent member of the dune shrub community occurring also on sand not sufficiently stabilized for invasion by the shrubs. It is stricdy coastal, ranging from Coos Bay, Oregon, southward to Point Sur, Monterey County, California. The large genus Artemisia is widely distributed over the northern hemis- phere, and the species that have been described from North America are very numerous. In the recent treatment by Hall and Clements (1923), many of these are reduced to units of lesser rank. A. campestris is made up of a complex of forms, quite di- Strand and Dune Flora of the Pacific Coast 173 verse in superficial aspect, which have been described by various authors as separate species. The species as a whole ranges over North America, Asia, Europe, and Northern Africa. Hall recog- nizes six subspecies, all o£ which have consistent ranges. Subsp. pycnocephala is the most distinct and is isolated in range from the others. DISCUSSION Species of Strand and Shifting Dunes geographic relations Of the 25 species of strand and moving dunes, 12 range beyond the region under discussion in various directions, and 13 are endemic. The 12 with extraterritorial ranges fall into five groups. Discussion of these and of the endemics will be followed by generalizations concerning the origins of the strand and shifting dune flora. 1. Species tvith more or less perfect circumpolar distribution, ranging various distances southward along the coasts of the Atlantic and Pacific. Elymus arenarius Lathy rus japonicus Honckenya peploides Mertensia maritima We have no certain knowledge of the preglacial distribution of the species of this group. We may, however, point to certain places where they may have survived during the period of Pleistocene glaciation. The southern parts of their coastwise ranges on the Atlantic and Pacific immediately suggest themselves. In addi- tion, there is no apparent reason why they may not have lived upon the Arctic coasts of America and Siberia. In Alaska, except in the southern part, the lowlands and shores were ice-free. Elymus, Honc\enya, and Mertensia live today upon the shores of Greenland, where the edge of the ice sheet lies very close to i«4 W. S. Cooper the sea. Much o£ the arctic archipelago was unglaciated, but here conditions must have been almost prohibitive in their severity. In Siberia, Kamchatka and the Taimyr Peninsula were glaciated (Woldstedt, 1929), but the arctic coast as a whole was probably free of glacial ice. The extended and fairly continuous ranges of today strongly suggest a wide preglacial distribution and re- advance from several centers. As the ice sheets receded, these species probably moved forward almost as fast as the shores became open to colonization. In Glacier Bay, Alaska, at the present time, Elymus and Mertensia are following less than half a century behind the disappearance of the ice. Honckenya and Lathyrus invade somewhat more slowly (Cooper, 1931). On the Pacific coast of North America there was advance southward from arctic Alaska and northward from Washington. For Mer- tensia it seems almost necessary to assume an arctic center, for its present range ends at the Queen Charlotte Islands. On the Atlantic coast of North America, invasion was probably from the south, and it was doubdess from here that Elymus and Lathy- rus reached the region of the Great Lakes. The arctic coast of Canada may well have been invaded both from the Adantic and from Alaska. In Europe the southern Adantic coast must have been the starting point, although for Elymus and Honc\enya arctic Siberia also may have been a center. For all species except Honc\enya, the present ranges on both sides of the Adantic lie entirely within the glaciated regions, so that these ranges must have shifted bodily northward. This is a perfectly reason- able assumption; the range of Glehnia actually underwent such a movement. Distribution of the species of the group on the Pa- cific coast of Asia may not have been seriously disturbed during the Pleistocene. Elymus and Honc\enya have most nearly re- gained their assumed preglacial territory. There is an apparent Strand and Dune Flora of the Pacific Coast 175 break in Mertensia across arctic Siberia, which may never have been occupied; in Lathyrus there is a gap across arctic Canada as welh It is o£ interest to note that upon the coast of southern Alaska and British Columbia, where glaciation must have completely exterminated plant life, all four species have effected a fairly thorough reoccupation. They grow wherever suitable habitats occur. This has been possible because they are primarily inhab- itants of the upper beach and not dependent upon accumulations of dune sand. All four species have given rise to varietal forms in different parts of their extensive ranges, the most important of which have been noted in a previous section. Doubdess some of these have arisen in connection with the Pleistocene interruption of their ranges; satisfactory treatment of this problem must await more detailed knowledge. 2. Species with close relatives on the Atlantic coast of North America and in central Alaska and eastern Siberia. The two species of Tanacetum comprise this small group. They resemble the species of the first group, especially Elymus and Lathyrus, in being represented by very closely allied forms on the Atlantic coast of North America, the Great Lakes, and in Alaska and eastern Siberia. The group, however, does not extend to Europe and is not present on the arctic and subarctic coasts of America. Moreover, the species of Alaska and Siberia is not coastal. The point of origin cannot be determined, but it seems clear that a formerly continuous range was broken up by gla- ciation. The present-day forms may well have arisen through isolation. 3. Species occurring also on the coasts of eastern Asia. This group comprises two species with remarkably similar dis- 176 W. S. Cooper trjbution: Carex macrocephala and Glehnia littoralis. Their ranges on the American and Asiatic coasts correspond rather closely; both are absent from the Alaska Peninsula and the Aleutian Islands. On the American coast both are abundant south of the glacial boundary and occur on Vancouver Island. North of there, they have been found in only a few scattered localities, partly because exploration has been incomplete, but principally because suitable habitats along this coast are few and widely separated. Both species require accumulating sand. Northward invasion of the glaciated region has thus been diffi- cult and slow. The fossil record of Glehnia 400 miles south of its most southerly station of today, in connection with its invasion of the glaciated area, proves that the range of this species has shifted bodily northward in postglacial time. It is reasonable to assume that the same thing has happened to other plants, for example: Carex, Tanacetum, Elymus, Honc\enya, Lathyrus. The ultimate origin of Carex macrocephala, with no near rela- tive in its very cosmopolitan genus, is veiled in mystery; Glehnia seems quite certainly to be descended from American stock. Another parallel between the two is that both have developed, under isolation, into pairs of forms, Asiatic and American, that may be specifically distinct. 4. Species occurring also in South America. Juncus lescurii Fragaria chiloensis Carpobrotus chilensis (also Lathyrus japonicus) Though similar in present range, these species have had quite diverse histories. Juncus is an almost cosmopolitan genus, but /. balticus, probably the closest relative of /. lescurii, is mainly of the northern hemisphere. Fragaria chiloensis also belongs to a Strand and Dune Flora of the Pacific Coast 177 genus that is typically northern. The same is true of Lathyrus japonicus, which, although found in South America, falls more naturally into group i. Carpobrotus chilensis, on the contrary, belongs to a group centering in South Africa which has spread eastward and westward, finally reaching our territory by way of South America. It is the only species that has certainly entered from that direction. 5. Species of general distribution upon subtropical coasts. Convolvulus soldanella, the most widely distributed of all our strand species, is, judging from the general range of the family, of subtropical origin. 6. Endemics, The species of this group, confined to the area under discus- sion, show various extraregional affinities. The following eight species are quite definitely derived from stock which has long inhabited the more or less arid parts of western America: Atriplex leucophylla Lupinus littoralis Abronia latifolia Oenothera cheiranthifolia Abronia maritima Franseria bipinnatifida Abronia umbellata Franseria chamissonis The remaining five represent genera mainly north temperate in distribution, present in western America but not characteristic of the adjoining arid regions: Poa confinis Polygonum paronychia Poa douglasii Lathyrus littoralis Poa macrantha From the analysis here presented, the following generaliza- tions concerning the origins of the flora may be made. Fourteen species of eleven genera, some endemic, some with 178 W. S. Cooper extraterritorial extensions of range, are of subarctic or north temperate origin : Poa confinis Honckenya peploides Poa douglasii Fragaria chiloensis Poa macrantha Lathy rus japonicus Carex macrocephala Lathyrus UttoraUs Elymus arenarius Mertensia maritima Juncus lescurii Tanacetum camphoratum Polygonum paronychia Tanacetum douglasii Nine species of six genera, all (except Glehnia) endemics, are derived from the neighboring arid and semiarid hinterland : Atriplex leucophylla Oenothera cheiranthif olia Abronia latif olia Glehnia littoralis Abronia maritima Franseria bipinnatifida Abronia umbellata Franseria chamissonis Lupinus littoralis The remaining two species have entered from South America or the subtropics: Carpobrotus chilensis Convolvulus soldanella FEATURES OF LOCAL DISTRIBUTION Species density. — Maximum species density occurs from mid- dle Oregon to Mendocino County, California. Within this stretch 22 of the 25 species occur, with a maximum number of 19 at any one point. The only species not present within this strip are Carex macrocephala, Abronia maritima, and Mertensia maritima. Northward we find a species density of 18 as far as Cape Flattery, beyond which there is a sharp decline to 8 at the center of Vancouver Island. Farther north there is a very gradual decrease to a minimum of 5 in southern Alaska. South from Mendocino County, decrease is fairly regular. The only excep- tion is in the vicinity of Monterey Bay, where there is a sudden Strand and Dune Flora of the Pacific Coast 179 drop from 14 to 10. At the Mexican boundary the species density is 7; at Socorro, 5; and beyond Bahia de Sebastian Vizcaino but I species, Abronia maritima, remains. As far as Socorro, no com- pensating strand plants appear to replace those dropping out, except that Ailenrolfea occidentalis (Wats.) Ktze. and species of Salicornia tend to become builders of foredune hillocks. Certain tropical strand species appear farther south— for example, Ipo- moea pes-caprae — but the extent and character of the strand flora of the Mexican coasts is not at present ascertainable. In more general view, the region of high species density may be considered as extending from southern Vancouver Island to Monterey Bay. Yet this is by no means the region where the greatest number of individuals of the type appear. Foredune hillocks, accurate indicators of the number and vigor of these species, are almost absent on the Oregon coast. They become gradually more evident southward, and attain their greatest development in southern California, where species are compara- tively few but individuals are abundant. Two factors may provide an explanation. Constant and severe summer winds in the north are inimical to foredune formation both because they make accumulation of sand difficult and be- cause they increase transpiration from the plants. Again, in the north, frost kills back the plants in winter, so that accumulations made during the summer are dissipated, whereas in the south growth is continuous and the hillocks constantly tend to increase in size. Particularly important in the south are the frost-sensitive succulents Abronia, Carpobrotus, Convolvtdus, and Franscrta, which there are able to flourish throughout the year. Effects of barriers. — For the most part, upon the coast of the United States and Baja California the strand-dune environment occurs in strips of varying length separated by bluffs, cliffs, bays, i8o W. S. Cooper and rivers. Most of the interruptions are short and obviously in- effective as barriers to coastwise migration. It is conceivable that a break in the environment of considerable magnitude might be effective. The Santa Lucia mountain range constitutes such a possibility, because it presents a precipitous face to the sea fifty miles in length, with very few and insignificant accumulations of sand. At the northern end five species reach their southern limits — Poa douglasii, Ely m us arenarius, Polygonum paronychia, Lathyrus littoralis, and Franseria chamissonis. A similar but less imposing stretch of cliffs in southern Humboldt and north- ern Mendocino counties has no effect at all. North of the Canadian boundary the situation is very differ- ent. Here, owing to the abruptly mountainous coast and over- steepening of slopes caused by glaciation, suitable habitats for strand plants are the exception rather than the rule; they are of limited extent and separated by wide distances. The abrupt drop in species density near the glacial boundary is surely significant. Species from the south have invaded or reinvaded the glaciated area for a slight distance only or not at all. No more than two species dependent on accumulating sand — Car ex and Glehnia — have gone far north, and these, so far as collections show, have established themselves at only a few points. Elymus, Honc\enya, Fragaria, Lathyrus, and Mertensia, requiring merely beach con- ditions, are much more thickly distributed. Fewness and wide separation of favorable habitats, together with shortness of avail- able time, has certainly been effective in hindering northward migration. Climate can have borne no causal part in this, for there is no sudden change in any factor corresponding with the abrupt drop in species density. Climate, however, must be effective in a more general way in limiting the ranges of species, here as elsewhere. Freezing tem- Strand and Dune Flora of the Pacific Coast i8i peratures doubtless set potential limits to northward migration of sensitive southern species, but south of the glacial boundary, as I have pointed out, the effect of this factor is merely to con- vert potential ever-growers into winter-dormant perennials. In- creasing dryness southward may be a factor in the eUmination of certain species. Those ranging farthest south are superficially the most xeric. COMPARISON WITH THE ATLANTIC COAST OF NORTH AMERICA Two facts stand out clearly : likeness between the floras of the two coasts in the north, and unlikeness in the south. All four species in the subarctic group are present also on the northern Adantic coast, together with a closely allied species of Tana- cetum. Three of these five range farther southward on the Adantic coast: Mertensia, io°; Honckenya, 7°; Lathyrus, 1°. Elymus goes 5° farther south on the Pacific coast. The ranges of Tanacetum on the two coasts are closely similar. There are but two important strand plants on the northern Adantic coast that are not native on the northern Pacific, and both of these have become naturalized there. Cakjle edentula already behaves like a native, and Atnmophila arenaria is fast making itself thor- oughly at home. Four species of the northern Pacific coast are not native upon the Adantic: Carex macrocephala, Juncus les- curii, Fragaria chiloensis, and Glehnia littoralis. It is of interest that the Asiatic form of Carex macrocephala has become estab- lished in one or two places on the New Jersey coast and is spread- ing rapidly (Fernald, 1930). The general strand flora of the southern Adantic coast differs absolutely from diat of the southern Pacific. Five gtvitr^— Carex, Polygonum, Atriplex, Croton, Oenothera— occur in common, but the parallel is no more than a coincidence. 1 82 W. S. Cooper Species of Stable Dunes Of the eighteen species in this group, representing sixteen genera, all but two are endemic. Cardionema ramosissima occurs also in western South America; Artemisia campestris var. pycnoce- phala is excluded from the list of endemics solely because of the wide concept of the species here adopted. The group is not so stricdy coastal as are the strand species. Nemacaulis dentidata, Mucronea californica, and Cardionema ramosissima occur also in the lowlands of southern California, Corethrogyne filagini- folia is found there and in the south Coast Ranges and Sierras, and Croton californicus in the Great Basin and eastward to Texas. Six of the genera it^xts^Tittd—Eriogonum, Eriophyllum, Nemacaulis, Mucronea, Monardella, Corethrogyne— 2.1^ con- fined to western North America. If a narrower view of ge- neric limits were adopted, Hor\elia (Potentilla) and Ericameria (Haplopappus) would be added to this group. Lupinus and Cas- tilleja, though not confined to western Nordi America, attain their greatest development there. Atriplex, Croton, and Arte- misia are widespread, but all are abundandy represented in west- ern America. Chorizanthe and Cardionema occur also in South America. It is quite evident that almost all members of the group are of local origin. Differentiation within certain important genera of the arid interior has given rise, along the coast, to forms more or less stricdy confined to the ocean shore. Two genera, Nema- caulis and Corethrogyne, are monotypic; in the latter, however, there has been differentiation into a multitude of minor forms more or less distinct. Both are so closely connected with related genera that we are justified in assuming local origin. All but one species of the group must therefore be placed with the list of nine Strand and Dune Flora of the Pacific Coast 183 species of strand and shifting dunes to which local origin has been assigned. Cardionema ramosissima alone, like Carpobrotiis chilensis, seems to have originated in South America. SUMMARY The flora of strand and dunes on the Pacific coast of North America comprises two groups distinguished by environmental preferences: (a) species of strand and shifting dunes; {b) species of stabilized dunes. The latter constitute a distinct community only on the southern portion of the coast; the corresponding vegetation in the north is not of a strictly maritime type. A province such as the one here treated, linear in form, fre- quendy interrupted, unstable in the extreme, cannot be expected to be a center of evolution, to give rise to a distinctive flora, to become a focus for centrifugal migration. Instead, the population of such a province is made up in part of species differentiated from neighboring inland stock in response to the special condi- tions of the shore, and in part of species that have migrated, mainly along the shore, from other maritime provinces where they have been differentiated from inland stock. Thus, on the Pacific coast of North America we find first, a group of 26 species definitely derived from stock of the local hinterland, representing either genera confined to the region or local sections within widespread genera. This group includes one-third of the species of strand and moving dunes and prac- tically all those of stabilized dunes. The genera represented are mainly characteristic of arid and semiarid habitats. Of the re- maining 17 species, 14 (all of strand and moving dunes) are of subarctic or north temperate origin. Most of the genera which these species represent occur in the neighboring regions inland, but are widespread, and their local representatives are not closely 184 W. S. Cooper related to the maritime species. Three species (2 of shifting, i of stabiHzed dunes) are derived from the subtropics and South America. Considering the strand and shifting dune flora alone, species density is greatest in southern Oregon and northern CaUfornia. This center does not correspond with the region where the great- est number of individuals occurs, which is in southern Cali- fornia. In the unglaciated portion of the coast, physiographic barriers to migration have been of slight importance. In the glaciated region, fewness and wide separation of suitable habi- tats, combined with shortness of available time, have greatly hindered reinvasion. The strand floras of the Atlantic and Pa- cific coasts of North America are strongly similar in the north and entirely different in the south. BIBLIOGRAPHY* Brown, N. E. 1925. Mesembryanthemum and some new genera separated from it. Garden- er's Chron., 78:433. 1928. Mesembryanthemum and allied genera. Jour. Bot., 66:322-327. BUCHENAU, Fr. 1906. Juncaceae. In Engler: Das Pflanzenreich, IV, 36. Chaney, R. W,, and Mason, H. L. 1933. A Pleistocene flora from the asphalt deposits at Carpinteria, California. Carnegie Inst. Wash. Publ., 415:45-79. Cooper, W. S. 1931. The seed-plants and ferns of the Glacier Bay National Monument, Alaska. Bull. Torrey Bot. Club, 57:327-338. DONAT, A. i93i.Ueber Pflanzenverbreitung und Vereisung in Patagonien. Ber. deutsch. Bot. Ges., 49:403-413. Ferguson, A. M. 1901. Crotons of the United States. Rep. Mo. Bot. Gard., 12:33-73. Fernald, M. L. 1909. The variations of Arenaria peploides in America. Rhodora, 11:109-115. 1923. The native tansy of Newfoundland. Rhodora, 25:13-16. 1930. Carex macrocephala and C. anthericoides. Rhodora, 32:9-11. 1932. Lathyrus japonicus versus L. maritimus. Rhodora, 34:177-187. Goodman, G. J. 1934. A revision of the North American species of the genus Chorizanthe. Ann. Mo. Bot. Gard., 21:1-102. GURKE, M. 1897. Borraginaceae. In Engler und Prantl: Die natiirlichen Pflanzenfamilien, IV, 3, a:7i-i3i. Hall, H. M. 1907. Compositae of southern California. Univ. Calif. Publ. Bot., 3:1-302. 1928. The genus Haplopappus. Carnegie Inst. Wash. Publ., 389. Hall, H. M., and Clements, F. E. 1923. The phylogenetic method in taxonomy. Carnegie Inst. Wash. Publ., 326. Hoffmann, O. 1894. Compositae. In Engler und Prantl: Die natiirlichen Pflanzenfamilien, IV, 5=87-387- * Limitation of space forbids citation of the numerous publications from which distributional data have been derived. [185] 1 86 W. S. Cooper Holm, T. 1922. Contributions to the morphology, synonymy, and geographical distribu- tion of arctic plants. Rep. Canadian Arctic Exped., 1913-18, V: Botany. Ottawa. HuLfiN, E. 1927-30. Flora of Kamchatka and the adjacent islands. Kungl. Svenska Vet. Akad. Handl., Ser. 3, Bd. 5, nos. i, 2; Bd. 8, nos. i, 2. Jepson, W. L. 1925. A Manual of the Flowering Plants of California. Berkeley, Calif. MacBride, J. E 1916. Notes on certain Borraginaceae. Cont. Gray Herb., Harvard Univ., n.s. 2, 48:39-58. Mackenzie, K. K. 1 93 1. Cariceae. In North American Flora, 18, pt. 2. Macloskie, G. 1903-06. Flora patagonica. Rep. Princeton Univ. Exped. to Patagonia, 1896- 99, 8, pts. I, 2. Macoun, J. M., and Holm, T. 1921. Vascular plants. Rep. Canadian Arctic Exped., 1913-18. V: Botany. Ottawa. Macoun, John 1883-90. A Catalogue of Canadian Plants. Montreal. Mathias, Mildred E. 1928. Studies in the Umbelliferae. I. Ann. Mo. Bot. Gard., 15:91-108. 1930. Studies in the Umbelliferae. III. Ann. Mo. Bot. Gard., 17:213-476. Munz, p. a. 1928. Studies in Onagraceae. II: Revision of North American species of sub- genus Sphaerostigma, genus Oenothera. Bot. Gaz., 85:233-270. Peter, A. 1897. Convolvulaceae. In Engler und Prantl: Die natiirlichen Pflanzenfamilien, IV, 3, a: 1-40. Reiche, K. 1 896. Flora de Chile. Santiago. 1907. Grundziige der Pflanzenverbreitung in Chile. In Engler und Drude: Die Vegetation der Erde, VIII. Rydberg, p. a. 1898. A monograph of the North American Potentilleae. Mem. Dept. Bot., Columbia Univ., II. 1915. Carduaceae (Helenieae, Tageteae). In North American Flora, 34:81-180. 1916. Carduaceae (Tageteae, Anthemideae). In North American Flora, 34:181- 288. 1922. Carduales (Ambrosiaceae, Carduaceae). In North American Flora, 33, pt. 1:1-110. Strand and Dune Flora of the Pacific Coast 187 Skottsberg, C. J. F. 1910-1916. Botanische Ergebnissc der schwedischen Expedition nach Pata- gonien und dem Feuerlande, 1907-09. Kungl. Svenska Vet. Akad. Handl., Bd. 46, no. 3; Bd. 50, no. 3; Bd. 56, no. 5. Standley, p. C. 1909. The AUioniaceae of the United States, with notes on Mexican species. Cont. U. S. Nat. Herb., 12:303-389. 1918. AlHoniaceae. In North American Flora, 21, pt. 3:171-254. St. John, H. 1915. Elymus arenarius and its American representatives. Rhodora, 17:98-103. VOLKENS, G. i893.Chenopodiaceae. In Engler und Prand: Die natiirlichen Pflanzenfam- ilien. III, i, a:36-9i. WOLDSTEDT, PaUL 1929. Das Eiszeitalter: Grundlinien einer Geologie des Diluviums. Stuttgart. The Genetic Phytogeography of the Southwestern Pacific Area, with Particular Reference to Austraha By Ludwig Diels FROM THE TIME o£ J. D. Hookcr's famous Introductory Essay, On the Flora of Australia, its Origin, Affinities, and Distribution (1859), it has been known that a large, Autochthonian element was a prominent feature of the Austral- ian flora, and that there were strong Indo-Malayan affinities and some relations to the antarctic regions. This evidence appeared to favor the assumption that Australia had been isolated for a long time, that it had had early connec- tions with the Malayan Archipelago, and that more recendy it had been influenced by the south. Some years ago, however, this doctrine was challenged by Wegener in his well-known theory of the displacement of the continents. According to his idea, there was, in Palaeozoic times, a univer- sal continent Pangaea. In the middle Jurassic, Antarctica-Aus- tralia began to separate from this continent and to drift south- eastward. It was still connected with South America in the early Tertiary. Not until the older Quaternary did Australia separate from Antarctica and move northward, and, even at that time, a deep sea existed between Australia and the Malaysian Archi- pelago. Now, if Australia had been joined to the southern regions for the long time from the beginning of the Tertiary period to the first part of the Quaternary, namely, during the epoch of the evolution of Angiosperms, the southern element would be domi- nant in the Australian flora. Let us see whether this is the fact. [189] 100 Ludwig Diets This southern element contains two layers of somewhat difler- ent character. The first is the Antarctic element sensu strictiore. It is composed of microthermic plants growing in southernmost South America, on the subantarctic islands, and on the higher mountains of New Zealand, Tasmania, southeastern Australia, and New Guinea. These species have been listed by Hooker, Engler,Albofl[,Skottsberg,and others. Skottsberg,the last author to treat them exhaustively, finds 49 natural orders represented in the Antarctic class; it is to be kept in mind, however, that the affinities of all these plants are not yet setded. It is an important fact that nearly all the Antarctic species in Australia have been found in the southeastern part (including Tasmania) only, and that these species are represented also in New Zealand, at any rate by vicarious forms. Both countries seem to have been fed from the same regions, situated between S Lat. 50° and 60° at a period when access was easier than at present. Another layer of the southern element contains species which need more warmth than the true Antarctic plants. They are present in lower altitudes and latitudes of Australia, New Guinea, New Caledonia, and New Zealand. Because they are taxonomically more distinct from one another than are the true Antarctic species, they are considered older types. Examples are offered by Araucaria, Podocarpaceae, Cupressaceae, Monimi- aceae, Escallonioideae, and Cunoniaceae. The area of a number of these types is similar to that of the Antarctic species so far as they are restricted to eastern Australia. Many genera, however, spread to New Guinea, Melanesia, and often to the Malayan islands or even more distant regions. Others are distributed all over the Australian continent (Callitris), or are found in south- western Australia (for example, Podocarpus, Aphanopetalum). From these facts it is obvious that these genera are no recent Phy to geography of the Southwestern Pacific 191 immigrants, but old inhabitants of the continent. They belong to the Autochthonian element of the Australian flora. Their ex- istence proves that the Autochthonian flora included members of groups which, at the present day, are distributed principally in the southern hemisphere, but it is impossible to infer that they originated in the south. A great many genera of the Autochthonian element of the flora of Australia (often including New Guinea, Melanesia, and New Zealand) do not display any clear affinities, namely : Tremandraceae Liliaceae § Johnsonieae Liliaceae § Lomandreae Liliaceae § Calectasieae Liliaceae genus Hewardia Haemodoraceae genus Haemo- Euphorbiaceae § Stenolobeae Sapindaceae § Dodonaeaceae Stackhousiaceae Thymelaeaceae genus Pimelea dorum Amaryllidaceae § Conostylideae Passiflora § Disemma Amaryllidaceae genus Campy- Myrtaceae § Chamaelaucieae nema Orchidaceae § Diurideae Santalaceae § Anthoboleae Proteaceae § Persoonioideae *Balanops *Eupomatia *Himantandra Halorrhagaceae genus Hal- orrhagis Umbelliferae § Xanthosiinae Epacridaceae Scrophulariaceae genus Duboisia Scrophulariaceae genus Antho- cercis Labiatae § Prostantheroideae Saxifragaceae genus Eremosyne Rubiaceae genus Opercularia Saxif ragaceae genus Tetracar- Goodeniaceae paea Saxifragaceae genus Bauera *Byblis Rosaceae genus Stylobasium Leguminosae § Podalyrieae *Cephalotus Stylidiaceae Compositae genus Olearia Compositae genus Calotis Compositae genus Lagenophora Compositae genus Brachycome Compositae § Angiantheae 192 Ludwig Diets This list contains not only isolated monotypic genera (*), but also polymorphous groups. We do not know where to look for their affinities, and this corroborates the opinion that they were separated from other tribes at an early date. Hence it appears that Australia, where they were evolved, was isolated for a consider- able length of time and not connected with southern regions dur- ing this important period of its history. Thus, for the southern element of the Australian flora, I find : (i) That in eastern Australia, several true Antarctic genera and species are found which may date from late immigrations. (2) That only a very few other southern threads appear in the tissue of the Autochthonian flora. (3) That in the overwhelming majority of this Autochthonian flora, no southern relations are indicated. These data are not compatible with the principal assumption of Wegener's theory that Australia had drifted to the north quite recendy, and had approached the continental block of south- eastern Asia only yesterday. In order to test further the value of this assumption, it will be useful to examine the Asiatic or Malesian elements in the flora of Australia. Such Asiatic features present themselves in the more distant relations of some Autochthonian genera, namely : Casuarina Eucalyptus Chenopodiaceae § Kochiinae Sterculiaceae § Buettnerieae Chenopodiaceae § Atriplicinae Myoporaceae genus Eremophila Acacia § Phyllodineae Rutaceae § Boronieae In Casuarina and Acacia § Phyllodineae, it is doubtful whether the species indigenous in Malaysia originated in Australia, or, on the contrary, whether the Australian species originally came from the north. In other genera (Rutaceae § Boronieae, Eucalyp- Phytogeography of the Southwestern Pacific 193 tus),theevidenceis in favor of the northern origin, the Boronieae being speciaUzed Xanthoxyleae, and the genus Eucalyptus a progressive derivative of the Malesian Syzygium type. The argument deduced from the genetic resemblance of such genera is strengthened by a great many examples of relations which, in analyzing the flora, we recognize as Malaysio-Aus- tralian. These relations are found not only in orders peculiar to the Palaeotropics (for example, in Pandanus and Musa), but also in many orders common to Asia and America. Even here the Australian genera or species are more closely allied to the Asiatic forms. This is true for : Cycadaceae Monimiaceae § Mollinedieae Palmae Menispermaceae § Cocculinae Araceae Simarubaceae Dioscoreaceae Loganiaceae Zingiberaceae Gesneriaceae Anonaceae The same feature is to be observed in many endemic genera of the Australian flora. From this fact it is evident that these relations are not attrib- utable to recent events, but to conditions of a more remote period. This is confirmed by the extension of, for example, Pitto- sporaceae, Pritzelia, and Dioscorea hastifolia to the central parts of the continent and to southwestern Australia. Therefore, the connection between Malaysia and Australia must have lasted for a long time in different forms, and may con- tinue to the present day, at any rate in species with efficient means of distribution. As a consequence, this Malaysian compo- nent comprises groups of each degree of classification. There are numerous genera distributed from southern India to eastern Australia, and even many species are found everywhere from the iQ^ Ludwig Diels Malay Peninsula at least to Queensland. Not a few of them may have increased their area by the help of man, spreading with him to the east. But this cannot apply to all; we must assume that such genera as Nepenthes and Terminalia, and such orders as Sapin- daceae, Asclepiadaceae, and Rubiaceae, reached the boundaries of their extension by the natural means of propagation, even though this happened perhaps only a short time ago. From the character of the Malaysian elements, it is apparent that there were connecting pathways from early times. We know that the situation and the condition of these pathways changed, and that they were impassable at certain times and in certain sections. For this reason, the relations between Malaysia and Australia do not appear in all groups of animals and plants in equal manner and in equal strength. But there is no doubt that the degree and the variety of those relations do not agree with the assumption of the displacement theory that Australia entered into connection with Asia only at a late date. If that hypothesis were correct, no peculiar Australian groups of Malaysian affinity would exist. The Malaysian component would prove to be modern. As a matter of fact, such peculiar groups are by no means missing. The whole Malaysian component is not modern, but one of its parts is. These facts cannot be explained by the notions of the displacement theory. In the modern time of earth's history, during the evolution of Angiosperms, we must assume that Australia had a position similar in relation to the other great continents to that which it holds today. An independent evolu- tion of old groups must have been going on in Australia. Besides, influences of southeastern Asia would have commenced early and would be going on up to the present time, whereas the ex- tensions of the Antarctic influences would have affected Aus- traha only temporarily. The Role of the Terrestrial Alga in Nature By F. E. Fritsch THE ESSENTIALLY AQUATIC NATURE o£ the plants classcd as Algae is likely to cause one to lose sight o£ the fact that a very considerable number of these plants lead a ter- restrial existence under conditions of life that differ very pro- foundly from those obtaining in the aquatic habitat. The major- ity of the terrestrial Algae are of relatively lowly organization, many being unicellular, colonial, or exhibiting a simple filament- ous habit, and it is only in Trentepohlia and in some of the Blue- green Algae (Stigonemataceae) that indications of a higher differentiation are seen. No terrestrial alga attains to the com- plexity of structure and of reproductive methods met with in the advanced Phaeophyceae and Rhodophyceae. The terrestrial algal flora is essentially recruited from the classes Chlorophyceae, Xanthophyceae (Heterokontae), Bacillariophyceae (Diatoms), and Myxophyceae (Cyanophyceae), although some species of Porphyridium (Bangiales) are terrestrial and Pascher^** has described a terrestrial member of the Chrysophyceae {Geo- chrysis) . Broadly speaking, three groups of terrestrial Algae can be distinguished: (i) those which are found between the particles beneath the surface of the soil (subterranean community ),'^ (2) those which live on or just beneath the soil surface (surface community), and (3) those which live on other terrestrial substrata — rocks, tree trunks, etc. — in part well raised above the soil level (aerial community). No sharp line can be drawn be- tween these three communities, but on the whole they are ex- * Superior figures refer to items in the bibliography at the end of this essay. [ 195 ] 196 E E. Fritsch posed to increasing liability of desiccation. Nor is it possible to demarcate them rigidly from mud-inhabiting forms such as Botrydium, Protosiphon, etc., whose normal vegetative phase is, however, developed only in the presence of considerable quantities of moisture.* A rather special community is that which occurs amid mosses in diverse habitats and consists, in large part, of Desmids"^ and Diatoms,"' ^ the latter diflFering rather markedly according to the degree of dryness of the habitat. The Subterranean Community The subterranean community, which is found in the water con- stituting the films around the soil particles and occupying the smaller interspaces between them, comprises manifold species, many of which have been recorded as occurring at considerable depths.'" "' ^'' ^' " In the present state of our knowledge, this com- munity appears to be composed of a number of distinct elements. First, it includes certain common members of the surface community. A recent investigation of naturally occuring British soils from 2- and 8-inch depths, conducted in my laboratory by Miss J. James, has shown the universal presence of Chlorococcum humicolum and species of Hormidium,-\ both of which are typ- * Yet, West's "association of irrorated rocks" (West,*^ p. 422) can hardly be regarded as terrestrial because the forms involved, although some are terrestrial, are probably normally well supplied with liquid water. t Here should also be included moss protonema,^ which appears almost in- variably in soil cultures. Bristol^ records Ulothrix subtilis var. variabilis as a com- mon subterranean alga, although it has not been recorded by other investigators and no species of Ulothrix appeared in Miss James's cultures. I am not clear as to the nature of this species of Ulothrix, but strongly suspect that further study will show it to be one of the terrestrial Hormidiums. Similarly, I am inclined to think that the older records of the occurrence of species of Pleurococcus in the subterranean flora are a result of confusion with such forms as Pleurastrum and possibly with some of the unicellular Xanthophyceae (Pleurochloris, etc.), which recent research has brought to our notice. I am even doubtful whether Pleuro- coccus occurs as a member of the surface community. The Role of the Terrestrial Alga in Nature 197 ical surface forms. The species of Chlorella, Coccomyxa, Dac- tylococcus, and Stichococcus found in the subterranean com- munity, although not so universally present, probably belong to the same element. Blue-green Algae (especially the species of Phormidium) and the majority of the Diatoms should probably also be included here.^ Second, there are found in the subterranean community a small number of forms which, so far as present evidence goes, are more or less characteristic of this habitat; among these are Gongrosira terricola, and certain Yellow-green forms {Bo- try diopsis, Pleurochloris, Monodus, Monocilia, Bumilleria, etc.), as well as, perhaps, a few of the Diatoms. Divers of these forms have not so far been found except in soil cultures or commonly, at least, have been met with only in such cultures, suggesting the subterranean regions of the soil as the usual habitat. The third element is what I should like to describe as the casual one (Bristol,' p. 570), including species of Chlamydo- monas, AnJ^istrodesmus, Vaucheria, etc., which are only of occa- sional occurrence and probably owe their presence in the soil to wind-borne resting stages which fall on the surface and after- ward are washed down by rain (see below, p. 201; also Fritsch"^). A problem of considerable interest concerns the distribution of Algae in a soil. The first serious attempt to determine this quantitatively was made by Bristol,^ who employed a method of progressive dilution of a soil suspension, the number of organ- isms in each gram of soil being then computed statistically.' She arrived at the conclusion that there were no appreciable diflFer- ences in the horizontal distribution of soil Algae in a uniform soil, but that there were marked quantitative differences in the vertical distribution. She found an abundant algal population 198 E E. Fritsch in the top inch* and an equal or even greater number in the fourth inch, whereas samples from the two-inch and three-inch depths and from depths below four inches contained far smaller numbers. Such qualitative differences as were observed were quite irregular and appeared to be of no general significance. Moreover, by manurial treatment the proportion of the diverse Algae present can be materially altered."" Bristol," p. 572, points out that the four-inch depth is known to be that of the greatest biological activity of most soil organisms, but it may be doubted whether such a conclusion, which refers to Bacteria, Protozoa, etc., can be legitimately extended to the pigmented Algae. Except in the immediate neighborhood of the soil surface, the Algae inhabiting the soil must be in complete darkness, at cer- tain times and, although in cultivated soils agricultural opera- tions will tend to bring the deeper-living forms nearer to the surface, there is litde that is apt to cause such a transference in natural soils. The activities of earthworms and other animals are perhaps more likely to lead to the destruction of the Algae than to their transference (cf ., however, Esmarch,^ p. 253) . It is a well- known fact that many of the simpler Algae are capable of living in prolonged darkness and even of remaining green under these circumstances, provided that suitable organic nourishment is available. One of the most valuable studies in this respect is that of BristoF on a species of Scenedesmus (S. costulatus Chod. var. chlorelloides Bristol-Roach) isolated from the soil. She showed that, under conditions permitting normal photosynthesis, this alga carries out no heterotrophic nutrition, whereas in dimin- ished light intensity an increasing amount of carbohydrates is utilized and in darkness this of course is the only source of carbo- hydrate nourishment. However, the capacity for growth is less; • It is not quite clear whether this included the actual surface layer. Phytogeography of the Southwestern Pacific 199 it diminishes progressively with increasing heterotrophic nutri- tion and in darkness is only about one-half that observed when full photosynthesis is maintained. As Bristol points out, there are likely to be considerable differences between different Algae in this respect^ and, indeed, it is known that some Green Algae prosper better when organic nutriment is available than when it is not. It is therefore credible that subterranean Algae will find a means of sustenance and even of active growth when buried in the darkness of the soil, but there are two considerations that must be borne in mind in this connection (cf. also Petersen,^ p. 16). First, the Algae will be competing for organic nourish- ment with the other living organisms of the soil and it seems unlikely that this competition usually will be a successful one. Second, there is no evidence so far that the kinds of organic sub- stances that have proved efficacious in the artificial cultivation of soil Algae are available in the soil, and there is therefore no direct proof that an alga buried in the soil in darkness will find the means of sustenance. This problem has recently been approached by Petersen,"^ who has endeavored to ascertain whether two soil- Algae {Pleurochloris magna and Nitzschia Ktitzingiana f. ter- restris), inoculated into sterilized soil, exhibited any capacity for growth in the dark. His tentative conclusions are in the negative and, although the process of sterilization may conceivably have induced alterations in the nature of the organic material avail- able, the experiment certainly affords some evidence that these two soil forms do not find suitable material for sustenance in the soil in the absence of light. Esmarch,^^ p. 259, also failed to obtain any evidence of growth in Blue-green Algae buried in the soil. These investigations show that, until definite proof has been obtained for a number of the regularly occurring soil Algae, 200 E E. Fritsch it is not wise to conclude from experiments under laboratory conditions that such forms will be able to grow and multiply when buried in the soil, although it may well be that at certain times conditions for active growth may obtain. For diverse reasons I am inclined to believe that the subter- ranean Algae are in large part, if not entirely, surface-growing forms that are washed down into the soil by heavy downpours of rain. Thus buried, they may perhaps occasionally find the necessary means of sustenance and so multiply for a time, but I think it is at least doubtful whether there is any marked activity on the part of such subterranean forms. It is significant in this connection that the most frequent of the subterranean Algae are known to be common surface forms (see also the occurrence of moss protonema), and are unicellular types with rounded cells {Chlorococcum) or filamentous forms that readily fragment into short lengths {Hormidium, Stichococcus). It is also significant that common filamentous surface forms that do not show such ready fragmentation {Prasiola, Zygogonium ericetorum) y so far at least, have not been recorded beneath the soil surface. It is curious, however, that the soil-inhabiting Mesotaeniaceae (species of Mesotaenium and Cylindrocystis) hitherto have not been found in subterranean habitats. Esmarch^'' also arrived at the conclusion that the subterranean Blue-green flora is com- posed of the same forms as occur at the surface. Here the capa- city to propagate by motile hormogones and the known liking for saprophytic nutrition on the part of these forms will no doubt tend to lead them to a direct penetration from the surface into the upper layers, as apparently occurs especially in cultivated and in calcareous soils. As to the Algae mentioned above, that so far have rarely or never been found except in cultures of soil samples (Gongrosira The Role of the Terrestrial Alga in Nature 201 terricola, Monocilia, etc.), it is possible that some of these are specially equipped, in a way at present not recognized, for life beneath the surface. However, it is equally probable that they, too, are actual surface forms, rarer than the others, and not so far discovered in their normal habitat. An intensive study of the surface of the soil has as yet been undertaken only in Denmark and Iceland by Petersen,^' ^^ whose investigations showed that a far larger number of surface-growing Algae exist than had previously been recognized. The same investigator^ records two species of Bumilla'ia, hitherto mainly observed in cultures of sub- terranean soil-Algae (especially B. exilis), from the surface of soils of Hammer Bakker, and Botrydiopsis minor from the sur- face of Iceland soils.^* These, as well as certain of the Chloro- phyceae recorded, were not macroscopically visible. It is noticeable that a considerable number of the forms so far mainly recorded from beneath the surface of the soil are Xantho- phyceae (Heterokontae). Certain members of this group are known to be intolerant of strong illumination, and it is possible that the soil forms do not inhabit the actual surface, but live in the feebler light just beneath (cf. also below). This would ac- count for their not having been observed, so far, in the study of actual surface growth. The resting stages of aquatic Algae are known to be to a large degree wind-dispersed. Only a small proportion of such wind- borne spores will be deposited in other adjacent waters. Large numbers of them will fall on the soil surface and become im- mersed in the soil during showers of rain (Fritsch,^ p. 254) . Soil animals may also be instrumental in their burial. No doubt many such resting stages will perish without further development, but some may survive and, in my opinion, will amply account for the occurrence of the casual element among the soil Algae (species 202 E E. Fritsch of An\istrodesmus, C/jlamydomonas, Tribonema, etc.). In fact, the occurrence of such aquatic types in soil cultures seems to me to provide further evidence in support of the view that the sub- terranean Algae are forms washed down from the surface. I will frankly agree that it is difficult, in this interpretation of the subterranean Algae, to find an explanation for the fact, estab- lished by Bristol and referred to above, that there is often a marked increase in the population of soil Algae at the four-inch depth. This depth may, however, mark a physical boundary of some kind (for example, increase in compactness of the soil) beneath which the downward passage of algal cells occurs more slowly so that an accumulation takes place at this level. In this connection it is significant that some (though not all) of the highest numbers recorded for the four-inch depth by Bristol,' pp. 572, 573, followed a period of heavy rainfall. The alternative to the view that the placing of Algae in the soil is a result of downward-washing during rain is to suppose that motile reproductive stages actively move downward in the soil and find in certain layers (namely, the four-inch level) con- ditions specially favorable for development and multiplication. Once buried beneath the surface in darkness, there is no appar- ent reason why they should not move downward as well as up- ward. For such forms as the species of Chlorella and Sticho- coccus, however, in which no motile reproductive stages occur, the only conceivable method of transference in the soil is a downward one caused by the washing effect of rain. Moreover, it still remains to be shown that, in forms such as Chlorococcum, zoospores are formed within the soil; it is possible that, if re- production occurs there, it is by means of aplanospores. A scrutiny of the mere superficies of the soil, even in wet weather, is hardly likely to disclose all the surface-living algal The Role of the Terrestrial Alga in Nature 203 growth. It is very probable that some of these forms Uve beneath the actual surface in a region where photosynthesis is still pos- sible, but where better conditions in regard to moisture supply and protection from desiccation exist; also, some of these forms may be shade-loving species (see the endolithic forms referred to below). Moreover, the more specifically saprophytic forms may be influenced in their distribution by the availability of organic matter. A study of the surface community, therefore, should include not only the forms visible at the surface, but also all those encountered at depths to which light is found to pene- trate. It is not improbable that a microstratification of the surface- living soil Algae may be found to obtain, depending on their immediate light requirements and on their relative degree of susceptibility to desiccation. No data are at present available concerning the depth to which light can penetrate beneath the surface of a compact soil, but it is doubtful whether there is anything but darkness below the top few centimeters. Forms such as Chlorococcum and Hormidium appear to grow on the actual surface. How far there is direct (as opposed to passive) penetration from here into the illuminated layers of the soil has not as yet been ascertained. The forms growing at the surface are of course exposed to considerable desiccation during dry periods, and many of them become con- spicuous only in wet weather. Then, however, they often appear in remarkable numbers. On peaty and sandy soils Zygogonium ericetorum may form an intermittent covering over acres of ground, whereas Prasiola and Hormidium often occur exten- sively on heavier soils, the former especially where there is or- ganic pollution. These forms are conspicuous by virtue of the wefts which they produce. Unicellular types are more difficult to recognize unless they occur in abundance, but on damp sur- 204 E E. Fritsch faces evident growths of Chlorococcum may be found, and forms such as Coccomyxa and Stichococcus, though more occasional, sometimes occur in conspicuous strata; a common habitat is the basal parts of tree trunks just above the soil level." Striking sheets of Blue-green Algae are seen, especially in wet periods, on cultivated soils and calcareous soils, and Petersen^^ records for Iceland extensive sheets of Phormidium autumnale wherever the ground is rich in organic nutriment (see also West," p. 421). It remains to be discovered whether, at times when such growths are not evident on the surface, they are persisting in the illumi- nated subjacent layers, to aid in the resuscitation of a stratum in wet weather. The numerous forms of soil Diatoms, character- ized in general by their small dimensions and their capacity for motility,^' ^^ are no doubt mainly forms living in the illuminated region of the soil, though possibly only a few are present at the actual surface. The occurrence of a rich surface algal flora, especially in wet weather, affords, therefore, ample material from which the sub- terranean element can be supplied, although it is evident that the agencies bringing about downward transference operate differently on the various members of the surface community (p. 205) . Bristol' has shown that a large proportion of the Algae in the soil are in a vegetative condition, or at least in a condi- tion susceptible to the effects of desiccation. This is perhaps what one would expect of an Alga carried down into darkness and gradually losing its food reserves because probably its capacity to withstand drought depends, partly at least, on a high osmotic pressure and also apparently on the presence of fatlike sub- stances."^' ^' ^' *°' *^ It is not necessary, however, to suppose that the cells that survive are in the form of spores. The vegetative cells of diverse terrestrial Algae are known to possess a great capacity The Role of the Terrestrial Alga in Nature 205 to resist desiccation,"^" ^ and the surviving cells may be those that have recently been w^ashed downward and have not yet lost the equipment which enables them to withstand drought. In fact, if they are plentifully stocked with food reserves as a result of photosynthesis at the surface, after burial they may possibly carry on a certain amount of multiplication until these reserves are exhausted, the more so because moisture conditions will be more favorable than they normally are at the surface. But it seems highly probable that, in many instances, when the food reserves are used up, the alga is doomed to gradual death unless it falls a prey to the Protozoa present. The fact established by Johann- son,'" that the production of carbon dioxide from soils is nearly always greater during the day than during the night, would seem to imply that even the surface-living Algae are insufficient in quantity to affect materially the gas relations. Soil Algae have been supposed to play a role in increasing the fertility of the soil but, clearly, if they grow and multiply at all beneath the illuminated zones of the soil, they can do so only at the expense of their own reserves or of the organic material present. In the deeper layers, therefore, they cannot increase fertility unless through their action they render certain of the organic ingredients more readily accessible to the remaining population of the soil. For this view, however, there is at present no evidence. On the other hand, the carrying down of surface forms to the deeper layers will probably add materially to the organic content of the soil because these forms have built up their bodies and no doubt have often stocked them with food reserves in an autotrophic manner. The inclusion of their substance in the soil, whether direct or indirect, must enrich it in organic material. If the previous conclusions are correct, the importance of Algae in soil economy lies mainly in their growth at the surface, 2o6 F. E. Fritsch where photosynthesis is possible, and therefore an intensive study of surface-growing Algae and of their conditions of existence is in the future likely to produce results more valuable to science than a study of their occurrence in the subterranean habitat. Soil Algae have also been credited with fulfilling another function, namely, in respect to nitrogen fixation. A direct fixation of nitro- gen by Algae, and particularly by Green Algae, is certainly highly unlikely, and no data brought forward in this connection have stood the test of a critical examination.''*' It is a demonstrated fact, however, that nitrogen-fixing Bacteria in cultures bind an increased amount of nitrogen in the presence of Algae. It is sug- gested that a possible reason for the beneficial effect of the Algae is that they remove the nitrogenous products of the activity of the Bacteria or provide in their mucilage envelopes a carbo- hydrate supply for the Bacteria. Clearly, however, the value of Algae in relation to nitrogen fixation beneath the illuminated zone of the soil must depend, in large measure, on their being in a condition of active growth and multiplication, and of that I believe no adequate proof has so far been presented. At pres- ent, therefore, I am inclined to doubt that soil Algae are funda- mentally important to the nitrogen-fixing Bacteria except in those layers of the soil which are accessible to light and where rapid growth and multiplication of the Algae can take place. I am fully cognizant that the suggestions presented here con- cerning the source and nature of the subterranean algal com- munity, although supported by a certain amount of fact, require to be proved by diverse direct investigations. I feel, however, that there is so much that is debatable in existing views concerning the subterranean algal flora that a critical consideration of the available facts was desirable. I hope that an experimental testing of the accuracy of my suggestions will not be long delayed. The Role of the Terrestrial Alga in Nature 207 The Surface Community In the few surface Algae that have been investigated in this con- nection, there is a very marked degree of adaptation to the con- ditions of desiccation to which they are exposed.^' ^' ^' *^' " It may particularly be emphasized that they possess the faculty of with- standing drought without appreciable change and without the assumption of special resting stages. There is also evidence that they readily absorb atmospheric vapor as well as liquid water, and in this way probably are able, in large measure, to tide over periods devoid of actual precipitation and to resume active growth and multiplication as soon as wet weather sets in. The special machinery that appears to be connected with the peculiar properties of these forms has been adequately described else- where^' ^' ^ and need not be dealt with here. There is some evi- dence to show that the special properties of such Algae enable them during periods of drought to compete successfully with the higher plants (Piercy,*° p. 515). Apart from the role which such surface-growing soil forms probably fulfill after being washed down into the soil, they must often lead to the production of surface humus after death. This must be especially true when their nature is such that they are not carried down in a living condition. Attention has also been directed (West and West,^ p. 303) to the importance of surface wefts of Algae such as Porphyrosiphon Notarisii and Zygogo- nium ericetorum in consolidating the surface, and in affording in damp weather a seed bed on which smaller seeds, etc., find a ready means of germination. The role of surface forms as pri- mary colonizers of new ground has frequently been empha- sized.^^' " In this connection we may note the descriptions of the first stages in the recrudescence of growth on Krakatoa" and the \ 2o8 ^ E. Fritsch preliminary steps in the colonization of burnt heath in England.^ In both places what appears first is in large part a covering of gelatinous palmelloid forms that in wet weather constitutes a thin but practically continuous mucilaginous covering on the otherwise bare surface. This must aid materially in the establish- ment and early germination of the seeds and spores of the col- onizing plants. It is probable that a more extensive study of primary colonization of newly exposed soil will show that mem- bers of the surface algal community are almost always the first to appear. To what extent such forms may also be active in the disintegration of the particles at the soil surface and immediately below it, is at present unknown. The Aerial Community Turning last to what I have at the outset called the aerial com- munity, I think we should include here all algal growths that do not occur on the soil proper, that is, not only those which are found on vertical surfaces (rocks, tree trunks) above the soil level, but also those inhabiting more or less horizontal surfaces of a similar nature. In all these the substratum will furnish rela- tively Htde water and the growth of Algae must be in large measure dependent on moisture obtained direcdy from the at- mosphere. Among the most important of the rock-growing Algae are the Myxophyceae,*' "^^ "' '^' "' *^ the terrestrial members of which are indeed seen at their best in such a habitat; species of Trentepohlia also thrive but are of secondary importance, and other Green Algae and Xanthophyceae are usually quite sub- sidiary except apparently on some kinds of sandstone." On trees the most important forms are Chlorococcales, Pleurococcus, and species of Trentepohlia, the last-named especially in the Tropics or in very humid regions such as parts of the Monterey Penin- The Role of the Terrestrial Alga in Nature 209 sula. A rich flora of diverse green forms sometimes occurs on trees and woodwork,** but as a rule the growth is very monotonous. In respect to their role in nature, the rock-inhabiting forms are of special interest. Blue-green Algae in damp, mountainous tracts in temperate regions give one some idea of their cap- abilities for growth on rock surfaces,"' " but, in order to realize their full possibilities, one must observe them in the damp Trop- ics. The growths there have been familiar since Welwitsch's description,*^ in 1868, of the "Pedras Negras" or black rocks of Pungo Andongo in Angola, where the dark color is caused by the prolific growth of Scytonema myochrous (Dillw.) Ag. var. chorographicum West. Bews,* p. 553, has described similar ex- tensive coverings of Blue-green Algae from the Drakensberg region in Natal, where the communities of diverse Myxophyceae "cover enormous areas of the cliffs, especially the cave-sandstone cliffs and are strikingly apparent from a distance as longitudinal black strips During the very dry weather the algal covering dries up and peels from the rock in small flakes, but it does not disappear to any great extent or leave the rock again bare" (cf. also ^*), in this respect differing from the growth that Welwitsch has recorded in Angola. I myself have described similar phenom- ena on a smaller scale in Ceylon,^"* and McLean,^"" speaking of the Pedra da Gavea, west of Rio de Janeiro, says, "The larger part of the rock-surface is covered by blue-green Algae which form great dark stripes upon it!' He also records "a species of Trentepohlia (T. aurea) which is extremely abundant on the damper patches'.' The so-called "Tintenstriche" recorded by diverse workers on the rocks and precipices of various mountain- ous tracts in Europe,"' "' " are caused by similar growths of Blue- green Algae along vertical lines down which there is a more or less constant trickle of water. 210 E E. Fritsch Such dense and extensive mats o£ aerial Algae are usually supposed to play an active part in rock erosion, and there is every reason to believe that this is correct, but at present there is litde positive evidence of it. The peeling off of the growth v^^hich Welwitsch, Bews, and others describe as occurring regularly in dry weather may well act on the rock surface in a way similar to that recorded by Frye^ for certain Lichens, but exact investi- gation is required. It is not known whether in all such occur- rences a direct corroding eflFect may be exerted on the under- lying rock with the help of substances excreted from the Alga, as has been established for Myxophyceae inhabiting calcareous rocks (cf. below). On horizontal rock surfaces the dense wefts or sheets (Phormidium) of Blue-green Algae are bound sooner or later to provide a sufficient thickness of humus for the growth of Bryophyta, Ferns, and small Phanerogams, and sometimes this no doubt also occurs in connection with more or less vertical substrata. When adequate moisture is available for a consider- able length of time, successive phases of development of these mats may be traced ;^^'^^ a sufficiently thick growth is established in a short space of time to afford a starting point for the develop- ment of higher types of vegetation. In short, although no doubt the rigors of the habitat often prevent development beyond the algal mat, there is every reason to suppose that a similar growth preceded the establishment of Phanerogamic vegetation in the more sheltered valleys and on the lower slopes of the cliffs where such growths are now observable on the upper parts. On calcareous rock surfaces in the Alps and the Dolomites, Bachmann^ and Diels" have recorded both an epilithic and an endolithic community of Algae, composed in large part of Chroococcales with occasional filamentous Blue-green forms and species of Trentepohlia (cf. also^*). The epilithic community is The Role of the Terrestrial Alga in Nature 211 confined to those points at which a slow trickle of water normally descends the rock face, whereas the endolithic community ap- pears to be of two types. The type described by Diels inhabits very narrow, outwardly not recognizable clefts in the rocks which apparently owe their origin to atmospheric agencies. These aerial Algae may be described as chasmolithophytes." They differ in general from those found on die exposed surface and are no doubt shade forms. On the contrary, the community described by Bachmann and others" consists of Chroococcales occupying small pits in the rock surface which are believed to result from a direct solvent action of acid excreted by the Alga on the calcareous substratum (true endolithophytes) . When such growths are numerous, the surface layer of the rock to a depth of 1.5 mm. becomes quite soft and porous, so that a direct erosive action of the alga on the substratum is evident. But, even in respect to the chasmolithophytes, one cannot doubt that the presence of the Algae will lead to a gradual widening of the clefts and progressive erosion of the surface.^ On the whole, the chemical nature of the underlying rock does not appear to have much effect on the aerial algal commu- nity, which is apparendy more influenced by the nature of the surface, the amount of available moisture, and the intensity of insolation. Fremy" has made a study of the composition of the aerial algal growth on siliceous and calcareous rocks in Nor- mandy, and mentions as the only striking difiference the absence of Stigonema minutum on the calcareous as opposed to its fre- quent occurrence on the siliceous rocks (cf. *°). He points out that the hardness of the rock definitely influences the abundance and nature of the algal growth. Hard, dry rocks bear no growth at all when exposed to full insolation (cf.also"), but when shaded they support Pleurococcus and Trentepohlia aurea, forms which 212 E E. Eritsch are also found on rather softer dry rocks subjected to full insola- tion. When shaded, these softer dry rocks bear a blue-green growth (Gloeocapsa spp., Tolypothrix byssoidea f. saxicola, Stig- oneina minutum) which, with the difference noted above, is found alike on siliceous and calcareous rocks of a certain degree of hardness. Ercegovic^^ regards Scytonema myochrous and Calothrix parietina as photolithophytes because they always occur on places exposed to strong sun. Diatoms rarely form con- spicuous growths on relatively dry rock surfaces. Schade," how- ever, records yellowish brown coverings caused by Eragilaria capucina Desm. (cf. also ^). Apart from obvious algal coverings, Falger" has shown that Algae may also be found on rock surfaces that afford no macro- scopically visible evidence of growth. He suggests that these primary colonizers prepare the way for the more obvious growths. Again Myxophyceae are dominant, although some Chlorophyceae and Diatoms are also found. Sandstone is much richer than other kinds of rocks and here, as in Fremy's observa- tions, a direct effect of the degree of porosity of the substratum, implying a varied retention of moisture, is evident. Adequate data on the moisture content of exposed rock surfaces are, how- ever, at present scarcely available, although Schade" has calcu- lated that in sandstone the water content may amount to 12 per cent of the weight of the rock. The forms inhabiting trees are probably of less importance in the economy of nature, but physiologically are perhaps the most highly adapted terrestrial Algae. The faculty that Pleurococcus possesses of withstanding extreme and prolonged drought and other unfavorable conditions is almost proverbial. According to Schmid,*^ this Alga does not stand continuous immersion in water and normally satisfies all its moisture requirements by The Role of the Terrestrial Alga in Nature 213 means of atmospheric vapor. Probably the same is also true to a large degree of Trentepohlia (Irgang,** p. 15). Note: This essay was completed in July, 1934. Since then Miss James's paper referred to on page 196 has been published (see Beih. bot. Centralbl., 53, A, 519- 553, 1935). There has also appeared a detailed study of soil Algae by J. B. Peter- sen (Dansk. Bot. Arkiv, 8, no. 9, 1935) in which the author arrives at conclusions which to a considerable extent agree with those propounded above with respect to the subterranean community. BIBLIOGRAPHY* ^ Bachmann, E. 1915. Kalklosende Algen. Ber. deutsch. bot. Ges., 33, 45-57. ' Beger, H. 1927. Beitxage zur Oekologie und Soziologie der luftlebigen (atmophy- tischen) Kieselalgen. Ibid., 45, 385-407. ' Beger, H. 1928. Atmophytische Moosdiatomeen in den Alpen. Vierteljahrschr. naturf. Ges. Zurich, 73, Beibl. 15, 382-404. (Festschrift Hans Schinz.) * Bews, J. W. 1917. The plant ecology of the Drakensberg Range. Ann. Natal. Mus., 3, 511-565. ^ Bristol, B. M. 1920. On the alga-flora of some desiccated English soils. Ann. Bot., 34, 35-80. " Bristol Roach, B. M. 1927. On the Algae of some normal English soils. Jour. Agr. Sci., 17, 563- 588. (Cf. also Proc. and Papers ist Internat. Congr. Soil Sci., 3, 1927-) ' Bristol -Roach, B. M. 1927. Methods for use in studying the Algae of the soil. Abderhalden, Handb. biol. Arbeitsmethod., 11, pt. 3. ^ Bristol Roach, B. M. 1927. On the carbon nutrition of some Algae isolated from soil. Ann. Bot., 41, 509-517- ° Bristol Roach, B. M. 1928. On the influence of light and of glucose on the growth of a soil alga. Ibid., 42, 317-345- " Bristol, B. M., and Page, H. J. 1923. A critical enquiry into the alleged fixation of nitrogen by Green Algae. Ann. Appl. Biol., 10, 378-408. " Diels, L. 191 4. Die Algen-Vegetation der Siidtiroler Dolomitriile, etc. Ber. deutsch. bot. Ges., 32, 502-526. " DORST, E W. 1 93 1. Beitrage zur Oekologie der protococcoiden Algen. Ber. oberhess. Ges. Natur. u. Heilk. Giessen, 14, 105-135. * This is not an exhaustive list of the literature on terrestrial Algae; however, such a list can be readily obtained from the bibliographies in the papers cited. [214] The Role of the Terrestrial Alga in Nature 215 " ERCEGO\ac, A. 1925. La vegetation des lithophytes sur les calcaires et les dolomites en Croatie. Act. Bot. Inst. Univ. Zagreb., i, 64-114. " ESMARCH, E 1 910. Beitrag zur Cyanophycecnflora unserer Kolonien. Jahrb. Hamburg, vv^iss. Anstalt., 28, Beih. 3. ^ EsMARCH, E 1914. Untersuchungen iiber die Verbreitung der Cyanophyceen auf und in verschiedenen Boden. Hedwigia, 55, 224-273. " Falger, E 1922-23. Die erste Besiedelung der Gesteine. Mikrokosmos, 16, 13 £. (Cited from Bot. Centralbl., 3, 339-340, 1924.) " Fremy, p. 1925. Essai sur I'ecologie des Algues saxicoles aeriennes et subaeriennes en Normandie. Nuov. Notarisia, 36, 297-304. " Fritsch, E E. 1907. The role of algal growth in the colonisation of new ground, . . . Gcogr. Jour., 531-548. " Fritsch, E E. 1907. A general consideration of the subacrial and freshwater algal flora of Ceylon. Proc. Roy. Soc. London, B, 79, 197-254. ^ Fritsch, E E. 1922. The terrestrial alga. Jour. Ecol., 10, 220-236. ^ Fritsch, E E. 1922. The moisture relations of terrestrial Algae, L Ann Bot., 36, 1-20. ^ Fritsch, E E. 1 93 1. Some aspects of the ecology of freshwater Algae. Jour. Ecol., 19, 233-272. ^ Fritsch, E E., and Haines, E M. 1923. The moisture relations of terrestrial Algae, IL Ann. Bot. 37, 683-728. ^ Fritsch, E E., and Rich, E 1924. Freshwater and subaerial Algae from Natal. Trans. Roy. Soc. S. Africa, 11, 297-398. ^ Fritsch, E E., and Salisbury, E. J. 191 5. Further observations on the health association on Hindhead Common. New Phytol., 14, 1 16-138. =« Frye, E. J. 1926. The mechanical action of corticolous Lichens. Ann. Bot, 40, 397-417* (See also ibid., 38, 175-196, 1924.) " GiSTL, R. 1933. Erdalgen und Dungung, . . . Arch. f. Mikrobiol., 4, 348-378. 2i6 E E. Fritsch "® Gronblad, R. 1933. A contribution to the knowledge of subaerial Desmids. Comment. Biol. Soc. Sci. Fennica, 4, no. 4. ^ Rowland, L. J. 1929. Periodic observations of Trentepohlia aurea Martius. Ann. Bot., 43 173-202. ^ Irgang, E. 1927-28. Beitrage zur Kenntnis von Trentepohlia lolithus. Fedde, Reperto- rium, Beih., 51, 1-24. "^ JOHANNSON, N. 1929. Rhythmische Schwankungen in der Aktivitat der Mikroorganismen des Bodens. Svensk. Bot. Tidsskr., 23, 241-260. ''^ McLean, R. C. 1919. Studies in the ecology of tropical rain-forest, . . . Jour. Ecol., 7, 5-54. ^ MooRE, G. T, and Carter, N. 1926. Further studies on the subterranean algal flora of the Missouri Botani- cal Garden. Ann. Missouri Bot. Gard., 3, 1 01-140. ^* Pascher, a. i93i,Eine braune, aerophile Gallertalge und ihre Einrichtungen fiir die Verbreitung durch den Wind. Beih. bot. Centralbl., 47, I, 325-345. ^ Petersen, J. B. 1915. Studier over danske aerofile alger. Mem. Acad. Sci. et Lettres Dane- mark, VII, 12, no. 7. ^^ Petersen, J. B. 1928. The alga-flora of some soil-samples from Iceland. Dansk. Bot. Arkiv, 5, no. 9. ^^ Petersen, J. B. 1928. The aerial Algae of Iceland. Botany of Iceland, 2, 327-447. ^® Petersen, J. B. 1932. The algal vegetation of Hammer Bakker. Bot. Tidsskr., 42, 1-48. ^° Petersen, J. B. 1932. Ueber das Wachstum von Erdalgen. (Vorl. Mitteil.). Planta, 17, 15-21. *° PlERCY, A. 1 91 7. The structure and mode of life of a form of Hormidium flaccidum A. Braun. Ann. Bot., 31, 513-537. " PUYMALY, A. 1924. Recherches sur les Algues vertes aeriennes. Bordeaux. *^ Printz, H. 1920, Subaerial Algae from South Africa. Norsk. Vidensk. Selsk. Skrift (1921), no. I. The Role of the Terrestrial Alga in Nature 217 " ROBBINS, W. W. 191 2. Algae in some Colorado soils. Bull. Agr. Exper. Station, Colorado, 184, 24-36. ** SCHADE, F. A. 1913. Pflanzenokologische Studien an den Felswanden der Sachsischen Schweiz. Engler bot. Jahrb., 48, 1 19-210. *^ SCHMID, G. 1927. Zur Oekologie der Luftalgen. Ber. deutsch. bot. Ges., 45, 518-533. *^ SCHORLER, B. 1914. Die Algenvegetation an den Felswanden des Elbsandsteingebirges. Abh. naturw. Ges, Isis, Dresden, 3-27. " Treub, M. 1888. Notice sur la nouvelle flore de Krakatau. Ann. Jard. bot. Buitenzorg, 7, 213-223. *^ Welwitsch, F. 1868. The Pedras Negras of Pungo Andongo in Angola. Jour. Travel and Nat. Hist., I, 22-36. *^ West, G. S. 1 91 6. Algae. Cambridge Botanical Handbooks. ^ West, W., and G. S. 1897. Welwitsch's African Freshwater Algae. Jour. Bot., 35, i f. The Plant as a Metabolic Unit in the Soil-Plant System By D. R. HOAGLAND* THE RELATION o£ the plant to the soil as a medium for plant growth concerns the general botanist, the ecologist, the agriculturist, the forester, and indeed all who seek an understanding of plant growth, plant distribution, or the artificial modification of the plant's nutrient environment. But it is seldom that the soil-plant system is contemplated as a whole in its physiological aspects. Too frequently the absorbing root cells of the plant are regarded, apart from their water-absorbing functions, as merely passive receptacles for the dissolved mineral constituents of the soil. Chief emphasis is frequently placed on the solubility of these constituents in various solvents, and the quest for simple laboratory methods of evaluating soil conditions is unceasing. Single factors, for example, hydrogen ion concen- tration, are often assigned great weight in explanations of plant distribution or adaptation to soil conditions. Many other pro- cedures are employed, in which some special phase of physico- chemical equilibrium in the soil system is broadly interpreted in physiological terms. Such methods of investigation do frequently yield results of local practical application, or occasionally useful generalizations, yet they are essentially nonphysiological, and therefore inade- quate, in that they fail to deal with the soil-plant system as one profoundly influenced by the growth and metabolism of the • It is desired to acknowledge the participation of T. C. Broyer and other mem- bers of the staff of the Division of Plant Nutrition, University of California, in the experimental v^^ork cited herein. [ 219 ] 220 D. R. Hoagland plant itself. This latter view recognizes a system of reciprocal re- lations of extraordinary complexity. At times it may appear that this complexity renders hopeless the attainment of any real com- prehension of the phenomena involved, but the development of methods of investigating plant growth under conditions which permit control of the principal factors in the environment of the root and shoot has renewed faith in the possibility of resolving some of the problems. Many past investigations lacking any elements of controlled experimentation have led to vast increases in observational data, and in general information, but not to secure advances in the understanding of basic principles. The indispensability of the ecological or agricultural approach to the study of soil-plant relationships is not open to question. The whole natural environment of the plant cannot be imitated in the laboratory, and important physiological problems may first be suggested in the field. Yet the knowledge gained from ob- servations made under natural conditions seldom suffices to establish the physiological nature of the processes involved. Such knowledge is indeed often self-limited, since new methods of studying plants in the field, and new points of view in the inter- pretation of data gained by ecological and agricultural investi- gation, are generally developed in the course of experiments in which some element of exact laboratory control enters. Field research and laboratory research should be regarded as different means of attacking the same basic problems, not as separate branches of plant science. Methods of Controlled Experimentation The complete control of the environment of a plant in an experi- mental procedure is obviously beset with manifold technical difficulties, but it has been demonstrated that these difficulties The Plant as a Metabolic Unit 221 may be overcome in sufficient degree to permit the realization of certain ideals o£ quantitative measurement, despite the com- plexity of the biological system. As an illustration, experiments conducted in this laboratory may be cited. Wheat plants w^ere grown in chambers controlled with respect to light, temperature, humidity, air flow, and nutrient solution. In an earlier form of these chambers (Davis and Hoagland),'* control was available only within a limited range of conditions, but a more elaborate installationf makes feasible the accurate control of humidity and temperature over a wide range, illumination of regulated in- tensity and duration being supplied by Mazda lamps. By the use of such chambers, experiments can be replicated at will, with a very small probable error. Therefore it becomes possi- ble to compare with quantitative accuracy many known com- binations of variables in successive experiments. Using a pure line seed, "standard" sets of plants may be grown at any time which are essentially identical, as judged not only by the super- ficial criteria of growth, but also by organic and inorganic com- position, as has been shown recendy by A. R. DavisJ and his coworkers. It is, of course, recognized that there are formidable obstacles to the wide extension of this kind of experimentation. Chief among these obstacles is the practical difficulty of producing artificial light of suitable intensity and quality. The growth of many species of plants, under the kind of artificial illumination usually employed, is very different from that occurring under ordinary natural illumination. However, present difficulties are not insurmountable and there exists a rich field for exploration. * Superior figures refer to items in the bibliography at the end of this essay. tThis apparatus, erected by the Carriere Corporation, in consultation with Professor A. R. Davis and others, will be described later. % Unpublished data. 222 D. R. Hoagland Although the artificial environment necessarily departs in many respects from any natural environment, certain general princi- ples governing the response of the plant to its environment can be developed by laboratory research and may find an application in the study o£ plants in the field. The solution of certain physiological problems requires the complete and accurate control of environment of the whole plant, but other problems, including most of those to be discussed in this paper, may be usefully studied for the present through methods of partial control, accomplished in the greenhouse or laboratory. Water and sand culture technique afford indispensa- ble means of controlling the nutrient environment of the root. It is true that this technique has been employed by plant physi- ologists for many years, but only recendy have some of the factors been properly recognized and controlled. The impor- tance of root aeration and of certain chemical elements formerly considered unessential, may be mentioned. Partial control of the aerial environment serves effectively for certain purposes. Duration of illumination can be increased through the use of artificial light at night; the intensity and quality of the natural light can be regulated to some degree by the use of special light filters, or screens; partial control of humidity or of air movement can be accomplished in suitable glass chambers. Furthermore, accurately controlled experiments may be made with isolated tissues, such as storage organs or root systems, where the factor of illumination does not immediately enter. These methods of quantitative control are capable of leading to an understanding of certain interrelations between the plant and its environment which cannot be gained from observations on plants growing under uncontrolled conditions. Quantita- tively controlled experiments have an obviously broad signifi- The Plant as a Metabolic Unit 223 cance to the student of plant physiology whether his interest lies in ecology, or agriculture, or in specific plant functions, such as transpiration, translocation, photosynthesis, respiration, or ab- sorption of nutrients. Suitable control of environment and quan- titative measurement of biochemical changes in plants subjected to this control offer a promise that plant physiology may develop in due course some of the attributes of an exact science, at least so far as the methods of exact science are at present applicable to biological systems of such great complexity. It is proposed to analyze certain problems of the soil-plant system with reference to experiments conducted in accordance with the views just outlined. For the greater part the discussion will be based on data recently acquired by a group of investi- gators working in California. Nature of the Process of Accumulation of Mineral Solutes by Plants It is now generally understood that absorbing plant cells possess the power of accumulating certain chemical elements derived from the nutrient medium, not merely by fixing these elements in organic combination, but also by storing them in vacuolar sap in inorganic form. This frequently implies a movement of solutes against a concentration gradient, and work must there- fore be done by the cell in this process. Observations on Valonia," Nitella,^ and other plant cells supporting this conclusion have been discussed so often in recent years that no recital of the arguments need be presented at this time. The conception of protoplasmic activities as indispensably related to the absorption and accumulation of essential or non- essential mineral solutes has been fundamentally advanced by the work of Steward" in his investigation of storage tissues. It 224 ^' R- Hoagland was clearly shown that ability to absorb solutes against concen- tration gradients (accumulation) is dependent upon a state o£ intense cell metabolism, which is reflected in the processes o£ aerobic respiration. This does not mean, however, that carbon dioxide production is a direct measure o£ the power of accumu- lation; in fact, there is reason to suppose that certain other meta- bolic transformations not readily measurable, but reflected by the ability of cells to grow and divide (according to Berry and Steward'"), may be more directly related than respiration, as determined by carbon dioxide production, to the process of absorption and accumulation of solutes. In any event, a suitable supply of oxygen is requisite to those activities which must be set in action before the cell can remove solutes from a dilute outer solution and transfer them to an inner solution of higher concentration. Furthermore, these metabolic processes must pro- ceed at a certain minimal rate in order that the cell may retain solutes already accumulated in the sap. The applicability of these general principles to growth of plants in the soil has been demon- strated by extensive studies conducted in this laboratory in recent years on the nature of absorption and accumulation of mineral solutes by root cells. A new approach is oflFered to many prob- lems of availability of nutrient elements. Oxygen Requirement for Accumulation OF Solutes by Root Cells One out of many experiments will illustrate the indispensable requirement of oxygen for the process of accumulation of mineral solutes by root cells. In this experiment barley plants were grown in culture solution for three weeks and then the entire root systems were excised, a very large number of uniform plants being used. One set of excised roots was placed in an The Plant as a Metabolic Unit 225 experimental solution containing potassium bromide* and potas- sium nitrate, with a rapid stream of air passing through the solution for 10 hours. Another set of roots was subjected to a similar solution, but nitrogen gas instead of air was passed through it. A third set of roots served for the study of initial composition. Roots were frozen and thawed, and sap expressed, concentrations of nitrate, potassium, and bromine being then determined, and conductivity of the sap also. It is shown in figure i that very large increases in concentra- tions in the sap of the ions named above, accompanied by corre- lated increase in conductivity, occurred when air was passed through the solution in which the roots were immersed, but ac- tual salt accumulation was not significant when nitrogen gas was substituted for air. The nitrogen gas contained a very small amount of oxygen as an impurity, apparendy sufficient to pre- vent, over the brief experimental period, loss of solutes already accumulated by the root cells. The conclusion from this and other experiments, in brief, is that metabolic activities of root cells requiring oxygen bring about the very rapid intake of electrolytes, so that there are built up in the cell sap far higher concentrations of the ions concerned than exist in the external solution. As already stated, work is done by the cell in the process of accumulation. Diverse types of living plant cells represented by storage organs, certain aquatic plants, and root systems of higher plants all seem to have the same general requirements for solute accumulation. The mecha- nism by which energy is used in the transport of solutes against concentration gradients is still without an explanation in accord with all experimental evidence. * Bromine is a very convenient element for studies on accumulation, and does not differ from the ordinary nutrient elements in its relation to the general physiological processes involved. 226 D. R. Hoagland D INITIAL CONC. IN SAP ABSORPTION WITH ABSORPTION WITH AIR O X o X S ft. >■ O 3 O Z O u Fig, I. Effect of aeration on accumulation of salt by excised roots of barley plants. Each set of root systems represents i68 plants previously grown in a complete culture solution from May 31 to June 26. Absorption period, lo hours; tempera- ture of solution, 24° C. ^ . , 1 • (KBr .0075 M) Experimental solution: .' .^^^ . -^A 3000 cc. ^ |Ca(N03)2 .0025 M^ pH of culture: initial, 6.0; final, air 5.8; N2, 6.3. During the absorption period, a rapid stream of air was passed through one solution and purified nitrogen gas through the other. The nitrogen gas, however, retained traces of oxygen. Some oxygen was also present in the tissue and solution at the beginning of the absorption period. It should be noted that much of the nitrate absorbed was reduced, and that potassium absorbed in association with nitrate remained in the sap in equilibrium with organic acid anion formed in the course of metabolism. Numerous consequences arise from the principle that absorp- tion and accumulation o£ mineral elements are dependent on protoplasmic activities. In the first place, the oxygen requirement of root cells and the oxygen-supplying power o£ the soil assume a new and specific significance. No matter what the concentra- tion of solutes in the soil solution may be, these solutes are not The Plant as a Metabolic Unit 227 physiologically available to the plant unless oxygen is supplied to the roots, and carbon dioxide removed, at the necessary rates. Our general knowledge of the oxygen requirements of roots leads to the supposition that in this regard great differences exist among plants of different types. Investigations of this laboratory suggest that barley plants can maintain active accumulation of mineral solutes at relatively low oxygen tensions, considerably lower than that of the atmosphere. But similar oxygen tensions may be inadequate for certain plants, and some probably have oxygen requirements much lower than that of barley. The supply of oxygen to the roots may not be derived exclu- sively from the soil atmosphere. Cannon^ has concluded that translocation of oxygen from the leaves to the root system may occur during periods of photosynthetic activity. If this is an important source of oxygen for root activity, then illumination has a role in the absorption of mineral elements not previously suggested, and anatomical adaptations facilitating the transfer of oxygen assume significance in the study of the mineral nutrition of plants. The importance of this transfer would depend upon the physical structure of the soil, rainfall, activities of microor- ganisms, and all other factors which determine the power of the soil to supply oxygen to the roots. In connection with root activities, the effects of carbon dioxide concentrations in the soil atmosphere are likewise of interest. Such concentrations may rise to relatively high levels at certain times. It has not been shown that absorption of mineral solutes is necessarily depressed by carbon dioxide concentrations likely to be present in the soil atmosphere, but the whole relation of carbon dioxide-bicarbonate equilibrium to root growth and solute absorption requires further careful study. It is obvious that the soil system is influenced not only by the metabolic activi- 228 D. R. Hoagland ties of root cells, but also by those of soil microorganisms. The latter aspect of soil-plant interrelations, although it is of utmost importance, cannot be discussed in this brief review. Energy Supply for Accumulation of Solutes Since energy exchanges are involved in the processes of accumu- lation of solutes, a supply of available carbohydrate must be fur- nished to root cells and synthesis and translocation of sugars are therefore closely interrelated with mineral nutrition. Preliminary evidence is available that depletion of the supply of sugar or other available carbohydrate in the root system is accompanied by a retardation or inhibition of the absorption of mineral solutes and that feeding sugar to the roots, under otherwise suitable conditions, restores, at least in part, the accumulating power of the root cells, lost as a result of sugar depletion. The mineral elements absorbed from the culture medium are, directly or indirectly, indispensable to the photosynthetic activi- ties of green cells, and beneficent or vicious cycles in the soil- plant system may be established. For example, a limitation in the supplying power of a soil for potassium may lead to an impairment of leaf function (but not necessarily because potas- sium has a direct role in the process of photosynthesis) and con- sequently to a diminution in the rate of supply of carbohydrate to the root system. A retardation of root growth or metabolic activity may follow with a corresponding retardation of the rate of accumulation of potassium from the soil solution. A further starvation of the shoot for potassium ensues, and so the cycle proceeds, with a trend toward some sort of equilibrium between shoot and root. Injury or death of the plant may occur, however, before such an equilibrium is attained. The amount of fruit or seed production, which is determined The Plant as a Metabolic Unit 229 in part by climatic and biological factors, will affect the require- ment of the plant for potassium or other nutrients and therefore affect estimates of the adequacy of soil conditions. Heavy fruit- ing will tend to diminish the amount of carbohydrate available for root activities, and, according to the views already presented, lessen the capacity of the root system for absorbing nutrients. The soil does not possess a fixed power for producing plant growth, but has instead a physiological value which cannot be adequately appraised without reference to the whole soil-plant system, and in relation to internal plant factors and to climatic environment. A basis for these statements has been established primarily through experiments with plants of agricultural inter- est, but the general principle is of wide application. Certain Factors of the Climatic Environment AND Root Growth Other interrelations of root and shoot exist which are much more obscure. In growing young barley plants in the greenhouse at different times of the year, under a given uniform nutrient supply, it has been observed that both the quantitative relation of root to shoot and the nature of the root system, as reflected by development of fine lateral roots, vary in a highly significant manner, depending on seasonal conditions. The relationship is illustrated in figure 2. The proportionate yield of roots is far higher in summer than in winter. Without discussing details of the experiments at this time, it may be said that not temperature, but some factor of illumination, seems to determine the extent and nature of the root system. It is not a question of ordinary photoperiodism, but of the quality or intensity of the light. Recent investigations on plant hormones^' ^ suggest that growth determinants of this type, the production of which de- 230 D. R. Hoagland i/> z 160 150 UO 130 O 120 UJ 110 100 90 IIJ Ik 80 70 60 50 JAN. MAR. MAY JULY SEPT. NOV. JAM. Fig. 2. Yields (fresh weights) of shoots and roots of barley plants grown in culture solution, in greenhouse, at different periods of the year; i68 plants grown in each set for periods of approximately 3 weeks. Roots, before weighing, were centrifuged by a uniform technique in order to remove adhering moisture. Ca(N03)2 .0025 M KNO3 .0025 M MgSO* .001 M KH2PO4 .0005 M Fe added at intervals as iron tartrate. Composition of culture solution: pends on suitable illumination, may possibly cause the observed effects on root development. For the particular experiments now under discussion, a simpler explanation cannot at present be excluded, in terms of a supply of organic nutrients, synthe- sized when light of suitable intensity and quality is available, although it does not seem that the problem is one merely of car- bohydrate supply. Whatever the mechanism may be, it is clear that a suggestion emerges for the study of a specific influence of the light factor on root growth and therefore on the relation of the plant to the soil. The Plant as a Metabolic Unit 231 Temperature Influences The expectation is that the metabolic activities of root cells gov- erning the accumulation of solutes should have a high tem- perature coefficient over certain ranges of temperature, and ample experimental evidence is now available to support this expectation. In barley roots, for example, the accumulation of mineral elements has a high temperature coefficient in the range 6° 0.-24° C, but the rate of accumulation may increase with temperature up to 30° C, or higher, at least over a brief period of time. The lower temperature limit for accumulation has not yet been determined, but it is probably not far below 5°-6° C. It is, however, interesting that accumulation of mineral solutes can proceed slowly even at this low temperature. A typical example of temperature influence on accumulation (for potassium) is given in figure 3. Barley plants were grown in nutrient solution for approximately three weeks and then the root systems were excised. These excised root systems were im- mersed in solutions of potassium salts and subjected to varying temperatures for limited periods of time, under conditions of aeration suitable for high respiratory activity. The temperature coefficients indicated are of a high order and in no way sug- gestive of simple diffusion processes. It is evident, rather, that temperature is influencing metabolic rates and that these are controlling the rate of accumulation of potassium. When the en- tire plant is studied, other complicating factors enter. The results cited are not restricted to the barley plant. Very similar temperature relations have been observed in experiments with Nitella cells and storage tissues. The temperature coeffi- cients for the accumulation of different elements are not neces- sarily identical, but it must be emphasized that temperature 232 D. R. Hoagland 90 80 70 ft. ^ 60 o * 50 o «• u . ..i 40 ly W W» J 30 < J M — te. s ^ « 20 lO ,..«»-' » y ^^ / / / / ^ Y 12 18 24 TEMPERATURE OF SOLUTION -DEGREES C SO Fig. 3. Effects of temperature on accumulation of potassium in sap by excised barley root systems; each set represents 168 plants, previously grown in a com- plete culture solution for 21 days. Absorption period, 10 hours. Experimental solution: .005 M .005 M .002 M KBr KNO3 Ca(N03)2 The temperatures of the solutions were maintained as shown within approxi- mately plus or minus 0.2° C. effects are not restricted to any special type o£ ion. The absorp- tion of both anions and cations is influenced by temperature. The importance of soil temperature with special reference to ecology or to practical agriculture is apparent. Seasonal and diurnal changes in soil temperature must influence enormously the rates of absorption of mineral solutes within certain ranges of temperature, and under conditions otherwise favorable to absorption. The general significance of this factor has been emphasized by Lundegardh/ but as yet the relationship of soil The Plant as a Metabolic Unit 233 temperature to mineral nutrition has received but scant general attention, and adequate data on temperature coefficients for plants of diverse types do not exist. It may be reasonably assumed that temperature limitation of mineral solute absorption some- times constitutes a limiting factor for plant growth. And yet, high temperatures hasten depletion of carbohydrate reserves through accelerated respiration. When plant growth is being studied over an extended period, rate of photosynthesis and of translocation of carbohydrate to the root system, as well as rate of translocation of salts from root to shoot, should be considered in the evaluation of soil temperature effects. Hydrogen Ion Effects Few factors in the plant's environment have received so much discussion as hydrogen ion concentration, but with each addition to our knowledge it becomes clearer that hydrogen ion concen- tration cannot profitably be interpreted in terms of an isolated variable. Leaving aside the well-known effects of hydrogen ion concentration on the solution of certain physiologically impor- tant chemical elements of the soil, there remains a highly compli- cated interrelationship between hydrogen ion concentration of the soil solution, buffering system of the soil, metabolism of the plant, and accumulation of mineral solutes. Some of the recent studies of this and other laboratories lead to doubt whether, within a wide range of values, external hydro- gen ion concentration per se is as important a factor as many investigators have beUeved. It is evident from data already avail- able that the effects of hydrogen ion concentration must be re- examined with reference to root temperature, oxygen supply of the culture medium, and other factors controlling rates of root metabolism. The view now suggested is very different from that 234 ^' ^' Hoagland derived from the assumption that hydrogen ion concentration is mainly operative through some relatively simple physicochem- ical system which may be explained by reference to Donnan equi- libria, protein isoelectric points, hydrogen ion gradients, and other factors of the kind. Transpiration and Absorption of Mineral Solutes The relation of the absorption and transpiration of water by plants to the absorption of mineral elements, or to their move- ment within the plant, has been discussed by plant physiologists decade after decade. Yet the most recendy published articles re- flect a striking diversity of opinion with respect to the function of transpiration in the economy of the plant — whether it is a necessary evil or is beneficial. There remains great uncertainty concerning many features of the movement of water and solutes within the plant's conducting systems, but several aspects of the absorption of mineral elements in relation to absorption and transpiration of water can be clarified on the basis of simple and convincing experiments. Metabolically active root cells, when in a low salt condition, have the capacity to accumulate mineral solutes with great ra- pidity from a dilute solution, even when the root is severed from the shoot, as has already been demonstrated. Under suitable con- ditions and over a brief interval of time, almost as great a quan- tity of certain mineral solutes can be absorbed from a culture solution by the excised roots alone as by the whole plant. Trans- piration is therefore not involved in the first step of the process by which plants remove solutes from the soil. The plant as a whole being considered, root cells, when in a state of high metabolic activity and not already saturated with mineral solutes, may absorb some ions from the culture solution The Plant as a Metabolic Unit 235 t 7 « 6 u > ^ § 4 ui E z 2 z ^ o o N03 1 ■ PINAL CONCENTRATIONS LJ ORIGINAL CONCENTRATIONS HCO1 H2PO4 Fig. 4. Relation of water and solute absorption; barley plants (duplicate sets of 168) grown for 19 days in a complete culture solution. Experimental period of absorption, 24 hours. Solutions used, same as described under figure 2. Correction made for very slight loss of water caused by evaporation from solu- tion; 825 cc. of water absorbed by each set of plants, from original volume of 3900 cc. Each set of plants absorbed the following total quantities of the various ions (in milligram equivalents) : K 1 1.5 NO3 22.0 Ca 2.6 SO4 0.6 Mg 1.5 H2PO4 0.2 HCO3 determined by titration method. Some organic acid radicles may have been present. at a proportionally much faster rate than water and as a neces- sary consequence the nutrient solution becomes diluted with respect to these elements. A recent experiment (fig. 4) illustrates this point. The roots of young barley plants were immersed in an aerated nutrient solution for twenty-four hours, without any addition of water. The arrangements permitted only a very slight amount of water to escape by direct evaporation from the solu- 236 D. R. Hoagland tion. At the end of the period, the residual solution was analyzed and compared with the original solution. The concentrations of potassium and nitrate were greatly re- duced, potassium almost to zero, while calcium, magnesium, and sulphate ions were present in the residual solutions in slighdy higher concentrations than in the original solution. Obviously, solutes were absorbed much more rapidly than water with re- spect to the first concentrations mentioned. Furthermore, the general balance of ions was completely changed by the absorp- tive processes of the plant. Bicarbonate ions, present at the begin- ning in only negligible concentration, entered or were formed in the culture solution in large concentration as a result of those root activities which brought about a larger intake of anions than of cations. The metabolic activities of the root system, through their relation to the absorption of mineral solutes, thus effectively and continuously disturb the equilibrium conditions of the solid and solution phases of the soil. The conclusion has been generally accepted in recent years by plant physiologists that absorption of mineral solutes is inde- pendent of transpiration, but much of the earlier evidence for this conclusion (growing plants in shade and sunlight, and de- termining ash content, and the like) has not permitted the same unambiguous interpretation as may be placed on the results from simpler experiments of the type referred to just above. A satisfactory understanding of the phenomena requires an appreciation of the role of root metabolism in accumulation of mineral solutes from dilute solutions. Although the independ- ence of the processes of transpiration and of mineral solute ab- sorption is established, it has not been proved that transpiration does not indirectly influence absorption of solutes over an ex- tended period. A certain upper limit of concentration exists for The Plant as a Metabolic Unit 237 root cells, no matter how intense their metabolism, and the con- tinuance of absorption from a nutrient solution depends upon movement of solutes from the root to the shoot. This raises once more the question of the path of translocation of mineral solutes. Doubdess, most investigators in the field of botany feel that the assumption may be safely made that these solutes are always moved upward in the transpiration stream, but this assumption is not accepted by all investigators. Curtis* has presented evidence with respect to woody plants which indi- cates very strongly, he thinks, that movement occurs primarily in the living cells of the phloem and not in the xylem system. Experiments with certain annual plants, however, lead to the conclusion that movements of inorganic solutes of large magni- tude do occur in the xylem system, and this may be the chief, possibly the sole, path of conduction in the plants concerned. If so, transpiration would have a function in the removal of solutes from the root system and would therefore influence the effi- ciency of root activities in the process of absorption of mineral solutes. Whatever may be the path or the mechanism of trans- location, growth and metabolism must be concerned in the ac- cumulation or utilization of mineral solutes in green cells just as in root cells. The distribution of mineral elements in different tissues cannot be explained without reference to the metabolic activities of living cells, in relation to solute accumulation, and to the effects of climatic environment upon these activities. With respect to another aspect of translocatioii, it has been demonstrated that a very direct relationship frequendy exists between root pressure and rate of accumulation of mineral ele- ments from solution.* The rapid accumulation of electrolytes by the root system establishes an osmotic gradient resulting in * Unpublished data by D. R. Hoagland and T. C. Broyer. 238 D. R. Hoagland exudation from cut surfaces or in guttation, under appropriate atmospheric conditions. The guttation water, and especially the exuded sap, from cut stems, may contain certain solutes in much higher concentrations than those of the culture solution. Upward movement of mineral solutes can therefore occur under the in- fluence of the forces of root pressure, closely related to the meta- bolic activities of root cells. The results of some recent experi- ments suggest that such solutes enter into the xylem system by a polarized process involving metaboUc activities of root cells. Soil and Plant Interrelations Having surveyed in a general way the nature of the process of absorption of mineral elements by plants, we must now give some further consideration to certain interpretations in terms of soil-plant relations, without, however, discussing the chemical and physical system of the soil— a subject of extreme complexity. The plant is generally compelled, by virtue of the properties of the soil environment, to remove essential mineral elements from dilute solutions. Some of the essential elements will ordi- narily be present in the soil solution in extremely low concen- trations. It is necessary for the rapidly growing plant to acquire certain of these elements at rates which are extremely high in comparison with those of the concentrations available. The plant does in fact function as an active concentrating agent. (Its re- markable efficiency in this respect has been referred to earlier.) Active root growth also constandy brings into play new soil sur- faces which yield solutes to the plant, and this exploration of the soil is likewise dependent upon metabolic activities of the whole plant. These considerations apply not only to the essential ele- ments of older history, but also to other essential elements only recendy recognized as essential, such as boron, copper, and zinc. The Plant as a Metabolic Unit 239 The respiration of root cells, an activity which must proceed in order that accumulation of solutes may take place, is also a source of hydrogen ions, which can bring about the solution of bases or of other elements from the solid phase of the soil. Great importance thus is assigned to the displacement by hydrogen ions of calcium, magnesium, and potassium from those colloidal complexes of the soil in which the properties of base exchange chiefly reside. In this displacement we find a very direct relation between metabolic activities of root cells (also, therefore, an indirect relation with activities of leaf cells) and the dissolving of certain essential elements from the solid components of the soil system. There are thus two important general consequences of root metabolism as related to nutrition with respect to cations — production of hydrogen ions, and disturbance of equilibrium through absorption of bases by root cells. The latter process is of course selective, potassium ions being withdrawn usually at a far greater rate than calcium or magnesium ions. Among the anions absorbed by plants, nitrate is ordinarily of greatest quantitative importance and this ion has a special sig- nificance for soil-plant interrelations as determined by metabolic activities of the plant. The nitrate is often removed from solution in excess of bases, leaving bicarbonate or organic acid ions derived from the tissue in concentration equivalent to that of the excess nitrate removed. In effect, an exchange of ions occurs. Nitrate ions may be removed from solution and concurrently undergo reduction, or they may be accumulated in the cell sap as nitrate ions and subsequendy reduced. Recent experiments on barley plants indicate that the accumulation of nitrate ions as such is controlled by the same metabolic activities that control accumulation of other ions. As conditions become less favorable for aerobic processes, the removal of nitrate ions from solution (as 240 D. R. Hoagland differentiated from accumulation of nitrate ions in the cell sap) may be less affected than the removal of other ions, the reason ap- parently being that the absorption of nitrate accompanied by re- duction does not depend upon an oxygen supply in the same way that the accumulation of nitrate and other ions does. Accord- ingly, the extent of the selective removal of nitrate may be a function, not merely of the character of this ion, and of the asso- ciated basic ions, but also of the kind of metabolism occurring in root cells.* Furthermore, the experiments of Lundegardh and Burstrom,' and of investigators in California, show that respira- tion may be accelerated in the process of absorption of certain ions; particularly the nitrate ion, as results obtained in this labo- ratory make plain. This again illustrates the complexity of the reciprocal relations existing between the soil and the plant. In the accumulation of mineral elements by root cells, inter- ionic relations play an important role. It has been more than sufficiently shown that these relations are not limited to strong solutions, nor to interrelations of cations. The explanation of the results cannot be given satisfactorily in terms of a simple physicochemical system, for the reason that we are dealing with a dynamic system in which the nature of many of the metabolic reactions is still unknown. Interionic effects, nutrient deficien- cies, and the relative rates of absorption of ions of varied char- acter should be reinvestigated with reference to oxygen and available carbohydrate supply to the root system, temperature of soil, and related factors. Such an investigation must form a part of any comprehensive plan of attack on the nature of soil-plant interrelations. * The activities of microorganisms living on root tissues require additional study, but at present the metabolism of the root cells seems to be primarily concerned. The Plant as a Metabolic Unit 241 Accumulation of Mineral Solutes and Plant Buffer Systems It is beyond the scope of this contribution to enter into a general discussion of the function of mineral elements in plant growth, but some comment on the mineral nutrition of the plant in re- lation to the buffer system of the cell sap is useful, as illustrating another problem of soil-plant systems. There is a fundamental difficulty in the quantitative examina- tion of the bufiFer systems of complex plant tissues in that some method of expressing sap from such tissues must be employed and the resultant fluid is a composite and to some extent altered sap. Nevertheless, important information has been gained. The tendency to maintain specific hydrogen ion concentrations, within a rather narrow range, is very marked in many plant tissues, as can be observed even under the highly imperfect condi- tions of present experimentation. Phosphates and salts of organic acids are of especial importance in most plant buffer systems. The metabolic activities of the cell determine, on the one hand, the accumulation of the mineral components of the buflfer sys- tem, and on the other hand, the transformations of organic sub- stances to provide organic acid radicles. Oxygen and available carbohydrate supply, temperature, and other factors already dis- cussed, are essentially concerned in both aspects of the processes which operate to maintain a suitable buffering system in the sap. The predominate anion capable of ready metabolic trans- formation is the nitrate ion. Nitrate and bases enter the plant and nitrate undergoes more or less rapid reduction followed by the synthesis of organic forms of nitrogen. A basic residue is left and this is neutralized by organic acids. The nature of the mineral and organic complexes of the soil determines the inherent supplying power of the soil for basic 242 D. R. Hoagland and acidic ions, but the metabolic activities of the plant govern the rates at which absorption can occur. The maintenance of suitable internal hydrogen ion concentrations in plant tissues is dependent on the proper functioning of the system as a whole. The measurement merely of the reaction of the external solution gives but a very faint insight into the role of hydrogen ion con- centration in plant growth, and undue simplification of the prob- lem may readily lead to fallacious conclusions. General Conclusions This survey is very incomplete, yet it is believed that the evidence presented by way of illustration suffices to establish the general thesis proposed in the Introduction. Many problems of the soil- plant system can be reoriented if the view be adopted that the plant not only occupies the soil, but also constitutes a metabolic unit in the whole environment, aerial and subterranean. This is not a vague and sterile generalization, but on the con- trary, one which leads to specific and fruitful suggestions for future experimentation. The soil is regarded as a medium which must provide a suitable environment for root metabolism. Soil temperature and aeration assume a definite importance in the study of solute absorption. More specific reasons than have been apparent heretofore are found for investigating the relation of leaf metabolism and of the demands of reproductive processes to the metabolism of root cells. The indirect effects of the quality, intensity, and duration of light on root development are recog- nized as requiring far more intensive study than has yet been accorded to them. Transpiration becomes of interest in relation to changes in the concentration of solutes in the soil solution and to the aeration of the soil. Transpiration or other factors influ- encing rate of movement of mineral solutes out of the root sys- The Plant as a Metabolic Unit 243 tern indirectly modify the rate of intake from the soil solution and the problem of translocation requires investigation with this in mind. There is need of further knowledge of carbon dioxide production by root cells, as influenced by soil aeration, soil tem- perature, carbohydrate synthesis and translocation, and related factors. The discussion may be appropriately closed by emphasizing again the general view that problems of the soil-plant system cannot be solved without the aid of patient and laborious experi- mentation under conditions which permit the requisite control of the plant's environment and the quantitative measurement of processes or of their end products. Such researches are indis- pensable to the development of scientific knowledge in the field of geobotany or of practical agriculture. Summary The metabolic activities of the plant are reviewed with special reference to the absorption of mineral elements. Methods of con- trolled experimentation are considered and certain results cited. In connection with the process of solute accumulation, emphasis is placed on the supply of carbohydrates available to root cells, and on the temperature and oxygen-supplying power of the root medium. Relations of transpiration to absorption of mineral solutes, relation of absorption of these solutes to plant bufiFer systems, and other interrelations of the soil-plant system are discussed. LITERATURE CITED ' BouiLLENNE, Ray, and Went, F. 1933. Recherches cxperimentales sur la neoformation des racines dans les plantules et les boutures des plantes superieures. Ann. du Jardin Botanique de Buitenzorg, 43:1-178. * BouiLLENNE, Ray, and Prevot, P. 1934. Recherches cxperimentales sur le phenomene de neoformation chez Begonia Rex Putz. Bull, de la Classe des Sciences, Acad, royale de Belgique, 20:723-737. ' Cannon, W. A. 1932. Absorption of oxygen by roots when the shoot is in darkness or in light. Plant Physiology, 4:673-684. * Curtis, O. F. 1929. Studies on solute translocation in plants. Experiments indicating that translocation is dependent on the activity of living cells. Am. Jour. Bot., 16:154-168. ° Dams, A. R., and Hoagland, D. R. 1928. An apparatus for the growth of plants in a controlled environment. Plant Physiology, 3:277-292. ' Hoagland, D. R. 1930. The accumulation of mineral elements by plant cells. Contributions to Marine Biology (Stanford Univ. Press), 1 31-144. ' LUNDEG.\RDH, H. 1 93 1. Environment and plant development (Edward Arnold and Co., Lon- don). ® LuNDEGARDH, H., and Blt<.str6m, H. i933.Atmung und lonen Aufnahme. Planta-Archiv. f. wiss. Botanik, 18:683-699. ' Osterhout, W. J. V. 1933. Permeability in large plant cells and in models. Ergebn. d. Physiologic und experimentellen Pharmakologie, 35:967-1021. ^° Steward, F. C. The absorption and accumulation of solutes by living plant cells: 1932. 1. Experimental conditions which determine salt absorption by a stor- age tissue. Protoplasma, 15:29-58. 1932. II. A technique for the study of respiration and salt absorption in stor- age tissue under controlled environmental conditions. Ibid., 15:497- 516. [244 ] The Plant as a Metabolic Unit 245 1932. III. (with Wright, R. and Berry, W. E.) The respiration of cut discs of potato tuber in air and immersed in water, with observations upon surface: volume effects and salt accumulation. Ibid., 16:576-611. 1932. IV. A quantitative interpretation with respect to respiration and salt absorption. Ibid., 17:436-453. 1933- V. Observations upon the effects of time, oxygen, and salt concentra- tion upon absorption and respiration by storage tissue. Ibid., 18:208- 242. 1934. VI. (with Berry, W. E.) The absorption of potassium bromide from dilute "solutions by tissue from various plant storage organs. Ann. Bot., 48:395-410. Malaysian Phytogeography in Relation to the Polynesian Flora By E. D. Merrill IN ANY GENERAL TREATMENT o£ the flora of Polynesia, whether this be merely a taxonomic consideration or a discussion of the phytogeography of the region and the relationships of the flora to that of surrounding lands, it is manifest that the vege- tation of the islands that form the western boundary of the Pacific is a major factor to be considered. At the present time, representatives of between 1400 and 1500 more or less generally recognized genera of flowering plants and ferns are known from Polynesia. However, this estimate includes New Caledonia, which is remarkable for its very specialized flora, and comprises the very numerous exotics introduced by man. More than iioo of these genera occur also in Malaysia. Essentially, the flora of Polynesia, whether we consider that of the low or the high islands, is Malaysian in origin, with certain characteristic Australian elements which also occur in Malaysia and particularly in Hawaii, with a small admixture of American forms. No single family of plants is confined to Polynesia, and the number of endemic genera is relatively small. Specific en- demism in the low islands is very limited, most of the species being those of wide tropical distribution. With the exception of the weeds and cultivated plants introduced directly or indirectly by man, the flora of the low islands is chiefly composed of species having special adaptation for dissemination by ocean currents. However, in each of the groups of high islands in Polynesia there is a distinctly high specific endemism, which is in general true of oceanic islands, but the relationships of these endemics [ 247 ] 248 E, D. Merrill are in large part with the genera and species characteristic of the tropical lands bordering the western Pacific. The purpose of this paper is not to theorize concerning how or when these vari- ous types, or their ancestors, reached Polynesia, but rather to point out certain obvious facts in present-day distribution and briefly to discuss the bearing that our limited knowledge of the geological history of the Malaysian region has on this problem. The so-called monsoon region in the Asiatic tropics has its typical flora, and no sharp distinction may be drawn between that of tropical Asia and that of the great archipelago that lies between Asia and Australia. Whether many of the genera char- acteristic of this region originated in what is now continental Asia and migrated into Malaysia, or vice versa, is a question that can scarcely be settled definitely and is perhaps of slight impor- tance; unquestionably there have been extensive intermigrations in both directions. We do know, however, that there are a num- ber of characteristic Asiatic (continental) types in Malaysia, occurring at both low and high altitudes, and that these types, as a rule, do not reach Polynesia. Some of the numerous Asiatic genera having indigenous representatives in Malaysia are : Taxus, Pinus, Agrostis, Anthoxanthum, Bromus, Deschampsia, Muhl- enbergia, Poa, Asparagus, Lilium, Ophiopogon, Salix, Ulmus, Arenaria, Anemone, Ranunculus, Thalictrum, Sedum, Astilbe, Deutzia, Sycopsis, Photinia, Rosa, Gleditschia, SJ^immia, Buxus, Sarcococca, Pistacia, Berchemia, Hypericum, Viola, Daphne, Ac- anthopanax, Sanicula, Primula, Lysimachia, Fraxinus, Ligus- trum, Acer, Gentiana, Sweertia, Trigonotis, Vaccinium, Rhodo- dendron, Gaultheria, Ajuga, Mo si a. Salvia, Alectra, By t ho phy- lum, Hemiphragma, Sopubia, Lonicera, Peracarpa, Ainsliaea, Anaphalis, Aster, Cirsium, Lactuca, and Solidago. There are many others both at low and at high altitudes. Malaysian Phytogeography and Polynesian Flora 249 The intermigrations of the early angiosperms in what is now tropical Asia and Malaysia doubtless took place at a time when the continental area extended far to the south, probably in the late Cretaceous and early Tertiary, and this is the period when the more widely dispersed Asiatic types, or the ancestors of cer- tain Asiatic types now found in New Guinea, extended their ranges. There seems to be little or no evidence of direct Asia- New Guinea land connections since the early Tertiary. The evidence is that, in later times, there were an eastern and a western route of intermigration as between Asia and Malaysia, the former from southeastern Asia and Formosa south through the Philippines at least as far as Celebes, and the latter from India through Burma and the Malay Peninsula to Java, Sumatra, and Borneo, with apparently a secondary paralleling line through what is now the Andaman Islands to Sumatra and perhaps Java. These routes were doubtless in operation at a relatively early time, probably in the Tertiary, so far as the ancestors of our modern flora were concerned. At a later period, particularly from the Pliocene-Pleistocene into the Recent, there must have been a very active interchange of plants between Asia and west- ern Malaysia, when the Sunda Islands (Java, Sumatra, and Bor- neo) and the Palawan-Calamian group in the Philippines were at times definitely a part of the Asiatic continent. The Malay Archipelago lies wholly within the tropics. This great equatorial group of islands extends from the northwest to the southeast, from northern Sumatra to eastern New Guinea, a distance of about 4000 miles. Throughout this vast region the climatic conditions are essentially uniform; for the most part, precipitation is ample and, in certain areas, even extreme, and the average temperatures at equal altitudes are strikingly uni- form. In the Malay Peninsula, Sumatra, Java, Borneo, the larger 250 E. D. Merrill islands of the Philippines, Celebes, the Moluccas, and New Guinea, there are numerous mountains from 5000 to 8000 feet high, with peaks at least 10,000 feet high in Sumatra, Java, Celebes, Luzon, and Mindanao. The maximum elevations are 13,700 feet in Borneo and about 16,000 feet in New Guinea. These high mountains form "islands" within the tropics where, once established, species adapted to temperate and subtemperate conditions have persisted. Certain parts of the archipelago have prolonged alternating wet and dry seasons in each year, and others have a rather uniform distribution of rainfall in all months of the year. In the great islands, there are no regions that approach arid or even semiarid conditions, and there is perhaps only one part of the entire region where the restricted annual rainfall appreciably affects the vegetation. This part comprises some of the Lesser Sunda Islands. The rainfall — its amount and its seasonal distribution — is the most important single factor, aside from altitude, which affects the distribution of vegetative types in all Malaysia. At low altitudes, the remarkably uniform temperature has had litde limiting influence on the distribution of plants. The prevailing winds, which affect vast areas for sev- eral months each year, are of minor importance to vegetative types except as they affect the seasonal distribution of rainfall. There are no striking differences in soil types except in limited areas. Therefore, neither precipitation, humidity, temperature, nor soil conditions can cause the vegetation of any major part of the Archipelago to be strikingly different from that of any other major part at a similar altitude. Yet there are remarkable differences in dominant families, genera, and species between the known flora of any two great islands, or contiguous groups of islands, particularly between eastern and western Malaysia. It is true that most of the families Malaysian Phytogeography and Polynesian Flora 251 of plants having native representatives in Malaysia are generally distributed throughout Insulind, and the same is true of many of the dominant genera. Yet certain families, such as the Diptero- carpaceae,* that dominate the primary forests of western Malay- sia, become distinctly unimportant in eastern Malaysia, and, conversely, a considerable number of genera of the Myrtaceae that occur in eastern Malaysia, fail to reach w^estern Malaysia. Some of these genera are Eucalyptus, Mearnsia, Xanthostemon, Osbornia, and Xanthomyrtus (the last two also are found in Borneo). Furthermore, there is a remarkable development of specific endemism; every large, medium-sized, and even small island in the group has a very high percentage of species that, so far as we know at present, are found only on that island, or on a certain part of it, or on a restricted group of neighboring islands. Had there not been some active limiting factor, there is no evi- dent reason why a great many of these thousands of local species should not now have more general geographic distribution, or why certain families and genera should be dominant in or lim- ited to specific parts of the Archipelago. This limiting factor seems to be found in the geologic history of Malaysia. From the Tertiary, a more or less continuous insular area existed in some part of this region, the straits and arms of the sea inhibiting the natural dispersal of a great many species of plants and animals. We know little about the climatic conditions of Malaysia in past geologic times, but we may deduce from the present-day vegetation that for a long time the climate has not been strikingly different from that which obtains today. Certainly there have been no periods in which the low-altitude vegetation was affected * Merrill, E. D., 1923. Distribution of the Dipterocarpaceae. Origin and rela- tionships of the Philippine flora and causes of the differences between the floras of eastern and western Malaysia. Philip. Jour. Sci., 23:1-33, /. 1-8. 252 E. D. Merrill by cold weather, and no extensive areas in which arid or semi- arid conditions prevailed. In other words, since the beginning of the development of our modern vegetation, Malaysia as a whole apparently has been a continuously tropical region with a re- markably equable climate, ample rainfall, and relatively high humidity. From the limited paleobotanic record, we know this has been true of certain parts of the region in which the Pliocene- Pleistocene plant-remains perfectly match the plants of today, but we can only surmise that this is true for the entire region. Setchell* has suggested that persistently dry areas, or the alter- nating wet and dry seasons characteristic of certain parts of the Archipelago, may be just as important to plant distribution as the geologic history of the region, a suggestion that cannot be ignored; yet it is significant that we find throughout Malaysia, more or less in alternation, constantly wet and alternatingly wet and dry regions; for example, western Java is contrasted to eastern Java, and all the east coast of the Philippines to the west coast. Even in regions dominated by alternating wet and dry seasons there are found certain areas, particularly in the vicinity of high mountains, where precipitation is more or less equably distributed through all months of the year; such areas occur on almost every island in Malaysia which has mountains 4000 to 5000 feet high or higher. It has been known for many years that the fauna of the Sunda Islands is essentially Asiatic, and that as one proceeds eastward the Asiatic mammals, fresh-water fishes, and certain groups of birds, reptiles, and insects become less and less evident. On the contrary, there are various striking Australian types of insects, birds, and other animals in New Guinea and neighboring ♦Setchell, W. A., 1929. The Wallace and Weber Lines: a suggestion as to climatic boundaries. Proc. Fourth Pacific Sci. Congr., 3:311-321. Malaysian Phytogeography and Polynesian Flora 253 islands. In general, this same statement applies to the plants. Mere proximity to Asia or Australia is, in part, the cause of this condition, but a brief consideration of geologic history will pro- vide a fuller explanation. The Tertiary was the period of the development and dissem- ination of mammals. The mammals were very poorly repre- sented in Australia, where their place was taken by the more primitive marsupials. It seems evident that, in earlier geologic times, Australia was definitely united with Asia, but was sepa- rated from it before the mammals became dominant. This in part explains the plant and animal forms, long-continued isola- tion favoring local development uninfluenced by direct migrants from Asia. In pre-Tertiary times, intermigration was possible until the late Cretaceous; since that time most northwestern- southeastern intermigrations have been inhibited by the con- stant archipelagic conditions of certain parts of Malaysia, and since the late Cretaceous or early Tertiary there apparendy have been no direct land-connections between Asia and Australia. In the Tertiary, there were great physical changes in a large part of Malaysia, as is evidenced by the greatly developed Tertiary sedi- ments in the Archipelago, but always, somewhere, there were apparently large land areas in existence. MolengraaflF has clearly indicated the probable geologic history of this region since the Tertiary.* There have been two fairly stable continental areas and an intermediate unstable insular area. The Asiatic bank, or Sunda shelf, extending southward from Asia, comprising approximately 1,850,000 square kilo- meters, carries upon it the Malay Peninsula, Sumatra, Java, and the smaller islands eastward to Bali, Borneo, and the Palawan- * Molengraaff, G. A. F., 1921. Modern deep-sea research in the East Indian Archipelago. Geogr. Jour., 57:95-121. f.1-9. map. 254 ^' ^' ^^^^^^^ Calamian group in the Philippines. In its fauna and flora this entire region is essentially Asiatic or, we might say, Malay- Asiatic. Marked by the 200-meter isobath, the eastern boundary of this shelf is, in its southern part, the famous Wallace's Line, so named by Huxley, supposedly demarking the limits of the Asiatic and the Australian faunas and a part of the eastern boundary of the ancient Malay-Asiatic continent, or Sundaland. The Australian bank, or Sahul shelf, much smaller than the Asiatic one but having practically the same history, carries upon it the great island of New Guinea. The western boundary of this ancient continent, or Papualand, approximates the position of Weber's Line, proposed by Pelseneer as a substitute for Wallace's. The insular area between Wallace's and Weber's Lines, includ- ing the Lesser Sunda Islands, the Moluccas, Gilolo, Celebes, and all the Philippines except the Palawan-Calamian group, has had an entirely different history. All or part of this intermediate region has been insular since the beginning of the Tertiary. There are no large shelf seas; on the contrary, the straits and seas between the various islands are marked by great deeps. This entire region has been subject to several great elevations and depressions. It has been and still is unstable, and, in contrast to the more stable Sundaland to the west and Papualand to the southeast, it is orogenetically active. Its flora and fauna are made up in part of relic species and their descendants, and in part of infiltrations from the west, north, and southeast. East of Sumatra, Java, and Borneo, Asiatic types sharply decrease; many do not extend to this intermediate area, and still more do not reach New Guinea. West of New Guinea, the Australian types decrease and most of them do not reach Sundaland, although a striking series has attained a northward distribution into the more distant Philippines. Malaysian Phytogeography and Polynesian Flora 255 In relation to Polynesia, it is naturally the eastern part o£ Malaysia that is o£ the greatest biological interest, for it is from this region that most of the Malaysian types of plants now found in Polynesia have been derived. There has been in Polynesia a certain amount of dissemination of plants by v^ater, but the number of species having adaptations to this end is very limited. Other plants have been disseminated through the medium of winds and migratory birds. There are, however, many genera characteristic of Malaysia-Polynesia that have no manifest adap- tations for any of these methods of dissemination. One is forced to postulate a different distribution of land areas at some time in the past, to explain the present-day distribution of plants and animals. These land areas probably were of considerable size, not the narrow land-bridges which have been widely scattered over the Pacific basin by some theorists. Effective north and south distribution is limited by tempera- ture, but within the humid tropics this factor is inoperative. Thus, except for a very few high-altitude plants in Luzon, there is no evident connection between the Philippine and the Japa- nese floras. Even the species distributed between Formosa and the Philippines are few in number and none is dominant in either region; and, strikingly, none of the characteristic Papuan and eastern Australian types, so well developed on the Philip- pines, reaches Formosa.* The north and south distribution in eastern Malaysia is more extensive than is generally realized. Such characteristic species as Uncinia rupestris Raoul and Blechntim fraseri Luerss., origi- nally known only in New Zealand, now stand thus : the Uncinia, New Zealand and the Philippines; the Blechnum, New Zealand, * Merrill, E. D., 1923. Die pflanzengeographische Scheidung von Formosa und den Philippinen. Bot. Jahrb., 58:599-604. 256 E. D. Merrill the Philippines, and Borneo. Among genera, about eight species of Ascarina are distributed in New Zealand, New Caledonia, and western Polynesia, and two species in the Philippines; about twenty-three species of Quintinia in New Zealand, New Cale- donia, Australia, and New Guinea, and one in the Philippines; about ten species of Mooria {Cloezia) in New Caledonia and one in the Philippines; about twelve species of Xanthostemon in New Caledonia, three in Australia, two in New Guinea, one in Celebes, and three in the Philippines; three species of Phyllo- cladus in New Zealand, one in Tasmania, and one in the Philip- pines and northeastern Borneo. Numerous eastern Australian types extend into eastern Malaysia as far north as the Philip- pines, including representatives of such diverse genera as Centro- lepis, Thysanotus, Pater sonia, Phrygilanthus, Drimys, Clianthus, Pleiogynium (also in Fiji and Vavau), Stac\housia, Campto- stemon, Eucalyptus, Oshornia, Pimelea, Xanthostemon, Calo- gyne, Microlaena, Didiscus, and others, with a few more wide- ly distributed. Such eastern Malaysian-Philippine genera as Spiraeopsis, Wallaceodendron , Reinwardtiodendron, Neotrewia, Strophioblachia, Myrmephytum, Sararanga, Heterospathe, Po- thoidium, Monophrynium, Papualthia, Cubilia, Euphorianthus, Tristira, Schuurmansia, Everettia, Mearnsia, Anompanax, Di- morphanthera, Loheria, and Nycticalos range from the Philip- pines to Celebes and the Moluccas, and some to New Guinea. It is significant that, although all these genera have species that are more or less widely distributed in the Philippines, few extend to western Malaysia and still fewer to the Asiatic con- tinent. Among those that have the more extended range are: Drimys (Borneo), Centrolepis (Borneo and Indo-China), Thy- sanotus (Malay Peninsula, Indo-China, and China), Clianthus (Indo-China and Hainan), Stac\housia (Sumatra and Caroline Malaysian Phytogeography and Polynesian Flora 257 Islands), Calogyne (China), Osbornia (Borneo), Didiscus (Bor- neo), and Strophioblachia (Indo-China). No representative of any of these genera reaches Formosa, yet most of them extend to northern Luzon and some even into the Batan and Babuyan Islands, between Luzon and Formosa. In general, the eastern Malaysian types have not extended north of the Philippines, and their western migration also stop- ped there, as the eastern migration of many western Malaysian types ceased in Borneo, the Palawan-Calamian group in the Philippines, and the Philippines proper. This is entirely in accord with the supposed geologic history of the Malaysian region. The constantly insular area in the un- stable region lying between the Lombok Passage and Macassar Strait, extending north along the western boundary of the Philip- pines to the west and east to the islands off the western end of New Guinea, including the Lesser Sunda Islands, Celebes, the Moluccas, Gilolo, and all the Philippines except the Palawan- Calamian group, effectively inhibited a general east and west migration across this region, not only in the Recent and the Pleistocene, but also during the Pliocene and probably some of the Tertiary. Undoubtedly there was some late east and west dis- tribution across the southern part of this unstable insular area but, in general, the routes of migration seem to have been north- east from Sumatra, Java, and Borneo into the Philippines, and thence south and southeast to Celebes, Gilolo, and New Guinea, and vice versa. The species that succeeded in extending their ranges apparendy were forced to travel along the two longer sides of a triangle, its base approximately in the position of the Lesser Sunda Islands and its apex somewhere in Luzon. When a genus is described from material collected in a certain place and is known only from that region for many years, we 258 E. D. Merrill more or less automatically accept it as a group characteristic of that region. If a representative of it is later found in another area, we are apt to consider it as an extraneous entity there. Euca- lyptus is such a genus. It is tremendously developed in Australia, has a very few species in New Guinea, and one which extends to the Bismarck Archipelago, the Moluccas, Celebes, and the Philippines. We are justified in accepting it as an Australian ele- ment in the other regions. The same is true of the few phyllod- inous species of Acacia outside of Australia. The one native species of Ulmus in northern Sumatra, two species of Pinus in Sumatra and the Philippines, and one species each of Taxus and Gleditschia in the Philippines and Celebes may be unquestion- ably accepted as Asiatic (continental) types in Malaysia. But consider the following genera: The genus Vavaea was described in 1846 for a single species from Vavau in the Tonga Islands. When the first Philippine species was found, it was considered to be a Polynesian element in the Philippine flora; and yet seven species are now known in the Philippines, one in Java, three in New Guinea, one in the CaroHne Islands, and still only three in Polynesia. This supposed Polynesian element in the Philippines and Malaysia is really a Malaysian element in Polynesia. Couthovia was even more mistakenly located. This genus was proposed in 1858 for two species, one from Hawaii and one from Fiji. In 1888 a third species was described from New Guinea, and ten years later a fourth from Celebes which, four years after, was found in the Philippines. Thus we had a "Polynesian" element in the New Guinea-Philippine-Celebes flora until the New Guinea flora be- came better known. Eleven species of Couthovia are now known from New Guinea and two from the Caroline Islands. Couthovia thus resolves itself into a Papuan genus which has extended its Malaysian Phytogeography and Polynesian Flora 259 range north into the CaroHne Islands (two species), southeast to New Caledonia (one species), east in Polynesia to Fiji and Hawaii (two species), and northwest into Celebes and the Philippines (one species). Dolicholobium was described in i860 from Fiji. At present five species of it are known in Fiji, eight in New Guinea, one in the Solomon Islands, and one in the Philip- pines. Like Vavaea and Conthovia, this genus seems to be not so much a "Polynesian" as a Papuan type which apparently has extended its range north from New Guinea into the Philippines, as well as east into Polynesia. May not such genera as Joinvillea and Tetraplasandra, for a long time known only from Polynesia, have the same history .f' The first is now also known from the Malay Peninsula, Sumatra, Borneo, and the Philippines, and the second from the Philip- pines and Celebes. As our knowledge of the tremendously rich and really not well-known Malaysian flora increases, we may expect to find, particularly in eastern Malaysia, various other generic types now considered to be Polynesian. Future investigations may radically modify our beliefs con- cerning the geologic history of this region, but in general the present-day distribution of both plants and animals is in support of the geologic history as it is now understood. There is no sharp line separating the floras and faunas of eastern and western Malaysia nor those of Asia and Australia, except for limited groups; all proposed lines break down more or less when all groups of plants and animals are considered. Apparently both eastern and western Malaysian elements have introduced into the unstable insular area separating the more stable Sundaland and Papualand, which is what logically would be expected when an insular group is situated between two continental areas and the islands are separated by relatively narrow arms of the sea. 260 E. D. Merrill It may be that Wegener's hypothesis of continental shift has its bearing on both the separation and the intermingling of the AustraUan and Asiatic faunas and floras in Malaysia, if it be assumed that Australia, formerly more widely separated from Asia, moved toward Asia and, in so doing, crushed the south- eastern borders of an ancient, larger Asiatic continent. Before the shift took place, intermigrations were impossible; after the shift, plants and animals could move to a limited degree in both direc- tions. This hypothesis has been applied to the problem of plant distribution in Malaysia by Doctor Lam.* We can only theorize concerning the underlying causes for the peculiar distribution of Australian types, which extend into Poly- nesia and eastern Malaysia, including the Philippines, but do not reach to any marked degree into western Malaysia. Our knowledge of the tremendously developed Malaysian flora and fauna is limited, and any conclusions based on the data now available concerning the distribution of known genera and spe- cies may be radically modified as exploration progresses and the compilation of geographic distribution data becomes more com- plete. Conclusions based on the known distribution of any one group, be it genus, family, or any other category, are notoriously unsafe. Conclusive generalizations can be safely based only on the consideration of all groups, and here the specialist fails be- cause of his inability to master and evaluate the details of geo- graphic distribution of organisms outside of his own special field. It is safe to assume, however, that, whatever have been the causes leading to the present-day distribution of Malaysian life, no investigator of the fauna and flora of Polynesia can safely * Lam, H. J., 1930. Het genetisch-plantengeografisch Onderzoek van den indischen Archipel en Wegener's Verschuivingstheorie. Tijdschr. Nederl. Aard- drijksk. Genootsch., 47:553-581; niaps 6-7. Malaysian Phytogeography and Polynesian Flora 261 ignore the biology of Malaysia from which, as a center of origin and dispersal, it seems certain that a large number of the present Polynesian types of plants and animals have been derived. Within Malaysia, in large measure because of its geological history, there developed two great centers of origin and dis- persal; one was the Borneo-Java-Sumatra-Malay Peninsular part of ancient Sundaland, the other the New Guinea part of ancient Papualand. Plants and animals extended their ranges from these two centers: from the first, most of them went north into the Asiatic continent and northeast into the Philippines; and from the second, they went north through Gilolo, the Moluccas, and Celebes into the Philippines, south into eastern Australia, and east into Polynesia. Plant Communities of the World By Eduard Rubel COMMUNITY is the sociologic unity of any rank from the lowest to the highest. In a short synopsis only the widest ranks can be discerned, but one must tell how one builds them up and understands them. As fundamental unit one has the "association!' Its definition, proposed by Flahault and Schroter and adopted by the International Botanical Congress of 1910, held in Brussels, runs, "An association is a community of certain floristic composition, of uniform habitat conditions, and of uniform physiognomy!' This definition has proved very satisfactory, and is better than all the others which since have sprung up here and there. By composition is meant not only the list of species; it includes their abundance, dominance, sociabil- ity, constancy, fidelity, vitality, periodicity, rhythm. The con- stants and characteristics are the main feature; they form the normal characteristic combination of species. Something so very ecologic as an association can only have its conditioning cause from an environment of a certain ecologic uniformity. Phys- iognomy is the effect of habitat. Habitat is the ecology of en- vironment; physiognomy is the ecology of the vegetation itself, which expresses itself by its morphology, including its anatomy (life form of the dominants) . Species combine to genera, these to families (characterized by the ending -aceae), these to orders, ending in -ales (femininum plurale), to classes, and to other subdivisions. Quite analogous associations, ending in -etum, combine to a "federation" as Du- Rietz proposed to the International Congress, or "alliance" as Braun-Blanquet calls it, with the ending -ion (originally pro- [263] 264 Eduard Riibel posed by C. E. Moss in 1910). The following unity is the associa- tion "order" ("subformation" DuRietz) with the ending -etalia (neutrum plurale) introduced by Koch. The recognition of the community orders of the world is not yet far advanced; very much remains to be done. Gapless further unions are not yet possible. With the highest unities of communities the situation is different; they come from another line of research — ^from the attempt at classification. This is an old task. Here, as everywhere in every science, works the fundamental need of the human mind for order, for grouping, so to be able to apprehend the infinity of things existing side by side. Only in orderly fashion can mind render account of things. An investigator needs a synthesis of his field of research, a system. We have systems in philosophy, we have systems of the plants of the world; we need a system of the plant communities of the world. A plant community is something eminendy ecologic; to ecol- ogy, then, must be paid the first attention in arranging a system. A very essential point is connected with the law of minimum and the replaceability of factors. The law of minimum, founded by Liebig for cultures, is also valid for natural plant communi- ties— and for one's own self-discernment. Any factor can only be effective when other factors also act sufficiently. It decides the factor which is present in a minimum. A soil may be highly nutrient, yet the nutrient substances do not work if moisture is lacking. Water and sun accomplish nothing if there is no nutri- ment. The minimum factor becomes the deciding one. Which it is, varies. Therefore, a classification may not be founded on a single factor, nor on a few factors. The single factor, moreover, acts difFerendy according to other factors accompanying it. Abso- lute numbers in a factor say very little; all is relative. Fifty centi- meters of rainfall acts hygromorph in oceanic Ireland, xero- Plant Communities of the World 265 morph in continental Algiers. Temperature modifies the mois- ture factor. Factors not only modify others; they may replace them. Replaceability of factors is very important. It makes it possible for a vegetation to thrive in localities where various factors have changed, and this allows the spreading of plant communities. For the investigator, it makes the characterizing of vegetation difficult. Language employs factors for character- izing. A locality is judged differently, according as we empha- size the changed or the unchanged factor. Formerly, vegetation was grouped climatically and the change of situation by edaphic factors was neglected. Edaphically conditioned vegetation did not fit. These difficulties are solved if the replaceability of fac- tors is duly taken into consideration. The replaceabilities modify thoroughly the law of minimum. A replacing factor may raise the minimum of another factor or may lower it — Roosevelt has annihilated the hampering factor of lameness by an extraordinary factor of will. A species or a community may be well adapted to a climate and thrive if un- limited soil is at its disposition. But it may lack strength to battle against the competition of invaders. Good, thriving, natural vegetation in New Zealand has been dispossessed by invaders accidently brought into the country by man — invaders which had a stronger power of competition and therefore raised higher than the existing conditions the needed minimum for mainte- nance of the old vegetation. Dispossession was the result. Taking replaceability into account, edaphic factors find their modifying value in climatic bases for classifying vegetation. The entire ecology may be more or less grasped, and the system becomes more and more natural.* * More explicitly than can be done here, but still concisely, all these questions of general sociology are treated in my Soziologie ("Geographie der Pflanzen" Handworterbuch derNaturwissenschaften, Bd. IV, p, 1044; Gustav Fischer, Jena.) 266 Eduard Riibel After trying all the systems published, I have come to prefer the following one, which is an emanation, a continuation of the system of Brockmann-Jerosch and Riibel, of 1912. The entire vegetation of the world can be included ecologically in compre- hensive formation-classes, as follows: The formation-classes of the world 1. Pluviisilvae 2. Pluviifruticeta 3. Laurisilvae 4. Laurifruticeta 5. Durisilvae 6. Durifruticeta 7. Ericifruticeta 8. Aestisilvae 9. Aestifruticeta 10. Hiemisilvae 1 1 . Hiemif ruticeta 12. Aciculisilvae 13. Aciculifruticeta 14. Duriherbosa 15. Semper virentiherbosa 16. Altherbosa 17. Emersiherbosa 18. Submersiherbosa 19. Sphagniherbosa 20. Siccideserta 2i.Frigorideserta 22. Litorideserta 23. Mobilideserta 24. Rupideserta 25. Saxideserta 26. Aquerrantia 27. Solerrantia 28. Aererrantia Rain forests Rain scrub Laurel-leaved forests Laurel-leaved scrub Sclerophyllous forests Sclerophyllous scrub Ericaceous scrub (heaths) Summergreen forests Summergreen scrub Raingreen forests (monsoon forests) Raingreen scrub Needle-leaved forests Needle-leaved scrub Hardgrass prairie Evergreen grassland (meadows) Tall herbage Emersed swamp vegetation Submersed swamp vegetation Moss moor (Hochmoor) Dry deserts Cold deserts Strand steppes Wander deserts Rock deserts Stone (and wood) deserts Phytoplankton (water floaters) Phytedaphon (soil floaters) Phytaeron (air floaters) Plant Communities of the World 2.6'j The system here given accords with physiognomic-ecologic considerations. The innate ecology is seized in order to give expression to sociologic gradation, to prograding stages of organ- ization. In this the system of communities is quite analogous to the systems of plants, which also are built up in progressive stages of organization. At the bottom of the list we put the floating communities, of very loose union. The least stable are those floating in the air, then those floating in the water, then those suspended in the ice. Also, small primitive organisms in and on the soil may be called floating. Above stands the stone and wood vegetation, still mosdy one-layered without root com- petition; above these, in turn, the rocky grounds and other deserts with some open vegetation — open, it may be, on account of the rocky ground, the mobility of the ground (dunes, screes, alluvials), the cold climate (open tundra), the dry climate. Higher stand the closed communities: first, the swampy ones — submersed aquatic, emersed aquatic, moors; then the few- layered communities on land — tall herbage, meadows, prairies, short-grass communities. The highest organized types are the several-layered woody communities of shrubs, and especially of trees, forming the complicated forest communities, differing according to the climates. For the purpose of a synopsis, it will be well to follow up these larger formation-classes, their ecology, morphology, and distri- bution. Of associations and other smaller unities, only examples can be given.* i.Pltwiisilvae: Rain forests. — ^Pluviisilvae are forests rich in epiphytes whose dominant plants consist of evergreen and often * A more detailed account of them will be found in my Pflanzengesellschaften der Erde (Verlag Hans Huber, Bern; 1930), where also a multicolored world map shows the distribution of climatic formation-classes. Brockmann-Jerosch, who did the map, has published a large wall map of the formation-classes (Justus Perthes, Gotha). 268 Eduard Riibel stem-flowering (caulifloral) trees, usually without bud protec- tion and with glabrous and glossy foliage. The climate this forest needs is hot oceanic, as found in the tropics. The tropical coun- tries have an annual precipitation of from two hundred to four hundred centimeters, distributed rather evenly the year round. The mean annual temperature is high and uniform, the differ- ence between the warmest and the coldest months being only from one to six degrees Centigrade. The vegetation looks lux- uriant. As many as four or five stories are developed. It is the phanerophyte climate of Raunkiaer, in which high-grown species are plentiful. Warming (1892) counted four hundred species of trees within three miles' distance in Lagoa Santa, Brazil. Light is a minimum factor within the forest. The leaves are evergreen, though the single leaf lives not much longer than one year ; but the defoliation is irregular the year round, so that the forest never stands bare. Rain forest is distributed in the tropics of Australia, Ceylon, the Sunda Islands, Annam, New Guinea, Central America, northern Brazil, and Africa. The sociologic unities have been studied so far only in very few localities. From British Guiana we have a good detailed study by Davis and Richards. They distinguish five consocia- tions. While in general in tropical forests the normal condition is that no one single species reaches plain dominance, four of these types have a dominant and only one follows clearly the plan of mixture. The Mora forest dominated by the strongly buttressed Mora excelsa occupies the lowest ground. It forms a broad strip along both banks of the Moraballi Creek and along the branch creeks. On the lower slopes of the smaller flat- topped hills lives the Morabukea forest type dominated by Mora gonggrijpii. This is much the darkest type of forest. The lower hilly land is covered by the mixed-forest association. Many Plant Communities of the World 269 species together form dominance; the most prominent are Pentaclethra macroloba, Licania venosa, Licania laxiflora, Li- cania heteromorpha perplexans, Eschtveilera sagotiana, and Ocotea rodioei, followed by about eighty-five less plentiful canopy-tree species. The majority of these are the same as in the Morabukea and Greenheart types. Ocotea rodioei, the Green- heart, dominates the next type, growing on reddish brown sand ''reefs!' The most distinctive type, structurally and flor- istically more sharply distinguished from the other types than they are from one another, is the Wallaba association dominated by Eperma falcata, growing on the white sand of the high hill- ridges. The number of trees to the unit of area is very large; big trees are scarce; buttressing is almost absent. From the Blue Mountains of Jamaica, Shreve gives an account of montane rain- forest with three dominant species — Clethra occidentalism Vac- cinium meridionale, Podocarpus urbanii — which make up half the tree canopy. 2. Pluviifruticeta: Rain scrub. — Pluviifruticeta are scrubs rich in epiphytes whose dominant plants consist of evergreen shrubs, usually without bud protection and with glabrous and glossy foliage. Climatically there is no reason why scrub instead of forest should prevail. Edaphic and biotic checking cause scrub under the same tropic climate. Rain scrub is very limited. In the flood region of tropic shores lives the characteristic Mangrove of shrubs and small trees. Karsten writes about them: Root sprouts break forth from the stem and from the lower branches and reach, in an elegant bend, the soil; the stem of the Rhizo- phores dies away and the "tree" stands only on its stilt roots — a very curious sight. So in spite of a short trunk these "trees" are positively of shrub rank. At high tide only the crowns are visible, at low tide one sees them standing on a dense tangle of 270 Eduard Riibel root stems. Only few species can endure the strenuous habitat condition. Two associations are well known. The American Mangrove consists only of Rhizophora fnangle, Laguncularia racemosa, Avicennia tomentosa, and A. nitida. The association grows on the tropic mud beaches and covers the shores as far north as southern Florida. The eastern Mangrove, Rhizophore- tum mucronatae, is much richer in species. It has its center in the Malay Archipelago. Impoverished association specimens are found as far distant as southern Japan and New Zealand. J. Laurisilvae: Laurel-leaved forests. — In Laurisilvae the domi- nant trees, if dicotyledons, have evergreen leaves which are usually glabrous, bright green, and at right angles to the incident light, relatively large intercellular air spaces often being pres- ent and the buds protected; if gymnosperms dominate, these bear broad evergreen leaves or scale-formed ones covering the twigs tile-fashion. These forests continue from the rain forests poleward and upward where the rain forests can no longer thrive on account of diminishing average temperature and in- creasing temperature variations, but like them they are corre- lated with regular rainfall, relatively high humidity, and uni- form temperature — in all, a temperate oceanic climate. These forests are distributed in Pacific North America, Florida, Chile, Patagonia, the Canary Islands (beautiful laurel forest), Madeira, Portugal, Cape Province (Knysna), the Himalayas, Japan, New Zealand. The misty forests of middle and northern Florida and Chilean Valdivia are of this kind. They are mosdy dominated by evergreen beeches, as, for example, Nothofagus betuloides. The climate has a mean temperature of only five to nine degrees Centigrade, but only nine degrees difference between winter and summer, with a great deal of precipitation all the year round. Plant Communities of the World 271 A number of conifer forests place themselves in the Lauri- silvae on account of their physiognomy and distribution, and apparently their ecology, which unfortunately is litde known. The inner constitution of this systematic group does not allow a real broad leaf, such as that of the dicotyledons, but they do what they can in differentiating from the needle type. Scale-like spreading and tile-like covering of the mass of closely knitted leaves make something similar to a large broad leaf, glabrous with a certain dim gloss, and at right angles to the incident light. Thuja gigantea and Tsuga heterophylla have such forms. The same result is reached by the broadening of needles and the placing of one near another like a mosaic to make up big leaves, as in Sequoia sempervirens. An ecologic anatomical research concerning these forests is highly desirable; here they have only been placed by analogy; because in distribution and habitat they are similar to the well-known dicotyl laurel woods, it is assumed that they need a similar climate. Such predictions have proved right in the past (e.g., concerning heaths) ; may future studies prove this one right also. Examples of these forests are found in the northwestern United States, in the Coast Range and the western Cascades with their mild winter (Seatde, mean January temperature, +4.3°), cool summer (mean, July, 17.7°), small difiFerence (13.4°), great quantity of fog (only 25 to 40 per cent of the possible sunshine), and grizzly "Scotch" rain. A dense canopy of giants with elegantly hanging branches makes a wonderful impression. The hemlock, Tsuga heterophylla, with a mosaic of parted needles and a sideways swing of branches, gathers all possible light in the dim, dusky wood. Still broader in foliage is the scale-leaved Thuja gigantea. In the Cascades, fire has not rarely destroyed these associations and a Douglas fir forest has sprung up as fire succession. The laurel leaf also 272 Eduard Riibel occurs in the undergrowth. Of evergreens I noted, for example, Gaidtheria shallon, Linnaea americana, Berberis nervosa, Viola semper vir ens, Chimaphila umbellata, Vaccinium par vi folium, and Pyrola bracteata. I should like to know whether the western coast of Canada and Alaska have such forests, with ecology and sociology to rank them among Laurisilvae. An ardent desire for studies! Near San Francisco is the Sequoietum semper virentis in the coastal mountains where cold sea-winds produce a fog belt for the greater part of summer. The redwood spreads glossy needle twigs, causing reflexes like regular laurel leaves; more so does Arbutus menziesii. There occurs also an azalea. Rhodo- dendron occidental, the laurel-leaved shrub Berberis nervosa, and the low winter green, Gaultheria shallon. South America's Araucaria forests belong also to this class of formation. ^.Laurifruticeta: Laurel-leaved scrub. — This is the same as Laurisilvae, but scrub instead of forest. In the oceanic southern hemisphere this glossy-foliaged scrub occurs upward to the wood limit, forming, as in South America, a scrub belt above the forest. The same is done in the Caucasus by the association of Rhododendron caucasicum. On the whole, such scrubs are not so independent in the northern hemisphere, but for the most part are biotically influenced, or, in a cooler climate, they take refuge as underwood in forests whose dominants belong to a climatically cooler type, yet create under their canopy a locally more favorable climate. This telescoping of types presents a com- plication. To be cited are the true laurel scrub Lauretum nobilis in the Mediterranean, the Prunetum laurocerasi, and Rhododen- dretum pontici in the Caucasus. 5. Durisilvae: Sclerophyllous forests. — The dominants inDuri- silvae are trees with sclerophyllous foliage or green axes function- ing as leaves. Sclerophyllous leaves are evergreen, mechanically Plant Communities of the World 273 stiffened, and leathery, usually reduced in size, often hair-clad. They avoid hot sun radiation by standing obliquely to the light or presenting only their edges directly, as in Eucalyptus. Duri- silvae thrive in the typical Mediterranean climate. A moist, mild, oceanic winter is there combined with a dry, hot, continental summer, and spring and fall rains occur, and winter rains in the southern parts. On the whole it is not a good tree climate; scrub has a greater importance in it than in other climates. Durisilvae occur in California, in the Mediterranean, in Aus- tralia. Around the Mediterranean are the forests of stone oak, Quercus ilex, and cork oak, Quercus suber, the olive groves, the date palm, the Casuarina forest; in Australia, the Jarra forest with dominating Eucalyptus marginata, the Karri with Eucalyptus diversicolor, the Wandoo with Eucalyptus redunca, and others. In California, Cooper worked out four associations of relative constancy and wide distribution. The moistest areas of Duri- silvae are occupied by the tan-bark oak forest, the Pasania densi- fiora — Quercus chrysolepis — Arbutus menziesii association, and Pasanietum densiflorae, characteristic of the lower altitudes of the northern Coast Ranges. The Quercus agri folia — Arbutus association is characteristic of the outer central Coast Ranges from Mendocino to Los Angeles County, and is particularly well developed in the San Francisco Bay region and southward to the Santa Monica Mountains. The Quercus agri folia — lobata association occurs in the broad valleys and gende foothill slopes of the central Coast Ranges. Higher altitudes are occupied by the Quercus chrysolepis — \elloggii association, which belongs to the higher Coast Ranges and southern California mountains and to the middle altitudes of the Sierra. I was shown a beauti- ful association of light sclerophyllous forest with grassy under- growth when MacDougal took me up into the Santa Catalina 274 Eduard Riibel Mountains. The forest, 1450 meters above sea level, v^as domi- nated by three evergreen oaks — Quercus etnoryi, Quercus art- zonica, and Quercus oblongi folia. 6. Durifruticeta: Sclerophyllous scrub. — Durifruticeta are scrubs with sclerophyllous foliage or green axes, switch plants. In Australia this scrub is for the most part climatically condi- tioned, but in the other continents biotic and edaphic factors have originated and enlarged the scrub areas enormously. From Australia is known the eucalyptal Mallee-scrub, the acacial Mulga-scrub, the Brigalow-scrub, and others; from Cape Prov- ince, the Fynbosch. The Mediterranean Macchia and Garigue is to be divided into a series of federations and associations so far not entirely unraveled. The most luxuriant is the Arbution, in which are the Arbutetum, dominated by the glossy Arbutus unedo, the Ericetum arboreae, and the Fruticiquercetum iUcis, dominated by the shrubby-formed Quercus ilex. In the cist rose federation may be counted the Cistetum monspeliensis, the Cistetum ladaniferi, and the Cistetum salvifolii. The Garigue of Spain includes different types of the Tomillares; the Dwarf- palm scrub, Chamaeropetum humilis, is also a good association; in Greece the Phrygana types, in southern France large areas of Quercetum cocciferae, the Kermes oak, dominating. In California the "climax chaparral association" (Cooper) is the dominant community over the whole of the southern Coast Ranges and the mountains of southern California and northern Lower California. Adenostoma fasciculatum is for the most part the dominant species, but many evergreens are conspicuous and quite a few consociations will be recognized within the asso- ciation. Many species of Arctostaphylos and of Ceanothus are prominent. With its center in the northern Sierra, the mosdy biotically conditioned (fire) "Conifer forest chaparral associa- Plant Communities of the World 275 tion" covers the middle altitudes. Species of Quercus and Arcto- staphylos and others are preponderating. 7. Ericifruticeta: Ericaceous scrub (heaths). — Ericifruticeta are scrub communities whose dominants bear ericoid leaves. The cricoid, rolled-up structure gives a tubular form which makes the leaves stiff without the necessity of special mechanical tissue. The heaths hve in a cool, moist, oceanic climate. They are ex- tremely well developed in Norway, and they follow the coast regions of Great Britain, Germany, and France to the Canary Islands, through cool, oceanic, subalpine belts of mountains. They look xeromorph, but Stocker has proved that they are mesophytes to hygrophytes. These studies have proved that we were right in placing these communities here by their habitat, at a time when the ecology was yet far too little known. The smallness and ericoid form of the leaves stiffens them against being torn and wilted by the very heavy winds of these countries. The xeromorph appearance is an unintended by-eif ect. The soil is leached and poor in salts; and heavily transpiring evergreen heath scrub gets out the best of it, as the vegetative season is prolonged by the oceanic climate with mild winters. Calluna heath was the association which awakened the first phyto- sociological ideas in Alexander von Humboldt. It occurs on wide stretches in Great Britain, the Netherlands, northwestern Germany, and southwestern Sweden. A moister association is the Ericetum tetralicis; a special Cornwall heath is the Ericetum vagantis. Other heath communities are found in Portugal, in the upper part of the fog belt of the Canary Islands, and else- where. A well-developed heath system are the arctic and the alpine dwarf shrub communities. Remarkable for richness in species (about four hundred species of Erica) are the heaths of Cape Province. 276 Eduard Riibel 8. Aestisilvae: Summer green forests (deciduous forests). — The dominant trees become leafless regularly in the cold season; their buds are always protected. A warm moist summer allows them a very high water balance. Aestisilvae are the widely ex- tended forests of middle Europe, eastern Asia, and eastern America. The beech woods of Europe are being closely studied at present in many countries. The preeminent forest of the east- ern United States and southeastern Canada is the beech-maple forest; toward the drier Middle West it changes through a transi- tion of maple-elm forest into extensive oak forests. Deciduous oak forests occupy the stretches of moderately continental cli- mate and struggle for area with the more continental prairie. The ratio of precipitation to evaporation is 80 to 100 per cent, and in beech forests it is more than 100 per cent. In southern Switzerland, below the beech belt on calcareous soil are oak forests (Quercus sessili flora) ; on siliceous soil, chest- nut forests {Castanea vescd). Wet or moist forests in running water are the "Auenwalder" of Europe, forests of Alnus incana with willows, many lianas such as hops (Humulus lupulus) and Clematis vitalba. The equalizing action of water permits the Populus communities to follow the rivers far into the African desert, and Populus fremonti and Fraxinus tomei in the river bed at Tucson into the Arizona desert. In stagnant water, Alnus glutinosa dominates the English "Carr" and the "Bruchwald" of the European continent. Near these may be ranged the Taxo- dium distichum community of the Great Dismal Swamp in Virginia and North Carolina. g. Aestifruticeta: Summer green scrub. — Summer scrub is biotically and edaphically conditioned; a few of its constituents semiclimatically. Where the climate permits closed deciduous- leaved summer vegetation, this suffices, as a rule, for tree Plant Communities of the World 277 growth. Well known is the federation Quercion pubescentis, or "Shiblyac" a sub-Mediterranean, submontane scrub widely distributed in Bulgaria, Rumania, and Jugoslavia, and in Italy westward to the Spanish Monte Bajo and eastward to the Cau- casian countries. Dominants are Quercus pubescens, Berberis vulgaris, Viburnum lantana, Syringa vulgaris, Cotinus coggy- gria, and Paliurus austrdis, among others. Similar scrub of oak, Rhus, Fendlera, and Cercocarpus, I have passed in the Wasatch Mountains, as well as near Palmer Lake on the mesa. In the Alps, moist siliceous slopes where avalanches discour- age the growth of breakable wood are clad with green-alder scrub (Alnus viridis), sl thicket of very flexible twigs. lO.Hiemisilvae: Raingreen forests (monsoon forests). — The dominant trees of these forests shed their foliage in dry hot summers, but burst into leaves with the beginning of a rainy season. Leaves are mosdy small and compound in accordance with the continental tropic climate unfavorable for tree grov^h. Green bark often assists in assimilation. In Africa these rain- green forests — thorn forest, savannah forest, or monsoon forest — cover very large areas from Portuguese West Africa through northern Rhodesia to the Congo, Tanganyika, and North Bechuanaland, as well as on the south side of the Sahara. Many species of Acacia are dominants. They are also frequent in Asia and in South America, as, for example, the catinga and monsoon forest on the dry calcareous hills of Minas Geraes. A Prosopetum juliflorae is conspicuous on the permeable sand and clay soil of the southern shore plains of Jamaica. ii.Hiemifruticeta: Raingreen scrub. — Inward on the conti- nents, trees give way to bushes. This scrub fringes the deserts and enters them at edaphically moister points. The association of the screw bean and mesquite (Prosopetum pubescentis) cover 278 Eduard Riibel in the Salton Sink in southern California areas is influenced by the level of underground water. For the greater part of the year this scrub stands leafless, but when rain falls the bushes produce very quickly their small compound acacious leaves. Between these bushes I noted Atriplex canescens and Atriplex lentiformis and, conspicuous in the undergrov^h, saltgrass, Distichlis spicata. Important for the association are also Acacia greggii, Acacia constricta, Flourensia cernua, and Larrea mexi- cana. J 2. Aciculisilvae: Needle-leaved forests. — The dominant trees bear needles. The needle leaf is small, linear, mostly with special mechanical strengthening, and, with few exceptions (Larix), evergreen. Many conifer forests belong in this association, but I have also found conifers in the Laurisilvae and Taxodium in the summer forests. The climate is cold and continental, the vegetative season short. The forests grow in the subalpine belt and the subarctic zone, in the inner parts of continents and in inner alpine valleys, edaphically, where the soil is too poor for broad-leaved forests — on sand, on flat ground, in swampy ground poor in oxygen. Large parts of the northern hemisphere between forty-five to fifty and between sixty-five to seventy degrees of latitude bear such forests. The Piceion excelsae covers enormous stretches in Europe and Asia. The unpretending Pinus silvestris can form communities almost anywhere, but has hardly the strength to compete with other trees for which the habitat is good enough. The highest tree-bearing areas in the Alps belong to larch and aroUa pine forests {Larix europaea, Pinus c em bra). In the northwestern United States is a subalpine Aciculisilva of Abies lasiocarpa and Tsuga pattoniana. In more continental and more southern Colorado the formations lie higher and Plant Communities of the World 279 separated, a subalpine Pinetum aristatae from 3600 to 3400 meters altitude and the Piceetum engelmanni from 3400 to 2700 meters. In the same belt in California, dominants are Pinus jeffreyi and Abies magnifica from 2500 to 2200 meters altitude. To the lower part of that belt probably belongs also the Wash- ington forest, between 1400 and 600 meters altitude, of Pinus monticola and various species of Abies. In the more arid area Pinus ponderosa forms forests around 1000 meters altitude in Washington and from 1900 to 2700 meters in Colorado. Moister areas of the same belt are governed in California from 1500 to 2200 meters altitude by Abies concolor (in these forests are the "Big Tree" groves). This forest prevails in Arizona around 2400 meters altitude. In the same belt live enormous Pseudotsuga forests, but the Douglas fir spreads so widely by reason of biotic factors that its forests derive from diverse causes. The lowest parts of oceanic Washington have been mentioned in the dis- cussion on Laurisilvae. A remarkable phenomenon occurs at the tree limit in the Cascades. While the ascending crests are clad with trees, per- petual snow lies on open flat places. In these mountains the climatic snow limit lies at 2000 meters altitude, the tree limit at 2300 meters. This brings about the very interesting situation that these important lines not only are near each other, but actually transgrade. In the Swiss Alps the forest limit remains at 700 to 1000 meters below the perpetual snow limit; but in the very oceanic climate of Washington State, forests can thrive above the cUmatic snow-line, and its alpine vegetation, which only begins above the forest limit, thrives even higher. Phanero- gams are known there up to 3200 meters altitude, that is, 1200 meters above the snow limit. A part of the subalpine belt and the whole of the alpine belt fall here into the niveal belt. We 28o Eduard Riibel feel carried back to the glacial period. What interesting prob- lems wait here for study! If the Alps in the glacial age had a climate similar to that of western Washington today (the glacial age snow limit of the Wallis coincides with Washington's of today), the greater part of the Alpine flora may have survived. The country north of Washington and into Alaska would surely give still better parallels. A study of this vegetation in respect to these questions of ecologic dependencies is ardently desired, nay, even demanded! 7 J. Aciculifruticeta: Needle-leaved scrub. — These unpretend- ing communities occur where more pretentious ones give way : in the mountains near the tree limit, and lower down when edaphically the conditions are less favorable, as on desolate calcareous slopes and badly aerated raw humus soils. The best known is the dwarf-pine scrub (Pinus mugo), which covers much area in the Alps, partly edaphically, in larger part biot- ically enlarged by forest-felling in former times. i^.Duriherbosa: Hard grass prairie. — Duriherbosa are grass- lands in which the dominant species have vegetative organs more strengthened by the development of the mechanical tissue than by turgescence, and which die down in winter. The climate is fairly continental with dry periods (semiarid to semihumid). Essential is the seasonal repetition of the precipitation; there are early summer rains just at the vegetative season, which en- courage the growth of a grass sward. Two groups are to be distinguished — one winter-cold, the other warm (savannah); in common are the deciding dryness and the summer rains. Hungarian and southern Russian sward steppes are conspicu- ous, called "pussta" in Hungary, "black earth steppes" in Russia. Dominant grasses are Stipa capillata and St. pennata, Andro- pogon gryllus, and, on the best black earth, Festuca vallesiaca Plant Communities of the World 281 and Koeleria cristata. Similar communities are found in locally dry continental areas in middle Europe, as in the Wallis of Switzerland; a fine Festucetum vallesiacae, and related unities, are found similarly in the Bohemian foothills and in the Jugo- slavian Velebit Mountains. The most famous communities of this class are the prairies and great plains of America, so well studied by Clements, Shantz, Livingston and Shreve, Weaver and Fitzpatrick, and others, that I need not go into details. The decided continentality of the Duriherbosan climate must rely on the strong change of dry and wet without great tempera- ture differences. The areas with rich savannah grasslands cover wide stretches in Brazil, the Argentine, Australia, and especially in Africa. With increasing coldness, the communities merge toward steppe prairie, and where the climate becomes less hostile to trees one finds gradual transitions to Hiemilignosa. 75. Sempervirentiherbosa: Evergreen grassland (meadows). — The turgor for the most part suffices to keep the dominant grasses and herbs in their shape, and mechanical stififening is therefore slight. These meadows have no pronounced winter rest; they sprout whenever temperature permits. They prevail in a cool oceanic climate poleward beyond the tree limit, in the mountains above the tree limit, edaphically in flood areas of sea and rivers and on windy coasts. Most meadows, all middle European meadows of the plains, are biotically conditioned (half-culture) and would turn into forests if left alone. Many leaves die in winter, but younger ones in all stages are always present and new ones are formed. The plant is evergreen, but the single leaf is not. Whenever I dug in the Engadine in the cold months when the temperature of the air was far below zero, I found the local habitat under the heavy snow cover to be around zero and the vegetation green, always ready to 282 Eduard Riibel assimilate if the temperature should rise above zero. Swiss sociology has always favored the study of meadows, from the time of Stebler and Schroter's classical studies. It has much to do with alpine as well as lowland agriculture. Arrhenatheretaha, manured meadows, are about the same all over Europe, since similar practice has brought about a conver- gence everywhere of the same factors. Caricetalia curvulae and Salicetalia herbaceae, "Schneetalchen"-Gesellschaften, grouping around the twist sedge association (of Cm-ex curvula) and the herb willow association (of Salix herbacea), are the highest alpine and best studied meadow communities. The latter are also well distributed in Scandinavia. 16. Altherbosa: Tall herbage. — Tall-growing herbs dominate in this. It is an exacting vegetation, greedy for humous, nutrient- rich soil, and it occurs where forest and scrub for one reason or another are absent. The Adenostylion or "Karflur" in the sub- alpine belt of the Alps is developed in an Adenostyletum alliariae, a Peucedanetum ostruthii, a Chaerophylletum villarsii. Mam- moth tall herbage, as much as four meters high, occurs in the Caucasus under the lead of Heracleum mategazzianum. In the Arctic, Angelica archangelica dominates, man-high in well- sheltered dales in Greenland. The "lair herbagej' Rumicion alpini, around the alpine dairy huts on the overmanured soil shows as dominant the big Rumex alpinus or Senecio alpinus, or, in the Caucasus, Tele\ia speciosa. A successional stage of tall herbage after wood-felling or forest fire is dominated by Epi- lobium angusti folium, called ''willow herb" or "willow rose" in Europe, charming with its red spikelets in the little wood open- ings, but detestable as "fireweed" over wide American lands where lumber companies have devastated the country by their logging operations. Plant Communities of the World 283 ly.Emersiherbosa: Etnersed swamp vegetation. — The domi- nants in this community are marsh plants which root in water- covered or saturated soil, but which have their leafy shoots above the water surface sufficiendy stiff to hold themselves erect. Water is an equalizing element, watery habitats are similar even in very diflerent general climates; water plants are therefore ubiquitous and their communities are similar in different countries. The order of Molinietalia coeruleae consists of communities which have their roots, but not yet the stems, in the water. They make what are called sour meadows; agriculture gains litter. In the Molinion federation the association of the pipe grass, Molinia coerulea, is the most important; in the Parvicaricion, one counts in Switzerland the Schoenetum nigricantis; Juncetum sub- nodulosi, Caricetum lasiocarpae, the important widespread Cari- cetum fuscae, the alpine Eriophoretum scheuchzeri, and the alpine Trichophoretum caespitosi and pumili. The Phragmite- talia are new-land builders, pond communities. Of the Magni- caricion elatae by far the commonest is the Caricetum elatae (others are the Caricetum inflatae and Caricetum vesicariae), beautifully described in 1863 by Kerner as the "Zsombek" of the Hungarian lowlands, and by Pound and Clements on the Missouri, and by Schroter in Switzerland. Still more widely dis- tributed is the Phragmition, with the Scirpeto-Phragmitetum communis, well known in many countries, as, for example, East Norfolk in England, and the enormous areas in the Danube and Volga deltas. If the water contains NaCl, one finds on all sea- shores similar communities, the associations Atriplicetum por- tulacoidis, Juncetum maritimi, Glycerietum maritimae, and very commonly Salicornietum herbaceae and Spartinetum. /S. Submersiherbosa: Submersed stvamp vegetation. — This consists of macrophytic aquatic plants with submerged or float- 284 Eduard Riibel ing leafy shoots. The influence o£ dimate is minimal. Differences in temperature are still of some influence, but rather on the floristic composition than on the ecology. Salt content divides groups; also the rooting in mud or fixation on stone. Therefore we have four large groups or orders — the mud-rooting sweet- water Potametalia or Limnaeas, the mud-rooting salt-water Enalids, the stone-fixed Limno-Nereids or Encyonematetalia, and the Halo-Nereids; and to be added to these are the swim- mers Pleuston or Hydrocharitetalia. Warm water knows as Limnaea the beautiful Victoria regia with its associates; colder water, the lake roses (water lilies) Nymphaea and Nuphar; in the alpine belt, Sparganium angusti- folium and Hippuris vulgaris and Ranunculus flaccidus; in the north, Stratiotes aloides; in America, Eriocaulon septangulare or Elodea canadensis; in Lake Titicaca, Ceratophylls and Pota- mogetons; in deeper waters, Charas and Nitellas. EnaUds are the shore-covering subaquatic sea-grass meadows of Zostera manna and Z. nana; in the Mediterranean, Posidonietum oceanicae. Chlorophyceae and Cyanophyceae cover the stones of lake shores, as Limno-Nereids. Halo-Nereids are plentiful. In north- ern Europe live the well-studied algal communities of Fucaceae, Laminariaceae, and others. I never forget the amusing California community of sparse Pelagophycus gigantea with their myste- rious floating heads, and the subaquatic dwarf-palm forest of Postelsia palm ae for mis. Ecologically, we have to place here also the Pleuston, the float- ing sargasso meadows of the ocean, and the water lens, Lemne- tum, of ponds. ig.Sphagniherbosa: Moss moor (Hochmoor).— These com- munities depend on aerial water rather than telluric; they form swamps and produce their substratum by their own vegetation. Plant Communities of the World 285 especially by a Sphagnum covering. The climate is cool oceanic. Moss moors have a wide distribution in Scandinavia, Great Britain, Ireland, and in northwestern Germany in Europe. Northern North America also has Hochmoor, and, in the southern hemisphere, so do the Antarctic islands, Patagonia, Tasmania, New Zealand. The habitat is peculiar. Osvald groups these communities in Europe into four large types or federa- tions— probably: (i) the western "ground covering," Eriopho- rum vaginatum Hochmoor, in England, Scotland, and Ireland; (2) the flat moss moor, Sphagnum rubellum Hochmoor, in southwest Sweden, along the Norwegian west coast, in low alti- tudes of the northern Scandinavian mountains, and in northern Finland; (3) the truest moss moor, Sphagnum medium Hoch- moor, in southern and central Sweden, southwestern Finland, the Baltic countries, northern and northwestern Germany, and the mountainous areas of middle Europe (cold windy valleys of the pre- Alps and cold moist Jura heights in Switzerland) ; (4) wood-moss moor. Sphagnum fuscum Hochmoor, with Pinus silvestris, Ledum palustre, Sphagnum an gusti folium, in eastern Sweden, Poland, Finland, Russia — in all the less oceanic areas. 20.Siccideserta: Dry deserts. — ^These are open communities. Larger plants do not exert any marked influence on any lower story which may be present. They are conditioned by the dryness of the climate. They lie in the great subtropical high-pressure zone of the earth. Temperatures may be hot or cold; the low amount of precipitation and its irregularity are decisive. Great deserts are the Eurasian, North African, western American, Australian, and South African. Artemisia steppes: The Caspian (Asia) Artemisia scrub desert includes an Artemisietum maritimae incanae and, in saltier dales, Artemisietum pauciflorae. Well known are the extensive 286 Eduard Riibel sagebrush deserts in western North America — an Artemisietum tridentatae with a Kochietum vestitae on loam and an AtripUce- tum confertifohae on a lower, moister level. The corresponding communities in Algeria-Tunisia are the Artemisietum herbaeal- bae and the Stipetum tenacissimae halfa grass-scrub desert. The halfa grass is stiffened by cellulose; its leaves are pluriennial, and therefore ecologically it is a shrub. Where conditions take on a more extreme character, there are deserts of dry squarrose shrubs with reduced foliage. Most shrubs are thorny and have small, gray-green, leathery, ever- green or only raingreen leaflets. These are the thornbush deserts of Algeria, Tunisia, Egypt, and a most unusual one in the alpine belt of Teneriflfa, the Retama Blanca desert of Spartocytisus supranubius. In South Africa is the large Karoo desert. Still more exacting on the vegetation is dryness and salt content com- bined. We find salty dry deserts in southern Russia, in Tunisia, Algeria, and around Great Salt Lake in Utah. Prominent in this class are the succulent deserts. In Africa, suc- culent Euphorbiaceae are building up communities; in America (that is, in Mexico, Texas, Arizona, and New Mexico ),Cactaceae. In the Bajada, Larrea tridentata is dominant, accompanied by the oddest forms of Opuntias. On the slighdy inclined deep-soil talus stands the Parkinsonietum microphyllae including the local miracles, Cereus giganteus, the fleshy telegraph pole, and the stem-assimilating thorny ocatilla, Fouquiera splendens. 21. Frigorideserta: Cold deserts. — The vegetation of these des- erts consists mostly of herbaceous perennial, rarely woody, plants, usually of fresh green color, and frequently of tufted, rosette, cushion, or creeping habit. They are climatically condi- tioned. Cold has the effect of dryness, since water for a great part of the year is in the form of ice and snow, unavailable for plant Plant Communities of the World 287 growth. Cold deserts are the most polar communities, hori- zontally the token of the high arctic zone, vertically of the niveal and subniveal belt. The alpine group has strong sun radiation in the short season; the arctic, less radiation but very long days; other factors are about the same. In the High Alps such desert communities on fixed rubble are a wonderful symphony of colors. In the siliceous Androsacetum alpinae gleams the deep blue Gentiana bavarica subacaulis, changing with the delicate rose of Androsace alpina cushions and the sky-blue heavens' herald, Eritrichium nanum. Other communities are the Oxyrietum digynae, the Allosoretum crispi, and, on calcareous soil, Arabidetum coeruleae. The New Zea- land niveal belt has the "vegetable sheep" Haastia pulvinaris. To the arctic cold deserts belong open tundra, the "fell field" of Warming, and the "rocky flat" of Ostenfeld. 22. Litorideserta: Strand steppes. — These are deserts developed under the influence of the neighboring sea where the atmos- phere is full of salty spray. Rock-cliff strand steppe is the Sene- cionetum cinerariae, which is well developed on Mediterranean shores. As dominant, glitters the silk-hairy Senecio cineraria. Slighdy darker is the also hairy, abundant, Helichrysum angus- ti folium. Corsica's lies Sanguinaires have in this association their endemics, Nananthea perpusilla and Evax rotundata. Southern France has on its calcareous shore a similar association, the Arte- misietum gallicae. The surge-sprayed lava rocks of the Canary Islands bear a most unusual type of this class, the Limonietum (Statice). A large number of purely endemic Limoniums spread their flower-veils in the most delicate gradations from lilac to blue to violet to red, and color for a short springtime the brown rocks. Flat shore-strand steppe is dominated in northern Europe by Salicornia herbacea with Festtica thalassica, or^on more sandy 288 Eduard Riibel soil, Salicornia perennis and Atriplex portulacoides. On the Mediterranean shores, Salicornia fruticosa is the main plant; in southwestern Australia, Salicornia australis; on Java, Ipomoea pescaprae; on the Moluccas, Canavalia; on Asiatic shores, Spini- fex squarrosus. 2^.Mobilideserta: Wander deserts. — The desert character o£ open vegetation is caused by the mechanical action of the un- stable substratum. The mobility of the soil is a peculiar factor with formidable influence on the vegetation, demanding a par- ticular ecology against its being torn away or buried. According as the mobility of soil is caused by wind, by gravity, by water, or by man, there are different groups : shifting dunes, steep screes, and shifting river alluvions. Ammophiletum arenariae is the main shifting-dune association all over Europe and America, with more or less slight variations. The English Lancaster coast has a foredune dominated by Agropyron junceum; Michigan, one dominated by Calamovilfa Ion gi folia. Salt-water and sweet- water dunes are closely related. The calcareous Alps have a fine sliding rubble vegetation around the Thlaspeetum rotundifolii, with Trisetum disticho- phyllum, Papaver alpinum sendtneri (white), and other plants, a fine scree on the Pilatus, for example. The eastern central Alps have an association with the pretty yellow Papaver aurantiacum. On the screes on Mount Garfield in the Rocky Mountains I saw, especially. Paronychia and Silene acaulis. On alluvial ground in the Swiss Alps, Myricaria germanica is dominant from the plain up to 2000 meters altitude. In the higher localities on glacial river alluvium a subassociation is dominated by the red big-flowering Epilobium fleischeri. Somewhat similar to the alluvium turned over about once a year by the high waters are the artificially turned-over soils, the Plant Communities of the World 289 arable ground. Treated every year in the same way, this ground has developed rather stable communities in two orders. The Segetalia, sown cultures, plowed once a year, have in middle Europe a weed community, Agrostemmatetum githagonis. Gen- erally distributed is Agrostemma githago; brilliantly blue and red shine Centaur ea cy units and Papaver rhoeas; characteristic also are Spergularia arvensis, Pisum arvense, Lunaria vulgaris. Olitoria, hoed cultures, truck crops, are both plowed and worked over, in the season, by hoeing. These crops, such as potatoes and other vegetables, are not only manured and hoed, but often are watered also, and vegetation is somewhat hydrophilous. We can call the association Chenopodietum polyspermi, by reason of the frequent Chenopodium polyspermum accompanied by Pamcum cr us gain, Mercurialis annua, Euphorbia peplus, Amaranthus ascendens, A. retroflexus. Polygonum persicaria, etc. 24. Rupideserta: RocI^ deserts. — This is the chomophyte vege- tation of rocks which is dependent on detritus on the rock or in crevices, and thus is clearly influenced by the rock habitat. (When crevices become large, or detritus plentiful, this latter influence ceases and we have stony forests or grasslands.) In the detritus on the rock the habitat of the exochomophytes is rather xeric, and in the crevices with the chasmochomophytes, or chas- mophytes, it is not. The prime condition is edaphic; climatic factors noticeable are altitude and exposure, both of strong in- fluence on the local climate. Rock deserts are found locally almost everywhere in mountains, in erosion valleys, and on slopes and cliffs in the lowlands. In the Swiss Alps are to be distinguished the Potentillion caulescentis, on a calcareous habi- tat and with a montane-subalpine association characterized by Potentilla caidescens, Kernera saxatilis, Asplenium rutamuraria, Lilium croceum, and Globularia cordifolia, among others, and 290 Eduard Riibel an alpine-niveai belt association dominated by the firm cushions of Androsace helvetica, accompanied by Draba tomentosa and others. The corresponding siliceous-rock vegetation gravitates in the subalpine belt around Aspleniiim septentrionale and sev- eral Sedum and Sempervivum species, and in the alpine around Androsace mtdtiflora, Fhyteuma carestiae (eastern), Ph. humile (western), and Primula hirsuta. 2^. Saxideserta: Stone {and wood) deserts. — This is the Litho- phyte vegetation of stones and wood, especially bark, growing direcdy on firm, not disintegrated substratum. Bark vegetation is similar to and often identical with the stone vegetation, espe- cially with calciphobes (Gams). Mosses, and especially lichens, cover the rocks as well as bark in multicolored mosaic. The com- munities have been well studied in Fenno-Scandinavia by Ser- nander, DuRietz, Hayren; in die Swiss Alps by Eduard Frey, Ochsner, Gams. The other large group are the Algetalia, algal communities coloring more or less perpendicular rocks, espe- ciallv Gloeocapsae and related species. 26.Aquerrantia: PhytoplanI{ton {water floaters). — This com- munitv consists of microphytes free-floating in the water. Com- munities specialize with the content of salt, oxygen, humus. ^\lso the snow contains its own kryoplankton according to its more solid consistency, and the floating of this growth is less mobile. 2'j.Solerrantia: Phytedaphon {soil ^oa^^y) .—Microphytes, myxomyceae, schizomycetes, mucorineae, form communities in the soil. They neither root nor adhere. One may call it a floating on soUd ground, just as the chemical engineer speaks of "solid solutions!' 2S.Aererrantia: Phytaeron {air ^o^/dr/) .—Microorganisms float freely in the air; their connection is extremely loose, wind causes rapid changes, yet perhaps they occur in groupings. Antarctic Plants in Polynesia By Carl Skottsberg "From the southeast, there is a slight but plausible indication of immigration;'— W. A. Setchell, Paths of dispersal, in Phytogeo- graphical Notes on Tahiti, 7926. THE RECOGXiTiox o£ an old Antarctic flora element, at least among the Phanerogams, dates back nearly one hundred years to the time when J. D. Hooker, on the famous cruise of the "Erebus" and "Terror" under James Ross, became aware of the biological resemblance between regions so remote from one another as Magellania, New Zealand, and the various islands on the threshold of the Antarctic Sea, and refused to be- lieve that this could be satisfactorily explained merely by assum- ing an exchange across the \^'ide expanses of water under the present phvsiographical conditions. Hooker had a presentiment that a natural explanation was hidden in the imknown history of the immense land mass known as Antarctica. Ross had dis- covered fossil wood of an extinct gynmospermous tree in Ker- guelen, where at the present time no trees can grow, and it did not seem too audacious to postulate the former existence of an Antarctic land flora where there now is solid inland ice. By and by, proofs came that the Arctic regions had harbored a rich, warm-temperate or even subtropical vegetation. It became only a question of time when Antarctica would surrender its buried treasures. C. A. Larsen, the famous Norwegian whaling captain, collected the first fossil wood in western Antarctica in 1892, and ten vears later the Swedish Antarctic Expedition, with Larsen in command of the vessel, discovered rich deposits from the Jurassic and Tertiary containing distinct connecting links between the [291 ] 292 Carl Skpttsberg floras of the lands on either side o£ the south polar continent. I was in the party, and I have taken an interest ever since in the history of the Antarctic flora and have tried to collect data bear- ing upon it. In 1915 I published an annotated list illustrating the intimate relations between subantarctic America and New Zea- land."^* After I came to examine some of the Pacific floras (espe- cially in Hawaii, where a small number of Antarctic genera occur, a strange sight in a Paleotropical insular world), I desired to take up the study of this element. The taxonomic basis is clear in respect to some groups, but in respect to others the position and affinities are not so well known and conclusions therefore are less safe. I have just had occasion to show how important mono- graphic studies are, if we would arrive at something like definite results."' Time has not permitted me to make a complete study of the Antarctic flora, its history and migrations. I shall restrict myself here to occurrences of the representation of presumably Paleantarctic genera by geographically isolated species in the high Polynesian islands. These occurrences are few but very striking, and they offer a number of interesting and troublesome problems. Much more numerous are those Antarctic genera which, in the Pacific region, extend to New Zealand, south- eastern Australia, New Guinea and, perhaps, Malaya without occurring in Polynesia. They do not seem to present exacdy the same problems as do the genera which are the subject of the present study. Roads have been easier along the extreme western border. Among the Antarctic-Pacific genera, more than one type of distribution may be distinguished. It is convenient to recognize one tricentric and one bicentric group. The former is represented in the Australasian, American, and African sectors; the latter is * Superior figures refer to items in the bibliography at the end of this essay. Antarctic Plants in Polynesia 293 absent from Africa (including Madagascar and the Mascarene Islands). If there were land connections between Antarctica and the surrounding continents, the African bridge certainly dates far back and its existence is altogether problematic; to those who do not believe in such connections, the remoteness of Africa from Antarctica and the scarcity of intervening islands explains why Antarctic types are so scarce in that sector. It is surprising that they exist at all. We shall begin with a tricentric type, Astelia, a true Paleant- arctic genus if there ever was one. In my sense,^" it consists of 3 subgenera with 7 sections and 23 species. Its present headquar- ters are in New Zealand, where all subgenera and 4 of the sec- tions occur, including 9 species. Two of the species extend to Auckland and Campbell Islands (or it may be more correct to say that the subantarctic species extend to New Zealand), and I to Warikauri (Chatham). There are species clearly related to the Neo-Zelandic ones in New Guinea (i), Tasmania (i), the Australian alps (2, i a variety of the Tasmanian sp.), New Cale- donia (i), and the Society Islands (i). A separate section is formed by the 7 Polynesian species, 6 in Hawaii and i in the Marquesas, all closely related to one another. Finally, one species is endemic in subantarctic America, including the Falkland Islands, and one in the Island of Reunion off the east coast of Madagascar. Consequendy, two of the three centers are repre- sented by only one species each. But this does not make them less important. Each of the species forms a monotypical section. I have come to the conclusion that Astelia hemichrysa of Reunion represents the oldest existing type in this genus. The other mem- bers of the same subgenus are in the Auckland Islands, Campbell Island, New Zealand, Tasmania, Australia, New Guinea, and the Society Islands. A close study of distribution details and 294 ^^^^ S^oUsberg affinities, as shown in the sketch map accompanying this essay (p. 295), has given some interesting results. The position o£ sect. Periastelia (Hawaii and the Marquesas) is o£ special im- portance. From its geographical position, it seems natural to look for its closest relatives in the southwestern Pacific. But the Tahitian species belongs to a different subgenus. In New Cale- donia and New Zealand we find members o£ the same subgenus, but they form another section not nearly related to Periastelia. We discover its nearest ally where we would not be likely to look for it, in the Magellanic Micrastelia. Simply because this occurs in America, we are not entitled to classify the Hawaiian species as belonging to an "American" element. Micrastelia is a stranger in the neotropical flora. Collospermum, which I segregated from Astelia, perhaps should be mentioned here because it goes back to astelioid ances- tors and is a Pacific genus of Antarctic affinity. It has 3 species in New Zealand, i in Fiji, and i in Samoa. The distribution of Gunner a is significant. The Indomalayan- Australasian sector is the home of 2 of its 6 subgenera; Pseudo- gunner a, with a single Malayan species of wide range; and Milligania, with 8 species in New Zealand, i in Tasmania, and I in New Guinea. The African sector is inhabited by subgenus Perpensum with 4 species, i in South Africa, i on Kilimanjaro and Usambara, i in Abyssinia, and i in central Madagascar. (These may be regarded as subspecies of one species, but, even so, they have a long geologic history behind them.) The Amer- ican sector is the richest. It has 3 subgenera: Misandra, with 3 species, ranging from the Falkland Islands and Tierra del Fuego to the tropical Andes; the monotypical Osteni gunner a^ recently discovered in Uruguay and probably less distant from the Neo- Zelandic Milligania than from any other subgenus; and Pangue Antarctic Plants in Polynesia 295 O Sect. Palaeastelia "j @ Sect. Asteliella \ Subg. Euastelia ® Sect. Desmoneuron J 0 Sect. Tsoneuron $ Sect. Periastelia C) Sect. Micrastelia # Sect. Tricella 9 Gen. Collospermum Lines of aflfinity ==s Uniting stations of one species Subg. Astellopsis Subg. Tricella ' Gen. Astelia AC — Auckland and Campbell Islands Ch — Chatham (Warikauri) Island P —Fiji Islands H — Hawaiian Islands M — Marquesas Islands NC — New Caledonia R — Reunion S — Samoa T — Society Islands Fig. I . Distribution of Astelia and Collospermum. 296 Carl S\ottsberg ("Panke" is a nonsensical corruption), with about a dozen species along the Andes, 2 in Costa Rica, i in Brazil, and 3 in Juan Fernandez, to which the 2 Hawaiian species are most closely related, whereas they have nothing whatever to do with the west Pacific Psetido gunner a. The Paleantarctic character of Gunnera is affirmed by the fairly close relation between Mis- andra and Milligania. The distribution of Acaena is equally striking. G. Bitter, the monographer of the genus, recognized 10 sections. Numbers I (13 sp.), IV (i), and VI (i) are confined to the American sec- tor; Number V has 6 species in subantarctic America, i in Masa- fuera, and i in Tasmania; VII has 26 species in South America, I in California, and i in Australia and Tasmania; VIII has 56 species in America (the Andes, Patagonia, Fuegia, i extending to South Georgia, Kerguelen, etc.), 2 in the African sector, 3 in New Zealand, and i polymorphic species in New Zealand and the surrounding islands, Tasmania, Australia, and New Guinea. Section II is endemic in the African sector (Cape, i sp.). Thus, this sector contains one endemic section, i endemic species of Section VIII in each of the islands of Tristan da Cunha and New Amsterdam, and endemic subspecies of a Magellanic species in Kerguelen, Crozet, and Prince Edward Island. The Australian- Neo-Zelandic sector has two endemic sections— IX with i species and X with 2 species. In addition, as we have already seen, V is represented by i species (Tasmania), VII by i species, and VIII by 4 species. Evidendy the present center, with respect to both the number of species (97) and of sections (7), lies in extratropi- cal South America. However, 5 sections are found in the Aus- tralian-Neo-Zelandic sector (9 sp.), and 2 in the African sector (4 sp.), and, if we beUeve that the stations in these two sectors are nothing but "branch offices" from the headquarters in South Antarctic Plants in Polynesia i