PRAIRIE CLUSTERS EN SOUTHWESTEEN MICHEGAN: A STUDY EN PLANT GEOGRAPHY The“: for the Degree of M. A. MICHIGAN STATE UNIVERSITY Kenneth Richard Robinson 3 969 MSU ggluaNING MATERIA_L_S_: P1ace in book drop to LJBRAfiJES remove this checkout from ”- your record. FINES wil] be charged if book is returned after the date stamped be10w. ____ ,____..___.___ rgt .1”?- .5352" $20 as; .1 SEP 2 0 2004 ‘ ' 0531_0 0 4 '- 5:1: ",W’T’." ":3, SW] 5 ‘37 1‘5; \ ‘ [a H ‘| ' w" , ' . ,tw ‘ W7 ‘ W. e- W " W "“1! ii‘ (/13: ' - ,_ x ~ _ qty: . 5’1: ' H" }5‘\,;f;{jg/j,fi)z.’? 0o! 17 a}; W “' WAR 1 2 196‘- WI? 0 8282733 ABSTRACT PRAIRIE CLUSTERS IN SOUTHWESTERN MICHIGAN: A STUDY IN PLANT GEOGRAPHY By Kenneth Richard Robinson The distribution of vegetation in southwestern Michigan is distinguished by the occurrence of islands of prairie communities within the climax deciduous forests of the region. These grassland remnants have persisted through time in spite of the present climatic regime which provides adequate precipitation for the growth of trees. Previous investigators have studied the species composi- tion, as well as pedologic and pyric variables, in the Michigan prairies. However, no comprehensive explanation for this biogeographic anomaly has been presented. The probable source of prairie flora in Michigan was the tall-grass "prairie peninsula" immediately south- west of the state. On the basis of palynological evidence, it appears that postglacial climates in the central United States became increasingly warmer and drier, culminating in a "xerothermic" period between 3,500 and 5,000 years B.P. Prairie grasses invaded southwestern Michigan and displaced portions of the climax forest by processes of Kenneth Richard Robinson succession. The genesis of prairies in Michigan may be considered as one element in a sequence of vegetational changes in response to climatic fluctuations through time. The grasslands in southwestern Michigan may be de- fined as "mesic" prairie associations on the basis of in- dicator species and edaphic factors, especially drainage conditions on the relatively level outwash plains. Post- xerothermic increases in precipitation have initiated a re—advance of trees upon the prairie sites. It is probable that the grasses would have been entirely replaced by de— ciduous trees in time. Such a succession has apparently been delayed by several factors which have given the grasses a competitive advantage over trees. These factors are: l. seasonal fluctuations in climate; 2. extensive mattings of fibrous grass roots in conjunction with the somewhat droughty nature of the Warsaw soils; 3. subsurface drainage conditions on the outwash plains; 4. fires and human agricultural activities. Certain edaphic variables are significant in ex- plaining the present distribution of vegetation in south— western Michigan. Of particular importance are soil moisture conditions which result in the relatively level prairie sites being neither too moist nor excessively dry. In contrast with the sandy soils and clay loams of the study area, the Kenneth Richard Robinson prairie soils have medial values for infiltration and field capacity. A critical variable in the persistence of prai- rie grasses appears to be the existence of a small but distinct water deficiency which is related to drainage conditions on the outwash plains. In general, the prairie associations seem to be well adapted to tolerate the local environment which is humid throughout the year but with a xeric period from July to September. The transitional character of southwestern Michigan in terms of climate, vegetational communities, and edaphic variables, suggests that the prairies lie within an ecotone. Although increases in precipitation appear to favor trees, the transitional nature of the area seems to be as conducive to the viability of grasses as to the development of climax forests. The favorable environment on the outwash plains, in terms of drainage, has enabled the prairie remnants to persist as "post-climax" enclaves within advancing deciduous forests. PRAIRIE CLUSTERS IN SOUTHWESTERN MICHIGAN: A STUDY IN PLANT GEOGRAPHY By Kenneth Richard Robinson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Geography 1969 ;,/ Approved imit«uwfuui(pil.____g L3 Qfél/{p é-ifvé? ACKNOWLEDGMENTS I wish to acknowledge, with gratitude, several persons at Michigan State University who have provided assistance in the preparation of this thesis. Professor D. Brunnschweiler, Department of Geography, originally proposed the Michigan prairies as a challenging topic for investigation and, subsequently, guided the study to suc- cessful completion with good advice and patience. Profes- sor J. R. Harman, Department of Geography, provided invaluable assistance with the climatological and edaphic aspects of the problem. Professor L. M. Sommers, Chairman, Department of Geography, made arrangements for financial aid through a graduate assistantship which permitted me to continue with graduate studies. Helpful suggestions during the initial phases of research were given by Professors J. E. Cantlon and S. N. Stephenson, Department of Botany and Plant PatholOSY, Pro- fessors A. E. Erickson and I. F. Schneider, Department of Soil Science, Professor A. T. Cross, Department of Geology, and Professor C. E. Cleland, Department of Anthropology. Cartographic advice was provided by Professor C. Akatiff and Mr. J. P. Brackin, Department of Geography. ii Finally, sincere appreciation to my family in Orlando, Florida, especially my mother, who has made personal sacrifices in order to allow me to acquire a university education. iii TABLE OF CONTENTS Page LIST OF TABLES . . . . . . . . . . . . . . . . . . . vi LIST OF ILLUSTRATIONS . . . . . . . . . . . . . . . vii INTRODUCTION . . . . . . . . . . . . . . . . . . . . 1 Chapter I. A SURVEY OF PRAIRIE RESEARCH IN THE UNITED STATES . . . . . . . . . . . . . . . . . 3 Grasslands of the Central United States . . . . . . . . . . . . . 3 The Michigan Prairies . . . . . . . . 8 Conclusions . . . . . . . . . . . . . 9 II. PRAIRIE REMNANTS IN MICHIGAN . . . . . . . . 10 Location . . . . . . . . . . . . 10 The Michigan Prairies: Definition and Classification . . . . . . . . . . . . 11 III. CLIMATIC ELEMENTS AND THE MICHIGAN PRAIRIES . . . . . . . . . . . . . . . . 23 Genetic Influences . . . . . . . . . . . 23 Regional Climate . . . . . . . . . . . . 2A Climatic Variability . . . . . . . . . . 26 Prairie- Forest Transition . . . . . . . . 30 The Water Balance . . . . . . . . . . . 32 Relationships Between Climate and Prairies . . . . . . . . . . . . . . . 37 IV. THE PALEOBOTANICAL RECORD . . . . . . . . . 39 Pollen Profiles in the Great Lakes Region . . . . . . . . AO Significance of the Palynological Evidence . . . . . . . . . . . A5 Xerothermism . . . . . . . . . A6 Dating the Xerothermic Period . . . . . . A7 iv Chapter Page V. GEOMORPHIC AND EDAPHIC ELEMENTS IN THE PRAIRIE REGION . . . . . . . . . . . . . 50 Surface Configuration . . . . . . . . . 50 Pedology . . . . . . . . . . . . . . . . 52 Soil Moisture . . . . . . . . . . . . . 55 VI. THE PRAIRIES OF MICHIGAN: A SYNTHESIS . . 67 Prairie Genesis . . . . . . . . . . . . 67 Reaction and Displacement . . . . . . . 68 Prairie Ecesis in Michigan . . . . . . . 68 Prairie-Forest Competition . . . . . . . 71 Prairie Persistence . . . . . . . . . . 72 Summary and Conclusions . . . . . . . . 77 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . 8O LIST OF TABLES Table Page 1. Prairie species in Southwestern Michigan . . l8 2. Categories of prairies and species composition in Wisconsin . . . . . . . . 19 3. Phenological values for Lower Michigan . . . 27 A. Mean water deficits for selected stations in the study area . . . . . . . . . . . . 3A 5. Generalized sequence of climate and vegetation in the Great Lakes Region . . Al 6. Moisture and infiltration rates for selected soil profiles in Southwestern Michigan . . . . . . . . . . . . . . . . 57 7. Mean infiltration rates and field capaci- ties for soils in the study area . . . . 65 vi Figure LIST OF ILLUSTRATIONS Page Distribution of Prairie Remnants . . . . . . l2 Grasslands of the United States . . . . . . l5 Moisture Variability in Southern Michigan . . . . . . . . . . . . . . . . 36 Distribution of Selected Pollen Profiles . . . . . . . . . . . . . . . . 42 Surface Formations . . . . . . . . . . . . . 51 Major Soil Associations . . . . . . . . . . 54 Soil Moisture and Infiltration . . . . . . . 6A vii INTRODUCTION The pattern of vegetational distributions in southern Michigan is distinguished by "islands" of stable prairie communities which are scattered within an area of deciduous climax forest. In 1811, these prairie sites were described by a settler as "tracts of land which have been without the appearance of timber since any white inhabitant ever saw them" (Lossing, 1811, p. l). The absence of trees from sites seemingly capable of support- ing hardwood forests represents a biogeographic anomaly which has never been fully explained. Investigations of prairie communities in other areas of the United States have stressed single-factor theories for prairie genesis, including: 1. the postglacial evolution of grasses as cli- max communities on level terrain, 2. the dominance of grasses upon soils whose parent materials possess a high content of lime, 3. the influence of cycles of drought, H. the establishment of prairies following the burning of hardwood forests by Indians. In plant geography, single-factor explanations are seldom conclusive. The complex nature of many vegeta- tional communities implies a multiple-factor relationship with the habitat, in which plants, climate, and soils conjoin within the environmental matrix. Therefore, it is logical to assume that several factors have operated, at varying degrees of intensity, to produce prairie vegetation. This study was undertaken to investigate the genesis of grassland vegetation in southwestern Michigan and its persistence in the form of relict prairies. Research was conducted to fulfil three objectives: 1. to survey the literature concerning North American grasslands in general and Michigan prairies in particular and to provide a bibliography for general reference, 2. to survey certain elements of the natural environment which may contribute to prairie persistence in Michigan, 3. to present a preliminary synthesis for the interrelationships of climate, soil moisture, landforms, and prairie vegetation in southwestern Michigan. CHAPTER I A SURVEY OF PRAIRIE RESEARCH IN THE UNITED STATES The prairies of North America occupy a large area of the continent and have an historical value as sites of settlement. Brief descriptive accounts of the prairies by early settlers appear in various county histories and local newspaper files. However, there is a scarcity of geographical literature about them. In general, the tall- grass prairies of the North American grassland biome have been the subject of investigation by scientists other than geographers. Grasslands of the Central United States One of the first noteworthy accounts of the eastern prairies of the United States explained these grasses as products of repeated burning by fire (Caton, 1876). A general classification of American prairies according to tall, mixed, and short-grass species was presented by Pounds and Clements (1898). Brief references to tall-grass prairies were included in a subsequent study of the post- glacial origins and migrations of flora and fauna in the eastern United States (Adams, 1902). u During the first quarter of the twentieth century, research activities in the prairie region were, to a large extent, general in scope. These studies included two sur- veys of prairie vegetation in Iowa and its persistence (Shimek, 1911; 1925), a study of environmental variables in the Illinois prairies (Sampson, 1921), the species com- position of a prairie climax formation in Minnesota (Berg- man, 1923), the species composition of a tall-grass prairie in Kansas (Schaffner, 1913), and a qualitative account of the grassland region in the central United States in terms of species, soils, and human occupance (Schaffner, 1926). To these broad surveys were added more detailed descriptions of the prairie flora and environment. A significant contribution during this period was the con- cept of a vegetational continuum which reflected possible variations in the environment. This principle was initially presented in a study of the mixed-grass prairies which showed a gradation into short-grass steppes to the west and tall-grass prairies to the east (Bruner, 1921). A description of the vegetational history of the Middle West (Gleason, 1923) suggested that the tall-grass prairies were one component in a continuum of vegetation between the drier, short-grass areas to the west and the humid, eastern deciduous forests. Other studies considered the tall-grass prairies as a zone of transition between the semi-arid, short-grass steppe and the eastern forests (Aikman, 1928) as well as the influence of pedologic factors which might favor the persistence of prairies (Cowles, 1928). These investigations culminated in Tran- seau's classic delineation and description of the eastern humid grasslands as a "prairie peninsula" (1935). Other studies were more concerned with the environ- mental variables which might favor the persistence of prairie vegetation. In essence, direct cause-and—effect relationships between the environment and the occurrence of grasslands were sought. Research activities considered the effects of environmental factors, such as soil and drainage, upon the ecesis and growth of various species of grasses (Clements & Weaver, 192h) as well as the evalu- ation of metabolism and growth rates of plants as measures of environmental influences (Weaver, 192A). Some compre— hensive surveys of species distributions in the tall-grass prairies of Nebraska (Steiger, 1930) and Kansas (Weaver & Fitzpatrick, 193E) were conducted in an attempt to deter- mine possible variations in the environment that might favor grasses rather than other forms of vegetation. The Kansan prairies were also studied in terms of their areal extent, specific interrelationships of various grasses, and possible degrees of interaction between each species of grass and edaphic factors in the region (Albertson, 1937; 1938; 19h1). Subsequently, quantitative techniques were employed in an investigation of the relationship between different species of grasses and the amount and quality of soil nutrients (Shively & Weaver, 1939), a study of competition between wheat grass and relict prairie vege- tation in Nebraska (Weaver, 1942), and a phytosociological analysis of a tall-grass prairie region in Oklahoma (Rice, 1952). The relationship between grasslands and water deficiencies has been a major consideration in the evalu- ation of the viability of prairie vegetation. Weaver and Albertson (1936; 1939; l9hOa; 19hOb; 19h3; 19hh) conducted a thorough investigation of the relationships between grasslands in the central United States and the period of drought between 1933 and l9hO. Their studies consid- ered the general effects of the drought upon prairies in Iowa, Nebraska, and Kansas; changes in species composition and the eventual deterioration of some prairie communities under conditions of extreme aridity; possible relation- ships between decreases in soil moisture and the denuda- tion of grasslands; the distribution of grasses and forbs at the end of the drought; and the degree of recovery of the prairies from the drought. They concluded that the amount of water which was actually available to prairie grasses, especially during the period of germination, was a critical variable in the ability of the grasses to sur- vive during conditions of drought. Similar studies also considered the physical milieu of the American prairies in relation to the influ- ence of non-periodic cycles of drought. These studies included a description of the Great Plains grasslands and the effects of drought (Clements & Chaney, 1936), an evaluation of grassland viability during periods of drought in the tall-grass prairie region (Savage, 1937), and a consideration of the impact of human populations and cli- matic cycles of dryness upon tall-grass prairies (Clements, 1938). Recent investigations have considered the effects of grazing upon prairies. In Missouri, it was shown that differences in floristic composition exist between grazed and ungrazed prairie vegetation (Drew, 19h7). Mentzer (1951) described a succession of herbaceous species which occurred on a Nebraskan prairie after the cessation of grazing activities. The environmental variables in the North American grasslands were summarized in comprehensive volumes by Hanson (1938), Malin (19h7), and Weaver (195M). Subse- quently, the eastern tall-grass prairies have been studied in terms of their vegetational evolution in space and through time. Two notable contributions have been the concept of vegetational changes as a response to climatic fluctuations in the post—Pleistocene period (Deevey, 19h9) and the hypothesis that non-periodic fluctuations in the flow of westerly winds across North America have encour- aged a "grassland climate" in the central United States (Borchert, 1950). The problem of "relict colonies" of prairie grasses in Ohio, Michigan, and Illinois has been related to post-Pleistocene modifications in the environment (Braun, 1950). A consideration of precipitation effective- ness and drainage conditions has permitted the classification of prairies in Wisconsin into five categories according to the degree of available moisture for plant growth (Curtis, 1955; 1959). Finally, a recent hypothesis has postulated that prairie persistence may be related to the fact that the prairie peninsula has acted as a "filter barrier" to post- glacial migrations of deciduous forests (Benninghoff, 196A). The Michigan Prairies The literature concerning the Michigan prairies is very limited. Brief accounts of prairie species are given in volumes by Cole (1901), Beal (190A), Pepoon (1907), Hebert (1934), Darlington (19h5), and Hanes (19u7). Ecological studies of wet prairies have been conducted near Ann Arbor (Gleason, 1917), on Harsen's Island (Hayes, 196A), and near the city of Kalamazoo (Brewer, 1965). A comparative survey of the historical and ecological vari- ables in the vegetation of southern Michigan described the distribution of grassy patches within the deciduous cli- max forests of the region (Quick, 1923). The areal extent of these prairies was determined through land survey records which were used to reconstruct the patterns of vegetation in southern Michigan at the time of white set— tlement (Kenoyer, 1930; l93h; l9hO; 19h3). The occur- rence of "dry" prairies in southwestern Michigan was examined with emphasis upon the pedologic characteristics 9 of the prairie sites (Veatch, 1928). More recently, an ecological analysis of the prairie remnants in Newaygo County described pyric and anthropeic elements which might have contributed to the evolution of these prairies (Hauser, 1953) - Conclusions The survey of literature indicates that the pri— mary emphases in these studies have been upon species composition of prairie communities, soil characteristics of the prairie regions, the possible influences of water deficiencies, and the role of fire in the persistence of prairie grasses. In spite of the relative abundance of scientific studies of the American prairies, few studies have been devoted to the Michigan prairies and no widely applicable explanation of their persistence has been pre- sented. An holistic approach which includes an evaluation of the climatic and edaphic elements in the local environ- ment should contribute toward an ultimate solution of the state's "prairie problem." CHAPTER II PRAIRIE REMNANTS IN MICHIGAN An investigation of the evolution and persistence of prairies in Michigan is hampered by the dearth of present-day examples. A humid climate, level terrain, and relatively fertile soils encouraged the appropriation of the prairie sites by early settlers for agricultural purposes. With the exception of minute fragments along railroad rights-of-way, in uncultivable corners of farm- steads, and in pioneer cemeteries, virgin prairie plants are non-existent in the state today. Location The majority of prairie clusters in southwestern Michigan lie within a nearly rectangular tier of counties between hlo HO' and M20 25' north latitude and 840 MS' and 860 h6' west longitude. The counties include Berrien, Branch, Calhoun, Cass, Kalamazoo, St. Joseph, and Van Buren. Lake Michigan lies along the western margin of the study area. The state of Indiana lies immediately to the south. The seven counties contain the major concentrations of prairie remnants in Michigan. The precise location of each prairie was determined from the original survey records 10 11 and is plotted on Figure 1. The prairies in Newaygo County, which have been described by Hauser (1953),are excluded from this study. The Michigan prairies lie just within the eastern apex of the nearly triangular region of contiguous prairie vegetation which extends into parts of northwestern Indi- ana and southwestern Michigan (Figure 2). This region of tall-grass prairies occupies an intermediate position between the semi-arid, short grasses to the west, and the eastern, deciduous forests. Throughout the eastern por- tion of the prairie triangle, grasses and trees inter- mingle. In general, the grasses tend to occupy the rela- tively level to undulating interfluve areas, whereas drier forest species, especially oak and hickory, tend to occupy the slopes of valleys or lie on the rolling terrain adjacent to streams (Aandahl et_al:, 1960). The Michigan Prairies: Definition and Classification Field reconnaissance in the study area revealed that the Michigan prairie sites are relatively level and largely devoid of trees. Cultivation has gradually des- troyed or severely altered nearly all of the original vegetative cover. Therefore, it is necessary to evaluate the use of the term "prairie" as it is applied to these sites. "Prairie" is a French word which designates a meadow form of vegetation (Voigt & Mohlenbrock, l96h). H ohdwwm OJ 12 a m. coon com o. mu.¢.<¢u 0k >wx .. on on LI 2‘ a. “as: Q a? 3:33.13 . . a-.. J. a )\ ' _ —' SUO.I U» \Wfiv . ”N‘ . m...... Wm. m w z 1024.me rimmed. . mmqo szmwm route-o. _ w on . . So “ 2.6.3.3.. Nu. . J n .. c . 3.30:». 0.. .N‘ Go. m 2...: . .& ath3QJOu . on ~33: hm . 'N ‘ON .4 _ 13...:- ,. O 00 O . . My .323 p n \ ‘ u _ .938 was“ op _ 2;. .E ' e . O . . O h 7/ I to . o .22.... a.» . _ J .1: u - IIIII II In! II I: III. llLlll.|O\-‘ III. e __ . 0 mm . _ 51:. 1 [23.5013 _ :33. Z. T m.‘ . on‘ _ . comma: zopzuo .2 90 23933 . 425$. zumam z<> H . . ’40 ‘ucISu on. if. )5... . 4; a . . 3,} UJPP<¢ . u Nn¢2(JXU.I . wx14 4430 0 zw>_m_m m_m_r extended periods of time. These long waves are spaced erl such a way that mean upper level troughs occur over the €36isstern half of North America. This condition enhances 23 2A cyclogenesis and the advection of air masses from distant source areas into the Great Lakes Region. For example, water vapor transfer into southern Michigan is largely a result of the flow of air from the Gulf of Mexico (Benton & Estoque, 1954). At the same time, the lack of topographic obstacles between the Gulf of Mexico and the Arctic Ocean allows the entrance of warm and cold air masses into the mid-continental areas with little or no modification. The interaction between these air masses is a major factor in regional cyclogenesis. Further discussions of the genetic climatology of Michigan may be found in Hodgins (1960), Niedringhaus (1966), and Trewartha (1966). These genetic elements produce a condition of cli— matic instability in the interior of the North American continent during all seasons of the year. Although short- term climatic fluctuations are a product of passing cyclones, long-term variations may reflect hemispheric modifications in the long-wave train in the upper atmosphere. Signifi- cant departures from mean values of temperature and pre~ cipitation over large portions of the continent may result. If pronounced climatic departures occur in the summer, normally the period of greatest environmental stress for vegetation, such climatic variation may become a limiting factor for particular plants. 3B§gional Climate Southwestern Michigan, including the study area, 11613 a climate which is indicated by the symbol Dfa in the 25 Koeppen classification as modified by Trewartha (1954). The Dfa climate, usually designated as a microthermal, humid continental-warm summer type, is defined as one in which the mean temperature of the warmest month exceeds 71.601 and that of the coldest month less than 320. With respect to precipitation, this climatic type has no marked dry season (i.e., f). In general, mean precipitation is relatively well distributed throughout the year with no month receiving less than 1.58 inches (Messenger, 1962). More specifically, the study area lies near the northern limit of the Dfa region (Villmow, 1952). As a result, mean climatic conditions characteristic of the Q£b_subregion (with mean temperature of the warmest month below 71.60) to the north are occasionally observed in the study area. These climatic fluctuations produce non-periodic, latitu- dinal shifts in the Dfafgfb'boundary across southern Lower Michigan. According to Niedringhaus (1966), the study area lies within both the Southern Lake Michigan—Lowland Pro- vince and the Southern Lower Peninsula Interior Subprovince. These climatic subregions have the highest mean annual temperatures in the state, with moSt stations recording values of 470 and above. Stations in the extreme south— western and southeastern areas of the state experience mean annual temperatures of 500. A normal frost—free lAll temperature values are given in degrees Fahrenheit. 26 period in the study area of 150 - 170 days is indicative of the relatively warm temperature regime during the year. Furthermore, the southwestern subregions have more pre- cipitation in May and June and less precipitation in the late summer than other regions in Michigan. The mean annual snowfall ranges from 40 - 60 inches throughout much of the study area, although higher values may be observed near the shore of Lake Michigan. Climatic differences between the study area and the northern Lower Peninsula are illustrated in phenologi- cal data provided by Visher (Table 3). In contrast with the Northern Hardwood Region, the Oak-Prairie Region of southwestern Michigan has (i) a longer period of summer temperatures, (ii) a longer growing season, (iii) shallower depths of soil freeze and frost penetration, (iv) a shorter duration of snow cover. Snow-melt normally occurs by March 1, or one month earlier than in the Northern Hard- wood Region. It is possible that early run-off and evapora- tion of snow-melt may render some of the winter precipita- tion unavailable to the plants during periods of late spring water need. Furthermore, these data suggest that the cli- matic conditions are more moderate in the study area than over the remainder of the Lower Peninsula. Climatic Variability High rainfall intensities and non-periodic varia— tions in precipitation are characteristic of the Great Table 3. Phenological Values for Lower Michigan (After Visher, 1943; 19A?) Northern Oak-Prairie Variable Hardwood Region Region 1. First day of warm weather (daily mean temperature above 500 F.) May 1 April 1 2. First day of summer (daily mean temperature above 680 F.) July 1 June 15 3. Average number of days with summer temperatures 50 75 4. Average length of growing season (days) <139 > 168 5. End of summer temperatures August 15 September 1 6. First snowfall December 1 December 15 7. Depth of soil freeze 3—6 feet l%-3 feet 8. Depth of frost penetration (January) > 40 inches (40 inches 9. Number of days with temperatures below 320 F. > 30 < 30 10. End of snow cover April 1 March 1 27 28 Lakes Region (Trewartha, 1966). Although the single pre- cipitation minimum in winter and the early summer maximum are indicative of the continental climate of the interior United States, considerable fluctuations about the mean values do occur. Hodgins (1960) has summarized the nature of pre- cipitation variability in his study of the genetic clima- tology of the Great Lakes Region. He found, for example, that South Bend, Indiana, adjacent to the study area, received 5.82 inches of rainfall in July, l95h. This amount of precipitation was 3.32 inches above normal for the month. Furthermore, 2.62 inches of precipitation were received in a two-day period. Hodgins emphasized that certain stations in Indiana received as much as 6 inches of precipitation in July while nearby stations may receive less than 1 inch during the same period. Significant temporal variations of precipitation do occur in northern Indiana and southwestern Michigan although no long-term patterns are evident. A marked decrease in precipitation during June and July occurs at the majority of stations in Michigan (Brunnschweiler, 1961). Such a decline may be related to the influence of the Great Lakes (Lautenhiser, 1953). During the warm season, the Great Lakes act as a "heat sink." The chilling effect of the water surfaces may cause increases in mean atmospheric pressure, particularly in late spring and early summer when the air-water 29 temperature contrast is most pronounced. These conditions might enhance the stability of traveling anticyclones, which would favor the relative decline of mean precipita- tion in July and August. Although Hodgins (1960) has disputed the "lake effect", the hypothesis does provide at least a partial explanation for periods of subsidence and stability in the study area. Precipitation values in the southwestern Lower Peninsula exhibit the highest degree of variability in Michigan with an annual range of 14 percent to 22 percent (Niedringhaus, 1966). The greatest amount of variability occurs in the summer months, when the highest co-efficient of variation (75 percent) for July precipitation is recorded for Benton Harbor on Lake Michigan. It may be significant that the greatest precipitation variation occurs during the time of greatest soil moisture need and active plant growth. Niedringhaus (ibid.) has also summarized the thermal variability in Michigan. In the Lower Peninsula, mean annual temperatures tend to be higher near the Great Lakes. For example, the 250 isotherm for January extends parallel to the Lake Michigan shoreline for nearly half the extent of the Lower Peninsula. Mean temperature values for July range from 740 in the southwestern and southeastern Lower Peninsula to less than 630 in the northeastern Upper Penin- sula. He also found that areas nearer the Great Lakes have fewer days with extreme temperatures than interior stations. For example, Muskegon has a mean maximum temperature of 3O 560 and a mean minimum temperature of 40°. Fifty miles inland, Greenville has a mean maximum temperature of 58° and a mean minimum temperature of 38°. In the state, absolute maximum temperatures vary from 110° at Saginaw to 980 along the shoreline of the Upper Peninsula. Abso- lute minimum temperatures range from -l4° in the south- western and southeastern Lower Peninsula to -51° in the Upper Peninsula. These data indicate that the study area experiences a more moderate range between climatic extremes throughout the year than other regions in Michigan. Prairie-Forest Transition The prairie clusters in Michigan occur in a zone of transition between the areas of nearly continuous pre- settlement forests to the north and east, and areas of grasslands to the southwest in portions of Indiana and Illinois (Braun, 1950). Transeau (1935) investigated this "prairie peninsula" and found certain climatic elements which are unique to it. His conclusions are: 1. Annual and seasonal precipitation values are lower whereas temperature and evaporation values are higher in the grasslands than in adjacent forested regions; 2. Midsummer relative himidity in the grasslands is lower than in adjacent forested regions; 3. Precipitation in the grasslands occurs mainly during the growing season (June) and becomes more irregular in successive months; 31 4. Temporal distribution of precipitation in suc- cessive years varies more in the grasslands than in the forested regions. Borchert (1950) also investigated the climate of the prairie peninsula. He found that certain climatic con- ditions are characteristic in the region. These include: 1. less total winter precipitation in the grass— lands than in the forested areas to the east. The precipi— tation gradient is steeper along the transitional boundary between grasses and trees than in the prairies or areas of deciduous forests; 2. low rainfall and limited snow cover over the prairie region. A steep mean water-snowfall gradient occurs along the boundary between grasses and forests; 3. increases in summer rainfall eastward from the short-grass region to the tall-grass prairies. The increases reflect increasing intensities of rainfall rather than increases in the number of rainy days (of. Hodgins, 1960, P. 113); 4. a higher degree of rainfall variability and summer drought in the prairies than in the forested regions northeast and southeast of the prairie-forest margin. The irregular periods of drought in the prairie region appear to be associated with abnormally strong flows of westerly winds (Borchert, 1950). Under these circumstances, humid Gulf air, moving northward, is deflec- ted eastwards before entering the prairie region. Also, 32 summer cyclones moving on the Alberta track tend to deliver precipitation to the northern hardwood forests of the Great Lakes Region rather than to the grasslands. Borchert designated the area which is strongly influenced by the westerly flow of air as the "prairie wedge." He considered it to be essentially co-terminous with Transeau's "prairie peninsula." Borchert also mapped the average departures from normal temperatures in July for major drought years. The isoline representing a positive anomaly of 2° during years of drought passes through the prairie zone in southwestern Michigan. This is an indication that climatic conditions characteristic of the prairie peninsula can occur occa- sionally in the study area. As a genetic explanation for the Michigan prairies, Borchert's theory is incomplete. He was concerned with the grassland climate as a whole and did not intend to explain specific problems such as relict plant communities in the prairie-forest transition zone. His hypothesis does not account for the specific distribution of vegetation types which occur under the same regional climate in south- western Michigan. On the other hand, it does suggest that climatic variability may be higher here than elsewhere in the state. The Water Balance A complex aspect of the prairie problem is the relationship between persistence of grassland vegetation 33 and the availability of moisture. For example, the study area experienced the highest mean annual precipitation for Michigan, approximately 36 inches, between 1932 and 1952 (Brunnschweiler, 1962). Jenny (1941) has suggested that mean annual precipitation values between 31 and 36 inches would theoretically permit a luxuriant growth of trees in the prairie-forest transition area of the eastern United States. However, prairie clusters persist in the study area despite the annual precipitation regime which is amenable to the growth of deciduous forests. The higher temperatures and the precipitation variability in southwestern Michigan may increase the potential for summer droughts. In order to evaluate this possibility, it is necessary to consider some aspects of the water balance for the study area. Initially, Thornthwaite (1948) applied the concept of evapotranspiration in his calculations of the water balance for selected areas of the world. By definition, evapotranspiration is the total moisture loss due to sur- face evaporation and transpiration from plants. Potential evapotranspiration is an empirical measurement of the amount of water which would be lost from a surface com- pletely covered with vegetation if the existing soil moisture regime is near field capacity. Messenger (1962) found that potential evapotran- spiration values in Michigan decrease from south to north in the Lower Peninsula. Mean potential evapotranspiration BA values for the northern Lower Peninsula are less than 550 mm. Mean values for southwestern Lower Michigan range from 650 mm to 675 mm, although computed values for Berrien, Cass, Branch, and St. Joseph counties exceed 676 mm when considered individually. The southwestern counties consti- tute an area not only of relatively high mean annual pre- cipitation but high values of potential evapotranspiration as well. According to Thornthwaite, the amount by which the actual and potential evapotranspiration differ in any month is the moisture deficit for that month. Table 4 gives the mean water deficits for nine stations in the study area. Table 4. Mean Water Deficits for Selected Stations in Study Area. (After Messenger, 1961, unpublished data) Mean Deficit Mean Deficit for Dry Station 1931 - 1952 50% of 20 Years; 1931—58 Battle Greek 1.06" 2.08" Coldwater .88" 1.94" Gull Lake .85“ 1.84" Kalamazoo .95" 2.51" Three Rivers .98" 1.37" Paw Paw .88" 1.87" South Haven 1.11" 2.08" Eau Claire .90" 1.65" Benton Harbor 1.12" 2.34" The mean deficits for the dry 50 percent of 20 years between 1931 and 1958 ranged from 1.37 inches at Three Rivers to 2.51 inches at Kalamazoo, and occurred 35 chiefly during the months of August and September. These months are normally the period of mean maximum temperatures in the study area as well. When the mean water deficits are compared with the co-efficients of variation for mean annual precipitation as computed by Niedringhaus (1966) and shown in Figure 3, a relationship between the two vari- ables is apparent. Although the association is not precise, higher mean water deficits tend to Occur in those areas with higher variability in precipitation. Climatic conditions in southeastern Lower Michigan are somewhat analogous to those in the study area, yet mesic prairies have not been described there. The south- eastern counties have higher mean temperatures in summer as well as greater water deficits than the prairie counties in the southwestern corner of the state (Figure 3). For example, Messenger (1962) calculated mean water deficits for the dry 50 percent of 20 years between 1932 and 1958 in which the highest value, 6.9 inches, occurred at Mid- land, Michigan, where no mesic prairies have been observed. The patches of "wet" prairies at Ann Arbor and Harsen's Island may be sedge meadows, similar to the "wet" grasses which Finley and Potzger (1952) described along the Kanka- kee River in Illinois. It is probable that mesic grasses, which originally entered southwestern Michigan from the prairie peninsula, did not reach southeastern Michigan. 36 MOISTURE VARIABILITY IN SOUTHERN MICHIGAN i l i 2 f i E .0 l8 0 20 4o Mlle: MEAN ANNUAL WATER DEFICIT co-EFFICIENT 0F VARIATION FOR 30 - 50 mm MEAN ANNUAL PRECIPITATION 50 ' 70 mm —20 70 ' 90 mm (After NIodrInqhous, I96 6) (Mm Messenger, IS 62) Figure 3 37 Relationships Between Climate and Prairies An evaluation of the climatic data for Michigan indicates that the variable nature of the climate may be partially related to the persistence of prairie vegetation in the state. However, the climate can not be considered as the dominant environmental element. If it were, those climatic variables which enhance dessication should dis- courage tree cover, at least on drier sites in the study area. The persistence of prairie clusters might be inter- preted in terms of the transitional character of the envi- ronment in southern Michigan. The contributing factors include: 1. the relatively great year-to-year variability in precipitation; 2. the year-to-year variations in temperature patterns resulting in periods of water deficiency and summer drought; 3. the proximity of the study area to the "prairie peninsula" and the possibility that intervals of genuine grassland climate could occur in southwestern Michigan as a result of an interaction of the above factors. The high annual precipitation in southwestern Michigan would appear to favor a more or less continuous forest cover in the study area. On the other hand, nega- tive precipitation departures in the summer season may 38 allow other environmental factors, such as edaphic varia- tions, to assume a selective influence upon the local distribution of plant communities. To account for the disjunct mesic prairies in the study area, both past plant distributions and edaphic elements in the study area must be reviewed. CHAPTER IV THE PALEOBOTANICAL RECORD The broad distribution of plant taxa is generally controlled by climate. Therefore, it is reasonable to assume that important and extended climatic modifications may be reflected by variations in the geographic range of certain plants through time. At this point, a brief exami- nation of paleobotanical data from the Great Lakes Region may help to clarify the extent to which prairie species were once distributed throughout the study area, as well as provide evidence of possible climatic oscillations through time. Palynology is a relatively recent technique in Ameri— can plant geography. This technique utilizes fossil pollen which may be obtained from selected profiles at various lo- cations on the earth's surface. Analyses of the fossil pollen permit tentative reconstructions of past plant dis— tributions. Consequently, correlations between climatic and vegetational fluctuations may be made. Five principles of plant geography, originally pre- sented by Gleason (1923), are germane to the Michigan prairie problem: 39 A0 1. Isolated areas of vegetation are to be inter- preted as the products of former plant migrations; 2. The migration of a species depends upon en- vironmental changes. These migrations are indicated by the occurrence and distribution of relict colonies, by ecologi- cal and taxonomic evidence, and inferences about past climates; 3. The present range of any species is not necessarily final, even without further fluctuations in the environment; A. Deductions and inferences concerning stages in plant migrations and distributions are reached in reverse order to their occurrence, from present to past. The accuracy of every deduction depends upon the accuracy with which the preceding stage has been interpreted. Thus, each preceding stage usually reveals less detail and has a higher probability of error; 5. The evidence from all available sources must be combined to build a plausible and probable sequence, culminating in the present distribution of the vegetation. Pollen Profiles in Great Lakes Region The locations of fifty documented pollen profiles,l obtained in the Great Lakes Region between 1932 and 1956, 1References for the palynological studies are given in the bibliography. A1 are given in Figure A. The majority of these profiles were taken from sites in southern Michigan and Indiana. The data may therefore be useful in the reconstruction of pre- vious modifications of the local climate and vegetation. Although individual variations exist in the quan— tities of pollen and thicknesses of horizons, all of the palynological spectra suggest that changes in climate have occurred in the Great Lakes Region through time (Table 5). Table 5.—-Generalized sequence of climate and vegetation in the Great Lakes Region. (After Sears, 1932; 19A2; Deevey, 19A9;and Flint, 1953). Period Time Climatic Regime Vegetation 3000 B.C. Recent to Warm and Humid Maple—Beech Present 5000 B.C. Atlantic to Warm and Dry Oak-Hickory 3000 B.C. and Grasses 8000 B.C. Boreal to Cool and Dry Pine 5000 B.C. Cool and Moist Spruce-Fir Late Glacial Cold (periglacial) Tundra(?)l lTundra pollens are poorly defined or absent in the fifty pollen profiles. A2 DISTRIBUTION SELECTED POLLEN OF PROFILES II. “I 6) ® ,1. WISCONSIN G) (9 I I I ! I I s 63 o I €93 j qfin’ I OHIO ILLINOIS ; (DID @- 2° a i. ‘3 O i. I INDIANA 3. i GD I N G) SITE 0F POLLEN PROFILE KEY TO POLLEN 0 50 IOO PROFILES - Soc l MIIes _1 STUDY AREA Page 43 Figure 4 Figure A.(Cont'd.)--Pollen Profiles in Great Lakes Region Location Wisconsin 1. Baraboo 2. Gibraltar 3. Trout Lake A. Sheep Ranch 5. Gravel Pit 6. Sunken Highway 7. Four Mile Lake 8. Draper Indiana 9. Fox Prairie 10. Lake County 11. Lake Cicott l2. Cranberry Pond l3. Bacon's Swamp 1A. Kokomo 15. Otterbein l6. Loon Lake 17. Altona 18. Lakeville 19. Round Lake 20. Yountsville 21. Mill Creek 22. Leroy 23. Shoe Lake 2A. Shipshewana 25. Culver 26. Jeff 27. Reed 28. Pinhook 29. Merrillville Reference Hanson, H. C. and C. T. Borheis, (1937) Truman, H. V., (1937) Potzger, J. E. and R. R. Richards, (19A2) Potzger, J. E., (19AA) Potzger, J. E., (19AA) Potzger, J. E., (19“4) Potzger, J. E., (19AA) Potzger, J. E., (19AA) Prettyman, R. L., (1937) Artist, R. C., (1936) Smith, w. H., (1937) Barnett, J., (1937) Otto, J. H., (1938) Howell, J. W., (1938) Richards, R. R., (1956) Moss, B. W., (19AO) Moss, B. W., (19A0) Hamp, F. A., (19A0) Hamp, F. A., (19AO) Swickard, D. A., (1942) Swickard, D. A., (19A2) Oliver, J. L., (1951) Oliver, J. L., (1951) Keller, C. 0., (19A3) Keller, C. 0., (19A3) Keller, C. 0., (1943) Griffen, C. D., (1950) Guennel, G. K., (1950) Guennel, G. K., (1950) Location Michigan 30. Hartford 30. Hartford 31. South Haven 32. Sodon Lake 33. Meeuwsen Farm 3A. Berens 35. Duck Lake 36. Redman Mint Farm 37. Dow Marsh 38. Carroll Farm 39. Austin A0. Blue Lake Al. Mud Lake A2. Schroeder A3. Shelby AA. Powell A5. Farwell A6. Bear Lake A7. Chandler Marsh A8. Mud Lake Ohio A9. Mud Lake Illinois 50. Turtle Lake AA and J. E. Potzger, and J. E. Potzger, Reference Oxvald, H., (1935) Zumberge, J. H. (1956) Zumberge, J. H. (1956) Fuller, G. D., (1935) Potzger, J. E., (19A8) Potzger, J. E., (19A8) Potzger, J. E., (1948) Potzger, J. E., (19A8) Potzger, J. E., (19A8) Potzger, J. E., (19A8) Potzger, J. E., (1948) Potzger, J. E., (1948) Potzger, J. E., (19A8) Potzger, J. E., (19A8) Potzger, J. E., (19A8) Potder, J. E., (19A8) Potzger, J. E., (19A8) Parmalee, G. W., (1947) Parmalee, G. W., (1947) Parmalee, G. W., (1947) Sears, P. B., (1931) Griffen, C. D., (1951) A5 The pollen profiles are similar in two aspects. First, they indicate that previous climatic regimes appar- ently have been accompanied by transformations in vegeta- tional climaxes through time. Secondly, the general order of vegetational change, following the recession of Wisconsin glaciation, has been from coniferous forests to deciduous forests, with at least one intervening period of grasslands. Tundra is associated with a periglacial period in the generalized sequence. Dansereau (1957) has suggested that a low, treeless tundra vegetation probably developed along the edges of the glaciers. The absence of tundra pollen from most of the profiles may be a function of the fact that tundra does not produce pollen in quantity. There- fore, it is often impossible to detect tundra pollen in the deeper horizons of the profiles.1 Significance of the Palynological Evidence The pollen profiles are significant in an analysis of the Michigan prairies for three reasons: 1. Although it is not always possible to distin- guish accurately between fossil pollen of prairie grasses and other grasses such as Zizania and Calamogrostis, most of the pollen spectra show increases in grasses at the upper 1The problem of detecting tundra pollen in North American profiles is discussed by Sears (1935, pp. 37-51) and Voss (193A, pp. 3-A3; 1937, pp. 119—135). A6 levels, in association with particular tree pollens. 2. Tree pollens, due to their relative abundance, are probably the most useful indices to past climates (Charlesworth, 1957). The present ranges of Abigs, nggg, and Piggg are limited to northern Michigan. However, pol- len spectra from southern Michigan, Indiana, and Illinois contain pollen of these boreal species in the lower hori- zons. It seems evident that these taxa were widespread across the study area at some time following ice retreat. 3. The oaks (Quercus spp.) and hickories (Cagyg spp.) are associated most frequently with the horizons con- taining grass pollen. Presuming that the relative ecolog- ical roles of these two genera have not changed, the presence of oak, hickory, and grass pollen may be one indication of I I O O 0 l a former drler env1ronment ln Mlchlgan. Xerothermism The eastern United States was influenced by at least one period of warm and dry climate following the 1". . . The oak group is most nearly related and most closely associated with the prairie group. The prai— ries are rarely contiguous to any other forms of arboreous vegetation. The group includes both 'oak openings' and denser oak forests. There are sufficient reasons, however, for separating them into two classes, as they indicate dif— ferent, though allied agricultural capabilities. The oak Openings are most nearly related to the prairies, while the oak forests graduate to the Oak—Maple Group, Maple Group, and Maple-Beech Group." (Chamberlin, 1882, p. 177) A7 recession of the Wisconsin glaciers. Specific lines of evidence include: 1. distinct increases of Quercus species in the upper horizons (Zones B to C2) of pollen profiles through— out the eastern United States (Sears, 1932); 2. the occurrence of disjunct floral and faunal populations of the grasslands which are now isolated on the Atlantic coast. According to Schmidt (1935), grass— land animals migrated to the coast by way of a "steppe corridor" of grasses extending from a source area in the central United States; 3. fossil indications of former forests in the tundra region of Labrador as well as the disjunct occurrence of certain Appalachian woodland plants in the uplands of New England (Fuller, 1935; Sears, 1935a, 1936b; Hanson, 1937; Voss, 193A). The presence of oak and grass pollen in the Great Lakes profiles may be construed as evidence for a former drier climate. This dry period has been designated as the "Xerothermic Period" and has been described by Gleason (1923), Transeau (1935), Sears (19A2), Deevey (19A9), Charles- worth (1957), and Benninghoff (196A). Dating the Xerothermic Period Radio-carbon techniques provide an estimation of the length of time for each stage of a vegetational sequence. A8 Two profiles (South Haven Bog and Hartford Bog, Zumberge and Potzger, 1956) in the study area allow certain inferences to be drawn regarding the development of prairies in south— western Michigan. The profiles show successions of vegeta— tion from Abies-Picea to Pinus to Quercus, with definite increases of grass pollen in the upper horizons. From carbon-1A analysis, an approximate date of 3,500 to A,000 B.P. has been established for the horizons containing oak and grass pollen. This stage is generally designated as the peak of the xerothermic period. This approximation places the dry period at variance with other estimates. Flint (1953) assigns an age of 5,000 B.P. for a postglacial dry period which affected eastern North America. Antevs (1953) suggests a range of time be— tween A,500 and 7,000 B.P. for such a xerothermic period. On the other hand, Benninghoff (196A) has emphasized the difficulty of dating the xerothermic period in the Great Lakes Region using carbon-1A analysis. Citing a fossil from a spruce-fir horizon in Kalamazoo County, Michigan, dated 13,000: 600 B.P., and with non-arboreal pollen constituting 22.8% of the total pollen count, he suggests that grassland elements may have been present in Michigan as part of an early postglacial invasion. At the present time, the evidence remains inconclu- sive. Consequently, precise dates for a xerothermic period and initial grassland occupance can not be assigned. The A9 palynological evidence does suggest that many taxa now associated with prairie environments were at one time wide- spread throughout much of Lower Michigan.l These distribu- tions were probably in response to postglacial climatic adjustments. 1Moore (1960) discusses the ability of grasses to assimilate silica which strengthens and hardens stems. Subsequently, silica accretions (Opal phytoliths) are in- dicators of the prominence of grasses in an ecosystem. The strong concentrations of Opal phytoliths in the loam soils of Kalamazoo and St. Joseph Counties indicate a long period of grassland occupation. Thus, the silica accretions sup- port the palynological and radio-carbon data. CHAPTER V GEOMORPHIC AND EDAPHIC ELEMENTS IN THE PRAIRIE REGION Climate determines the potential range of a vege- tational community within a region. On the other hand, tOpography, soils, and drainage affect the viability and relative abundance of any species after ecesis has occurred. In humid areas with adequate precipitation, the differentia- tion between forest and grassland communities is determined by conditions of site (Carpenter, 19A0; Buell & Cantlon, 1951). The persistence of prairies in Michigan must be investigated as a function of certain geomorphic and edaphic elements. Surface Configuration The topography of southwestern Michigan is primarily the result of advances and recessions of glaciers during the Wisconsin epoch of the Pleistocene Period. In the prairie area, moraines are the most distinctive geomorphic features (Figure 5). These moraines are products of deposition by two glacial lobes. 0n the west, the Michigan lobe formed the Lake Border,.Valparaiso, Kalamazoo, Sturgis, Lagrange, and Tekonsha moraines. The Saginaw lobe formed the Kalamazoo, 50 m omdmfim 51 k2422wm w_m_ OHLHBLQ mamoa macfiao> ”HHom .usHQ sweepso "andwawoooe .sossoo coacsoaom .ccosoaoocom .3HHm .mae .mH soaooom .smm Lo Nam "sofloaooq H opfim .Aaoma nsomxoflflm Ugo movaoqsom moum¢v cmmHQOHS campmoz Inpsom EH moafigosd Haom oopooaom now moves COHmepHflmcH one endpmflozII.o© canoe 58 AowmpcOoLOQv w.H wm.: H.HH m.mm I 5.0m pcmpcoo hmao . . ommpcooLOQ I aa.oa os.as m:.ss m mm a am Moaoaaao eawsa . . . . . Apcxsav I I I I I m.m Apc\sav HHOm pozIIHprHQH I I I - I A.@ AEH\eAV Haoa scoIIHcHoasH . . . I . AmopocdmoEpmv I mm N 3: m :m 0 0: Hm GQSPWHOE HMHpHCH coamepHHMCH Azfiqu A__:_.~|:®MV A..®MI:OMV AZOMI=WAQNV A..N\~©NI2@V AzmIISOv mo Ho mm mm Hm ¢ monocH CH mspdoo ocm mcomflmom HHom .QOHpmpowo> oaaflmpd mamoa OHQHH0> ”HHom .ocfimnoe Loo: camad nmmzpso wcfipmHSOCD "madmawOQoe .zpcsoo oonmEmHmm .OHLHmLm osopc .zaam .mme .5 coapoom .rzz mo xmm mo r3m ”soapmooq m cosh .Aamma .QomeHam paw movaoczom ampm¢v cmeQOHS campmoz Iapsom :fi mOHHQOLQ HHow Umpooaom pom moans COHpBAOHHmcfi Ono manpmflozll.nm canoe 59 Aomepnoonodv m.m w.H H.m N.Hm m.NH m.> unopnoo meao I . . . . . Aomepnoonodv a: o mo m pm NH Hm :H mm :H zpfioeoeo oaofim I I . . . . Ans\sflv om ma mm m as 0 mm H mpfiaflneoEnem I I I . I . Ahc\snv H m o : Haoa oosIIHoHoHsH I I I . I . Asexsfiv m o H o Haoa ssoIIHoaoHsH . . . . . . AmonondmoEpev mm a am m me 3 am a so as as on osseaaos HesoAsH nOHpeanHmnH A:MH:AV Azflfizlznmv A::ml=Hmv AzHNI:NHv Azmdlzwwv Azwml:ov mo Ho mm mm am as mononH nH mnpdem one mnoNHnom afiom .nOApepowo> phonon meeoa monem OONeEeHeM "aflom .He>oa maneon mnfiead nmezpno ”andeanQOB .sossoo ooaeeofioe .BOHm .mme .a soepoom .ezz no new "soaocooq m opflm .Aamma .nomeHnm one noofionnom nonmev nemfln0fiz nAOpmoz Inpnom nH moafinond Hflom oopoofiom pom mopen nOfipenuHHmnH one onepmflozII.om oHQeB 60 nomeunoonomv ma.o m.HH 0.0m N.Hm m.:a m.m unopnoo zeao . . . . oweunoonod I am.ss on an as ma em ma as on Mofioaoao onwae . . . . . nn\nH I ea an em m on m as m HA an sosMAcaoewoa I . I I . Ahexsev I e m m NH finch oosIIHerHsH . I I . Ann\nflv I I m a a ma HHom maoIIHoHonsH I . . . . . AmononnmoEpev pm :H ma ma om ma Hm Ha mm :H onepmfioe HerHnH noapenpawmnH AtmmlzmmHv A:WMI=N\HJNV A:N\H:NI:~\HN.HV A..N\~N.HI23HV A23HI:HHV Azflfllzcv 0 mm me He me ea mononH nH mnpoom one mnONHnom HHom .nOHpepowo> phonon mEeOH monem ooNeEeHem "HHom .wnfipeanonn mnHeHd nmezpno ”andeanQoe .sossoo soomom .om .zmam .mme .em soflooem .emz no e32 usofloooq : opfiw .Aaoma snowaHnm one noofionnom noumnv newflnoflz nnopmoz Inpnom nH moaflmond HHom oopooaom non woven nOHpeanHmnfi one endpmflozll.oo oHoeB 61 precipitation more readily if they are dry at the beginning of precipitation. As more rain falls, the ability of the soil to infiltrate water decreases as soil pores become clogged by soil particles or the expansion of clays. The permeability data are mean values for infiltra- tion in each profile (Table 6). In general, the infiltra— tion rates are low in the A horizons and increase considerably in the B3 horizons. Site 3 is a deforested area and shows the lowest permeability values for the Ap horizons of the four test samples. The low rate of infil— tration may reflect the effect of rain beat in closing soil pores in the A horizon of this profile. On the other hand, Site A lies under tree cover and has high permeability in the Ap horizon with 19.11 inches per hour. Permeability values decrease in the B2 horizons of Sites 1, 2, and A. A similar decline is evident in the B1 horizon of Site 3. The decline in permeability is apparently related to the accumulation of claysl in the B horizon. Ac- cording to Veatch (1927), the Michigan prairie soils have approximately 25% Clay in the B horizon. The clay content values in the prairie test sites (25.7% and 26.8% in the B2 horizons, respectively) agree with this approximation. 1The clay layer is called "hardpan" by the local inhabitants in the prairie area. However, it does not ap— pear to be a true hardpan such as may be found, for example, in the savanna regions of South America. 62 The forest soils have a slightly lower amount of clay (21.7% in the B2 horizons). The clay layer in the B horizon retards but does not impede percolation of water. The increase in per- meability in the B3 and Cl horizons of the prairie and forest soils reflect the well-drained character of the substrata beneath the outwash plains. Field capacities are also shown in Table 6. These values represent the percentage of moisture which remains in each horizon after gravitational water has drained away. The forest soils tend to have lower field capacities in the A and B2 horizons than the prairie soils whereas the values in the lower horizons of all profiles are not significantly different. However, the sample is too small to derive any significant conclusions. Prairie grasses form a very dense sod (Transeau, 1935). The matted grasses, fibrous root systems, and abun— dant humus in the A horizon retain water more readily than do the forest soils. As a result, field capacities appear to be higher in the prairie soils, especially in the humus layer. During years of drought or unusually low precip- tation, the grasses also make fuller utilization of the available ground water at depth.1 The ability to utilize l"The herbaceous vegetation which covers the prairies is furnished with an immense number of very strong roots, far, more than cultivated grasses. These form a complete mat on the surface and penetrate to great depths-—often met with con— siderable size at 6 to 8 feet. The extraordinary system of 63 available moisture is a function of the fibrous root systems which are more extensive than the shallower root systems of the contiguous trees. The relationship between mean rates of infiltration (permeability), field capacities, and the distribution of prairie clusters in the study area is shown in Figure 7. Field capacities are given in inches.1 In general, the infiltration rates are inversely prOportional to the water—holding capacities of the soils. For example, the upper four feet of the profiles in the sandy shoreline along Lake Michigan have the highest infil- tration rates (8.0 to 12.0 inches per hour) and the lowest field capacities (A.5 to 8.0 inches) for southern Michigan. These values reflect the capacity of the sandy soils to achieve field capacity quickly because water percolates through the profile at faster rates than would normally occur in soils with higher percentages of silt or clay. Selected data for the ranges between minimal and maximal values for infiltration and field capacities in the study area_are presented in Table 7. capillary roots with which they are furnished enable them to remain green and vigorous during periods of drought." (Caton, 1876, p. Al-A2) 1For example, a cubic foot of dry prairie soil may have a field capacity of 2.5 inches. This value means that the volume of soil (one cubic foot) would yield 2.5 inches of "field" water in a flat-bottomed container, one foot square, after gravitational water has been removed. 6A w onsmflm Gem. .5335 use 332.com .0340 .. 0.0m - 0.0. .. 0.0. . 0.0. .. 0.0. . 9n. .. rim. - 0.0. s 0.0. I 0.0 .22: 0.0 - no .0: v .233 >tocd<0 0.5.... .............. ....... ..... oooooo ..... ..... ZO_H