THE UNIVERSITY -*L '* OF ILLINOIS LIBRARY The person charging this material is re¬ sponsible for its return on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may . . . * EXPERIMENTAL VEGETATION THE RELATION OF CLIMAXES TO CLIMATES BY Frederic E. Clements and John E. Weaver THE UHHUr f]r thf M 11 i£Z5 mmssn m mu & Published by the Carnegie Institution of Washington Washington, December, 1924 CARNEGIE INSTITUTION OF WASHINGTON Publication No. 355 Judd & Detweiler, Inc. WASHINGTON, D. C. ^VX 2,1,'* \ CONTENTS. 5SU5 C 5^ t O \S) U"> FAGE. List of illustrations. v 1. Introduction. Scope and plan. 3 Principles. 3 Methods..... 4 Values. 5 The grassland climax. 5 Tests of unity. 6 Further evidence. 7 Relationship of the associations... Climatic unity. 8 Objectives of the present investigation 9 Stations. 10 Location. 10 True-prairie stations. 11 Mixed-prairie station. 12 Plains station. 13 Physical factors. 15 Rainfall and holard. 15 Temperature. 17 Wind and evaporation. 18 Detailed methods. 21 Germination of seed. 21 Depth of planting. 21 Methods of sowing and planting.. 23 . Experiments During 1920 . 27 High prairie, Lincoln. 27 Surface sowing. 27 Trench sowing. 27 Quadrat sowing. 27 Root habits of seedlings. 29 Summary of planting results. 31 Sod transplants. 31 Mixed prairie, Phillipsburg. 31 Surface sowing. 31 Trench and quadrat sowing. 33 Short-grass plains, Burlington. 33 Surface sowing. 33 Trench and quadrat sowing. 33 Sod transplants. 34 Summary of experiments. 34 Seeding in cultivated areas at all stations. 35 True prairie. 35 Root development. 35 Mixed prairie and short-grass plains. 36 Experiments at other climatic stat ions. 36 Subclimax prairie, Nebraska City. 36 Character. 36 Results. 38 Mixed prairie, Colorado Springs. . 39 Character. 39 Results. 39 Experiments at edaphic stations.. 40 Gravel-knoll station. 40 Character. 40 iii PAGE. 2. Experiments During 1920.— Cont. Gravel-knoll station— Cont. Behavior. 41 Low-prairie station. 42 Character. 42 Behavior. 43 Salt-flat station. 44 Character. 44 Behavior. 45 Swamp station. 45 Character. 45 Behavior. 46 Summary. 47 Experiments. 47 Physical factors. 48 Experiments during 1918 and 1919.. 49 Season of 1918. 49 Season of 1919. 50 Sod transplants. 51 Gravel-knoll. 51 Low and high prairie. 52 Mixed prairie. 55 3. Experiments During 1921. 56 Physical factors. 56 Rainfall. 56 Water relations. 56 Temperature. 59 Planting results. 60 Surface sowing. 60 Trench sowing.. 60 Denuded quadrats. 61 Root development in denuded quadrats. 63 Seedling transplants. 65 Summary. 65 Sod transplants. 66 Root development. 67 Experiments at other stations, 1921. 70 Physical factors. 70 Rainfall and holard. 70 Evaporation and temperature. . 71 Planting results. 71 Surface sowing. 71 Trench sowing. 74 Growth in cultivated soil. 75 Root development at Lincoln... 76 Sowing in denuded quadrats. . . 78 Root development at Peru, Nebraska. 80 Seedling transplants. 80 Summary. 81 Sod transplants. 83 Gravel-knoll. 83 Low prairie. 84 Salt-flat. 85 Salt-basin. 86 Swamp. 86 Regional transplants. 88 568008 IV CONTENTS. PAGE. 4. Experiments During 1922. 89 Physical factors. 89 Rainfall. 89 Holard. 89 Temperature. 89 Humidity. 91 Evaporation. 92 Planting results. 92 Surface sowing. 92 Trench sowing. 93 Denuded quadrats. 95 Summary. 96 Seedling transplants. 97 Summary. 98 Sod transplants. 98 Experiments at other stations, 1922. 99 Physical factors. 99 Water relations. 99 Planting results. 101 Surface sowing. 101 Trench sowing. 102 Denuded quadrats. 102 Seedling transplants. 103 Effect of competition. 104 Summary. 105 Sod transplants.. 105 5. Experiments During 1923. 109 Development of seedlings and trans¬ plants. 109 Physical factors. 109 Water relations. 109 Survival results. 110 Surface sowing. Ill Trench sowing. Ill Denuded quadrats. 112 Seedling transplants. 112 Experiments at edaphic stations.... 112 Survival results. 112 Surface sowing. 112 Trench sowing. 113 Denuded quadrats. 113 Seedling transplants. 113 Summary of survival. 113 Sod transplants. 115 Phytometric results. 116 Transpiration and growth.' 116 Plan. 116 Methods. 116 Conditions and results at Lin¬ coln. 117 Physical factors. 117 Transpiration and increase in leaf-area. 119 Conditions and results at Phil- lipsburg. 119 Installation. 119 Physical factors. 119 Transpiration and increase in leaf-area. 120 page. 5. Experiments During 1923— Cont. Phytometric results— Cont. Conditions and results at Bur¬ lington. 119 Installation. 119 Physical factors. 121 Transpiration and increase in area. 121 Summary and conclusions. 123 Transpiration and growth of communities. 123 Transpiration from natural cover and crops. 123 Objectives. 123 Methods. 124 True prairie, series 1. 125 Installation... 125 Results. 126 True prairie, series 2. 126 Installation. 126 Results. 127 Mixed prairie. 127 Installation. 127 Results. 128 Short-grass plains. 128 Installation. 128 Results. 129 Summary. 130 Growth of natural cover and crops. 131 Objectives. 131 Plan. 131 Results for 1920. 132 Results for 1921. 133 Results for 1922. 135 Results for 1923. 136 Summary.. 136 Structural reponse of dominants and subdominants. 137 Scope and significance. 137 6. R£sum£. 138 Plan and methods. 138 Communities. 139 Physical factors for 1920. 140 Ecesis during 1920. 140 Physical factors for 1921. 141 Ecesis during 1921. 142 Physical factors for 1922. 142 Ecesis during 1922. 143 Behavior during 1923. 144 Behavior of trees and shrubs. 146 Germination and survival at the cli¬ matic stations. 147 Germination and survival at the edaphic stations. 148 Survival of sod transplants. 151 Phytometric methods. 151 Attainment of objectives. 152 Bibliography. 154 Behavior tables. 155 LIST OF ILLUSTRATIONS. PLATES. Plate 1. A. Detail of high prairie, showing the luxuriant estival society and the amount of forage produced in June. B. General view of high prairie at Lincoln, Nebraska. Plate 2. A. Mixed-prairie station at Phillipsburg, Kansas. B. Corner of short-grass plains station at Burlington, Colorado. Plate 3. Root system of year-old Andropogon scoparius ( A ) and Stipa viridula (B), upland cultivated soil at Lincoln; maximum penetration 4 and 2.7 feet respectively. Plate 4. Year-old Elymus canadensis (A), Bouteloua racemosa (B), and Andropogon nutans ( C ); maximum penetration about 3, 3, and 4 feet respectively. Plate 5. A. Gravel-knoll station, Lincoln. B. Low-prairie station, Lincoln, showing transplanted sods. Plate 6. A. Low-prairie station in September showing height of the subclimax dominants. B. Salt-flat station, Lincoln; the chief grass is Distichlis spicata. Plate 7. A. View of sod transplants in the swamp at Lincoln, showing water standing on the surface. B. Subclimax prairie at Nebraska City, showing rank growth of Stipa spartea and Ceanothus ovatus held in check by mowing. Plate 8. A. A single season’s growth of Andropogon nutans from block of sod (left) and from seed (right) on low prairie at Lincoln. B. Year-old Bouteloua hirsuta grown on high prairie at Lincoln. Plate 9. A. Year-old seedlings of Andropogon scoparius (right), A. nutans (center), and A. furcatus (left) grown on high prairie; plants about 6-8 inches tall. B. Root system of year-old Andropogon furcatus from cultivated lowland, Lincoln; scale in feet. Plate 10. A. Bulbilis dactyloides sod in high prairie. B. Andropogon nutans in September of first season, showing marked growth with¬ out competition. Plate 11. A. Calamovilfa longifolia in September of first season, grown in rich silt-loam. B. Bouteloua gracilis one year old, grown in cultivated soil. Plate 12. Roots of year-old Muhlenbergia pungens (A) and Calamovilfa longifolia (B) grown in rich silt-loam; depths of penetration 2.5 and 6 feet respectively. C. Andropogon halli at the end of the first year in silt-loam. Plate 13. Andropogon nutans planted in denuded quadrats at Nebraska City, (A) in 1922, (B) in 1920; photographed June 3, 1922. Plate 14. A. Two-year-old Liatris punctata (left) and L. scariosa (right) grown without com¬ petition in cultivated soil: the largest plant had 107 heads. B. Year-old Desmodium canescens and two-year-old Onagra biennis ( C) growm in cultivated soil at Lincoln; maximum heights 5 and 7 feet respectively. Plate 15. A. Sod-cores and details of installation, Burlington. B. Clip-quadrat in Bulbilis-Bouteloua short-grass, Burlington. v VI LIST OF ILLUSTRATIONS. TEXT-FIGURES. page. 1. Map showing location of stations in the four grassland communities. 10 2. Mean annual precipitation at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines). 15 3. Average daily temperature at Lincoln (solid line), Phillipsburg (long broken lines) and Burlington (short broken lines), 1920; here, as elsewhere, the numbers denote the first, second, third, or fourth week in the month, respectively. 17 4. Average day (heavy lines) and night temperatures (light lines), at Lincoln (solid lines), Phillipsburg (long broken lines), and Burlington (short broken lines), 1920. 17 5. Average daily soil-temperatures at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1920. 18 6. Hvgrothermograph records from Lincoln (upper) and Burlington (lower), June, 1921; light lines temperature, heavy lines humidity. 19 7. Average day (light lines) and night humidity (heavy lines) at Lincoln (solid lines), and Burlington (broken lines), 1920. 20 8. Average daily evaporation at Burlington (short broken lines), Phillipsburg (long broken lines), and Lincoln (solid line), 1920. 20 9. Monthly precipitation at Lincoln 1920-1922; the monthly mean is shown by the heavy cross-bars. 27 10. Plants of Bouteloua gracilis (A), Bouteloua hirsuta ( B ), and Sporobolus asper ( C) 44-days old; scale 1 foot. 28 11. Bouteloua gracilis 3 months old. 29 12. Bouteloua hirsuta 3 months old. 29 13. Stipa spartea 3 months old. 30 14. Agropyrum glaucum 3 months old. 30 15. Monthly precipitation at Phillipsburg 1920-1922; the monthly mean is shown . by the heavy cross-bars. 32 16. Monthly precipitation at Burlington 1920-1922, the monthly mean is shown by the heavy cross-bars. 32 17. Root of Liatris punctata at end of first season’s growth. 35 18. Monthly precipitation at Nebraska City 1920-1922; the monthly mean is shown by the heavy cross-bars. 37 19. Average daily evaporation, high prairie (solid line), and low prairie, Lincoln (short broken lines), and Nebraska City (long broken lines), 1920..... 48 20. Average daily evaporation on high prairie (upper line) and low prairie (lower line), 1920. 52 21. Average day and night temperatures at Lincoln (solid lines), and Colorado Springs (broken lines), during 1919. 53 22. Average daily evaporation at Colorado Springs (upper line), and Lincoln (lower line), during 1919. 54 23. Average daily evaoraption at Burlington (short broken lines), Phillipsburg (long broken lines), and Lincoln (solid line), 1921. 56 24. Average daily temperature at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1921. 57 25. Average day (heavy lines) and night temperatures (light lines), at Lincoln (solid lines), Phillipsburg (long broken lines), and Burlington (short broken lines), 1921. 59 26. Average daily soil temperatures at depths of 3 and 18 inches respectively at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1921. 62 27. Liatris punctata from Phillipsburg, on July 1 of second year after planting. ... 63 28. Bouteloua hirsuta from Phillipsburg, on July 1 of second year after planting... 63 29. Andropogon scoparius from Phillipsburg, on July 1 of second year after planting. 64 30. Roots of Andropogon nutans after the block of sod had been transplanted for 2.5 months at Burlington. 64 31. Roots of Elymus canadensis at the end of June of the second year after the block of sod had been transplanted and watered at Burlington. 64 LIST OF ILLUSTRATIONS. VII TEXT-FIGURES— Continued. page. 32. Roots of Panicum virgatum at the end of June of second year after the block of sod had been transplanted and watered at Burlington. 69 33. Average daily evaporation on high prairie (solid line), gravel-knoll, Lincoln (short broken lines), and at Nebraska City (long broken lines), 1921.. 71 34. Pinus ponderosa 2.5 months old. 76 35. Liatris punctata 2.5 months old. 76 36. Robinia pseudacacia 2.5 months old. 77 37. Root system of Gleditsia triacanthus less than 3 months old. 78 38. Average daily temperatures at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1922. 91 39. Average day (heavy lines) and night temperatures (light lines) at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1922. 92 40. Average daily soil temperatures at depths of 3 and 12 inches respectively at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1922. 93 41. Average daily evaporation at Lincoln (solid lines), Phillipsburg (long broken lines), and Burlington (short broken lines), 1922. 94 ' . EXPERIMENTAL VEGETATION THE RELATION OF CLIMAXES TO CLIMATES BY Frederic E. Clements and John E. Weaver EXPERIMENTAL VEGETATION. 1. INTRODUCTION. SCOPE AND PLAN. The investigations that have led to the development of the phytometer and the experimental method in vegetation are companion efforts in the en¬ deavor to organize ecology as a quantitative science. No subject has suf¬ fered more from the lack of training and experience or from the absence of basic procedure and adequate perspective. It has become obvious that this condition can be remedied and actual progress insured only by instrumental and experimental methods and by a combination of intensive and extensive studies that will check superficiality on the one hand and broaden limited horizons on the other. It is certain that much work of purely superficial or local character will continue to be done under the name of ecology, but the touchstones of instrument, quadrat, and experiment afford a ready means of eliminating such papers from consideration. No study deserves to be called ecological that does not deal with the cause-and-effect relation of habitat and organism in a quantitative and objective manner. The task of focusing the efforts of the ecologist upon the habitat as cause and the plant and community as effect and reaction furnishes the theme for four related studies, namely, phytometers, experimental vegetation, climatic cycles, and changes of climate and climax. The first two constitute an endeavor to employ plant and community as the best measures of cause and reaction, while the last two emphasize the fundamental significance of change as the proper approach to this complex of problems. PRINCIPLES. The basic guides in this task are the cause-and-effect relation on the one hand and the cyclic nature of change on the other. The former is paramount in every ecological problem, since investigation that does not deal directly and constantly with causes lacks the very essence of science. However, in dealing with processes in a way that is at once complete and synthetic, ecology must not overlook the fact that both cause and effect must be inseparably associated and that effect sometimes affords the readier approach to the prob¬ lem. It is equally essential that ecology be treated as inherently dynamic, that change in some degree or at some rate be regarded as universal, and that such changes be recognized as recurrent or cyclic, both as cause and in effect. The fact has elsewhere been emphasized that effects in turn produce reactions, which consequently become causal to the extent that they modify the original causes (Plant Succession, 79). This constitutes the basic cycle which finds its expression in the larger cycle of the sere. Thus, succession is an intrinsic function of all vegetation, though the shifting of climaxes is a matter of thou¬ sands or tens of thousands of years rather than of scores in the subsere or hundreds in the prisere. 3 4 INTRODUCTION. The cyclic changes of the habitat are likewise reflected in the individual and the species. Adjustment and adaptation are clearly cyclic processes, and this is even more evident in the major evolutions of groups and floras. As the consequence of reaction and competition in the sere and of climatic shift in the major seres, invasion is necessarily cyclic and finds its typical expression in terms of dominant and relict phases. Finally, cyclic changes serve to link the present with the past in such a broadly continuous fashion that the gap between them disappears and such terms as ecology and paleo-ecology are seen to be mere conveniences. METHODS. The methods of experimental vegetation are based on the processes inher¬ ent in the natural development of vegetation, but segregated and controlled in such a way as to reveal the value and significance of each individual step or process. This is even true of migration, regarded merely as transport, the difference in the role of man as an unintentional or purposeful agent being immaterial, except as to the number and fate of the migrules. The most exact evidence as to the nature and success of invasion and ecesis is doubtless that to be obtained in the form of natural experiments, but except in the case of propagules these are difficult to discover outside of ecotones and are alto¬ gether absent from the typical areas of great associations. Hence, the only feasible procedure is to approximate the method of natural experiment as closely as possible, bearing in mind that the numbers must be large enough to insure against complete loss. Moreover, it must be recognized that arti¬ ficial aids to ecesis, such as watering, screening, etc., fall well within the vari¬ ation of factors in nature and are in consequence more or less natural. The basic methods of experimental vegetation are as follows: (1) sowing seeds; (2) planting seeds or propagules; (3) planting seedlings; (4) transplant¬ ing adult plants of various ages; (5) transplanting small communities or por¬ tions of communities. The converse of this is to be found in the method of denuded quadrats and transects, but as this is primarily a question of ecesis within a community, it has not been extensively employed in the present study. The accessory methods involve the use or manipulation of (1) com¬ petition; (2) physical factors; (3) protection against animals; (4) instruments; (5) phytometers; (6) seasons and cycles. Time also enters the equation, inas¬ much as germination and the ecesis of seedlings are periods of the greatest stress. Various ways are available for reducing or eliminating competition and modifying the physical factors, but the following have proved the sim¬ plest and most satisfactory, viz: (1) sowing on the surface in the midst of natural vegetation; (2) sowing or planting in trenches, by which competition is prevented for a short time; (3) sowing or planting in denuded quadrats, which eliminates competition for a much longer period, but renders the water relations usually less favorable at the same time that it improves light con¬ ditions; (4) transplanting adult plants into living cover or denuded areas, with similar consequences as to competition, water, and light; (5) improving the conditions for germination, establishment, or survival by watering, shad¬ ing, thinning, draining, etc. A full battery of recording instruments is indis¬ pensable and must be supplemented by determinations of chresard, evapora¬ tion, and light intensity in all cases, and of salts and air-content whenever GRASSLAND CLIMAX. 5 these are controlling. Equally essential are phytometers of various types to serve as interpreters between the physical instruments and the natural cover on the one hand, and the transplants on the other. Furthermore, the differences between seasons and the wet and dry phases of the climatic cycle must be taken into definite consideration, as they are often greater than the climatic differences between the several grassland associations or even between formations. Finally, a clear distinction must be drawn between persistence and dominance, the former indicating the potential and the latter the actual role of a species in the climax. VALUES. The basic value of the experimental method in vegetation is evident, but its far-reaching significance warrants detailed consideration. While its first importance inheres in its being indispensable in the search for causes, the problems of the origin, unity, and relationship of formations and associations and their correlation with climate can be solved objectively and permanently in no other way. In fact, the objectivity afforded by comprehensive and repeated experiment is the paramount reason for its constant and universal use. The opinions and hypotheses arising from observation are often inter¬ esting and suggestive and may even have permanent value, but ecology can be built upon a lasting foundation solely by means of experiment. This does not mean that observation and reconnaissance have no value, but such results are to be regarded as provisional, pending adequate experimental study. Still less does it mean that hypothesis and theory are to have no part in an exact and quantitative ecology. Such a view is possible only to those who ignore the decisive role that theory has played in the development of modern chem¬ istry and physics and are unfamiliar with its essential stimulus in the ad¬ vance of biology. The study of vegetation demands the use of experiment to even a greater degree than that of the plant, owing to the complexity and extent of the great communities and the consequent opportunity for making general observations and forming unchecked opinions. Moreover, it must be fully recognized that the intensive-extensive method of investigation applies with peculiar force to the plant community, since it was developed for the express purpose of turning complexity and extent to the advantage of thorough and detailed analysis. As a procedure, it has been fully justified by the experience of years in the vegetation of western North America, in which the extensive results of constant field studies have given perspective and balance as well as new points of departure to the results of the laboratory and station, and the latter have served to refine and definitize the natural experiments and observations over a vast territory. THE GRASSLAND CLIMAX. The development of the views as to the nature and structure of the grass¬ lands of North America illustrates the need of objective methods of determin¬ ing vegetation units and their relationships. This is all the more convincing, since the ecological investigation of the prairie and plains has been the work of a group with the same general training and outlook. In the first analysis of the grassland, Pound and Clements (1898:243, 1900:347) recognized two 6 INTRODUCTION. prairie formations, viz, the prairie-grass and the buffalo-grass formations, a bunch-grass formation of the sandhills, and a meadow formation. In the light of successional studies, the last two are to be regarded as subclimaxes. In a few years (Clements, 1902) it had become clear that the prairie-grass or Stipa-Agropyrum formation and the buffalo-grass or Bulbilis-Bouteloua formation were the two great communities of the prairie-plains region. This was essentially the view also of Shantz (1906, 1911) and of Pool (1914). This conception was maintained in Plant Succession (180, cf. note) after many additional years of successional research. However, the developmental concept of the formation had broadened its scope and afforded a clearer view of its structure. As a consequence of the special study of these relations, it became necessary to abandon the view of two separate grassland formations and to recognize a single formation composed of several associations. Mean¬ while, it had become increasingly evident that the Agropyrum consociation of the Northwest was closely related to the Stipa-Agropyrum prairie. This was first suggested by finding the three dominants associated from Washing¬ ton to Montana during the field work of 1914, a view adopted by Weaver, who had worked intensively in this community (1917:40). This idea was confirmed by further studies in 1917, but the true relationship was obscure until it became certain in 1918 that Stipa setigera and S. eminens were the original bunch-grasses of California. As a consequence, it proved possible to recognize a fourth grassland association, composed of bunch-grasses and characteristic of the Pacific region of winter precipitation (Plant Indicators, 115). By means of extensive field work and the comparative study of com¬ munities, the actual relationship of the grassland units was determined, but the experimental method might well have accomplished in 5 years what obser¬ vation required 20 years to attain. TESTS OF UNITY. In reaching the conclusion that the grasslands constitute a single climax formation and not several more or less related but distinct units, recourse was had to a number of tests, all as objective as possible. “The conclusion that the grassland is a single great climax formation is based in the first place on the fact that the three most important dominants, Stipa, Bouteloua, and Agropyrum , extend over most of the area, and one or the other is present in practically every association of it. This would seem the most conclusive evidence possible, short of actual vegetation experiments, that the grassland is a climatic vege¬ tation unit. Equally cogent is the fact that these dominants, together with Carex, Bulbilis, and Koeleria, mix and alternate in various groupings throughout the Stipa- Bouteloua association. Indeed, this association appears so conclusive as to the gen¬ eral formational equivalence of these seven dominants that it is regarded as the typical or base association. In addition, the characteristic societies either extend through several of the associations or are represented by corresponding communities belonging to the same genus. The relation of the associations to such subclimax species as Andropogon scoparius, Calamovilfa longifolia, Aristida purpurea, and Elymus sitanion further confirms the relationship of the dominants. The most obvious differ¬ ence between the various associations is exhibited by the tail-grass prairies, Stipa- Koeleria poium, and the short-grass plains, Bulbilis-Bouteloua poium. Yet these are closely related, as shown not only by the criteria given above, but also by their geographical contact. Still more eloquent is the fact that grazing favors Bouteloua and Bulbilis at the expense of Stipa and Agropyrum, and thus frequently converts the GRASSLAND CLIMAX. 7 base association of Stipa-Bouteloua into a pure short-grass cover. Concrete evidence of this has been obtained in widely separated areas and has led to the working hypoth¬ esis that a pure short-grass community is partly if not largely a response to over- grazing.” (Plant Indicators, 115.) FURTHER EVIDENCE. In the six years since the above was written, a large amount of additional evidence has been accumulated as to the unity of the grassland and the close phylogenetic relationship of its associations, quite apart from the experi¬ mental results brought forward later. In fact, the detailed field study of the ranges of the dominants makes it clear that the associations are even more closely related than was at first supposed, and the problem becomes one of maintaining several of these units as distinct rather than one of justifying their inclusion in the same formation. The number of dominants common to contiguous units has steadily increased, as well as the areas concerned. The most unique of the associations, the short-grass plains, has been shown to be a recent modification of mixed prairie due to overgrazing and has been more closely connected with the desert plains. The widespread dominants that occur in all the climatic associations have increased in number and now com¬ prise Stipa comata, Agropyrum glaucum, Bouteloua gracilis, B. racemosa, Sporobolus cryptandrus, and Koeleria cristata. Even more striking has been the evidence drawn from relicts and from sea¬ sonal variations arising out of the climatic cycle (Clements, 1921, 1922, 1923). These are the results of natural experiments on a scale and over an area that can not even be approximated by the experimenter, and their value is of the greatest, especially when supplemented by control experiments. In addition, the repetition and checking that are so indispensable are not only provided by recurrence in thousands of localities, but particularly also by the wide variation in behavior from season to season at the wet and dry phases of the climatic cycle. Of the first and the most direct importance are the relicts produced by overgrazing and by rodents, since these factors have operated directly upon the normal climax. Such relicts permit the complete recon¬ struction of the original community, both as to structure and extent. Indeed, this is usually done in the most graphic fashion by nature herself during sea¬ sons or phases of excessive rainfall, and when a drought period of several years is followed immediately by a year of exceptional rainfall, as happened from 1916-1919, a revealing picture is obtained. The expansion and con¬ traction of relict areas at such a time reproduces in miniature what occurs during the wet and dry phase of a major climatic shift extending over a period of a thousand years or more. In the light of grazing relicts and cyclic changes it has proved possible to reconstruct all the grassland associations in detail as they were before the historical period in the West. Furthermore, recon¬ struction has been carried much farther back to a time probably 10,000 to 20,000 years ago, when the differentiation of the grassland mass into the various associations had barely begun. In short, the relict method confirms the view of the unity of the grassland climax by disclosing several stages in the differentiation of the original mass and affording a measure of the extent to which this has progressed. This is the theme of the volume on the changes of climate and grassland, and hence it will not be further elaborated here. 8 INTRODUCTION. RELATIONSHIP OF THE ASSOCIATIONS. The above suffices to indicate what appears to be clearly proven by the detailed evidence, namely, that the relationship of the six communities of the grassland is a phylogenetic one. This seems almost obvious in the case of the subclimax prairie and the short-grass plains. The latter is demon¬ strably a grazing subclimax of the mixed prairie, though such a hard-pan area as that about Burlington may be an actual relict of a short-grass climax developed during the last major dry phase. On the other hand, the sub¬ climax prairie is just as clearly an eastward extension of the taller meadow- grasses from the true prairie, made in response to the same climatic shift and profiting by fire, especially during the prehistoric period. These changes are still so recent that the detailed steps can be readily followed, and similar adjustments can be traced between the other associstiona. However, the latter have been largely differentiated at an earlier period, and the phylo¬ genetic and geographic connections are less broad. The clue to the relationships of the four actual associations is to be found in the mixed prairie, which most nearly represents the original formation, and consequently affords a measure of the divergence of the others. The true prairie stands closest to the mixed prairie, Agropyrum glaucum and Koeleria cristata being common dominants, and the closely related Stipa spartea and S. comata, and Sporobolus asper and S. cryptandrus being reciprocal dominants in the two respectively. The chief difference lies in the fact that the short- grasses of the Stipa-Bouteloua association are usually absent, or form a frag¬ mentary layer in the true prairie, except where grazing has favored their spread. The bunch-grass association resembles the true prairie in the absence of the short-grasses, but its endemic dominants are naturally different, e. g., Agropyrum spicatum, Stipa setigera, and Festuca ovina in place of Stipa spartea and Sporobolus asper. Its cognate relation to the mixed prairie is shown by such common dominants as Stipa comata, Agropyrum glaucum, and Koeleria cristata, and by the persistence to-day at the higher altitudes in the mixed prairie of such bunch-grasses as Stipa setigera, S. eminens, and Festuca ovina. The desert plains bear a close resemblance to the short-grass condition of the mixed prairie, the reduction of the upper layer of tail-grasses being more a matter of drought than of grazing. The close relationship of the two is shown by the dominance of Bouteloua in both, as well as by the fact that such tail-grasses as Sporobolus cryptandrus, Bouteloua racemosa, Stipa pen - nata, and S. comata persist as an upper layer in the desert plains wherever protection permits. Moreover, the relict areas in and about the Mohave Desert and Death Valley prove beyond doubt that the desert plains and bunch- grass associations were once in contact and that at an antecedent period they were commingled and probably identical (Clements, 1922, 1923). CLIMATIC UNITY. The fundamental fact in the relation of climaxes to climate is that this must be determined by plant judgments rather than by human ones (Plant Indicators, 116). This is necessarily truer of grasses than of trees, since their life-habits enable them to minimize or escape the rigors of winter as well as the dangers of the dry season of the year. Consequently, it should be regarded as neither strange nor perplexing to find such great grassland dominants as Bouteloua gracilis, Stipa comata, and Koeleria cristata ranging through a num- OBJECTIVES. 9 ber of climates in the human sense. The human conception of a climate is loose enough at best, and it regularly ascribes much greater importance to temperature than to water relations. On the contrary, grasses are more responsive to differences of rainfall and evaporation than to those of tempera¬ ture, Bouteloua gracilis in particular ranging much more extensively in lati¬ tude than in longitude as a climatic dominant. This is confirmed by the construction of dominance maps of the chief grasses, which disclose the fact that the interruption or disappearance of a consociation is primarily a matter of water, a conclusion entirely supported by the present investigation. In addition, it must be recognized that obvious differences in the amount and distribution of rainfall may have slight significance for grasses. It is the amount and distribution during the growth period of the grass dominants that controls, and this is equally true of the effect of evaporation. The mixed prairie may range from 25 to 10 inches without undergoing any essential change in composition, owing to the fact that the period of development is shortened to fall within the limits of optimum water conditions. The exist¬ ence of two rainy seasons, as in southern Arizona, seems to indicate an impor¬ tant difference in climate, but the growth of the grassland is still an effect of the much more adequate summer rains. The rainfall of the Pacific Coast is of the winter type and at first thought appears entirely different. However, this is only apparently true in so far as the native bunch-grasses are con¬ cerned. These bear the same general relation to the spring rains as that pre¬ vailing in the mixed prairie, namely, development begins during the rains, followed by maturity and the drying of the aerial parts in late spring or early summer respectively. The differentiating effect of the peculiar distribution of rainfall has been recorded in the characteristic life-form, though the progressive assumption of the bunch-grass habit is revealed by the fact that it is found in practically all the dominants of the desert plains and in nearly all the tail-grasses of the mixed prairie. OBJECTIVES OF THE PRESENT INVESTIGATION. As has been already indicated, the primary purpose of these researches is to develop an experimental method that will combine the maximum of demon- strability and objectivity. Because of its size and apparent complexity, vegetation affords an open field for prepossession and interpretation, and its study can be made scientific in the proper degree only by the resolute employ¬ ment of experiment and measurement. The special ends of this investigation are many, chief among which are the demonstration of the unity of the grass¬ land climax and climate, and the growing differentiation of the associations, partly in response to the subclimates and partly to the changes wrought directly or indirectly by man. It has also proved possible to cast further light upon the relationship of climatic and edaphic areas in the same region, as well as upon the significant role of climatic cycles and communi'y shiftings in this. A special endeavor has been made to analyze competition in minuter detail and to evaluate more fully the effect of light intensity in a grassland cover. Par¬ ticular attention has been directed to germination, seasonal ecesis, and final survival, especially in relation to annual variations in climate. Finally, the entire study has been designed to exemplify a new aspect of the phytometer method, in which native species and communities have received the major emphasis. 10 INTRODUCTION. STATIONS. Location. The stations were chosen to represent typical conditions in each of the four grassland communities occurring in the vast area between the Missouri River and the Rocky Mountains. Lincoln, Nebraska, was selected as repre¬ sentative of true-prairie conditions. Aside from the matter of convenience in the use of the facilities of the Department of Botany of the University of Nebraska, the vegetation and environment of this region are better known than anywhere else in the grassland formation, due especially to continued study by Pound and Clements (1897, 1898, 1900), Clements (1905, 1907, 1916, 1920), Pool (1914), Weaver (1917, 1919, 1920, 1921, 1922, 1923), and others. Phillipsburg, in north-central Kansas, was selected as representative of mixed-prairie conditions, while Burlington, in eastern Colorado, was chosen in the short-grass plains. Phillipsburg lies about 190 miles southwest of Lincoln, and Burlington about 180 miles further west and south, 10 miles Fig. 1 . —Map showing location of stations in the four grassland communities. west of the Kansas-Colorado State line (fig. 1). The altitude at the several stations rises from 1,100 feet at Lincoln to 1,900 feet at Phillipsburg and 4,160 feet at Burlington. Precipitation, the chief factor in determining the type of vegetation, ranges from 28 to 23 and 17 inches at the respective sta¬ tions, decreasing westward. A fourth station was maintained at Nebraska City, Nebraska, about 50 miles southeast of Lincoln, near the Missouri River. Although the rainfall of 33 inches is sufficiently ample to produce scrubland or woodland, except for grazing, mowing, fires, etc., this station lies just within the border of the subclimax grassland and was chosen largely because of its ready accessibility. In addition to the four stations mentioned, considerable experimental work was carried on at the mixed-prairie station at the foot of Pike’s Peak near Colorado Springs, the Alpine Laboratory being used as a base. At Lincoln STATIONS. 11 a whole series of stations, ranging through gravel-knoll, high prairie, low prairie, salt-flat, swamp, and cultivated fields, was maintained. Moreover, reciprocal transplants from Tucson, Arizona, and Berkeley, California, have been made at many of the stations. True-prairie Stations. The high-prairie station at Lincoln (plate 1) was located on a rather flat hilltop about 60 feet above the general level of the flood-plain of Salt Creek and 2 miles north of the city. The fertile soil is of the type commonly called loess, but is really a glacial drift. It is a silt-loam belonging to the Marshall series, with a moisture equivalent of about 28 per cent and a maximum water capacity of approximately 60 per cent. The mechanical and chemical analyses of soils taken from an adjacent field in the same area (Weaver, 1920:140) indicate its general characteristics (tables 1 and 2). Table 1 —Mechanical analysis of high-prairie soil. Depth of sample. Coarse gravel. Fine gravel. Coarse sand. Me¬ dium sand. Fine sand. Very fine sand. Silt. Clay. 0.0 to 0.5 foot. p. ct. 0.0 p. ct. 0.0 p. ct. 2.0 p. ct. 1.9 p. ct. 6.0 p. ct. 26 1 p. ct. 39 3 p. ct. 24 7 0.5 to 1.0 foot. 0.0 0.0 1.2 2.2 3.8 21.8 38.4 32.6 1 to 2 feet. 0.0 0.0 2.7 2.9 4.8 19 6 45 6 24 4 2 to 3 feet. 0.0 0.0 5.6 6.9 10.1 23.1 32.8 21 5 3 to 4 feet. 0.0 0.0 7.0 8.8 12.8 23.5 28.3 19.6 Table 2. —Chemical analysis of high-prairie soil. [Digestion with HCL (sp. gr. 1.115) for 120 hours.] Depth of sample. Insolu¬ ble resi¬ due. Solu¬ ble salts. Vola¬ tile mat¬ ter. Iron and alu¬ minium oxid. Cal¬ cium oxid. Mag¬ nesium oxid. Phos¬ phorus pent- oxid. Nitro¬ gen. 0.0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. p. ct. 76.87 75.70 76.17 77.80 p. ct. 17.12 18.58 19.08 18.46 p. ct. 6.01 5.72 4.75 3.68 p. ct. 13.20 14.25 14.72 14.03 p. ct. 0.68 .70 .75 .86 p. ct. 1.19 1.32 1.68 1.69 p. ct. 0.13 .12 .12 .15 p. ct. 0.159 .134 .079 .045 The soil is fine in texture, being composed mostly of silt and clay, and is sufficiently supplied with calcium to lack acidity. Determinations through¬ out a number of years show that the subsoil is usually moist to great depths, although at infrequent intervals during drought periods the holard may be reduced below the wilting-coefficient of Briggs and Shantz (1912) to depths of 4 to 5 feet, leaving only a small chresard for vegetation. Root excavations and bisects show that the plants are not only rooted deeply, but also that the root-systems of different species form layers in the soil, the shallowest one ending at about 3 feet and an intermediate one at 5 feet, while a third layer extends far below this level (Weaver, 1920:28, 40). 12 INTRODUCTION. The vegetation is distinctly of the tail-grass sod type. Andropogon sco- parius, Stipa spartea, Koeleria cristata, and Bouteloua racemosa are the chief grasses, although Andropogon fmeatus and nutans occur more or less along with Poa pratensis. The interstitial Panicum scribnerianum and the relict Bouteloua gracilis are of much less importance. Prevernal societies are repre¬ sented by Antennaria campestris and Carex pennsylvanica. Such vernal bloomers as Astragalus crassicarpus, Baptisia bracteata, Senecio plattensis, and Nothocalais cuspidata are abundant, while the variety and abundance of esti¬ val herbs indicate favorable growth conditions throughout the early summer. Chief among these are Psoralea floribunda, Erigeron ramosus, Brauneria pal¬ lida, Meriolix serrulata, and Achillea millefolium, although a host of others occur (Pound and Clements, 1900; Weaver and Thiel, 1917). Species of Solidago, Aster, Liatris, Helianthus, Kuhnia, etc., constitute the most impor¬ tant autumnal societies, the subdominants during summer and autumn giving a truly kaleidoscopic appearance. An average grass-level of 6 to 8 inches and an upper story of forbs at 15 to 22 inches is attained by June 1, although the flowerstalks of Stipa and the later grasses and herbs are 2.5 to 3.5 feet tall (plate 1). Mixed-prairie Station. This occupies an area just south of Phillipsburg, quite typical of the gently rolling topography, on a hillside which slopes gently southward (plate 2a). The fertile soil is a mellow, dark-brown, very fine sandy loam of the Colby series. At a depth of 12 to 15 inches it is slightly lighter in color and con¬ tains enough clay to be quite sticky, although when wet it is dark in color to a depth of 2 feet. Below this level it is light yellow and shows its loess origin throughout. The first 4 feet have a water-holding capacity of about 66 per cent. As is true of most soils of semiarid regions, it shows no acidity at any depth, and the mellow subsoil is very deep. Table 3. —Mechanical analysis of soil at Phillipsburg, Kansas. Depth of sample. Coarse gravel. Fine gravel. Coarse sand. Me¬ dium sand. Fine sand. Very- fine sand. Silt. Clay. p. ct. p. ct. p. ct. p. ct. p. ct. p. ct. p. ct. p. ct. 0.0 to 0.5 foot. 0.0 0.0 0.3 0.2 1.2 43.5 35.8 19.0 0.5 to 1 foot. 0.0 0.0 0.0 0.2 0.5 44.4 32.8 22.1 1 to 2 feet. 0.0 0.0 0.0 0.2 0.3 39.7 34.0 25.8 2 to 3 feet. 0.0 0.0 0.0 0.3 0.6 41.2 31.9 26.0 3 to 4 feet. 0.0 0.0 0.0 0.1 0.2 37.5 31.4 30.8 — Repeated excavations for the examination of the roots of native and crop plants from 1919 to 1921 showed that the soil was moist to 8 feet. This condition is assumed to be abnormal for the region, but it is directly to be correlated with the excessive precipitation of 1919 (Weaver, Jean, and Crist, 1922:77). However, this station is somewhat subject to drought, the holard to a depth of 4 feet being sometimes reduced approximately to the hygroscopic coefficient, about 10.6 per cent. Under these conditions the native vegetation is rooted almost or quite as deeply as in the true prairie (Weaver, 1920:93). The vegetation is essentially mixed prairie, the tail-grasses alternating with STATIONS. 13 or forming a layer above the short ones (plate 2a). Andropogon scoparius, nutans, and furcatus often form more or less continuous irregular sods varying from 6 inches to 7 feet in diameter, from which short-grasses may be almost entirely excluded, while Agropyrum glaucum likewise frequently occupies large areas rather exclusively. Bouteloua racemosa and Elymus canadensis are other important tail-grasses. Alternating with these are similar or on drier slopes even larger areas of Bulbilis dactyloides and Bouteloua gracilis, the intervals being nearly devoid of vegetation, often to the extent of one-fourth of the surface. More usually, however, the two kinds of grasses are inti¬ mately mixed, the short-grasses often showing a strong tendency toward the bunch habit. Car ex filifolia and stenophylla supplement the understory of grasses, the leaves of which reach an average height of about 4 inches (before flower-stalk production), as contrasted with the midsummer tail-grass level 4 to 10 inches above. A taller open layer of Psoralea tenuiflora at 2 feet characterizes much of the area in late June, when societies of Ratibida colum- naris and Morongia uncinata are also conspicuous. However, as emphasized by Clements (1920:138), the mixed prairie shows its xerophytic tendency by less numerous and less extensive societies. Antennaria campestris, Astrag¬ alus crassicarpus, Nothocalais cuspidata, Anemone caroliniana, Senecio plat- tensis , Vida americana, etc., are all represented, even if sparingly, in the spring and early summer, but the absence of Viola, Stipa, Koeleria, and Brauneria is at once noted, and the presence of Astragalus mollissimus, Oxy- tropis lamberti, Malvastrum cocdneum, Opuntia fragilis, 0. camanchica, Aristida purpurea, and Plantago purshi indicates a more xerophytic type of vegetation. The sub dominants of midsummer are usually smaller and less abundant than eastward, while the autumnal aspect likewise lacks many species common to the true prairies. The part of the mixed prairie used for experimental planting was carefully selected with due regard to a proper balance between tall and short grasses. Plains Station. This is located just north of Burlington, on a vast level tract (plate 2b) . The soil is a rich, brown, fine sandy loam, very compact and hard when dry. It has a water-holding capacity of 65 to 70 per cent to a depth of 4 feet. At a depth of 2 to 2.5 feet it is underlaid with a so-called hard-pan. Soil analyses show that the concentration of colloidal clay and carbonates in the subsoil is sufficient to give rise to a hard-pan, i. e., a much more compact intercalated stratum of soil, upon its becoming completely dried out (Weaver and Crist, 1922). Silt constitutes about one-third of the soil at all depths, while the sand decreases and the clay increases in amount to 4 feet (table 4). Table 4. —Mechanical analysis of soil at Burlington, Colorado. Depth of sample. Coarse gravel. Fine gravel. Coarse sand. Me¬ dium sand. Fine sand. Very fine sand. Silt. Clay. Hygro¬ scopic co¬ efficient. p. ct. p. ct. p. ct. p. ct. p. ct. p. ct. p. ct. p. ct. p. ct. 0.0 to 0.5 foot.. 0.0 0.0 0.0 0.1 2.6 48.6 33.4 15.3 10.9 0.5 to 1.0 foot.. 0.0 0.0 0.0 0.1 2.2 49.1 32.5 16.1 10.9 1 to 2 feet. 0.0 0.0 0.0 0.2 1.9 46.7 32.0 19.3 12.2 2 to 3 feet. 0.0 0.0 0.0 0.1 1.5 45.5 31.0 21.9 12.0 3 to 4 feet. 0.0 0.0 0.0 0.1 0.9 42.2 34.2 22.6 11.4 14 INTRODUCTION. Chemical analyses show that carbonates are practically absent in the surface soil, but increase rapidly with depth, and often reach concentrations of 5 or 6 per cent in the hardpan layer, which appears somewhat chalky in color. The soil is not at all acid; the carbon dioxid increases very rapidly with depth and is high at 2 to 4 feet. This soil is rich in phosphorus and potassium and has a sufficient supply of nitrogen, and hence all the essential raw materials are present in abundance. Table 5. —Chemical analysis 1 of soil at Burlington , Colorado. Depth of sample. Acidity. Carbon dioxid. V olatile matter. Phosphorus pentoxid. Sulphur trioxid. Potassium oxid. Nitro¬ gen. 0.0 to 0.5 foot.. 0.5 to 1.0 foot.. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. p. ct. None. None. None. None. None. p. ct. 0.03 0.30 1.71 2.10 2.60 p. ct. 4.67 3.13 3.11 3.34 2.84 p. ct. 0.189 0.504 0.428 0.406 0.525 p. ct. 0.007 0.017 0.006 0.006 0.005 p. ct. 2.32 2.39 2.45 2.51 2.22 p. ct. 0.184 0.130 0.101 0.086 0.084 1 Phosphorus determinations were made by digestion with HN0 3 and HF, sulphur by fusion with Na 2 C> 2 , potassium by fusion with calcium carbonate, and nitrogen by the modified Gun¬ ning method. Water penetrates very slowly and run-off is usually high in this fine- textured soil. Shantz (1911) has shown that the average run-off from the short-grass sod at five stations in this region was 37 per cent of the total rainfall, while the maximum run-off reached 55 per cent. Consequently, the actual precipitation of 17 inches is a poor index of the efficiency of the rainfall. After heavy rains, 3 days were required for water to penetrate to a depth greater than 6 inches. The excellent root development of native plants in the surface 1.5 to 2.5 feet of soil enables them to absorb water readily, and further prevents a deeper penetration of rain. Determinations of the holard through a series of years (1920 to 1923) show that the soil is seldom moist below 2 feet, while by midsummer the amount above this level is frequently reduced to the hygroscopic coefficient. When this happens, the short-grass cover dries out and “cures” on the ground. Closed mats of Bulbilis dadyloides, usually mixed with Bouteloua gracilis , make up fully 90 per cent of the vegetation, forming a carpet seldom over inches deep, exclusive of the flower-stalks (plate 2b). Agropyrum glaucum is frequently associated with the short-grasses, but overgrazing and drought have resulted in a dwarf habit and it forms flower-stalks sparingly, except in years of more than normal rainfall. These grasses, with their widely spreading roots, occupy the soil so thoroughly that relatively few important sub¬ dominants are present. Most conspicuous among these, and increasing in abundance where overgrazing has occurred, are Aristida purpurea, Opurdia camanchica, O.fragilis, 0. polyacantha, Festuca octoflora, Plantago purshi, and Schedonnardus paniculalus. Erysimum asperum and Psoralea tenuiflora also are often abundant, while small, poorly developed societies of Astragalus crassicarpus, Malvastrum coccineum, and Ratibida columnaris are infrequent. Decreased size, vigor, and number are characteristic of most of the species when compared with their growth in more favorable conditions. CLIMATES. 15 While some of the species of the meager flora are rooted entirely in the surface soil, others reach depths of 4 feet or more, especially the dominant grasses and the legumes. However, even these are relatively shallow when compared with the great depth attained by species in the moist subsoil of the true prairie. PHYSICAL FACTORS. Rainfall and Holard. The grassland associations concerned lie in the region of summer rainfall; by far the greater part of the rain falls during the growing-season and only about one-tenth during the three winter months (fig. 2). Such a seasonal distribution of the precipitation is very favorable to the growth of grasses. The normal decrease of 5 inches at Phillipsburg in comparison with Lincoln and the further decrease of 6 inches at Burlington are quite evenly dis¬ tributed through the season. Moreover, the type of rainfall is similar through¬ out, consisting usually of heavy showers, often of rather short duration, though this is more marked upon the high plains than eastward, where che rains are more general. At all stations it results in much run-off, but this is particularly heavy on the compact soil at Burlington. However, in the plains region more of the precipitation falls in light showers of 0.2 inch or less, which are of practically no value in increasing the holard. At each station the holard has in general been very similar at any particular time during the three seasons (1920 to 1922), and the comparison of a single season’s data will suffice here (table 6). Fig. 2. —Mean annual precipitation at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines). At Lincoln a sufficient amount of water to promote good growth was avail¬ able at all depths and at all times. At the mixed-prairie station, July and early August were periods of drought and at times of actual deficiency. The holard at Burlington was favorable until June, but after this time marked defi¬ ciencies were of frequent occurrence. However, the value of water-content to the plant is not determined entirely by its quantity, but largely also by the 16 INTRODUCTION Table 6 .—Holard in excess of the hygroscopic coefficient at the several stations, 1920. Lincoln, Nebraska. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 10. 17.6 16.1 14.1 10.0 8.6 Apr. 21. 17.5 22.4 20.3 May 5. 20.5 19.1 19.7 June 9. 18.7 20.3 June 16. 5.9 12.5 14.8 15.6 June 23. 4.9 7.1 9.6 July 14. 22.0 16.8 9.1 13.2 14.7 July 28. 7.2 3.9 6.3 Aug. 5. 2.5 4.0 5.4 9.7 11.6 Aug. 12. 8.7 5.7 4.8 Aug. 19. 26.3 7.7 3.2 Aug. 31. Continued heav 1 v rains: no sanmles taken. Wilting coef. 14.0 12.9 12.6 10.5 9.1 Hygroscopic coef. 9.5 8.7 8.6 7.1 6.2 Phillipsburg, Kansas. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. May 7. 14.7 15.3 12.5 14.7 14.0 June 2. 4.8 6.6 11.6 11.8 13.0 June 10. 7.6 9.7 9.0 June 24. 6.6 3.1 5.0 9.0 July 1. 1.9 4.1 2.8 July 9. 1.5 2.8 1.6 July 21. -1.4 0.1 -0.2 2.3 5.3 Aug. 4. 0.7 7.7 -0.3 0.5 3.5 Aug. 18. 12.6 12.3 5.4 Aug. 26. 4.0 4.1 0.4 2.0 5.5 Wilting coef. 13.3 13.3 13.4 13.5 13.1 Hygroscopic coef. 10.6 10.6 10.9 10.6 10.7 Burlington, Colorado. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 15. 16.7 13.7 11.1 4 9 1 8 June 3. 2.3 5.2 7.3 6.9 2 9 June 12. -1.8 -0 1 2 8 June 25. 7.4 2.5 1 8 1 4 -2 5 July 2. -1.6 -0.5 0 0 July 8. -2.9 -1.3 -2.2 July 20. -0.7 -2.7 -1.7 2.5 -0.6 Aug. 5. 4.6 2.7 -2.7 -0.6 -3 1 Aug. 19. -3.1 -3.1 -0.5 Aug. 24. -0.8 -2.1 -1 9 -0.6 14 5 -1 4 Wilting coef. 13.3 13.3 14 0 14 0 Hygroscopic coef. 10.9 10.9 12.2 12.0 11.4 CLIMATES. 17 rate of loss both through the plant and by surface evaporation. These in turn are controlled by humidity as affected by temperature, wind, etc., all of which are more or less integrated in the evaporation. June July August the numbers denote the first, second, third, or fourth week in the month, respectively. Fig. 4. —Average day (heavy lines) and night (fight fines) temperatures, at Lin¬ coln (solid fines), Phillipsburg (long broken fines), and Burlington (short broken fines), 1920. Temperature. Because of differences in elevation, which more than offset those of latitude, spring usually opens about 7 to 10 days later at Phillipsburg, and 18 to 23 18 INTRODUCTION. days later at Burlington, than at Lincoln. These stations are 800 and 3,000 feet higher respectively than Lincoln. Air temperature was usually 5° to 7° less at Burlington than at Phillipsburg, and also lower at the latter than at Lincoln. Likewise, the average day temperatures were highest at Lincoln and lowest at Burlington, and this same general relation held for average night temperatures (figs. 3 and 4). The night temperatures at Burlington varied from 45° to 67° F. These were more or less unfavorable to growth, but their chief effect was probably through humidity, differ¬ ences of temperatures as such probably having little effect upon the type of grassland. Fig. 5. —Average daily soil-temperatures at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1920. Soil temperatures were highest at Lincoln (70° to 77° F.) and lowest at Burlington (64° to 72° F.) during the first half of June, but by the last week this relation was reversed, the soil at Burlington remaining warmest through¬ out the season (fig. 5). The average weekly differences were often as great as 6° to 8° F., the soil at Lincoln being coldest, at Phillipsburg intermediate, and the dry soils at Burlington having the highest temperatures. The much greater daily range of both temperature and humidity at the Burlington station as compared with that at Lincoln is shown in figure 6, and the average day and night humidities for 1920 are given in figure 7. Conditions at Phillipsburg were more or less intermediate. This combination of high temperature and low humidity, which occurs rather regularly in the afternoons at Burlington, when coupled with dry soil, obviously promotes transpiration and depresses growth. Wind and Evaporation. The wind movement is much greater at Burlington than at either of the other stations and is an important factor in desiccating both plants and soil. An average day velocity of 8 or 10 miles per hour at a height of 0.5 meter is quite usual, while periods of several days with a velocity of 20 to 30 miles per hour are not uncommon. The amount of wind is less at Phillipsburg and much less at Lincoln (for example, a daily average of 4 miles per hour from CLIMATES. 19 July 13 to September 19, 1916). The evaporation, which in a measure inte¬ grates the factors of radiant energy, humidity, and wind movement (fig. 8), was greatest throughout the season of 1920 (23 to 60 c. c. average daily evaporation from white cylindrical non-absorbing atmometers) at Burlington, THURSDAY FRIDAY. SATURDAY. SUNDAY. Fig. 6.—Hygrothermograph records from Lincoln (upper) and Burlington (lower), June 1921; light lines temperature, heavy lines humidity. intermediate at Phillipsburg (11 to 32 c. c.), and least at Lincoln (9 to 25 c. c.). Similar constant differences of evaporation rates were obtained the following seasons. 20 INTRODUCTION. In consequence, the conditions for plant-growth as regards rainfall, holard, temperature, humidity, wind, and evaporation are normally most favorable at Lincoln, intermediate at Phillipsburg, and least favorable at Burlington. These conditions are indicated by the native vegetation and borne out by the growth of crop plants (Weaver, Jean, and Crist, 1922). June July August Fig. 7. —Average day (light lines) and night humidity (heavy lines) at Lincoln (solid lines), and Burlington (broken lines), 1920. Fio. 8.—Average daily evaporation at Burlington (short broken lines), Phillips¬ burg (long broken lines), and Lincoln (solid line), 1920. METHODS. 21 DETAILED METHODS. Germination of Seed. In these experiments the fruits and seeds of a large variety of grasses’ forbs, 1 shrubs, and trees from a wide range of habitats were employed. These were collected in quantity, when ripe, chiefly at Lincoln, Nebraska City and Halsey, Nebraska; Phillipsburg, Kansas; and Burlington and Colorado Springs, Colorado. In addition, others were secured from the vicinity of Berkeley, California; Tucson, Arizona, and in lesser quantities from other sources. After drying, the seeds were placed in large cloth bags and kept over winter in a dry, well-ventilated, unheated building at Lincoln. A new crop of seeds was secured each season. After threshing and cleaning the seed, germination tests were conducted in pots and flats n the greenhouse of the University of Nebraska, and to a less extent in moist chambers in the laboratory. This was primarily for the purpose of eliminating the species that germinated rarely or not at all, for it proved possible to sow seeds even with a low viability so thickly as to secure a good stand. Seeds from plants growing in sand, viz, Calamovilfa longifolia, Muhlenbergia pungens, Redfieldia exuosa, etc., were germinated in sandy soil. The vitality of the seeds differed greatly from season to season, as would be expected. Among various causes for this, climatic conditions at the time of anthesis may have been a potent factor. In some cases, early frosts on low¬ lands proved harmful to the unripe seeds of late species of grasses. While the 1920 crop of Kuhnia glutinosa gave no germinable seeds, that of 1921 yielded 84 per cent. This is the highest record for any species, values of 20 to 25 per cent ( Psoralea tenuiflora, Ratibida columnaris, Stipa viridula, Amorpha canescens, Desmodium canescens, and Liatris scariosa ) being very good, while a germination of 10 to 15 per cent was quite usual. Some of the legumes germinated very slowly. A few species that did fairly well in the greenhouse {Amorpha canescens and Aster multiflorus) gave no germination under field conditions (1922). Depth of Planting. In dealing with seeds so variable in size and natural depth of planting (viz, Sporobolus asper, Aster, and Solidago as compared with Stipa spartea and Gleditsia triacanthus), great care was exercised to place the seed at such a depth as to favor germination and establishment. The results of a single greenhouse experiment upon this phase of the problem are of interest (table 8). In this experiment 50 seeds of each species were planted at various depths in loam soil of good tilth, except where the seeds, because of small size or low viability, were used in larger but equal (measured) numbers. Sandhill species were grown in sandy loam. These results show that depth of planting has a pronounced effect upon germination and establishment. This is due in most cases to the actual failure of the seed to germinate at increased depth of soil and in others to its inability to grow through the greater thickness of soil and still have sufficient 1 The term “forb” is here used for herbs other than grasses in order to afford a clear cut dis¬ tinction and at the same time avoid an awkward phrase. It is derived from the Greek-Latin root which appears in 4°pP4. an d herba (*ferb—). 22 INTRODUCTION Table 7. —Germination of seeds . 1 Species. 1919. 1920. 1921. 1922. 1923. Ar>pr neonmdo^. Fair. Fair. Acer saccharinum. Fair. Seeds fro- Good.... A ornnvrnin erlaiieiiTn . . . zen; no crop Poor. Good.... None. Arrmmlia canescens. Good.... Andropogon con tortus 5 . Poor. fnrnftt.ns . Poor . _ . Poor. Poor. Poor. halli . Poor. Poor. nnt.a.ns . Poor .... Good. . . . Fair. Fair. Fair. ReonariiiR. Poor . . . Fair. None. . . . None.... Poor. Arcremono nlatvceras 1 * 3 . Poor. Ariatida. rmrnnrpa 1 . Good.... Poor. Excellent. Good.... Excellent. Aater mnltiflorus. None.... Fair. None. .aalieifoliiiR. None.... None.... Astrn.palns crassicarpus. None. . . . Fair. Roiiteloiia. bromoidea 6 . Good.... erionoda, 3 . None. . . . gracilis. Good.... Excellent. Poor. Excellent. hiranta. Poor . . . Excellent. Fair. Good.... None. racemosa. Fair. Poor. Excellent. Poor. racemosa 5 . Good.... Fair. rnthroeki 5 . None.... Rrannpria pallida. None. . . . Poor. None. Bulbilis dactyloides. Fair. Calamovilfa longifolia®. Fair. Very poor Poor. GaRRia chamaecrista. Fair. Cornus canadensis 2 . Very poor Very poor Gorylna n.mericana 2 . Desmodium canescens. Excellent. Elymus canadensis. Good. . . . Excellent. Excellent. Excellent. Fraxinus lanceolata 2 . Poor. Fair. Fair. Gleditsia triacanthus. Fair. Good.... Good.... Glycyrrhiza lepidota. None. . .. None.... Helianthus petiolaris 3 . None. . .. rigidus. None.... None. . . . None. Hilaria rigida 5 . None. . . . Koeleria cristata. Excellent. Excellent. Kuhnia glutinosa. None.... Excellent. Excellent. Lespedeza capitata. Poor. Fair. Fair. Fair. Liatris punctata.. . Excellent. Good.... Good.... Good. scariosa. Excellent Excellent. Good. Muhlenbergia pungens 6 . Fair. Fair. Onagra biennis. Fair. Excellent. Excellent. Panicum virgatum. Poor Fair. None. . . . None.... Norte. Petalostemon candidus. Fair. Excellent. Pdnus ponderosa 7 . Good. . . . Good.... Good.... Psoralea tenuifiora 3 . None. . . . Good.... Ratibida columnaris 3 . Good.... Redfieldia flexuosa®. None.... Fair. Rhus glabra. None.... None.... Robinia pseudacacia 2 . Good. . . . Good.... Salvia pitcheri. None. Solidago missouriensis. None.... Fair. serotina. None. . . . None.... rigida. Fair. Sporobolus cryptandrus 3 . None.... Good.... asper. Excellent Good . . , Excellent. Stipa comata 3 . Poor. Good.... Poor. Poor. spartea. None.... Poor. Poor .... None. viridula 3 . Good. . . . None. . . . Fair Excellent. None.... Good. Symphoricarpus occidentalis. vulgaris. None. Ulmus americana. Good.... No crop, frozen Good.... 1 All seeds were collected at Lincoln, unless otherwise indicated. 1 Seed from Pennsylvania. * Seed from Arizona. 1 Seed from Colorado Springs. « Seed from sandhills, Halsey, Nebraska. 4 Seed from Burlington. 1 Seed from Wyoming. METHODS. 23 reserve material to establish itself. The decrease in the number of seedlings that appeared above ground was due in part to this cause, but especially to damping-off and other abnormal conditions arising under control. With the grasses a depth of planting exceeding 0.5 to 1 inch was distinctly detrimental, and most of them, like the composites, did best at 0.12 to 0.25 inch. Un¬ doubtedly, aeration and frequency and amount of precipitation exerted a profound effect upon ecesis in the field. Even under uniform conditions, some seeds, e. g., Elymus canadensis , germinate and grow rapidly, while others are much slower, as w r ould be expected. Methods of Sowing and Planting. Seeds and fruits of plants were sown under three different sets of conditions at the several field stations, namely, surface-seeding, trench, and denuded quadrat. Surface-seeding consisted of selecting areas with a typical cover of vegetation in which wooden stakes bearing the names of the species were driven. Near the stakes and without disturbing the surface, a small quantity of seed was scattered. A little debris was added to simulate conditions during fall and wdnter after the seeds have fallen naturally from the plant. Table 8. —Germination and depth of planting. 1 Species planted Mar. 24. Depth in inches. No. of plants. Apr. 1. Apr. 4. Apr. 7. Apr. 14. Apr. 25. Remarks. Aristida purpurea.... 0.25 8 9 5 4 3 lvs. Apr. 25. 0.5 8 23 23 23 Same size as those at 0.25 inch 1 3 8 0 0 depth. 2 0 0 0 0 Agropyrum glaucum 1 . 0.5 2 16 38 49 49 4 to 5 in. tall with 3 lvs., Apr. 1 4 18 29 25 25 40 • Same size as preceding. 2 2 4 15 14 4 All smaller than preceding. 3 0 0 0 0 0 Amorpha canescens... 0.25 9 11 10 8 9 1 3 7 7 5 5 Smaller than preceding. 2 0 0 0 0 0 3 0 0 0 0 0 Andropogon halli 1 .... 0.25 5 2 4 2 3 2 in. tall, 3 lvs., Apr. 25. 0.5 0 1 1 2 2 Same as above. 1 0 3 2 2 2 Do. 2 0 0 1 3 3 Nearly as large. Andropogon nutans... 0.25 8 7 5 2 2 Tillering. 0.5 0 1 1 1 1 As large as preceding. 1 0 0 0 0 0 2 0 0 0 0 0 Bouteloua racemosa 1 .. 0.12 14 25 27 29 3 Progressively smaller to the 0.5 20 25 17 14 0 maximum depth. 1 1 25 25 32 17 2 0 8 15 8 3 Bouteloua hirsuta 1 .... 0.25 4 10 10 10 11 3 or 4 leaves; 1.5 in. tall. 0.5 4 7 9 7 7 1 1 4 4 0 0 2 0 0 2 0 0 Desmodium canescens. 0.25 2 6 16 18 18 Largest, with 5 leaves. 1 5 11 23 24 24 Smaller than above. 2 0 1 11 9 9 As large as 1 inch. 3 0 0 0 0 0 'Number of seeds not counted. 24 INTRODUCTION Table 8. —Germination and depth of planting —Continued. Species planted Mar. 17. Depth in inches. No. of plants. 1 Mar. 27. Mar. 29. Apr. 1. Apr. 4. Apr. 7. Apr. 14. Apr. 25. Remarks. Elymus cana- 0.12 5 8 15 14 13 13 Poorest. densis. 1 0.5 6 16 16 16 18 18 About same size. 1 0 24 24 24 29 27 2 0 13 26 28 28 22 Smallest. Kuhnia gluti- 0 12 Thick. Thick. Thick. Thick. Thick. Thick. nosa. 0 5 Thick. Thick. Thick. Thick. Thick. Thick. Smaller than above. 1 A few. Thick. Thick. Thick. Thick. Thick. Do. 2 0 0 0 0 0 0 Liatris scariosa. 0.25 25 27 22 18 18 16 15 4 to 5 in. tall. 1 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 Liatris punc- 0.25 1 12 12 13 13 12 11 2 to 3 in. tall. tata. 1 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 Muhlenbergia 0.12 20 20 12 8 8 7 2 tillers; 3.5 in. pungens. 1 tall. 0.5 20 40 50 44 35 35 1 12 16 12 12 10 10 Smaller. 2 6 14 20 20 20 20 Smallest. Onagra bien- 0.12 Very Very Very Very Very Very Very Excessive com- nis. 1 thick. thick. thick. thick. thick. thick. thick. petition. 0.5 13 20 28 35 40 Very Very Much competi- thick. thick. tion. 1 3 3 3 4 4 7 7 Largest rosettes of all. 2 0 0 0 0 0 0 1 Very small. Petalostemon 0.25 6 6 6 6 6 7 7 2 in. tall. candidus. 0.5 1 1 1 1 1 1 1 0.5 in. tall. 1 0 2 2 2 3 3 3 Smaller, very del- icate. 2 0 0 0 0 0 0 0 Ratibida col- 0.25 2 2 3 3 3 2 3 Rosettes with 4 umnaris. to 5 lvs. 0.5 2 2 3 3 3 3 4 Smaller. 1 0 0 0 0 0 0 0 Sporobolus as- 0.12 8 9 28 34 35 31 31 3 to 4 in. 4 lvs. per. 1 0.5 5 9 20 24 22 27 27 Large as above. 1 0 4 14 34 34 5 Pot waterlogged. 2 0 1 2 2 4 2 1 Smaller than above. Stipa viridula.. 0.25 25 35 39 40 38 33 33 4 to 5 in. 3 lvs. 1 20 22 22 24 23 23 23 Do. 2 0 0 3 2 4 3 2 Much smaller. 3 0 0 0 0 0 0 0 1 Number of seeds not counted. This was sufficient to keep the seeds from being blown away. Upon germina¬ tion, seeds sown under such conditions meet with keen competition from the established vegetation for water and nutrients below ground and for light above. The degree to which light intensity is reduced under the grassland cover has never been adequately emphasized, but frequently it is as low as 5 to 10 per cent, even early in the summer. METHODS. 25 The trench method of seeding was used in order to remove competition both below and above ground for longer or shorter periods. Long, narrow trenches, approximately 4 inches wide and deep, were cut in the native sod. The soil from the trench was then thoroughly pulverized and all of the larger roots and rhizomes carefully removed, after which it was again replaced in the trench Enough additional surface-soil was added so that when firmly compacted the level of the trench was the same as that of the surrounding surface This detail is of fundamental importance, since the amount of run-off or the amount of water running into the disturbed area is thus determined. Stakes bearing the species name were then driven in place at about 12-inch intervals and the seeds planted, the soil firmed, and a little loose surface-soil added as a mulch. The rapidity with which rhizomes and neighboring roots invade this newly prepared area varies considerably with the type of vegeta¬ tion and climate, as also the rate at which the grasses arch over the trench and reduce the light intensity In the case of denuded quadrats, the native sod was removed to a depth of 4 or 5 inches over an area about 2.5 feet long and 16 inches wide. The soil was pulverized, the plant parts removed, and the surface brought to such a height as to be even with the surrounding sod after settling. One and sometimes two species were planted in each quadrat thus prepared. Rhizomes of Corylus americana and Symphoricarpus vulgaris and occidentalis were transplanted in 1921. Obviously, root and rhizome invasion from the sides would affect the area in the quadrat much more slowly than that in the trench, and this was also true of shading. Consequently, the quadrat method obviates both root and especially shoot competition for a considerable period, but certain other factors more or less unfavorable to growth are introduced. Since the most critical period in the life of the plant is that of ecesis, the method of seedling transplants was also used. Seeds were germinated and the seedlings grown in an unheated greenhouse in loam in 2.5 to 4 inch flower¬ pots. The larger pots were used for those species which grew most vigorously and whose roots penetrated deepest. Planting was timed in such a manner that the seedlings were well established and 3 or 4 weeks old before trans¬ planting them into the grassland, usually early in May. At this time trenches were prepared in the way already described, except that the subsoil in the trench was well watered before replacing the soil. The trench was half filled with soil, and the seedlings were then transplanted without injury by simply removing the entire contents of the pot as a core, placing this at the proper level in the trench, and firming the soil about it. A sufficient number of plants were available so that any injured in removing the pot were discarded. After watering again, the trench was brought to the proper level and a thin layer of loose soil applied. If the weather required it, the trenches were watered for a 10-day period during establishment. As with the other methods of planting, the plants were thinned from time to time to reduce competition when necessary. All the seedlings were grown at Lincoln and transported to the other stations as needed. Finally, a fourth method was used, which consisted of transplanting mature plants or sods. Blocks of sod about 10 inches square were selected and cut with almost vertical sides to a depth of 8 or 10 inches and removed with the minimum disturbance of the underground parts above the level. The size 26 INTRODUCTION. varied somewhat, depending upon the root-habit and nature of the soil. In transplanting Bulbilis dactyloides, Bouteloua gracilis , Muhlenbergia gracil- lima, and other species with fine, dense surface roots, the blocks were often larger but not so deep. Conversely, transplants from sandy or gravelly soil were necessarily smaller. Many forbs were also transplanted, care being taken to wrap each block securely in burlap for transportation. By this method of sod transplants, species were planted reciprocally among the different stations; for example at Lincoln, Spartina cynosuroides was trans¬ planted from swamp to salt-flat, low prairie, high prairie, and gravel-knoll, care being taken to replant a control block in the swamp for a check. In transferring the blocks of sod, great care was taken to make a hole of the exact size, so that the transplant was in good contact at both the bottom and sides, just enough loose soil being a*dded and firmly tamped around the block to make the contact complete. The sods were placed 2 or 3 feet apart. Ordinarily, root and rhizome growth had more or less obscured the lines of disturbance after the first season, unless there were marked differences in soil color or texture. As far as possible, transplants were made preceding or coincident with the renewed growth and at a time when the holard was distinctly favorable. Moreover, green parts were clipped back, often level with the soil, in order to reduce water-loss. The surplus food materials of the older roots, rhizomes, etc., rendered vigorous growth possible, and in the main this was the most successful method employed, early-blooming species like Koeleria cristata and Stipa spartea alone sometimes failing to seed the first season. The others, such as Agropyrum glaucum and Bulbilis dactyloides, not only fruited, but often also extended their area by vigorous propagation. 2. EXPERIMENTS DURING 1920. HIGH PRAIRIE, LINCOLN. Surface Sowing. • Surface sowings of 6 species were made on the high prairie, April 16. The seed was obtained at Lincoln, as in all cases, except as otherwise stated, Stipa viridula from Colorado Springs being the only exception in this instance. Koeleria cristata and Stipa alone had germinated by May 15, and Andropogon nutans and A. scoparius by June 2, while the seeds of Elymus canadensis and Aristida purpurea (seeds from Burlington) did not germinate. By June 15, Stipa and A. scoparius were mostly dead or dying, and from this time there was a steady decline until August 30, when all had succumbed. It is interesting to note that, although heavy rains occurred immediately after sowing and at two other intervals, only two species had germinated by May 15. Holard determinations show no indication of failing moisture in the surface 6 inches until the middle of June (table 6), but this is no criterion of the alternate periods of wetting and drying undergone by the seeds in the surface inch of soil. The drought during June was marked, no efficient precipitation falling from the 7th until the 25th; in fact, the rainfall was less Fig. 9.—Monthly precipitation at Lincoln, 1920-1922; the monthly mean is shown by heavy cross-bars. than half of the normal (fig. 9). Drought weakened the seedlings, but not to such an extent that they were unable to survive during the favorable growing weather of July and August, and competition for light undoubtedly played the decisive role. By late in May the light intensity at the soil surface (where the 27 28 EXPERIMENTS DURING 1920. prairie was burned the preceding spring) was reduced to 12.5 per cent, with a range of 2.5 to 26 per cent as determined by the stop-watch photometer, and later in the season this was much further decreased. Trench Sowing. The various species sown by this method all germinated, except Koeleria. The June drought resulted in the loss of 6 of the 12 species, nearly all of which were in fine condition early in the month. Five of the six remaining species not only survived the summer, but became permanently established. The failure of the four introduced species is of interest, especially since the holard was quite favorable for growth throughout the season (table 6). The attenu¬ ated condition of certain species noted was due to competition for light; by midsummer (July 25) the light values were only 20 per cent, even at a height of 4 inches above the surface, under a moderate cover. Quadrat Sowing. With three exceptions, all of the species sown on the surface or in the trench were also planted in denuded quadrats. Two plantings were made, one on April 16 and another on May 4 (table 58). Although the June drought made its impress upon the seedlings, all but three species, Koeleria cristata Stipa comata, and S. coronata, survived the summer. In fact, the following season showed that all of these were permanently established, except the California stipas, which were winterkilled. It is interesting to note that grasshoppers, while not attacking the native seedlings, greedily ate the intro¬ duced ones (except Stipa viridula). This alone might constitute a biological barrier. Little difference was noted in the success of the seedlings planted at different dates. At both plantings the soil was quite moist and good rains soon followed. Even a short period of drought after germination is often disastrous, as can be understood from a study of the life-history of the seed¬ lings. Since it is extremely difficult to recover with certainty the delicate Fig. 10.—Plants of Bouteloua gracilis (A), Bouteloua hirsuta (B), and Sporobolus asper (C), 44-days old; scale 1 foot. > HIGH PRAIRIE. 29 roots among the dense tangle of the native sod, plants for this purpose were grown in a similar soil free from vegetation. In early growth these were comparable to those in the quadrats, though differences occurred later. Root Habits of Seedlings. Plants for root study were planted in fertile cultivated soil on April 20. The roots were first examined on June 3 and 4, when the plants were about 44 Fig. 11 .—Bouteloua gracilis 3 months old. Fig. 12 .—Bouteloua hirsuta 3 months old. days old. The plants of Bouteloua gracilis, Sporobolus asper, and Bouteloua hirsuta were 3 or 4 inches tall, each had 3 to 5 leaves, and most of them were beginning to tiller (figs. 10, a, b, c). The position of the seed and the pro¬ nounced branching of the single primary root are clearly evident, and the depth of 7 to 11 inches so early attained is significant. Tillering occurred simultaneously with the development of a secondary root system, a phenome¬ non which is common also to the cultivated cereals (Weaver, Kramer, and 30 EXPERIMENTS DURING 1920. Reed, 1923). The period of tillering, before the new roots have become well established in the moist layers of soil, is a very critical one for the plant, drought at this time often causing great mortality. Where a little erosion occurs, it may be plainly seen that the seedling is literally hanging on to life by a single thread. By July 11 to 14 the root development of Bouteloua gracilis, under the favor¬ able growth conditions and lack of competition, had a height of foliage of about 8 inches, while flower-stalks 12 to 18 inches tall bore spikes of flowers almost in bloom. However, the plants in the quadrats were only 4 to 6 inches tall. The primary root was still plainly distinguishable from the abundant widely MIXED PRAIRIE. 31 spreading or deeply penetrating ones of the secondary system. The working- level was about 20 inches; some roots reached depths of 33 inches (fig. 11). Bouteloua hirsuta was even more deeply rooted than the preceding, although its above-ground parts were scarcely so well developed (fig. 12). The root penetration of Stipa spartea was similar to that of the gramas, though the two or three roots of the primary system alone had penetrated deeply, the larger fleshy branches of the secondary system mostly ending at a depth of about 8 inches. The plants were 8 to 10 inches tall; the parent plants had about 5 leaves and some were well tillered (fig. 13). The development of Agropyrum glaucum is shown in figure 14. Aristida purpurea at this time had formed heavily tillered clumps 6 to 8 inches tall, which were only slightly better developed than those in the prairie quadrats. The roots had a working depth of 20 inches and a maximum penetration of 32 inches. Plants of Elymus with a height of 12 to 15 inches (compared with 5 inches in the quadrats) showed a similar root depth. Stipa viridula, which also grew more vigorously in these plats, with a height of 8 to 10 inches, had a working level of 16 inches, while the longer roots penetrated 33 inches into the moist soil. These data show clearly why a species, after the hazards of the first season are past, has much greater chances of survival. Summary of Planting Results. With respect to results on high prairie, the germination on the surface, in the trench and in the quadrats was 66, 92, and 100 per cent respectively, and the survival, 0, 45, and 80 per cent. Most of the introduced species were badly damaged by grasshoppers, although some in the quadrats survived the season but were winterkilled. Aside from these, all the species that persisted to the end of the first season became permanently established. Andropogon scoparius, A. nutans , and Bouteloua gracilis grew well in both trench and quadrat. Sod Transplants. Sods of the following species were planted on April 23, mostly in triplicate. All showed some drought effects during June, such as rolling of the leaves, but most species grew normally and all survived. Among those from the high prairie, Agropyrum glaucum, Elymus canadensis, Koeleria cristata, and Stipa spartea produced seed, while Andropogon scoparius and Bouteloua racemosa failed to form flower-stalks. Bouteloua gracilis from the gravel-knoll did not blossom, but Distichlis spicata from the salt-flat seeded profusely. Andro¬ pogon nutans was the only species from the low prairie which did not fruit, A. furcatus, Poa pratensis , Panicum virgaium, and Spartina cynosuroides all producing seed. MIXED PRAIRIE, PHILLIPSBURG. Surface Sowing. Seeding was done on May 7 with the same six species as those used at Lincoln. When examined on June 10, three species only had germinated, but these were quite abundant. By July 1, two, Koeleria cristata and Stipa viridula, had died, while Elymus was represented by a few slender narrow¬ leaved plants with dead tips, which succumbed by July 10. Good rains 32 EXPERIMENTS DURING 1920. occurred soon after planting and at two later periods in May. However, an examination of the rainfall data shows that the precipitation for June was much less than half of the normal, as at Lincoln (fig. 15). No rains sufficient even to wet the surface-soil fell after June 18 until the 14th of July, when drought again prevailed until late July (table 6). The attenuated condition of the longer-lived species was due to the competition for light. An average value of 28 per cent (range 2.8 to 80 per cent) was determined for this station on June 11. Fig. 15. —Monthly precipitation at Phillipsburg, 1920-1922; the monthly mean is shown by heavy cross-bars. Fig. 16.—Monthly precipitation at Burlington, 1920-1922; the monthly mean is shown by heavy cross-bars. SHORT-GRASS PLAINS. 33 Trench and Quadrat Sowing. The behavior of the same species, sown in the trench, is summarized in table 9. While all germinated, two died by July 1, another by August 4, and only two, Andropogon scoparius and A. nutans , grew throughout the season and became permanently established. These plants, with others, did much better in the quadrats. While all but Arislida germinated by July 1, several showed a high mortality due to drought, although only two species had died. Conditions became drier during July and only six species were represented by the end of the summer. All of them became permanently established, Bou- teloua gracilis forming flower-stalks the first season. Thus, there was a mortality of 100 per cent among plants where the seed was surface-sown, while 33 per cent survived in the trench, and 67 per cent in the quadrats. Andropogon scoparius and A. nutans did well in both trench and quadrats, but Bouteloua gracilis alone flowered the first year. No sods were transplanted at this station. SHORT-GRASS PLAINS, BURLINGTON. Surface Sowing. The account of the growth of surface-planted species at Burlington is brief. Of the 13 species sown on April 15, evidences of germination were found for only 6 on June 11, and the plants of two of these species were dead. By July 2, three species alone remained, and one of these, Liatris punctata, died before the end of the summer. Bouteloua gracilis merged into and became scarcely distinguishable from the native sod; Stipa viridula tillered heavily and was over 6 inches tall. Precipitation records show that showers (0.25 to 0.63 inch) fell in April after seeding, but this was followed by no efficient rainfall (0.15 inch or more) until the middle of May. Another drought period intervened until the 27th, when 0.25 inch of rain fell, after which drought ensued until June 18, all of this serving to emphasize the unfavorable conditions to which the plants were subjected (fig. 16). The almost constant lack of rainfall throughout the summer was enhanced by the frequent high winds, low humidity, and high evaporation (figs. 7 and 8). Trench and Quadrat Sowing. Thirteen species were sown in the trench on April 15. Evidence that Andropogon nutans and Elymus canadensis had germinated was found, but both species had died and none germinated later. With one or two exceptions the same species that were sown on the surface and in the trench were also planted in denuded quadrats. Their growth throughout the season is shown in table 60. Seven of the 13 species germinated; all but four were dead by July 2, and only three species, Andropogon nutans, Stipa viridula, and Liatris punctata, became established. A summary shows that 33 per cent of the surface-sown species that germi¬ nated survived; all of those in the trench succumbed, while in the denuded quadrats 43 per cent lived through the season and as usual became established. Stipa viridula survived both on the surface and in the quadrats, Bouteloua gracilis on the surface, and Andropogon nutans and Liatris punctata in the quadrats. 34 EXPERIMENTS DURING 1920. Sod Transplants. Blocks of sod of the following species, mostly in quadruplicate, were secured from the several stations at Lincoln and planted at Burlington: Agropyrum glaucum, Andropogon scoparius, A. furcatus, Bouteloua racemosa, Elymus canadensis, Koeleria cristata, Panicum virgatum, Poa pratensis, and Stipa sparlea. Before planting the sods in the holes made in the short-grass mats, these were filled with water, which was allowed to sink into the deeper layers before the block was tamped in place. The grasses were planted in two long rows, each having two blocks of each species, and several feet of unbroken sod thus intervened between the blocks. After planting, all were thoroughly watered, and this was repeated on one row at intervals of 2 or 3 weeks through¬ out the season, as proved necessary. It was found later that the watered row, which had been planted only 2 or 3 feet from a broken area, had received some run-off water from the furrow separating the broken field from the grassland. Although none of the transplants died, the drought affected the imwatered row in a number of ways, while even the watered plants did not attain normal development. Nearly all of the former exhibited a marked rolling of the leaves and even severe wilting at times, and the tips on several species as well as whole plants died back considerably. Elymus, Koeleria, Panicum, and Stipa flowered under both conditions, but the flower-stalks were usually fewer and decidedly shorter in the unwatered row, as was true of the foliage also. Thus, the maximum leaf-height of Panicutn under the two conditions was 13 and 20 inches respectively, while the height of the panicles ranged in the one case from 7 to 13 and in the other from 19 to 20 inches. Stipa, Koeleria , and Elymus seeded in both habitats, but Agropyrum and Andropogon furcatus in neither, while A. scoparius and Boutelaua racemosa each had a single flower- stalk in the watered area. The ultimate fate of these transplants is indicated on pages 68-70. Summary of Experiments. The results of the 1920 experiments at the three stations are summarized in table 9. Table 9. —Summary of experimental seeding, 1920. Method. Per cent of germination. 1 Per cent of germinated species established. Lincoln. Phillips- burg. Burling¬ ton. Lincoln. Phillips- burg. Burling¬ ton. Surface-sowing. 66 50 46 0 0 33 Trench. 92 100 15 45 33 0 Denuded quadrat .... 100 90 54 80 67 43 Average. 86 80 38 42 33 25 1 As indicated elswhere, it is entirely possible that some seeds germinated, but did not appear above ground, or died later and disappeared between visits to the stations, though it seems prob¬ able that few remnants were overlooked. Of the plants surviving, only three, Andropogon nutans, Bouteloua gracilis, and Stipa viridula, succeeded under at least one condition at all three stations. CULTIVATED AREAS. 35 .4 ndrcrpogon scoparius and Bouteloua hirsuta grew at all stations but Burlington, and Liatris punctata at all but Lincoln. Elymus, Bouteloua racemosa , and Aristida purpurea lived throughout the season at Lincoln only. In most cases a somewhat better growth occurred at Lincoln than westward. For example, Andropogon nutans was 7 to 10 inches tall at Lincoln, 6 to 9 inches at Phillips- burg, and 4 to 5 inches at Burlington late in August. Among the transplanted, about 70 per cent flowered at both Lincoln and Burlington, those at the latter station having been watered. Of the unwatered sods at Burlington, 53 per cent had flower-stalks, but some had only one each. "While the transplants at Lincoln developed quite normally, most of those at Burlington were much dwarfed, even when watered. SEEDING IN CULTIVATED AREAS AT ALL STATIONS. True Prairie. The germination and growth of the preceding species in cultivated soil at the several stations is instructive. In every case a good seed-bed was prepared by plowing, disking, harrowing, etc., and a part of the fields used in growing experimental crop-plants (Weaver, 1920,1922) was set aside for native vegeta¬ tion. "Weeds were removed from time to time by hoeing. In these experi¬ ments the trench method alone was used. The upland station at Lincoln was located on a broad, level hilltop of silt loam adjoining the upland prairie. .All but two species, Aristida purpurea and Bouteloua hirsuta, germinated, and all of the others except Koeleria came through the first season successfully. Because of the lack of the root competition which is regularly encountered by plants in the native prairie, these plants suffered less from midsummer drought. The tillering was much heavier and the plants were in all respects more vigorous. For example, Andro¬ pogon nutans was 7 to 10 inches tall in the prairie quadrats and 12 to 17 inches in the field on August 30. Elymus was 4 to 8 inches and 10 to 14 inches tall and Bouteloua gracilis 4 and 9 inches respectively in the two habitats. Like Bouteloua racemosa , both were putting forth flower-stalks in the cultivated field. They all seeded by October, together with Andropogon scoparius, although the flower-stalks of A ndropogon and Elymus were dwarfed. Root Development. The striking growth made by these native species in a single season emphasizes the effect of the removal of competition, as shown especially by the root devel¬ opment, e. g., Andropogon scoparius, with a maximum penetration of 4 feet (plate 3a). Fully mature plants excavated in the prairie in similar soil differed mainly in their greater depth of penetration, about 5.5 feet ("Weaver, 1919:4). Siipa viridula (plate 3b) had a well-developed root system, a few of the longer roots extending to the 33-inch level, but compared with the roots of mature plants in the native habit at the root system had scarcely Fig. 17.—Root of Liatris punctata at end of first season’s gr owth. 36 EXPERIMENTS DURING 1920. begun. At Colorado Springs they had a lateral spread of over 1.5 feet and a working level of nearly 11 feet. The roots of Elymus canadensis under¬ went a remarkable development, as shown in plate 4a, and as with mature specimens, most of the roots were in the first 1.5 feet of soil. However, this year-old plant exceeded in depth (34 inches) any excavated in grassland, and especially in the large number and profuse branching of its roots. Liatris punctata , though only 3 inches tall and with but 2 leaves, possessed a tap-root that penetrated to a depth of 3.3 feet (fig 17). A considerable supply of reserve food had been stored in the enlarged portion of the tap, which had a maximum diameter of 5 mm. Mature plants reached depths of 6 to 15 feet. Both tops and roots of Bouteloua racemosa made exceptional growth; one plant was photographed and excavated on August 25, 1919, only 124 days after the seed was planted in the greenhouse on April 23, the seedlings having been transplanted to the field on May 10 (plate 4b). Mature plants reach¬ ed depths of 5.5 feet. Stipa comata and Andropogon nutans attained working- levels of 18 and 28 inches respectively and maximum depths of 34 and 48 inches. The root habit and extent in both cases were similar to those of mature plants (plate 4c). Mixed Prairie and Short-grass Plains. At Phillipsburg, six species were planted on May 7 in a field adjoining the prairie station. Andropogon nutans, A. scoparius, and Koeleria cristata did not germinate, Aristida purpurea grew until midsummer only, while Elymus canadensis and Stipa viridula both prospered. By July 10 the two tillers on each plant of Elymus were practically as tall as the parent, but, unlike the plants at Lincoln, none produced flower-stalks. At Burlington the plantings were made on the same day as in the grassland, viz, April 15. The following either did not germinate or died soon after germinating: Andropogon nutans, A. scoparius, Aristida purpurea, Bouteloua gracilis, B. racemosa, Koeleria cristata, Liatris punctata, Panicum virgatum, Stipa comata, and S. setigera. Stipa viridula made excellent growth throughout the summer, reaching a height of 8 to 12 inches. A thick stand of Elymus canadensis was obtained, but the plants died during the summer. To summarize, germination of 83, 50, and 18 per cent respectively was obtained at the several stations going westward. Among those which germi¬ nated, Koeleria died at Lincoln and Aristida at Phillipsburg. Elymus was successful at these two stations, but flowered only at Lincoln. Stipa viridula succeeded at all three stations, doing quite as well or even better westward. Seven other species grew at Lincoln but not elsewhere. The holard of the cultivated fields decreased in nearly the same proportion as in the grassland at the several stations. The response of the seedlings to the different climates was in close agreement with that of the seedlings and transplants in the grassland. EXPERIMENTS AT OTHER CLIMATIC STATIONS. Subclimax Prairie, Nebraska City. Character. The subclimax-prairie station was located near Nebraska City, on a gentle east slope somewhat over halfway down one of the large rolling hills char- SUBCLIMAX PRAIRIE. 37 acteristic of the topography of the region. It was situated in an area of native grassland of over 125 acres in extent. The soil is a dark-brown fine silt loam, consisting of glacial materials intermixed with loess. Below 1.5 feet it is quite yellowish in color and shows more plainly its loess origin. The high per¬ centage of clay and silt is indicated by its hygroscopic coefficient, which is about 12 per cent; hence it is very retentive of moisture. As already indi¬ cated, the precipitation at this station is approximately 5 inches greater (33 inches) than at Lincoln. This additional moisture promotes the growth of a more luxuriant grassland vegetation and permits the presence of a large number of herbs and shrubs found rarely or at least much less abundantly in true prairie (plate 7b). 7 6 5 4 3 2 1 0 11 1| ill 4n Nov Dec Fig. 18. —Monthly precipitation at Nebraska City, 1920-1922; the monthly mean is shown by heavy cross-bars. As shown in figure 18, April had about twice the normal amount of precipita¬ tion. Moreover, this was well distributed and probably accounts for the high percentage (80) of germination. However, the rainfall of May was far below normal, while that of June was less than one-third the usual amount, only two rains falling (0.35 inch on the 15th and 0.70 on the 25th). July was likewise a dry month. The dominant grasses are Andropogon nutans, A. Jurcaius, A. scoparius, and Stipa spartea. Koeleria cristata, Poa pratensis, Bouteloua racemosa, Panicum virgatum, and Spartina cynosuroides, which runs far up the ravines, are of lesser importance. The regular occurrence of A. furcatus and Panicum on highland and their excellent development indicate more favorable conditions for growth than in true prairie. The seasonal aspects and herbaceous societies resemble those described for Lincoln (Pound and Clements, 1900; Thornber, 1901). Throughout the grassland, even on the ridges, postclimax com¬ munities of Ceanothus ovatus , Amorpha canescens, and Rosa arkansana occur, being held in check by annual mowing. Symphoricarpus vulgaris and occi- 38 EXPERIMENTS DURING 1920. dentalis, Corylus americana, Cornus asperifolia, Rhus glabra , and Xanthoyxlum americanum invade the grassland from the vantage-ground of fences, hedge¬ rows, and ravines. In unmown and unpastured areas these with others, including many trees, form thickets or groves overrun with Vitis, Rubus, Ampelopsis, Celastrus, Smilax, etc., indicating the approach to climax condi¬ tions. Results. On April 2, ten species were sown on the surface of an area mown the preceding fall. Aristida purpurea and Bouteloua gracilis failed to germi¬ nate. Stipa viridida, S. comata (like Aristida , from seed collected at Colorado Springs), Koeleria cristata , and Liatris punctata died by midsummer, while A. scoparius, A. nutans, Bouteloua hirsuta, and Elymus, after suffering some¬ what severe mortality, came through in good condition and became perma¬ nently established. Table 10. —Holard in excess of hygroscopic coefficient, Nebraska City, 1920. Date. 0.0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 2. 24.2 21.3 18.3 14.2 12.6 June 6. 14.2 16.9 17.9 .... .... June 14. 16.7 10.6 .... .... .... June 21. 8.4 6.7 12.8 .... .... June 30. 8.9 7.9 .... .... .... July 4. 9.8 4.7 7.1 . . . . . . . . July 19. 4.5 5.3 7.3 8.9 9.9 July 28. 2.3 3.4 4.9 . . . . . . . . Aug. 4. 0.3 1.2 3.4 6.1 7.6 Aug. 12. 9.7 2.6 3.1 .... . . , . Aug. 19. 28.5 3.9 2.4 .... . . . . Aug. 28. Hygroscopic 12.4 9.5 5.2 3.3 5.5 coefficient. 12.1 11.7* 12.3 13.7 12.9 The attenuation of the leaves of the seedlings became noticeable early in June. By this time the average level had reached 11 inches, above which Stipa (3 feet), Euphorbia corollata, Silphium integrifolium, and various other herbs occurred. Light values ranged from 2.5 to 40 per cent. The behavior of the species planted in the trench was exceptionally good. All germinated and all but four lived through the season, owing to the rather favorable chresard of early summer, which promoted rapid root growth (table 10). The same species planted in the quadrats did much better. All germinated, became established, and lived throughout the summer. As usual, a fairly high mortality occurred among most species at the critical period, when the primary root system was supplying a maximum of transpiring area before the secondary one became established. As a consequence, the seedlings thrived best or held out longest about the edges of the quadrats. In solving this problem a comparison of conditions within the disturbed area, as compared with that in the undisturbed grassland, is illuminating. r MIXED PRAIRIE. 39 Table 11 .—Comparison of chresard and soil temperatures in denuded quadrats and in adjacent undisturbed grassland. Station. Date. Depth. Grass¬ land. Chresard. Denuded quadrat. Differ¬ ence. Remarks. inches. p. ct. p. ct. p. ct. July 14 0 to 6 22.0 14.0 8.0 Lincoln, high prairie. * ..Do. . . 6 to 12 16.8 12.1 4.7 June 14 0 to 6 6.8 11.3 -4.5 r Do... 0 to 6 10.7 19.9 -9.2 Lincoln, low prairie.. . .Do. . . 6 to 12 13.8 20.2 -6.4 June 22 0 to 3 6.9 -0.6 • • • • ,• Do. . . 3 to 6 • • . . 13.3 • ■ • « June 21 0 to 3 2.0 -4.2 • • * • Nebraska City. < . . Do. . . 3 to 4 • • • • 17.9 , , . , . .Do. . . 0 to 3 • • • • -2.6 • • • • Phillipsburg. J May 19 0 to 3 • • . • 5.8 • • • . . .Do. . . 0 to 6 5.4 12.9 7.5 Temperature. Station. Date. Depth. Remarks. Grass- Denuded Differ- land. quadrat. ence. inches. °C. °C. °C. Lincoln, high prairie. May 5 Under surface 27.6 16.2 11.4 1 p. m., clear, hot. Lincoln, low prairie. . . .. Do. . . . . .Do. 22.5 19.5 3.0 12 noon, clear, hot. Lincoln, gravel-knoll. June 14 3 26.5 21.2 5.3 2 p. m., clear, hot. Lincoln, high prairie. .. . Do. . . Under surface 37.0 41.0 4.0 Do. Lincoln, low prairie. . . . . Do. . . 1 34.0 47.0 13.0 2 p. m., burned area. Do. . . .Do. . . 3 30.0 38.0 8.0 Do. Lincoln, gravel-knoll. . . .Do. . . 1 41.5 46.0 4.5 2 p. m., clear, hot. Do. . . . Do. . . 3 37.5 41.0 3.5 Do. Nebraska City. June 21 Under surface 58.0 38.2 19.8 2 h 30 m p. m., clear, hot. Do. . . .Do. . . 3 37.6 32.4 5.2 Do. It is evident that it is often much drier in the quadrats, the soil hotter, and the humidity much lower. The average daily evaporation is from 15 to 40 per cent higher in the bare area. These conditions are obviously unfavorable for seedlings. The partial shade about the edges of the quadrats lowers the temperature, lessens wind movement, increases the humidity, and undoubt¬ edly accounts for the seedlings growing here when drought kills them else¬ where. Mixed Prairie, Colorado Springs. Character. The station at Colorado Springs was located on a hillside sloping gently southward and in the midst of a great expanse of mixed prairie, about 2 miles 40 EXPERIMENTS DURING 1920. east of the city. As typical for much of the region, the soil consists of a light- colored loam intermixed with some sand. It is very compact and run-off is high; at a depth of 6 to 10 feet it is underlaid with sand. The distribution of the rather meager precipitation (15 inches) throughout the year is very similar to that for other stations. The holard of the surface foot is about 45 per cent and the echard 8.2 per cent. Studies of the water-content throughout a number of years show that it is frequently reduced to the hygroscopic co¬ efficient in the surface layers, and the soil may become very dry to a depth of at least 4 feet. The high evaporation, which is often twice as great as in true prairie, accentuates drought. The extremes of day and night tem¬ perature are similar to those at Burlington. The high day temperatures cause low humidity (frequently 5 to 15 per cent) in the afternoon and cor¬ responding water-loss from both plants and soil. The climate is reflected in the composition of the grassland, both in the absence of many true-prairie species and the relative dwarfing of the others that grow in this habitat. As a result of close grazing before fencing the station, Bouteloua gracilis was quite dominant. Other shorfc-grasses or sedges were Muhlenbergia gracillima and Carex pennsylvanica. Stipa comala, Andro- pogon scoparius, Agropyrum glaucum, Aristida purpurea, Koeleria cristata, Bouteloua racemosa, and Andropogon furcatus all reappeared after cattle were excluded, the latter only sparsely. Artemisia frigida, Aragalus lamberti, Lupi- nus argenteus, Senecio aureus, Abronia fragrans, Chrysopsis villosa, Psoralea tenuiflora, and Argemone platyceras all form extensive societies. Ratibida columnaris, Opuntia camanchica, 0. fragilis, and Thelesperma trifidum were somewhat abundant. Most of the grasses are rooted at a depth of 3 to 5 feet, while many forbs penetrate much more deeply, some to 12 feet. This grass¬ land never reaches the luxuriance of true prairie and consequently light played a ver}^ minor role in the following experiments. Results. The seeds of 9 species were planted in quadrats, but not until June 17. All were from Lincoln except Aristida purpurea and Stipa viridula. Bouteloua gracilis, B. hirsuta, and Koeleria failed to germinate, while Liatris punctata died soon after germination. Andropogon scoparius, A. nutans, Elymus canadensis, and Stipa viridula became established, but like the others were watered from time to time. On June 9 the following sods from Lincoln were transplanted, being abun¬ dantly watered at the time of planting and at several intervals afterward: Agropyrum glaucum, Andropogon furcatus, A. scoparius, Bouteloua racemosa, Bulbilis dactyloides, Elymus canadensis, Koeleria cristata, Panicum virgatum, and Stipa spartea. Although they made a rather poor growth, due partly to the late period of transplanting, all survived the first season. EXPERIMENTS AT EDAPHIC STATIONS. Gravel-Knoll Station. Character. A comprehensive series of edaphic stations was located at Lincoln, all within a distance of a half-mile, ranging from gravel-knoll through high prairie, low prairie, lowland cultivated field, and salt-flat to swamp. GRAVEL-KNOLL. 41 The gravel-knoll station, which was only 300 feet south of the high prairie, occupied the crest of a steep hill about 60 feet above the valley of Salt Creek (plate 5a). The soil is a drift and consists of very porous coarse sandy to gravelly loam with a water-holding capacity of only 40 per cent of its dry weight. At a variable depth, about 4 feet, it is underlaid with clay. The hygroscopic coefficients at the several depths, together with the chresard during 1920, are given in table 12. Table 12 .—Holard in excess of hygroscopic coefficient. Date. 0.0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 21. 13.3 7.3 13.3 May 5. 14.2 5.5 15.1 . . . . . . . . June 9. 4.6 5.1 .... .... .... June 16. -0.2 0.0 4.1 6.6 .... June 23. -1.1 -0.2 1.8 .... .... July 15. -0.9 1.7 2.6 1.1 5.8 July 29. -1.8 -0.6 8.1 . . . . . . . . Aug. 5. -1.9 -1.7 0.5 1.4 7.2 Aug. 12. 3.3 -1.3 0.4 . . , . . . . . Aug. 19. 8.6 2.3 0.3 .... .... Aug. 31. Continued heavy rains; no samples taken. Hygroscopic coefficient. 5.2 4.9 3.1 2.8 7.4 The low water-content is due not alone to the nature of the soil, but also to the steep and generally south slope, which produces a high run-off and greatly increases the temperature. These factors, together with greater wind move¬ ment and higher evaporation than at the high-prairie station, subject both soil and vegetation to frequent drought. The area was dominated by rather open mats of Bouteloua gracilis intermixed with B. hirsuta, which Stipa spartea, Andropogon, and other tail-grasses were invading with difficulty from their foothold in the silt loam farther down the slope. However, the transition from one soil type to the other is quite abrupt and the ecotone correspondingly sharp, a narrow belt of what is essentially mixed-prairie occupying this area. Behavior. Of the seven native species sown in the trench on April 16, Aristida purpurea and Liatris punctata failed to germinate. Elymus canadensis and Koeleria cristata were all dead by July 15, together with Stipa coronata and S. setigera, dominants introduced from California. The single remaining clump of Stipa viridula was disturbed by rodents and died later, leaving only Andropogon nutans and A. scoparius as survivors of the first season. Among these the mortality was very high; the leaf-tips of many plants were dead, while the best plants were only 4 to 6 inches tall. In the quadrats the same species did even more poorly, only five germinating. One of these died by June 2, while all of the others, after a period of leaf-rolling, wilting, and progressive drying of the leaves, succumbed by the middle of July or not long afterwards. Sods of Agropyrum glaucum, Andropogon scoparius, Bouteloua racemosa, Elymus canadensis, Koeleria cristata, and Stipa spartea were transplanted from high prairie to gravel-knoll on April 20. Andropogon furcatus, A. nutans, Boa pratensis, Panicum virgatum, and Spartina cynosuroides were also trans- 42 EXPERIMENTS DURING 1920. planted here from the low prairie. Duplicate and often triplicate blocks of sod were used. One lot of each was watered thoroughly at two or three of the most critical periods during the summer. On June 15, none of the sods having been watered since transplanting, Andropogon furcatus and Panicum were badly wilted and the leaves on Andropogon were rolled, while both Koeleria and Stipa had blossomed earlier than elsewhere and the plants were drying at the base. Two weeks later, no rain having occurred until the 25th, all of the unwatered plants had either ripened seeds and were drying {Koeleria, Poa, Stipa), or had revived as a result of the rains, but were again in a wilted state. During August, one sod of Poa and another of Andropogon scoparius died, while the unwatered Agropyrum, Andropogon furcatus and another A. sco¬ parius were nearly dead. The plants with a rather shallow root habit, viz, Koeleria, Stipa, and Elymus, suffered less, probably because they could com¬ pete more successfully with the gramas for water in the surface soil. However, Bouteloua racemosa and even Spartina survived in fair condition, although dwarfed. In addition to the three early bloomers already mentioned, Elymus and Agropyrum were the only ones that flowered. The ultimate fate of these species is indicated later (p. 83). The low chresard, or its entire absence during certain periods of drought, is shown in table 12. Low-prairie Station. Character. This station was located on a level tract at the foot of the hill about 60 feet below the high-prairie and gravel-knoll stations and only a quarter of a mile southward (plate 5b). The soil is a fertile, dark-colored silt loam of the Wabash series. It is very fine in texture, being composed mostly of silt and clay (tables 13 and 14), and with the Truog test showed no trace of acidity. The greater amount of volatile matter and the greater nitrogen-content at all depths than in soil from the upland indicate more favorable conditions for plant growth. Table 13. —Mechanical analysis of soil from lowland cultivated area. Depth of sample (feet). Coarse gravel. Fine gravel. Coarse sand. Me¬ dium sand. Fine sand. Very fine sand. Silt. Clay. Moisture equivalent. 0.0 to 0.5.... 0.1 0.4 2.2 1.8 5.0 25.0 41.3 24.3 27.7 0.5 to 1.0 ft. . 0.3 0.7 2.1 2.2 5.0 25.4 38.8 25.8 27.9 1 to 2. 0.2 0.3 1.3 1.5 3.7 21.4 40.8 31.0 30.6 2 to 3. 0.0 0.1 0.4 0.5 1.7 19.2 43.4 34.7 32.9 Table 14. —Chemical analysis of soil from lowland cultivated area. Depth of sample (feet). Insoluble residue. Soluble salts. Volatile matter. Iron and aluminium oxids. Calcium oxid. Magne¬ sium oxid. Phospho¬ rus pent- oxid. Nitrogen. 0.0 to 0.5.. 79.34 12.96 7.70 9.57 0.68 0.75 0.13 0.218 0.5 to 1. . . 79.63 13.66 6.71 10.27 0.63 0.77 0.10 0.187 1 to 2. 78.11 15.83 6.06 12.11 0.64 1.01 0.08 0.135 2 to 3. 74.78 19.82 5.40 15.20 0.76 1.27 0.09 0.082 LOW PRAIRIE. 43 Moreover, the chresard is constantly greater than on the high prairie, though the area is well-drained and the soil is never waterlogged. Table 15. —Holard in excess of hygroscopic coefficient. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 10. 28.7 22.3 21.8 18.7 18.8 Apr. 21. 26.1 • • • • 22.1 • • • • • • • • May 5. 33.1 25.6 23.9 • • • • a a • • June 9. 25.7 24.0 • • • • • • • • June 16. 13.6 16.6 19.5 26.4 June 23. 9.3 11.5 20.4 • • • • July 14. 27.4 21.1 17.0 19.5 22.7 July 28. 17.6 12.5 15.7 • • • • a a a a Aug. 5. 2.3 8.2 11.5 12.0 23.6 Aug. 12. 15.4 12.4 11.8 • • • • a a a a Aug. 19. 21.8 8.8 10.1 a • • ■ a a a a Aug. 31. Continued heavy rains; no samples taken. Hygroscopic coefficient. 11.8 11.1 10.3 10.2 11.9 The low-prairie area is dominated by a few species less xerophytic than those occurring on the high prairie. Many of the species of the high prairie are absent, but are replaced in part by others of a more mesophytic kind. The dominant grasses are Andropogon furcatus, Panicum virgatum, Andropogon nutans , and Spartina cynosuroides, each of which often covers small areas with a pure or nearly pure growth. Poa pratensis is also very important, but is overtopped by the taller grasses, which reach a height of 5 or 6 feet (plate 6a). Characteristic forbs are Solidago canadensis , S. rigida, Glycyrrhiza lepidota, Aster multiflorus, A. salicifolius, Physatis heterophylla, Polygonum muhlenbergi, Artemisia gnaphalodes, Achillea millefolium, and Callirrhoe alcaeoides. The rank growth forms a dense cover and makes ecesis in the area very difficult. Bisects show that the plants are rooted deeply, most of them reaching depths of 5 to 12 feet. Root-layers at 3 and 5 feet are quite as distinct as those on the high prairie. Behavior. Surface sowings of eight species were made on April 10. Two failed to germinate, while Bouteloua gracilis germinated but soon died. By June 15 all were very delicate and much attenuated, owing to the low light intensity. A month later Liatris punctata had been shaded out and the others were slender and pale in color. Stipa viridula and Andropogon scoparius died later, leaving attenuated individuals of A. nutans and Koeleria surviving the first season. Koeleria had disappeared by the following spring, but A. nutans made a good growth and merged into the native sod. The light intensity under the cover of tail-grasses was so reduced as to constitute the major factor in competition, in view of the high holard. On June 14, 1922, in a portion of the prairie burned over the preceding spring, the light intensity under vegetation of average density and with an average height of 12 inches, was only 6 per cent. Some seedlings were growing in even deeper shade. By July 25 the average had dropped to 3.5 per cent; under a thin cover it was 12 per cent, but where the vegetation was very dense, only 1.5. 44 EXPERIMENTS DURING 1920. Seeds planted in the trench at the same time did somewhat better. Although three species failed of germination, all of the others, viz, Andropogon nutans, Bouteloua hirsuta, Elymus canadensis , and Stipa viridula, came through the season, although somewhat blanched and attenuated. Thus, A. nutans had reached a height of 7 to 8 inches as compared with a growth of 5 or 6 inches on the high prairie. Light in the trench in June varied from 5 to 10 per cent near the edge to 50 or 60 per cent in the center. By late July this was reduced to 4.5 to 16 per cent, even at 4 inches above the surface. Of the eleven species sown in denuded quadrats, all germinated, and all but one, Stipa coronata , came through the season in excellent condition. In fact, except for Stipa setigera, which was winterkilled, all became established. The marked development shows the possibilities of growth in this habitat when sufficient light is available. Little difference was apparent between the earlier and later sowings. A comparison of growth in the denuded quadrats of the prairie and in the adjoining cultivated area is of interest. Both were seeded on the same dates, weeds being kept out of both areas. The native vegetation in the field was consequently free from any competition. Aristida purpurea, Stipa comata, S. eminens, and S. setigera failed to germinate, as did also the sowing of Andropogon nutans, A. scoparius, and Koeleria cristata made on April 9. All others made an excellent growth, even exceeding that in the quadrats. For example, the height in inches attained by August 30 by typical representatives in field and grassland respectively were: Stipa viridula, 20 and 5 to 7; Koeleria cristata, 3 to 4 and 2 to 3; Elymus canadensis, 14 to 16 and 7 to 11. Moreover, the species in the cultivated area were more heavily tillered and three, Andropogon nutans, A. scoparius, and Elymus, came into blossom in Sep¬ tember. The root systems were well developed; the roots of Elymus reached a maximum depth of 4 feet and those of Andropogon scoparius were abundant to 3 feet, while some extended to 4.5 feet. The roots appeared finer than on older plants and were not so well-branched near the tips. The ample water- content of the fertile soil in the cultivated area, which was similar to that on low prairie, together with the lack of competition, accounts for the excellent growth. Blocks of sods of the same species used on the gravel-knoll (except Andro¬ pogon nutans) and secured from the same sources were transplanted into low prairie from March 22 to April 24. Bulbilis dactyloides from overgrazed low prairie, Distichlis from salt-flat, and Bouteloua gracilis from gravel-knoll were also included. The prairie had been mown the preceding year. During the first season all of the plants flourished, including the short-grasses, Bulbilis and Bouteloua gracilis. The former produced stolons abundantly, but none were able to become rooted. Because of light conditions, the height growth of the shorter grasses was somewhat emphasized, while that of the taller species was quite normal for lowland. All the transplants flowered the first season, except Spartina. The effect of a reserve food-supply in the roots and rhizomes is shown by comparing a single season’s growth of Andropogon nutans from a block of sod in the low prairie and from seed sown in the lower culti¬ vated plats (plate 8a). Salt-flat Station. Character. This station is located on a level tract of land just below and only 200 feet south of the low prairie. Owing to a small percentage of sodium chloride which SALT-FLAT AND SWAMP. 45 has deflocculated the clay (table 16), it is of a very different consistency from that of the low prairie (Hall, 1920: 283). Moreover, the adsorption of the Na ion in part has left the soil in a slightly acid condition, according to Truog's test. Table 16 .—Acidity and salt-content at salt-flat station. Depth of sample. Acidity. 1 Total soluble salts. 2 Chlorides. p. ct. p. p. m. p. ct. p. p. m. 0.0 to 0.5 foot. Slight. 0.17 1,700 0.06 600 0.5 to 1.0 foot. Very slight . 1.14 11,400 .18 1,800 1 to 2 feet.... ... Do. .46 4,600 .10 1,000 2 to 3 feet.... . . .Do. .50 5*000 .12 1*200 3 to 4 feet.... .. .Do. .64 6,400 .15 1,500 1 Neutral to litmus in water-extract. 2 No carbonates were present. The above conditions, as the holard is very similar to that in the low prairie, have greatly affected the type of vegetation. This consists of an open growth of salt-grass, Distichlis spicata, which is frequently mixed with a sparse growth of Agropyrum pseudorepens (plate 6b). In places occur scat¬ tered plants of Poa pratensis; other species, especially of forbs, are practically excluded. Agropyrum is much dwarfed, and with Distichlis forms a layer seldom exceeding 5 to 8 inches in height. The sparse flower-stalks are only 8 or 10 inches high, and consequently light plays a minor role in deciding the fate of transplants. Behavior. The same species as in the case of the low prairie and from the same sources as before, were transplanted into the salt-flats, with the exception of Bulbilis. Andropogon furcatus, A. scoparius, and Panicum virgatum had been transplanted during the spring of 1919. The first season they made a poor growth and none put forth flower-stalks; from the outset in 1920, Stipa appeared yellow and dwarfed. All of the others suffered more or less severely during the summer drought. Late in June nearly all showed more or less wilting and some yellowing. In July the ground was cracked on one or more sides of the blocks of sods, a phenomenon due to local soil texture and one which did not occur elsewhere, even in the dry soils at Burlington. However, none of the transplants died. Compared with their growth in the adjoining low pi airie, nearly all were greatly dwarfed and lacked vigor and but 6 of the 13 species blossomed. These were Agropyrum, Distichlis, Bouteloua gracilis, Elymus, Koeleria, and Stipa. The flower-stalks were usually short and the inflorescence smaller than normal. Swamp Station. Character. The swamp station is located in the valley at the foot of a hill 0.25 mile north of the high-prairie station. During 1920 water stood above the soil- 46 EXPERIMENTS DURING 1920. level throughout April, May, and most of June. For the remainder of the season the soil was saturated at nearly all times to near the surface. The vege¬ tation consisted chiefly of a rank growth of Spartina cynosuroides, with Scirpus atrovirens and Heleocharis palustris. Phalaris arundinacea and relict Typha latifolia, with several species of Juncus and Carex , played a minor role. Adjoining the wettest area and making the transition to the grassland was a zone of almost pure Poa pratensis. Sods were transplanted into both areas. In the latter the water stood above the surface during April and May, while * throughout much of the remainder of the season the soil was saturated or nearly so almost to the surface. Vegetation in both areas grew rank, but aeration was obviously a factor of equal or greater importance than light. Behavior. Sods of the same species used on the low prairie were placed in the wettest portion of the swamp and in the adjoining Poa zone on April 24. Water stood on the surface of the swamp until after June 15. By June 2, Andropogon scoparius and Bouteloua gracilis had died, and the leaf-tips were dead on A. furcatus, Elymus, and Stipa, while Koeleria , which was heading out at 4 to 8 inches high, had most of its leaves dead. Spartina and Panicum were thriving. By the middle of June the whole area was badly overgrown. Heleocharis had reached a height of 20 and Scirpus atrovirens 30 inches, while both Poa and Spartina overtopped all the transplants. Light values ranged from 2 to 10 per cent. Bouteloua racemosa had died and Stipa was nearly dead. The basal leaves on Agropyrum and Elymus had died, on the latter to a height of 9 inches, and Distichlis was much attenuated. The flower-stalks of Koeleria had rotted off at the base and only a few green leaves remained. Koeleria, Stipa, A. furcatus, and A. nutans completed the mortality list for the first season. Spartina had flower-stalks 5 feet tall. A few poor, slender plants of Agropyrum, Distichlis, Poa, and better ones of Panicum survived. The stems of Elymus rotted off near the ground, but new shoots came out from the base. Panicum did not seed. The other survivors, except Spartina, had gone through a period of anthesis earlier in the season. Aside from a deficit of oxygen, low light intensities must be considered in accounting for these results. During July and August a dense growth of Spar¬ tina, Scirpus, Phalaris, etc., to a height of 4 feet or more so completely over¬ shadowed the transplants that they were found with difficulty, and light values ranged between 1.8 and 7.5 per cent. Transplants in the Poa zone grew under slightly less unfavorable conditions in regard to both deficient aeration and reduced light. However, Andropogon scoparius and A. nutans both died in May and Koeleria and Stipa in July. Some of the other species had dead basal leaves and were more or less attenuated. All of the survivors except A. scoparius, A. furcatus, Bouteloua gracilis, and B. racemosa produced seed. Sods of Stipa setigera and S. eminens from California were transplanted to both high and low prairie on April 27. They were watered from time to time as necessary. During May and early June all showed slight to fair growth, but by July 15 three had died at each station and the others were growing poorly. All had died on the high prairie by August 9, while only one of each species survived on the lowland. The leaves on these did not exceed 2 or 3 inches in SUMMARY. 47 height. They remained alive until fall, when they were winterkilled. Sods of several species from Arizona were transplanted into high prairie on June 2 and were also watered from time to time. By June 15 some plants showed a new growth, but a month later half of them were dead and the others were doing poorly. These died during August, with the exception of a single Bou- teloua, which produced a flower-stalk 20 inches high; however, this plant was winterkilled. Summary. Experiments. A summary of the planting experiments is given in table 17. Omitting the data from Colorado Springs, where the plants were watered, the per¬ centage of germination was greatest at Nebraska City and in the low prairie at Lincoln, as was also that of establishment, if the cultivated area is omitted. Establishment was lowest, 15 per cent, on the gravel-knoll. Of the species which frequently failed to germinate, Aristida failed at 6 places of planting, Bouteloua hirsuta, B. gracilis, Koeleria , and Liatris at 2 or 3 each, and Andro- pogon scoparius at 1 . Andropogon nutans grew in 9 of the 10 places sown; A. scoparius, Elymus, and B. hirsuta in 6, while Koeleria and Stipa viridula grew in 4 of 5. In general, the percentage of germination, as well as that of establishment, increased from surface to trench to quadrat, and inversely with competition for water and light. Table 17 .—Summary of planting experiments, 1920. P. ct. of species germinating. P. ct. of germinated species established. r Method of planting. Gravel- knoll. Low prairie. Cultivated land. Nebraska City. Colorado Springs. Gravel- knoll. Low prairie. Cultivated land. Nebraska City. Colorado Springs. Surface. 75 80 33 • 50 Trench. 71 57 67 100 30 100 100 60 Denuded quadrat.. 71 100 100 56 0 91 100 80 Average. 71 77 67 93 56 15 75 100 70 80 Including with these data those from the high prairie at Lincoln, Phillips- burg, and Burlington, and arranging the stations in order of the average per¬ centage of germination under the different methods of sowing, the sequence is that shown in column 1 of table 18. Column 2 gives the arrangement based on the percentage of the germinated species that became established, the low-prairie station ranking ahead of Nebraska City. The difference between the percentage of germination in high and low prairie is much less than the difference in that of establishment. The use of the most favorable method of planting, i. e., the denuded quadrat, gives practically the same sequence of stations. 48 EXPERIMENTS DURING 1920. Table 18 .—Comparison of germination and growth at all stations } 1920. Average per cent of germination. Average per cent of establishment. Per cent of establish¬ ment in denuded quadrats. Nebraska City. 93 High prairie, Lincoln.... 86 Phillipsburg. 80 Low prairie, Lincoln.... 77 Gravel-knoll, Lincoln.... 71 Burlington. 38 Low prairie. 75 Nebraska City. 70 High prairie. 42 Phillipsburg. 33 Burlington. 25 Gravel-knoll. 15 Nebraska City.100 Low prairie. 91 High prairie. 80 Phillipsburg. 67 Burlington. 43 Gravel-knoll. 0 In accounting for the above results, it remains only to compare the habitats at Nebraska City, low prairie, and gravel-knoll a little more fully. Those of the high prairie (Lincoln), Phillipsburg, and Burlington have already been shown to decrease in all conditions favorable for plant growth in the order mentioned. Physical Factors. A comparison of the holard at the gravel-knoll and Burlington stations shows that it is not greatly different, both very frequently falling to the hygroscopic coefficient (tables 12 and 6). In the first 6 inches especially it is often reduced by surface evaporation to even a lower point. The erosion of the soil about the seedlings was often pronounced on the gravel-knoll and in this respect growth conditions were less favorable than at Burlington. The cover of short- grasses on both areas indicates xerophytic conditions, the vegetation often drying and turning brown late in July or in August. The chresard was con¬ tinuously higher on low than on high prairie, a difference of 5 or 10 per cent in favor of the former not being unusual. The low prairie was also better supplied with water than the Nebraska City station, though the hygroscopic coefficient w r as at no time reached in either habitat, a margin of 3 to 10 per cent or more at all depths usually being present. The average daily evapora¬ tion was also slightly less in the low prairie, Nebraska City being intermediate between the low and high prairie stations at Lincoln (fig. 19). Fig. 19.—Average daily evaporation, high prairie (solid line), and low prairie, Lincoln (short broken lines), and Ne¬ braska City (long broken lines), 1920. However, a comparison of the average weekly humidity at these three stations shows this to be much higher at Nebraska City (68 to 82 per cent) than on the low prairie (57 to 79 per cent), the high prairie being lowest of all (58 to 73 per cent). The average weekly humidity at the three stations was i EXPERIMENTS. 49 73.S. 71.9. and 65 .4 respectively from June 3 to August 26. This factor seems suffi cient to account for Nebraska City exceeding the low--prairie station both in per cent of germination and of establishment in denuded quadrats. In summary, two sod transplants died from drought on the gravel-knoll All showed repeated rotting and dying back of the leaves, including those Altered from time to time, and the hve which blossomed did so earlier than elsewhere. Drying at the base was often quite as pronounced as at Burlington. AH of the transplants nourished on the low prairie and all but Spcrtina blos¬ somed. On the salt-fats the transplants suffered more or less severely during drought, as shown by the wilting and yellowing of leaves. None died, but all were considerably dwarfed. The dower-stalks were short on the species that blossomed and the inflorescence smaller than normal. The sods in the swamp suffered, from deficient aeration especially and also from deficient light. Five high-prairie species and two from the lew prairie died. All but Spcmbu and Paver--. .. umr..- . shewed the erects of deficient aeration, the basal leaves of c* several clams dying and the rower-stalks rotting off rear the base. Sir of the 13 produced seed, some or much abbreviated rower-stalks. Sods in the Pckl core cd the swamp fid somewhat better, only three high-prairie and one low- prairie steedes succumbing two other high-prairie species failed to dower. N earlv ah sh;wed the erects of deficient aeration and shade, but to a less degree than those in the swamp. The transplants at Colorado Springs were made late in the season -June 9 and watered at intervals: ah survived, but nine made a good growth. Exseeimests Dueixg 1918 no 1919. *> <~v Of 19 IS Feme preliminary experiments were conducted during 1918 ani 1919. and an account of the results seems desirable for the sake of completeness. On April 6. seeds of the following species were sown in 4-inch pots in rich loam soil in the greenhouse: .irrm^ rkucmn. Aodr?poeeri /urcatu.?. A. rerun?. A. sc&parius. Aristide oLi.;nr.tha . Baiticloua h:mru . and 8yer>:o. ae osper. When the roc-ts had reached a depth of 3 to 5 inches ani the plums were wed established, the latter were transferred to trenches S inches wide and 6 inches deep. This was done on April 26 in both gravd-knoll and high prairie. 4 to 14 pets cf each species being transplanted at each station. They were wmered from time to time until June 1. when they were f.rally weeded and given no further attention until late in September. On October 25 to 27 the plants were counted, the roots excavated, and measurements taken. Tne summer was exceedingly dry and only one plant, eye uve . survived on tne gravei- knoiL In the high prairie all of the An skim died, ad but one of Amcymivu, and over half of fkrceiovu and ep-sr;eol.-?. while only tne trree species of Ardmpceon came through with small mortality. Fci.reTu.u t ecu had reached an average height of 6.5 inches and most of the plums had seeded. The bulk of the root system plate was in tne drst S inches of soil, although seme roots penetrated considerably deeper. A frycc T■verne attained a height of 5 or 6 Irenes and the coarse roots were traced well below the 18-inch level: A. scope:rim was c to 10 inches tall; the central plant in plate 9 a shews the major portion of me root system, tne roots being much utter and mere abundant than in the preceding species. 50 EXPERIMENTS DURING 1920. A. nutans grew to a height of 7 or 8 inches (plate 9a). The largest of the three survivors of Sporobolus asper reached a height of 10 inches, the plants were in seed, and the roots penetrated more than 15 inches, the lateral spread reaching distances of 8 to 10.5 inches. The difficulty of extracting the entire root system in grassland has already been mentioned. Because of competition the shoot growth was less than in cultivated soil, and it seems probable that the root system was also shortened. These root systems, which were only partially recovered, afford a striking contrast to those of similar species of a single year’s growth in cultivated soil (p. 35 and 36). Season of 1919. During 1919, Andropogon furcatus was sown in quadrats on May 10 and again on June 15 in high prairie, low prairie, gravel-knoll, and in the lower cultivated plats. Those on the high prairie which were watered on May 28 at the same time as the others, did very well, and little difference was noted be¬ tween the two plantings. In June 1920, both quadrats contained plants 6 to 8 inches tall, which reached 7 to 10 inches by fall, but no flower-stalks appeared. A good sod, almost merging into the native prairie, was developed during 1921, but no species flowered. By June 1922, some invasion had occurred, and later in the season, as well as in 1923, flower-stalks were produced, although rather sparsely. Those on the low prairie made a better growth the first season, but, because of the plowing of the area, were transplanted to salt-flat and the new low-prairie station the following spring. Both lots continued to thrive (pp. 44 and 45). Plantings of this same species on the gravel-knoll died the first season or the following winter. Those made in the lower cultivated plats did best of all. The roots of the plants were excavated on August 26, when they had a working depth of 2 feet and a maximum depth of over 3 feet (plate 9b). A few stalks flowered and set seed. The vegetative growth by August of the second season was 2 feet high. Flower-stalks, some of which were 44 inches tall, occurred in abundance and good seed was developed. When about 3 weeks old, duplicate seedlings of Aristida purpurea, Stipa viridula, and Elymus canadensis were transplanted in the usual way on May 5, into high and low prairie, gravel-knoll, and upland and lowland cultivated areas. Bouteloua racemosa was also included in the last two sta¬ tions. All of those on the gravel-knoll died before the end of August, after repeated wilting and drying throughout the summer. Those on the low prairie grew well until fall, but were plowed up the following spring. On the high prairie all lived through the first season, but did not survive the winter. Aristida and Stipa both died on the cultivated upland, while Bouteloua and Elymus made an excellent growth and were in full bloom August 1. By August 15, Bouteloua had reached a height of 23 inches (plate 4b). Clumps of Elymus 1.5 to 2 inches in diameter at the base were common, and as many as 66 stalks were counted in a single clump. The dense root system (plate 4a) reached a depth of 2 feet and a maximum lateral spread of 22 inches. All the plants grew very well in the lower cultivated plots, except Aristida, wdiich did rather poorly. Bouteloua came into full bloom late in August, after making an excellent vegetative growth, and it matured seed abundantly. Elymus also blossomed, and by August 9 of the second season had flower- SOD TRANSPLANTS. 51 stalks 58 inches tall with very large spikes. Stipa viridula made an excellent growth, but did not seed until the next year. By August 9, 1920, the seed was ripening on flower-stalks over 5 feet tall and in spikes 14 inches in length. The root system at this time had reached a depth of over 5 feet and had a lateral spread of 18 to 20 inches near the surface. Roots were very abundant to 2 feet. In fact, the fine growth and great size attained by native species under favorable conditions emphasize the controlling effects of competition in the native grassland. Sod Transplants. Gravel-knoll. The following sods were transplanted to the gravel-knoll during the spring of 1919: Agropyrum glaucum, Andropogon furcatus, Bulbilis dactyloides, Ely?nus canadensis, Koeleria cristata, and Panicum virgatum. They were watered a few times until established. Agropyrum, Bulbilis, Elymus, and Koeleria all blossomed somewhat earlier than elsewhere, the flower-stalks being somewhat dwarfed. During July all dried out badly, A. furcatus and Koeleria dying by the end of the season. Four sandhill dominants, Andropogon halli, Calamovilfa longifolia, Muhlenbergia pungens, and Sporobolus cryptan- drus, were secured from the extra-regional sandhills near Central City, Nebraska, and transplanted into the high prairie on June 9. Great difficulty was encountered in holding the sand in place about the root systems. The new shoot growth was clipped back as usual to lessen transpiration, but in spite of repeated watering, none survived. Table 19 .—Holard in excess of hygroscopic coefficient on the gravel-knoll, 1919. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 28. 24.0 25.3 20.5 19.0 12.0 May 10. 7.3 5.6 7.1 7.9 1.1 May 27. 1.0 1.7 3.6 • • • • « • • • June 16. 7.0 6.6 .... • • « • .... July 1. 0.2 0.1 • • • • .... . . . . July 8. -1.1 -0.4 2.4 3.5 • • • • July 14. -1.7 -0.1 • • • • .... .... July 28. 1.6 1.7 6.7 11.4 7.1 Aug. 26. -0.7 6.6 7.2 6.7 7.1 Sept. 6. Hygroscopic -0.9 -0.5 0.4 4.1 8.3 coefficient. 5.2 4.9 3.1 2.8 7.4 An examination of the rainfall records shows that while May and June had ample rainfall, July was extremely dry. The total rainfall was only 0.38 inch (normal 3.83 inches), 0.30 inch of which fell on the last day of the month. The rainfall for August was also an inch below normal. The low holard of the gravelly soil in which the plants were growing is shown in table 19. On July 14 the native grama grasses ( Bouteloua gracilis and B. hirsuta) began to dry badly and two weeks later all were brown and cured on the ground. During 1920, which was a more favorable season, Agropyrum produced a thin crop of flower-stalks 34 inches high, while Bulbilis flowered profusely and put forth numerous stolons 8 to 12 inches long, some of which rooted in the 52 EXPERIMENTS DURING 1920. grama sod. Only 2 stalks of Elymus about a foot tall were left; Panicum did poorly, reaching a height of but 12 to 14 inches. Undoubtedly by this time most of these species were rooted rather deeply in the clay subsoil at about 4 feet depth. In 1921, Agropyrum again blossomed, but not profusely, while Bulbilis seemed quite at home, blooming luxuriantly and extending its territory nearly a foot into bare areas. One of the two stems of Elymus put forth a flower-stalk at 18 inches height; Panicum remained dwarfed (13 inches) and did not blossom. During 1922 all made an excellent growth, exeept Elymus, which was winterkilled. As usual, Agropyrum blossomed early; Bulbilis had spread until it occupied nearly twice its original area, and Panicum flowered at 18 to 24 inches. All dried earlier in the late summer than elsewhere, owing to the severe drought in August. The next year was an excellent one for growth, and Agropyrum spread over 30 inches into the adjacent grassland, flowering profusely at 2.5 to 3 feet. Bulbilis spread even more widely than before, but the stolons rooted with difficulty in the heavy growth of grasses; it also flowered profusely. Panicum likewise increased its area slightly, but blossomed at a height of only 27 inches. Low and High Prairie. The same species, except for Panicum and Andropogon furcatus, but in¬ cluding A . scoparius and Stipa spartea, were transplanted to the low prairie during April and early May 1919. All flourished and bloomed profusely except Stipa. No further data were obtained, since this part of the prairie was broken the following spring. A similar lot of sods, with the addition of Bouteloua gracilis from the gravel-knoll and Panicum, Poa, Andropogon Fig. 20.—Average daily evaporation on high prairie (upper line) and low prairie (lower line), 1920. furcatus, and Spartina from the low prairie, were transplanted to high prairie. Because of the severe drought during July, the transplants wilted somewhat and both the production and size of flower-stalks and inflorescence were reduced. The chresard on both high and low prairie is given in table 20. The high evaporation rate (daily average 30 to 50 c. c.) on the high prairie during the drought period, and the relatively lower one on low prairie (15 to 28 c. c.), together with the evaporation throughout the season, are compared SOD TRANSPLANTS. 53 in figure 20. Greater wind movement at the upper station was an important factor in causing the difference in evaporation, although humidity played a large share also. From July 7 to August 18 the average daily evaporation was 64 per cent greater at the upland station. Table 20 .—Holard in excess of hygroscopic coefficient on high and low prairie at Lincoln , 1919. High prairie. Low prairie. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 28. May 5. 18.5 19.8 22.2 18.2 19.1 15.6 13.8 9.8 11.4 14.2 24.4 26.1 26.5 17.3 16.1 17.5 16.1 19.5 May 27. 8.2 13.7 11.2 .... .... 8.8 17.8 16.4 .... .... June 16. 16.6 16.6 .... .... .... 24.5 18.6 .... .... July 1. 11.5 10.0 9.5 9.9 10.0 11.7 11.1 17.1 17.1 18.5 July 8. 2.3 5.3 9.2 10.7 9.2 5.4 5.7 16.8 17.1 20.6 July 14. 0.0 3.5 5.2 .... .... 4.4 5.6 9.6 .... .... July 28. -0.6 1.2 1.6 7.1 11.1 0.1 1.9 5.1 13.3 18.5 Aug. 11. 4.4 2.0 2.0 3.3 7.0 6.8 3.6 5.9 8.1 18.1 Aug. 26. 10.4 3.2 2.2 5.4 5.4 17.2 6.1 7.3 8.9 17.9 Sept. 6. 1.2 3.3 3.4 5.6 8.3 -0.2 4.4 2.6 9.3 18.9 Sept. 26. 14.5 15.0 5.2 2.6 5.7 17.2 15.2 5.8 10.0 16.2 Hygroscopic coef- ficient. 9.5 8.7 8.6 7.1 6.2 11.8 11.1 10.3 10.2 11.9 June .23 30 7 July 14 21 28 4 August 11 18 25 Sept 1 8 Fig. 21.—Average day and night temperature at Lincoln (solid lines), and Colorado Springs (broken lines), during 1919. Thermograph records of soil temperature at the two stations at a depth of 6 inches show no marked differences. The daily range at the lowland station was usually 4° to 7° F.; in the drier soil of the upland station 4° to 12° F. The minimum soil temperatures were 2° or 3° F. lower at the former station and the maximum 3° to 9° F. higher on the upland. The soil temperatures varied from 53° F. (average 60° to 65° F.) late in May, to 85° F. (average 70° to 78° F.) in August. Except in early spring, when the low temperatures on the lower 54 EXPERIMENTS DURING 1920. wetter soil may retard growth, and the higher temperatures at all depths on the upland may facilitate absorption, it is not probable that the effects of tempera¬ ture on the development of vegetation at the two stations are very different. A comparison of the two sets of thermograph records shows that differences of air-temperature greater than 5° F. are rare at the two stations. It seems cer¬ tain that such small differences in temperature variations in the range of growth conditions (minimum 65° F. late in May, maximum 105° F. in July and August) would be almost negligible in the development of natural grassland. The average day and night temperatures throughout the season, together with those at the Colorado Springs station, are compared in figure 21 and the average daily evaporation rates at the same station in figure 22. Notwith¬ standing the adverse conditions, all the transplants, including Spartina , bloomed more or less profusely, with the exception of Andropogon furcalus, Koeleria, and Stipa. Fig. 22. —Average daily evaporation at Colorado Springs (upper line), and Lincoln (lower line), during 1919. During 1920, all of the transplants did quite well, except two of the Stipas which were not well developed. Bulbilis made an excellent growth, blossoming profusely by the middle of June (plate 10a), and both Panicum and Spartina flowered abundantly at heights of 30 to 44 inches. In fact, Andropogon scoparius and A. furcatus were the only species that did not blossom, but they made a vegetative growth of 8 to 12 inches. By the following season, Andro¬ pogon furcatus had almost merged into the common sod. Koeleria was nearly all dead; two of the Stipas were behaving normally, while the other two did not blossom and were in very poor condition. Elymus was repre¬ sented by only one or two stalks, but all of the other species were developing quite normally. On August 23, when Panicum virgatum was blooming profusely with flower- stalks 3 to 4 feet tall, a trench was dug to a depth of 11 feet and the roots of this 3-year-old transplant examined. Below 3 feet the silt-loam soil became more sandy and below 6 feet it changed to a sandy gravel with many small rocks. It was quite moist at all depths. The roots penetrated to a maximum depth of 9.8 feet, and were fairly abundant to 7 feet. The general root habit and branching were about normal (Weaver, 1920), but the degree and extent of branching were considerably more pronounced than in the low prairie. SOD TRANSPLANTS. 55 However, it did not approximate the remarkable network of branches dis¬ played by this species in the drier soils at Burlington (fig. 32). Spartina cynosuroides was also excavated from the same trench at this time. It had made a good growth, notwithstanding the rolling of leaves from time to time as a result of drought. Many leaves reached a height of 40 inches. It had increased its area about one-third by rhizome propagation, but the root¬ stocks had not developed many new shoots. Roots were abundant to 9 or 10 feet and were densely branched throughout, perhaps having a third more branches than those in the lowland, and branches were longer and rebranched to a greater degree than those growing in wetter soil. Koeleria and Poa died during the winter of 1921-22. The following summer all the rest, except Andropogon scoparius, made a good growth, and flowered rather profusely, except the very late bloomers, which were somewhat retarded by the late summer drought. Panicum extended its territory nearly 3 feet, but few of the others gained in area. During 1923 all did very well, owing to good growth conditions, Andropogon furcatus merging completely with the native sod. All flowered and set seed, the height of the inflorescence being about normal. Mixed Prairie. On June 19, 1919, the following sods from Lincoln were transplanted to the mixed-prairie station at Colorado Springs: Agropyrum glaucum, Andro¬ pogon furcatus, Bulbilis dactyloides, Elymus canadensis (2 sods), Koeleria cris- tata (3 sods), Panicum virgatum, Poa pratensis, and Stipa spartea (3 sods). Before planting, the hollow made for each sod was filled with water that was allowed to sink away before the sod was planted, tamped in place, and again watered. Owing to watering at 2-week intervals, all did fairly well the first season. In June 1920, cattle broke in and closely grazed the entire fenced area. However, most of the sods lived throughout the Season, but made a poor growth, Bidbilis and Agropyrum alone coming into blossom. One Elymus and one Koeleria died late in the summer. During 1921, perhaps as a result of repeated grazing, the remaining Elymus and the two remain¬ ing koeierias died, as did also one of the stipas; Bulbilis alone came into blos¬ som. The area was grazed again in 1922. This repeated grazing, combined with drought, resulted in a very poor growth, Poa succumbing and Bulbilis alone blossoming. Because of further grazing in 1923, it was clearly evident that most of the plants (except Bulbilis) were rapidly losing ground, Agropyrum, Panicum, and Andropogon furcatus being represented by remnants only. 3. EXPERIMENTS DURING 1921. PHYSICAL FACTORS. Rainfall. The season of 1921 at Lincoln was quite favorable as to precipitation. Rainfall for April, June, and July was slightly above normal, while August was slightly below normal, as was May also (fig. 9). An examination of the records shows that the showers were exceptionally well distributed, no marked drought periods occurring, and at no time did the vegetation become dry or brown, even on the gravel-knoll. At Phillipsburg an excess of 0.14, 0.98, and 1.30 inches fell during April, May, and June respectively. The July precipitation was slightly above normal and that of August and September slightly below (fig. 15). A drought period occurred between June 9 and 28, during which time no efficient mois¬ ture fell. At Burlington the precipitation for April was nearly twice normal (and 0.86 inch more than at Lincoln). May fell 0.91 inch below normal; June was somewhat above (0.58 inch); July had less than half the usual amount (only 1.15 inches); August had an excess of 1.19 inches, but September a deficit of more than one-half the normal (fig. 16). No efficient rainfall oc¬ curred (i. e., showers of over 0.15 inch) between March 7 and April 15, April 17 and May 19, June 22 and July 4, and July 8 and August 7. By July 24 Fig. 23.—Average daily evaporation at Burlington (short broken lines), Phil¬ lipsburg (long broken lines), and Lincoln (solid line), 1921. J the short-grasses were about half-dried and brown, while two weeks later not only the grama and buffalo grasses were curled and brown, but the wheat- grass heads were dry as well. Water Relations. The above conditions were reflected in the holard at the several stations (table 22). At Lincoln, a margin of 5 per cent or more (usually 8 or 10) 56 PHYSICAL FACTORS. 57 existed at all times and at all depths to 4 feet. At Phillipsburg the echard was approached once in July and twice in August, actually reaching the danger-mark to a depth of 4 feet late in that month. Conditions at Burling¬ ton were, as usual, much worse. At no time was water available in the third Table 21. — Humidity at Lincoln and Burlington , 1921. For week ending— Average day. Average night. Average daily. Lincoln. Burlington. Lincoln. Burlington. Lincoln. Burlington. Apr. 21. 53.9 64.7 59.3 Apr. 28. 53.7 • • • • 81.0 • • • • 67.3 .... May 5. 57.4 37.0 86.4 71.5 71.9 54.2 May 12. 57.6 • • • • 73.5 .... 65.5 • • • • May 19. 56.1 .... 73.9 .... 65.0 • • • • May 26. 43.7 , , . . 64.2 .... 53.9 • • • • June 2. 57.2 39.4 79.4 78.7 67.9 59.0 June 9. 63.2 .... 84.4 .... 73.8 • • • • June 16. 61.4 50.4 81.7 87.0 71.5 68.7 June 23. 61.7 .... 80.2 .... 70.9 • • • • June 30. 60.4 30.0 81.7 72.1 71.0 51.0 J uly 7. 66.0 50.9 83.6 81.0 73.5 65.8 July 14. 58.2 .... 84.3 , . . . 71.2 .... July 21. 57.7 42.5 80.4 75.5 69.0 59.0 July 28. 63.0 40.0 78.4 78.3 70.7 59.1 Aug. 4. 66.7 38.4 85.1 72.6 75.9 55.5 Aug. 11. 62.0 42.8 81.9 79.8 71.8 61.3 Aug. 18. 65.3 49.3 83.4 83.6 74.3 66.4 Aug. 25. 58.0 .... 83.4 .... 70.7 .... Sept. 1. 62.0 .... 86.4 .... 74.2 .... or fourth foot, while after June 30 it was depleted repeatedly above the hard- pan (at about 2 feet). A complete record of humidity was not obtained, but a comparison of the average daily humidity at Lincoln and Burlington at the several intervals (table 21) show T s that without exception the air during the day at Burlington was 10 and often 30 per cent less humid than at Lincoln. Fig. 24. —Average daily temperature at Lincoln (solid line), Phillipsburg (long broken lines), and Bur¬ lington (short broken lines), 1921. 58 EXPERIMENTS DURING 1921. Owing to the lower night temperatures at Burlington (altitude 4,160 feet), the humidity at this less critical period averaged only slightly lower than at the true-prairie station. Table 22. —Holard in excess of the hygroscopic coefficient at the several stations, 1921. Lincoln, Nebraska. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 9. 20.2 17.9 13.7 13.3 12.5 May 18. 18.0 19.7 16.7 15.0 14.5 May 25. 6.2 13.2 • • • • .... • • • • June 1. 7.9 10.3 .... .... • • • • June 7. 23.0 18.6 14.9 14.5 17.2 June 22. 8.8 14.1 15.1 .... .... June 29. 7.0 8.8 11.8 15.2 ■ • • • July 13. 13.4 20.1 10.4 .... .... July 20. 21.0 18.5 12.7 13.1 16.8 July 28. 7.5 7.3 . . , . • • • • • • • • Aug. 3. 10.2 11.0 11.3 .... .... Aug. 10. 16.4 17.9 12.0 12.1 26.8 Aug. 23. 10.8 12.9 8.5 11.4 8.8 Aug. 31. 8.5 7.0 5.1 .... • • • • Hygro. coeff. 9.5 8.7 8.6 7.1 6.2 Phillipsburg, Kansas. Apr. 28. 15.8 14.5 11.4 5.9 2.3 May 19. 16.8 13.0 9.8 6.4 0.7 May 25. 12.2 10.9 .... .... .... June 9. 19.8 15.8 13.9 14.5 13.0 June 22. 5.3 8.6 10.8 11.7 9.3 July 19. 1.6 2.2 • • • • .... .... July 21. 8.2 3.3 3.9 4.6 • • • • July 25. 1.6 2.0 0.9 2.7 4.9 Aug. 4. 3.5 4.4 .... • • • • . • • • Aug. 9. 0.1 0.8 0.6 2.1 2.1 Aug. 17. 20.9 7.5 1.5 1.3 .... Aug. 30. -1.6 -0.4 -0.6 0.0 0.3 Hygro. coeff. 10.6 10.6 10.9 10.6 10.7 Burlington, Colorado. Apr. 30. 9.6 10.7 8.6 0.0 1.3 May 20. 6.4 5.6 4.7 -0.7 -2.2 June 10. 10.3 0.0 0.7 -0.5 -0.6 June 30. -2.6 -1.6 — 1.5 -1.0 -1.3 July 17. -1.0 -1.0 -3.6 .... .... July 24. -3.5 -2.2 -3.4 -1.8 -0.4 Aug. 8. 2.0 -0.7 -2.1 -0.3 -1.0 Aug. 17. 16.9 11.7 -2.4 -2.0 .... Aug. 29. -0.3 0.5 -2.1 -0.4 -1.0 Hygro. coeff. 10.9 10.9 12.2 12.0 11.4 The average daily evaporation at the several stations is shown in figure 23. The rather uniform rate at Lincoln (daily average 8 to 27 c. c.), as compared with the erratic and excessively high one at Burlington (18 to 62 c. c.), empha- PHYSICAL FACTORS. 59 sizes the precarious conditions for plant growth at the latter station. Con¬ ditions at Phillipsburg were intermediate (8 to 43 c. c.). Temperature. . The average daily air-temperature was in general lowest at Burlington, while that at Phillipsburg after the third week in June was somewhat higher than at Lincoln (fig. 24). During May and June, day temperatures were highest at Lincoln and lowest at Burlington, but after July 1 this was usually reversed. Night temperatures throughout the season were distinctly Fig. 25.—Average day (heavy lines) and night temperatures (light lines), at Lincoln (solid lines), Phillipsburg (long broken lines), and Burlington (short broken lines), 1921. lower at Burlington by 5° to 10° F., those at the other stations being very similar. The soil-temperature at a depth of 3 inches was invariably lower at Burlington than at Phillipsburg. Later in the season soil-temperature was correlated with the holard, the soils becoming progressively drier and warmer. A series of readings in cropped areas (oat field) to a depth of 4 feet is instructive in this connection (table 23). Table 23. —Soil temperatures , 1921. Depth in feet. April 28- 30. May 19-21. June 9-10. June 22. June 30. Lin¬ coln. Phil¬ lips¬ burg. Bur¬ ling ton. Lin¬ coln. Phil¬ lips¬ burg. Bur¬ ling¬ ton. Lin¬ coln. Phil¬ lips¬ burg. Bur¬ ling¬ ton. Lin¬ coln. Phil¬ lips¬ burg. Bur¬ ling¬ ton. °C. °C. °C. °C. °C. °C. °C. °C. °C. °C. °C. °C. 0 to 0.5. . . 15.4 24.0 12.0 21.0 20.0 22.2 22.5 21.4 19.8 21.2 23.2 29.2 0 5 to 1. . . 13.0 23.1 11.0 19.0 17.8 15.0 21.8 20.2 18.8 21.1 22.4 26.1 1 to 2. 12.0 15.0 10.5 13.4 16.4 14.0 20.2 18.8 16.8 20.1 21.5 23.9 2 to 3. 12.0 14.0 10.5 13.1 15.0 12.4 18.6 17.2 15.8 18.8 21.2 21.8 3 to 4. 10.8 12.0 10.0 12.0 13.5 11.8 17.8 16.0 15.0 17.4 20.5 20.0 60 EXPERIMENTS DURING 1921. Isolated readings at similar depths showed the temperature of the grassland soil (below the first foot) to be very similar to that in the cultivated areas. To summarize; as during 1920, conditions for plant growth were much more favorable at Lincoln and least favorable at Burlington, while those at Phil- lipsburg were intermediate, holard and humidity being the controlling factors. PLANTING RESULTS. Surface Sowing. Seeds of 14 species were sown on the surface of the high prairie at Lincoln on April 20. Aristida purpurea, Elymus canadensis, and Lespedeza capitata failed to germinate. Bouteloua gracilis, Koeleria cristata, Liatris punctata, and Sporobolus asper germinated rather abundantly, but died before the end of June, the seedlings being much attenuated. The light values were low; by May 5 they fallen to 16 to 27 per cent, as the prairie had not been mown the previous autumn. There was not only considerable dead grass, but the new growth was rank and flower-stalk production greater than usual. Before the end of July, Bouteloua hirsuta, Bromus inermis (which was badly eaten by grasshoppers), Liatris scariosa, and Pinus ponderosa were added to the mortality list. Robinia pseudacacia held out until August, as did also Stipa viridula. Andropogon nutans alone survived the season, making a fair growth 6 to 9 inches high and becoming indistinguishable from the native sod. At Phillipsburg the surface plantings made on April 28 did very much better, 83 per cent living throughout the summer. After planting, showers were opportune at both stations, all 12 species germinating at Phillipsburg. Unquestionably the light relation was much more favorable at the latter station. Reduced light and surface drought were the chief factors to which plants gaining a foothold in climax grassland were subjected. The same species as those used at Phillipsburg, with the addition of Sporobolus, were sown on the surface at Burlington, April 30. None had germinated by May 20, and only 5 of the 13 germinated at all. These were Bouteloua racemosa, B. gracilis, Bromus inermis, and the two species of Liatris. All were found in greater or less abundance on June 10, but had disappeared by the 30th. The fate of surface-sown plants of the preceding year is of interest. At Lincoln and Phillipsburg none survived the first season. By May of the second year at Burlington, Bouteloua gracilis had merged into the native sod. Stipa viridula, the other survivor, made a good growth (8 to 10 inches) and tillered heavily, standing out quite plainly in the short-grass turf. Its leaves were tightly rolled and somewhat wilted from time to time, and it did not produce flower-stalks. During 1922 it grew vigorously, but was much smaller than similar plants in denuded quadrats. By the end of the season it was clear that competition with the native short-grasses was too severe, Stipa rapidly losing ground. However, it held out through the very favorable season of 1923, being represented by only a few culms 6 to 8 inches high, while the leaves of grama were 6 inches tall all around it. It was clearly evident that it would soon lose in the unequal struggle. Trench Sowing. Although poor at all stations, the results from trench planting April 20 to 30 were best at Lincoln. Of 12 species used, all but Pinus ponderosa germi- TRENCH AND DENUDED QUADRATS. 61 nated; Elymus, Koeleria, Bromus, and Stipa viridula succumbed before the end of June, the last two being badly eaten back by grasshoppers. The slender plants of Lespedeza and Liatris punctata, the latter reaching a height of 5 or 6 inches, died in August. Andropogon nutans, B. hirsula, and Sporobolus came through in excellent shape, tillering heavily and forming a rather dense sod from 2 to 6 inches tall. Bouteloua gracilis was represented by remnants only, and Liatris scariosa by a dozen one-leaved slender plants. In the trench at Phillipsburg, Bouteloua racemosa, B. gracilis, B. hirsuta, Koeleria, Lespedeza, Liatris scariosa and punctata, and Pinus failed to ger¬ minate, while Bromus and Elymus died in June and Stipa viridula in July. A single plant of Robinia pseudacacia with 5 leaves and a height of 4 inches lived through the summer but was winterkilled. The only species that became permanently established were Andropogon nutans and Aristida purpurea; the former was represented in late summer by 4 or 5 good clumps, 3 tolO inches tall. Of the same species sown in the trench at Burlington, Bromus and Elymus alone germinated, and very sparsely, but both died during the June drought. The development of the plants grown in the 1920 trench at Lincoln was above the average. During 1921, Andropogon nutans, A. scoparius, Bouteloua hirsuta, and B. racemosa all made excellent growth. The boutelouas reached heights of 3 to 8 inches, and the andropogons of 5 to 10 inches, the latter forming a fairly dense sod. Bouteloua gracilis was represented by a single remnant, which was winterkilled. During 1922 the four remaining species made a fair growth, the boutelouas alone flowering and producing a single flower-stalk each. These grew well in 1923, though Andropogon scoparius and Bouteloua racemosa were quite sparse. The height growth ranged from 4 to 14 inches, but none had developed flower-stalks late in August. At Phillipsburg, Andropogon scoparius and A. nutans alone survived the first season. During 1921 they flourished, forming dense bunches and reaching heights of 14 and 24 inches respectively, but they were favored by some water running into the trench. A. nutans developed few flower-stalks. Both made an excellent growth in 1922, extending the area of the bunches, but Andro¬ pogon scoparius alone flowered. In 1923 they again grew rank, reaching heights of 2 feet, and both flowered abundantly. At Burlington there were no survivors of trench sowing. Denuded Quadrats. Twenty-eight species were planted in the quadrats at Lincoln (high prairie) on April 14. Eight failed to germinate. Elymus and Stipa viridula died in June, Agropyrum and Bromus inermis in July, and Acer negundo, A sac - charinum, and Calamovilfa in August; of these, Bromus and Stipa were badly damaged by grasshoppers; 13 species survived. Corylus americana and Symphoricarpus spp., all of which survived, were propagated at the several stations from rhizomes obtained from Weeping Water, Peru, and Lincoln, Nebraska rather than from seed. Among the 28 species planted in quadrats at Phillipsburg, 9 failed to germinate. Among these were Koeleria cristata, Liatris scariosa, Muhlenbergia pungens, and Onagra biennis, none of which germinated at Lincoln and all but Muhlenbergia failed in the quadrats at Burlington. Calamovilfa, Sporobolus asper, and S. cryptandrus died in June, Corylus and Pinus in July, while Stipa 62 EXPERIMENTS DURING 1921. viridula succumbed in August. Except for Calamovilfa, the losses here and at Lincoln were of different species. Thirteen species survived. Of 25 species planted on April 30 at Burlington, only 6 did not germinate. Ten species disappeared in June, another in July, and 4 others in August, Andropogon nutans, Calamovilfa longijolia, Gleditsia triacanthus, and Stipa viridula being the only ones to survive the season. A comparison of the growth of Gleditsia and Andropogon, the only species that lived at all three stations, is instructive. By the end of the summer the former was 4 to Fig. 26.—Average daily soil-temperature at depths of 3 and 18 inches respectively at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1921. 8 inches tall at Lincoln with 12 to 13 leaves 5 to 10 inches high, with the same number of leaves at Phillipsburg, and only 2 to 3 inches high with 4 to 6 leaves at Burlington. Andropogon had formed a dense sod 4 or 5 inches high at Lincoln, one 12 to 18 inches high at Phillipsburg, while at Burlington the scattered growth was only 2 to 7 inches in height. Of the survivors in the 1920 quadrats at Lincoln, Stipa setigera and S. eminens were winterkilled (table 58). Andropogon nutans and A. scoparius flourished, forming good sods 10 to 13 inches tall, the former having flower- stalks at 2 feet. Stipa viridula and Aristida purpurea developed good bunches 6 or 7 inches in height. Bouteloua gracilis and Elymus blossomed, the Boute- loua tufts being 4 to 6 inches tall and those of Elymus 12 to 15 inches. All made a good growth during 1922, but Andropogon scoparius and Bouteloua gracilis alone came into blossom, probably owing to the severe drought in late summer. None died during the very dry winter following (table 58). At Burlington, only 3 species, Andropogon nutans, Lialris punctata, and Stipa viridula, survived the 1920 season in the denuded quadrats. The vicissitudes of these plants are of interest (table 60); wilting and rolling of the leaves, dying back, and discoloration were of usual occurrence. However, they were well rooted and all survived the summer of 1921, notwithstanding DENUDED QUADRATS. 63 the ravages of grasshoppers and drought. In 1922, all but Stipa viridula blossomed. The flower-stalks of Andropogon nutans were 1 to 2.5 feet high late in August, while Liairis flowered profusely. Stipa exhibited many fine clumps 12 to 18 inches tall in 1923. As in the other quadrats, its area had been scarcely invaded, but it did not blossom. All the species prospered and seemed permanently established, at least as long as competition with the short-grasses did not occur (table 60). Root Development in Denuded Quadrats. Fig. 27 .—Liatris punctata from Phillipsburg, on July 1 of second year after planting. From July 1 to 3 several species from the 1920 quadrats at Phillipsburg were excavated and the root development studied. Liatris punctata had 5 leaves and a height of 5.5 inches. The portion of the root which began 2 or 3 inches below the surface was 5 mm. in diameter for a distance of about an inch, beyond which it tapered rapidly and descended more or less vertically downward to a depth of nearly 4 feet (fig. 27). Nine branches came off from the tap in the first 8 inches of soil, none of which exceeded 17 inches in length. Like the tap-root, they were fairly well supplied with short branches. Bouteloua hirsuta formed fine, well-tillered clumps 6 to 9 inches tall. The root system was characterized by its wide-spreading habit in the surface soil. The delicate, thread-like, exceedingly well-branched roots reached a distance of 16 to 18 inches on all sides of the plant, the longest ones often lying in the surface inch or two of mellow soil (fig. 28). The deeper roots, which ran Fig. 28. —Bouteloua hirsuta from Phillipsburg, on July 1 of second year after planting. 64 EXPERIMENTS DURING 1921. obliquely downward, extended to depths of 12 and a few to 24 inches. Many new thick, w r hite roots only 2 to 4 inches long came off from the new tillers. Bouteloua gracilis had well-developed clumps 4 to 7 inches high. The fine fibrous root system was very similar to that of the preceding, especially in the number and extent of fine lateral branches. However, it differed in having a smaller lateral spread (about 12 inches) and greater depth of pene¬ tration. Many of the roots reached 2 and some nearly 3 feet in depth. Andropogon scoparius produced densely rooted plants 8 to 12 inches tall and so w T ell tillered that they formed a continuous sod. The dense network of roots and branches reached a working level of 30 inches and a maximum 29 30 31 Fig. 29 .—Andropogon scoparius from Phillipsburg, on July 1 of second year after planting. Fig. 30.—Roots of Andropogon nutans after the block of sod had been transplanted for 2.5 months at Burlington. Fig. 31.—Roots of Elymus canadensis at end of June of second year after the block of sod had been transplanted and watered at Burlington. depth of 37 inches. Many new roots of the current season’s growth were only 6 to 12 inches long. The last 3 to 8 inches of these, as well as those of many of the rapidly growing ones of 1920, were 1 to 1.5 mm. thick, glistening white, and destitute of branches (fig. 29). Stipa viridula was represented by bunches 4 to 9 inches in height. The roots were abundant to a depth of 2 feet, some extending to the 4-foot level. They had a lateral spread of 12 to 14 inches on all sides of the plant. These data are sufficient to show the fairly deep root penetration of the species by midsummer of the second year in denuded quadrats. It also explains how they persist so tenaciously during periods of drought after they have had a year’s growth. The removal of the preceding species left Andropogon nutans alone in the 1920 quadrats. It grew very well, reaching a height of 15 to 18 inches by fall and producing a few flower-stalks 40 inches tall. During 1922 it made ex- SEEDLING TRANSPLANTS. 65 cellent growth but did not blossom, while in 1923 it reached a height of 2 feet and flowered abundantly. Seedling Transplants. Seedlings of 13 species were transplanted into high prairie on May 11, when 3 or 4 weeks old (table 61). As at all stations, these seedlings were thoroughly watered when planted and during a period of 10 days thereafter whenever necessary. Notwithstanding this care, two weeks later two of the four lots of Andropogon nutans, Aristida purpurea, Bouteloua racemosa, and Koeleria cristata had died. Agropyrum glaucum, Elymus canadensis, Stipa comata, and S. viridula were eaten back by grasshoppers and many individuals had died. One lot of Bouteloua gracilis had also died, and Andropogon fur- catus, Stipa spartea, and Liatris punctata were the only ones which had not suffered a high mortality. The soil at this station is of such a type that it becomes quite compact following heavy rains or watering. This is detri¬ mental to seedlings, as well as germinating seeds, since the surface crust hinders aeration and offers more or less of a mechanical barrier to tender growing parts. When dry it cracks and in this way injures the plants by mechanical tearing of roots, rhizomes, etc., near the surface. At Phillips- burg and Burlington the soil is of such texture that it is mellow at all times. However, all species were represented until the middle of June, after which Stipa comata and S. viridula died, and Agropyrum, Aristida, Elymus, and Liatris also succumbed during July. The other 7 species came through the season in fair to excellent condition, although much reduced in number. Bouteloua hirsuta flowered. A similar lot of seedlings of the same age was transplanted at Phillipsburg on May 19. These were in groups varying from 2 to 6 per species (table 62). By June 9 all were flourishing and most of them continued to do so throughout the month. By July 25, Agropyrum, Elymus, and Stipa spartea had suffered heavy losses; Koeleria, Stipa comata, and S. viridula were nearly all dead, and by the end of the summer they had disappeared. Agropyrum, Andro¬ pogon furcatus, Bouteloua gracilis, Elymus, and Liatris scariosa were nearly all dead or dying, the last having been eaten to the ground. However, A. nutans, B. hirsuta, Liatris punctata, and Stipa spartea were in fair to excellent condition. Seedlings were transplanted at Burlington on May 20 (table 63). By June 10, Koeleria had died, while Agropyrum, Liatris scariosa, and Stipa viridula had suffered a high mortality. However, the other species were in fairly good condition. By the end of the month, Liatris scariosa and Stipa viridula had succumbed and most of the other species had lost heavily. Many had their leaves rolled or were badly wilted. By the end of July, A. nutans, B. hirsuta, and B. gracilis alone were barely alive and these died during August. Summary. The sequence of germination was the same as in preceding years (table 24), i. e., Lincoln being highest (81 per cent), Phillipsburg second (68 per cent), and Burlington last (43 per cent). On the basis of establishment, however, Phillipsburg, unlike the preceding year, ranked first with 60 per cent, Lincoln second (40 per cent), and Burlington third (7 per cent) The 66 EXPERIMENTS DURING 1921. unusually favorable rainfall at Phillipsburg during spring and early summer has already been pointed out, as well as the relatively favorable light con¬ ditions for surface-sown plants (p. 60) . On the basis of survival of seedlings, Phillipsburg also ranked first, none surviving at Burlington. As to the surface-sown plants of 1920, none were left at either Lincoln or Phillipsburg for growth during 1921, but Stipa Table 24 .—Summary of planting experiments at the three stations, 1921. Method of planting. P. ct. of species germinating. P. ct. of germination established. Lincoln. Phillips¬ burg. Burling¬ ton. Lincoln. Phillips¬ burg. Burling¬ ton. Surface sowing. 79 100 38 9 83 0 Trench. 92 43 15 45 50 0 Denuded quadrat.... 71 61 76 65 63 21 Average. 81 68 43 40 65 7 Seedlings. 54 75 0 viridula survived at Burlington. Among the 5 survivors of the 1920 trench planting at Lincoln, none died during 1921; 1 was winterkilled in 1921-22, but the others were growing fairly well at the end of the summer of 1923. At Phillipsburg only 2 of the 6 species planted lived, but both of these grew well, flowered from time to time, and persisted to the end of 1923. None survived the 1920 summer at Burlington. Among the 8 survivors in the 1920 quadrats at Lincoln, all lived throughout the next summer, but 2 were winterkilled. The 6 remaining species did well until the fall of 1923. At Phillipsburg all but 1 were excavated for root study. This plant made an excellent growth and flowered from time to time until the end of 1923. The 3 species at Burlington lived not only through the 1921 season, but also during 1922 and 1923, all but Stipa viridula coming into blossom. Sod Transplants. Between March 20 and 31, sods of 16 species were planted on high prairie (table 64). Distichlis was secured from the salt-flats, Bulbilis from overgrazed low prairie, Bouteloua gracilis and B. kirsuta from gravel-knoll, Andropogon furcatus, A. nutans, Panicum virgalum, and Poa pralensis from low prairie, and Spartina cynosuroides from the swamp. The rest were high-prairie species. All made a good growth during the summer, except Distichlis, which was nearly all dead by August 31, and all blossomed except the three Andro- pogons, Bouteloua kirsuta, B. gracilis, Koeleria, and Spartina. During 1922 all not only survived, but Bulbilis, Panicum, and Spartina increased their area. However, by the end of August, Agropyrum, Distichlis, and Elymus were doing very poorly, owing in part to the severe drought. Twelve of the 16 species blossomed (table 64). Koeleria died during the following winter. The rest survived the following summer, but by August Bouteloua gracilis, Agropyrum, Distichlis, and both lots of Elymus were in straits. Bulbilis, Stipa, and Panicum alone produced flower-stalks before the final check was SOD TRANSPLANTS. 67 made in August. This lack of vigor was probably a result of the extremely dry late summer, fall, and winter preceding (p. 109). The 1920 sods made fair to good growth, except for one block of Andropogon furcatus and two of Distichlis, which were nearly dead by the last of August. The following did not blossom: Agropyrum, A. furcatus, Bouteloua gracilis, B. racemosa, Distichlis, Spdrtina, and Stipa. During 1922 a severe drought occurred in late summer and both plantings of Distichlis and Elymus and one of Koeleria succumbed. Agropryum, one Andropogon scoparius, and Poa pratensis were in very poor condition at the end of the summer, while Andro¬ pogon nutans, A. scoparius, Poa, and Stipa alone had seeded. The following winter one Stipa died, but all of the others made a fair to good growth during this fourth summer. On August 25 only the following had blossomed or showed signs of developing flower-stalks: Poa, Bouteloua gracilis, B. racemosa, Koeleria, and Panicum. Sods of the same species as those used at Lincoln during 1921, except Bouteloua gracilis, Bulbilis, Distichlis, and Spartina, were secured from the various stations at Lincoln and transplanted into the short-grass sod at Bur¬ lington on April 15. As in 1920, they were placed in duplicate rows, one being thoroughly watered 5 times during the season (table 65). By May 20 the effects of drought were apparent. The leaves on many species were rolled and the leaf-tips dead. Others had been frayed by whipping in the wind, a very common phenomenon in this region among both native and crop plants. Koeleria and Stipa were almost dead. The plants in the watered row were in better condition. On June 29, the root development was examined in two species, which had been growing for a period of 75 days. Andropogon nutans was in fairly good shape and had made a growth of 4 to 6 inches. The block of sod was found to be in close contact with the soil on all sides. The old roots had practically all died, though in a few instances laterals from them had continued growth. Many new white roots had originated from the rhizomes and ended at 6 to 12 inches depth. They were evidently growing rapidly, since 3 to 6 inches of the tips were unbranched; nearer their origin they were thickly beset with short branches. Several longer roots penetrated from 18 to 32 inches, but none spread much laterally, and all were well branched (fig. 30). The soil was moister under the block of sod than elsewhere to a depth of 20 inches. Panicum virgatum was growing nicely and had reached a height of 10 inches. As before, the old roots were nearly all dead, but many new ones descended vertically and were fairly abundant to 24 to 30 inches. At the last depth they came in contact with the dry hard-pan. They were densely beset with laterals to near the tip. By June 29, Stipa had died in the unwatered row and Elymus and Koeleria were nearly dead. Even some of the species in the watered row were doing poorly, especially Poa and the three andropogons. By late July Andropogon nutans and Poa had died in the watered row and several other species were in bad condition. In the unwatered row, Elymus and Poa had succumbed,while Andropogon furcatus, Bouteloua hirsuta, and Koeleria died soon after. Autumn found only 7 of the 11 species alive in the watered row, and these were repre¬ sented by mere remnants of the original fine blocks of sod. Of the survivors, Agropyrum, A. scoparius, Bouteloua hirsuta, Elymus, and Stipa were high- 68 EXPERIMENTS DURING 1921. prairie species, and A. furcatus was from the lowland. Agropyrum alone had blossomed. Of the 11 duplicate unwatered species (2 having been dug up) only Agropyrum, Bouteloua racemosa, and Andropogon scoparius remained, and Agropyrum alone had a single flower-stalk. During 1922, both Agropyrum and Bouteloua racemosa in the unwatered row survived, although both suffered severely from drought, and Andropogon scoparius was winterkilled. Agropyrum put forth flower-stalks mostly less than a foot high, and the spikes were very much dwarfed. In the lot that were formerly watered, Andropogon scoparius died in July. The other 6 species made a fair to good growth, often showing the effects of drought in the rolled leaves or dead leaf-tips, as well as by the dwarfing of the whole plant. Bouteloua racemosa, B. hirsuta, Agropyrum, and Elymus alone blos¬ somed. Andropogon furcatus, Bouteloua hirsuta, and Bouteloua racemosa died before the next spring. The single survivor of the unwatered lot, Agropyrum, reached a height of only 11 inches the following summer, and did not blossom or extend its territory into the surrounding short-grasses. Agropyrum in the other lot grew only slightly better, Elymus was represented by only 2 shoots, and Stipa by a small c'ump, while Bouteloua racemosa alone made a good growth and seeded. Although the season was unusually favorable for growth, the results indicate that one more year would probably have been sufficient to eliminate most if not all of the transplants. As to the growth of the 1920 transplants at this station, it may be recalled that none died during the first season, even in the unwatered row, but all suffered from drought and were dwarfed. During 1921 none were watered. Root Development. On June 28-29 the root development of several species was examined. Elymus, which had blossomed the preceding season, had made a fair growth in the unwatered row, reaching a height of 21 inches with flower-stalks appear¬ ing. The old roots had not renewed their growth, but laterals on them were functioning (fig. 31), and in one case a large lateral had replaced the cut main root. It was traced to a depth of 20 inches. Many new roots reached a depth of 18 inches, while a few penetrated to the hard-pan, about 30 inches deep. All were quite profusely branched to near the tips, and the lateral spread was about 12 inches. Some new roots, evidently of the current year’s growth, extended just through the old block of sod. Andropogon furcatus was excavated in the unwatered row. The original sod, which was in complete contact with the soil, was a foot square and 8 inches deep. The new growth of leaves was about 8 inches high. Most of the old roots had died, but new ones had grown out thickly from the rhizomes. Many of these penetrated to the hard-pan layer at 28 to 30 inches, but only slightly into it. All of the roots were well branched, but were very dry, like the soil in which they grew. The dry, mellow loess below the hard-pan layer (which was 8 to 12 inches thick) contained roots of the short-grasses and Psoralea. There was very little lateral spread of the roots beyond the area occupied by the original block of sod. Koeleria, Stipa spartea, and Panicum virgatum were excavated in the watered row, which had also received some run-off water because of its loca¬ tion near a furrow separating an adjoining plowed area. The clump of SOD TRANSPLANTS. 69 Koeleria was 3.5 inches in diameter and had flower-stalks 12 inches tall, with rather normal spike development at a height of 6 to 15 inches. The roots, although very fine, were easily distinguished from those of the short-grasses by their yellowish color and slightly larger size. A great mass of profusely branched roots filled the soil to a depth of about 18 inches, a few penetrating 8 inches deeper. A lateral spread of 14 inches in the surface inch of soil was determined, the occupation of a considerable area on all sides of the plant being characteristic of the species. Stipa spartea had made a growth of 2 feet, with 5 flower-stalks, some of which were 34 inches tall, and formed a representative clump. The roots reached a maximum depth of 34 inches in the fairly moist soil, many ending at or above 2 feet. The lateral spread was about normal (10 inches), but the roots as a whole were more profusely branched than is usual in its native region. Panicum virgatum, which flowered the preceding season, had made a good growth, reaching a height of 18 inches. The root development was remarkable. The coarse roots, 2 to 4 mm. in diameter, pursued a nearly vertically downward course to a maximum depth of 7.3 feet, spreading only a little near the surface. The profound branching to the extreme depth of penetration is shown in figure 32. Branches at the rate of 25 per inch were not uncommon, and many of the roots were branched to their tips. In fact, the profuse branching, which undoubtedly was a response to the dry soil, can scarcely be overemphasized. It stood out in marked contrast to the rather poorly developed lateral roots found on plants growing in lowland soil near Lincoln (Weaver, 1919:4). Notwithstanding the root growth of these transplants, the severe drought conditions caused the death of several species. One sod of the watered Koeleria and one in the un¬ watered row succumbed. In the unwatered row, 1 Andropogon scoparius was winter-killed, 2 individuals of Elymus , and all 3 of Poa died. Agropyrum was the only species that increased its territory perceptibly. It invaded the short-grass sod on all sides for a distance of about 18 inches, the new shoots from the rhizomes reaching a height of 8 to 13 inches. How¬ ever, none of these or those of the original block bore flower-stalks. Bouteloua racemosa, Elymus , Stipa, and Panicum (in the watered row only) bore a few flower-stalks and seeded, as did also Poa and Koeleria , the Fig. 32. —Roots of Panicum vir¬ gatum at end of June of second year after the block of sod had been transplanted and watered at Burlington. 70 EXPERIMENTS DURING 1921. last in the unwatered row also. All suffered severely from spring and summer drought, and in general growth was poor. Most of the species, including such late bloomers as A. scoparius, A. furcalus , and Panicum, took on their late autumn reddish color, dried out, and went into winter condition by August 1. As a whole, those that had not been watered the preceding season, and hence were less well established, suffered most. During 1922 none of the plants died in the lot which had been watered, but Koeleria of the unwatered lot succumbed in July. Of the former, Agropyrum had extended its area 3 or more feet on all sides, even into the densest sod. Although it flowered profusely in its original area at a height of about 2 feet, only a few dwarfed flower-stalks appeared in the invaded area. This was the only species which extended its area, though, on the other hand, the short-grasses were unable to invade markedly. Because of the severe drought of July and August, the grasses at this time mostly took on their late autumn or winter color and habit. Aside from Agropyrum , Stipa, Elymus, Koeleria, and Bouteloua race- mosa blossomed; the last, unlike the others, had flower-stalks and inflorescences of about normal size. In the row that was not watered the preceding year, conditions were much more severe. Agropyrum possessed only a few flower- stalks, while Panicum put forth a few small panicles at 8 inches height. Elymus and Koeleria also blossomed. The very dry fall and winter following took severe toll. Of the 15 lots of plants in the area formerly watered, 4 died, while among those less well established 8 succumbed, leaving Agropyrum as the sole survivor. Notwith¬ standing the excellent growing conditions of the following summer (1923), it had no flower-stalks, but had sent out a few rhizomes to distances of 2 to 3 feet. The rest of the plants received extra water from a furrow which was only 2 feet distant and separated the unbroken and broken land. Hence, their growth was very good. Agropyrum had spread nearly 4 feet, but no flower-stalks occurred in this sodded area. Elymus had also extended its area over a foot, but another block of the same species died. Panicum , Bouteloua racemosa, and Stipa blossomed. Among the forms introduced with the sods, Erigeron ramosus blossomed the second year (1921), as did also Sporobolus asper, though both died later. Brauneria pallida in one instance made a good vegetative development only, while in another case it blossomed normally the second year after transplanting before succumbing. Aster multiflorus occurred in several blocks of sod and usually spread a foot or more by rhizomes; it blossomed profusely. In some cases it persisted 4 years near the furrow, but in others it soon succumbed. EXPERIMENTS AT OTHER STATIONS, 1921. Physical Factors. Rainfall and Holard. Studies on experimental vegetation were continued in the series of edaphic stations at Lincoln, at Nebraska City, and Colorado Springs during 1921. The general conditions of precipitation at Lincoln have already been given (p. 56). The season at Nebraska City was one of drought. The precipitation from February to July was decidedly below normal, that of April, May, and June being 0.86, 1.56, and 0.52 inches respectively below the mean. However, EDAPHIC STATIONS. 71 July had an excess of 5.17 inches, but August was slightly below normal. No efficient rain fell between May 9 and 26 or from June 18 to July 2. The holard at the Nebraska City, low-prairie, and gravel-knoll stations is given in table 25. Notwithstanding the decreased rainfall, the soil at Nebraska City usually had a margin of 10 per cent above the hygroscopic coefficient. In fact, a more constant supply was maintained here than on the low prairie at Lincoln, where the chresard in the surface 6 inches was practically exhausted during the last half of June. An abundant supply was at all times available below the 6-inch level. Even on the gravel-knoll conditions were quite favorable as compared with the preceding year. However, the xerophytism of this habitat is shown by the small chresard, usually only 1 to 6 per cent, even at depths of 3 or 4 feet. Evaporation and Temperature. The average daily evaporation at Nebraska City (fig. 33) was considerably lower (often 33 per cent) than on the high prairie. In general, evaporation on the gravel-knoll was greater than on high prairie. This was especially true during the earlier and later parts of the season. The average daily air-temperature at Nebraska City was usually higher than that at Lin¬ coln, and during the first 3 weeks of June it ranged from 2° to 11° F. higher. This was largely due to the much higher average night temperatures (often 6° to 10° F.), since the day temperatures nearly always averaged lower than those at Lincoln. The temperature of the soil at 18 inches was constantly 1° to 3° warmer. Fig. 33.—Average daily evaporation on high prairie (solid line), gravel-knoll, Lincoln (short broken lines), and at Nebraska City (long broken lines), 1921. Planting Results. Surface Sowing. During 1921, surface sowings were made on the low prairie at Lincoln and at Nebraska City. Of the 9 species planted at Lincoln on April 23, Aristida purpurea and Bouteloua gracilis did not germinate. Elymus canadensis had died by June 27, as the following did also during July and August: Lespedeza capitata , Liatris punctata , L. scariosa, Sporobolus asper, and Stipa viridula. Koeleria cristata was still alive in September, but was missing the following 72 EXPERIMENTS DURING 1921. May. Owing to the low light intensities, all of the seedlings were extremely attenuated as compared with those growing in the quadrats, and were unable to make any real growth. The ability of Koeleria to endure shading was noted Table 25 .—Holard in excess of the hygroscopic coefficient at Nebraska City, Lincoln low prairie , and gravel-knoll , 1921. Nebraska City. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 * feet. 3 to 4 feet. Apr. 9. 24.4 21.4 18.5 11.7 8.7 Apr. 23. 19.4 20.1 • • • • • • • • • • • • May 6. 10.7 16.5 15.6 .... , , , , May 20. 10.1 16.1 16.8 12.6 12.8 June 3. 23.8 21.6 19.5 .... .... June 24. 12.0 9.9 15.0 13.6 13.8 July 15. 9.2 14.9 14.3 11.5 9.6 Aug. 13. 14.8 10.5 16.0 15.9 10.1 Aug. 26. 12.8 10.2 10.9 12.5 9.1 Hygro. coeff. 12.1 11.7 12.3 13.7 12.9 Low Prairie. Apr. 9. 20.8 25.2 Apr. 23. 28.4 23.3 21.5 20.5 29.7 May 18. 10.5 24.8 21.6 22.3 23.0 May 25. 9.7 20.0 .... .... • • • • June 1. 6.8 16.4 .... • • • • .... June 7. 9.7 21.7 21.3 23.9 22.4 June 22. 1.4 16.0 19.2 .... .... June 29. 0.3 11.1 14.2 16.8 .... July 20. 25.4 23.0 15.3 15.2 18.9 July 28. 18.2 17.9 .... • • • • . . . . Aug. 3. 15.0 12.1 14.5 .... .... Aug. 23. 12.3 14.9 11.8 13.7 19.3 Hygro. coeff. 11.8 11.1 10.3 10.2 11.9 Gravel-Knoll. Apr. 9. 6.5 3.9 5.0 5.0 1.0 May 18. 6.3 5.3 6.6 9.2 1.7 May 25. 0.1 -0.2 • • • • • • • • .... June 1. 3.3 -0.3 .... .... .... June 7. 7.8 6.4 11.8 5.7 1.2 June 22. -0.1 0.4 3.4 4.5 3.9 June 29. 1.9 3.3 3.1 3.3 .... July 13. 4.5 2.6 3.7 • • . . .... July 20. 4.8 3.1 2.1 4.7 9.6 July 28. 0.4 1.3 • • . • .... .... Aug. 3. 4.8 4.0 2.7 .... .... Aug. 10. 3.1 2.3 2.0 6.7 10.8 Hygro. coeff. 5.2 4.9 3.1 2.8 7.4 in many instances, this adaptation probably resulting from its smaller stature. Light values at the soil-surface in the several stations are summarized in table 26. EDAPHIC STATIONS*. LIGHT VALUES. 73 Table 26 .—Light values at the several stations. Cover. High prairie, Lincoln. May 28, 1922 :* Andropogon scoparius. Psoralea tenuiflora floribunda Psoralea, Kuhnia glutinosa.. Brauneria pallida. Astragalus crassicarpus. July 25, 1921: Mixed grasses. Average density. Light cover. Medium, 4 inches above surface. Light. Trench, medium cover. Swamp, Lincoln. May 28, 1922: Spartina cynosuroides. Phalaris arundinacea. . Mixed grasses, etc.... Mixed grasses, average. August 24, 1922: Spartina cynosuroides Densest places. Medium density.... Most open. Nebraska City. June 3, 1921: 3 Stipa spartea. Andropogon furcatus. Andropogon scoparius. Liatris scariosa. Grasses, Amorpha canescens. Rosa arkansana.. Brauneria pallida.. A. scoparius, Koeleria cristata A. scoparius, A. nutans.. Stipa spartea. Very open grass mixture. Ceanothus ovatus. A. furcatus at side of quadrat. June 3, 1922: Very open grass mixture. Euphorbia corollata, grasses.. Rosa arkansana. Baptisia bracteata. Vernonia fasciculata. A. furcatus, Desmodium. Stipa spartea, Andropogon. . . Koeleria cristata, Poa praten- sis. Ceanothus ovatus. Per cent. 15 20 to 26 6 10 to 15 2.5 to 3.5 8.3 5.0 9.5 15.0 17.0 20 to 23 7.5 6.5 4.5 to 20 11.2 1.5 to 2.6 4 to 5.7 11.4 to 16.6 18 11.5 17.5 to 26.5 7.5 8 to 11.5 2.5 7 15 20 8 to 10 60 1.5 15 40 to 45 20 6.5 6.5 . 8 9 9 22 to 26 2 to 2.5 Cover. Low prairie, Lincoln. May 28, 1922: Mixed grasses: Above 1921 growth. Average. Under old growth. June 14, 1922: 2 Surface seeding, average den¬ sity. Side of 1922 trench: Spartina cynosuroides. Verbena stricta. Light, mixed. Andropogon furcatus.. Grasses, Art. ludoviciana.. . Lighter mixed cover. South edge of 1922 quadrats. . . West edge of 1922 quadrats... . July 25, 1922: Mixed grasses: Average density. Lighter. Very dense. 4 inches above trench surface: Medium. Lighter. Dense. Phillipsburg. June 10, 1922: Bouteloua gracilis and Bul- bilis dactyloides. Psoralea tenuiflora. Elymus canadensis. Andropogon furcatus. A. scoparius. Bouteloua, Gaura, Lygodes- mia. Verbena stricta. Grasses: Edge of 1922 trench. Edge of 1921 quadrat. . . . Burlington. June 11, 1922: Bulbilis dactyloides 4 . Bulbilis, B. gracilis, open 4 . Bulbilis, very dense 4 .. Agropyrum glaucum 6 Stipa spartea. Short-grasses: Malvastrum coccineum 4 . Erysimum asperum 4 Edge of 1922 trench. Edge of 1920 quadrat. Edge of 1922 quadrat. Per cent. 7.5 to 28 16 3.4 to 14 5 to 7 12.5 5 12 7 to 12 8 to 13 26 15 to 27 10 2.5 to 4 12 1.5 11 16 4.5 60 to 80 35 24 3 to 8 4 to 5 10 11.5 30 to 33 20 to 26 45 to 50 71 to 83 25 20 to 43 15 43 26 56 20 37.5 to 45 1 Burned over in late winter. 2 This portion of the area burned over in late winter. 3 Mowed the preceding September. 4 Photometer sunken so as to be level with soil surface. 6 Transplanted sods. 74 EXPERIMENTS DURING 1921. Of the 9 species sown on the surface at Nebraska City, Liatris scariosa and Pinus ponderosa failed to germinate, squirrels having eaten the seeds of the latter. A single plant of Elymus was found, but it died before June. By June 24, Liatris punctata and Bromus inermis were also dead, the latter having been eaten to the ground by grasshoppers. Sporobolus asper and Stipa viridula (also somewhat eaten) died in later summer. Bouteloua gracilis and B. hirsuta survived the season, reaching a height of 1 or 2 inches, tiller¬ ing somewhat, but persisting as very delicate seedlings. Of the 1920 surface sowing (p. 43), Koeleria alone survived the first sum¬ mer on the low prairie, but was then winterkilled. At Nebraska City, 4 species, Andropogon scoparius, A. nutans, Bouteloua hirsuta, and Elymus cana¬ densis survived. During 1921 a single small weak survivor of B. hirsuta remained alive until July 15, when it succumbed. The other species had become indistinguishable from the general plant cover by midsummer. Trench Sowing. Plants in the trench at Nebraska City did even more poorly than those on the surface. Aristida purpurea, Bouteloua gracilis, B. hirsuta, Liatris scari¬ osa, Pinus ponder osa, and Stipa spartea did not germinate, squirrels having eaten the seeds of Pinus. By June 24, Elymus canadensis, Liatris punctata, and Sporobolus asper, all represented by a few plants only, had died. Calamo- vilfa longifolia, which had germinated abundantly, also succumbed in July. Bouteloua racemosa survived the first season, tillered rather freely, and reached a height of 4 to 7 inches; however, it was winterkilled. Andropogon nutans alone survived, and it was represented by a good stand that had reached a height of 10 to 15 inches by fall. Out of 11 species planted in a trench on the gravel-knoll April 20, Aristida, B. gracilis, Koeleria, Lespedeza, and Liatris scariosa failed to germinate. A few of each of the following species germinated, the two last dying almost at once and the rest during June: Elymus, Liatris punctata, Sporobolus, Stipa viridula. Pinus died in July and Andropogon nutans alone survived, being represented by a few plants 4 to 5 inches tall in September. These lived over the winter, but were doing poorly in May 1922, and all died in June. Six of the 8 species planted in the low-prairie trench germinated, but Bouteloua gracilis, Lespedeza capitata, and Liatris punctata died soon after. Elymus canadensis succumbed to dense shading in June, and Stipa viridula, which was represented by only 3 delicate plants, in July. Aristida purpurea, simi¬ larly undeveloped and few in number, consisted of a few remnants 4 to 6 inches tall in late summer, but unfortunately these were dug up in trans¬ planting sods. Of the seedings in the trench on the gravel-knoll during 1920, Andropogon scoparius and A. nutans were the only survivors, and these were in a very poor condition. Both made a feeble growth and died late in June 1921. In the low-prairie trench, 4 species survived the first season (1920), but all showed the effects of shading. Of these, Stipa viridula and Bouteloua hir¬ suta made a poor growth the next spring and died in July. Andropogon nutans and Elymus canadensis, although much attenuated, made a good growth, reaching heights of 12 to 15 inches. In 1922 a small clump of each reached a height of about 32 inches, but neither blossomed, owing to the EDAPHIC STATIONS: CULTIVATED AREAS. 75 dense shade. However, the following summer both blossomed at a height of 2.5 feet. At Nebraska City all but 4 of the 1920 species planted in the trench survived, but Bouteloua gracilis died the following winter. Liatris punctata and Elymus were both quite abundant in May 1921, but succumbed during June and July. Andropogon nutans and A. scoparius came through the second season in a thriving condition, reaching a height of 5 to 9 inches. A sparse growth of Bouteloua hirsuta also survived the season, but the plants were quite delicate. During 1922, Bouteloua succumbed during August drought, after making a good growth earlier. The two Andropogons formed good sods 10 to 13 inches tall; both grew well in 1923, but neither blossomed. Growth in Cultivated Soil. Native species made remarkable development when grown in narrow rows (trench method) in fertile cultivated soil, kept at a favorable holard at all times, and free from competing species (table 27). These were planted on April 22 and 23 on a level area of soil, the physical and chemical composition of which was almost identical with that of the cultivated lowland area already described (p. 41). Potatoes had been grown on the area the preceding year. A month before planting a good seed-bed was formed by plowing and repeated Table 27 .—Growth of species in cultivated soil , 1921. Species. Aug. 2. Oct. 11. Agropyrum glaucum. . . Excellent growth, 14 in. tall. Excellent growth, IS in., no flower- stalks. Andropogon nutans. . . . Excellent, 24 to 32 in. Lvs. 2 ft., flower-stalks 3 to 5 ft., seed ripe. Aristida purpurea. Fine, 12 in. tall. Fine bunches 1 ft., flower-stalks 18 to 20 in., abundant, seed ripe. Bouteloua gracilis. Excellent, 1 ft., flower- stalks to 2 ft., abundant blooming. Lvs. 16 to 18 in., flower-stalks 23 to 33 in., seed ripe. Bromus inermis. Good 10 to 15 in. Lvs. 9 in., dense growth, flower-stalks 18 in. Calamovilfa longifolia.. Excellent 20 to 24 in. Fine development, 24 to 27 in.; flower-stalks 3 to 5 ft.; seed ripe. Elymus canadensis.... Fine, flower-stalks 2 to 2.5 ft., heads 5 to 5.5 in. long. Fine 15 to 18 in., heads abundant at 30 in. Gleditsia triacanthus... Good, 7 to 14 in. Good, 13 to 20 in., stems 7 to 10 mm. diameter. Lespedeza capitata.... Good, 4 to 5 in. Good, 5 to 6 in. Liatris punctata. Good, none over 5 in. Good, 4 to 6 in., max. of 7 lvs. Muhlenbergia pungens. Excellent, 6 in. Flower-stalks 15 to 18 in. seeded; bunches 12 to 16 in. diameter, 8- in. lvs. Onagra biennis. Rosettes abundant, 2 to 3 in. tall, 6 to 7 Ivs. Rosettes 5 to 8 in. diameter, 2 to 3 in. tall. Robinia pseudacacia. .. 2 to 3.5 ft., excellent. 5 to 5.5 ft. stems over 0.5 in. diam¬ eter. Sporobolus asper. Excellent, 20 in. Bunches 3 to 3.5 in., flower-stalks abundant, 3 to 4 ft., seed ripe. Stipa spartea. Good growth. Lvs. 15 to 22 in., very broad; no flower-stalks. Stipa viridula. 20 to 22 in., fine. Excellent; no flower-stalks. 76 EXPERIMENTS DURING 1921. harrowing (Weaver, Jean, and Crist, 1920, p. 80). The growth made during the first season by Andropogon nutans and Calamovilfa longifolia is shown in plate 10b and 11a, and that of Bouteloua gracilis in plate 11b. Root Development at Lincoln. Pinus ponderosa reached a height of 1.5 inches when 45 days old (June 6). The first whorl of leaves was as long as the cotyledons, while the strong, vertically descending tap-root had penetrated to a depth of 13 inches. The root was 1 mm. in diameter, brown in color, and had no laterals over 4 cm. long. These were usually about 2 cm. in length, white near the end, and entirely unbranched (fig. 34). Liatris punctata of similar age had a single leaf, 2 to 4 inches high, m addition to the cotyledons. The glistening white tap-roots, which had already stored some food in their enlarged upper parts, tapered rapidly and reached depths of 10 to 16 inches. They descended almost vertically and gave off but few short, unbranched laterals (fig. 35). Robinia pseudacacia seedlings 45 days old had about 6 compound leaves and a height of 5 inches. The tap-roots, which were about 1 mm. in diameter, tapered rapidly, but regained their maximum diameter again in the vigorously growing unbranched portion of the tip. The much branched tap-root de¬ scended irregularly to a depth of 12 to 15 inches; branching began about an inch below the surface and continued at the rate of 10 to 20 per inch, some of the horizontal laterals having a spread of 5 inches (fig. 36). The root development of Gleditsia triacanthus was even more remarkable. On July 12, when 81 days old, the trees had a height of 7 to 9 inches and 15 to 17 compound leaves. The tap-roots reached depths of 30 to 40 inches. The course was rather directly downward, long branches being given off profusely to a depth of 20 inches. Below this for 9 to 12 inches many shorter EDAPHIC STATIONS: ROOT DEVELOPMENT. 77 ones arose, while for the rest the tap-root was unbranched. The great spread of the longer horizontal laterals (12 to 18 inches), their large number and their characteristic branches, which as often pursued an upward as a down¬ ward course, are shown for one of the largest and best developed root sys¬ tems in figure 37. The rosettes of Onagra biennis were 5 to 8 inches in diameter and 2 to 3 inches tall on October 5. The glistening white tap-roots, 5 or 6 mm. in diameter, penetrated downward in a somewhat devious course, tapering to 1 mm. in width at 1.5 feet, but reaching a depth of 39 to 44 inches in the stiff wet clay. Branches to the number of 100 were produced just below the soil surface and extended to a depth of 1 foot. These varied from 1 cm. to 6 inches in length in a few; they ran off rather horizontally and were poorly rebranched. Below a foot the branching was much less pronounced (3 to 6 per inch of tap-root) and the laterals short. In the deeper soil the tap-root took on the appearance of a delicate white thread. Near the end of the growing- season (October 5 and 6), the root Fig. 36 .—Robinia pseudacacia 2.5 months old. development was examined in Muhlenbergia pungens and Calamovilfa longifolia, both characteristic sand¬ hill species but growing here in rich silt-loam. The former was repre¬ sented by bunches 12 to 16 inches in diameter and about 8 inches in average height. Flower-stalks 15 to 18 inches tall were abundant (plate 12a). The root development was marked, roots being traced in the clay sub¬ soil to a maximum depth of 2.5 feet; they were very abundant in the first 22 inches. Many new roots, with thick, white, rapidly growing tips ended in the first foot, some of these originating from the prostrate stems. All branched profusely upon entering the soil. However, there were not so many surface roots as described for sandhill specimens (Weaver, 1920:89), but this may have been due to the age of the plants. Fine laterals were exceedingly numerous, as many as 30 to 50 per linear inch. While most of these were only 1 cm. or less in length, others were 8 to 9 cm. long All were profusely and minutely branched, the larger ones to the third and fourth order, many even to the tips, although some that were growing rapidly were wide and free from branches. Calamovilfa longifolia had made an excellent growth, having leaves 24 to 27 inches high and an abundance of flower-stalks at 24 to 27 inches. Many tough, wiry rhizomes, thickly covered with long scales and tipped with buds an inch long with very sharp, hard points, extended out on all sides of the 78 EXPERIMENTS DURING 1921. clumps for distances of 6 to 8 inches. Multitudes of tough, wiry roots pene¬ trated the soil vertically or obliquely downward to depths of 5 feet and some extended beyond a depth of 6 feet. Beginning just below the soil-surface and extending to near the tips of the larger roots, laterals well provided with Fig. 37.—Root system of Gleditsia triacanthus less than 3 months old. branches to the third and fourth order were abundant. Most of the major laterals were only an inch or two long, but with the main roots they thoroughly occupied the soil beneath and for several inches on all sides of the clumps (plate 12, a and b). The excellent growth of all of these native species under favorable condi¬ tions, 73 per cent of the grasses bearing seed the first season, emphasizes the keen competition prevailing in stabilized grassland. Sowing in Denuded Quadrats. Fourteen species were planted in denuded quadrats on the gravel-knoll, April 14. Koeleria, Liatris 'punctata, L. scariosa, and Muhlenbergia pun - EDAPHIC STATIONS: DENUDED QUADRATS. 79 gens did not germinate, while Lespedeza capitata, Pinus ponderosa, Sporo- bolus asper, and Stipa viridula germinated in small numbers (except the last, which was abundant), but all died before the end of June. At this time the mortality among the remaining species was also pronounced. Bouteloua gracilis and Cala7novilfa longifolia died in August. Symphoricarpus vulgaris, which was propagated from transplanted rhizomes, was nearly dead; Elymus canadensis was represented by a single plant 9 inches tall, Aristida purpurea by 4 small clumps, and Andropogon nutans, which had formed a thin sod, was in poor condition. Thus, while 71 per cent of the species germinated or grew from rhizomes, only 40 per cent of these survived the first summer. Of the 16 species sown in the low prairie, Koeleria, Liatris punctata, L. scariosa, and Symphoricarpus vulgaris failed to germinate or grow. Stipa viridula died during June, but the other species (except Robinia pseudacacia, which died in August) lived throughout the summer. Nearly all made a fair growth, but showed the effects of shading, the trees having thin, broad, pale leaves, while those of several of the grasses were much attenutaed. At Nebraska City, 7 of the 22 species failed to germinate. The seeds of Pinus ponderosa had been dug up and eaten by squirrels. Bouteloua gracilis, Lespedeza, Liatris scariosa, and Muhlenbergia died in June, as did also Bromus inermis and Elymus canadensis, both of which were eaten to the ground by grasshoppers. All of the others made good growth, shading at this station being less pronounced than on the low prairie. Andropogon nutans formed a dense sod (plate 13). Calamovilfa was all dead but a few plants, and Robinia was represented by a single seedling, grasshoppers having damaged the good stand of early summer. On the low prairie all the species germinated and all but one came through the season in excellent condition. Stipa setigera (seed from California) was winterkilled; the rest made a fine growth until the end of July, Aris¬ tida and both species of Bouteloua blossoming. Andropogon scoparius, A. nutans, A. furcatus, Koeleria, and Elymus formed good sods and, like Stipa viridula, had tillered heavily. A. scoparius, A. nutans, A. furcatus, and Elymus continued to thrive, reaching heights of 18 to 24 inches, Koeleria grew to a height of 7 to 11 inches, and Bouteloua gracilis and B. hirsuta put forth flower-stalks 18 to 20 inches tall. Stipa viridula was so badly shaded that half of the plants died, as did nearly all the tips of the leaves on the remaining plants. Autumn found Aristida purpurea also nearly dead or dying. In 1922, one lot of Aristida died in May and one of Stipa in June, while Bouteloua hirsuta succumbed in August. Although the others made a good growth for the most part, Andropogon nutans, A. scoparius, Aristida, Bouteloua gracilis, and Elymus alone flowered. During the summer of 1923, Aristida, Bouteloua gracilis, Stipa viridula, and one lot of Koeleria died as a result of the dense shade. While all three andropogons showed a normal development and formed dense sods, Koeleria and Elymus were clearly losing ground. At Nebraska City, all of the species in the 1920 quadrats germinated, became established, and lived throughout the summer. However, the mortality among the individuals was high and in general the stand was not as good as on the low prairie. All but Koeleria survived the second season. Grasshoppers kept this species eaten back and also made consider¬ able ravages on Stipa viridula, Liatris punctata, and Elymus. Most of the 80 EXPERIMENTS DURING 1921 . grasses tillered heavily, but no sod was formed comparable to that on the low prairie. Neither did they make an equal height-growth, Andropogon reaching heights of only 8 to 15 inches and Bouteloua 4 to 10 inches. Andro¬ pogon scoparius and Bouteloua hirsuta were the only species that seeded. Stipa viridula was winterkilled in 1921-22, but all the rest came through the following summer. Aristida and Elymus did poorly, but the others made fair to excellent growth, although flower production was not common, prob¬ ably owing in part to the dry fall. These 3-year-old quadrats were mostly invaded by Poa pratensis, species of Car ex, and Helianthus rigidus, but none to the extent of causing serious competition. In 1922-23, Elymus was winter- killed, while all the others not only made a very good growth, but some also produced seed. At Colorado Springs, 7 of the 9 species planted in denuded quadrats in 1920 germinated, and of these 4 survived the first season. During 1921, Andropogon nutans and Elymus canadensis died, but Andropogon scoparius formed an open tufted sod 2 inches high, which covered nearly the whole quadrat and reached a height of 4 inches by 1923. Stipa viridula also did well, tillering freely and reaching heights of 6 and 12 inches in 1922 and 1923 respectively. Root Development at Peru, Nebraska. A number of species were grown in a well-prepared seed-bed in a cultivated field near Peru, Nebraska, a station about 20 miles south of Nebraska City. The mellow silt-loam soil, underlaid at a depth of 1 to 1.5 feet with a loess of very loose texture, not only absorbs water readily, but has a high water¬ holding capacity. This ranges from 57 to 64 per cent and is rather uniform to a depth of at least 4 feet, the same type of subsoil extending to depths of many feet. The mechanical analysis of this soil shows that it is approximately one-half silt, while the remainder is composed almost entirely of very fine sand and clay. Weaver has shown that in this soil-type roots of species common to both true and subclimax prairie penetrate more deeply than in the stiffer silt loam at Lincoln (1919:15). Normally, the rainfall is about 5 inches greater than at Lincoln; however, during 1921 it was not only deficient, but also poorly distributed. During April, the rainfall was nearly 3 inches below normal, 70 per cent of it falling at one time. For May it was 1.6 inches below normal, 72 per cent of it falling between May 7 and 10. June, with 3.4 inches of precipitation, practically all of which occurred before the middle of the month, had a deficiency of 1.4 inches. The holard was much below nor¬ mal and the poor growth of crops marked the season as one of distinct drought (Weaver, Jean, and Crist, 1922:78), though the native species made a fairly good growth. When they had reached an age of 90 days (July 18-19), the development of the plants both above and below ground was recorded (table 28). The development below ground was marked and agreed well with pre¬ vious findings as to the deep-seated nature of mature root systems in this mellow loess soil. Seedling Transplants. Seedlings of 14 species were transplanted to the gravel-knoll on May 9. Agropyrum glaucum, Aristida purpurea, and Stipa viridula had died by May 25, Elymus canadensis and Koeleria cristata succumbed early in June, and EDAPHIC STATIONS! SUMMARY. 81 Bouteloua racemosa and Stipa comata later in the month. Liatris punctata and Stipa spartea died in July and Andropogon furcatus in August. By the end of the summer Bouteloua gracilis and Liatris scariosa were represented by remnants only, and both species died later in the year. Bouteloua hirsuta came through the summer with only one small clump about 3 inches tall and Andropogon nutans with two, 6 or 7 inches high, the survival for the first season being 14 per cent. Table 28. —Development of seedlings at Peru, Nebraska, July 19, 1921. Species planted Apr. 19. Development of shoots. Work¬ ing depth of roots. Maxi¬ mum depth of roots. Lateral spread of roots. in. in. in. Andropogon nutans.. Good stand; luxuriant growth; ave. height 11 in.; max. height 16 in. 36 48 Aristida purpurea. . . Good growth; height 4 to 6 in.; 12 to 30 tillers per plant. 26 41 • • Bouteloua gracilis. . . Good stand; lvs. 4 to 8 in. high; exceed¬ ingly well tillered; flower-stalks about 9 in. tall. 22 37 8 Elymus canadensis. . Vigorous plants; height 10 to 16 in.; 3 to 4 tillers per plant; heads appearing. 27 42 9 Liatris punctata. Good growth; 3 to 5 leaves per plant; 3 to 4 in. tall. 47 Stipa viridula. Thin stand, vigorous growth, ave. height 7 to 9 in.; 6 to 10 tillers per plant. 18 26 12 Seedlings that were transplanted to low prairie at the same time made a much better growth. All grew well, but by the end of June Agropyrum, Boute¬ loua racemosa, and Elymus particularly, showed the effects of shading. In July, through error, a part of the prairie into which the transplant area ex¬ tended was mown. Those in the area of better illumination made a fine growth, the shorter boutelouas reaching heights of 4 to 8 inches and the andropogons 6 to 12 inches. In the shaded area Agropyrum died and the other species made but a poor to fair growth, the leaves being thin and atten¬ uated; the loss of only 1 of the 13 species is remarkable. All of the seedling transplants at Nebraska City did very well until about the middle of June, when Agropyrum, Koeleria, and Liatris punctata died and by July 15, Bouteloua racemosa had likewise succumbed. The other boutelouas, Aristida, and the andropogons, however, were in a flourish¬ ing condition. Elymus, Stipa comata, and S. viridula had been badly eaten back by grasshoppers and were represented by remnants only, as were also Liatris scariosa and S. spartea. Elymus and Stipa viridula died in August, and S. comata was found to be in very poor condition in the autumn, but the other 7 species were doing quite well. Summary. A summary of the experiments for this year is given in table 29. The results from the Colorado Springs station are omitted because the area was broken into by cattle and closely grazed. Owing to the drought at Nebraska City, the percentage of germination is lower here than in low prairie at Lin- 82 EXPERIMENTS DURING 1921. coin. However, the percentage of establishment averaged slightly higher at the Nebraska City station, although growth was somewhat poorer. The germination averaged slightly lower on the gravel-knoll (where surface sow¬ ing was omitted) than at Nebraska City, while the establishment was decidedly Table 29. —Summary of sowing experiments, 1921. Method of planting. Percentage of germination. Percentage of germinated species established. Gravel- knoll. Low prairie. Nebraska City. Gravel- knoll. Low prairie. Nebraska City. Surface. 78 78 14 29 Trench. 55 75 50 17 17 33 Denuded quadrat.... 71 75 68 40 83 60 Average. 63 76 65 29 38 41 Seedlings. • • 14 92 57 lower. The survival of transplanted seedlings was highest on low prairie (92 per cent), intermediate at Nebraska City (57 per cent), and least on the gravel-knoll (14 per cent). Of the survivors of the surface-sown species of the preceding year (1920) on low prairie, Koeleria was winterkilled. One of the 4 survivors at Nebraska City died. Of those sown in the trench, both survivors of 1920 on the gravel- knoll died, 2 of the 4 on low prairie, and 3 of the 6 at Nebraska City. Among the species planted in the denuded quadrats during 1920, none persisted on the gravel-knoll. One of the 10 survivors on the low prairie died during 1921, and also 1 of the 10 at Nebraska City. Table 30. —Comparison of germination and growth at the several stations, 1921. Average per cent of germination. Average per cent of establishment of germinated species. Lincoln high prairie. 81 Lincoln low prairie. 76 Phillipsburg. 68 Nebraska City. 65 Lincoln gravel-knoll. 63 Burlington. 43 Phillipsburg. 65 Nebraska City. 41 Lincoln high prairie. 40 Lincoln low prairie. 38 Gravel-knoll. 29 Burlington. 7 Per cent of establishment in denuded quadrats: Per cent of establishment of seedlings: Lincoln low prairie. 83 Lincoln high prairie. 65 Phillipsburg. 63 Nebraska City. 60 Gravel-knoll. 40 Burlington. 21 Lincoln, low prairie. 92 Phillipsburg.,. 75 Nebraska City. 57 Lincoln, high prairie. 54 Gravel-knoll. 14 Burlington. 0 If the data from the other stations are included and arranged in the order of the average percentage of germination (table 30), they prove to be in general agreement with those of the preceding year (p. 48), with two excep¬ tions. Owing to the drought, Nebraska City ranks after Phillipsburg, in¬ stead of heading the list, while Phillipsburg falls behind low prairie in rank. When the stations are arranged according to the average percentage of establishment, Phillipsburg ranks highest with 65 per cent (largely because EDAPHIC STATIONS! SOD TRANSPLANTS. 83 of the excellent establishment of surface-sown plants). The Lincoln stations and Nebraska City are about the same (38 to 41 per cent), while the gravel- knoll and Burlington have changed in relative position, the latter giving the lowest percentage of establishment. On the basis of establishment in de¬ nuded quadrats, the same sequence occurs as during 1920, except that Ne¬ braska City then ranked first and the gravel-knoll last. If an average is obtained from all four criteria for growth, the stations aline them¬ selves as follows: Low prairie 72, Phillipsburg 67, high prairie 60, Nebraska City 56, gravel-knoll 37, and Burlington 18. Sod Transplants. Gravel-knoll. Between March 20 and 31, 18 blocks of sod, including 14 species from the various Lincoln stations, were transplanted to the gravel-knoll. Distichlis spicata died in June, while Agropyrum glaucum, Andropogon sco- parius, and two lots of Koeleria cristata died during August, notwithstanding the relatively favorable season for growth. With the exception of Poa pratensis, all of the other species made a fair growth, but the scarcity of water was indicated by the rolling of leaves and frequent wilting (table 25). Andro¬ pogon furcatus and A. nutans reached a height of a foot and Panicum virgatum one of 1.5 feet, but did not put forth flower-stalks. Bouteloua hirsuta and racemosa had a few flower-stalks each at 1 and 2 feet respectively, while Bulbilis blossomed profusely in June, but like the other species did not increase its area. Elymus bore several fine heads at a height of 2 to 2.5 feet and Spartina cynosuroides from the swamp flowered at 27 to 34 inches; Poa and Stipa also blossomed. Poa and both clumps of B. racemosa were winterkilled. Elymus died during the August drought, but all of the others made a fair to good growth, although adversely affected by the late summer drought. Seven of the 8 species bloomed, Andropogon furcatus from the low prairie alone failing to put forth flower- stalks. All survived the drought of the following fall and winter. Panicum alone made a poor growth the following season, while Andropogon furcatus, Bulbilis , and Spartina spread a foot or more beyond their original area. By August 25, Stipa and Bulbilis were the only species that gave signs of having flowered or preparing to flower, a striking contrast to the preceding year, when drought promoted reproduction. Even during the very favorable year of 1923, but more especially during preceding seasons, sod transplants on the gravel-knoll were all much shorter than elsewhere. They also began renewed development earlier and produced flower-stalks sooner. Of the species planted in the spring of 1920, although suffering severely from drought, only one sod of Poa and another of Andropogon scoparius in the unwatered row died (table 66). One of Koeleria was winterkilled, while a sod of Elymus and another of Koeleria died during August of the second season. Practically all of the species, including Panicum virgatum and Spartina cynosuroides, made a fair growth, rolling of leaves and wilting occurring much more rarely than in transplants of the current year, owing to the well-established root systems. Exceptions to this occurred in the case of Agropyrum, both lots of which did very poorly, as did also one sod of Elymus and the remaining block of Koeleria. After the middle of August, when the 84 EXPERIMENTS DURING 1921. short-grass cover of the knoll began to dry, most of the tail-grass sods began to change color and enter the winter condition also. One lot each of A. furcatus, A. scoparius, B. racemosa, Elymus canadensis, and Poa pratensis and all three lots of Panicum virgatum had flowered and set seed, although the flower-stalks were fewer in number and smaller than normal. Of the species that seeded the first season after transplanting, viz, Koeleria, Stipa, Agro- pyrum, Elymus, Panicum, and Poa, only the last three seeded the second year. The single remaining lot of Koeleria and Poa were both winterkilled and one Agropyrum died in June 1922. Both lots of Elymus died later in the summer, as did also one of Andropogon furcatus. The effect of drought upon the other species is shown in table 66. So well were these species rooted, however, that the dry fall and winter of 1922-23 took but a single Andropogon scoparius. The growth during 1923 was marked, Panicum doubling and Andropogon nutans tripling its area. Low Prairie. Fifteen species of grasses were transplanted into the low prairie late in March 1921. The rank growth of the native tail-grasses produced a very dense shade (plate 6a). Bouteloua hirsuta, Distichlis spicata, and Koeleria cristata had died by late summer and Agropyrum had made but poor growth. Bouteloua gracilis and Bulbilis dactyloides both reached a height of 12 inches, the leaves being much attenuated, but neither was able to increase its territory. Andropogon furcatus, A. nutans, Bouteloua racemosa, Spartina cynosuroides, and Stipa spartea all developed about normally, but none blos¬ somed. Bulbilis, Poa, Andropogon scoparius, Elymus, and Panicum bore well-developed flower-stalks and seeded rather profusely. During the follow¬ ing May and June, owing to the dense cover and resulting competition, Andropogon scoparius, Bouteloua racemosa, B. gracilis, and Stipa spartea, all high-prairie species, died. Bulbilis died in July, while by the end of the summer Agropyrum, one Elymus , and Poa were represented by remnants only. Elymus, Panicum, and Spartina alone blossomed. All these survived the dry winter, but by August 1923, Agropyrum was dead and Poa nearly shaded out. Both lots of Elymus were heading at 40 to 42 inches high, while the other subclimax species, viz, Spartina, Panicum, Andropogon nutans, and A. furcatus, were all flourishing. As to the 1920 transplants at this station, all made a good growth, and all except Spartina cynosuroides bloomed (table 67). By May 18, 1921, Bouteloua racemosa, one lot of Koeleria, and two of Stipa spartea were dead. Many others, particularly Agropyrum, Bulbilis, Distichlis, and Stipa, were suffer¬ ing from shading and invasion, but all other species made a good growth during June and July. By the end of the summer, one Agropyrum had died, and the other was much attenuated. Bulbilis, Distichlis, and Koeleria were nearly dead, being represented by remnants only, and Bulbilis alone had blossomed. However, all of the other species had made an excellent growth. The three andropogons exhibited abundant flower-stalks 3 to 5 feet high, and Bouteloua gracilis had leaves a foot tall, but flowered sparingly. Elymus and Panicum possessed flower-stalks in profusion 4 to 5 feet tall, and Stipa seeded at 33 to 36 inches, while Spartina made a good growth but did not blossom. By May of the following spring (1923), one Andropogon scoparius and Koeleria had died in addition to Distichlis. Panicum virgatum and Andro- EDAPHIC STATIONS: SOD TRANSPLANTS. 85 pogon nutans and one lot of A.furcaius were indistinguishable from the general cover. Poa pratensis and another Andropogon scoparius died in July, and Agropyrum glaucum and Bulbilis by August. At this time the following were in rather poor condition: one Andropogon scoparius, both plants of Bouteloua gracilis, and Elymus and Koeleria, but the rest had made a good growth. By the spring of 1923, Bouteloua gracilis and one Andropogon scoparius were dead. By the end of this, the fourth season, sufficient time had elapsed for a fair adjustment of the species to the low-prairie habitat. Although the remaining high-prairie species (two lots of Stipa and one of Koeleria) were still in fair condition, the last Andropogon scoparius was in straits. The other species of the subclimax prairie, viz, Andropogon furcatus, Elymus canadensis, and Spartina cynosuroides, were flourishing, while Andropogon nutans and Panicum virgatum had already become indistinguishable from the native sod. The following had entirely lost out in the struggle for light: Agropyrum, Bouteloua gracilis, Bulbilis, Poa, Bouteloua racemosa, and Dis- tichlis spicata. As a consequence, these experiments clearly reveal the effect of competition in sorting out species and stabilizing vegetation. Salt-flat. Sods of each of the preceding species were also transplanted into the salt-flat late in March, 1921. During June most of the plants suffered from drought, great cracks appearing on the sides of the transplanted blocks, owing to the peculiar nature of the soil (p. 44). Growth was very poor, wilting and even dying of parts being not uncommon. Throughout July, growth was slow, most of the sods making only a sparse growth and Spartina alone flourishing. Koeleria and Stipa failed to blossom or bore only a few dwarfed flower-stalks, and the inflorescence of Bulbilis was also shorter than normal. By the end of the summer, the leaves of Andropogon furcatus, A. nutans, and Bouteloua racemosa were dying back, while Panicum virgatum and B. hirsuta were nearly dead. These, like Agropyrum, B. gracilis, and Spartina, had failed to blossom. Andropogon scoparius, Elymus, and Poa had dwarfed flower-stalks and Distichlis alone developed normally. As a result of the drought of the following June, all the sods were in poor condition, large cracks occurring around most of the transplanted blocks, even those that had been transplanted 2 or 3 years, and many of them were very dry. One Koeleria died in June and the other two did very poorly. Poa and Panicum were represented only by remnants at the end of the season and the growth in many of the other blocks was very sparse. Practically all of the plants were dwarfed and only 4 species blossomed, viz, Andropogon scoparius, Bulbilis, Elymus, and Koeleria; on these the inflorescences were much dwarfed. During the very favorable growing-season of 1923, another Koeleria and Stipa died and Andropogon furcatus, A. scoparius, Bouteloua racemosa, Panicum virgatum, and Spartina cynosuroides were represented by remnants only. Seven species seeded, although not abundantly. The sods transplanted into the salt-flat during 1920 made a poor growth the first season, though 8 of the 13 species blossomed (table 68). Compared with the growth of similar species in the adjoining low prairie, nearly all were greatly dwarfed. However, they survived the first season. During May 1921, Andropogon scoparius, A. nutans, Stipa spartea, one lot of Elymus 86 EXPERIMENTS DURING 1921. canadensis, and one Koeleria cristata were in poor condition or dying. How¬ ever, many of the others, including Panicum and Spartina, were doing quite well. Throughout June and July they grew fairly well, except A. scoparius, which was mostly dead and badly invaded by Distichlis. All lived through¬ out the second season but Bouteloua racemosa, though none grew normally except Distichlis. Agropyrum, A. scoparius, A. nutans, B. racemosa, Spar¬ tina, and Stipa failed to flower. The dwarfed condition of the remaining species, all of which blossomed, was well represented by A. furcatus, which bore flower-stalks only about 2 feet high. Panicum was an exception, with a few stalks reaching a height of 40 inches, but its vegetative growth, as for all of the other species, was much below normal. Elymus canadensis failed to grow in 1922, while one Panicum virgatum died in June. Although the plants in general made a fair growth, they were dwarfed and several species did poorly. Andropogon furcatus, Distichlis, Koeleria, Panicum, and Poa blossomed. Owing to an excellent holard, all the species except Elymus, Stipa, and Andropogon nutans made a good growth during 1923. Consequently, it seems evident that several species can tolerate the saline areas dominated by Distichlis and Agropyrum, as was clearly shown during the wet year of 1923. As a result, the salt-basin is being gradually invaded by various prairie species. Salt-basin. A similar lot of sods was transplanted to an area in the salt-basin just west of Lincoln, which is dominated by Dondia depressa. This halophyte occu¬ pies an intermediate zone between the less saline Distichlis area and the wetter and more saline one of Salicornia herbacea. Repeated analyses of the soil in the Dondia area show that it has a salt-content of 1.8 to 2.0 per cent, prac¬ tically all of which is sodium chloride. The holard in spring and early summer is more or less favorable for growth. Dondia, Salicornia, and Distichlis were also transplanted early in April, together with the sods of the grasses. Within 20 days the transplants from non-alkaline soil began to lose their green color and the tips of the leaves to dry, and after another 20-day period all were dead. The three halophytes did very well throughout the season, except Salicornia, which suffered apparently from drought late in the summer. Swamp. Sods of the same species were transplanted into the swamp in March 1921. Owing to the building of a dam higher up the ravine, conditions were markedly different from the preceding year. The soil in the wettest area was not saturated to the surface throughout the spring, although water stood on the top during the first week in May. For much of the remainder of the season the area was relatively dry, the holard being so low that water could not be pressed out by hand from the surface 6 inches. The Poa zone was also drier than the preceding year. This area, which was nearly pure bluegrass at the beginning of the experiments (1920), was now invaded by a rank growth of Spartina, and light conditions were less favorable than for¬ merly. In the swamp all of the grasses made a very good growth until the end of June, Bulbilis flowering profusely, but Stipa and Koeleria, other early bloom- EDAPHIC STATIONS! SOD TRANSPLANTS. 87 ers, failed to bloom. At this time Spartina was 3 feet high all around the sods and the transplants began to show the effects of shading, the light values falling as low as 2 to 5 per cent. By the end of July the surrounding vegeta¬ tion was 4 feet tall and some of the sods were found with difficulty. The soil was wet, and hence many of the plants were slender, with narrow leaves. Late in August, Agropyrum, Bouteloua hirsuta, B. gracilis, Distichlis, and Koeleria were dying back or mostly dead, some being represented by rem¬ nants only. They were all attenuated and none had produced flower-stalks. Andropogon nutans, A. scoparius, B. racemosa, Spartina, and Stipa had all made excellent growth, but none produced flowers. Bulbilis, Elymus, Poa, and Panicum also grew well and blossomed about normally. The sods in the Poa zone, which were planted at the same time, made a good growth early in the season. By the end of June, Spartina, Cyperus, Mentha, etc., had reached a level of about 3 feet and the shade was dense. Bulbilis and Bouteloua gracilis both had delicate slender leaves 10 to 14 inches tall, while those of Koeleria reached 15 inches. None had blossomed, but Stipa had a few fruits which were smaller than normal. On August 31, one- half of Bouteloua gracilis was dead and Distichlis was nearly so. However, most of the other species, though slender, had made in almost every case a much better growth than in the wetter part of the swamp. Six of the 13 species, viz, Andropogon scoparius, A. nutans, Stipa, Distichlis, Elymus, and Panicum, blossomed, in contrast to 4 of the 14 species in the swamp. In 1922, the swamp area was much drier than during preceding years; in fact there was nowhere excess holard or deficient aeration, light being the controlling factor. Spartina was over 2.5 feet tall by the middle of June and 4 feet in August, covering the area with a dense growth. Conditions were very similar in the Poa zone. The leaves on practically all of the plants were very narrow, long, and delicate, Andropogon scoparius did not appear in the spring, and Koeleria died in August. The following early bloomers produced flowers: Bulbilis, Koeleria, Stipa, and Poa, while the subclimax dominants Elymus and Panicum also blossomed. By June of 1923, Boute¬ loua hirsuta, Bulbilis, and Distichlis were dead, while Bouteloua gracilis, B. racemosa, Elymus, and Stipa vanished later in the summer. This left a few remnants of Agropyrum and well-developed plants of Andropogon nutans, Panicum virgatum, and Spartina cynosuroides at the end of the third season. In the Poa zone all survived the 1922 season, but the following species died the next summer: Andropogon nutans, A. scoparius, Bulbilis, Bouteloua gracilis, Distichlis, and Elymus. Agropyrum,, Bouteloua racemosa, and Stipa spartea were represented by remnants only, while Andropogon furcatus, Pani¬ cum, and Spartina were the only species that had developed normally. The fate of the sods planted in the swamp during 1920 is instructive. A. scoparius, A. fur catus, A. nutans, B. gracilis, B. racemosa, Koeleria, and Stipa died the first season, while many of the others came through in very poor condition (table 69). In May of the following spring Agropyrum and Elymus, and especially Poa, showed signs of distress. Spartina developed nor¬ mally, while Panicum and Distichlis were in good health. Both Agropyrum and Poa had died by the end of July; autumn found remnants only of Distich¬ lis and Elymus, but Panicum and Spartina had made a good growth and seeded. During 1920, Koeleria, Stipa, Andropogon scoparius, and A. nutans 88 EXPERIMENTS DURING 1921. died in the Poa zone. The following spring Agropyrum glaucum and Bouieloua racemosa failed to appear, while B. gracilis died in June after a feeble growth. Distichlis succumbed in August, but Elymus, Poa, Panicum , and Sparlina made a fairly normal growth, all ripening seed. Since the swamp was much drier and the air-content higher, the chief factor inhibiting growth was not aeration but light (p. 73). Spartina be¬ came indistinguishable from the surrounding vegetation. Panicum, blos¬ somed normally, Elymus bore a single flower-stalk 2 feet high, while Distichlis did very poorly. In the Poa zone, Poa and Elymus made an excellent growth, the former merging into the common sod and the latter flowering at a height of over 4 feet. Panicum and Spartina developed normally in every way. In 1923 Distichlis died, but the three low-prairie species developed very well, the 4 species in the Poa zone also prospering. The results are in accord with the expectation, since all the species except those of the mesophytic subclimax prairie, viz, Andropogon furcatus, Elymus, Panicum, and Spartina, had lost in the struggle during the period of 3 or 4 years. Regional Transplants. As to the fate of the sods transplanted to Colorado Springs in 1920, both blocks of Stipa spartea had died by the end of 1921, as well as one each of the following species: Andropogon furcatus, A. scoparius, Bouieloua racemosa , Elymus, Koeleria, and Panicum. This was chiefly due to drought, but partly also to overgrazing, stock having broken into the inclosure at several times. In 1922, Bulbilis alone blossomed, repeated grazing again weakening the other plants. By the autumn of 1923, grazing having been continued, the last lot of Elymus had died, several of the other species were represented by rem¬ nants only, and all were in rather poor condition. During 1921, sods of Stipa setigera, S. eminens, Poa tenuifolia, and Melica imperfecta from California were transplanted into high and low prairie re¬ spectively on June 23. These were received at Lincoln on May 18, but were kept in a garden and watered frequently before the large blocks of soil con¬ taining the transplant were removed to the native grassland. Here they were again watered from time to time as necessary. Only a few made a feeble growth and all on the high prairie were dead by July 30. On the low prairie four lots put out new shoots 3 to 5 inches long, but these died by the end of the summer. On July 20 the following species from Arizona were transplanted in quad¬ ruplicate into a garden at Lincoln and kept well watered: Bouteloua eriopoda, B. rothrocki, B. bromoides, and Aristida divaricata. All made some growth. In September one clump of B. eriopoda bore flower-stalks about 15 inches tall, and one B. bromoides possessed leaves 9 inches high and flowered abundantly at 15 to 22 inches. 4. EXPERIMENTS DURING 1922. PHYSICAL FACTORS. Rainfall. The season of 1922 was fairly favorable for growth, except the latter part, when severe drought occurred at all the stations, being relatively less marked at Burlington. At Lincoln the combined precipitation for March and April averaged about normal; that of May and June was approximately 1 and 2 inches below the mean respectively, July had an excess of 2.4 inches, but only 0.7 of the 3.7 normal rainfall of August occurred (fig. 9). At Phillipsburg the April precipitation was nearly 1.5 inches above normal (no report for May), June, July, August, and September showing deficiencies of 2.4, 0.6, 1.8, and 1.6 inches respectively (fig. 15). The Burlington precipitation was about half the normal during March (April report missing), 0.4 inch in excess during May, while June gave an excess of 1.4 inches over the mean. July had a deficiency of 0.8, August an excess of 0.5, and September was also far below the normal (fig. 16). Spring and early summer drought periods, i. e., rainfall less than 0.21 inch, occurred at Lincoln from April 11 to May 8 and May 26 to June 24. At Phillipsburg a 10-day drought occurred in April (May report missing) and no rain fell from June 1 to 26. The severity of the late-summer drought at both Lincoln and Phillipsburg is well illustrated by the fact that by the end of August the late-developing Andropogons were already half dried and appeared to be entering the winter condition. No efficient precipitation fell at Burlington from May 2 to 19 and 20 to 29 (April report missing) or June 2 to 15. Holard. A study of the water-content (table 31) shows that a margin of 5 per cent (more usually 7 to 11 per cent) above the hygroscopic coefficient existed at all times to a depth of 4 feet in the high prairie at Lincoln (except in the surface 6 inches on June 22). At the mixed-prairie station a good water-content prevailed to 3 feet in the spring, but the margin of safety was much less at nearly every determination than at Lincoln. Late in June, and again throughout July and August, drought prevailed, no water being available often to depths of 4 feet. Condi¬ tions at Burlington were not very different, the soil in spring and early summer being as usual quite moist. However, the deficiencies were very marked and practically continuous after June 18. Temperature. Air-temperatures in general were lowest at Burlington, 7° F. higher at Lincoln during early May, and usually 3 to 5° higher during the rest of the season; those at Phillipsburg exceeded those at Lincoln by 2 to 5°, especially after June 7 (fig. 38). The night temperatures at Burlington were decidedly lowest (by 2 to 10°), those at Lincoln intermediate, and at Phillipsburg highest (fig. 39). A similar general relation, but one less marked and with overlappings, holds for the day temperatures. Soil-temperatures at a depth of 3 inches were obtained until the middle of June, after which the thermo¬ graph bulbs were buried at a depth of 12 inches and left for the remainder of 89 90 EXPERIMENTS DURING 1922. the season. In figure 40 it may be seen that the moister soil at Lincoln was also that with the lowest temperature, although the differences, as in the case Table 31 .—Holard in excess of the hygroscopic coefficient at the several stations, 1922. Lincoln, Nebraska. Date. 0 to 0.5 0.5 to 1 1 to 2 2 to 3 3 to 4 foot. foot. feet. feet. feet. Apr. 30. 15.3 15.8 14.7 10.9 9.1 May 10. 10.1 13.7 15.3 . . . . .... May 17. 12.0 13.1 14.1 12.1 7.8 June 7. 6.4 11.2 10.9 . . . . . . . . June 14. 6.8 11.3 10.5 8.4 6.6 June 22. 3.5 6.3 11.1 .... .... July 6. 5.7 4.7 8.9 8.0 5.4 July 13. 17.8 11.7 8.1 . . , . , . . . July 20. 16.4 10.7 6.2 .... .... July 27. 21.2 14.7 7.5 . , . . . • • • Aug. 3. 8.2 10.1 7.4 8.0 6.0 Aug. 10. 5.6 6.2 7.4 .... . . . . Aug. 17. 0.9 1.8 0.8 . . . , .... Aug. 24. 3.2 5.4 5.1 7.1 5.3 Hygroscopic coeff.... 9.8 10.9 10.1 10.0 10.3 Phillipsburg, Kansas. Apr. 28. 18.8 19.4 14.1 1.5 1.3 May 19. 5.4 10.2 9.3 6.2 0.4 June 11. 1.7 9.5 7.8 8.2 0.5 June 17. 1.4 4.4 .... .... . . . . June 24. -1.6 0.9 .... .... . . . . June 30. 10.8 1.1 3.7 5.1 4.7 July 17. -1.7 0.2 -0.5 -0.2 0.9 July 28. -0.1 1.6 , . . . . . . . . . , . Aug. 3 . 0.5 0.7 -0.5 -0.1 0.0 Aug. 11. -1.5 -1.8 -1.3 .... . . . . Aug. 17. -3.7 -1.5 .... . , . . . . , . Aug. 27. -2.6 -1.4 0.7 -1.7 0.2 Hygroscopic coeff.... 10.6 10.6 10.9 10.6 10.7 Burlington, Colorado. Apr. 30 . 17.6 16.9 8.5 -1.0 -1.0 May 20. 19.8 9.6 8.7 6.8 -0.4 June 4. 12.2 10.3 8.3 6.8 . . . . June 11. -0.1 1.2 4.5 4.9 -1.1 June 18. 0.4 1.1 3.8 .... .... June 25 . 0.2 0.0 1.2 2.1 -0.2 July 2. 11.1 -1.8 -2.6 -0.6 -1.3 July 16. -2.5 -1.9 -0.5 0.0 -1.0 July 23. 6.7 -1.1 -0.1 1.1 . . . . July 30. -0.7 -0.7 -0.2 . . . . . . . . Aug. 4. 5.4 -1.5 -1.9 -0.4 0.6 Aug. 20. -1.7 -1.9 -1.1 -0.2 . . . . Aug. 26. -0.3 -2.1 -2.7 -1.9 -0.4 Hygroscopic coeff.... 10.9 10.9 12.2 12.0 11.4 of air-temperatures, are not great enough to be of much importance in the establishment and growth of native vegetation. PHYSICAL FACTORS. 91 Humidity. Continuous records of humidity were not obtained during 1922, but a comparison of the hygrograph records from Burlington and Lincoln, from which the average day and average night humidities have been obtained, shows striking differences. With few exceptions, the air at the former was Fig. 38.—Average daily temperatures at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1922. 13 to 25 per cent drier by day (table 32). Owing to the high altitude (4,160 feet) and cool nights at Burlington, differences in night humidites were less marked. Frequent isolated readings at Phillipsburg, when compared with the records at the other stations, showed that they were usually intermediate. Table 32 .—Average day and night humidites at Lincoln and Burlington, 1922. Date Average day humidities. Average night humidities. Lincoln. Burlington. Lincoln. Burlington. p. ct. p. ct. p. ct. p. ct. May 1 to 7. 44.7 45.5 73.5 77.3 8 14. 50.5 26.2 77.5 58.2 15 21. 44.5 39.7 70.9 74.5 22 28. 77.4 64.5 89.0 87.9 29 to June 4. 64.2 • • • • 90.2 • . » • June 5 to 11. 58.3 45.8 87.5 75.4 12 18. 60.6 40.5 86.6 74.2 19 25. 48.8 37.2 73.2 73.0 26 to July 2. 61.4 • • • • 85.3 • • • • July 3 to 9. 60.8 51.0 78.5 82.2 10 16. 71.0 • • • • 93.0 • • • • 17 23. 70.6 44.5 89.4 78.5 24 30. 72.2 46.1 92.3 78.3 31 to Aug. 6. 66.0 .... 90.2 • . . • Aug. 7 to 13. 55.3 61.0 85.3 90.6 14 20. 54.0 42.8 81.0 62.1 21 27. 50.2 • • • • 80.0 • • • • 28 30. 56.3 .... 87.8 .... 92 EXPERIMENTS DURING 1922. Evaporation. The average daily evaporation (fig. 41) shows again the relative xerophytism of the several communities, as well as the periods of stress in June and August, although that at the two western stations remained rather high throughout the growing-season. The evaporation at Lincoln ranged from 9 to 33 c. c., at Phillipsburg 8 to 41 c. c., and at Burlington from 15 to 47 c. c. In general, the losses were higher than for the preceding year (fig. 23). Fig. 39.—Average day (heavy lines) and night temperatures (light lines), at Lincoln (solid line), Phil¬ lipsburg (long broken lines) and Burlington (short broken lines), 1922. To summarize, the season of 1922 was less favorable for growth than 1921, and deficiencies in rainfall and soil moisture were marked, being pronounced even at Lincoln, the least xerophytic of the stations, during late summer. Humidity and holard were again the controlling factors in plant growth, conditions as regards both being progressively more severe westward. PLANTING RESULTS. Surface Sowing. Twenty-two species of grasses, forbs, and trees were planted on the sur¬ face of the high prairie at Lincoln on April 18; 82 per cent of the species ger¬ minated, only Aster muitiflorus, Bouteloua gracilis , Solidago missouriensis, and Sporobolus asper failing to do so. However, germination was considerably delayed because of dry weather. All of the species did at least fairly well until July, Lespedeza, Liatris spp., Onagra, and Pinus having succumbed by the 14th, while 4 other species were suffering badly. The drought period, SURFACE AND TRENCH SOWING. 93 May 26 to June 24, was quite severe. During the latter part the long-establish¬ ed native grasses on the high prairie had rolled leaves and Erigeron ramosus was drying, many without flowering. It seems probable that most, if not all, of the surface-sown species would have succumbed except for regular watering. Late August found Bouteloua hirsuta, Kuhnia, Ratibida, and Robinia added to the mortality list. Thus, the mortality during the first summer was 50 per cent, notwithstanding regular watering at critical periods throughout the season. Liatris scariosa died in August 1922, and Bouteloua hirsuta did very poorly, but the 5 remaining species flourished, except during the late- summer drought, Aristida purpurea, Bouteloua racemosa, and Elymus cana¬ densis producing seed. Andropogon nutans, Bouteloua hirsuta, and Bouteloua racemosa died during the dry fall and winter. During the wet summer of 1923 the shade was very dense. At the end of the season Elymus canadensis was represented by a single weak plant; Liatris punctata, which was growing in a dense sod, was over a foot tall, but failed to blossom; but Aristida purpurea, Fig. 40.—Average daily soil temperatures at depths of 3 and 12 inches respectively at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1922. the other survivor, developed normally and produced an abundance of seed. It seemed only a question of a year or two more, however, until all three would disappear from the stabilized grassland. Trench Sowing. Twenty-three of the same species sown on the surface were also planted in a trench on the high prairie at Lincoln on April 18. Seven, viz, Amorpha can- escens, Aster multifiorus, Bouteloua hirsuta, B. gracilis, Onagra biennis, Rati¬ bida columnaris, and Solidago missouriensis, failed to germinate. Of the 70 per cent germinating, Agropyrum glaucum and Liatris punctata died before the end of the June drought and many others wilted badly. Lespedeza capitata, Muhlenbergia pungens, Petalostemon candidus, and Sporobolus asper succumbed by the middle of July, while the remaining ten species (63 per cent) lived throughout the summer. Most of them, however, were in poor condition, being badly dried by August 21. 94 EXPERIMENTS DURING 1922. Plants in the trench at Phillipsburg did much more poorly. Of the 23 species planted on April 29 only 8 (35 per cent) germinated. One-half of the trench was rather densely shaded by Bouteloua racemosa and Andropogon furcatus, and germination in it was very low. Only about hah of those that germinated did so abundantly. Liatris scariosa and Sporobolus asper died by June 30, and Agropyrum glaucum in July, while Elymus canadensis, Des- modium canescens, and Kuhnia glulinosa succumbed in August. Andropogon nutans and Bouteloua racemosa were the only ones that survived the first summer. Although the prairie had been burned over the preceding winter through accident, by May 27 the general level of the grasses had reached 5 or 6 inches, while an upper story of Psoralea, Helianthus, Erigeron, and flower- stalks of Boa and Siipa at 15 to 22 inches added to the shade. Consequently, light early became an important factor. c.c 40 30 20 10 0 Fig. 41. —Average daily evaporation at Lincoln (solid line), Phillipsburg (long broken lines), and Burlington (short broken lines), 1922. Twenty-two species were sown on the surface of the mixed prairie at Phil¬ lipsburg on April 29, the spring being late at all stations. Of these, 14 per cent, including Aster multiflorus, Lespedeza capitata, and Solidago missouri - ensis, failed to germinate. By June 30 (no efficient rain having fallen dur¬ ing June 1 to 26), 4 forbs, Pinus, and Sporobolus had succumbed. On August 3, Agropyrum was found dead, and by the last of the month Amorpha, Aristida, Petalostemon, and Ratibida had also succumbed. This left 8 species (42 per cent) surviving, of which at least 4 were in very poor condition. Shading was an important factor. As early as June 30 the grass- level was 8 to 14 inches high, with subdominant herbs such as Psoralea at a level of 24 inches. Twenty of the same species grown at Lincoln and Phillipsburg were also sown on the surface of the short-grass sod at Burlington on April 30. Ten species (50 per cent) failed of germination. In fact, only Andropogon nutans, Robinia pseudacacia, Desmodium canescens, and Liatris scariosa germinated at all abundantly. Some species lay dormant until late in June, being stimulated to grow by a heavy shower. By the middle of July only one plant of Liatris scariosa remained alive, and it succumbed later. The follow¬ ing species alone survived in both the true and mixed prairies: Andropogon furcatus, A. nutans, Bouteloua racemosa, Desmodium canescens, and Elymus DENUDED QUADRATS. 95 canadensis. Shading played a minor role at Burlington, except in protecting the plants from too rapid water-loss, since the short-grass sod was only 3 inches tall by July 1. The fate of the surface sowing of 1921 is of interest here. At Lincoln 9 per cent survived, but none at Burlington, while, owing to opportune showers and to a much more favorable light relation at Phillips- burg than at Lincoln, 83 per cent survived the first season. Establishment and growth in the mixed prairie, however, were better than on the short- grass plains. Among 22 species planted in the trench at Burlington on April 30, 68 per cent failed of germination. However, 5 of the 6 species that grew germinated rather abundantly. Robinia pseudacacia , Desmodium canescens , and Kuhnia glutinosa died by the middle of July, and Agropyrum glaucum in August, while 5 plants of Andropogon nutans alone survived the summer. Andropogon nutans was the only species that survived at all three stations, reaching a height of 1 to 3 inches at Burlington, 5 to 8 inches at Phillipsburg, and 5 to 9 inches at Lincoln. Bouteloua racemosa survived at Phillipsburg and Lincoln, and 8 other species at Lincoln alone. As to the 1921 trench plantings, the best results were obtained at Lincoln, where 5 of the 11 species that germinated survived the first season. Bouteloua gracilis died the following June, but the rest made fair to good growth. During the favorable season of 1923 growth was excellent, Andropogon nutans and Sporobolus asper reaching heights of 8 to 10 inches, while Bouteloua hirsuta and Liairis scariosa were 5 to 7 inches tall, but none flowered. Of the 6 species that started growth at Phillipsburg, only Andropogon nutans and Aristida purpurea became permanently established. Both did quite well during 1922, but Andropogon , which was densely shaded, died in late summer. Aristida was represented in 1923 by a small clump which blossomed profusely at a height of 22 inches. No species planted in the trench in 1921 survived at Burlington. Denuded Quadkats. Thirty-five species of grasses, forbs, shrubs, and trees were planted in denuded quadrats in the true prairie on April 18. Among these, ten (29 per cent) failed to germinate. Brauneria pallida died early in June and Acer negundo before the June drought was broken. During this period many other species were wilted or drying. By the middle of July, Agropyrum, Elymus, and Ulmus had also died, while Andropogon halli and Calamovilfa longifolia were added to the mortality list by August 22. However, 18 spe¬ cies or 72 per cent survived the first summer, the list including trees and forbs as well as grasses. Many of these were seriously affected by drought. Of 32 species planted on April 29 in denuded quadrats in the mixed prairie, 10 (31 per cent) failed of germination. By June 10, Lespedeza and Liatris scariosa had died, and Agropyrum , Muhlenbergia, and Ulmus disappeared by the end of the month. Robinia dried out during July and Acer negundo early in August. The end of August found Desmodium , Elymus , Onagra, and Pinus added to the list of non-survivors. This left 11 species, 50 per cent of those which germinated, that survived the first summer. As usual, most of these showed by dead leaf-tips, rolled leaf-blades, or browning, the effects of intermittent periods of drought. 96 EXPERIMENTS DURING 1922. Twenty-eight species were sown in denuded quadrats at Burlington on April 30; 50 per cent of them failed of germination. Among these were Amorpha, Aster, Liatris, Ratibida, Redfieldia, and Solidago, all of which had also failed at Lincoln and Phillipsburg, notwithstanding they showed fair to good germination in the greenhouse as well as under other methods of planting. Seeds of certain species kept on germinating until July. By this time Acer saccharinum and Andropogon furcatus had died, but no further losses were recorded until August, when Agropyrum and Robinia also succumbed. In fact, except for the death of Aristida, no further loss of an entire species occurred, although many plants were eaten off by grass¬ hoppers, badly wilted, or entirely destroyed by drought. Thus, 64 per cent survived, showing at once the better conditions furnished by this method as compared with that of surface or trench sowing. The surface of the soil in these quadrats, as was true of those 2 or 3 years older, was always mellow; Bulbilis had started to invade the older ones from all sides, but the invasion was very incomplete, even on areas denuded for over 2 years. While the grama in the buffalo-grass sod was flowering rather sparsely at 6 to 12 inches, that around the edges of the quadrats, owing to an increased holard due to lessened competition, was 6 to 20 inches tall. Summary. Andropogon nutans, Bouteloua gracilis, Bouteloua racemosa, Gleditsia tria¬ canthus, Kuhnia glutinosa, and Panicum virgatum grew throughout the sum¬ mer at all three stations. The following survived the summer at two: Des- modium canescens, Muhlenbergia pungens (Burlington and Lincoln), and Acer saccharinum, Andropogon furcatus, Aristida purpurea, Petalostemon candidus, and Sporobolus asper (Phillipsburg and Lincoln). Andropogon scoparius, Liatris punctata, Onagra biennis, Pinus ponderosa, and Robinia pseudacacia survived only at Lincoln. A comparison of the results shows that Bouteloua gracilis made the best growth at Burlington (5 to 13 inches) and the poorest at Lincoln (2 to 5 inches). Gleditsia grew poorest at Burlington (2 to 5 inches) and about the same at the other stations (3 to 7 inches). However, all of the others, Andropogon nutans, Bouteloua racemosa, Kuhnia glutinosa, and Panicum virgatum, showed the best growth at Lincoln (4 to 12 inches), intermediate at Phillipsburg (3 to 9 inches), and least at Burlington (1 to 5 inches). The fate of the species planted in quadrats in 1921 is of interest here. At Lincoln, 13 of the 20 species that germinated survived the first season, but the following spring 7 of these failed to come up. The 6 remaining species did very well throughout the next summer, but none came into blos¬ som. In 1923, Andropogon nutans formed a dense sod; Gleditsia was repre¬ sented by a single remnant, and Stipa spartea by a scattered growth. The other species did fairly well, but all were much shaded and none produced seed. At Phillipsburg, 19 species germinated and 13 survived the first sum¬ mer; 6 of these, including 4 species of trees, were winterkilled, and Stipa setigera died by the end of 1922. Agropyrum and both species of Sym- phoricarpus did but poorly; Andropogon nutans formed a dense sod, while Aristida and Elymus both ripened seed. Of the 19 species which germinated SEEDLING TRANSPLANTS. 97 at Burlington in 1921, only 4 survived the first summer. Gleditsia was winter- killed and Calamovilfa was accidentally destroyed. Andropogon nutans grew well during both 1922 and 1923 and reached a height of 6 to 9 inches, but Bul- bilis had invaded rather extensively. Stipa viridula died during the severe winter of 1922-23. Seedling Transplants. Seedlings of 20 species were transplanted into the high prairie on May 17 and watered freely from time to time, as weather conditions demanded, until August 23. Notwithstanding this aid to establishment, Aristida purpurea and Petalostemon candidus died early in June, Agropyrum glaucum and Solidago missouriensis in July, and Psoralea tenuiflora flori- bunda and Stipa viridula following the drought in August. Several other species were represented by mere remnants, 14 (70 per cent) surviving the summer. As sometimes happens under cultivation, Onagra biennis grew a flower-stalk the first season, but failed to blossom and died in the fall. On May 19, seedlings of 19 species were transplanted into the mixed prairie at Phillipsburg. As at the other stations, they were about 15 days old. In planting, water was placed in the bottom of the trench and after it had settled down, moist soil was added around each root-mass until the trench was two- thirds filled, w T hen a second watering was given. The trench was then filled with drier soil and finally covered with a dry mulch. At all stations the plants were watered for about a week thereafter. Agropyrum, Bouteloua racemosa , and Stipa viridula died during the June drought, and Psoralea by the first of August, but all the others made a good growth, the grasses tillering freely and Bouteloua gracilis and B. hirsuta blossoming. The mortality was only 22 per cent. Seedlings at Burlington did very poorly. Of the 19 species planted here on May 20, only 5 survived, 74 per cent succumbing to the unfavorable condi¬ tions of the short-grass plains; 1 died early in June, 5 more by July 2, 7 others before the first week in August, and another later in the same month. On August 26, wilted or half-dead clumps of Bouteloua racemosa, Stipa viridula, and Sporobolus asper, and better lots of Bouteloua hirsuta and B. gracilis, 2 to 3 inches tall, alone remained. The history was one of repeated wilting and revival only to wilt again, the conditions causing transpiration at this station often being so severe that vegetation wilted even in soil of good water-content. Under such conditions growth was poor. Only three spe¬ cies survived at all the stations, viz, Bouteloua gracilis, B. hirsuta, and Sporo¬ bolus asper, even under this favorable method of transplanting. As to the 1921 seedlings, 6 of the 13 species transplanted into high prairie at Lincoln died the first summer, while 2 more were winterkilled (table 61). The 5 remaining species grew throughout 1922, all surviving the drought. However, Andropogon nutans and Stipa spartea died the following winter, leaving only the three species of Bouteloua. All these made a good growth in 1923, forming five clumps 5 to 8 inches tall. At Phillipsburg, 9 of the 12 species survived the first season and 1 died the following winter (table 62). At the end of 1922 Liatris scariosa had died and Agropyrum and Andropogon nutans were doing very poorly. The rest made a good growth, Bouteloua hirsuta alone producing flower-stalks. Andro- 98 EXPERIMENTS DURING 1922. yog on nutans and Agropyrum died the following winter. The two shorter gramas both made a good growth in 1923 and produced seed. Andropogon furcatus was densely shaded and did only fairly well; Stipa spartea was repre¬ sented by a few narrow, densely shaded leaves only, while Liatris punctata had 5 fine plants about a foot high that blossomed profusely. None of the 1921 seedlings at Burlington survived the first season. Summary. The average germination under all methods of planting was similar to that of the two preceding years, Lincoln being highest (74 per cent), the mixed- prairie station second (63 per cent), and the short-grass plains last (44 per cent). However, as regards surface sowing, Phillipsburg ranks slightly ahead of Lincoln, and this in spite of frequent watering at the latter station. The explanation for this seems to lie in soil structure, that at Phillipsburg being much mellower, and hence it does not crust when alternately wet and dry. Table 33. —Summary of planting experiments, 1922. Method of planting. Per cent of germination. Per cent of establishment of germinated species. Lincoln. Phillips¬ burg. Burling¬ ton. Lincoln. Phillips¬ burg. Burling¬ ton. Surface sowing. 82 86 50 50 42 0 Trench. 70 35 32 63 25 14 Denuded quadrats.. . 71 69 50 72 50 64 Average. 74 63 44 62 39 26 Seedlings. 70 78 26 This probably accounts for the 8 per cent margin in favor of the seedling trans¬ plants, since in all cases the percentage of germination falls off to the west¬ ward. Sixty-two per cent of the species that germinated at Lincoln became estab¬ lished, 39 per cent at Phillipsburg, and only 26 per cent at Burlington. The same sequence held for all methods of planting, except that establishment in the quadrats at Burlington exceeded by 14 per cent that at Phillipsburg. Both the lowest percentage of germination and establishment occurred in the trench, the denuded quadrats ranking highest in percentage of estab¬ lishment. Sod Transplants. Large blocks of soil containing 26 species of grasses and forbs were transplanted into high prairie at Lincoln March 22 to April 5. All the transplants made a good growth except Psoralea tenuiflora floribunda, which died before the middle of June, probably because of the cutting of its deep tap-root, Silphium integrifolium, which succumbed by the end of August, and Andropogon scoparius which grew poorly. In addition to 10 of the grasses, the following forbs blossomed and set seed: Anemone cylindrical Brauneria pallida, Grindelia squarrosa, Liatris scariosa, Solidago missouriensis, S. rigida, and Vernonia fasciculata; only 7 per cent of the entire lot died. EDAPHIC STATIONS. 99 Eighteen blocks of sod, representing 13 species of grasses, were transplanted from Lincoln into the short-grass sod at Burlington on April 15. Although none died, the growth was with few exceptions rather poor, being characterized by scattered shoots, frequently with rolled leaves and dead leaf-tips. Flowers were present in 61 per cent as compared with 67 per cent of the grasses transplanted at Lincoln. In all cases, however, the flower- stalks were 3 to 10 inches shorter and usually few in number, while the inflorescence itself was dwarfed. For example, Agropyrum headed at Bur¬ lington at 11 inches, while at Lincoln the flower-stalks were 21 to 27 inches tall; Elymus headed at 12 to 18 inches in the Great Plains, but in the true prairie at 28 inches. The vegetative growth was proportionately dwarfed. EXPERIMENTS AT OTHER STATIONS, 1922. Physical Factors. Water Relations. Studies were continued during 1922 at the series of edaphic stations at Lincoln, and at Nebraska City and Colorado Springs. The general conditions of precipitation at Lincoln have already been given. The season at Nebraska City was one marked by June and August drought. April precipitation was nearly normal, May showed a deficiency of 1 inch, June of 2.6 inches, and, although July had an excess of 3 inches, August had a rainfall of only 0.7 inch instead of the normal 3 inches (fig. 18). Drought periods occurred on April 11 to 30, May 6 to 20, May 25 to June 10, and June 11 to 25. The water-content in excess of the hygroscopic coefficient at Nebraska City and Lincoln low prairie is given in table 34, where that of the high prairie is added for purposes of comparison. A study of the table shows that notwithstanding the decreased rainfall, the soil at Nebraska City usually had a margin of at least 8 per cent chresard below the first foot of soil. Exceptions to this occurred after the first week in June and again in August. However, the chresard in the surface foot, which most critically affects seedlings, was practically exhausted on June 17, and was very low on August 17, as well as at certain other periods. As a whole, the holard was less favorable than during 1921. Water-content on the low prairie was more favorable than that at Nebraska City and considerably in excess of that on the high prairie. At no time was there a margin of less than 7 per cent available at any level, and it was usually 10 to 12 per cent. Conditions on the gravel-knoll were much less favorable than during the preceding year, as was evidenced by the drying of the grama grasses during a drought when no water was available in the first 2 feet of soil. The average daily evaporation at Nebraska City was considerably lower than on the high prairie at Lincoln. During the 4 months (last half of May, June, July, and August) the average daily rates were as follows: 12.4 and 15.1, 18.4 and 29, 12.7 and 13.5, 14.4 and 20.2 c. c. respectively. Isolated humidity readings taken from time to time and compared with the hygrograph record at Lincoln for the same hour showed conditions not greatly different from from those of the preceding year. 100 EXPERIMENTS DURING 1922 Table 34. —Holard in excess of hygroscopic coefficient in 1922. High Prairie. Date. 0 to 0.5 foot. 0.5 to 1 foot. 1 to 2 feet. 2 to 3 feet. 3 to 4 feet. Apr. 30. 15.3 15.8 14.7 10.9 9.1 May 10. . 10.1 13.7 15.3 May 17. 20.0 13.1 14.1 12.1 7.8 June 7. 6.4 11.2 10.9 June 14. 6.8 11.3 10.5 8.4 6.6 June 22. 3.5 6.3 11.1 July 6. 5.7 4.7 8.9 8.0 5.4 July 13. 17.8 11.7 8.1 July 20. 16.4 10.7 6.2 July 27. 21.2 14.7 7.5 Aug. 3. 8.2 10.1 7.4 8.0 6.0 Aug. 10. 5.6 6.2 7.4 Aug. 17. 0.9 1.8 0.8 Aug. 24. 3.2 5.4 5.1 7.1 5.3 Hygroscopic coeff.... 9.8 10.9 10.1 10.0 10.3 Low Prairie. Apr. 30. 18.4 10.4 20.6 12.5 18.2 May 17. 13.4 16.1 17.0 17.0 17.6 May 31. 26.7 21.3 18.7 June 7. 13.2 14.6 15.2 June 14. 10.7 13.8 15.1 13.8 19.0 June 22. 6.9 9.4 15.4 June 29. 21.9 10.2 11.0 July 6. 13.4 8.0 6.8 12.8 16.6 July 13. 28.1 16.8 11.5 July 20. 26.3 21.1 11.2 July 27. 30.6 28.7 15.3 Aug. 3. 17.5 14.6 11.8 11.1 15.9 Aug. 10. 10.2 15.8 13.7 Aug. 17. 7.9 9.8 11.0 Aug. 24. 7.8 6.8 11.1 11.8 15.9 Hygroscopic coeff.... 10.0 9.6 9.2 11.1 10.8 Nebraska City. Apr. 22. 17.9 19.3 18.3 15.0 13.3 May 13. 8.7 16.2 17.4 14.4 13.2 May 20. 4.4 11.0 16.1 June 3. 18.4 18.8 17.0 15.0 14.3 June 10. 4.3 9.4 12.3 June 17. -0.3 1.5 7.9 11.2 June 21. 2.0 4.2 10.4 10.3 10.2 June 28. 4.4 0.8 5.0 12.2 11.8 July 5. 8.5 0.9 5.9 July 12. 25.6 21.6 17.7 10.2 10.3 July 19. 7.6 13.4 9.9 July 26. 5.5 10.3 9.1 Aug. 1. 18.2 16.3 13.1 9.9 9.5 Aug. 8. 11.9 12.7 10.4 Aug. 15. 1.5 6.1 7.8 Aug. 22. 11.2 7.1 4.1 Aug. 31. 3.7 4.8 6.8 8.0 3.4 Hygroscopic coeff.... 12.1 11.7 12.3 13.6 12.9 EDAPHIC STATIONS! SURFACE SOWING. 101 Planting Results. Surface Sowing. Of typical grasses and forbs, 19 species were sown on the surface of the low prairie on April 20, 1922; unlike the high prairie and gravel-knoll, this area had not been burned. A dense tangle of grass-leaves and flower-stalks covered the ground and formed a loose mulch to a depth varying from 2 to 4 inches. This not only greatly affected the light relations, but also the air was of much higher humidity and lower temperature below the mulch, and the soil was cooler. For example, at 2 p. m. on a clear day (May 28) an average of four thermometer readings at a depth of an inch under the mulch gave a temperature of 68° F., while that at a similar depth in denuded quadrats was 95° F. Under these conditions the plants germinated slowly, Desmodium canescens, Liatris punctata and scariosa, and Petalostemon candidus alone appearing above ground by May 17. Moreover, by May 26 the new growth of the dominants was 10 to 15 inches tall, so that light relations played an exceedingly important role from the first. However, only 3 species, Aster multiflorus, Ratibida colum- naris, and Solidago missouriensis, failed to germinate. Weak plants of Amorpha canescens, Onagra biennis, and Sporobolus asper lasted only until the middle of June, and Kuhnia glutinosa, Liatris scariosa and punctata, and Lespedeza capitata died within the next 30 days, all showing the effects of shading. Bouteloua hirsuta succumbed by August and B. gracilis, B. racemosa, Desmodium canescens, and Elymus canadensis during this month. This left remnants of Aristida purpurea and Andropogon nutans, as well as rather good growths of Andropogon furcatus and Petalostemon candidus . The total mor¬ tality was 75 per cent during the first summer. The severe competition for light on the low prairie was chiefly responsible for this; by June 7 the surface and trench plantings were already badly shaded, while by June 22 the general grass level was 18 inches; by September 1, that of the foliage reached over 2 feet and the flower-stalks 43 inches. In the part of the area unmown since 1919, growth was even greater. On the surface at Nebraska City, 20 species were sown on April 22. Spring opened late and the soil was in excellent condition as regards holard. Nine species were above ground by May 13. Aster multiflorus, Bouteloua gracilis, Lespedeza capitata, Onagra biennis, Solidago missouriensis, and Sporobolus cryptandrus failed to germinate. By June 21, most of the plants were badly wilted and Liatris scariosa and Pinus ponderosa had succumbed. Robinia pseudacacia died in July, Agropyrum glaucum and Petalostemon candidus by August, while Amorpha canescens and Ratibida columnaris were found dead on August 31. Only 50 per cent survived. These included slender, delicate specimens of Andropogon nutans, A. furcatus, Bouteloua hirsuta , B. racemosa, Elymus canadensis, Kuhnia glutinosa, and Liatris punctata. A similar lot of seeds was sown on the surface at Colorado Springs on May 5, but these, like those planted in the trench and denuded quadrats, showed no germination when examined on May 25 and again on July 4, owing to unfavorable holard. As regards the 1921 surface sowings, it may be recalled that none survived, owing to the dense shade. At Nebraska City, Bouteloua gracilis and B. hirsuta alone were represented by delicate seedlings at the end of the first 102 EXPERIMENTS DURING 1922. season. However, they held out during the summer of 1922, although very much attenuated. During 1922 they were still very slender and delicate, but they survived this third season of growth, although densely shaded, indicating an unexpectedly high tolerance of shade for the short-grasses. Trench Sowing. Seeds of 20 species were planted in a trench on the low prairie on April 20. Amorpha canescens, Aster multiflorus, Onagra biennis, Ratibida columnaris, Sporobolus asper, and Solidago missouriensis did not germinate. Nearly all of the others grew in considerable abundance, except Bouteloua hirsuta and Liatris punctata, which were represented by a few plants each, all dying before the end of June. The others did quite well during the rainy month of July, Liatris scariosa, Lespedeza capitata, and Muhlenbergia pungens disappearing by August. Nine species (64 per cent of those that germinated) became permanently established, and although slender were in fairly good condition at the end of the summer. A similar lot of seeds, representing 21 species, was planted in a trench in subclimax prairie on April 22. Eight species, including 3 which grew at Lincoln, failed of germination. Agropyrum glaucum, Muhlenbergia pungens, Petalostemon candidus, and Pinus ponderosa died before June 20, the last having been dug up by rodents. Andropogon nutans, A. furcatus, Liatris scariosa and punctata died before the middle of July, and Elymus canadensis in August. This left a few weak plants each of Bouteloua racemosa, Kuhnia glutinosa, and Robinia pseudacacia, but a good sod of Andropogon nutans 7 to 9 inches high. Of the 1921 trench plantings on the low prairie, remnants of Aristida purpurea alone survived the first season, and these were accidentally destroyed the next spring in transplanting. At Nebraska City, Andropogon nutans alone was alive the following spring. In 1922 it reached a height of 13 inches, and did quite as well the next season, but it failed to blossom. On the gravel- . knoll, Andropogon nutans alone survived, but although it grew well early the next spring, it died the following June. Denuded Quadrats. On April 20, 27 species were sown in denuded quadrats on the low prairie. Five did not germinate. These belonged to different species from those failing of germination in the trench, and included Brauneria pallida, Corylus americana, Fraxinus lanceolata, Gleditsia triacanthus, and Ratibida columnaris. During the June drought, Agropyrum, Aristida, Aster, Lespedeza, Liatris punctata, and Solidago succumbed; Elymus, Muhlenbergia, and Ulmus also died before August 1. This gave a mortality of 40 per cent, though most of the survivors were in good condition at the end of the summer. Among 32 species planted at Nebraska City, 10 did not germinate. Muhl¬ enbergia and Petalostemon died before June and the following as a result of the June drought: Acer saccharinum, Aristida, Liatris scariosa and punctata, Ratibida, and Ulmus. Later, Agropyrum and Lespedeza were lost, giving a total mortality of 46 per cent. Among the survivors were 4 species of trees (two of which, Acer negundo and Gleditsia triacanthus, were in fine condition), 2 forbs, and 6 grasses. In general, they were in good condition. By the EDAPHIC STATIONS! SEEDLING TRANSPLANTS. 103 end of August the average height-level of the grassy vegetation was 16 inches, shade being dense over most of the quadrats. However, very little invasion had occurred. As to the denuded quadrats planted in 1921, 10 of the 12 species that grew on the low prairie survived the first summer. Acer negundo, A. saccharinum, and Lespedeza capitata failed to grow the following spring, and Elymus died in June, as did also Bouteloua gracilis, perhaps as a result of the surface soil washing away from the roots. All of the others except Corylus ameri- cana made a good growth, the latter being badly eaten by grasshoppers. All were densely shaded and none flowered. Corylus did not survive the dry fall and winter following, while Aristida died during the summer of 1923, evidently being unable to longer endure the dense shade. At the end of the summer Sporobolus was represented by remnants only. Symphoricarpus made only a fair growth, while Andropogon nutans alone developed normally, reached a height of 25 to 30 inches and merged into the native sod. At Nebraska City only 8 of the 15 species that grew survived the first summer. Gleditsia and Robinia failed to appear in the spring of 1922 and Calamovilfa died early in June. All of the other species made a good growth, Andropogon merging into the native sod. None of the quadrats were much invaded. No flower-stalks were produced by any of the plants. In 1923 the 6 remaining species all did quite well, several of them seeding rather abun¬ dantly. On the gravel-knoll, 40 per cent of the species that grew survived the first season. During 1922 some of the Lespedeza seeds, sown the preceding spring, germinated, but these and the few feeble shoots from Symphori¬ carpus rhizomes were all dead by midsummer. Aristida and Elymus both seeded, but, like Andropogon, suffered severely from the August drought. During the following dry fall and winter Elymus died, but in 1923 both Aris¬ tida and Andropogon made a good growth and seeded. Seedling Transplants. Seedlings were transplanted both on the low prairie and at Nebraska City, 22 pots containing 18 species being used in growing the plants for each station. The transplanting was done at Lincoln on May 17. One lot of Agropyrum died in June and another in July, together with Elymus, Koeleria, Liatris, and Stipa spartea; Ratibida, Kuhnia, Argemone, Onagra, and Psoralea succumbed in August, making a total loss of 55 per cent. Most of the survivors came through the season in good condition, owing to the adequate water-content even during the August drought. At Nebraska City, Agropyrum, Koeleria, Lespedeza, Solidago, and Stipa viri- dula died before the end of June. Three others failed in July, viz, Aristida, Glycyrrhiza, and Liatris, and August added Psoralea to the list, making a total of 41 per cent. Most of the survivors grew well, Bouteloua hirsuia putting forth flower-stalks. To complete the history of the 1921 seedling transplants, only 14 per cent survived on the gravel-knoll, constituted by Bouteloua hirsuta and Andro¬ pogon nutans. During 1922 these made only a poor vegetative growth, fre¬ quently wilting, and did not put forth flower-stalks. By August 1923 they had reached a height of 8 inches, neither coming into blossom. 104 EXPERIMENTS DURING 1922. On the low prairie a single species died the first season, but by the next spring Aristida purpurea and Liatris scariosa were dead and Bouteloua race- mosa was represented by a single plant which died in June. The rest made a fair to excellent growth, except Stipa comata. The shade was dense and Elymus canadensis alone grew above the general level and headed at 35 inches. No others blossomed. During the summer of 1923 the following species were so deeply shaded that they died: Bouteloua hirsuta, B. gracilis, Liatris punctata, Stipa viridula, and S. comata. Elymus headed at about 3 feet and Stipa spartea seeded at 2 feet, but neither Andropogon nutans, which did poorly, nor A. furcatus showed signs of flower-stalks by the last of August. Table 35. —Growth of species in cultivated soil. Species. Degree of development during first season. Agropyrum glaucum. Good growth of foliage; maximum 2 feet; no flower-stalks. Andropogon furcatus. Fine bunches; foliage 18 in.; a few flower-stalks 36 to 38 in. halli. Foliage 16 in.; flower-stalks 30 in. (cf. plate 12 c). nutans. Excellent growth; 20 in., seeding abundantly at 3 to 4.5 feet. Aristida purpurea. Large bunches; foliage 11 in.; flower-stalks 18 in. Bouteloua gracilis. Dense growth; leaves 14 in.; flowering profusely at 15 to 22 in. hirsuta. Leaves 10 in.; flower-stalks appearing Aug. 8; later 12 in. tall; abundant. racemosa. Good growth; ave. height 13 in.; flower-stalks 3 to 3.5 feet. Calamovilfa longifolia. . .. Foliage 16 in.; no flower-stalks. Desmodium canescens.... Flower-stalk appearing July 10; fruited abundantly at 5 feet (cf. plate 14 b). Elymus canadensis. Foliage 2 feet; flower-stalks 4 to 5 feet; very large heads. Kuhnia glutinosa. 34 in. tall; fruiting abundantly. Lespedeza capitata. Fine plants, 20 to 24 in.; seeded abundantly. Liatris scariosa. 3 to 6 in.; many leaves per plant. Muhlenbergia pungens.... 9 in.; flowered and seeded profusely at 14 to 16 in. Onagra biennis. One rosette 28 in. in diameter; many smaller ones; many had flower-stalks 5 to 5.5 feet tall; blossomed. Panicum virgatum. Average height foliage 29 in.; seeded abundantly at 3 to 3.5 feet. Petalostemon candidus. . . Excellent growth; very much bunched stems, 8 to 12 in. tall. Pinus ponderosa. Five plants, 4 to 5 in. tall. Robinia pseudacacia. Stems nearly an inch in diameter and 7.5 feet tall. Sporobolus asper. 9 in. tall; flower-stalks 14 to 16 in.; seeded profusely. Stipa viridula. Foliage about 16 in. tail; maximum 26 in.; no flower-stalks. At Nebraska City, 6 of the 13 species of seedlings died the first summer; Regardless of this heavy mortality, all the species survived the summer of 1922, notwithstanding the drought. However, all suffered more or less severely in late summer, Stipa comata being represented by remnants only at the end of the season. Bouteloua hirsuta and B. gracilis alone blossomed. The seedling area, although quite shaded by the adjacent vegetation, was not much invaded. In 1923, Stipa comata became badly invaded and died. All the others made a good growth. Effect of Competition. In order to determine the effects of competition upon the development of the species investigated, many of them were grown in a cultivated area at Lincoln, kept free from weeds by hoeing, and watered from time to time (table 35). EDAPHIC STATIONS! SOD TRANSPLANTS. 105 The fine growth made in a single season by these species is in striking con¬ trast to that under even the most favorable method of planting in the grass¬ land and emphasizes the striking effect that competition plays in limiting plant development under natural conditions. Plants growing for a second year under cultivation made no less remarkable growth. Bouteloua gracilis had an average height of foliage of 19 inches, spread 15 inches on either side of the trench, and some of the flower-stalks were 32 inches tall. Agropyrum glaucum was 29 inches tall, with a maximum height of 44 inches, and had spread 6 feet by rhizome propagation. A single plant of Onagra biennis had 3 stems which reached a height of 7 feet and 12 others of lesser height, all from the same rosette (plate 14). Summary. A summary of planting experiments is given in table 36, where for purposes of comparison the high prairie is included. The germination of surface- sown seed was not greatly different at the three stations (70 to 84 per cent), Table 36. —Summary of planting experiments , 1922. Method of seeding. Per cent of germination. Per cent of establishment of germinated species. High prairie. Low prairie. Nebraska City. High prairie. Low prairie. Nebraska City. Surface sowing. 82 84 70 50 25 50 Trench. 70 60 62 63 64 33 Denuded quadrat.... 71 81 69 72 60 54 Average. 74 75 67 62 50 46 Seedlings. 70 r 45 59 that on the low prairie being highest. Germination in denuded quadrats was also about 10 per cent greater on the low prairie than elsewhere, Nebraska City ranking third. Averaging the results of the three methods, 75 per cent germination occurred on low prairie, 1 per cent less on high prairie, and 8 per cent less at Nebraska City. Because of the effect of the luxuriant vegetation of the low prairie in reduc¬ ing the light intensity, a factor which quite outweighed the higher water- content, establishment averaged lower here (50 per cent) than on high prairie (62 per cent) and was least at Nebraska City (46 per cent). The last fact can be directly correlated with drought, especially when it is recalled that the surface sowing at Lincoln was watered. Similar conditions hold for the seedling trans¬ plants. A survey of the data shows clearly that while surface sowing gave the greatest per cent of germination, the best establishment occurred in the denuded quadrats. Compared with the western stations, both the per cent of germination and establishment averaged higher. Sod Transplants. Twenty-eight blocks of soil, comprising 25 species, were transplanted from the several habitats about Lincoln into the low prairie on March 22. These included 11 subdominant forbs, as well as all of the most important grasses. 106 EXPERIMENTS DURING 1922. Agropyrum glaucum grew poorly, not attaining a height greater than 11 inches and failing to flower. Astragalus crassicarpus, Brauneria pallida, and Psoralea tenuiflora floribunda, all provided with strong tap-roots neces¬ sarily cut off in transplanting, languished and died within a few weeks. All of the rest, however, came through the season in good condition, 75 per cent of the dicotyls and 50 per cent of the grasses flowering. A lot of sods, including 11 species from Arizona, were transplanted into low prairie early in May. The blocks were very small, being only 3 or 4 inches in diameter, and it seems probable some were dead when transplanted. In spite of sufficient watering the following did not grow: Aristida divaricata, A. purpurea, Bouteloua eriopoda, B. hirsuta, B. racemosa, Hilaria mutica, H. cen¬ chroides, Sporobolus wrighti, and Valota saccharata. Andropogon saccha- roides put forth 3 green shoots, which reached a height of 11 inches before they died in August. Bouteloua bromoides and Hilaria cenchroides formed good clumps 5 to 8 inches high, the former bearing flower-stalks 8 to 13 inches tall. However, both were winterkilled, notwithstanding the fact that they were well covered with a mulch of dead grasses. The following species from Arizona were also transplanted into high prairie at the same time: Aristida divaricata, A. purpurea, Bouteloua eriopoda , B. bromoides, B. hirsuta (all of which died promptly), and Hilaria cenchroides . The last made a sparse growth, reaching a height of 8 to 10 inches in August, but died under a mulch during the following winter. Sod of 14 species of grasses from both high and low prairie were trans¬ planted into the salt-flat near the low-prairie station late in March. Andro¬ pogon nutans, A, furcatus, Panicum virgalum, and Sporobolus asper did very poorly, making only a sparse growth, yellowing and wilting from time to time. By June 22 not only this lot of transplants, but also those of previous years were in very poor condition. Large cracks occurred around most of the blocks of sods and some were almost dried out. Shading played rather an unim¬ portant role, since the height of the Distichlis was only 18 inches. Sporo¬ bolus died late in July, and the others made a fair to poor growth only. Nearly all were more or less dwarfed and of a light-green color, and the inflorescences were often reduced. In fact, only 5 species (36 per cent) seeded; these were Bidbilis dadyloides, Bouteloua gracilis, Elymus canaden¬ sis, Koeleria cristata, and Stipa spartea. On April 5, 14 species were transplanted into the former swamp area. Silting up since the building of a dam, as well as the enormous losses by transpiration from the rank growth of vegetation, had decreased the water- content. On June 7 the holard at any depth did not exceed 22 per cent, which was the amount in the surface 6 inches; by June 22 it had fallen to 15 per cent or less in the surface 2 feet. It seems evident that there was no excess moisture or deficient aeration and that light was the controlling factor. By May 6, Spartina was 8 to 10 inches tall, and it had reached a level of 32 inches by the middle of June. In August a rank growth of Spartina, Phalaris, etc., 4 feet tall, covered the area; light readings at this time gave values of 1.5 to 11 per cent. Andropogon furcatus, A. scoparius, and Panicum virgatum, al¬ though starting out well, were badly invaded by the swamp dominants and died by the middle of July, as did Andropogon nutans later in the summer. All suffered severely from shading, as was evidenced by the attenuated leaves EDAPHIC STATIONS: SOD TRANSPLANTS. 107 and slender flower-stalks of the 5 species that blossomed, viz, Agropyrum, Bulbilis, Elymus, Koeleria, and Stipa. The species of the following genera came through the season in fair to good condition: Agropyrum, Bulbilis, Elymus, Koeleria, Spartina, and Stipa. Bouteloua gracilis, B. hirsuta, B. racemosa, and Sporobolus were in poor condition by autumn. Duplicate blocks of sods of the following species were shipped from Lincoln and transplanted at Colorado Springs on March 28: Andropogon scoparius, A. furcatus, A. nutans, Elymus canadensis, Sporobolus asper, Stipa spartea, Koeleria cristata, Bouteloua racemosa, B. gracilis, B. hirsuta, Bulbilis dacty- loides, Panicum virgatum, Spartina cynosuroides, and Agropyrum glaucum. All were growing vigorously on May 21, but were considerably eaten by grasshoppers and doing rather poorly by July 4. A final check on August 24 showed that Bulbilis dactyloides alone had blossomed, and no flower-stalks were appearing on any of the other species. Sporobolus asper had died, and Elymus, Andropogon furcatus, and B. racemosa were eaten back to mere rem¬ nants. Triplicate sods of Stipa viridula, Muhlenbergia gracilis, M. gracillima, and Sporobolus cryptandrus were shipped from Colorado Springs and trans¬ planted in the cultivated plats at Lincoln on June 22. The sods were very small and the plants died in spite of ample watering. Large duplicate blocks of sods were shipped from Lincoln to Berkelej^, California, which has an average rainfall of 24 inches, occurring chiefly in the winter months, and a long dry season in summer and autumn. These were transplanted on December 4 (1921), the species employed being Andro¬ pogon scoparius, A. furcatus, A. nutans, Elymus canadensis, Stipa spartea, Koeleria cristata, Bouteloua racemosa, and Bulbilis dactyloides. The sods were set on a dry west exposure in a heavy clay soil covered chiefly with introduced weeds, such as Arena fatua, Brassica campestris, and Raphanus saiivus. Owing to an abundance of rain, conditions at the time of transplanting were very favorable for growth. The original vegetation was bunch-grass prairie composed mostly of Stipa setigera, as indicated by the relicts of this. Some attention was given the sods in 1921 and 1922, at first a little water, and then occasional removal of the competing weeds. Most of the grasses did not survive the first dry season, but Elymus canadensis made a good growth. Bulbilis dactyloides persisted for a time, and Poa pratensis spread at first, but succumbed when water was withheld. Even Elymus was unable to sur¬ vive the exceptionally dry spring and summer of 1923, and the plot was finally destroyed by the great fire of September 1923. The failure of the grasses to establish themselves was due chiefly to inability to pass the long dry seasons and to competition as a contributing factor, but the experiment was on too small a scale to justify the conclusion that none of the species could be established. A number of grasses from the Santa Rita Reserve near Tucson, Arizona, were planted under the same conditions at Berkeley. The sods were set April 3, 1921, and comprised Andropogon saccharoides, A. coniortus, Aristida divaricata, A. purpurea, Bouteloua bromoides, B. hirsuta, B. racemosa, B. roth- rocki, B. eriopoda, Hilaria cenchroides, H. mutica, and Sporobolus wrighti. All of these made some growth in 1921, but by the autumn of 1923 all had succumbed except the following: Andropogon saccharoides grew vigorously in 1922, reaching a height of 4 feet, flowered, and set seed abundantly; in 1923 108 EXPERIMENTS DURING 1922. the 3 plants were growing and flowering, with a slight tendency to increase but not spreading over the ground, and no seedlings were noticed; Hilaria mutica made an attempt to establish itself and 1 of the 3 tufts was still living; Aristida divaricata was in the same condition as the Hilaria; Bouteloua bro- moides made a good start and flowered in 1922, but disappeared during the long dry summer of 1923; Sporobolus wrighti grew each season, and although it did not spread, was holding its own, even against the weedy grasses with which it had to compete during the spring and early summer. It is instruc¬ tive to note that the grasses from the desert plains grew better than those from the true prairies, a fact readily explained by the greater similarity in climate and climax between Arizona and California. 5. EXPERIMENTS DURING 1923. DEVELOPMENT OF SEEDLINGS AND TRANSPLANTS. It remains only to trace the history of the 1922 plantings through the period from September 1922 to the end of the summer of 1923. I Physical Factors. Water Relations. The fall and winter of 1922 constituted a period of great drought. Some idea of its intensity may be gained by an examination of the rainfall data in table 37. Table 37 .—Precipitation from August 1922 to August 1923. Burlington. Phillipsburg. Lincoln. Total. Departure from mean. Total. Departure from mean. Total. Departure from mean. August. 3.04 + 0.50 1.13 -1.77 0.69 -3.02 September.... 0.16 -1.23 0.94 -1.58 2.12 -0.52 October. 0.18 -0.75 0.44 -1.04 1.42 -0.40 November.... 1.44 + 1.01 0.40 -0.28 2.62 + 1.77 December.... 0.00 -0.63 0.00 -0.82 0.06 -0.61 January. 0.00 -0.26 0.00 -0.44 0.14 -0.48 February. 0.10 -0.36 0.16 -0.65 0.66 -0.04 March. 0.36 -0.47 1.31 +0.52 1.77 +0.44 April. 1.66 -0.40 3.40 + 1.07 2.21 -0.56 May. No report. .... 7.86 +4.85 3.31 -0.94 June. 3.03 +0.26 4.94 + 1.03 5.25 +0.93 Julv. 4.33 + 1.56 2.34 -0.62 1.88 -1.95 August. 4.89 +2.30 1.85 -0.95 .5.50 + 1.79 Although Burlington had an excess of 0.5 inch in August, drought at the other stations began during this month. Precipitation was below normal at Burlington (except during November) until May, and the same was true at Lincoln until June, except a slight excess in March, while at Phillipsburg it was exceedingly low until March. However, it should be noted that the spring and summer rainfall was high at the western stations, although below normal at Lincoln during April and May and again in July. The season was one which promoted excellent growth. Not only was the rainfall generally above normal, but the rains were also well distributed, drought periods west¬ ward being fewer and shorter than usual. This was indicated by the excellent growth of Bouteloua gracilis , which reached a height of 12 to 18 inches at Burlington and also by the fact that the short-grasses remained green or dried only slightly throughout the entire summer. Conditions were slightly less favorable in August at Lincoln and at Phillipsburg, where the three grama grasses ripened seed. On the whole, vegetation grew well at Lincoln, the grasses on the gravel-knoll remaining green all summer. An examination of the chresard taken at intervals in the various stations (table 38) is illuminating in this connection. The very low water-content on the high prairie in October is remarkable; on the gravel-knoll practically no water was available to 4 feet, and it was not 109 110 EXPERIMENTS DURING 1923. high even on low prairie. By March 1 a good water-content was found at both prairie stations and available moisture to the amount of 3 to 7 per cent was present throughout the season. The adequate holard in the surface layers at Burlington during July and August was very unusual. At Phillipsburg, as already indicated, it was drier in August. Because of the drought of autumn and late winter, many species under all methods of planting and transplanting were killed (table 39). Table 3S .—Holard in excess of hygroscopic coefficient at the several stations, 1923 (unless otherwise indicated ). Station. Date. Depth in feet. 0 to 0.5 0.5 to 1 1 to 2 2 to 3 3 to 4 i Lincoln: High prairie. Oct. 27,1922 1.8 3.7 3.0 2.8 3.8 Do. Mar. 1 18.1 18.3 12.1 10.7 7.3 Do. June 2 10.4 16.6 16.7 11.5 8.-2 Do. July 23 3.0 4.3 5.4 7.1 7.5 Do. Aug. 3 4.2 5.4 5.0 6.4 8.2 Do. Aug. 23 12.5 16.4 6.3 7.1 8.7 Gravel-knoll. Oct. 27 -1.5 1.0 -0.8 1.5 3.1 Do. June 2 -0.6 1.0 3.1 9.8 7.6 Burlington. July 6 0.4 -0.7 0.2 0.8 -0.6 Do. July 19 10.0 0.0 -0.3 -0.2 -0.2 Do. Aug. 21 5.8 4.5 -0.2 -0.7 -0.6 Phillipsburg. June 20 11.3 13.3 11.1 12.6 11.9 Do. Aug. 22 -0.6 1.5 0.7 1.6 5.6 Nebraska. May 27 20.9 20.0 17.7 13.9 12.7 Do. Aug. 28 22.7 20.0 9.5 6.7 7.7 Lincoln: Low prairie. Oct. 27 3.5 5.8 6.2 8.3 13.2 Do. Mar. 1 22.4 21.1 11.8 11.3 16.4 Do. Aug. 23 11.4 18.4 13.4 8.0 14.0 Survival Results. An examination of the data in table 39 reveals a number of interesting facts, the two most important of which are as follows: Losses at the three major climatic stations were greatest at Burlington, intermediate at Phillipsburg, and least at Lincoln. The losses at all stations averaged highest for the plants of a single year’s establishment, next for those 2 years old, and least for those established for 3 seasons. With a single exception, survival was greatest at all 3 stations under the method of denuded quadrats, this even exceeding that of transplanting blocks of sods. Surface sowing gave the lowest survival. Survival on the low prairie and at Nebraska City was very similar, following the sequence stated above as regards length of establishment. Except for 1920 (when all averages were very high) it exceeded that on the high prairie at Lincoln, and naturally that at the western stations. Survival on the gravel- knoll of species grown for at least two summers was remarkably high, though few species were concerned, with the exception of sod transplants. SURVIVAL RESULTS. Ill Surface Sowing. Nine of the 18 species that germinated on the surface of the high prairie at Lincoln died by the end of August. Because of the exceedingly dry autumn and winter, 6 species were winterkilled, leaving only 33 per cent of the survivors of the first summer. Petalostemon candidus had only two very delicate plants 3 to 5 inches tall by autumn, but Andropogon furcatus and A. nutans had both made a good growth, reaching a height of 10 or 12 inches. Table 39 .—Comparative survival at the different stations during fall and winter of 1922-23. Method of planting. Lincoln, high prairie. Phillips¬ burg. Burling¬ ton. Method of planting. Lincoln, low prairie. Nebraska City. Gravel- knoll. 1920. 1920. Surface sowing.... , , , . . . 100 Surface sowing.... ... ... ... Trench planting. . . 100 100 • . • Trench planting . . . 100 100 ... Denuded quadrats. 100 100 100 Denuded quadrats. 100 88 ... Seedlings. 60 Seedlings. Sod transplants... . 94 ... 52 Sod transplants... . 79 94 Average. 98 100 84 Average. 93 94 94 1921. 1921. Surface sowing. . . . ... 50 ... Surface sowing.... ... 100 ... Trench planting. . . 100 50 . . . Trench planting. . . . . . 100 ... Denuded quadrats. 100 100 50 Denuded quadrats. 80 100 67 Seedlings. 60 68 Seedlings. 100 88 100 Sod transplants.... 94 63 Sod transplants.... 100 100 Average. 89 72 57 Average. 93 97 89 1922. 1922. Surface sowing.... 33 38 ... Surface sowing.... 17 57 Trench planting. . . 40 50 0 Trench planting. . . 78 50 Denuded quadrats. 78 72 22 Denuded quadrats. 85 75 Seedlings. 79 50 20 Seedlings. 70 85 Sod transplants.... 86 17 Sod transplants.... 80 Average. 63 53 15 Average. , 66 67 ... In the mixed prairie at Phillipsburg, 8 species, 4 of which were in very poor condition, survived the first season. Andropogon nutans, A. furcatus, Des- modium canescens, Elymus canadensis, and Kuhnia glutinosa were winter- killed, leaving only the 3 species of Bouteloua. None of these had seeded by the following August, but all were in fair condition, having reached heights of 4 to 8 inches. None of the surface-sown plants survived the summer at Burlington. Trench Sowing. Of the species sown in a trench on the high prairie, 63 per cent lived through the first season, most of them being badly dried out late in August; 5 of the 10 did not grow the next spring, viz, Andropogon furcatus, Aristida purpurea, Elymus canadensis, Liatris scariosa, and Rohinia pseudacacia. Andropogon nutans and Pinus ponderosa had succumbed by August of 1923, leaving only several fine clumps of Bouteloua racemosa and 4 good plants of Desmodium canescens, each about 8 inches tall. Plants in the trench at Phillipsburg were represented in the fall of 1922 by Andropogon nutans and Bouteloua racemosa only. The last was winterkilled, while Andropogon was represented by very much attenuated plants 8 inches tall, the shade being very dense again during 1923. At Burlington, Andropogon nutans alone survived the summer, but it died the following fall or winter. 112 EXPERIMENTS DURING 1923. Denuded Quadrats. Of the 25 species that grew in the denuded quadrats on the high prairie, 18 survived the first summer. Of these, Aristida purpurea, Desmodium canescens, Onagra biennis, and Pinus ponderosa succumbed by spring. Andro- pogon scoparius and Muhlenbergia pungens, which were represented by a single clump each, died before September 1923. Acer saccharinum was killed back to within 6 inches of the soil, 16 of the trees dying; however, like all the other species, they made a good growth during the summer. Exceptions to this were Bouteloua gracilis, Sporobolus asper, and Robinia pseudacacia, which did poorly. None of the species blossomed. The quadrats were only slightly invaded. It is interesting that not only grasses and forbs, but also certain trees, survived the second summer. At Phillipsburg only half of the species that germinated in the denuded quadrats lived to the end of the first summer. Andropogon furcatus, Gleditsia triacanthus, and Acer saccharinum were winterkilled. Although all quadrats were densely shaded by a rank growth of Bouteloua gracilis 18 to 24 inches high, the 8 remaining species all did well. Aristida purpurea failed to blossom, as did also Andropogon nutans and Panicum virgatum. The rest, viz, Bouteloua racemosa, B. gracilis, Sporobolus asper, and Petalostemon can- didus, seeded, the last at a height of 2 to 2.5 feet. At Burlington, 64 per cent survived in the denuded quadrats. Because of the very dry fall and winter that ensued, only 2 of the 9 species showed renewed growth the following spring. Bouteloua gracilis formed excellent clumps and flowered profusely in 1923. Elymus canadensis, the other survivor, was represented by a single stalk which headed at 16 inches. Seedling Transplants. Of the seedlings transplanted into high prairie in the spring of 1922, only 70 per cent survived, notwithstanding that they had been regularly watered throughout the summer. Of the 14 survivors, Elymus, Petalostemon, and Argemone were winterkilled. Onagra died during the following summer, but all of the others made a good growth, though none came into blossom. Of the 19 species of seedlings grown in the mixed prairie, 15 survived, but 8 of these were winterkilled. Although the rest made a good growth during the following summer, Bouteloua gracilis and Ratibida columnaris alone came into bloom. At Burlington only 5 of the 19 species trans¬ planted survived. All but one were winterkilled, viz, Sporobolus, Bouteloua racemosa, B. hirsuta, and Stipa viridula, Bouteloua gracilis alone surviving. This did fairly well in 1923, but produced no flower-stalks. EXPERIMENTS AT EDAPHIC STATIONS. Survival Results. Surface Sowing. On the low prairie, owing to dense shading, only 25 per cent of the 1922 surface-sown plants survived the first summer. Aristida, Andropogon furcatus, A. nutans, Bouteloua gracilis, B. racemosa, and Petalostemon candidus did not grow the following spring. Andropogon nutans alone survived, and at the end of the following summer the slender plants were a foot high. EDAPHIC STATIONS! SURVIVAL. 113 At Nebraska City, 50 per cent of the species survived. Bouteloua hirsuta, Elymns canadensis, and Kuhnia glutinosa were winterkilled and Bouteloua racemosa died by the end of the next summer. Andropogon nutans, A. fur- catus, and Liatris punctata alone survived. In all cases the plants were very much attenuated and it seemed doubtful if they would ecize. Trench Sowing. As regards the trench plantings in the low prairie, 9 species, i. e., 64 per cent of those that germinated, grew throughout the first summer. Aristida and Kuhnia were winterkilled and Andropogon furcatus died the following summer. All the remaining species were very slender, Elymus and Petalo- stemon being represented in August by remnants only. Andropogon nutans, Bouteloua gracilis, B. racemosa, and Desmodium canescens, which made up the remainder, reached heights of 6 to 8 inches, but none seeded. Of the 21 species planted in the Nebraska City trench, only 4 survived, and of these Kuhnia glutinosa and Robinia pseudacacia were winterkilled. Andropogon nutans and Bouteloua racemosa were each represented by clumps of rather slender plants at the end of the following summer. Denuded Quadrats. On the low prairie at Lincoln, 60 per cent of the 1922 plantings in quadrats survived, Amorpha canescens and Onagra biennis alone succumbing dur¬ ing the following winter. At the end of August 1923, the following were in very poor condition, being represented by remnants only: Bouteloua gracilis, B. hirsuta, Sporobolus asper, and Acer saccharinum. The 7 remaining species, although somewhat slender, had made fair to good growth, but none seeded. At Nebraska City, 12 species survived the first summer. A fourth of these, including Andropogon furcatus, Desmodium, and Kuhnia were winterkilled, and Robinia pseudacacia died later. All of the remaining species made good growth, Andropogon nutans forming a dense sod. Seedling Transplants. Of the species transplanted as seedlings into the low prairie, 45 per cent survived the first summer, but among these Stipa viridula, Petalostemon candidus, and Kuhnia were winterkilled. All the rest, including the 3 gramas, Sporobolus, Desmodium, and Andropogon nutans, prospered until the end of the growing-season. At the Nebraska City station, 13 species (59 per cent) lived to the end of the first summer. Petalostemon and Ratibida were winterkilled. All of the others made a good growth because of the favorable season, with the exception of Onagra, which died in midsummer. Summary of Survival. The fate of the 1922 plantings by the end of 1923 is shown in table 40. Under all methods of planting or transplanting, the survival at the end of the second season of growth was greatest at Lincoln, intermediate at Phillips- burg, and least at Burlington. Of the several methods of planting, survival in the trench was lowest and surface sowing next, while the denuded quadrats were only slightly more successful than the seedling transplants. The average survival on the low prairie at Lincoln was greater than at Nebraska City, but the latter was nearly equal to that on high prairie at Lincoln. On low prairie and at Nebraska City the success of the several methods of seeding was in the same order as at the other stations, except that trench planting 114 EXPERIMENTS DURING 1923. outranked surface sowing. The average survival of the seedling transplants at these stations was less than on high prairie (where they were watered the first season), but exceeded that in mixed prairie. An examination of the species which gave the greatest percentage of sur¬ vival is of interest. As to surface sowing, Petalostemon, Andropogon nutans, and A. furcatus alone survived at Lincoln, and the three boutelouas at Phillips- burg, so far as the three major stations are concerned. However, Andropogon nutans survived in the two stations at Lincoln as well as at Nebraska City, and A. furcatus on high prairie and at Nebraska City. Liatris punctata was the only other survivor and was found at Nebraska City alone. Table 40. —Summary of survival of 1922 plantings in August 1923. Method of seeding. Lincoln. Phillips¬ burg. Burling¬ ton. Lincoln (low prairie). Nebraska City. Surface sowing. 17 16 0 6 21 Trench. 13 13 0 43 15 Denuded quadrats. 48 36 14 50 36 Average. 26 22 5 33 24 Seedlings. 60 37 5 39 45 In the trench, Bouteloua racemosa and Desmodium were the sole survivors at Lincoln, and Andropogon nutans at Phillipsburg, while, as before, none were found at Burlington. But Bouteloua racemosa also survived at Nebraska City and on the low prairie and Andropogon nutans at both of these stations also. Desmodium persisted on low prairie as well as on high prairie. The only other survivors occurred on the low prairie alone, viz, Elymus, Petalostemon , and Bouteloua gracilis . With respect to denuded quadrats, Bouteloua gracilis persisted at all five stations, B. racemosa and Andropogon nutans at all but Burlington, Petalo¬ stemon, Panicum , and Sporobolus at three, and Andropogon furcatus, Kuhnia, Liatris scariosa, Gleditsia, Acer saccharinum, and Elymus at two, while the following were found at only one station each: Robinia, Bouteloua hirsuta, Desmodium (at Lincoln), Pinus, and Acer negundo (at Nebraska City). Among the seedlings, Bouteloua gracilis occurred at all the stations, and B. hirsuta and Sporobolus at all but Burlington. Liatris scariosa, Andropogon nutans, Bouteloua racemosa, Stipa spartea, and Desmodium were found at three stations, while Ratibida columnaris occurred at two. The following were represented at one station only: Liatris punctata, Koeleria (at Lincoln), Petalostemon (at Phillipsburg), Kuhnia, and Elymus (at Nebraska City). Arranging the species in order of their survival under all methods of planting at the several stations gives Andropogon nutans 14, Bouteloua gracilis 12, B. racemosa 11, Sporobolus and Desmodium each 7, Petalostemon and Bouteloua hirsuta each 6 , Andropogon furcatus and Liatris scariosa each 5, Elymus canadensis 4, Panicum virgatum, Kuhnia glutinosa, and Stipa spartea each 3, Liatris punctata, Gleditsia triacanthus, Acer saccharinum, and Ratibida colum¬ naris each 2, while the following occurred only once: Robinia pseudacacia, Pinus ponderosa, Acer negundo, and Koeleria cristata. EDAPHIC STATIONS! SURVIVAL. 115 Bouteloua gracilis was the only species which survived under some method of planting at all 5 stations. The following persisted at both Lincoln and Phillipsburg, but not at Burlington: Petaloslemon, Andropogon nutans, Bouteloua racemosa, B. hirsuta, Panicum, Sporobolus, Stipa spartea, Ralibida, and Liatris scariosa. Sod Transplants. As to the 1922 sod transplants on the high prairie, the following died during the fall or winter: Anemone , Grindelia, Helianthus, and one lot of Bouteloua gracilis, while Astragalus crassicarpus died before autumn. Aside from the grasses, nearly all of which made an excellent growth, it is interesting to note that several forbs, viz, Liatris scariosa, Solidago missouriensis, and S. rigida, blossomed profusely. At Burlington the fall and winter were so extremely dry that 83 per cent were winterkilled, including 9 species. This left a single sod of each of the following: Agropyrum, Bouteloua gracilis, Panicum virgatum, and Andropogon scoparius. The last died before the end of the summer. Agropyrum was represented by only 3 shoots and Panicum by 2, neither species blossoming, but Bouteloua gracilis made a good growth and put forth flower-stalks abun¬ dantly. The contrast between prairie and plains is consequently very sharp and the disaster met by planted or transplanted species in the latter area needs no emphasis. The indications in 1923 were that Bouteloua gracilis and possibly Agropyrum glaucum were the only species that would survive per¬ manently. Sods transplanted to the low prairie in 1922 were winterkilled to the extent of 5 species, while Agropyrum, Bouteloua gracilis, and Bouteloua hirsuta died the following summer. By the end of the season Koeleria and Anemone were nearly dead, due to the severe competition and dense shade. Although all of the other species were still prospering, the condition of the transplants made in 1920 and 1921 indicated that a year or two would suffice to eliminate the upland species and leave the taller subclimax grasses and forbs in pos¬ session of the area. None of the 1922 sods in the salt-flat were winterkilled, but Bouteloua hirsuta died the following summer. In August 1923, Koeleria was represented by mere remnants, while several of the other species evidenced by their paler color and dwarfed habit the uncongenial nature of the habitat, notwith¬ standing the rather wet season. However, all did fairly well, but several failed to bloom. As to the species transplanted into the former swamp area in 1922, Sporo¬ bolus and two lots each of Bouteloua hirsuta and B. racemosa failed to grow the following spring. The slough-grass cast a very dense shade, and by the end of 1923 Bulbilis, Elymus, and Bouteloua gracilis had died, while Agropyrum was represented by mere remnants. This left Spartina, which had made a normal growth, Stipa spartea, the flowers of which were still in the leaf-sheath in late August, and Koeleria, which failed to blossom. In another season it is probable that these two high-prairie species would also have been eliminated. The sods planted in 1922 at Colorado Springs did poorly because of grazing. By the end of the second summer, Andropogon furcatus, Panicum virgatum, and Bouteloua racemosa had died, and Elymus and Spartina were represented by mere remnants. None but Bulbilis had blossomed. 116 EXPERIMENTS DURIN& 1923. PHYTOMETRIC RESULTS. Transpiration and Growth. Plan. In order to obtain further light as to the climatic differences of the major stations in terms of functional response, a special investigation was made of transpiration and growth during the summer of 1923. Plants of Helianthus annuus, Avena sativa, Elymus canadensis , and Acer negundo were grown from seed or transplanted as seedlings into sheet-metal containers of appropriate shape and sufficient size to accommodate the root systems throughout the duration of the experiment. After the plants were well established, the leaf area and the weight of plant and container were determined. They were then installed at the several stations, together with a complete battery of instruments, for a period of 14 days, and measurements made of transpiration and growth in terms of increased area. The considerable differences in altitude and hence in seasonal development made simultaneous studies unde¬ sirable, and consequently the periods of observation were successive. This was also imperative because of the time and effort involved. b Methods. The individual plants of sunflower and box-elder were placed in cylindrical galvanized-iron containers 5 to 6 inches in diameter and 9 to 10 inches deep, filled with rich loam soil tamped firmly in place. A layer of one-half inch of coarse gravel in the bottom of the container covered an exit-tube, consisting of an automobile tire valve-stem with the inner end cut off and covered with a fine copper gauze soldered in place. The core of each tube had been removed. The tube was soldered in place with the threads projecting through the wall of the container, so that an exhaust pump could be attached for aerating the soil. The tube also assisted materially in watering the plants, the usual cap preventing loss during the intervals. The soil was well screened, brought to an optimum holard, and weighed at the time of filling the containers. By restoring the containers to their original weight from time to time, the holard was maintained at the desired level. To prevent loss other than by transpira¬ tion, the containers were furnished with a sloping metal top provided with a circular opening with the edges reamed upward, and large enough to receive a cork 2.5 inches in diameter. An effective seal was formed by boring a hole large enough for the plant stem, splitting the cork and fitting it into place after padding the sides of the opening with a little cotton. The seal was tested by a check container without a plant, but fitted with a wooden peg to simulate the plant stem. During the period of 14 days this did not lose water in an amount sufficient to be detected by a balance sensitive to 2 grams under a load of 7 kg. The containers for wild rye and oats were similar to those already described, except for the tops, which were furnished with a slit 5 inches long and 1 inch wide. The edges of the metal cut in making the slit were turned down into the container about a quarter of an inch on each side and a half-inch at the ends of the slit. These furnished supports for narrow strips of shellacked oak, which were held in place by thin wedges of similar material at each end. These were put in place after the containers had been filled and the soil pressed TRANSPIRATION AND GROWTH. 117 firmly under the metal tops. They were then coated with shellac, so that no openings remained except between the wooden strips, which narrowed the slit to about 10 mm. The seeds were planted through this opening, which was nearly filled with soil kept moist by frequent watering. After the plants came up, the remainder of the opening was filled with sand level with the top of the wooden strips. The containers were provided with felt tops, cut to fit around the slits, and sunken level with the soil. The efficiency of the sand-mulch in preventing water-loss was found to be high, only 6 grams escaping in a period of 2 weeks. Twelve plants of average size were selected from the several containers and the leaf-area determined. From this, and the number of plants in each container, the initial area of the group of plants in any container was calculated. The final area was determined in a similar manner. This method was used because it was quite impossible to determine the total area of the plants in place, as could be done with the dicotyls. Owing to cool, wet weather, the plants grew slowly, except during the last week of May. They were kept covered during rains and at night, watered from a burette, and aerated from time to time. To keep the metal containers from heating the soil, they were surrounded by sand and the top covered by a collar of felt about a centimeter thick, held in place by strips of adhesive tape. On June 1 the leaf-area was determined by means of solio prints and the planimeter, corks were inserted, and the containers (with collars removed) brought back to their initial weight after they had been transported to the station in the high prairie at Lincoln. Here they were placed in the soil and thoroughly covered after the collars had been replaced, the exposed plants being sheltered during rains. After the 14-day period of the experiment, following final weighings,the number of parent plants and number of tillers in each container was ascer¬ tained. Eighteen specimens of each group were then selected and their areas determined. From these data the final areas were calculated. Conditions and Results at Lincoln. Physical Factors. The period (May 31 to June 15) was exceptionally cool and wet. Only one entire day was sunny, and the sun shone for over half of the time on two others alone. During a 4-day interval the sun did not appear, while the total sun¬ shine over the whole period was approximately 5.5 days. A total of 4.9 inches of rain fell during 8 days, and rains occurred in the daytime on 6 days, during one of which it was necessary to keep the plants sheltered for the entire day. In fact, they were under cover for a total period of 30 hours of daylight in which rain was falling. The air-temperature during the first 7 days ranged from 55° to 91° F., while the humidity was never lower than 55 per cent. This was followed by a period of 4 days when the temperature remained at 58° to 65° F. and the humidity above 80 per cent. The remaining time was similar to that of the first 7 days. The average day temperature for the entire period was 70.1° F. and the day humidity 75 per cent. The soil temperature at a depth of 6 inches among the containers varied from 62° to 71° F. The total wind movement at a height of 39 cm. during the period was 1,282 miles, an average of only 3.6 miles per hour. The exceptionally mesophytic conditions are shown by the average 118 PHYTOMETRIC RESULTS Table 41. — Water-loss and growth at Lincoln. Helianthus annuus. Water added /*N TJ -tT c3 (grams). Final weight, June 15 (gran CO 3 OS a -d • M o GO Con¬ tainer. Initial weigl June 1 (gi June 4. June 12. June 14. Total water (grams). Initial area (sq.cm.). Final area. (sq. cm.) Average are 1 (sq.cm.). Loss per sq. (grams). P. ct. increa area. 1 6,748 300 250 100 6,464 934 229.1 960.5 594.8 15.7 319.2 2 7,053 300 250 100 6,700 1,003 266.9 1,135.5 701.2 14.3 325.4 3 7,035 300 250 100 6,690 995 237.9 1,142.4 690.1 14.4 380.2 4 6,793 300 250 100 6,587 856 167.0 1,029.1 598.1 14.3 516.2 5 6,942 300 250 100 6,769 823 156.7 1,075.9 616.3 133 586.6 6 6,966 300 250 100 6,660 956 258.5 1,099.4 678.9 141 325.3 7 6,898 300 250 100 6,646 902 201.0 1,127.9 664.5 135 461.1 8 6,700 300 250 100 6,452 898 202.7 921.1 561.9 159 354.4 9 6,974 300 250 100 6,629 995 177.3 1,124.4 650.8 153 534.2 10 6,800 300 250 100 6,516 934 244.2 928.8 586.5 159 280.3 11 6,980 300 250 100 6,655 975 275.1 1,205.7 740.4 132 338.2 12 6,962 300 250 100 6,672 940 187.5 1,106.1 646.8 145 489.9 Aver.... 145 409.2 A VENA SATIVA. Initial weight, June 1 (grams). Water added (grams). Final weight, June 1 (grams). 'g <8 3 <73 a T3 a • ri V CO Con¬ tainer. June 7. June 13. June 15. Total water (grams). Initial area (sq. cm.). Final area (sq. cm.). Average are (sq. cm.). Loss per sq. (grams). P. ct. increa area. 1 8,918 600 300 8,442 1,376 414.0 1,601.7 1 ,007.8 136 286.9 2 8,360 500 180 8,081 959 414.0 1,620.8 1 ,017.4 94 291.5 3 8,233 500 120 7,885 968 414.0 1,614.4 1,014.4 95 290.0 4 8,221 500 0 7,860 861 414.0 1,544.0 979.0 88 272.9 5 8,237 500 0 7,927 810 414.0 1,436.1 925.0 87 246.8 6 8,316 500 138 8,004 950 414.0 1,596.7 1 ,005.4 94 99 285.6 278.9 ElYMUS CANADENSIS. 1-31 5,273 100 0 5,186 187 145.4 492.6 319.0 59 238.8 2-40 5,010 100 100 4,896 314 187.6 635.7 411.6 76 238.8 3-41 5,010 100 100 4,916 294 192.3 651.6 421.9 69 238.8 4-28 5,206 100 0 5,070 236 131.3 445.0 288.1 82 238.8 5-43 5,180 100 100 5,044 336 201.7 683.3 442.5 76 238.8 Control. 5,093 0 0 5,087 6 238.8 Aver.... 72 238.8 Acer negundo. 1 4,355 4,306 49 38.6 67.6 53.1 92 75.1 2 6,104 . . • 6,049 55 42.2 107.6 74.9 72 155.5 3 5,910 , , , 5,872 38 36.2 45.6 40.9 93 25.9 4 6,363 80 6,300 143 99.2 241.2 170.2 84 143.1 5 5,933 5,878 55 58.8 72.6 65.7 84 23.4 Aver. . . . 85 84.6 TRANSPIRATION AND GROWTH. 119 daily evaporation, which was only 8.4 c. c. per day, in contrast to a normal daily loss of 20 to 30 c. c. during this portion of the growing-season (table 41). Transpiration and Increase in Leaf-are a. The behavior of plants of the four species with respect to water-loss and increase in area during the control period of two weeks at Lincoln is exhibited in table 41. Conditions and Results at Phillipsburg. Installation. Plants of the same species as those used in the preceding experiment were grown at Lincoln and shipped to the mixed-prairie station at Phillipsburg, where they were used experimentally for the 14-day period from June 18 to July 3. Some differences in the age and size of the various species from those at Lincoln were brought about by more or less favorable growing conditions during the few weeks preceding the beginning of the experiments, but the several species were as nearly as possible comparable with those previously used. The same methods were employed in both cases. Physical Factors. The period June 19 to July 3 was one of fairly typical weather for early summer at this station, when compared with that of the three preceding years for which factor data were obtained. Eight days were clear, except for a few floating clouds, and the sun shone on four others for over half of the time. Two days alone were sufficiently cloudy so that the sun shone for less than one- third of the day. Heavy showers occurred at night on June 18, 21, 26, and 29, as well as during the day on June 21, when it was necessary to cover the plants for 2 hours. Three days were so extremely hot, windy, and dry that it was desirable to shade the plants and protect them from the wind for periods of 2 to 5 hours. Notwithstanding an adequate holard, partial wilting occurred at these times, crop and ruderal plants suffering likewise. During one of the most severe days (June 24) the temperature was 98° F. and the humidity 42 per cent, while a south wind of 20 to 30 miles per hour was blow¬ ing. The air-temperature during the first five days of the period ranged from 65° to 98° F. and the humidity from 41 to 97 per cent. For the remainder of the time the variations were from 42° to 90° F. and 34 to 100 per cent humid¬ ity. The average day temperature for the period was 80.6° F. and the average day humidity 60 per cent. Soil temperatures at a depth of 6 inches varied from 74° to 82° F. The total wind movement at a height of 39 cm. during the 14-day period was 1,100 miles, an average of 3.3 miles per hour; moreover, most of this occurred during the day. The somewhat xerophytic conditions are indicated by the average daily evaporation of 25.5 c. c. Under these conditions of prevailingly clear, warm weather, growth was marked and the losses by transpiration were high (table 42). Conditions and Results at Burlington. Installation. The plants used at Burlington were also grown in appropriate containers at Lincoln for the usual period and shipped to Burlington, where they were used 120 PHYTOMETRIC RESULTS experimentally for the 14-day period from July 6 to 20. Methods similar to those utilized for the other stations were employed throughout. Owing to the abundance of grasshoppers, it was necessary to screen the inclosure, a fence of hardware cloth about 4 feet high, the top edge of which was turned outward and downward, being employed. This modified the conditions of wind and evaporation within to a considerable extent and shaded the plants for about 2 hours each day. Transpiration and Increase in Leaf-are a. Table 42 . — Water-loss and growth at Phillipsburg. Helianthus annuus. GQ S .. ^ 3 3 Water added (grams). "d > i i 1 l—-H c3 3 • S E Total wa (grams Initial ar (sq. cr Final are (sq. cn Loss per (grams P. ct. inc area. 1 6,862 50 200 300 200 6,627 • 985 65.89 860.86 463.38 212 1,206.5 2 6,974 50 200 300 200 6,639 1,085 52.84 925.73 489.29 222 1,651.9 3 7,030 50 200 300 200 6,724 1,056 57.11 934.08 495.59 213 1,535.6 4 7,036 50 200 300 200 6,628 1,158 53.75 727.39 390.57 296 1,253.3 5 7,061 50 200 300 200 6,797 1,014 54.91 870.29 462.60 219 1,484.9 6 7,058 50 200 300 200 6,718 1,090 51.29 960.98 506.14 215 1,773.6 7 7,045 50 200 300 200 6,760 1,035 50.39 945.36 497.88 208 1,776.1 8 7,074 50 300 400 200 6,717 1,307 91.86 1,133.60 612.73 213 1,134.0 Aver. . . 225 1,476.9 Avena sativa. 1 8,3S8 200 350 400 200 8,138 1,400 316.37 1,254.24 785.31 178 296.4 2 8,385 150 350 300 200 8,210 1,175 260.54 905.84 583.19 201 247.6 3 8,442 200 350 400 200 8,200 1,392 353.59 1,297.50 825.55 169 266.9 4 8,121 100 350 300 200 8,053 1,018 353.59 836.16 594.88 171 136.4 5 8,369 200 350 400 200 8,129 1,390 353.59 1,239.82 796.71 174 250.6 6 8,211 100 350 400 200 7,958 1,303 316.37 1,196.56 756.47 172 278.2 Aver. 178 246.0 Elymus CANADENSIS. 1 5,048 100 200 200 4,900 648 214.08 720.90 467.49 139 236.7 2 5,289 200 400 300 4,939 1,250 214.08 780.54 497.31 251 264.6 3 5,422 100 300 100 5,172 750 214.08 763.50 488.79 153 256.6 4 Check.... 5,291 4,741 200 400 300 4,913 4,733 1,278 8 214.OS 789.06 501.57 255 268.6 Aver. 199 256.6 Acer NEGUNDO. 1 6,190 200 100 6,068 422 128.94 336.70 232.82 181 161.1 2 6,079 ... 100 50 5,979 250 77.52 214.99 146.26 171 177.3 3 4,337 50 100 100 4,208 379 139.28 393.41 266.35 142 182.5 4 4,349 50 100 100 4,260 339 124.81 285.53 205.17 165 128.8 5 6,022 . . . 100 , . , 5,956 166 59.17 157.24 108.21 153 165.7 G 5,955 100 5,878 177 69.38 179.46 124.42 142 158.6 Aver. 159 162.3 TRANSPIRATION AND GROWTH. 121 Physical Factors. The weather conditions for the first four days of the experiment were very characteristic of the high plains during July, with day temperatures reaching 90° to 96° F. and falling to 60° or 65° F. at night, while the humidity ranged from 25 to 30 per cent in the afternoon to 80 or 85 per cent at night. However, this was followed by 10 days during which the atmosphere was more humid and cooler, due largely to an unusual amount of cloudy weather and rain. Rains fell in the evening or at night as follows: July 11, 0.02 inch.; 14th, 1.48 inches; 16th, 0.51 inch; 17th, 0.58 inch. Moreover, the heavy rains were general and greatly increased the humidity. One day was entirely cloudy, and four were cloudy half or more of the time, while on 3 others the sky was overcast more than one-fourth of the time, only 6 days being clear. With the exception of the first 4 days, clouds invariably obscured the sun after 5 or 6 p. m. The temperature ranged rather uniformly between 55° to 60° and 80° to 85° F., and the humidity usually reached 100 per cent at night (the vegetation being covered with rain or dew) and fell to 50 to 55 per cent during the after¬ noon (on one day to 36 per cent). The average day temperature and humidity for the period from 8 a. m. to 6 p. m. were 80° and 57 per cent respectively. While the temperature was nearly the same as the average day temperature for the same period during the three preceding years, as determined by hygrothermographs in the field, the humidity was 11 per cent higher. The soil temperature at a depth of 6 inches among the containers was 73° to 86° F. The wind movement at a height of 39 cm. during the first 4 days averaged over 7 miles per hour, but for the rest of the time it was much less (3.4 miles per hour). The average daily evaporation was 35.4 c. c., which was 5 c. c. lower than for the same period of the preceding years. It was necessary to shade the plants for an hour or so to prevent wilting during the first day or two. Transpiration and Increase in Area. Table 43 . — Water-loss and growth at Burlington. Helianthus anntjus. Con¬ tain¬ er. Initial weight, July 6 (grams). Water added (grams). Final weight, July 20 (grams). Total water used (grams). Initial area (sq. cm.). Final area (sq. cm.). Average area (sq. cm.). Loss per sq. dm. (grams). P. ct. increase in area. July 11. July 13. July 16. July 18. July 19. 1 6,526 100 100 300 300 300 6,444 1,182 30.62 1,235.16 632.89 187 3,933.9 2 7,572 100 100 300 300 300 7,637 1,035 22.22 775.20 398.71 259 3,388.7 3 6,591 100 100 300 300 300 6,409 1,182 29.59 1,159.44 594.52 199 3,818.4 4 7,538 100 0 300 300 300 7,617 921 21.58 650.91 336.25 274 2,916.2 5 7,599 100 0 300 300 300 7,612 987 20.20 802.33 411.27 240 3,871.7 6 6,600 100 100 300 300 300 6,517 1,183 26.10 1,314.35 670.23 176 4,935.8 7 6,692 100 100 300 300 300 6,522 1,270 21.91 1,354.01 687.96 185 6,079.9 8 7,545 100 100 300 300 300 7,500 1,145 27.78 868.16 447.97 256 3,023.7 9 6,477 100 0 300 300 300 6,446 1,031 28.17 1,114.18 571.18 181 3,855.2 10 6,636 100 0 300 300 300 6,754 882 18.22 955.05 486.64 181 5,141.7 Aver 214 4,096.5 122 PHYTOMETRIC RESULTS Table 43. — Water-loss and growth at Burlington —Continued Avena sativa. Container. Initial weight, July 6 (grams). Water added (grams). Final weight, July 19 (grams). Total water used (grams). Initial area (sq. cm.) j Final area (sq. cm.). Average area (sq. cm.). Loss per sq. dm. (grams). P. ct. increase in area. July 11. July 13. July 16. July 18. 1 8,850 200 250 200 8,722 778 92.92 566.94 329.93 236 510.1 2 8,732 150 100 200 150 8,593 739 100.33 721.32 410.83 180 618.9 3 8,702 150 100 200 200 8,583 769 85.48 642.81 364.15 211 652.0 4 8,803 100 100 150 100 8,734 519 38.93 506.17 272.55 190 1,200.0 5 8,915 150 100 200 200 8,816 749 89.21 523.91 306.56 244 487.3 Check. 8,742 8,734 8 . . Average. 212 693.7 Elymus canadensis. Initial weight, July 6 (grams). Water added (grams). /•—N CO a "d w Agropyrum. . ( 80 35.4 0.96 Phillipsburg.. . ( Bouteloua. .. . June 18 \Andropogon. . July 3 75 80.6 60 25.5 1.33 Lincoln. Andropogon .) Stipa.1 July 24 r 79 80 22.0 0.8.5 | Koeleria. Bouteloua. .. to Au&. 8 CLIP-QUADRATS. 131 Growth of Natural Cover and Crops. Objectives. The growth of a representative area of a community may be used as a climatic index as well as a measure of response in much the same way as transpiration. It possesses three distinct advantages over the latter, though the two are complementary and hence one can not replace the other. Growth demands no laborious installation, as it can be determined directly from the native or culture community in position. While it can be measured at any time, it yields the major values at the end of the growing-season, and thus does not require the services of a resident investigator. Moreover, it inte¬ grates the response for the whole season, though as a complex it permits less ready analysis than transpiration and is also less satisfactory for short periods. In short, it is the simplest and most convenient of all community phytometers when employed in the form of the clip-quadrat. This is the only practicable method, as the measurement of the individuals in a com¬ munity group is too time-consuming to be desirable. There are certain cases in which it is profitable to pull the individuals out with their roots, but this is hardly feasible in a close cover or a compact soil. The clip-quadrat is merely the usual one of a square meter in extent, from which the shoots are cut at any desired time. It may be either smaller or larger in order to meet special con¬ ditions, as in the case of crops planted in rows. The growth is regularly expressed in terms of dry matter, but in the case of grazing ranges or forage crops, the green weight should likewise be found. Finally, the clip-quadrat facilitates the analysis of community response to climatic factors by making it possible to measure growth during different portions of the season or its variations from season to season or from the wet to the dry phase of a climatic cycle, and to determine the part played by the various species in the total production. In the case of the grains, it is often preferable to select the individuals to be cut in accordance with the results desired, instead of taking all those in a particular area (Weaver, Jean, and Crist, 1922). This may be regarded as an aggregate clip-quadrat; it has the further advantage of permitting measure¬ ment of particular individuals throughout the season. Plan. Clip-quadrats were first installed at the three stations in 1920; these repre¬ sented high and low prairie in the true-prairie association, mixed prairie, and short-grass plains. They were again cut in 1921 and 1922 to determine the fluctuation from year to year. The first step in the simple procedure was to select a considerable number of quadrats in typical areas of each climax. The height and density of the cover, the abundance of dominant and sub¬ dominant species, the presence of layers, etc., were recorded and photographs made of certain of the quadrats. In some instances it has proved desirable to make a chart of these as well. The cover was then removed by cutting it near the surface and at a uniform level with a hand-clipper. It was collected, sorted, and then shipped into the laboratory to be thoroughly air-dried, after which the actual production was determined on the basis of the dry weight. This gave an expression of the growth of the community as a unit, as well as the role taken by each dominant or subdominant in this. As complete factor 132 PHYTOMETRIC RESULTS. records were obtained at each station, this made it possible to correlate the yield of community or species with the climate and the season at each. The physical factors for the several years have been given in the preceding chapters and hence are not repeated here. Since even the most uniform cover shows some variation in density, the clip-quadrats were selected with much care and in sufficient number to insure dependable results. As in all ecological studies that are adequate, i. e., causal and quantitative, rather than merely mathematical, this demands considerable knowledge of the community and can not be met by random selection. The best plan is to locate a proper proportion of quadrats in pure or nearly pure stands of each dominant whenever these are present and to distribute the others among the various mixtures. It is often desirable to take the sub¬ dominants into account in doing this, as their yield may be much greater than that of the grasses. Results for 1920. The growth response for this year was obtained from about 400 clip- quadrats, of which 50 were in the true prairie at Lincoln, 50 from mixed prairie at Phillipsburg, 50 from the short-grass plains at Burlington, and 30 from each of the cultivated fields at the various stations. The short-grasses at Burling¬ ton averaged 103 gm., those at Phillipsburg 290 gm. Wheat-grass ( Agropyrum glaucum) at Burlington yielded 398 gm. per square meter, while at Phillipsburg the yield was 480 gm. A mixture of short-grass and tail-grasses gave 244 gm. per square meter at Burlington and 470 gm. at Phillipsburg, and the mixed prairies at Phillipsburg produced a total yield of 439 gm. per square meter, while those at Lincoln averaged 452 gm. The average yield at Phillipsburg (439 gm.) exceeded that from the hilltops at Lincoln (361 gm.) or even the slopes (429 gm.) at this time (July 9), but did not equal that of the lowland, which was 564 gm. Moreover, an examination of the factor data shows that late-summer drought usually prevails at the western stations and good plant growth is not maintained. This is well illustrated by a second series of quadrats cut August 16 to 24. The average yield at Burlington was 196 gm. (more mixed short and tall grass quadrats being included than before), at Phillipsburg 310 gm., and at Lincoln 465 gm. On the basis of the 50 quadrats taken at each station during the season, the average total production, pro¬ ceeding from the drier western station eastward, was 183, 378, and 458 gm. respectively. Crops of oats ( Avena sativa), spring wheat ( Triticum aestivum), barley {Hordeum vulgare), alfalfa ( Medicago sativa), and white sweet clover ( Melilotus alba) were grown in plats adjoining the several grassland stations (Weaver, Jean, and Crist, 1922:76). Each kind of crop was planted from the same lot of seed and grown under conditions of farm practice common to the several localities respectively. The crop plants were thoroughly air-dried before weighing, and the relative height of the mature oats and wheat at the several stations, as well as the comparative yield from an average square meter, were determined. The yield is the average of 25 to 30 square-meter quadrats taken from each of the several plats at the three stations when the grain was ripe. 400 plants of alfalfa and 300 of sweet clover of average size were carefully selected at each of the stations, cut at the ground-line, thoroughly air-dried, and the dry weight ascertained. This was done during July and again in GROWTH OF NATURAL COVER AND CROPS. 133 August for the sake of obtaining comparative values between clip-quadrats and a group of selected plants (table 51). The height-growth of all crops was greatest at Lincoln, less at Phillipsburg, and least at Burlington, the height of the former averaging more than twice that of the latter. The maximum penetration of roots was usually greatest at Phillipsburg, next at Lincoln, and least at Burlington, a divergence readily explained by the greater holard at the first and the dry subsoil and hard-pan at the last. Table 51. —Growth and yield of crop-quadrats and plants, 1920. Crop and station. Date of harvest. Average height. Maximum depth of roots. Average yield in grams per sq. met^r. Weight of 1,000 kernels. Oats: Lincoln. July 12 feet. 3.0 feet. 4.8 706 gm. 20.1 Phillipsburg. July 20 2.6 6.0 379 16.6 Burlington. July 19 1.5 2.7 175 16.2 Wheat: Lincoln. July 15 3.2 4.8 740 29.8 Phillipsburg. July 20 2.3 5.8 322 9.1 Burlington. July 19 1.7 2.7 205 20.1 Barley: Lincoln. July 12 2.7 5.4 607 32.7 Phillipsburg. July 17 2.4 6.7 407 14.7 Burlington. July 19 1.7 2.9 176 23.9 Alfalfa: Lincoln. July 12 1.5 5.7 Dry weight of plants. 528 Phillipsburg. July 9 0.7 5.0 292 Burlington. July 8 0.4 2.3 122 Sweet clover: Lincoln. July 12 2.0 5.5 840 Phillipsburg. July 9 1.3 5.7 461 Burlington. July 8 0.4 2.8 213 Alfalfa: Lincoln. Aug. 9 1.8 5.9 739 Phillipsburg. Aug. 4 1.2 • • • 601 Burlington. Aug. 5 0.6 2.0 214 Sweet clover: Lincoln. Aug. 9 2.5 • • • 1,103 Phillipsburg. Aug. 4 1.7 • • • 869 Burlington. Aug. 5 0.8 2.7 323 The average yield of grain for the three cereals was about twice as much at Lincoln as at Phillipsburg, and nearly twice as much at the latter as for Burlington. A similar relation obtained for the first cutting of alfalfa and sweet-clover, but at the second cutting Phillipsburg more nearly approached Lincoln. Thus, the five crops decreased decisively in both height and yield from Lincoln westward in correspondence with the rainfall, holard, and humidity. Results for 1921. Two series of cuttings were made during 1921, one about July 1 and a second during the last half of August. Ten clip-quadrats each of Bulbilis dactyloides and Agropyrum glaucum were used at each of the stations early in the summer; at all of these both grasses had headed or blossomed at the time of cutting. 134 PHYTOMETRIC RESULTS. The buffalo-grass averaged about 4.5 inches high at Burlington, while the staminate spikes were 5.5 inches tall. In the mixed-prairie the foliage of this grass averaged 5 inches in height and the flower-stalks 6 inches. Bulbilis occurs in the prairies at Lincoln only where the tail-grasses are held in check by grazing; hence it was impossible to secure an ungrazed sod comparable to that at the other stations. The foliage was only 4.5 inches high and the effects of heavy overgrazing the preceding year were apparent. On July 1 the average yield of 10 quadrats at each of the respective stations was 207, 266, and 235 gm. per square meter. At Burlington the leaves of wheat-grass had reached an average height of 16 inches and the flower-stalks of 26 inches at this time, at Phillipsburg the respective values were 22 and 30 inches, and at Lincoln 24 and 32 inches. The average yield of 10 quadrats from each station was 400, 457, and 606 gm. respectively. Table 52 .—Growth and yield of crop-quadrats in 1921. Crop and station. Aver¬ age height. Average number stalks per sq. meter. Average number stalks per plant. Average length heads or panicles in inches. Average number heads per sq. meter. Average total weight dry matter per sq. meter. Maxi¬ mum depth. Oats: feet. gm. feet. Lincoln. 3.2 375 3.3 10.5 283 792 4.8 Phillipsburg. . . 2.8 353 2.9 9 269 366 5.3 Burlington.... 1.5 414 2.5 5 171 180 2.5 Wheat: Lincoln. 3.2 648 2.8 4 365 557 4.3 Phillipsburg. . . 2.6 475 1.8 3.5 211 314 4.5 Burlington.... 1.6 419 1.6 2.5 277 172 2.5 Barley: Lincoln. 3.1 384 3.7 3.5 306 622 4.6 Phillipsburg. . . 2.8 253 2.2 3.25 201 369 6.0 Burlington.... 1.3 255 1.6 2 197 122 2.5 On August 17, 12 quadrats cut in the high prairie at Lincoln yielded an average of 581 gm., and a similar number from the low prairie 929 gm., giving an average yield of 755 gm. This considerably exceeded the average weight of 11 tail-grass quadrats cut on August 30 at Phillipsburg, which was 477 gm. The average of the 44 quadrats from both cuttings at each station was as follows: Burlington 353 gm., Phillipsburg 402 gm., and Lincoln 603 gm. per square meter. Crops of the smaller cereals were again grown in 1921 under conditions similar to those already described for 1920. There was a regular decrease in height of the crop from the more humid to the more arid stations. The same general relation holds for the average number of stalks per square meter, except at Burlington, where many tiny stalks only 2 to 4 inches tall started growth relatively early and soon dried out, but remained until harvest. During 1920 the average number of stalks per square meter at Burlington was from one-third to one-half less than at the other stations, although the number at Phillipsburg often exceeded that at Lincoln. The average number of stalks per plant (1921) was in direct relation to the water-content of soil and other factors favorable or unfavorable to plant growth. In general, this relation held also during 1920. The average number GROWTH OF NATURAL COVER AND CROPS. 135 of heads per square meter and the average length of heads or panicles de¬ creased from Lincoln to Phillipsburg to Burlington respectively. An exception to this occurred in the case of the number of heads of wheat at Burlington when compared with Phillipsburg, while the difference in this respect in the case of barley was small. However, a clear gradation in the reduction of total dry weight from east to west is apparent, giving a direct correlation with differences in water-content and humidity. Results for 1922. A single series of cut quadrats was obtained in 1922; these were taken on August 4 to 15, beginning with the short-grass plains. A series of 7 quadrats of Bulbilis at Burlington yielded an average of 179 gm., while the average of a similar series at Phillipsburg was 260 gm. A series of mixed short and tall grass cuttings at these two stations yielded 263 and 365 gm. respectively. The differences in the height-growth of foliage and flower-stalks for both kinds of grasses were again like those recorded for the preceding years. A series of 8 representative quadrats of wheat-grass was also obtained at both Phillips¬ burg and Lincoln. The average height of the foliage and flower-stalks re¬ spectively was 15 and 26 inches at the former station, and 17 and 30 inches at the latter. The respective yields were 334 and 541 gm. per square meter. A series of 27 quadrats containing mixed tail-grasses, largely Andropogon scoparius, A. furcatus, and A. nutans, was secured from these two stations. Repeated measurements showed that the average height of the general level of the foliage was 3 to 8 inches greater at Lincoln (13 to 24 inches tall) than in the mixed prairie. The yield in grams per square meter was 287 and 413 at the two stations respectively. The average yield of all the quadrats for this year in the three grassland communities was 224, 311, and 447 gm. at Burlington, Phillipsburg, and Lincoln respectively. Table 53. —Growth and yield of crop-quadrats, 1923. Crop and station. Date of harvest. Height. Dry' weight. 1 Oats: in. gm. Lincoln. July 7 42 2,637.0 Phillipsburg.. . July 3 32 905.0 Burlington.... July 7 18 806.5 Wheat: Lincoln. July 7 42 2,395.0 Phillipsburg. . . July 3 29 1,296.0 Burlington.... July 7 24 801.5 Barley: Lincoln. July 2 43 1,671.5 Phillipsburg.. . July 3 37 1,292.5 Burlington.... July 7 26 1,038.5 1 Total of straw and grain. Five fields of maize were selected as representative of conditions about each of the three stations respectively, all of which were within a radius of 2 miles of the stations at which holard and other factor determinations were made. 10 stalks were then selected from various places in the field and the measure¬ ments recorded. The seventh leaf was selected for measurements of length and width. After completing the measurements on each stalk, the ear was husked 136 PHYTOMETRIC RESULTS. and placed in a sack, 10 ears being secured from each field. Finally, a typical stalk, as determined by the measurements of the 10 preceding, was selected, cut off at the ground line, cut into pieces, and after removing the ear but not the husks, was placed in another sack. These sacks were kept well ventilated, and after all had become thoroughly air-dried the weight of their contents was ascertained. Consistent and marked increases in height and diameter of stalk, height of ear, number and length of leaves, length and diameter of ear, as well as dry weight of stalks and ears, were found proceeding from west to east. The respective values for average weight of stalk and weight of ear were 320 and 82 gm. for Burlington, 413 and 110 gm. for Phillipsburg, and 496 and 183 gm. for Lincoln. Results for 1923. No clip-quadrats were taken from native vegetation in 1923, but three were made in plats of each of the grains at the three stations. As in previous years, there was a pronounced decrease in the height of all three species from Lincoln westward, the plants at Burlington averaging about half as tall as at Lincoln. With respect to weight, the differences were even greater, oats and wheat yielding a third as much as at Lincoln. The average production for the three crops was 2,234 gm. at Lincoln, 1,164 gm. at Phillipsburg, and 882 gm. at Burlington, in close correspondence with rain¬ fall, evaporation, and chresard. Table 54 .—Average yield of clip-quadrats in grams. 1920. 1921. 1922. Dominant type of vegetation. Lin¬ coln. Phil¬ lips¬ burg. Bur¬ ling¬ ton. Lin¬ coln. Phil¬ lips¬ burg. Bur¬ ling¬ ton. Lin¬ coln. Phil¬ lips¬ burg. Bur¬ ling¬ ton. Buffalo-grass. 290 98 235 266 207 260 179 Wheat-grass. • • • 541 500 606 457 400 541 334 • • • Mixed short and tall grasses. • • • 313 197 • . • • ■ • • • • • • • 365 263 Mixed tail-grasses. 458 410 755 477 413 287 Average (based on number of quad- rats). 458 378 183 603 402 353 447 311 224 Summary. In every case each grass or mixture of grasses yielded progressively less as the rainfall and holard decreased to the westward. The only exception is the case of buffalo-grass at Lincoln in 1921, and this has already been explained as an effect of grazing. Moreover, the averages for each year at the several stations show a graduated series, plant production increasing with increased efficiency of rainfall. However, it may be readily seen that the total yield at all of the stations was greater in 1921 than during the preceding or following year, an increase particularly noticeable in the case of the late-maturing tail- grasses. It was early determined that the water relations of soil and air were con¬ trolling, other factors being merely contributory. The yield of pure stands of GROWTH OF NATURAL COVER AND CROPS. 137 short-grasses (Bulbilis dadyloides and Bouteloua gracilis), wheat-grass (Agropyrum glaucum), mixed short and tall grasses, and mixed tail-grasses was found to decrease from the true prairie to short-grass plains directly with the chresard and inversely with evaporation. The same relation was de¬ termined not only for the smaller cereals (oats, wheat, and barley), but also for alfalfa and sweet clover, as well as for maize. The plant yield at each station during different seasons was closely correlated with the variations in rainfall and holard. Deficiencies in water-content were naturally most marked late in the summer, and their effect was especially to be found in the late- maturing tail-grasses at the eastern stations. The results obtained with crop-plants were comparable in practically every respect, demonstrating that the method was equally applicable to both native and cultivated species. Structural Response of Dominants and Subdominants. Scope and Significance. Comprehensive studies have been made of the leaf-structure of the great majority of the climax dominants of the grassland, as well as of a considerable number of serai dominants, subdominants, and trees. The material was obtained from both native and experimental plants, and in the case of the most important species represented the whole range of adjustment and adaptation, from salt-flat and swamp through low and high prairie to gravel-knoll at Lincoln, and from Nebraska City to Lincoln, Phillipsburg, Burlington, and Colorado Springs. As was to be expected, the grasses in general were very stable as to leaf-structure, the adjustment to the several climates occurring chiefly in function and growth, and hence in form. As a rule, distinct ecads were produced only in the series of edaphic stations with their more efficient differences, and the various species and genera naturally responded in very different degree even to these. Moreover, the effect of a particular habitat tended to be cumulative, and greater changes will probably appear in the course of the next few years. This is suggested by the existing adaptations shown by the leaves to the respective climates and recorded in the growth- forms and in the genera. With occasional exceptions, the leaf-structure is essentially the same for the dominants of each association, but differs markedly from one association to another. This seems to be a more recent adaptation that has been impressed upon the earlier one characteristic of each genus, with the consequence that it is necessary to distinguish three successive changes of varying degree in many dominants. Furthermore, this adapta- tional series affords a new approach to climatic shifts in the past and to the basic correlation of ungulate teeth and hoofs with changing vegetation. For these reasons, as well as because of the large amount of material available, it has seemed desirable to treat this phase of the investigation in a separate paper (Clements, 1924). 6. RESUME. PLAN AND METHODS. An endeavor has been made to determine experimentally the factors operating in the composition and sequence of the climax grassland associations lying between the Missouri River and the Rocky Mountains. Lincoln, Nebraska, was selected as representative of true prairie, Phillipsburg, in north-central Kansas, of mixed prairie, and Burlington, in eastern Colorado, of short-grass plains. The altitude rises from 1,100 feet at Lincoln to 1,900 feet at Phillipsburg and 4,160 feet at Burlington. Since the associations are zoned from east to west, precipitation is the chief factor in determining the type of grassland; it decreases through this series of stations from 28 through 23 to 17 inches annual mean. A fourth station was maintained in the edge of the subclimax prairie at Nebraska City, 50 miles southeast of Lincoln, where the precipitation is 33 inches. In addition to these climatic stations, a series of edaphic ones ranging through gravel-knoll, high prairie, low prairie, salt-flat, swamp, and cultivated field was maintained at Lincoln. Moreover, some reciprocal transplants from Colorado Springs, Colorado; Tucson, Arizona; and Berkeley, California, were made at several of these stations. Continuous records of the most important ecological factors, viz, precipita¬ tion, temperature of air and soil, humidity, and evaporation, were secured and frequent measurements were made of the holard to a depth of 4 feet, of light, and wind velocity, in addition to physical and chemical analyses of the soils. In the endeavor to interpret the various climatic and edaphic complexes in terms of plant activities, seeding and transplanting were carried out in different ways in the several communities. Seeds were sown in the undis¬ turbed grass cover, others were placed in especially prepared trenches, while still others were sown in denuded quadrats. Seedlings were transplanted and reestablished under favorable conditions, and finally, large blocks of sod of mature plants were also employed. Some individuals were favored by watering, while others were exposed unaided to the test of the new habitat. Frequent checks were applied, resulting in complete records of the plant’s activities. The investigation was begun in 1919 and continued until the close of the growing-season in 1923. Fruits and seeds of a large variety of native grasses, forbs, shrubs, and trees from a wide range of habitats were employed. These were kept dry, but subjected to winter temperatures at Lincoln, Nebraska, and the germinability of the seeds was determined previous to planting, more than 60 species being tested. Great variability in the vitality of the seeds from year to year was found, and this could be traced to differences in climate. A germination of 20 to 25 per cent was exceptional, while one of 10 to 15 per cent was quite usual. Experiments were also made to disclose the effect of depth of planting upon germination and establishment. With most grasses a depth exceeding 0.5 to 1 inch was detrimental, and most species, like the composites, did best at a depth of 0.12 to 0.25 inch. Surface sowing consisted of scattering the seeds in selected marked areas with a typical cover of vegetation, without disturbing the surface, and adding only enough debris to prevent the seeds from blowing away. Upon germina¬ tion the seedlings at once met the keen competition from the existing vegeta- 138 PLAN AND METHODS. 139 tion for water and nutrients below ground and for light above. Frequently the latter was as low as 5 to 10 per cent, even early in the summer. By planting the seeds in a trench 4 inches deep, for which the sod had been broken, the soil pulverized, and a good seed-bed prepared, competition both above and below ground was removed for a short time. Denuded quadrats 0.5 meter square were also employed, the native vegetation likewise being removed and a good seed-bed prepared to a depth of 4 inches. This method eliminated competition for light and also for water and nutrients for a time, but other unfavorable factors were introduced, such as higher temperatures and greater evaporation. To tide the plant over the most critical period in its life, that of ecesis, the method of seedling transplants was also employed. These were grown in 2.5 to 4 inch flower pots until they were 3 or 4 weeks old, when they were transplanted into a specially prepared trench without disturbing the root system and watered for a period of about 10 days. A fourth method, that of transplanting mature perennials, was also utilized. Blocks of sod 10 inches square and 8 inches deep were transplanted reciprocally among most of the stations, a control block being replanted in the area where the species was secured. As far as possible, sowing and transplanting were done preceding or coincident with the inception of new growth and at a time when the holard was distinctly favorable. COMMUNITIES. The vegetation at each station has been studied and described in detail, in connection with the measurement of physical factors. Ninety per cent of the cover at Burlington ( Bouteloua-Bulbilis association) consists of closed mats of short-grasses (mostly Bulbilis) forming a carpet seldom over 4 inches deep. At Phillipsburg ( Bouteloua-Stipa association) tail-grasses, chiefly Andropogon, alternate with or form a layer above the shorter buffalo and grama grasses, the former reaching a height-level of 14 inches by midsummer. In the true prairie ( Stipa-Koeleria association) the vegetation is distinctly of the tail- grass sod type. The chief genera are Stipa, Koeleria, Andropogon , and Sporobolus, with which are associated subdominants to form extensive societies. By June 1 the grasses have a height-level of 6 or 8 inches and the upper story of forbs attains a height of 15 to 22 inches. The subclimax prairie at Nebraska City (. Andropogon associes), in addition to the true-prairie species which grow somewhat more rank, is characterized by the regular occurrence of Andropogon furcaius and Panicum virgatum on highland, while the various potential scrub societies usually held in check by mowing and fire, indicate its t rue relationship to forest. The low prairie at Lincoln is dominated, almost to the exclusion of other species, by Andropogon furcatus and nutans, Panicum virgatum , and Spartina cynosuroides, and is actually an edaphic postclimax (Plant Succession, 109), closely resembling the subclimax grass¬ land. The swamp area is covered by the Spartina consocies bordered by a zone of Poa pratensis , but the bluegrass has been replaced by Spartina during the course of the experiment. The vegetation both here and in low prairie reaches a height of 4 or 5 feet. The salt-flat, on an area near the low prairie, had just enough sodium chloride, combined with an unfavorable soil structure, to exclude most prairie species, and is occupied by an open growth of Dis- tichlis spicata and dwarf Agropyrum glaucum , seldom exceeding 5 to 8 inches 140 RESUME. in height. The gravel-knoll is covered with mats of Bouteloua gracilis inter¬ mixed with Bouteloua hirsuta and often also with sparse tail-grasses in the intervals. These grow in porous sandy to gravelly drift soil, which forms the crest of a steep slope to the south. It is a miniature short-grass area of edaphic nature surrounded by true prairie. PHYSICAL FACTORS FOR 1920. A comparison of the physical factors during 1920 at the three major climatic stations shows that conditions for plant-growth in respect to rainfall, holard, temperature, humidity, wind, and evaporation were most favorable at Lincoln, intermediate at Phillipsburg, and least favorable at Burlington. These condi¬ tions are indicated by the native vegetation and are borne out by results of these experiments. Holard and humidity were found to be the controlling factors in plant growth, all others being secondary. The precipitation is not only 5 inches less at Phillipsburg and 11 inches less at Burlington than at Lincoln, but, owing to lighter showers, greater run-off, and increased evapora¬ tion, it is also progressively less efficient westward. While at Lincoln sufficient chresard was found at all times and at all depths to 4 feet for good growth, a period of midsummer drought occurred at Phillipsburg, and at Burlington the holard was favorable only until June (to a depth of 2 feet only), after which marked deficiencies were of frequent occurrence. Evaporation was lowest at Lincoln (9 to 25 c. c. daily), intermediate at Phillipsburg (11 to 32 c. c.), and highest at Burlington (23 to 60 c. c.). ECESIS DURING 1920. Both germination and establishment at the three stations were in the sequence of increasing water-content. Germination by all methods of planting averaged 86, 80, and 38 per cent, and establishment 42, 33, and 25 per cent at Lincoln, Phillipsburg, and Burlington respectively. Surface sowing gave both the poorest germination and poorest establishment, and seeding in denuded quadrats the best. In most cases better growth occurred at Lincoln than at the other two stations. Germination and growth in cultivated areas adjoining the several grassland stations were much better than in the sodded areas, but the sequence of the stations was the same. In order to more fully understand the causes for the success or failure of seedlings, the root habit was studied both in native grassland and in cultivated soil. During 1918, when seedling Bouteloua hirsuta, Andropogon scoparius, furcatus, and nutans, and Sporoholus asper had reached heights of 6 to 10 inches in the grassland at Lincoln, the bulk of the root systems was found to occur at no greater depths than 8 to 18 inches. In cultivated areas seedlings of Bouteloua, Sporoholus, and Liatris, for example, reached heights of 3 or 4 inches in 44 days and the roots reached depths of 7 to 11 inches. By mid¬ summer they were 8 to 18 inches tall, and the root depth 20 to 33 inches. The excellent growth made by all the species when properly spaced in fertile, cultivated soil with adequate holard emphasizes the keen competition pre¬ vailing in stabilized grassland. In the natural cover, as well as in the trench and denuded quadrat, most of the grasses require 2 or more years to produce seed, but under the former environment 73 per cent bore seed the first season. Because of the lack of root competition encountered by plants in the prairie, RESULTS FOR 1920 AND 1921. 141 the species in cultivated soil not only suffered less from drought, but also made a greater growth and tillered more heavily. For example, Andropogon nutans was 7 to 10 inches tall at the end of the first season in the true-prairie quadrats and 12 to 17 inches in the adjacent cultivated field. The tap-root of Gleditsia reached a depth of 40 inches by midsummer and that of Onagra 44 inches at the end of the first season, while Andropogon scoparius had a maximum lateral spread of nearly 1.5 feet, was rooted abundantly to 26 inches depth, and had a maximum penetration of 4 feet. These are typical examples of many findings. In the deep, mellow loess soil at Peru, Nebraska, the root penetration of seedlings was even greater. Of the 40 blocks of sod (9 species) transplanted at Burlington no species died, but only 53 per cent (unwatered sods) had flower- stalks, as compared with 70 per cent at Lincoln, where none of the 13 trans¬ planted species died (23 blocks). In the edaphic series, the low prairie exhibited a chresard 5 to 10 per cent greater than that of high prairie, and it also exceeded that at Nebraska City, where unusually dry weather prevailed. However, the humidity was higher at Nebraska City and the evaporation correspondingly less than on low prairie, which was more mesophytic than high prairie. The holard on the gravel- knoll and at Burlington was not very different, frequently falling to the hygroscopic coefficient, the short-grasses on both areas drying and turning brown late in July or in August. The average germination was highest at Nebraska City (93 per cent), intermediate on low prairie (77; 86 per cent on high prairie), and least on the gravel-knoll (71 per cent). The average of establishment was greatest in the cultivated area at Lincoln (100 per cent), 75 per cent on low prairie, 70 per cent at Nebraska City, and only 15 per cent on the gravel-knoll. Thus, all stations except the last exceeded high prairie (42), while Burlington with 25 per cent exceeded the gravel-knoll. As to sod transplants on the gravel-knoll, all showed repeated rolling and drying back of the leaves; only a few flowered, and these did so earlier than elsewhere. On low prairie all grew well and blossomed. The transplants suffered from drought in the salt-flats, exhibiting wilting and yellowing of the leaves and dwarfed stature. Seven species, mostly from upland, succumbed to insufficient aeration in the swamp, and four in the Poa zone. All survived at Colorado Springs, where they were transplanted late, but none made a good growth. In the case of Spartina cynosuroides transplanted from swamp to high prairie two years earlier, many leaves reached a height of 40 inches and the roots were abundant to 9 or 10 feet depth, being much more branched than in lowland. Panicum virgatum showed a similar growth. PHYSICAL FACTORS FOR 1921. The season of 1921 was very favorable for growth, no marked drought periods occurring at Lincoln, where at least 5 and usually 8 or 10 per cent of chresard existed at all times and at all depths to 4 feet. At Phillipsburg the echard was approached once in July and twice in August, no water being available to a depth of 4 feet in late summer. As usual conditions were much worse at Burlington. At no time was water available in the third or fourth foot of soil, while after June 30 it was depleted repeatedly above the hard-pan at a depth of* 2 feet. The humidity was highest at Lincoln and lowest at Burlington, where it often dropped to 10 or 20 per cent in the late afternoon. 142 RESUME. Evaporation increased inversely with humidity and soil moisture, ranging from 8 to 27 c. c. daily at Lincoln, 8 to 43 c. c. at Phillipsburg, and 18 to 62 c. c. at Burlington. The remaining conditions for plant growth were also most favorable at Lincoln, intermediate at Phillipsburg, and poorest at Burlington. ECESIS DURING 1921. The germination results agreed with those of preceding years, Lincoln being highest (81 per cent), Phillipsburg second (68 per cent), and Burlington last (43 per cent). However, owing to the unusually favorable rainfall following germination at Phillipsburg, establishment was highest here (65 per cent), intermediate at Lincoln (40 per cent), and least at Burlington (7 per cent). No seedling transplants survived at Burlington, but 54 per cent lived at Lincoln, and 75 per cent at Phillipsburg. Although the surface-sown seeds germinated better than those in the trench, establishment, as in 1920, was least on the surface, next in the trench, and best in denuded quadrats, the latter exceeding that of seedling transplants. As to the 1920 plantings, the average survival during 1921 was greatest at Lincoln and least at Burlington. An examination of the grasses and forbs during the second year in the quadrats at Phillipsburg showed that they were well established, the roots penetrating to depths of 2.5 to 4 feet. Sods on the high prairie all did well, but 4 species died at Burlington, and the rest, including those that had been repeatedly watered, were represented by mere remnants of the original fine blocks. Of those transplanted to high prairie in 1920, all survived and some increased their area, but at Burlington 1 died, while all suffered from drought and were considerably dwarfed. By mid¬ summer of the second season's growth in the short-grass plains, the roots were confined to the surface 2 to 2.5 feet, except in the watered area, where they penetrated 5 to 7 feet. As to the other stations, the season at Nebraska City was one of drought, though the soil usually had a 10 per cent chresard at all depths to 4 feet. This exceeded the water-content on low prairie, where the chresard in the surface 6 inches was nearly exhausted in June. Conditions on the gravel-knoll were more favorable than during 1920. Evaporation at Nebraska City and low prairie was lower than on the high prairie and was highest on the gravel- knoll. The average germination at the three stations was very similar (63 to 76 per cent) and in all cases less than on high prairie (81 per cent). The rate of establishment was in the following order: Nebraska City 41 per cent, high prairie 40 per cent, low prairie 38 per cent, and gravel-knoll 29 per cent, the last considerably exceeding Burlington (7 per cent). Seedling transplants gave the best growth on low prairie (92 per cent), next at Nebraska City (57 per cent), then high prairie (54 per cent), and least in gravel-knoll (14 per cent). Four species of sod transplants died on the gravel-knoll, and three were shaded out in low prairie, while all survived on the salt-flat and in the swamp, the latter being much drier than in 1920. Of the 1920 transplants, no species died on the gravel-knoll, one succumbed to the shade in low prairie, one died on the salt-flat, and 6 in the swamp and bluegrass zone. PHYSICAL FACTORS FOR 1922. The season of 1922 was fairly favorable for growth, except for the latter part, when severe drought occurred at all the stations, though it was relatively RESULTS FOR 1922. 143 less marked on the Great Plains. A margin of at least 5 per cent and more, usually 7 to 11 per cent, chresard existed at all times to a depth of 4 feet in the high prairie at Lincoln, while at the mixed-prairie station drought began late in June and continued throughout the season. There was often no chresard to a depth of 4 feet. The soil in spring and early summer was as usual quite moist at Burlington, but deficiencies were marked and practically continuous after the middle of June. As usual, the factors were most favorable in true prairie and least so on the Great Plains. ECESIS DURING 1922. As for the two years preceding, the average germination under all methods of planting was highest at Lincoln (74 per cent), intermediate at Phillipsburg (63 per cent), and least at Burlington (44 per cent). In surface sowing Phil¬ lipsburg slightly outranked Lincoln, due to differences in soil structure. This also accounts for the greater survival of seedling transplants at the former, the percentages being 78, 70, and 26 respectively. Establishment was decidedly highest in true prairie, viz, 62 per cent; it was but 39 per cent in mixed prairie, and 26 per cent in the short-grass plains. Germination aver¬ aged best on the surface, next in denuded quadrats, and least in the trench, but establishment was in the following order: denuded quadrats, trench, and surface, the last giving the lowest percentage. Bouteloua gracilis made the best growth in the quadrats at Burlington and the least at Lincoln, but for all of the other species the growth-rate was in the sequence of increasing water- content, i. e., eastward. Andropogon nutans, Kuhnia glutinosa, Bouteloua racemosa , and Panicum virgatum were 4 to 12 inches tall at Lincoln, 3 to 9 inches at Phillipsburg, and 1 to 5 inches at Burlington. The survival for the 1920 surface-sowings was 33 per cent at Burlington only; for the trench it was 33 per cent at Lincoln and Phillipsburg, and for the quadrats 55 per cent at Lincoln and 43 per cent at Burlington, the plants at Phillipsburg having been mostly excavated for root study. The average survival of 1921 plantings and seedling transplants was greatest at Phillipsburg and lowest at Burlington. Two species died out of the 26 trans¬ planted to high prairie in 1922; these were both dicotyls with strong tap¬ roots. All of the 13 species of grasses transplanted to Burlington survived, although they flowered less than at Lincoln and nearly all were dwarfed. Of the 1920 sod transplants, 2 died on high prairie during 1922, but only 1 at Burlington, although many more individuals out of the 40 transplanted blocks of sod succumbed here. None of the 1921 transplants died on the high prairie; several increased their territory, and 12 of the 16 species blos¬ somed. At Burlington several individuals died, but only 1 species vanished entirely. Agropyrum spicatum and Bouteloua racemosa were the sole survivors of the unwatered lot. As to the other stations, the season at Nebraska City was one marked by drought in June and August. However, the soil usually had a chresard of 8 per cent below the first foot, but that in the surface foot, which most critically affects seedlings, was practically exhausted two or three times during the summer, being less favorable than in 1921. Conditions as to the holard were more favorable on low prairie, where a considerable excess prevailed over that on high prairie. On the gravel-knoll the grama grasses dried during the 144 RESUME. drought, as no moisture was available in the first 2 feet of soil. The germina¬ tion of surface-sown seed was not greatly different at the three stations (70 to 84 per cent), this method of sowing giving the best germination. The average germination was 67 per cent at Nebraska City, 75 on low and 74 on high prairie. Establishment gave the usual sequence with respect to method of planting. It averaged 46 per cent at Nebraska City, 62 per cent on high prairie, and only 50 per cent on the more densely shaded low prairie, where light values ranged from 2 to 10 per cent. Seedlings did best on high prairie (70 per cent) where they were watered, next at Nebraska City (59 per cent), and poorest on low prairie (45 per cent). Three of the sod transplants on low prairie succumbed to the dense shade, as did 4 in the swamp, while one died on the salt-flat. Of the 1920 sods, 3 died on the gravel-knoll, 4 were shaded out in low prairie, and 1 died in the salt-flat, but no further mortality occurred in the rapidly drying swamp. Of the 1921 transplants, 3 more died on the gravel-knoll, 5 in low prairie, 1 in the salt-flat, and 2 in the swamp. BEHAVIOR DURING 1923. In 1923, a season with high spring and summer rainfall at the western stations, the plants made an excellent growth. Not only was the rainfall above the normal, but the rains were also well distributed, drought periods being fewer and shorter than usual. At Lincoln the precipitation for spring and summer averaged below normal. The preceding fall and winter had been one of unusually severe drought, and high losses were sustained at all the stations. In the major climatic stations these were greatest at Burlington, intermediate at Phillipsburg, and least at Lincoln. For example, losses of the 1921 plant¬ ings and transplants were 43, 28, and 13 per cent in the above sequence of stations. The mortality at all stations averaged greatest for the plants of a single year’s establishment, next for those 2 years old, and least for those established for 3 seasons, e. g., the respective losses at Burlington being 85, 43, and 16 per cent. The survival was greatest in denuded quadrats, this method even outranking that of transplanting blocks of sod. In the survival of surface-sown plants for all years Phillipsburg ranked first (14 per cent), Burlington second (11 per cent), and Lincoln last (9 per cent). The average germination was respectively 79, 45, and 76 per cent, Lincoln taking second and Burlington third place. The average germination in trenches decreased from Lincoln westward, the percentages being 85, 59, and 21 respectively. The average survival for all years at the end of 1923 was in the same order, the percentages being 27, 21, and 0 respectively. Germination in denuded quadrats was greatest at Lincoln for every year except one, and least at Burlington, with one exception, the percentages for the three stations being 81, 73, and 60. With a single exception, establishment was likewise in the same order, viz, 72 per cent in true prairie, 60 per cent in mixed prairie, and 43 per cent in the short-grass plains, and this was also true of survival at the end of 1923, the percentages being 44, 22, and 21 respectively. Hence, while surface-sowing ranked highest in germination, it resulted in the lowest estabishment. Transplanted seedlings usually did best at Phillipsburg, the average survival at the end of 1923 being 44 per cent as compared with 37 per cent at Lincoln and 3 per cent at Burlington. RESULTS FOR 1923. 145 Permanent establishment by the end of 1923 under the three methods of planting was 44, 27, and 9 per cent respectively at Lincoln, 22, 21, and 14 per cent at Phillipsburg, and 21,0, and 11 per cent at Burlington. This was always greatest in the denuded quadrats and next greatest in the trench. The results of seeding and transplanting at Nebraska City and on low prairie are not greatly different, although often somewhat better than on high prairie, while ecesis on the gravel-knoll was often only slightly higher than that at Burlington. At the latter, as on the gravel-knoll, drought was the deciding factor, while on low prairie, as well as at Nebraska City, light played the dominant role. Certain species were found to survive more frequently than others. Out of the total plantings at Nebraska City, Lincoln, Phillipsburg, and Burlington, beginning with 1920, Andropogon nutans survived in 35 different places by the end of 1923. The survival of other species was in the following order, the figures indicating the different number of places in which they survived : Bouteloua gracilis 18, B. racemosa 15, B. hirsuta 14, Sporobolus asper 11, Aristida purpurea, Elymus canadensis, and Liatris punctata 10 each, Andro- pogon scoparius 8, A. furcaius 7, Petalostemon candidus 6, Stipa spartea, Desmodium canescens, and Liatris scariosa 5 each, Kuhnia glutinosa 4, Panicum virgatum, Stipa viridula, Symphoricarpus occidentalis 3 each, Gleditsia triacanthus, Symphoricarpus vulgaris, Koeleria cristata, Ratibida columnaris, Robinia pseudacacia, and Acer saccharinum 2 each, and Stipa comata, Agropyrum glaucum, Pinus ponderosa, and Acer negundo each only 1. As regards sod transplants, by the fall of 1923 over half of the 1920 plantings at Burlington had died. Agropyrym glaucum alone remained in the un¬ watered area. The 1921 transplants, half of which were watered, lost 15 out of 22 individuals, Agropyrum and Bouteloua racemosa alone surviving in the unwatered area, while practically all of the rest showed clearly that they would not last much longer. 83 per cent of the 1922 transplants were winter- killed, and at the end of the second season Bouteloua gracilis and remnants of 2 other species alone remained. These results clearly reveal the uncongenial nature of the Great Plains, in so far as prairie species are concerned. Relatively few species died on the high prairie, even Spartina cynosuroides and Panicum virgatum forming seed. Of the 1920 plantings only 6 individuals had died, including 2 plants of Distichlis spicata, which never flourished out of its owtl saline habitat. Of the 1921 plantings a single one succumbed. Six forbs among the 1922 plantings died, and 1 lot of Bouteloua gracilis. Several of the species had considerably extended their area. On the gravel-knoll the mortality was much greater, although here likewise several low-prairie species became permanently established, their roots ex¬ tending well into the clay subsoil below 4 feet. Three species of the 1920 and 7 of the 1921 lot succumbed, but others planted in 1919 still survived, Bulbilis and Agropyrum being quite at home. Nearly all were somewhat dwarfed, and they began growth and matured earlier than at the other Lincoln stations. Exceedingly interesting results were obtained on low prairie. Of the 14 species and 28 blocks of sods transplanted in 1920, all flourished, including Bulbilis and Bouteloua gracilis. By the second season, however, 1 species was shaded out, 4 the next, and 1 in 1923, so that by the end of the fourth summer the dominants of the low prairie alone, viz, Andropogon nutans and furcatus, 146 RESUME. Panicum virgatum, Elymus canadensis, and Spartina cynosuroides were flourishing, Stipa, a single Koeleria, and Andropogon scoparius surviving from the high prairie. Of the 1921 transplants, 5 subclimax dominants developed normally, but of the 10 species remaining all had succumbed but 3 represented by mere remnants. Two or three years were sufficient for the tail-grasses to overshade and cause Bulbilis and Bouteloua gracilis and hirsuta to disappear. This applies also to several true-prairie dominants. Of the 1922 transplants a total of 11 species had succumbed by the end of 1923, the process of elimina¬ tion being well under way. On the salt-flats, while the plants were able in most cases to tolerate condi¬ tions (only 6 species among the three years’ plantings succumbing), they made a very poor growth, did not extend their territory, and seeded very poorly. For the most part they had a paler color and in every way reflected the unfavorable environment, except the indigenous Distichlis alone. Transplants in the swamp lost heavily, at first-from poor aeration and later from lack of light, after a dam was built above the experimental area and the swamp dried out. High-prairie species were most affected and died in greatest numbers, the flower-stalk often rotting off at the surface of the saturated soil. By 1923, Spartina, Panicum virgatum, and Elymus canadensis, all of which grew well, were the only survivors among the 13 species planted in the swamp, while these and Poa pratensis were the 4 survivors of a similar lot in the Poa zone, which was now overgrown with Spartina. During the course of the experiment it was found that species of widely different families exhibit the phenomenon of dormancy, the seeds lying in the soil for one or more years before germinating. Among the most con¬ spicuous cases the following may be enumerated: Robinia pseudacacia, Gledit- sia triacanthus, Petalostemon candidus, and Lespedeza capitata among legumes; Sporobolus asper, Andropogon nutans, Panicum virgatum, and Andropogon furcatus among grasses; and likewise Kuhnia glutinosa, Corylus americana, and Acer negundo. While the volunteer growth of these plants during the second or third year has not been recorded in detail, because of the risk of confusion, it may be stated that in general they suffered the same fate as the original plantings. BEHAVIOR OF TREES AND SHRUBS. The fate of trees and shrubs planted or transplanted by rhizomes into grass¬ land is of especial interest. The trees employed were Acer negundo, A. saccharinum, Gleditsia triacanthus, Robinia pseudacacia, and Pinus ponderosa, together with the following shrubs: Corylus americana, Symphoricarpus occi- dentalis, and S. vulgaris. At Burlington none survived longer than the first season, and at Phillipsburg the only survivors at the end of 1923 were the 2 species of Symphoricarpus planted in 1921. Corylus (delayed growth), Robinia, Gleditsia, and Acer saccharinum of the 1922 planting, and Robinia of 1921 survived on high prairie. On low prairie Symphoricarpus occidentalis of the 1921 planting and Acer saccharinum of 1922 alone survived. None lived on the gravel-knoll, but at Nebraska City both species of Symphoricarpus of 1921 and Acer negundo and Pinus ponderosa of 1922 were alive at the end of the summer of 1923. However, as shown by a series of earlier experiments with trees at Lincoln, the results of which are as yet unpublished, the possibility of CLIMATIC STATIONS. 147 tree-growth in stabilized grassland is almost nil, since tree seedlings are killed by drought on upland and are shaded out on lowland. GERMINATION AND SURVIVAL AT THE CLIMATIC STATIONS. The percentage of germination for each year and for each method and the percentage of survival for the first year of planting as well as succeeding years are given in table 55. The percentage of germination under surface sowing was greatest at Lincoln during 1920, and at Phillipsburg during 1921 and 1922, while it was least at Burlington for all three years. The averages for the three years are Phillipsburg 79 per cent, Lincoln 76 per cent, and Burlington 45 per cent (table 56). In regard to establishment, Burlington ranked first in 1920 (clearly an erratic case), Phillipsburg in 1921, and Lincoln in 1922. As to the survival of plants of all years, at the end of 1923 Phillipsburg was first with 14 per cent, Burlington second, 11 per cent, and Lincoln third, 9 per cent. The greater germination and establishment at Phillipsburg were probably due to the structure of the soil (p. 12). Germination in the trenches was highest at Phillipsburg in 1920, but at Lincoln for the other years, Burlington coming last. The average germination for the three seasons was greatest at Lincoln (85 per cent), next at Phillips¬ burg (59 per cent), and lowest at Burlington (21 per cent). Survival in 1920 and for 1923 was greatest at Lincoln, but in 1921 greatest at Phillipsburg, Burlington regularly ranking third. However, in the survival for all plantings at the end of 1923, the sequence is Lincoln, Phillipsburg, and Burlington, with respective percentages of 27, 21, and 0. Germination in the denuded quadrats was greatest at Lincoln during 1920 and 1922, and least at Burlington, except in 1921, when it exceeded Phillips¬ burg and Lincoln. The average germination for the three years at the several stations was 81, 73, and 60 per cent, in the order of decreasing rainfall. Establishment was also greatest at Lincoln every year and least at Burlington, except during 1922, when the latter exceeded that at Phillipsburg. The average establishment for the three years was 72 per cent in the true prairie, 60 per cent in mixed prairie, and 43 per cent in the short-grass plains. The percentage of survival for all plantings in quadrats for the three years was, in the above order, 44, 22, and 21. Transplanted seedlings did best at Phillipsburg both during 1921 and 1922 and poorest at Burlington. Moreover, the average survival at the end of 1923 was greatest at Phillipsburg (44 per cent), next at Lincoln (36 per cent), and least at Burlington (3 per cent). The explanation of the better survival in the mixed prairie seems clearly to be sought in the more favorable soil structure. With reference to the success of the various methods, surface sowing gave the poorest results as to germination in 1920, but the best the other two seasons, the quadrat ranking second (first in 1920), and the trench third. On the other hand, establishment was always greatest in the denuded quadrat, second in the trench, and least on the surface. The survival for all years at the end of 1923 was greatest in denuded quadrats, and least on the surface. At Lincoln under the three methods it was 44, 27, and 9 per cent, re¬ spectively, at Phillipsburg 22, 21, and 14 per cent, and at Burlington 21, 0, and 11 per cent. The seedling-transplant method was somewhat less suc¬ cessful than sowing in quadrats, but more satisfactory than sowing in trenches. 148 RESUME. The persistence of plants is high after surviving a single growing-season, but the mortality is also usually greater during the second winter and summer than for any year afterward. During the three years the average germination for all methods was greatest in true prairie (81 per cent), second in mixed prairie (70 per cent), and least at Burlington (42 per cent). The average establishment was, in the same sequence of stations, 48, 46, and 20 per cent, as was also the average survival at the end of 1923, viz, 27, 19, and 11 per cent. Thus, the experimental evidence clearly indicates that the best conditions for growth are to be found in true prairie and the least favorable ones in the short-grass plains, conditions in the mixed prairie being intermediate. GERMINATION AND SURVIVAL AT THE EDAPHIC STATIONS. Results from the series of edaphic stations show that the germination of surface-sown seeds was very similar in the low prairie and at Nebraska City during the three years, where it slightly exceeded that on the high prairie at Lincoln. The survival at the end of the first season averaged greater at both Table 55. —Per cent of germination and survival at all stations, 1920-1923. Surface sowing. Trench sowing. Station. P. ct. ger¬ mina¬ tion. P. ct. survival. P. ct. ger¬ mina¬ tion. P. ct. survival. 1920. 1921. 1922. 1923. 1920. 1921. 1922. 1923. 1920. Lincoln (high prai¬ rie) . 66 0 92 45 42 33 33 Phillipsburg. 50 0 . . . , . , 100 33 33 33 33 Burlington. 46 33 33 33 33 15 0 . . . . . . 1921. Lincoln (high prai¬ rie) . 79 9 (9) (9) 92 45 36 36 Phillipsburg. 100 • . 83 50 25 43 . , 50 33 17 Burlington. 38 • . 0 • • • • 15 • • 0 • . • • 1922. Lincoln (high prai¬ rie) . 82 50 17 70 63 13 Phillipsburg. 86 • . • • 42 16 35 . . • . 25 13 Burlington. 50 • • . . 0 • • 32 . , . . 14 0 1920. Gravel-knoll (Lin¬ coln) . 71 30 0 Low prairie (Lin¬ coln) . 75 33 (17) (17) (17) 57 100 50 50 50 Nebraska City. ... 80 50 38 (38) (38) 100 60 30 20 20 1921. Gravel-knoll (Lin¬ coln). 55 17 0 Low prairie (Lin¬ coln) . 78 14 0 0 75 17 Excavated for roots. Nebraska City. ... 78 29 29 29 50 , , 33 17 17 1922. Low prairie (Lin¬ coln). 84 25 6 60 64 43 Nebraska City. ... 70 • • 50 21 62 33 15 EDAPHIC STATIONS. 149 Table 55. —Per cent of germination and survival at all stations, 1920-1923 —Continued, Denuded quadrats, sowing. Seedling trans- Station. P. ct. ger¬ mina¬ tion. P. ct. survival. plants, p. ct. survival. 1920. 1921. 1922. 1923. 1921. 1922. 1923. 1920. Lincoln (high prairie). 100 80 55 55 55 Phillipsburg. 90 67 67 Exca vated Burlington. 54 43 43 for r 43 oots. 43 1921. Lincoln (high prairie). 71 65 30 30 54 39 23 Phillipsburg. 61 63 31 31 75 58 50 Burlington. 76 21 11 5 0 1922. Lincoln (high prairie). 71 72 48 70 50 Phillipsburg. 69 50 36 78 37 Burlington. 50 64 14 26 5 1920. Gravel-knoll (Lincoln). 71 0 Low prairie (Lincoln). 100 91 82 73 45 Nebraska City. 100 100 90 80 70 1921. Gravel-knoll (Lincoln). 71 40 30 20 14 14 14 Low prairie (Lincoln). 75 83 42 25 92 69 31 Nebraska City. 68 60 40 40 57 57 50 1922. Low prairie (Lincoln). 81 60 50 45 39 Nebraska City. 69 54 36 59 45 ¥ Note: Numbers in parenthesis indicate that the plants had merged into the native sod and were not traced further. stations (low prairie 24, Nebraska City 43 per cent) than on high prairie (20 per cent), while the average survival at the close of 1923 was respectively 8, 29, and 9 per cent. The average germination in the trenches for all years was least on the gravel- knoll, 63 per cent; it reached 64 per cent in low prairie and 71 per cent at Nebraska City, as compared with 85 per cent on high prairie. However, ecesis on the gravel-knoll was less than half that on high prairie (23 in contrast to 51 per cent), while in low prairie and at Nebraska City it was 60 and 42 per cent respectively. The survival at the end of 1923 was nil on the gravel- knoll, 31 per cent in low prairie, and 17 per cent at Nebraska City, as com¬ pared with 27 per cent on high prairie. Germination in denuded quadrats averaged 71, 85, and 79 per cent on gravel knoll, low prairie and Nebraska City respectively, in contrast to 81 per cent on high prairie. The corresponding percentages were 20, 78, and 71, as against 72 on high prairie, the survival at the end of 1923 being 10, 40, 49, and 44 per cent respectively. The ecesis of seedling transplants was 14 per cent on the gravel-knoll (1921 only), 68 per cent in low prairie, and 58 per cent at Nebraska City, as compared with 62 per cent on high prairie. The respective survival at the end of 1923 was 14, 35, and 48 per cent, as against 36 per cent on high prairie. Thus, 150 RESUME. Table 56.— Summary of percentages of germination and survival. Surface Sowing. Station. Lincoln (high prairie). Phillipsburg. Burlington. Low prairie (Lincoln). Nebraska City. Year of plant¬ ing. 1920 1921 1922 1920 1921 1922 1920 1921 1922 1920 1921 1922 1920 1921 1922 P. ct. germina¬ tion. 66 79 82 50 100 86 46 38 50 75 78 84 80 78 70 Aver¬ age. 76 79 45 79 76 Survival at end of first year. 0 9 50 0 83 42 33 0 0 33 14 25 50 29 50 Aver¬ age. 20 42 11 24 43 Survival at end of 1923. 0 9 17 0 25 16 33 0 0 17 0 6 38 29 21 Aver¬ age. 9 14 11 8 29 Trench Sowing. Lincoln (high prairie). Phillipsburg. Burlington. Gravel-knoll (Lincoln), Low prairie (Lincoln). Nebraska City. 1920 1921 1922 1920 1921 1922 1920 1921 1922 1920 1921 1920 1921 1922 1920 1921 1922 92 92 70 100 43 35 15 15 32 71 55 57 75 60 100 50 62 85 59 21 63 64 71 33 51 36 5 23 60 42 33 36 13 33 17 13 0 0 0 0 0 50 Excav. 43 20 17 15 27 21 0 0 31 17 Denuded Quadrats. Lincoln (high prairie). Phillipsburg. Burlington. Gravel-knoll (Lincoln). Low prairie (Lincoln). Nebraska City. 1920 1921 1922 1920 1921 1922 1920 1921 1922 1920 1921 1920 1921 1922 1920 1921 1922 100 71 71 90 61 69 54 76 50 71 71 100 75 81 100 68 69 81 73 60 71 85 79 80 65 72 67 63 50 43 21 64 0 40 91 83 60 100 60 54 72 60 43 20 78 71 55 30 s 48 < ! Excav. 31 36 43 5 14 0 20 45 25 50 70 40 36 44 22 21 10 40 49 TRANSPLANTS. 151 Table 56 .—Summary of 'percentages of germination and survival —Continued. Seedling Transplants. Station. Lincoln (high prairie) Phillipsburg. Burlington. Gravel-knoll (Lincoln).. . Low prairie (Lincoln) Nebraska City. Year of plant¬ ing. P. ct. germina¬ tion. Aver¬ age. Survival at end of first year. 54 70 75 78 0 26 14 92 45 57 59 Aver¬ age. 62 77 13 14 68 58 Survival at end of 1923. 23 50 50 37 0 5 14 31 39 50 45 Aver¬ age. 36 44 3 14 35 48 no consistent differences are to be found between the low-prairie station at Lincoln, which is really a postclimax stage in the true-prairie climate, and the subclimax prairie at Nebraska City, both averaging somewhat better than the high-prairie station at Lincoln. The gravel-knoll station, climatically a preclimax to the true prairie, excelled Burlington in percentage of germina¬ tion and establishment, but barely equaled it in survival at the end of 1923. SURVIVAL OF SOD TRANSPLANTS. While the number of species lost at Burlington from the 1920 transplanting was the same as that at Lincoln, the individual losses were far greater, owing to the transplanting of a larger number of blocks of each species at the plains station. The relative losses for 1921 were representative. Losses in individ¬ uals on the gravel-knoll were consistently high, while those on low prairie were similarly high, though the cause for the latter was not drought, but shade. The losses in the swamp were greatest during 1920, because of poor aeration due to an excessive amount of water, while later losses were caused almost entirely by the density of shade. While in general the high-prairie species were the ones to disappear from swamp and low prairie, no consistent loss of individual species occurred at the other stations. Excluding the losses of 1923, owing to the very dry fall and winter, the following species suffered total loss in as many instances as are indicated by the number: Koeleria cristata 14; Andropogon scoparius 7; Elymus canadensis, Poa pratensis, Agropyrum glaucum, and Bouteloua racemosa 6 each; Distichlis spicata 5; Stipa spartea and Andropogon nutans 4 each; Bouteloua gracilis, Panicum virgatum, Andropogon furcatus 3 each; and Bouteloua hir^uta, Bulbilis dacty- loides and Sporobolus asper 1 or 2 each. PHYTOMETRIC METHODS. In spite of the abnormal weather during 1923, the investigation of the transpiration of standard plants and sod-cores of native and cultivated spe¬ cies, and of the growth of natural cover and crops, as shown by cut-quadrats, yielded results of value in confirmation of those obtained by the methods of experimental vegetation. This was especially true of the cut-quadrats, as 152 RESUME. these also extended through a series of nearly normal years. They not only demonstrated the close dependence of grassland and crop communities upon rainfall and water-content, but also the intimate relationship of the various prairie climaxes and climates. The evidence drawn from germination, ecesis, function, and growth was corroborated in large measure by the adaptation shown in leaf-structure, though this was more or less obscured by the fixity of type in certain species. Table 57 .—Summary of sod transplants. Station. No. of No. of Loss of species (not individuals). species. trans¬ plants. 1920. 1921. 1922. 1923. 1920. Lincoln (high prairie). 13 23 0 0 2 0 Burlington. 9 40 0 1 1 0 Gravel-knoll. 11 28 0 0 3 0 Low prairie. 14 28 0 1 4 1 Salt-flat. 13 17 0 1 1 0 Swamp. 13 13 7 2 0 1 Poa zone in swamp. 13 13 4 4 0 0 1921. Lincoln (high prairie). 15 16 0 0 1 Burlington. 11 22 4 1 3 Gravel-knoll. 14 18 4 3 0 Low prairie. 15 16 3 5 1 Salt-flat. 15 18 0 1 1 Swamp. 14 14 0 2 7 Poa zone in swamp. 13 13 0 0 6 1922. Lincoln (high prairie). 26 30 2 9 JmJ Burlington. 13 18 0 10 Low prairie. 25 28 3 8 Salt-flat. 14 14 1 1 Swamp. 14 16 .... 4 6 ATTAINMENT OF OBJECTIVES. The conclusion seems to be warranted that the general objective of de¬ veloping the basic method of experimental vegetation so that it combines the maximum of demonstrability and objectivity has been achieved. While the methods involved require further refinement and testing, there can be little question that they are prerequisite to fundamental and permanent progress in the field of vegetation. The special objectives set have also been attained, though necessarily in varying degree, since such problems as those of the cycle and of competition will constantly unfold with advancing research. Both planting and factor results have confirmed the essential variety within unity that characterizes the grassland formation with its associations, and harmonize completely with the evidence from cyclic changes and relicts. They have further illuminated the relation of serai habitats and communities to those of the climax, and have likewise justified the concepts of subclimax, postclimax, and preclimax. The essentially subclimax nature of the low prairie has been corroborated, as have also the climatic and phylogenetic relationships of the subclimax prairie. The evaluation of competition as a controlling process in OBJECTIVES. 153 grassland has been attended by unexpected success, especially in connection with the persistence of subclimax prairie in regions of forest climate. With respect to natural migration, the experimental evidence lends no support to the assumption that single or scattered individuals regularly invade climaxes with success, but on the contrary indicates that all such invasion is a mass movement in response to wet and dry phases of climatic cycles. The demonstration of the phyletic relationship between subclimax prairie, true prairie, mixed prairie, and short-grass areas and between the corresponding climates in the first three cases appears conclusive, and in conjunction with relict and cyclic results will constitute the basis for future research in the grassland formation. 154 EXPERIMENTAL VEGETATION. BIBLIOGRAPHY. Clements, F. E. 1897. Peculiar zonal formations of the Great Plains. Am. Nat. 31:968. 1904. Development and structure of vegetation. Rep. Bot. Surv. Nebr. 7. 1904 2 . Formation and succession herbaria. Univ. Nebr. Studies 4:329. 1905. Research methods in ecology. 1907. Plant physiology and ecology. 1916. Plant succession. Carnegie Inst. Wash. Pub. 242. 1920. Plant indicators. Ib., Pub. 290. 1921. Aeration and air-content. Ib., Pub. 315. 1922. The original grassland of Mohave and Colorado Deserts. Carnegie Inst. Wash. Year Book 21:350. The method and principle of relicts. Ib., 351. Rainfall and climatic cycles. Ib . x 354. 1923 The original vegetation of Death Valley. Ib., 22: Application of the relict method. Ib., 317. Rainfall and climatic cycles. Ib., 318. -and E. S. Clements. 1921. Changes in vegetation. Ib., 20:405. 1922 Changes in grassland. Ib., 21:348, 22:316. -and G. W. Goldsmith. 1921-23. The phytometer method. Ib., 20-22. -. 1924. The phytometer method in ecology. -and H. M. Hall. 1919-23. Experimental taxonomy. Ib., 18-22. -and D. C. Lutjeharms. 1921-22. Slope-exposure studies. Ib., 20-21. -and J. E. Weaver. 1919-23. Transplant quadrats and areas. Ib., 18-23. Hall, A. D. 1920. The soil. An introduction to the scientific study of the growth of crops. Lutjeharms, D. C. 1924. Environmental differences on north and south slopes of a canyon. Pool, R. J. 1914. A study of the vegetation of the sandhills of Nebraska. Minn. Bot. Studies 4:189. Pound, R., and F. E. Clements. 1898. The phytogeography of Nebraska. 1898 2 . The vegetation regions of the prairie province. Bot. Gaz. 25:381. 1900. The phytogeography of Nebraska. Second edition. Shantz, H. L. 1906. A study of the mesa region east of Pike’s Peak: the Bouteloua formation. Bot. Gaz. 42:16. 1911. Natural vegetation as an indicator of the capabilities of land for crop produc¬ tion. U. S. Dept. Agr., Bur. PI. Ind. Bull. 201. Weaver, J. E. 1917. A study of the vegetation of southeastern Washington and adjacent Idaho. Univ. Nebr. Studies 17:1. 1919. The ecological relations of roots. Carnegie Inst. Wash. Pub. 286. 1920. Root development in the grassland formation. Ib., 292. 1924. Plant-production as a measure of environment. Jour. Ecol. 12:205; Carne¬ gie Inst. Wash. Year Book 22:312. -and J. W. Crist. 1922. Relation of hard-pan to root penetration in the Great Plains. Ecology 3:237. -. 1924. Direct measurement of water-loss from vegetation without disturbing the normal structure of the soil. Ecology 5:153; Carnegie Inst. Wash. Year Book 22:312. -, F. C. Jean, and J. W. Crist. 1922. Development and activities of roots of crop plants. Carnegie Inst. Wash. Pub. 316. - , Joseph Kramer, and Maud Reed. 1924. Development of root and shoot of winter wheat under field conditions. Ecology 5:26-20. -and A. F. Thiel. 1917. Ecological studies in the tension zone between prairie and woodland. Rep. Bot. Surv. Nebr. 1917. BEHAVIOR TABLES 155 m W « H O t-H > a w a Table 58 .—Sowing in denuded quadrats, high prairie, Lincoln, Apr. 16 and May 1920. —Continued. 156 EXPERIMENTAL VEGETATION Table 59. — Sowing in denuded quadrats, Phillipsburg, May 7, 1920. BEHAVIOR TABLES 157 1 During 1922 Andropogon nutans grew well but did not blossom, while in 1923 it made an excellent growth and flowered abundantly. 158 EXPERIMENTAL VEGETATION. 1. Andropogon nutans. 2 green shoots only, 15', little invasion. 2 small clumps, 8-15', will blossom. 9. Liatris punctata. 2 fine stalks, 12', thriving. Now 16', will flower abundantly. 13. Stipa viridula. Good, many well-tillered clumps, 12-15'. 12-18', thriving, very little invasion. Table 61. —Seedlings transplanted to high prairie, Lincoln, May 11, 1921. BEHAVIOR TABLES 159 Table 62. — Seedlings transplanted at Phillipsburg, May 19, 1921. 160 EXPERIMENTAL VEGETATION Table 63 .—Seedlings transplanted at Burlington, May 20 , 1921 . BEHAVIOR TABLES 00 tab P <5 > P •tt o 3 5 t3t: 2 oi 0 ) o) P£ ri 05 *3 bll a 05~ -O * • •£> f. o3 o3 c3 c3 c 0 > 0 ^QQ^P T3 S3 01 p ■Ofl e3 cj 05 05 PQ 05 a p >-> f—« CO I . CO “?H I * ro <* i N ' i :as . O . c3 .pO ! 0 ) o3 TJ CD > CO CM CO I CM -6 o O bfl **CM CD CD CD P> ' o o c> CO CM CD *3 o a CQ a p V p J 5 'm a p >> a o bo < oo co 3 0 ' cJ co oo cs 2 «.s « 2*8 ° a e.fS o? y S 33 ® ■*■{ -S os Oj ” C3 'O j§ Oj CO c a & g § §) m a 5 S 2 2 o o P P p a a_S2 o o o oo o 0 3131313 (h >H -*_5 .+_} -+_5 -*_5 XJd .co 3 2 3 co 03 o O o3 O •;-* -P> £3 <-> eg 5j — S3 r. . 2 « a d eg a w o a C 03 CO w CD B^’E'E ej oj M fli 4 J -p O ^ 2 fl-c O O O-b o 03 cj.&.S* .» O CL O o o -4-3 p o pp -M p £ o S—i bJO co • £ b * c*i rO bJO 'O' -P b-C bs e ss o e CO © * © © CO p P XD P p P 03 ,P CD © © —-o © e s t- o T—I P P -P P p o >? L "o P > P P P P CO 03 g >0 O “ eo © rP g .bp “’d dP CO • 5-. to $ * p +=> m o hC _ Oj ° D D ^>P-T • rp $ o > 03 3 *CO p o co © <3 23 O o i £ £ hJ <$ PQ CLEMENTS AND WtAVER PLATE 15 A. —Sod-cores and details of installation, Burlington. B. —Clip-quadrat in Bulbilis-Bouieloua short-grass, Burlington. 1 V A * >1 ✓ r I \ •• ■ • flu ' • * .