INFORMATION TO USERS This reproduction was made from a copy o f a docum ent sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this docum ent, the quality o f the reproduction is heavily dependent upon the quality o f the material subm itted. The following explanation o f techniques is provided to help clarify markings or notations which may appear on this reproduction. 1. The sign or “ target” fo r pages apparently lacking from the document photographed is “Missing Page(s)” . If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity. 2. When an image on the film is obliterated with a round black mark, it is an indication o f either blurred copy because o f movement during exposure, duplicate copy, or copyrighted materials th at should n o t have been filmed. For blurred pages, a good image o f the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame. 3. When a map, drawing or chart, etc., is part o f the material being photographed, a definite m ethod o f “ sectioning” the material has been followed. It is customary to begin filming at the upper left hand com er o f a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete. 4. For illustrations th at cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Departm ent. 5. Some pages in any docum ent may have indistinct print. In all cases the best available copy has been filmed. University Microfilms International 300 N. Zeeb Road Ann Arbor, Ml 48106 8503201 D o d g e, S h e r id a n L ee SOIL TEXTURE, GLACIAL SEDIMENTS, AND WOODLOT SPECIES COMPOSITION IN NORTHEAST INGHAM COUNTY, MICHIGAN M ichigan S tate U niversity University Microfilms International 300 N. Zeeb Road, Ann Arbor, Ml 48106 Ph.D. 1984 SOIL TEXTURE, GLACIAL SEDIMENTS, AND WOODLOT SPECIES COMPOSITION IN NORTHEAST INGHAM COUNTY, MICHIGAN By Sheridan L. Dodge A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OP PHILOSOPHY 1984 ABSTRACT SOIL TEXTURE, GLACIAL SEDIMENTS, AND WOODLOT SPECIES COMPOSITION IN NORTHEAST INGHAM COUNTY, MICHIGAN By Sheridan L. Dodge This investigation is concerned with assessing the effects of A and B2 horizon textures and the thickness of glacial till on the tree species composition of woodlots in the northeastern sector of Ingham County, Michigan. Three general research hypotheses are advanced and tested. It is hypothesized that, first, soil texture is unrelated to the species mix of regional stands and, second, as a result of this proposed relationship, that no differences exist between in A and B2 horizon soil textures beneath sugar maple-beech and oak-hickory woods. The last, and most important, research hypothesis is that the thickness of till is an important determinant of forest composition. Where till is relatively thick, sugar maple and beech are expected to predominate; where till is thin or lacking and glaciofluvial sediments lie near the surface, woodlots will be composed of oaks and hickories. Forty-eight woodlots, divided into an equal number of sugar maple-beech and oak-hickory, were randomly sampled for species make-up, a composite sample of both the A and B2 horizons, and the thickness of the till substratum. These Sheridan L. Dodge samples were then processed to obtain numeric data suitable for statistical processing. Woodlot species composition is expressed in terms of importance values and soil texture is specified by the soil fractions. Till thickness beneath each woodlot is determined by the average of two sample auger depth records for each stand. These data were then statistically analyzed with t-tests, principal component analysis, cluster analysis, and regression, both simple and multivariate. The results of these tests demonstrate that there are two distinct upland woodlot types, mesic and xeric, in the study area. In addition, there are no significant differ­ ences in A and B2 horizon textures beneath oak-hickory and sugar maple-beech woods. There is a significant difference in the thickness of the fine-textured substratum, however; sugar maple-beech stands are underlain by thicker till than are oak-hickory woodlots. Results of the regression analysis indicate little influence of soil texture on woodlot composition and, moreover, till thickness seems to have little effect on the distribution of tree species within the study region. Nevertheless, a number of problematic considerations make this conclusion on the relationship of till thickness and species composition tentative. ACKNOWLEDGEMENTS I wish to thank my major professor Dr. Jay R. Harman, for his advice, guidance, and criticism during all stages of the dissertation. Appreciation is also extended to Dr. Harold A. Winters and Dr. Richard E. Groop of the Department of Geography, Dr. Peter G. Murphy, Department of Botany and Plant Pathology, and Dr. Delbert Mokma, Department of Crop and Soil Sciences. They provided advice for field, laboratory, and statistical methods as well as constructive reviews of the dissertation. Finally, I am.indebted to Dr. Hunter S. Thompson of the Department of Gonzo Journalism for his sense of humor and "twisted" insights on American society. TABLE OF CONTENTS Page LIST OF TABLES ' V LIST OF FIGURES vii Chapter 1. INTRODUCTION 1 Possible Origins of Forest Association Patterns The Research Problem Hypotheses and Rationale Research Goals 2. THE STUDY AREA - NATURAL SETTING AND LITERATURE REVIEW Topography Glacial Patterns Soils Climate Presettlement Vegetation Development of Modern Forest Patterns Possible Causes of Vegetation Patterns Post-Glacial Climate Fluctuation Fire Disturbance The Influence of Geomorphology The Soil Factor Summary 3. METHODS 3 5 7 9 12 12 14 19 20 23 26 27 28 30 31 32 38 40 Rationale for Research Design Data Collection - Field Data Preliminary Survey Woodlot Selection Woodlot Sampling iii 40 41 41 42 46 iv Chapter Page Data Collection -Archival Data Data Aquisition - Soils Data Processing - Woodlot Data Data Processing -Soils Statistical Analysis Student's t-test Principal Components Classification of Component Scores Linear Regression 4. RESULTS Distribution of Woodlots Woodlots - Modern Composition Woodlots - Presettlement Composition A and B2 Horizon Textures Overburden Thickness Comparison of Soil/Substratum Sample Means PCA of Species Importance values PCA of Edaphic Data Matrix Cluster Analysis of Woodlot Component Scores Relationships Between Soil, Overburden, and Composition Summary 5. DISCUSSION Porest/Woodlot Patterns Woodlot Composition/Communities Effectiveness of PCA - Compositional Trends Woodlot Groupings Modification of the "Oak-Hickory" Label The Problem of Succession Evaluation of Hypotheses 6. SUMMARY AND CONCLUSIONS Summary Conclusions Suggestions for Future Research 52 54 55 58 60 60 61 64 67 71 71 74 77 79 82 84 85 89 90 95 98 100 100 112 112 116 124 130 131 138 138 141 143 LIST OF REFERENCES 148 APPENDIX 162 LIST OF TABLES Table I. 2. Page Selected climatic statistics Upland soils and associated southern Michigan 22 vegetation in 34 3. Mean importance values of oak-hickory woods 75 4. Mean importance values of sugar maple-beech woods 76 5. Mean frequencies of reproduction in oak-hickory woodlots 78 6. Frequencies of reproduction - Woodlot 1 78 7. Mean soil fractions 83 8. Explanation of variance in IV data matrix 86 9. Vegetation component loadings 88 10. JENKS hierarchical clusters 91 II. CLUSTAN hierarchical clusters 91 12. Equation Summaries 96 13. Importance value and frequency comparison of selected species 122 14. Mean importance values of disturbance indicators 123 Al. Importance value program 162 A 2 . Soil fractions program 166 A3. Presettlement woodlot composition 168 A4. Thickness of overlying fine-textured sediment 171 v vi A5. T-test results 174 A6. Explanation of variance inedaphic data matrix A7. Edaphic component loadings 176 A8. Component scores 177 A9. Woodlot township/range designation 179 176 LIST OF FIGURES Figure Page 1. Location of study area 13 2. Glacial landforms of study area 17 3. Presettlement forest of study area 24 4. Location of study woodlots 72 5. A horizon soil textures 80 6. B2 horizon soil textures 81 vii CHAPTER 1 INTRODUCTION General Forest Geography The southern peninsula of Michigan is dominated by two forest associations widespread in eastern North America, beech-sugar maple (Fagus grandifolia-Acer saccharum) and oak-hickory (Quercus-Carya). While the composition of the associations has been well documented, the exact boundaries of these forest regions are indefinite because investigators do not agree on the position of the zone that separates beech-sugar maple forests from the oak-hickory association. Braun (1950) and Vankat (1979) indicate that oak-hickory forests intrude eastward from northwest Indiana into the beech-sugar maple region of Michigan. This lobe of beech and sugar maple extends through Michigan's southern tiers of counties and into Ontario; the boundary between the two forest associations is, however, poorly defined (Vankat 1979). Kuchler's (19 64) map of potential vegetation adds to the uncertainty of forest distributions in southern Michigan. He depicts this portion of the state as a region of oak and hickory with "islands" or inclusions of beech and sugar maple within the former assemblage. A similar forest pattern is depicted on the maps of Veatch (1928, 1959), 1 2 especially the 19 59 map which is more detailed and of a larger scale than the Kuchler map. Given this increased information content, one can discern that Veatch also con­ cludes that a uniform distribution of beech-sugar maple and oak-hickory forests was not present in the pre-settlement vegetation of southern lower Michigan. Benninghoff and Gebben (19 60) propose a similar multiform distribution of beech and sugar maple, as well as oak and hickory, for this region of the state. While many vegetation maps designate Ingham County, located in southern lower Michigan, as part of the region of beech and sugar maple dominants, the nature of this larger scale pattern is also uncertain. Braun's (1950) map sur­ rounds Ingham County with large areas of beech-sugar maple forest as does the smaller-scale vegetation map of Veatch (19 28). The map by Kuchler (1964), a modified version of the Kuchler map (Barnes and Wagner 19 81), and Vankat's map (19 79) also seem to indicate that the county is part of the beech-sugar maple region. Yet, one could argue, based upon a different interpretation of these maps, that the area lies within an ecotonal zone between beech-sugar maple and oakhickory. This interpretation may also be supported by Veatch's map of 19 28 which depicts a small tongue of oakhickory forest that extends into the northeastern sector of Ingham County. When the county is viewed at a larger scale an even more complex pattern of vegetation is revealed. The 3 19 59 map by Veatch shows that the original forest pattern of the area was extremely intricate and beech-sugar maple, oakhickory, and other forest types freely intermixed. Possible Origins of Forest Association Patterns Explanations of the pattern of forest associations in southern lower Michigan are many and varied. The influ­ ential factors may range from climate and microclimate (Braun 19 50, Vankat 19 79) to edaphic (Braun, 19 38, 19 50, Vankat 1979) to geomorphological (Elliot 1953, Kenoyer 1930, 1934, Quick 1924, Society of American Foresters 1954, Veatch 19 28). Most of these studies conclude that the oak-hickory association occupies droughty sands and gravels that are found in outwash plains and some moraines containing coarsetextured sediments. Beech and sugar maple, in contrast, are associated with end and recessional moraines and till plains, glacial features that usually consist of finertextured tills. The correlation between forest type, landforms, and soil texture does not appear to apply, however, to portions of Ingham County. Upland oak-hickory woodlots in the north­ eastern sector of the county are not restricted to coarsetextured glaciofluvial sediments, nor does this species association seem to be correlated with specific glacial morphological features (Dodge 1981). My earlier 4 investigation revealed that oak and hickory, as well as beech and sugar maple, occur frequently on features that have been mapped as moraines and till plains (Martin 19 55, Lemme and Mokma 19 80) and these, in turn, are composed of till. Whereas the soils that are associated with this type of sediment are extremely varied, oak-hickory and beech-sugar maple woodlots are found with nearly equal probability on either loams or sandy loams (Dodge 1981). Soil, therefore, does not seem to play an important or, at least, an apparent role in determining the species association found at a given location. What then is the explanation for the distribution of tree species in northeastern Ingham County, especially when accepted factors do not appear to be important? Vanlier et al. (19 69) may provide a clue to the ques­ tion of cause and the resultant forest patterns in the study area. These authors state that extensive outwash deposits that have been buried by till are found in parts of Ingham County and also to the north in adjacent Clinton County. Within the northeast corner of Ingham County, the Coon Creek valley in Williamston Township is underlain by these buried sands and gravels and valley outwash. Nonetheless, the authors note that the extent of these sediments is not well known because of insufficient drilling and data. If till does overlie coarse-textured materials in Williamston Township, the sediment contact must lie at depths generally 5 greater than five feet because numerous augerings to that depth do not reveal the presence of sand or gravel (Dodge 1981). The depiction of these buried deposits is relatively infrequent on soil maps of the township, although they appear more commonly on soil maps of the two adjacent northern counties, Clinton and Shiawassee (Threlkeld and Feenstra 1974, Pregitzer 1978, Barnes et al. 1979). Despite their infrequent depiction on soil maps of the area, these buried sands and gravels may be an important determinant in the forest patterns of northeastern Ingham County. The Research Problem This investigation will examine the solum texture and the nature and depth of the substratum in order to determine whether these factors are spatially related to the distribu­ tion of upland tree species in northeastern Ingham County, Michigan. Sub-surface coarse-textured glaciofluvial sedi­ ment, as described by Whiteside et al. al. (1968) and Vanlier et (19 69), is the most likely primary causal factor within the study tract. These sands and gravels probably lie at depths greater than five feet and, therefore, are not indi­ cated on most soil profiles of the Soil Survey of Ingham County (Barnes et al. 1979). Although these coarse- textured sediments may have little expression in the solum and C horizons of local soils because of overlying drift, 6 beech and sugar maple may be sensitive to their effects, particularly where these sediments are relatively near the surface. This study, then, will ascertain the thickness of the overlying till, if present, and determine whether there is a correlation between variations in this fine-textured sediment and the forest patterns of the region. The relationship between species composition and solum texture within the study region is not completely under­ stood. Soil maps indicate that most of the upland portions of this area are underlain by relatively fine-textured soils (Barnes et al. 1979). Dodge (1981) found this same pattern in a Williamston and Locke Townships study tract; sandy loams were found beneath most of the sampled oak-hickory and sugar maple-beech woods. Extremely sandy soils were rare in those townships, but coarse-textured soils are locally common in the expanded study region of this thesis. Due to the strong influence of sand on tree growth and species selection, the species composition of woodlots found on these sands may be somewhat different from the oak-hickory woodlots examined in the 1981 study. Because the study area has been expanded and the soils may be more varied, I can not surmise that the findings of my 1981 investigation will necessarily pertain to the larger area. Thus, this study will also examine the soil texture beneath all woodlots. This examination will improve on my survey (Dodge 19 81) and perform a more exacting sampling of both the A and B2 7 horizons beneath selected oak-hickory and beech-sugar maple woodlots. These samples will then be analyzed to determine if there is a difference in soil texture associated with the two forest associations and if a correlation exists between texture and species composition of the region's woodlots. Hypotheses and Rationale I am hypothesizing that the relative thickness of the till that commonly overlies glaciofluvial sediments in this portion of the county will have a strong correlation with the type of forest association that dominates a given site. The rationale for this hypothesis is that rapid drainage and lessened water holding capacity of buried sands and gravels may generate moisture stress where the thickness of the overlying drift is insufficient to store adequate water to meet the growth needs of sugar maple and beech. These two species usually occupy those sites in southern Michigan with mesic soil moisture regimes [those soils with an "adequate" moisture supply throughout the growing season] Wagner 1981:43).^ (Barnes and Thus, a moisture stress would be The adequate designation of Barnes and Wagner (1983) is .not clearly defined. This moisture regime likely corre­ sponds to the "well-drained" and "moderately well-drained" U.S.D.A. drainage classes. In each case, water is available for plant growth throughout the growing season. 8 especially critical with the depletion of soil moisture as the summer progresses. Beech and sugar maple consequently, would likely occupy those sites where sand and gravel are not present or, at least, where these sediments are deeply buried beneath finer-textured tills. These locations, with their associated soil and substratum, would have a greater moisture holding capacity and an ability to retain this moisture for a longer period of time. The oak-hickory association in southern Michigan occurs on sites with xeric and dry-mesic soil moisture conditions where lack of available soil moisture (drought) is character­ istic. (Barnes and Wagner (1981). Therefore, oaks and hickories, unlike* beech and sugar maple, may dominate in the study area where coarse-textured materials have a surface expression or are relatively near the surface, covered only by a shallow layer of unsorted glacial sediment. tion likely makes these sites more droughty. This situa­ In summary, oaks and hickories are expected to be found on upland sites where till is thin or absent. It seems probable that beech and sugar maple are able to outcompete oaks and hickories and, thus, I expect beech and sugar maple to occupy those locations with thick and relatively moist, unstratified glacial drift. Soil texture of the A and B2 horizons, or the A & B where a B2 is lacking, is expected to show little relation­ ship with forest patterns within the study region because 9 soil maps and maps of glacial landforms indicate that similar soil types and glacial deposits are associated with both oak-hickory and beech-sugar maple woods. Therefore, there should be no major difference, on average, in the soil textures found beneath the two species groups. Only those soils that are extremely sandy are expected to be related to woodlot composition. Nevertheless, because these soils are spatially restricted within the study area, they are not expected to bias the overall relationship between soil texture and tree species composition. Research Goals Specific research objectives include: construction of a species composition map of selected upland woodlots in northeast Ingham County; creation of a statistical grouping of these woodlots according to composition; collection of data concerning the oak-hickory succession to the more mesic beech-sugar maple association; and an assessement of the possible soil and subsoil controls on species distribution within the study region. Previous field examination of these topics in the northeastern townships of Ingham County has not been well documented and, although Veatch (19 59) did produce a detailed map of forest cover, his work was not based on field delineation of forest types but rather on the spatial correlation between forest associations and soil. 10 Furthermore, Veatch was concerned only with the uppermost five feet of the soil and parent material; he did not consider the sediments beneath what was commonly considered to be the soil profile. The most important research goal is the assessment of the possible relationship between buried sands and gravels and the distribution of the forest associations found within the study region. To the casual observer, it might appear that the presence of oak^ and hickory forests on loamy soils constitutes an anomaly. upland Yet not only are these forests common within the study area, they appear to be per­ sisting, given that Veatch's map is accurate. Why do oaks and hickories occupy sites with soils that elsewhere support beech and sugar maple, the regional climax? Another goal is to establish in detail the areal of beech-sugar maple study region. extent and oak-hickory woodlots within the Although Veatch's 19 59 map is very detailed, his groupings were presettlement associations; the spatial extent and the actual species composition of present day associations may be quite different because of human disturbance. The study will also build upon past investigations of edaphic factors and their effects on tree species distribu­ tions in the beech-sugar maple region. Many earlier studies have only examined the soil textures of A horizons; in those studies in which the B horizon textures were included as 11 independent variables, soil fractions or textural classes were obtained from published soil surveys (Crankshaw et al. 1965, Lindsey et al. 1976). An improvement can be made by actual field sampling and laboratory analysis of the A and B2 horizons, and by adding one additional factor that has not appeared in previous edaphic studies of the beech-sugar maple region: the statistical and spatial relationship between substratum and tree species distribution. Examina­ tion of buried glaciofluvial sediments is probably the most innovative aspect of this thesis and the concept, as well as its related techniques, may prove useful in the investiga­ tion of plant geographical controls in other regions where textured discontinuities exist between the upper soil horizons and the substratum. CHAPTER 2 THE STUDY AREA - NATURAL SETTING AND LITERATURE REVIEW Ingham County is located in the south-central portion of the lower peninsula of Michigan and has an area of 1,445 square kilometers (558 square miles). The thesis study area encompasses all of the northeastern townships of the county, Meridian, Williamston, and Locke, as well as the northern tiers of sections in Alaiedon, Wheatfield, and Leroy town­ ships. In addition, the study region contains within its boundaries some sections of Bath, Shaftsburg, and Perry Townships in Clinton and Shiawassee Counties (Figure 1). This region of investigation is approximately 533 square kilometers (205 square miles) or about 52,480 hectares (131,200 acres). Topography The landscape of the study area is characterized primar­ ily by nearly level to undulating plains. Some portions, however, are hilly, especially in northern Meridian and Williamston Townships and southern Clinton and Shiawassee Counties (Veatch et al. 1941, Malik 1960). This type of topography is typical of the southern peninsula of Michigan 12 13 Shiawassee Looking Clinton Lansing Hasten* E ast Lansing Okemos mi. km. Ingham Figure 1. Location o f study a r ea . 14 (Vanlier et al. 1969). Swamps and wet river valleys are common but are not as widespread as the upland plains and hills. Most of the study tract has relatively low topographic relief and gentle slopes. Areas south of the Red Cedar River and in Locke Township on the western margins of the study area may have typical relief of 6-12 meters (20-40 feet) and slopes of 2-6 percent. North of the Red Cedar in Williamston Township, slopes are more pronounced and relief is usually greater than the gently rolling topography in the southern sections of the study area. Relief of 15-21 meters (50-70 feet) in the hillier sections is common and slopes of 12-18 percent are typical, although slopes may be as steep as 18-30 percent. Slopes in lowlands throughout the study region range from 0-3 percent. The highest elevation in the area, located in Williamston Township, is 299-302 meters (980-990 feet) and the lowest elevation is 250-253 meters (820-830 feet) along the Red Cedar in Meridian Township. Glacial Features The entire study region has been glaciated by the Wisconsinan continental ice sheet. The glacial history of the area is complex and not well understood, but ice had probably melted from the region by approximately 14,500 15 years ago (Farrand and Eschman 1974). The retreat of the ice sheet margin in this area may not have conformed to the traditional view of orderly ice margin retreat with its associated recessional moraines and till plains, al. vanlier et (19 69) note that strips of stagnant ice and remnant ice, situated parallel to the ice front, were probably left behind as the active ice margin retreated to the north. This form of ice sheet wasting may be the cause of the complex nature of the area's glacial sediments. The composi­ tion of these deposits is highly variable from place to place (Vanlier et al. 1969, U.S. Army Engineer District, 1970b). The glacial sediments that underlie the study area are primarily tills, although water-laid sands and gravels cover some areas, especially along former glacial melt-water drainageways (Martin 1955, Vanlier et al. 1969, U.S. Army Engineer District, 1970b). Deposits of till found in the western margins of the study region are very clayey (U.S. Army Engineer District, 1970b). In general, however, local till is relatively sandy and bouldery, a characteristic of tills associated with the Saginaw Lobe (Wayne and Zumberge 1965). In some places, this till is underlain by deposits of sand and gravel. This buried outwash, believed to be pre-Wisconsinan aged, underlies extensive areas of southern Clinton and northern Ingham counties (Vanlier et al. 1969). However, this type of sediment formation, in a different 16 interpretation, may mark a local readvance of the ice front or even a flow of till that buried adjacent Wisconsinan outwash deposits (Dorr and Eschman 1970). The thickness of the local glacial sediments is not uniform. This thickness generally ranges from 15 to 60 meters (50 to 200 feet) and the sediments are thinnest along the course of the Red Cedar River in Williamston and Okemos (Vanlier et al. 1969). Due to these deep overlying glacial deposits, nowhere in the study region is bedrock (Pennsylvanian Saginaw Formation) exposed at the surface. In the Red Cedar River valley, however, this formation is very near the surface due to local thinning of the glacial sediment (Vanlier et al. 1969, U.S. Army Engineer District 1970b). The principal glacial formations in the area are mor­ aines, till plains, and glacial channel outwash (Martin 1955) (Figure 2). Curvilinear Saginaw Lobe recessional mor­ aines of the Cary (Woodfordian) Substage [16,000 to 13,500 B.P. (before the present)], (Wayne and Zumberge 1965, Dorr and Eschman 1970) are situated in the northern part of Ingham County. These moraines, from south to north, are the Dimondale, Lansing, and Grand Ledge, and they are roughly parallel to one another on the western margins of the study region but appear to merge into a complex moranic system in Williamston and Locke townships. The landscape associated 17 flivet, .Looking Clinton s© &sssifc< 1S^s moraine till plain outwash plain Sgi&aSSSftSaSW^ Ingham Figure 2 . Glacial landforms o f study area ( a f t e r H.M. Martin 1955). 18 with these morainal belts ranges from low, broken knolls, especially in the case of the southernmost Dimondale moraine, to very hilly land where the moraines merge in the eastern portions of the study region (Leverett and Taylor 1915). The authors note that relief is usually less than 7.6 meters (25 feet) but it may range to 30 meters (100 feet) in the more rugged morainic zones. In most cases, however, morainal knolls are hardly distinguishable from till plain (Leverett and Taylor 1915). Most of the area is till plain. These till plains lie between recessional moraines and a very well developed example lies south of East Lansing, between the Lansing and Dimondale moraines. Till in this plain is generally clayey and stoney and poorly drained sites are common (Leverett and Taylor 1915, Veatch et al. 19 41). Till plains found in the study region vary in topography, but most are nearly level to undulating (Veatch et al. 1941). A limited area of valley outwash is located along the channel of the Red Cedar River and its northern tributaries to the east of the townsite of williamston. This area is underlain by thin sand and silt deposits that are deposited in long, narrow, former melt-water channels (Vanlier et al. 1969). 19 Soils Soils within the region are extremely varied because of the complex nature of the parent material, genetic, lithologic, and stratigraphic differences between the solum and the substratum, and the many different glacial landforms found in the region (Veatch et al. 1941). In general, loca­ tions with little slope are less well-drained than areas with greater slope, but texture of the glacial sediment is also important. Relatively flat areas underlain by coarse deposits may have soils that have low moisture holding capacity, and even on moraines with broken topography the variability of soil and moisture conditions may not be so much a product of slope, but rather differences between glacial sediments (Veatch et al. 1941). Upland mineral soil textures in the study area range from loams and sandy loams to loamy sands (Whiteside et al. 1968). Poorly-drained lowlands may have muck soils composed of accumulated organic debris, however. The substratum is highly variable and may have a texture markedly different from the soil horizons, a phenomenon sometimes refered to as a "two-storied" soil. The textural classes of the substra­ tum usually range down the textural triangle from loam to sand and gravel. Examples of the textural discontinuity 20 between the soil and C horizons are found in the Boyer and Oshtemo series in which sandy loam overlies gravelly sand, and in the Owosso series where sandy loam overlies clay loam. Well-drained and somewhat poorly-drained (U.S.D.A. Soil Conservation Service drainage class) soils within the region are alfisols and most are typic hapudalfs. Two exceptions are the Capac, an Aerie Ochragualf, and the Spinks, a Psammentic Hapludalf. Poorly and very poorly-drained soils are extremely varied in their taxonomic classification, with Alfisols, Mollisols, Inceptisols, Entisols, and Histosols all represented in this group of wetter soils. Climate The study area is part of the temperate-continental climate region of eastern North America (Trewartha and Horn 1980). The humid summers of this climate type are warm to hot and winters are cold, but not as long nor as severe as the cool summer climate zone that lies to the north of this region (Gabler et al. 1975). It has been suggested that south-central Michigan may actually lie near the fluctuating boundary between the hot summer and mild summer subtypes of the humid continental climate (Niedringhaus 1966). et al. Vanlier (19 69) have even described the region's climate as 21 continental semimarine because of the moderating influence of Lake Michigan on temperature and the lake's enhancement of precipitation and humidity. This concept is debatable, due primarily to the study area's interior location and distance from the lake, and Niedringhaus (1966) concluded that the climate of the interior of Michigan's southern peninsula is primarily land-controlled but that continentality (ratio between temperature ranges and latitude) is reduced by lake influence. Mean annual temperature for the entire study region lies between 8 to 9 C. (47-48F.) (Table 1). The warmest month is July and January is the period of minimum tempera­ ture. The growing season, or frost free period, is about 160 days and normally lasts from the first week of May to the first week of October. Mean annual precipitation ranges from 762 to 787 milli­ meters (30-31 inches) throughout the study region, but annual extremes from 483 to 1219 millimeters (19-48 inches) have been recorded. Precipitation amounts are significantly lower than stations nearer Lake Michigan and this decline in precipitation reflects the lessened influence of the lake on inland locations. Snowfall is not particularly abundant and the yearly average is about 762 to 1016 millimeters (30-40 inches) throughout the area. Table 1. Selected climatic statistics (U.S. Army Engineer District 1970a). Year of Record Mean Annual Temp. Mean Jan. Temp. Mean July Temp. Lansing 1864 8.3C (46.9F) -5.2C (22.7F) 21.80 (71.2F) 769.6mm (30.3 in) June 88.9mm Feb. 45.7 mm (3.5 in) (1.8 in) East Lansing 1957 8.6C (47.4F) -6.2 (20.9F) 21.30 (70.4F) 650.2mm (25.6 in)* July 101.6mm Dec. 25.4mm (4.0 1n) (1.0 In) Williamston 1931 779.8mm (30.7 in) June 99.1mm Dec. 40.6mm (3.9 in) (1.6 in) Station * may be local anomaly — — — Mean Annual Precip. Mean Monthly Precip. Max. Min. Mean Annual Snowfall 1193.8mm (47.0 1n)* — 779.8mm (30.7 in) 23 Precipitation maximum is the late spring to early summer and a secondary precipitation peak, albeit lower than the spring's, occurs in the autumn. This bimodal pattern is due to the northward and southward migration of the mean jet stream position in the spring and the fall. The result is a quickening of cyclonic activity in these seasons (Trewartha 1981). Precipitation amounts are also enhanced during these months by the advection of maritime tropical air masses into the region due to a general southwesterly air flow in the middle troposphere. Thus, passing cyclonic disturbances in the spring and fall interact with air of significantly greater amounts of moisture than do those of winter (Harman 1971, Harman and Harrington 1978). Presettlement Vegetation The presettlement upland forest of the study area consisted of three major associations: and sugar maple-beech (Veatch 1959) maple-beech" designation, oak, oak-hickory, (Figure 3). The "sugar in this case, may be more appro­ priate than the traditional "beech-sugar maple" because a number of studies have demonstrated that sugar maple is far more dominant in the area's mesic woods (Schneider 19 63, 1966, Beaman 1970, Frye 1976, Beach and Stevens 1980). Veatch apparently believed this situation was also true 24 fliver, .Looking Clinton sugar m aple-beech o a k -h ic k o ry ^v'v.ses lowland deciduous lowland coniferous Ingham Figure 3. P re se ttlem e n t f o r e s t o f stud y area ( a f t e r J.O . Veatch 1 9 5 9 ). 25 prior to settlement. Veatch's (1959) map, based primarily on tree species-soil relationships but also historical records, indicates that most of the uplands north of the Red Cedar River were either oak-hickory or oak. The oak woods were not extensive but were, rather, isolated or clustered patches within large tracts of mixed oak-hickory forest. Large areas of sugar maple-beech forest were present south of the Red Cedar. Exceptions to this general pattern were near the towns of East Lansing, Williamston, and to the southeast of Okemos. The two former areas were forested by sugar maple and beech, while the latter land to the south of the Red Cedar River was a region of oak and hickory. Lowland hardwoods were numerous on the eastern margins of the study region and along tributaries of the Red Cedar. Chandler Marsh north of East Lansing was also an area of swamp hardwoods and probably tamarack. Veatch et al. (1941) state that the patterns of pre­ settlement vegetation in Ingham County were closely related to soil types and the associated soil characteristics. Dense stands of hardwoods grew principally on loam and soils with high clay content while more droughty soils supported forests with a more open appearance. In general, sugar maple and beech grew on loams and clayey soils and oaks grew on sandy soils. 26 Development of Modern Forest Patterns All of Ingham County was surveyed under the General Land Office Survey System (Ordinance of 1785) by 1829. The first permanent settlement, however, was not until 18 34 at Stockbridge in southeastern Ingham County. These first inhabitants had to literally cut their way into the county because an access into the dense forest did not exist. This cutting commenced in the southern part of the county and pro­ gressed northward as settlement expanded into the northern townships (Malik 1960). Once these settlements were established, farmers cut away the surrounding forest and the newly cleared land was placed under cultivation. Cutting occured mainly in the winter months and it proceeded at a rapid pace. An example of the efficiency and effectiveness of forest clearing occured in Wheatfield Township during the spring of 18 36 when two brothers cleared 30 acres of dense forest in twenty-two days (Malik 1960). The rapid expansion of agri­ cultural land continued such that by 1859 much of the cultivatable land of the county was occupied. During the period from 18 38 to 18 59, the number of farms increased from 200 to 2,508 (Malik 1960). Most of the agricultural land was under cultivation by 1875 and the landscape was characterized by 27 rectangular fields and stump fences (Malik 19 60), fences that were the remains of the once extensive "virgin" forest. Only small remnants of this presettlement forest exist today and these are more appropriately called "woodlots" because of their small size. Due to forest destruction, only about 15 percent of Ingham County is forested in a form similar to that of the presettlement eras (Veatch et al. 1941) and nearly all is second growth (Beal 1902, Dice 19 31, Michigan 1941, Beaman 1970, Flanders 1971, Frye 1977, Beach and Stevens 1980). The one woodlot in the study area that is not greatly altered from its original state is Toumey Woods. Only occasionally was dead wood removed from this stand and grazing animals were excluded by a fenced border (Schneider 1963, 1966, Flanders 1971). Possible Causes of Vegetation Patterns The correlation between environmental variables and the distribution of the different upland forest associations within the study region has not been extensively investi­ gated. And yet while local studies are relatively few, numerous geographic studies of other areas in southern Michigan offer various interpretations of the distribution of mesic and xeric forests within this part of the state. 28 Post-Glacial Climate Fluctuations Holocene climate change is suggested by some scientists as an explanation of the mosaic of oak-hickory and sugar maple-beech forests in southern lower Michigan. Both Braun (1938, 1950) and Vankat (1979) have speculated that shifting climatic conditions since deglaciation have helped maintain an oak-hickory subclimax on moisture-stressed sites. Vankat (1979) concludes that this association is but a remnant of more extensive xerophytic forests of the warmer hypsithermal (8,000-4,000 B.P.). Braun (1938:521) believed that a general cooling of the post-hypsithermal climate and the resultant maladjustment in species ranges have created patterns where stands of oak-hickory and sugar maple-beech woods occur in juxtaposition. A few investigators have suggested that given the present climate conditions, the passage of time, and the adjustment of species distributions, the vegetation of the lower peninsula will become the sugar maple-beech climax (Whitford 1907, Quick 1953). However, those who conclude that the local forest pattern consists of an edaphically selected mosaic of climax communities would likely reject this interpretation. Among these are Livingston (1903) who stated that climate plays little role in the determination of forest types in southern Michigan. In addition, both 29 Friesner and Potzger (1934), in their study of forest patterns of south-central Indiana, and Charton's (1972) investigation in northwestern Indiana concluded that precipi­ tation is sufficient to enable the development of the theore­ tical sugar maple-beech climax. But, despite this condition, some sites in both study areas are forested with oaks and hickories. This association in these cases is considered subclimax and there is no evidence of an imper­ fect adjustment of species distribution to the present climate (Charton 1972). The nearly uniform temperatures and precipitation amounts throughtout the study region indicate that these climatic variables also play little role in the vegetation patterns within this area. These climate conditions are adequate for the successful establishment and growth of mesic tree species within the study region, and given the long relatively cool and moist post-hypsithermal period, I suspect that all analagous sites "should" be occupied by mesic species. But this is not the case because oak-hickory woods grow on sites with topography and soils that support stands of beech and sugar maple elsewhere in the study area (Schneider 1963). This situation occurs also in other areas of southern Michigan (Parmelee 1953). Thus, some other factor appears to be a more critical determinant in the local distribution of forest types. 30 Fire Disturbance A second factor, the Indian-ignited wildfire, is sometimes cited as a cause of the distribution of oak and hickory forests of southern Michigan. Schneider (1963, 1966) credits the Indian's use of fire as the most likely determinant of the present vegetation patterns. Evidence of Indian occupation is abundant along the course of the Red Cedar River; this valley had, in all likelihood, the largest aboriginal population in Ingham County (Malik 19 60). The site that is now the village of Okemos was the largest of the many Indian settlements in the area and, in addition, the remains of hunting camps and burial grounds are relatively numerous in the northeastern townships of the county. Despite the many known settlements, the number of inhabitants may never have exceeded 500 at any time (Fuller 1924 in Schneider 1963). Schneider (1963), in citing Curtis (1959), observed that a large native population was unneces­ sary for widespread fires and extensive disturbance of vegetation because, once out of control, the fires were difficult, if not impossible, to stop. Thus, those areas that were in unprotected locations were probably vegetated by subclimax oaks, hickories, and red maple (Beal 1902, Schneider 1963). 31 Parmelee (1953), too, speculates that fire played an important role in the preservation of the retrograde oak forest in some parts of southern Michigan and suggests that these fires may account for the presence of oak woods on soils that, in theory, could support the sugar maple-beech association. He observes, however, that the Ingham County survey of 1825-1827: "includes frequent reference to dense underbrush in the oak upland areas and makes no reference to fires. Moreover, the occasional occurrence of witness trees of such fire sensitive species as Prunus serotina, Ostrya virqiniana and Acer rubrum in the oak upland areas suggest further that fires were not important in determining forest composition in this particular county. " (Parmelee 1953: 164-166) This conclusion may not mean that fire played no role in the presettlement forest of the study area. The "open stands" and "open growth" of oaks, particularly black oak, that occu­ pied sandy soils (Veatch 1959, Veatch et al. 1941) might be explained not only by moisture stress but by relatively fre­ quent, localized fires on dry sites. The Influence of Geomorphology Livingston (1905) considered geomorphology an important factor in the determination of local forest type. Oaks and hickories, as well as beech and sugar maple, are often 32 associated with specific landforms throughout southern Michigan. The oak-hickory type is usually situated on outwash plains, sandy moraines, and dry sandy ridges (eskers and kames?), and the beech-sugar maple group, in general, is associated with moraines, till plains, and better-drained river bottoms (Livingston 1903, 1905, Wood 1914, Quick 1924, Dice 1931, Kenoyer 1930, 1934, 1940, Bingham 1945). But geomorphology does not appear to play a primary role in the determination of forest type within the investigative region of this thesis. A comparison of Veatch's (1959) and Martin's (1955) maps reveals that both associations, oak-hickory and sugar maple-beech, are located on all types of landforms within the study region. The Soil Factor Landforms are often highly correlated with soil type such that sandy soils may be associated with outwash plains, and loamy to clayey soils more likely develop on till plains and in morainal areas of southern Michigan. Thus, while forest types are associated with particular landforms, this relationship may be one of subordinate importance. The actual texture and related water retention capacity of the soils on a given landform are probably a more critical influence on forest distribution in much of the southern 33 lower peninsula (Livingston 1903, 1905). As a result, sites with low available soil moisture are usually occupied by species with lower moisture needs such as oaks and hickories, and this relationship is especially true on glaci­ ated terrain where soil moisture is insufficient for mesic beech and sugar maple (Braun 1938). Oak-hickory forests in southern Michigan usually are associated with dry, sandy and gravelly soils, soils that are normally low in humus and with ground water levels at relatively greater depths in comparison with mesic sites (Livingston 1903, Wood 1914, McCool and Veatch 19 24, Dice 1931, Veatch et al. 1941, Braun 1950, Vankat 1979, Eyre 1980), (Table 2). There may be clay in the substratum of these soils, but the clay is pervious and water drainage is rapid (Veatch 1928). This spatial correlation between oaks and hickories and sandy soils occurs in a number of areas throughout southwestern Michigan (McCool and Veatch 19 24, Kenoyer 1930, 1934) as well as in Kent and Oakland Counties (Livingston 1903, Bingham 1945). Beech and sugar maple, in contrast, usually grow on loams and clayey soils in southern Michigan (Table 2). These soils are well-drained, yet they retain relatively large amounts of soil moisture due in part to the rich, moist humus typical of these soils (Livingston 19 03, Wood 1914, Quick 1924, Braun 1950, Society of American Foresters Table 2. Upland soils and associated vegetation in southern Michigan (after Veatch et al. 1941). Soil Type* Characteristics Substratum Dominants Miami loam well-drained clayey glacial t i l l sugar maple, beech white oak, elm, white ash, hickory, basswood, red oak, black oak - dense stands Conover loam - s i l t loam somewhat poorly drained to poorly drained clayey glacial t i l l elm, ash, basswood, oaks with fewer beech, sugar maple, walnut, and butternut Hillsdale sandy loams well-drained pervious sand clay with sand, clay, gravel pockets oaks and hickories; few to abundant sugar maple, beech, elm, and cherry Bellefontaine sandy loam - loamy sand well-drained mixed/stratified sand, gravel, boulders, clay oaks and hickories Fox Sandy loam - loam well-drained sand/gravel oaks and hickories open stands Coloma loamy sand well-drained sand oaks and hickories Ottawa loamy fine sand well-drained clay oaks Table 2. (cont'd.). loamy sand well-drained sand/gravel oaks Oshtemo loamy sand well-drained sand/gravel oaks and hickories — open stands Berrien loamy sand well-drained clay or waterlogged sand oaks with hickory, beech, red maple, aspen Plainfield * old classification, superceeded by modern surveys 36 1954, Benninghoff and Gebben 1960, Vankat 1979, Eyre 1980). Presettlement forests of beech and sugar maple occupied heavy clay soils in Kent County (Livingston 1903) and also in Van Buren and Kalamazoo Counties (Kenoyer 1930, 1934). At the present time in Oakland County, sugar maple-beech stands are related to clayey soils derived from the sediments of former glacial lakes (Bingham 1945). An exception to this general pattern occurred in Barry, Calhoun, and Branch Counties where a low correlation existed between glacial till and sugar maple-beech forest. However, this species association was strongly correlated with Miami loam (Kenoyer 1940). Why there was a dichotomy between till and the Miami soil, which is normally derived from till parent material, is not explained by the author, but, in any case, the Miami loam has a high moisture holding capacity. Two woodlots within the study area have been inten­ sively studied, specifically in terms of soil, soil moisture, and plant relationships. Reimer (1953) and Schneider (19 63, 19 66) in their investigation of Tourney Woods (a sugar maple-beech woodlot on the Michigan State University campus) found that soil moisture levels did not fall below the amount necessary for plants to maintain growth during periods that ranged from 16 months to two years. Smith (1954) and Coltharp (1958) studied plant-soil water relationships ten miles to the northeast of Tourney 37 Woods at the Rose Lake Wildlife Experiment Station. They concluded that soil moisture was not a limiting factor in the same oak-hickory woodlot, but Smith (1954) did note that during his seven year experiment late summer moisture levels were near the permanent wilting point. The general relationships between soil and forest type may not be as clearly defined as many of these studies indi­ cate. The Society of American Foresters (1940, 19 54) has mapped the oak-hickory association on a variety of welldrained upland soils including loamy soils. In addition, Parmelee (1953) found that the extreme variability of soils in southern Michigan made it difficult to correlate woodlot species composition with soil type. Despite this diffi­ culty, he concluded that a continuum of xerophytic to more mesophytic oak woods does exist in this region. He found xerophytic oak stands associated with sands, loamy sands, and sandy loams, whereas more mesophytic oak woods were located on loams and silt loams. These finer-textured soils, the Miami and Hillsdale series, supported both the upland oak and beech-sugar maple forest type. Parmelee was not able to establish why these soils were associated with both forest types, but he speculated that it might be physical or chemical properties of the soils. He was especially curious about an undisturbed red oak-white oak woods in Ionia County, since severely cut-over. The 38 reproduction layer in that stand was sugar maple and Parmelee was puzzled by the long delayed trend to a mesic sugar maple-beech stand. He thought that some unknown factor, or factors, was retarding or even supressing the conversion of oak woods to the theoretical regional sugar maple-beech climax. Parmelee thus warned other investigators that "any reference to the oak upland type as an edaphic climax must be made with reservation" 1953:164). Like Parmelee, (Parmelee I too wonder whether the term "edaphic climax", as applied to the oak-hickory association of southern Michigan, is appropriate. Might the presence of oaks and hickories in the region of investigation be explained by different factors, either those mentioned in other research or some influence or influences not heretofore investigated? Summary The southern part of the lower peninsula of Michigan was once covered by vast forests of oaks and hickories as well as beech and sugar maple. Little remains today of this forest cover due to extensive agricultural development. The distribution of these forest types has been explained by 1.) variation in soil properties, climate, and 4.) fire. 2.) geomorphology, 3). None of these explanations seems 39 entirely satisfactory; all of the explanations have weak­ nesses and, even when taken in conjunction, the factors do not fully explain forest patterns in the study area. The upland oak-hickory and sugar maple-beech associa­ tions, as noted by Schneider (19 63, 19 66) and Dodge (19 81), occupy similar soils within the area of investigation. questions are then: The why does the oak-hickory association persist on soils that are edaphically equivalent, or at least very similar, to soils that support the sugar maplebeech association? And secondly, might the critical control be found not in the solum but rather in the deeper substra­ tum, a substratum that is sometimes of a different character than the solum (Veatch et al. 1941, Vanlier et al. 1969, Barnes et al. 1979)? CHAPTER 3 METHODS Rationale for Research Design A number of studies have discussed the influence of several environmental parameters on the distribution of upland forest associations of southern Michigan. And many more such influential variables, be they related to soil chemistry, plant-soil water relationships, or interspecific competition, probably await future study. An investigation of all these possible controls is beyond the scope of this thesis; such an investigation would be made more efficient and more operationally feasible by the elimination of pos­ sible controls of apparently secondary importance. A number of factors influential in other parts of southern Michigan do not seem highly correlated with the forest association patterns within the study region. For example, climate is uniform throughout the area and wild­ fire, which is infrequent at present, was probably not a significant disruptive force on vegetation in the county prior to pioneer settlement. Topography and landforms are also unlikely controls on upland species distributions. The topography of the region is not remarkedly distinguished by altitude and slope. Nor is vegetation likely a product of 40 41 glacial landform influence; both upland forest types, oak-hickory and sugar maple-beech, occur on till plain, moraine, and the limited areas of glacial outwash. What, then, are the controls that likely influence the upland forest types within northeastern Ingham County? I have selected two factors associated with the physi­ cal properties of the solum and parent material for this investigation: soil texture and thickness of a finer- textured overburden. Geographic patterns of A and B2 hori­ zon texture have been shown to be related to distribution of oaks and hickories and sugar maple and beech (Crankshaw et al. 1965, Lindsey et al. 1965, Lindsey and Escobar 1976). However, an examination of the phyotogeographical influence of a buried zone of coarse materials and its proximity to the surface has no apparent precedent in biogeographical or ecological literature. This observation, in itself, is an argument for inclusion of this potential control, but the known presence of extensive buried outwash within the study area makes the case for inclusion all the more compelling. Data Collection - Field Data Preliminary Survey A reconnaisance of the study region was conducted to determine the existing distribution of the upland forest associations. This survey was facilitated by the use of 42 topographic maps (Bath, East Lansing, Perry, Shaftsburg, and Webberville 7.5' Quadrangles) on which woodlots are distinctly depicted with green shading. A visual inspection was necessary to determine the species composition of these woodlots and this task was accomplished by observation from an automobile or by actual visitation of a given woodlot. Not all woods were inspected or even typed in this prelimi­ nary examination because only a general assessment and delineation of woodlot composition was necessary at this preliminary stage of the project. Areas roughly delineated as either oak-hickory or sugar maple-beech served as a pool for the selection of sample woodlots for more detailed study. Woodlot Selection A sample set of twenty-four woodlots was drawn from each population of the two regional upland forest associations: oak-hickory and sugar maple-beech. This selection process yielded a total of forty-eight woodlots, an arbitrary number adequate for statistical analysis yet not so large as to be impossible for one person to sample 2 during a reasonable period of a few months. All sample 2 Twenty-five of each forest type were originally sampled. However, one woodlot in each group was eliminated upon secondary examination of the species composition. I felt these woods were too disturbed for inclusion in the sample. 43 sites were located in northeastern Ingham County except for several oak-hickory woodlots in southern Clinton and Shiawassee Counties. The sample area was extended northward into this region because fewer than twenty-five suitable oak-hickory woodlots were found in the Ingham County portion of the study area. It should be noted that my woodlot selection may be controversial in that I have sampled from "associations". This controversy likely arises from the dispute among ecologists over the proper method of classifying vegetation (Mueller-Dombois and Ellenberg 1974, Vankat 1979). The traditional American (Clements) and European (Braun-Blanquet) view is that vegetation can be defined by associated dominant species. However, a more recently developed viewpoint contends species associations are nonexistent, but, rather, species respond to environmental gradients. Thus, the distribution of one species is quite independent of the distribution of another (Whittaker 1975, Spurr and Barnes 1980). The organsmic association, an extreme of the former concept, seems to me untenable, nor do I agree with vegetation scientists who, based on the idea of the gradient, conclude that meaningful vegetation classifications can not be constructed. As Whittaker (1975) notes, classification of vegetation is a product of the mind and grouping is justified by its usefulness. 44 I find the association concept is useful in this investigation for a number of reasons. First, some species do have similar ecological requirements, and, thus, they are located on the same sites; for example, beech and sugar maple (Spurr and Barnes 1980). Second, previous studies have concluded that the upland forest of southern lower Michigan can be divided into two primary classes: oak-hickory and sugar maple-beech (see Chapter 2). And third, although other groupings likely exist within the study region, the sugar maple-beech and oak-hickory associations appear to show the least alteration from disturbance, I have sought to minimize the effects of human-induced disturbance because it makes less clear the correlation between composition and environmental gradients. All woodlots were examined prior to any sampling for evidence of woodlot disturbance. Woods that were severely cut-over or where the second-growth canopy appeared to be dominated by disturbance indicators such as black cherry (Prunus serotina), bigtoothed aspen (Populus grandidentata), or red maple (Acer rubrum) were not used as study sites. Also, woodlots that were being actively grazed were not included in the sample set. However, it was impossible to eliminate all woodlots that had evidently been disturbed by grazing because nearly all have been used for such activity in the past (Beaman, 1970, personal communication with farm woodlot owner, 1981). 45 Other considerations in the selection process were phys­ ical size of the woods, upland situation, and access to the site. Only those woods greater than approximately 2 hec­ tares (5 acres) were included in the investigation sample set in order to eliminate the effects of "edge", a condition where border vegetation differs significantly from the less disturbed woodlot interior (Gysel 19 51, Charton and Harman 1973). Only those woods located on higher topographic posi­ tions were included in the woodlot sample group. mination was based upon a number of criteria: This deter­ site depiction on topographic maps, soil series and drainage class, and visual inspection of woodlot topography and canopy species. Woods with large areas of wet soils and poor drainage were not considered but sites with somewhat poorly-drained soils were included only if the woods were dominated by upland tree species and satisfied the other selection criteria. I was usually able to determine soil drainage conditions through examination of both the over­ story and the reproduction layer. For example, agglomera­ tions of such species as red maple, silver maple (Acer saccharinum), cottonwood (Populus deltoides), elms (Ulmus sp.), and swamp white oak (Quercus bicolor) were suggestive of wet soil conditions and such stands were not included in the sample. Some woods were suitable for investigation but were unusable due to various situations such as lack of access because of land posting, denial of 46 entrance by owners, and woodlot protection (as is the case with Sanford Natural Area on the Michigan State campus). Woodlot Sampling Information about the species composition of the sampled woodlots was obtained by the point-quarter (quadrant) method. This technique has the advantage of reducing the time necessary for sample collection, yet neither data nor accuracy are sacrificed when this method is used in randomly spaced populations (Brower and Zar 19 77). Quadrant points and data (species, diameter, and point-plant distance) were selected and recorded in the usual prescribed manner (Phillips 1959, Mueller-Dombois and Ellenberg 1974, Brower and Zar 1977) and only trees greater than or equal to 7.5 centimeters (3 inches) in diameter (DBH) were included in the sample. Transects were usually aligned parallel to the long axis of the woodlot and their locations were biased toward the interior of the woods so that the effects of "edge" were minimized. Care was also taken so that these traverses did not result in overlap of sample quadrants. In addition, tree species not recorded in the quadrants were noted for their presence in the stands. The number of sampled points and their associated quadrants ranged from eight in some smaller woodlots to twenty-four in the long ones, although the number of cases in the lower end of this 47 range were few. The number of sample points varied between woodlot, despite Cottom and Curtis's (1956) recommendation of at least twenty points, because of the need to avoid wet and highly disturbed areas. In addition, the maintenance of a fixed number of points within all woodlots may have increased the probability of quadrant overlap, particularly in the smaller woods. Woodlots with a relatively homogenous canopy composition were sampled with fewer points and this was particularly true in woods dominated by sugar maple. I ceased sampling in each woodlot when I concluded that further sampling would not add significant amounts of information about a given woodlot's species composition and structure. This decision was based upon visual examination of the species content of a woodlot and a comparison with the data I recorded in the transects through the woods. The seedling and sapling layer (less than 7.5 centi­ meters DBH, taller than 1 meter) was also examined. This stratum of all oak-hickory woodlots was sampled by use of the same quadrants employed for the tree sample and the species of the nearest individual in each quadrant was recorded. Only one sugar maple-beech woodlot was sampled for comparison purposes. The decision not to sample more than this number was based on the known uniformity of repro­ duction under the regional sugar maple-beech climax canopy; sugar maple has been shown to be the most important component of this stratum (Schneider 1963, 19 66, Rogers 48 1978). Observation of all sugar maple and beech woodlots within the study area confirmed this finding. The data obtained from this sampling may serve as a possible indicator of canopy/reproduction layer species uni­ formity in more xeric woodlots. A species similarity between the canopy members and the (reproduction) layer might suggest that the oak-hickory association is indeed a subclimax, whereas markedly different and more mesic repro­ duction might offer evidence that oak-hickory is but a pre­ climax to sugar maple and beech on more moist sites as Parmelee (19 53) suggests. Soil samples were extracted from the A and B2 horizons with a 10 centimeter (4 inch) bucket soil auger. In some woodlots underlain by the Spinks loamy sand, a B2 horizon is not present and a sample from the A & B horizon suffices instead. The samples were removed along random transects through each woodlot and again care was taken to avoid sampling in those areas with obvious wet soil conditions and very poor drainage. Ten to twenty holes were augered per woodlot with the actual number dependent on physical size of the woods and the amount of soil variability as determined by examination of soil maps and observation of the augered soil samples. The A and B2 samples taken from each hole were combined in plastic sample bags to form two composite horizon mixtures for each woodlot (Shickluna 1975, Marks and Harcombe 1981). 49 Data pertaining to the thickness of the fine-textured sediment and the presence of buried sands and gravels were also obtained through the use of a bucket auger. For this procedure, up to 3.7 meters (12 feet) of extension pipe was added to the bucket (0.3 meters long) as the depth of the bore hole increased. This process entailed drilling a 1.2 meter (4 feet) deep hole, extracting the auger if the sand and gravel layer was not encountered, adding another section of pipe, and further deepening the hole. This step was repeated with another extension if fine sediment was still present at the 2.4 meter depth. If the auger was still in till at the 4.0 meter depth, this figure was recorded as the depth of the fine-textured sediment at that site. The holes were not deepened beyond 4.0 meters due to the arduous nature of the work and because buried coarse material located at depths greater than approximately four meters probably does not exert significant control on the distribution of beech and sugar maple. These species do not usually extract moisture from depths greater than nine feet and, even in a very dry year, the four to nine foot depth zone is a moisture reserve used only when a moisture deficit occurs in the near-surface zones (Schneider et al. 1965). Two of these deep holes were located in each woodlot. However, this was not a simple task because many woodlots are underlain by very stoney sediments and the probability of a rock blocking the path of the auger is very high in 50 these cases. After the outset of deep augerings and during the first attempt to auger at a rocky site, I decided that the extreme stoniness of some thick tills would make handaugering to the 4.0 meter depth extremely difficult, if not impossible. I thus modified my thickness determination pro­ cedure so that, if the auger encountered a rock within the upper 2.4 meters (8 feet) of the sediment, data and randomly selected a new hole site. I recorded no This course was repeated at each rock strike above 2.4 meters. In six wood­ lots this procedure was modified because numerous attempts to obtain a second thickness reading was thwarted by seem­ ingly impenetratable stoney till. thickness reading was recorded. In these cases, only one But once the auger pene­ trated below the 2.4 meter depth, I recorded the depth at which the auger stopped, be it due to stone contact or because the auger handle was flush with the soil surface. In effect, this auger depth at rock-strike represents the known thickness of the till at that site. If sand and gravel do exist at depths greater than 2.4 meters, their combined influence on woodlot composition, especially mesophytes, is probably minimal (Schneider et al. 19 65) and the determination of the exact depth of the till is not critical to the investigation. Besides difficulties with rock impediments, some auger holes filled with water when the depth of the water table was reached. It was impossible to auger beyond this depth because of an inability to remove 51 the auger due to suction and repeated collapse of the hole and, as was the case with the stone obstructions, the depth of the hole was recorded as the thickness of the fine-textured overburden. The Spinks soil was a problem with respect to this phase of the data collection. This soil is derived from sandy parent material, probably glaciofluvial sediment, that is usually very close to the surface. A & B Therefore, the A and horizons are coarser than other soils of the area which are for the most part developed from glacial till. Due to the lack of a lithologically distinct fine-textured upper sediment layer, I regarded the Spinks A horizon, a loamy sand, as the fine-textured overburden. The decision whether to follow this practice was made ad hoc in each woodlot because, like other soils in the region, the Spinks is sometimes highly variable. Occasionally the A & B horizon had a relatively high clay content and in these cases I considered both the A and the A & B to represent the thickness of the finer-textured layer. 3 "A & B - Horizons that would qualify for A2 except for included parts constituting less than 50 percent of the volume that would qualify as B" (Soil Survey Staff 1975: 460). 52 Data Collection - Archival Data An examination of the Land Office Survey notes of 18 26 and 1827 was conducted at the Lands Division of the Michigan Department of Natural Resources. The purpose of this archi­ val work was to determine the probable presettlement species makeup of the sample woodlots, although no attempt was made to collect all possible data for construction of a comprehen­ sive map of the study area's original forest cover. A second aim of this data collection was to corroborate Veatch's (1959) map of original forest composition. The determination of presettlement forest type is important in the case of present day oak-hickory woods because a similar species composition of modern woods and the forest noted in the 18 26-18 27 survey would indicate that existent oak-hickory woods are not a result of postsettlement disturbance sugar maple-beech woods. Such a similarity of species of not only oak-hickory woodlots, but also sugar maple-beech, may provide evidence that present forest-soil and substrate relationships are applicable to the region's presettlement forest or, at least, the forests of the early nineteenth century. The original species data were obtained by first locat­ ing the sample woodlots on topographic maps. Some, but not all, of these woodlots are bounded by section lines and in these cases I consulted the survey field notes. The survey 53 traverse notes of the relevant section line were located, chain measures converted to feet, and the locations of marker ("witness") trees determined on the topographic maps. The species of the witness trees that were once located on present woodlot bounds were recorded, and when no marker trees corresponded to the location of sample woodlot, the species of the witness tree nearest to the site of the modern woodlot were used. I was not able to record species for woodlots not located on survey traverses. The field notes referred also to the general nature of the landscape and the forest types encountered along some of the survey lines and these observations were recorded for possible use in analysis and discussion of the thesis problem. Several interpretations of the surveyors' notes were necessary in the data collection process. Some archaic species names were translated into modern species termi­ nology. For example, "lyme", as well as "linden", is basswood (Tilla americana) and "yellow" oak is probably black oak, although this translation is not at all certain because references to "black" and "yellow" oak sometimes occur in close proximity on the same section traverse. This "yellow" oak might also be red oak (Quercus rubra) or possibly chinkapin (Q. muehlenbergii) (Barnes and Wagner 1981). Nevertheless, this translation problem occurred only once and I assumed this species was in fact black oak. Other survey terminology is "sugar land" which refers to 54 st sugar maple forest as does "l01- rate" and "good land". Land designated "2 rate" is oak woods. These conversions were deduced from references of the nature of the land and the type of forest the surveyors were in at the time of reference. Data Aquisition - Soils The composite soil samples from each woodlot were thoroughly mixed by hand in order to improve homogeneity. Clods were broken and large stones and organic matter such as roots, seeds, and leaves were removed. The composite samples were then divided by first quartering each sample and one of these quarters was, in turn, divided into four equal parts. Approximately a quarter of this final division filled a 0.6 liter (1 pint) container and all sample containers were removed to the Michigan State University Department of Crop and Soil Science Laboratory for additional preparation. Each sample was further crushed by a rolling pin and sifted through a 2 millimeter sieve to remove all stones and pebbles. All 9 6 samples, the composite A and B2 samples of forty-eight woodlots, were then prepared for the hydrometer method of particle-size analysis as described by Day (1965) and Mokma (1978). The organic fraction of the A horizon was not destroyed by the addition of hydrogen peroxide 55 (H20 2 ^ because this procedure was not considered criti­ cal for the analysis of the soil fractions. Day's recom­ mended procedures (19 65) were augmented by wet-sieving each prepared 40 grain sample through a 50 micron (0.05 grain) sieve. This sieving removed the sand fraction from the silt and clay portions of the sample which, once removed, was oven-dried, weighed, and recorded for later soil fraction calculations. Hydrometer readings were recorded in a constant tempera­ ture room following the procedure of Day (19 65) and Mokma (1978). Several trial runs were conducted before actual tests of the soil samples began. I used these preliminary sessions to determine the maximum number of suspension cylinders in which I could read the hydrometer with reasonable accuracy and efficiency during any one timed recording session. In addition to the recorded hydrometer readings of the suspensions at the necessary time intervals, I noted the temperature and hydrometer reading of a control cylinder filled with 1000 milliliters of distilled water and dispersal agent. Data Processing - Woodlot Data Importance values (IV) for all tree species P 7.5 centimeters DBH) in each woodlot were calculated from the data collected during the point-quarter surveys. The 56 within-stand species importance value is the summation of the relative density 1.5) on the third component and low second component scores. Only one other woodlot (No. 33) has a high third component score, but the highly disturbed state of this stand results in a "Disturbed 94 negative "Mesic/Xeric" component scores. But in those few cases in which the woodlot scores on the first component are positive, the scores on the "Disturbance" and "Dry-Mesic" components are negative. Toumey Woods is again an anomalous member of this "Xeric" group and this inclusion within the second cluster is probably due to the relatively low nega­ tive score on the "Mesic/Xeric" component. Woodlot 15 also deviates from the general characteristics of the "Xeric" cluster, not only because it is a sugar maple-beech stand, but also because the second component score of this stand is very high. This indicates the woodlot, based upon the presence and importance of flowering dogwood and serviceberry and the physical evidence of past utilization (junk, a somewhat open aspect, and the small stature of many trees), is in a fairly disturbed state. The third CLUSTAN cluster is characterized by those woods that have not only strong negative associations with the first component, but also very high positive "Distur­ bance" component scores. Thus, this cluster can be thought of as a "Disturbed Xeric" woodlot group. An exception to this general trend is Woodlot 31 with a negative score on the second component. The fourth group is a "Dry-Mesic" one and is distinguished by high scores ( 1.5) on the third component and low second component scores. Only one other woodlot (No. 33) has a high third component score, but the 95 Xeric" classification for this woodlot. The final cluster is but a class of one observation, woodlot 34. This woodlot has a very high score on the "Disturbance" component and this extreme measure accounts for this single element cluster. Relationships Between Soil, Overburden, and Composition The simple regression between the overlying sediment thickness and the species importance values resulted in three equations with significant r values. Y x = .014 + .014X Y 2 = .004 + .004X Y 3 = .223 - .014X where: Y^ = basswood IV Y 2 — bitternut hickory IV ^3 = pignut hickory IV X = overlying sediment thickness Thus, only twenty-five percent of the species (three of twelve) in the regression study group are statistically related to the overburden thickness. This relationship, however, is tenuous at best because very little of the variance associated with the dependent variables is 2 explained by any of the three equations. The r values in 96 each case are very low, explaining less than seven percent of the total variance (Table 12). Table 12. Equation summaries. Thickness vs. Species IV r 9 11 Std. Error Est. S1.S_r Sig a Sig b basswood .270 .073 .174 .031 .410 .032 bitternut hickory .248 .061 .060 .045 .420 .045 pignut hickory .254 .065 .184 .041 .001 .041 Thickness vs. "Mesic/Xeric 11 Score £ l£ Std. Error Est. .421 .177 3.156 Siq r S.tS a Sig b .001 .00001 .001 Edaphic Component I Scores vs. Black Oak IV Multiple r r£ Std. Error Est. F .409 .167 .079 1.204 Siq f .004 The regression between the woodlot scores on the "Mesic/Xeric" component and the overlying sediment thickness yields an equation with a significant r value. Y = 7.696 + 1.449X where: 97 Y = "Mesic/Xeric score X = overburden thickness And although this equation explains more of the variance associated with the dependent variable than does the IV-thickness regression, eighteen percent is still rela­ tively low (Table 18). The multiple regression analyses indicate that there is little relationship between the edaphic component scores and either the group of twelve species importance values or the "Mesic/Xeric" component scores. The black oak-edaphic compo­ nent score regression is the only equation with a signifi2 cant independent variable, but again the multiple r value demonstrates that only approximately seventeen percent of the total variance is explained by the equation (Table 18). Y = .409 - .243XX where: Y = black oak IV X^= edaphic component I scores The second independent variable, the scores on the second edaphic component, has an F statistic less than 4.04; there­ fore, it is insignificant and does not enter the equation. The most troublesome aspect of the equation is the nonnormally distributed errors between the estimated Y and the observed Y. This lack of normality of the residuals is indicative of a non-linear relationship between the 98 dependent variable (Y) and the independent variables (X). A number of linear transformations, as described by Kim and Kahout (1975), were employed in an attempt to overcome this non-linear relationship. The use of the logY transformation also resulted in a non-normal distribution of the residuals 2 and an even lower r of .104. LogX and 1/X conversions yielded regression equations with dependent variables with insignificant regression coefficients and, thus, no equations were output by the REGRESSION subprogram. The multiple regression between woodlot scores on the "Mesic/Xeric" component (Y) and the scores on the two edaphic components (X) show no significant relationship between the dependent and independent variables. None of the two independent variables remain in the equation after the stepwise regression procedure. Summary Sugar maple-beech and oak-hickory woodlots are located in distinct areas within the study region and, although the stand composition is somewhat similar to that of the preset­ tlement era, there are some differences. The sugar maple- beech association is an appropriate label for this mesic group of stands, but the modern oak-hickory community also is composed of a significant amount of red maple and black 99 cherry. A PCA of the importance value matrix of the study woodlots reveals discrete underlying gradients, or compo­ nents, within the data. Single variable and multivariate clustering of the woodlots was based upon the position, or score, of the stands along the gradients. The clusters associated with these gradients of 1.) moisture demand, as implied by species-component correlation, and 2.) composi­ tion change in response to disturbance, demonstrates that there are two major woodlot groups within the study area, "Mesic" and "Xeric". In addition, the "Xeric" cluster can be divided into subgroups with different species composition and disturbance responses. Soil texture beneath stands of sugar maple-beech and oak-hickory is not significantly different, but the overlying fine-textured sediment is significally thicker beneath sugar maple-beech woodlots. Despite this difference in thickness, regression analysis shows no strong trend between woodlot composition and sediment thickness, nor is there a relationship between soil texture and the species makeup of mesic and xeric stands. CHAPTER 5 DISCUSSION Forest/Woodlot Patterns The present distribution of sugar maple-beech and oakhickory woodlots corresponds to the upland forest patterns delineated by the Land Office Survey of 18 26-27. The loca­ tions of the modern upland communities are somewhat similar to the patterns mapped by Veatch (19 59), but his depiction and description of the presettlement upland forests of Locke Township seem simplistic. Veatch denotes the upland communi­ ties of this township as consisting primarily of oaks and hickories with infrequent sugar maple and beech while more moist sites were populated by more elm, ash, basswood, shagbark hickory and swamp white oak. However, present upland woodlots in this region and as far north as Perry in southern Shiawassee County are sugar maple-beech; only the extreme northwestern and western-most sections of Locke Township have oak-hickory woodlots. The land survey records indicate that at the time of settlement this entire region was a complex mosaic of sugar maple-beech and oak-hickory upland forests intermixed with lowland hardwoods and 100 101 tamarack swamps. And, like today, the western sections were described as "rolling oak land", but in the central and eastern portions of the township, where only sugar maple-beech stands are now present, oak woods appear to have abutted stands of sugar maple, beech, and other mesic species. The exact extent of these disparate communities can not be determined from the original survey field notes, but the dissimilar areas were probably not large, most likely a few square miles or less. Today, however, there seems to be little evidence of these former oak-hickory forests of central and eastern Locke Township; only sugar maple-beech woods remain. It is unlikely that these remaining stands are the product of succession from the previous oak woods because many of the woodlots contain what appear to be some old-growth (individuals that predate settlement) beech and sugar maple; woodlot 6 may be nearly all old-growth (conversation with land owner 19 81). Although the 150 years that have passed since completion of the original survey may be sufficient time for succession, I see little evidence of this process prior to settlement because the survey notes indicate that the oak lands of Locke Township had an understory composed of species like those in the canopy. There is no mention of sugar maple or beech understory in the presettlement oak forest. Furthermore, the locations and composition of 102 modern sugar maple-beech woodlots correspond to the area of sugar maple and beech forest as noted in the survey of 1826-27. As for the argument that sugar maple-beech woods would have been removed before oak woods because of the early settlers' bias for farming "sugar land", several owners of mesic woodlots told me that their woods had served as a "sugar bush". During earlier times, maple sugar was an important sweetener and the production of this material was a major regional activity. Malik (19 60) observed that approximately 208,000 kilograms (459,000 pounds) of maple sugar was produced in Ingham County from 1840 to 1874. Thus, the early settlers of Locke Township may have preferred to forgo the clearing of all mesic woods despite their notion that these stands indicated better crop land and, instead, may have cleared the oak land rather than sacrifice a source of maple sugar. Like the sugar maple-beech association, the area of the modern oak-hickory community corresponds to the presettle­ ment oak-hickory forest patterns outlined by witness tree notes recorded by the original surveyors of the study region. But Veatch (19 59) depicts an area of oak and hickory that was situated south of the Red Cedar River in northeastern Alaiedon Township and the northwestern sections of Wheatfield Township. Nonetheless, field reconnaissance of this area revealed no trace of oak-hickory woodlots or 103 even smaller, isolated clusters of oaks and hickories that might be remnants of the oak-hickory forest recorded by Veatch. Furthermore, examination of the 18 26-27 field notes reveals that relatively little oak was present in Alaeidon and Wheatfield Townships. Those oaks the surveyors did use as witness tress were members of forest that consisted of "sugar", beech, "linden" or "lyme", ash, and elm; this community was often described by the surveyors as "rolling sugar land". Veatch (19 59) also delineated two small oak inclusions within the sugar maple-beech forest of southwestern Meridian Township, but I again found no evidence of these stands and I do not believe they were present at the time of settle­ ment. One of these oak stands was supposedly near Hudson Woodlot (No. 1) and the other area was in the vicinity of Sandhill Woodlot (No. 15), (the location is difficult to determine given the scale of Veatch's map). Both of these areas are underlain by loamy sands (Zobeck 19 76, Barnes et al. 1979) which Veatch may have used as evidence of former oak woods. But sugar maple and beech are also capable of occupying sandy sites and this is very true in the case of Hudson Woodlot where the northern third of the woods is Spinks loamy sand. Based upon this evidence and that of the original survey, Veatch's depiction of oak and oak-hickory 104 forest sites south of the Red Cedar within the study region appears to be erroneous. In summary, the concept of determining original forest cover through correlation of species composition with soil type may continue to serve a useful function in general small-scale mapping as practiced by Veatch's (1959). But it has limitations for detailed, accurate large-scale depiction and description of presettlement species associations. Field observations of present woodlot composition and the examination of the historical record, in addition to an examination of local soil properties, are critical procedures useful in the differentiation of present and former forest compostion and patterns in this part of Michigan. Of all the study area woods, one sugar maple-beech stand is the most intriguing and problematic in terns of its location and variance from the observed patterns. This is Woodlot 24 which is located in Williamston Township (W 1/2, NE 1/4, SW 1/4, Sec. 9, T4N, RlE) and nearly surrounded by oak-hickory woodlots, although a similar but disturbed mesic woods lies just to the north (Sec. 4, T4N, RlE). Other than these two woods and some groves of sugar maple and beech on the banks of the Red Cedar River (Sec. 34, T4N, RlE), the nearest significant sugar maple-beech woodlots lie 5.9 kilometers (4.5 miles) to the southeast and southwest. 105 There is no evidence in the 18 26-27 survey notes that these two isolated sugar maple-beech woods were ever connected prior to the deforestation of the township; only white oak, black oak, and elm witness trees were recorded on the sectional transect between the two stands. Nor did the surveyors make any mention of a sugar maple-beech inclusion within the regional oak forest. There is no certain explanation for the isolation of these two sugar maple-beech woodlots, although three possi­ ble scenarios may account for the segregated nature of the two stands. These are: 1. The woodlots are remnants of a more exten­ sive pre-hypsithermal mesic species assemblage, 2. the sites were reoccupied by mesic species that persisted through the hypsithermal as rare elements of the xeric oak association or, 3. the stands are remnants of a posthypsithermal mesic riverine forest that was invading the xeric forest region at the time of settlement. Each of these hypotheses has strengths and weaknesses; none seems completely satisfactory. Parmelee (19 47), in an investigation possibly relevant to this problem, observed that post-glacial forest succes­ sion in the Lansing area was characterized by five periods (undated by Parmelee) of distinct pollen zonation. described them in this sequence: He 106 A. a spruce-fir maximum, B. a period of pine dominance, C. an era of early oak dominance followed by an increase of elm, beech, and probably sugar maple, D. a second oak maximum and the decline of elm, beech, and sugar maple and, E. a succession to a mesophytic forest of beech, sugar maple, and hemlock. This sequence of vegetation change is similar to that noted by Bailey and Ahern (19 80) at other sites in the central Great Lakes, although some of their sample sites located within the modern oak-hickory forest region lack the final mesophytic stage noted by Parmelee. The lack of radio­ carbon dates for Parmelee's periods make correlation with the other studies difficult but his C and D stages most likely mark the onset of the hypsithermal, or "temperature peak", that was characterized by a rise in temperature and a decrease in precipitation. The exact duration of this Holocene climatic interval is uncertain but it most likely commenced by 8000 - 9000 B.P. and ended with the onset of a cooler and more moist climate approximately 4000 B.P. (Bailey and Ahern 1980, Davis 1976, Delcourt and Delcourt 1980, Dauherty 1968, Webb and Bryson 1972, Wright 1968). Other estimates date this warm-dry period as long ago as 10,000 - 7000 B.P. 4000 - 2000 B.P. (Barnes and Wagner 1981) to as recent as (Dorr and Eschman 1977.), but, in fact, it 107 is most likely that there was no uniform beginning or end of the hypsithermic interval throughout central and eastern North America (Dauherty 19 68, Webb and Bryson 1972, Wright 1968). Although oaks achieved maximum dominance during the hypsithermal, oaks and hickories are known to have been present in southern Michigan as early as 10,0 00 - 9000 B.P. Beech and sugar maple, however, were later arrivals at about 8000 B.P. and had invaded the central lower peninsula by 7,100 to 7,900 B.P. (Bailey and Ahern 1980, Barnes and Wagner 1981, Bernardo and Webb 1977, Davis 1976, Kapp 1978). Thus, beech, sugar maple, oaks, and hickories were present at the onset, or at least the early stages of the hypsithermal, although sugar maple and beech were not dominant members of these middle Holocene forests (Bernardo and Webb 1977). Forest species of this period, like today, probably competed for sites and because beech was a successful invader of the closed oak forests (Davis 1976), it may have outcompeted oaks and occupied sites with more available moisture. It can also be inferred that sugar maple, too, was associated with beech on mesic locations, although this assumption is difficult to support because sugar maple is not a heavy pollen producer and is not well represented in the fossil pollen record (Davis 1965). 108 With the commencement of drier and warmer hysithermic conditions, species more tolerant of dryness would have become the more successful competitors. Parmelee (1947) observed this trend in his D period when water stress was associated with a marked increase in oak pollen while elm pollen declined, primarily due to the drying of hygric sites (Parmelee 1947). This drying trend and the increased dominance of oaks was common throughout the southern Great Lakes region (Bailey and Ahearn 1980). And this increased moisture stress must also have affected mesic sites within the Lansing area and likely explains the decline in beech and maple pollen in conjunction with the decline of elm. Following this argument, if one assumes that woodlot 24 and the associated mesic stand just to the north pre-date the hypsithermal, one can envision the persistence of mesic species on restricted moist sites during the peak of the warming and drying trend, estimated at approximately 6000 B.P. (Bailey and Ahern 1980, Emiliani 1972). These refugia could not have been more than a few acres in extent because even after a 4000 year long post-hypsithermal interval marked by a cooler and wetter climate the mesic stands in Sections 4 and 9 of Williamston Township were still so small as to have gone unnoticed by the Land Office Survey crews. And given that beech and sugar maple did not become abundant in the Great Lakes until thousands of years after their 109 arrival (Bernado and Webb 1977) and the long duration of the dry hypsithermal, I find the refugium scenario unacceptable. It seems more likely that oaks occupied the site of woodlot 24 during the long hypsithermal. Even today, the stream that flows through the stand is intermittent and the flow would likely have been more restricted during a warm-dry interval. In addition, the site is underlain by very coarse-textured soil and subsoil that would become substantially drier with decreased stream flow, a lower water table, and increased evapotranspiration. The probability is low that a small grove of beech and sugar maple could persist under such conditions for a few thousand years. Another explanation may be associated with the posthypsithermal period when topographically low sites became more moist due to increased water input and reduced rates of evaporation. This process would have allowed sugar maple and beech to reoccupy these wetter locations within the oakhickory portion of the study region. The seed source for this mesophytic reinvasion would have been the isolated and scattered individuals persisting within the oak community, assuming that these mesophytes had not been eradicated within the study area during the long period of warm and dry climate. 110 This model does not explain why so few low moist depressions and cooler north-facing slopes are occupied by mesic stands within the oak-hickory region of the study tract. Only one other stand of beech was observed during the field reconnaisance of the oak-hickory area but this consisted of only a few individuals around a kettle hole (N 1/2, SE 1/4, SW 1/4, Sec. 35, T5N, RlE). Some cooler and moister slopes are occupied by individual sugar maple or beech, but again the sites are not numerous, nor are there many individuals. Given the broken terrain of much of the oak-hickory forest region and the numerous kettles therein, more examples of sugar maple-beech woods would be expected if scattered individuals within the post-hypsithermal oak-hickory forests had served as a seed source that permitted reoccupation of moister sites by beech and sugar maple. A more attractive explanation is an invasion of mesophytes into the oak-hickory region north of the Red Cedar River as the hypsithermal waned. This event may have occurred during the 2000 - 1150 B.P. interval when there was an expansion of mesophytic hardwoods, especially beech, in the Great Lakes region as a result of increased moisture availability (Bernardo and Webb 1977, Bailey and Ahearn 1980). The likely invasion route would have been the stream, and associated moist habitats, that flows Ill southwestward through woodlot 24 to the Red Cedar. If this was the case, the beech and sugar maple along this drainage might have been noted by the surveyors during the winter of 1826-27. Unfortunately, the records contain no reference to the nature of the forests along this stream, although the surveyors did make reference to the stream itself. Another problem with this scenario is that much of the land drained by this stream was tamarack swamp at the time of settlement (U.S. Land Office 1826-27) and these large swamp areas would have presented significant barriers for upstream invasion of sugar maple and beech. These .species may have migrated around the peripheries of these hygric sites but the survey notes do not indicate this and extensive clearing of the swamps and surrounding uplands has erased all possible evidence along the hypothesized migration routes. Another weakness of this explanation is the lack of migration of beech and sugar maple into the oak-hickory forests along other tributaries of the Red Cedar. If these species could invade along one stream, then why not along others, particularly when the Red Cedar banks were occupied by beech and sugar maple at the time of settlement (U.S. Land Office 18 26-27, Veatch 1959)? The answer to why these two sugar maple-beech woods are isolated so far north of the Red Cedar will probably never be known because the ultimate evidence, the continuous forest itself, has been destroyed. 112 Woodlot Composition/Communities Effectiveness of PCA - Compositional Trends The principal component analysis of the woodlot importance value matrix confirms that there is a discontin­ uous distribution of species throughout the study area, but the identification and definition of an environmental grad­ ient by PCA is less certain. The bipolarity of the "Mesic/Xeric1' component loadings and the woodlot scores on this measure demonstrate that oak and hickory are not closely associated, either ecologically or spatially, with sugar maple and beech within the study region. However, it is not possible to definitively state that either this component or the "Dry-Mesic" axis is indicative of an under­ lying moisture influence in light of the regression analysis results. Nevertheless, the PCA does serve as a delineater of ecological relationships and groupings. The ordination, or simplification, of the data set is more likely taxonomic rather than ecologic, wherein the components not only repre­ sent a possible environmental gradient but also serve as a complement to community classification (Whittaker and Gausch 1977). In addition, the axes defined by a PCA may represent the response of vegetation to an influence other than such common limiting factors as light, temperature, moisture, and 113 nutrient availability. For example, the component may indi­ cate the relationship between plants with similar histories and responses to disturbance (Goodall 1954). This is the case with the "Disturbance" component that is characterized by high positive loadings of species that increase in importance in response to disturbance regimes. Despite the "success" of the PCA in the identification of three ecologically interpretable trends, there may be some problems inherent in the factoring of the data set. Peet and Loucks (1977) note that a linear relationship between the components and the underlying environmental influences is rarely the situation; instead, Gaussian and bimodal relationships are more common (Peet and Loucks 1977:490). And bimodality is very evident in the first and most important (in terms of explained variance) component, the "Mesic/Xeric" axis. In addition, the component axes may be linearly orthogonal in the PCA model although uncorre­ lated species response to multiple environmental influences is rare in nature (Kershaw 1973, Orloci 1973). Thus, the PCA model is much simplified and may mask non-linear responses to real-world controls. But as Orloci (19 7 3, 1978) observes, linearity must often be assumed because there is no real solution to the problem of non-linerity other than to limit PCA to a narrow range of species and environmental influences. Both Peet and Loucks (19 79) and 114 Noy-Mier and Whittaker (19 78) stress that only those data matrices with low species diversity can be successfully analyzed by PCA, but this argument is countered by Kershaw (1973). He states that a heterogeneous data set with two clearly identifiable species will be delineated by PCA, although subsequent axes may not be easily interpretable. And what is a homogeneous or a heterogeous data matrix? Peet and Loucks's (1979) matrix totaled sixteen species; is their data set significantly more homogeneous than the twenty-six species analyzed in this investigation? There seems to be little substantive and quantitative advice on this matter in the literature; the "rules" on matrix size are more subjective and qualitative in nature. Finally, Orloci (1978) points out that the percentage of the total variance within the data set that is explaiend by the components is an indicator of the "success" of the PCA. In this study, only about thirty-nine percent of the total is explained by three interpretable components. The remaining components and the associated explained variance indicate that a very complex set of underlying parameters influences the importance and distribution of species within the study area. This complexity of the component axes does not imply that there is a significant ecological or spatial response to all of these trends by the vegetation of the region. Moreover, usually only the first three components 115 are examined in investigations of this type because each of the remaining axes accounts for an insignificant amount of the variation within the data set (Kershaw 1973). Further­ more, the first component of the woodlot species PCA accounts for a major portion of the explained variance; the second and third components explain no more than one-third as much variance as the "Mesic/Xeric" axis. If, indeed, the first component represents the influence of moisture avail­ ability, as may the corollary "Dry-Mesic" component, then this environmental control is by far the major influence on the distribution of tree species within northeastern Ingham County. Despite the potential problems with this or any other PCA of ecological data, I feel the analysis is sufficently accurate in the simplification of the data matrix to three major trends. And while there is no confirmation of the research hypotheses by the PCA, the first and third compo­ nents do imply that moisture availability is a factor as originally hypothesized. The second component also demon- states that approximately 150 years of woodlot use in this part of the county has left a mark, in terms of species composition, even on many stands that were selected for study based upon their relative lack of disturbance. 116 Woodlot Groupings An analysis of the JENKS and CLUSTAN cluster output verifies that the woodlot selection process was true to design in that the stands are members of either of the two regional upland forest associations. This conclusion is based on the results of these two clustering routines wherein the sugar maple-beech woodlots are, for the most part, a distinct "Mesic" group and the oak-hickory stands are allocated space within one or more "Xeric" clusters. It may be argued that the woodlot selection process was biased in that only two observed associations were investigated. But based on extensive field observations and the tradi­ tional American view of community classification character­ ized by the dominant species (Whittaker 1978), I am confident that the major least-disturbed, upland tree associ­ ations within the study area are sugar maple-beech and oakhickory. The exact composition and species importances within the "Mesic" and "Xeric" clusters are highly variable, however. The variability of stand composition is not only observ­ able in the field but also evident from the statistical analyses. Differences in composition are most apparent, both statistically and otherwise, within the oak-hickory study group. This association can be subdivided into two 117 subtypes, one comprised of those species most tolerant of dry conditions and a second group that is characterized by species requiring somewhat more moisture. Most notable of these submesic species is red oak, a species that is not common in the woodlots of the "Xeric" cluster derived in the CLUSTAN analysis. Parmelee (19 53) also found that red oak is characteristic of these more mesic sites throughout southern Michigan. In addition, the "Dry-Mesic" woods contain more red maple and pignut and shagbark hickory than do the "Xeric" stands of primarily black and white oak. The "Dry-Mesic" woodlots are mainly located in the central Bath Township portion of the study area, but this spatial relationship may be merely coincidental due to exten­ sive forest clearing. The present location of these stands, then, might be because these woods were spared while others of this type were not. As for possible environmental causes of the location of these "Dry-Mesic" woods, I can not be certain because members of the group are too few in number for reliable statistical analysis. In any case, the over­ burden thickness does not appear to be a factor because two of the stands have above average till thickness in compar­ ison to all twenty-four oak-hickory woodlots, while the remaining two "Dry-Mesic" woods are underlain by finetextured sediment of below average thickness. As for a soil influence, Parmelee (19 53) concluded that the edaphic factor 118 was the key in differentation of oak upland forests in southern Michigan. However, the soil texture of the A hori­ zons of the "Dry-Mesic" group does not appear significantly different from all other oak-hickory woodlots examined in this study. The B2 horizon textures of this cluster do have a higher clay content than the remaining oak-hickory wood­ lots and it is this finer-textured horizon and associated moisture holding capacity that may account for the higher importance value of red oak, pignut and shagbark hickory, as well as red maple within these woodlots. It is just this finer-textured subsoil that Arend and Gysel (19 53) suspected has a profound control on the growth and distribution of the various oak species within southern Michigan. The most troublesome aspect associated with the cluster­ ing of woodlots is the inclusion of sugar maple-beech stands in "Xeric” clusters and oak-hickory woods in a "Mesic" group. The most obvious example of this is the case of Tourney Woods (No. 4). In both the JENKS and CLUSTAN results, it is classified not as a mesic stand, which the sugar maple and beech composition denotes, but, rather, a xeric stand more similar to the oak-hickory woods within the study region. The cause of this "mis-classification" lies in the method of calculation of the component scores, the attributes upon which the clusters are derived. SPSS determines the scores through multiplication of the 119 component-score coefficient (loading) matrix and the standardized importance value vector with the standard­ ization taking the form of: z^ = (IV^ - mean of IV^)/standard deviation of I V ^ Thus, in the case of Tourney Woods which has only two recorded species importance values (sugar maple and beech), the subtraction of the mean species IV from IV^, IVj, IV^, and so forth, results in a negative value for nearly every z^. And because there are many more high positive loadings than negative ones, the sum Of the matrix multipli­ cation (score coefficients x standardized IV1s) yields a negative "Mesic/Xeric" score for the woodlot. It is this negative score that causes this particular stand to be grouped with oak-hickory woodlots in both clustering examples. This computation of the component scores also accounts for the incorporation of sugar maple-beech woodlots 12, 17, and 20 into the JENKS "Xeric" group. In addition, woodlots 34 and 36 are included in the "Mesic" cluster by the JENKS computation and, yet, both stands are oak-hickory. Thus, the JENKS clustering technique is not as indicative of the actual nature of the study woodlots as is the CLUSTAN pro­ gram, the results of which appear to mis-classify only one stand, that being Toumey Woods. This discrepancy between the actual state and the statistical approximation of 120 reality shows that, despite the widespread use and admitted benefits of quantitative analysis of ecological data, good "old-fashioned" observation and common sense still serve biogeography and ecology. In addition to the mesic/xeric dichotomy within the data, the effects of disturbance within the study region, so often evident from field observation, are also indicated by the CLUSTAN computations. The key characteristic of the woodlots within these "Disturbed Xeric" and "SeverelyDisturbed Xeric" clusters is some combination of relatively high importance values for the five disturbance indicator species: sassafras, flowering dogwood, serviceberry, bigtooth aspen, and shagbark hickory. Consequently, these stands have high second component scores because the five species have high positive loadings on the "Disturbance" component. It might be argued that these clusters are not a product of disturbance but, rather, the chance concentration of the five species in a given woodlot. Nevertheless, as noted in Chapter 4, these species, because of their biology, are generally considered reliable indicators of some pre­ vious disruptive processes that have altered the forest composition. Two of these species, sassafras and bigtooth aspen, are intolerant, fast-growing, pioneer species that do not withstand competition very well and as a result are associated with the higher light levels found beneath 121 canopy gaps and abandoned fields. The remaining three species, flowering dogwood, serviceberry, and shagbark hickory, are more tolerant, slower-growing, and often very persistent in the understory and they are usually released upon some disruption of the overstory (Fowells 19 65, Barnes and Wagner 19 81). It is these very species, along with black cherry and red maple, that are often the most abundant species within the oak stands of southern Michigan, a region where two-thirds of the oak woods have been grazed or heavily cut (Gysel and Arends 1953). Moreover, an examina­ tion of the overstory and reproduction layer of the disturbed-cluster woods (Table 13) leads me to conclude that chance can not alone account for the much greater magnitude of importance and frequency of the five species within the designated disturbed stands. 122 Table 13. Importance value and frequency comparison of selected species. "Disturbed” vs. Other Oak-Hickory Woods within Study Region Mean IV Mean Frequency O-H "Disturbed" O-H sassafras .062 .004 .12 .06 bigtooth aspen .008 .004 • o o .01 serviceberry .016 .004 .40 .16 flowering dogwood .006 .000 .09 .03 shagbark hickory .020 .006 .03 o o "Disturbed" N = 9 Oak-hickory N = 15 • "Disturbed" It is notable that only oak-hickory woods are included within the "Disturbed" group of the CLUSTAN analysis. This is because the five species I have designated as indicators of past disturbance are not as common within the sugar maple-beech woodlots of the study area (Table 14). Why are these species so much more important in oak-hickory stands, given that some of the species may be competitive following disturbance on evenmesic sites(Barnes This situation may be due, inpart, and Wagner 1981)? totheabundant reproduction of very tolerant and successfully competitive 123 sugar maple within sugar maple-beech woodlots. Heavy maple reproduction is the case in even some of the more disturbed sugar maple-beech woodlots included in this investigation and this is especially true on the Michigan State campus (woodlots 1, 8, and 10). A coppice of sugar maple seedlings and saplings of uniform height has arisen where major disturbance (probably grazing) has ceased. Thus, based upon my observations, disturbance may alter the importance values of species within mesic woodlots, but it does not necessarily follow that species characteristic of xeric stands will become more dominant in mesic woods (Harman and Nutter 1973, Nutter 1973, Parmelee 1953). And the species so indicative of disturbance in oak-hickory stands appear to be outcompeted in sugar maple-beech woodlots regardless of the condition of each stand, be it relatively unaltered or subject to continual human-induced disruptive activity. Table 14. Mean importance values of disturbance indicators. Oak-Hickory Sugar Maple-Beech .073 002 017 000 serviceberry .024 .006 flowering dogwood .007 .002 033 001 sassafras bigtooth aspen shagbark hickory 124 Modification of the "Oak-Hickory" Label Until this point, I have used the commonly accepted term, oak-hickory, to describe the xeric stands within the study area. Obviously, there are variations of composition among the examined woodlots, most notably in the dominance of various oak species. This situation appears to be especially true on the sandier sites where black and white oaks are the major components of the canopy. However, there are two species other than oaks or hickories that are charac teristic of the xeric and sub-mesic woodlots of the region; these are red maple and black cherry. Neither red maple nor black cherry appears to have been a frequent member of the presettlement forest because the original surveyors made little reference to them. However, Parmelee (1953) did point out that this issue was uncertain, made all the more so as a result of Beal's (19 02) observation that red maple was common in the oak stands then situated on the site of Michigan State University. Parmelee concluded that red maple may be returning to a more important, yet still secondary, status within the woodlots of southern Michigan. However, in this portion of the state, I can not agree that red maple is of only secondary importance given that the average importance value (.504) of this species is exceeded only by black and white oak and is 125 nearly two times greater than that of pignut hickory and red oak. This situation is not uncommon within southern Michigan, for Gysel and Arend (1953) noted that red maple, as well as black cherry, are often the most common species on formerly oak-dominated sites. In southern Wisconsin as well, Larsen (19 53) found red maple was becoming more frequent in oak woods and he concluded that it may be replacing oak as the regional dominant. A recent study by Peet and Loucks (1977) also indicates that black cherry has more recently become a codominant, along with shagbark hickory, in the oak stands of the southern portion of Wisconsin. The reasons presented for these trends are numerous. It has been observed by many investigators that oak reproduc­ tion is often very poor while red maple and black cherry are more prolific reproducers. Disturbance, such as grazing or natural tree fall, may favor red maple and black cherry reproduction, especially during periods of drought. And in more open disturbed stands the reproduction of relatively intolerant cherry may be enhanced. In addition, selective cutting of woodlots for firewood or structural timber may also favor black cherry and red maple because they are less valuable commercially for these purposes. It has also been observed in southern Michigan that oaks are relegated to drier locations while red maple and black cherry tend to 126 dominate more moist areas that are probably former oak sites, an observation corroborated by the correlation of red maple and the "Dry-Mesic" component in this investigation (Fowells 1965, Gysel 1956, Gysel and Arend 1953, Harlow et al. 1979, Westveld 1949). Finally, Parmelee (1953) suggested that the trend to an increased dominance of red maple and black cherry may not be a product so much of disturbance but, rather, is representative of a long-term fluctuation of species importance about a midpoint that represents an optimal climax. Despite this view of oscillating climax species composi­ tion, I interpret the conspicuous presence of red maple and black cherry in both the canopy and the reproduction layer of the local xeric stands to be more likely a product of past disturbances. For example, repeated grazing, as noted by Westveld (19 49), leads to acorn destruction and the estab­ lishment of species with abundant seed production and relatively high tolerance. Given these conditions and repeated good seed years, red maple will likely have a high probability of successful restocking of disturbed xeric sites. And this success may be all the more probable given that red maple seedlings and saplings that grow on dry sites develop deep initial tap roots, a phenomenon that is unusual in red maples located in more moist areas (Fowells 19 65, Tourney and Korstain 1947). It is this characteristic that 127 may enable red maple to successfully compete with deeprooted and drought tolerant oaks. Repeated cutting of local woodlots has also altered the species composition. Since oak is a more valuable wood in terms of both lumber and fuel than either red maple or black cherry, selective removal of oak also probably accounts for the relatively high impor­ tance values of normally secondary species. Good examples of this are two non-study woodlots (NW 1/4, SE 1/4, Sec. 24, T4N, RlW and NE 1/4, SE 1/4, Sec. 24, T4N, R1E) where oak removal has resulted in an unnatural dominace of red maple. In addition to disturbance, the increased importance of red maple may be due, in part, to the reduction of a certain disturbance, specifically fire. Fire is usually considered an event that favors the establishment of pioneer species, but the survival of red maple, heretofore regarded as an indicator of disturbance, (Fowells 1965). is severely reduced by fire This species is very succeptable to burn damage, often resulting in the death of the individual. Although Parmelee (1953) and I found little evidence of presettlement fire in the original survey records of Ingham County, the apparent increased dominace of red maple may be explained by the lack of extensive and repeated wildfires in relatively recent times. Black cherry, in contrast to red maple, is an intoler­ ant species characteristic of secondary succession that 128 often declines and dies when overtopped. It produces some viable seed in every year and these are usually enough to produce abundant reproduction that is characterized by vigorous, although shallow-rooted growth (Fowells 1965). Thus, the lower importance value of black cherry in compar­ ison to red maple is likely due to its lack of tolerance and drought resistance on xeric sites. On the other hand, the very high frequency of cherry reproduction in oak-hickory woodlots is a result of the reproductive traits of this species. What will be the future canopy composition of the oak-hickory stands within the study region, given the dearth of oak reproduction? In all likelihood, the importance of red maple will continue to increase, but whether enough of the remarkable number of black cherry seedlings and saplings survive to ensure canopy codominance is problematic. Grazing of woodlots is no longer a common practice and this may result in a lower frequency of cherry reproduction due to increased competition from more tolerant red maple, oaks, and hickories. Nevertheless, continued cutting of woodlots, especially with disregard of silvicultural practices, may well induce replacement by opportunistic black cherry. Continued poor seedling establishment of both black and white oak may lead to a decline in their canopy importance despite the usual longevity of mature individuals. However, 129 several good seed years coupled with high germination rates will ensure continued white oak dominace because individuals of this species may remain supressed for many years and then, with release, resume rapid and vigorous growth (Forwells 1965). The continued dominance of black oak will be dependent not only on maple seed production and establishment, but also continued opening of the stands by disturbance because black oak is not very tolerant and usually dies when overtopped (Fowells 1965). In summary, the xeric woods of the region might more properly be described by a term other than "oak" or "oakhickory". Eyre (19 80) has designated forests of this part of the Middle West as the "White Oak-Black Oak-Northern Red Oak" cover type where hickory usually comprises less than ten percent of the stand. recognized. Nevertheless, local variants are Red oak is not a codominant, on the average, within area woodlots and a local type for this portion of Michigan would more accurately be described as the "Black Oak-White Oak-Red Maple" group. It is a non-traditional title, but one that is more apropos given the history and resultant composition of the woodlots in northeast Ingham County. 130 The Problem of Succession I have found little evidence resulting from this study that supports the supposition, maintained by some investiga­ tors (see Chapter 2), that oak-hickory woods of northeast Ingham County are destined to become mesic in character. This conclusion is based on the following findings: 1. Reproduction of sugar maple, beech, and bitternut hickory is infrequent in oak-hickory woodlots. This situation is quite unlike the abundant sugar maple reproduction Parmelee (1953) observed in certain oak stands of southern Michigan. 2. Scattered variously aged concentrations of sugar maple and beech in oak-hickory woodlots are situated primarily on compensated (cooler, with less evaporation and more available moisture) north-facing slopes (for example, Woodlot 42). Thus, these occurrences seem to be unlikely precursors of a widespread mesic trend. 3. Moreover, an examination of the Land Office Survey notes (1826-27) reveals that the composition of the understory of the primeval oak forest was remarkably similar to that of the overstory (white oak, black oak) at that time. These findings cast further doubt on the local applicability of the traditional view of autogenic succession [vegetation change induced by the vegetation itself (Spurr and Barnes 1980)] to a regional climatic monoclimax (Cowles 1899, Clements 1916). This model of succession has been 131 repeatedly criticized as too simplistic (Mueller-Dombois and Ellenburg 1974). The occurrence of two relatively stable and persistent forest types within my study region is indica tive, instead, of a regional polyclimax (Nichols 19 23) wherein succession is controlled, not solely by climate, but by "specific environmental and biotic conditions" (Spurr and Barnes 19 80:413). And the probable controls, in this case, are the thickness of till and the presence of a near­ surface, coarse-textured substratum beneath regional stands. Evaluation of Hypotheses I initially hypothesized that there is little relation­ ship between soil texture and the species composition of upland woods within the study region. This hypothesis is confirmed by the regression between soil texture (as represented by the woodlot scores on the components of the soil texture PCA) and the tree species importance values which revealed that the relationship between the two variable sets is extremely weak. In addition, I also proposed that, because of the hypothesized lack of a textural influence of upland species distribution, there would be no significant differences in the mean soil textures beneath the sugar maple-beech or oak-hickory woodlot groups. The results of the t-tests establish that 132 the soil texture of both the A and B2 horizons beneath oak-hickory and sugar maple-beech woods are essentially alike. Thus, this corollary hypothesis can be accepted in light of these results. These findings do not conform to many of the soilvegetation relationships evident in other upland regions of southern Michigan (see Chapter 2) where sandy sites gener­ ally support oaks and hickories while heavier soils are occupied by mesophytes such as beech and sugar maple. However, in northeast Ingham County the oak-hickory and sugar maple-beech associations are situated on relatively fine-textured soils which includes sandy loams, loams, and sandy clay loams. This is also true of the coarser-textured soils of the region; oak-hickory and sugar maple-beech woods are situated on both sands and sandy loams. This finding further corroborates the tendency for there to be little geographical or statistical relationship between soil texture and the distribution of upland forest associations within the study area. The reason for this lack of a reciprocal relationship between texture and vegetation leads to the final and most important hypothesis. I originally proposed a correlation between the species composition and the thickness of a zone of till that often overlies sands and gravels within the study region. While the results of the t-test demonstrate 133 that sugar maple-beech stands are underlain by significantly thicker till than are oak-hickory woodlots, there is little in the results of the regression analysis that establishes a strong statistical relationship between till thickness, and consequent near-surface glaciofluvial deposits, and woodlot species composition. Why then, despite a statistically demonstrable difference in till mantle thickness beneath each woodlot group, the apparent unacceptability of the hypothesized relationship? Several explanations are possible. First, the study region is characterized by a very complex soil pattern. In addition, deep augering reveals not only a highly variable thickness of till and near-surface proximity of sands and gravels but also an extremely varied nature of the materials themselves. The glacial sediments beneath a given woodlot might range variously from tills to heavy lacustrine clays to sands or gravels. With only two deep auger holes in each woodlot, the probability is high that the data collected from the deep augering did not necessarily represent the "average" arrangement of the sediments beneath each woodlot. This situation is particularly true at a site with a complicated history of glacial deposition, for example, a former ice-contact zone or a glacial drainageway. The end result is that the sample set of overburden thicknesses, while large enough for statistical evaluation, may be too 134 small to accurately represent the spatial complexity of the glacial sediments. A larger number of deep bore holes per woodlot would yield a more accurate determination of till thickness and the depth to buried coarse-textured materials. The end result of this increased sample might be a strengthening of the correlation between till thickness and woodlot species composition and distribution. Because the mean till thickness value is based on only two auger depth readings per woodlot, a random near-surface strike of an isolated sand and gravel deposit would necessarily reduce the mean till thickness in that stand. Or the random selection of an auger site which is underlain by uncharacteristicly thick till in a woodlot characterized by near­ surface sands will yield an average till thickness greater than expected. In effect, the former case dampens the corre­ lation between thick till and species characteristics of mesic sites while the latter situation lessens the intensity of the relationship between near-surface sands and oaks and hickories. An increased sample size per woodlot would decrease the likelihood of random and untypical samples masking the hypothesized relationships. A further complication is the impediment to deep augering caused by high water table and rock obstructions (see Table 9, Chapter 4). On a number of occasions the auger struck water at depths less than 4 meters (13 feet) 135 and in most of these cases strong suction nearly prevented removal of the bit. As a result, the sediment thickness was recorded as the final depth of the auger bit. Both this situation and the assignment of the depth at a rock obstruction to the till thickness record in a given woodlot tend to arbitrarily reduce the mean overburden thickness. This loss of accuracy adds additional problems in the interpretation of the influence of fine-textured sediment on the vegetation patterns in northeast Ingham County. The presence of a high water table also complicates the interpretation of the influence of glacial sediments. This is particularly true where water-bearing glaciofluvial sedi­ ments occur beneath mesic woodlots, such as Woodlots 19 and 24. The effect of this phenomenon runs counter to the hypothesis because in these cases near-surface sands and gravels support sugar maple and beech, not oaks and hickories. The presence of a high water table probably offsets the inherent dryness of coarse-textured sediments thus providing such moisture-demanding species such as beech, sugar maple, and basswood with a readily available supply of moisture. The seasonal adjustment of these high water tables is also difficult to interpret. It was impossible to adapt the sampling process to offset these changes because the augering was conducted over a period of nearly nine months commencing in April and ending in 136 December. Sediments that were dry in late summer might not necessarily be so in the spring and early summer. This situation would again be most critical in mesic woodlots where the implicit droughtiness of near-surface galciofluvial sediments is likely mitigated by readily available moisture in the early stages of the growing season. However, the effects of a high water table beneath oak-hickory stands is less clear. I suspect that the occurrence of such a condition is not very likely because large amounts of available water probably lead to the dominance of sugar maple and beech. This certainly appears to be the case in woodlot 24 (discussed earlier in this chapter), a mesic woods in a region of oaks and hickories. Moreover, I noted only one oak-hickory stand where the auger struck water at depths less than 4 meters (13 feet) and, thus, it seems to be an uncommon phenomenon in comparison with the sugar maple-beech woodlot group. A final problem with the analysis of the sediment thickness-species connection is that the relationship may be non-linear. As mentioned in Chapter 4, linear transforma­ tions failed to elucidate or improve the correlation of till thickness and species I V 1s. Although more complex transfor­ mation might be attempted, the use of complex polynomial equations for biological modeling is not always desirable. Mead (1971) supports this point of view and maintains that 137 these curves are not very interpretable, at least in terms of biological relationships. He concludes that simply using a complex equation to describe the unknown relationship between the dependent and independent variables is unacceptable. In other words, an investigator in this situation is merely "grasping for straws" in hopes of finding some relevance in the data. It would, as a result of Mead's critique, seem inappropriate to contemplate further "massaging" of the data with various regression models. In the end, the ecological significance of buried sands and gravels to regional forest patterns is still left some­ what uncertain. The relationship, as ascertained by my data sampling and analysis, seems rather weak, but I think that the complexity of the local glacial sediments makes a defini­ tive conclusion based only on my data impossible at this point. Nevertheless, the demonstrated pattern of near­ surface glaciofluvial material beneath oak-hickory stands and thick till under sugar maple-beech woods, as well as the influence of moisture availability implicit in the "Mesic/Xeric" and "Dry-Mesic" components, leads me to believe that the hypothesized relationship between substratum and forest type is valid and could be demonstrated with improvements in the research design. CHAPTER 6 SUMMARY AND CONCLUSIONS Summary This investigation was conducted in order to determine the relationships between soil texture, glacial sediment, and the distribution and composition of upland woodlots in northeast Ingham County, Michigan. To ascertain the rela­ tionships between these variables, three research hypotheses were tested. First, the soil texture of the A and B2 soil horizons was expected to have little correlation with the upland forest patterns within the study region. Second, and as an outgrowth of the first, I hypothesized that there would be no differences between the mean soil texture beneath oak-hickory and sugar maple-beech woods, the two major regional upland forest types. And third, the thick­ ness of glacial till, which often overlies glaciofluvial sediments in the region, was expected to be related to the species composition of local forest stands, wherein oaks and hickories should dominate where overlying till is thin or absent, whereas those stands underlain by relatively thick fine-textured glacial sediments would be composed primarily of sugar maple and beech. 138 139 Forty-eight woodlots, evenly divided between sugar maple-beech and oak-hickory, were randomly sampled for species composition, composite soil type, and the thickness of the fine-textured overburden. These field-collected data were then processed so that woodlot composition could be expressed in terms of respective species importance values. In addition, the texture of the composite soil samples was analyzed with respect to the percentages of sand, silt, and clay. The till thickness records for each woodlot were averaged to obtain a representative figure for each case. A supplemental examination of the original Land Office Survey records was conducted to ascertain the presettlement species composition of both the overstory and the reproduction layer of many study stands. I used a number of statistical methods to analyze the data; these include t-tests, principal components analysis, cluster analysis, and regression. The mean soil fractions and till thickness for each woodlot group were compared with Student's t-test while principal component analysis was used to extract the underlying statistical interrelationships of the species importance value matrix. In turn, the resultant woodlot component scores of this analysis were classified with two clustering methods, one a single variable algorithm and the other a multivariate technique. Finally, simple and multiple regressions were conducted to determine the 140 strength of the relationship between the independent vari­ ables of soil texture (measured in terms of stand scores on the resultant components of the PCA of the edaphic vari­ ables) and till thickness and the dependent measures of tree species importance values. Results of the t-tests establish that there are no significant differences in mean soil texture beneath sugar maple-beech or oak-hickory woods. However, there is a significant difference in average thickness of the till that underlies the study stands. Till tends to be thicker beneath sugar maple-beech stands while oak-hickory woodlots are underlain by thin till and near-surface sands and gravels. The PCA of the species importance value matrix extracted three major components, these composite indicators being designated as "Mesic/Xeric", "Disturbance”, and "DryMesic". Clustering of the woodlot scores on these three components resulted in a number of interpretable clusters. The JENKS method yielded "Mesic" and "Xeric" clusters while the multivariate CLUSTAN analysis resulted in five interpret­ able groups. These represent the major regional forest types, "Mesic" and "Xeric", with a third cluster encompassing those stands best described as the "Dry-Mesic" subtype. The remaining two clusters are characterized by oaks and hickories, as well as tree species indicative of past disturbance. 141 Simple and multiple regressions indicate a real correla­ tion between the predictor variables of soil texture and till thickness and the dependent species importance values which are representative of stand composition. None of the regression equations explain much of the variance associated with these dependent variables. Conclusions The results of this investigation demonstrate that sugar maple-beech and oak-hickory woodlots are not randomly dispersed within the study area. They are, instead, situated in large areas dominated by either beech and sugar maple or oaks and hickories. Oak-hickory woods are confined to a region north of the Red Cedar River in the central sector of the study area and this tract appears to extend northward into unsurveyed (in this study) sections of Clinton and Shiawassee counties. A region of sugar maple-beech stands encircles this oak-hickory core to the east, west, and south. And based on the examination of the witness tree records, this general distribution of woodlots is similar to the pattern of upland forest types evident at the time of settlement. The principal component and cluster analyses show that field-observable differences in woodlot composition are statistically demonstrable. And not only are there two rela­ tively undisturbed forest types within the region, but also 142 a third which is characterized by species in the relative center of the regional moisture requirement continuum. More­ over, the effects of disturbance, as revealed by the PCA and clustering, are very common within the portion of the study area dominated by oaks and hickories. These effects are so prevalent within the region that it seems nearly impossible to select a sample set of woodlots with negligible amounts of disturbance. Thus, a xeric forest subtype, incorporating a number of species characteristic of disturbed sites, is now a distinctive element of the regional landscape. A probable result of this persistent disturbance is the increased importance since the time of settlement of a num­ ber of species, particularly red maple. Red maple is now the codominant species in many oak-hickory stands and this is a situation unlike that of about 150 years ago when these woods were primarily white and black oak. At the time of the original survey, red maple was rarely, if ever, noted in the witness tree record. As a consequence of these recent changes, a more appropriate designation for these stands, stands that have been traditionally designated as oak or oak-hickory, would be oak-red maple. The expected end product of this dissertation, the con­ firmation of the research hypothese, was not entirely suc­ cessful. The lack of a strong correlation between upland woodlot species composition and the texture of the A and B2 143 horizons is a justified conclusion and, in essence, the aver­ age texture of these horizons is the same beneath both sugar maple-beech and oak-hickory woodlots. However, the results of the regression analyses do not establish the validity of the key hypothesis dealing with the influence of buried sands and gravels on stand composition. Several factors may mitigate the effects of this hypothesized relationship. These include the variability of till thickness beneath a given woodlot, insufficient sample size, the seasonal posi­ tion of local water tables, as well as non-linear relation­ ships between the responses of individual species to the effects of near-surface glaciofluvial sediments. Thus, the substantiation of the primary hypothesis remains an unre­ solved issue. Suggestions For Future Research Although this study has come to a conclusion, I would like to recommend additional investigative work that may lead to a more thorough understanding of the relationships between soil, glacial sediment and forest type within Ingham County. A possible first step would be a more intensive sampling of the thickness of the till that underlies most of the regional upland woods. The result of this increase in sample number per woodlot would be a dampening of the effects of anomalous samples, such as the occurrence of 144 rather thin till in a woods with otherwise thick till. This more detailed sampling procedure would be further aided by the acquisition of a powered auger and the help of one or more assistants in the field. Not only might the within- woodlot sampling be increased, but the study area might also be expanded. Parmelee (1953) observed‘that some of the oak stands included in his study were undergoing an apparently delayed process of mesophytic succession. But, in other cases, the oak composition of the stands was being maintained despite the occurrence of relatively fine-textured soils at those sites. Deep augers into the substratum beneath these woodlots (Parmelee noted that some of these woods had been destroyed prior to the completion of his dissertation) would determine whether the relationship between till and outwash is similar to that which occurs in Ingham County. Also, this arrangement of glacial sediments, wherein till overlies sands and gravels, and their effect on the distribution of forest types in southern Michigan is more commonplace than once realized. If the regression approach to this problem proves unworkable, then a more direct assessment of moisture stress may be in order. Elton (1970a, 1970b), Charton (1972), and Charton and Harman (1973) have employed tree ring analysis to appraise moisture stress on upland tree species and a similar approach could be used in my study region. A 145 working hypothesis, in this case, might be that where till is relatively thick (mesic sites), one would expect a moderate moisture stress, whereas on sites with near-surface sands and gravels (xeric sites), the tree ring record would indicate more periods of moisture stress. A dendrochronological investigation of this type might be a rewarding tack for future biogeographical research in this part of the state. Another method of ascertaining the effects of moisture stress on composition is the reciprocal planting of beech, sugar maple, and selected oak and hickory seeds and seedlings in sugar maple-beech and oak-hickory woodlots. As a result, one might expect, based on the findings of this thesis, a relatively low germination and/or survival rate of beech and sugar maple on xeric sites. On more mesic sites, where soil moisture is available throughout the growing season, survival rates of all plantings would be expected to be higher. If this expectation is confirmed, the result would suggest that the lack of abundant sugar maple and beech reproduction in oak-hickory woods is a function of an inherently dry substratum. Despite the likeihood of relatively high seedling survival rates on mesic sites, why is there a scarcity of natural oak and hickory reproduction in sugar maple-beech woodlots? The answer may not be so much related to substratum and soil moisture, but the lack 146 of suitable seed sources and the relatively lower tolerances of oaks and hickories in comparison to beech and sugar maple. In any case, this type of investigation might yield new insights on ecological relationships in southern lower Michigan. Finally, with regard to the question of succession in the local oak woods, one could take a more definitive position on this subject if it were examined as the primary research objective. This problem might be examined through use of the trenched plot, as utilized by Craib (19 29) and Tourney and Kienbolz (19 31). These scientists contended that premature death of seedlings on certain sites was due to a lack of moisture, primarily as a result of root competition. By severing all roots that entered a number of study plots, Tourney and Kienbolz (19 31) observed that vegetation on the plots became more luxurient over time, as well as more woody and mesic in nature. A series of such experiments in a number of local oak stands might enable one to assess the possibility of a potential mesic trend and to determine if such a tendency is retarded by interspecies moisture competition or whether it is forestalled by the nature of the substratum. If interspecies moisture competition proves to be limiting the traditional view of succession induced by organic matter accumulation and a change to a cooler and 147 more moist microclimate may not be applicable in this part of the state. On the other hand, continued xeric condition in the trenched plots would be further evidence of moisture stress induced by near-surface sands and gravels. LIST OF REFERENCES LIST OF REFERENCES Arend, J. L. and L. W. Gysel. 1951. Some effects of soil and topography on oak yields in southern Michigan. Lake State Forest Experiment Station Technical Notes 365. Bailey, R. E. and P. J. Ahearn. 19 80. A late- and post­ glacial pollen record from Chippewa Bog, Lapeer Co., MI: further examination of white pine and beech immigration into the Central Great Lakes Region. In Geobotany II, ed. R. C. Romans, pp.53-74. New York: Plenum Press. Barnes, B. V. and W. H. Wagner, Jr. 1981. Michigan trees: a guide to the trees of Michigan and the Great Lakes region. Ann Arbor: The University of Michigan Press. Barnes, J. R . , W. L. Bowman, G. H. Earle, Jr., R. J. Engel, D. F. Gibbs, S. G. Holcomb, and G. Threlkeld. 1979. Soil Survey of Ingham County, Michigan. United States Department of Agriculture, Soil Conservation Service. Washington: United States Department of Agriculture. Beach, J. H. and W. D. Stevens. 19 80. A study of Baker Woodlot II. Description of vegetation. Michigan Botanist 19:3-13. Beal, W. J. 19 02. Some changes now taking place in a forest of oak openings. Annual Report of the Michigan Academy of Science 4:107-108. Beaman, J. H. 1970. A botanical inventory of Sanford Natural Area I. Physical Environment. Michigan Botanist 9:147-164. Benninghoff, W. S. and A. I. Gebben. 1960. Phytosociological studies of some beech-maple stands in southern lower Michigan. Papers of the Michigan Academy of Science, Arts, and Letters 45: 83-91. 148 149 Bernardo, J. C. and T. Webb III. 1977. Changing patterns in the holocene pollen record of northeastern North America: a mapped summary. Quaternary Research 8:64-96. Bingham, M. T. 1945. Flora of Oakland County. Institute of Science Bulletin 22. Cranbook Braun, E. L. 1938. 19:515-522. Ecology Deciduous forest climaxes. ________ . 1950. Deciduous forests of eastern North America. New York: The Free Press. Brower, J. E. and J. H. Zar. 1977. Field and laboratory methods for general ecology. Dubuque: Wm. C. Brown Co. Charton, F. L. 1972. "A dendrochronological appraisal of climatic, edaphic, and floristic patterns in north­ western Indiana". Ph.D. Dissertation. Michigan State University. Charton, F. L. and J. R. Harman. 1973. Dendrochronology in northwestern Indiana. Annals of the Association of American Geographers 63:302-311. Clements, F. E. 1916. Plant succession: an analysis of the development of vegetation. Carnegie Institute of Washington Publication 242. Cited in Forest Ecology, 3rd edition, S. H. Spurr and B. V. Barnes, p.413-414. New York: John Wiley and Sons. Clifford, H. T. and W. Stephenson. 1975. An introduction to numerical classification. New York: Academic Press. Coltharp, G. B. 19 58. "Some effects of commercial-type clearcut on soil and water relationships on a small wooded watershed". Ph.D. Dissertation. Michigan State University. Computer Centre, University College of London. 1975. CLUSTAN. London: University College of London. Cottam, G. and J. T. Curtis. 1956. The use of distance measures in phytosociological sampling Ecology 37:451-460. 150 Cowels, H. C. 1899. The ecological relations of the vegeta­ tion on the sand dunes of Lake Michigan. Botanical Gazette 27:95-116, 167-202, 281-308, 361-391. Cited in Forest Ecology, 3rd edition, S. H. Spurr and B. V. Barnes, p.413-414. New York: John Wiley and Sons. Craib, J. L. forest. 25. 1929. Some aspects of soil moisture in the Yale University School of Forestry Bulletin Crankshaw, W. B., S. A. Quadir, and A. A. Lindsay. 1965. Edaphic controls of tree species in pre-settlement Indiana. Ecology 46:688-698. Curtis, J. T. Madison: 1959. The vegetation of Wisconsin. University of Wisconsin Press. Dagnelic, P. 1973. Factor analysis. In Ordination and classification of communities, ed. R. H. Whittaker, pp.223-248. The Hague: Junk. Daubenmire, R. F. 1974. Plants and environment. York: John Wiley and Sons. New Daugherty, H. E. 1968. Quaternary climatology of North America with emphasis on the State of Illinois. In The quaternary of Illinois, Special Bulletin 14, ed. R. E. Bergstrom, pp.61-69. Urbana: University of Illinois. Davis, M. B. 19 65. Phytogeography and palynology of north­ western United States. In The quaternary of the United States, eds. H. E. Wright, Jr. and D. G. Frey, pp.377-401. Princeton: Princeton University Press. ________ . 1976. Pleistocene biogeography of temperate deciduous forests. Geoscience and Man 13:13-26. Day, P. R. 1965. Particle fractionation and particle size analysis. In Methods of soil analysis: part I , ed. C. A. Black, pp.545-567. Madison: American Society of Agronomy. Della-Bianca, L. and D. F. Olson, Jr. 1961. Soil-site studies in Piedmont hardwood and pre-hardwood upland forests. Forest Science 7:320-329. 151 Delcourt, P.A. and H.R. Delcourt. 1980. Vegetation maps for eastern North America: 40,000 yr. B.P. to the present. In Geobotany II, ed. R. C. Romans, pp.123-166. New York: Plenum Press. Dice, L. R. 19 31. A preliminary classification of the major terrestrial communities of Michgian Papers of the Michigan Academy of Science 14:217-240. Dodge, S. L. 1981. "Oak-hickory and beech-maple woodlots of northeastern Ingham County: a preliminary report". Research Paper. Department of Geography. Michigan State Unviersity. Dorr, J. A., Jr. and D. F. Eschman. 1970. Geology of Michigan. Ann Arbor: The University of Michigan Press. Elliot, J. C. 1953. Composition of upland second growth hardwood stands in the tension zone of Michigan as affected by soils and man. Ecological Monographs 23:271-288. Elton, W. M. 1970a. "Forest and lake breeze patterns and the Laporte, Indiana, rainfall anomaly". Ph.D. Dissertation. Michigan State University. ________ . 1970b. Forest distribution and soil texture along the Valparaiso Moraine in northwestern Indiana. Geographical Bulletin 1:41-54. Emiliani, C. 1972. Quaternary hypsithermals. Research 2:270-273. Quaternary Eyre, F. H. 1980. Forest cover types of the United States and Canada. Washington: Society of American Foresters. Farrand, W. R. and D. F. Eschman. 1974. Glaciation of the Southern Peninsula of Michigan: a review. Michigan Academician 7:31-56. Ferguson, R. 1977. Linear regression in geography: concepts and technigues in modern geography. No. 14. Norwich: Geo Abstracts. 152 Flanders, R. A., Jr. 1971. "Temporal associations among woodland plants in southern Michigan". M. S. Thesis. Michigan State University. Fowells, H. A., ed. 1965. Silvics of forest trees of the United States, (Forest Service Handbook 271). Washington: United States Department of Agriculture. Friesner, R. C. and J. E. Potzger. 1934. Contrasts in certain physical factors in Fagus-Acer and QuercusCarya communities in Brown and Bartholomew Counties, Indiana. Butler Unviersity Botanical Studies 4:1-12. Frye, D. M. 1977. A botanical inventory of Sandhill Woodlot, Ingham County. I. The vegetation. Michigan Botanist 15:131-140. Fuller, G. N., ed. 1924. Historic Michigan (Ingham County)# Vol. III. Dayton: National Historic Association. Gabler, R. E . , R. Sager, S. Brazier, and J. Pourciau. 1975. Introduction to physical geography. San Francisco: Rinehart Press. General Land Office Survey. Michigan, Vol. 7. 1826-1827. 1826-1827. Field Notes, N & W Field Notes, N & W Michigan, Vol. 31. Gleason, H. A. 19 26. The individualistic concept of the plant association. Bulletin of the Torrey Botanical Club 53:7-26. ________ . 19 39. The individualstic concept of the plant association. American Midland Naturalist 21:92-110. Goodall, D. W. 1954. Objective methods for the classifica­ tion of vegetation. Australian Journal of Botany 2:304-324. Grieg-Smith, P. 19 64. Quantitative plant ecology. York: Plenum Press. New Groop, R. E. 1980. JENKS: an optimal data classification program for choropleth mapping. Michigan State University Department of Geography Technical Report 3. 153 Gysel, L. W. 1951. Borders and openings of beech-maple woodlands of southern Michigan. Journal of Forestry 49:13-19. ________ . 1956. Measurements of acorn crops. Science 2:305-313. Forest Gysel, L. W. and J. L. Arend, 1953. Oak sites in southern Michigan: their classification and evaluation. Michigan state College Agricultural Experiment Station Technical Bulletin 236. Harlow, W. H., E. S. Harrar, and F. M. White. 1979. Textbook of dendrology. New York: McGraw-Hill Book Co. Harman, H. H. 1967. Modern factor analysis. University of Chicago Press. Chicago: Harman, J. R. 1971. Troposheric waves, jet streams, and United States weather patterns Associations of American Geographers Resource Paper II. Harman, J. R. and J. A.' Harrington. 1978. Contrasting rainfall pattens in the upper middle west. Annals of the Association of American Geographers 68:402-413. Harman, J. R. and M. D. Nutter. 1973. Soil and forest patterns in northern lower Michigan. The East Lakes Geographer 8:1-12. Hausch, B. and W. H. Lyford. 1956. White pine growth and soil relationships in southeastern New Hampshire. New Hampshire Agricultural Experiment Station Technical Bulletin 95. Hill, M. D., R. G. H. Bunce, and M. W. Shaw. 1975. Indica­ tor species analysis, a division polythetic method of classification, and its application to a survey of native pine-woods in Scotland. Cited in Multivariate statistical methods, L. Orloci, p.278. The Hague: Junk. Hough, A. F. and R. D. Forbes. 1943. The ecology and silvics of forests in the high plateaus of Pennsylvania. Ecological Monographs 13:299-320. Kapp, R. D. 1978. Presettlement forests of the Pine River watershed (central Michigan) based on original land survey records. Michigan Botanist 17:3-15. 154 Kenoyer, L. A. 1930. Ecological notes on Kalamazoo County, Michigan, based on the original land survey. Papers of the Michigan Academy of Science, Arts, and Letters 11:211-218. ________ . 1934. Forest distribution in southwestern Michigan as interpreted from the original land survey (1826-32). Papers of the Michigan Academy of Science, Arts, and Letters 19:107-111. ________ . 1940. Plant associations in Barry, Calhoun, and Branch Counties, Michigan as interpreted from the original survey. Papers of the Michigan Academy of Science, Arts, and Letters 25:75-77. Kershaw, K. A. 1973. Quantitative and dynamic plant ecology. 2nd ed. New York: American Elsevier Publishing Co. Kim, J. O. 1975. Factor analysis. In Statistical Package for the Social Sciences, 2nd ed., N. H. Nie et al., pp.468-527. New York: McGraw-Hill Book Co. Kim, J. 0. and F. J. Kahout. 1975. Multiple regression analysis: subprogram regression. In Statistical Package for the social sciences, 2nd ed., N. H. Nie et al., pp.320-367. New York: McGraw-Hill Book Co. Klecka, W. R . , N. H. Nie, and C. H. Hull. 1975. cal package for the social sciences primer. McGraw Hill Book Co. Statisti­ New York: Korstian, C. F. 1927. Factors controlling germination and early survival in oaks. Yale University School of Forestry Bulletin 19. Kuchler, A. W. 1964. Potential natural vegetation of the United States. New York: American Geographical Society. Larson, J. A. 1953. A study of an invasion by red maple of an oak woods in southern Wisconsin. American Midland Naturalist 49:908-914. Lemme, G. D. and D. L. Mokma. 1980. Soil classification. East Lansing: Department of Crop and Soil Sciences, Michigan State University. 155 Leverett, F. and F. B. Taylor. 1915. Pleistocene of Indiana and Michigan and the history of the Great Lakes. United States Geological Survey Monograph 53. Lindsey, A. A., W. B. Crankshaw, and S. A. Quadir. 1965. Soil relations and distribution map of the vegetation of presettlement Indiana. Botanical Gazette 126:155-163. Lindsey, A. A. and L. K. Escobar. 1976. Eastern deciduous forest, Vol. 2: beech-maple region, (Natural history theme study No. 3). Washington: National Park Service. Livingston, B. E. 19 03. The distribution of the upland plant societies of Kent County, Michigan. Botanical Gazette 35:36-55. Cited in Aims and methods of vegetation ecology, D. Mueller-Dombois and H. Ellenberg, p.24. New York: John Wiley and Sons. ________ . 1905. The relation of soils to natural vegeta­ tion in Roscommon and Crawford Counties, Michigan. Botanical Gazette 39:22-41. Cited in Aims and methods of vegetation ecology, D. Mueller-Dombois and H. Ellenberg, p . 24. New York: John Wiley and Sons. McCool, N. M. and J. O. Veatch. 1924. Sandy soils of southern peninsula of Michigan. Michigan Agricultural Experiment Station Special Bulletin 128. McGahan, M. W . , J. H. Brown, W. P. Gould, H. C. Lammpe, and R. E. Sheaven. 19 61. Site condition and tree growth analysis in Rhode island forests. Rhode Island Agricultural Experiment Station Bulletin 357. Malik, J. A. 19 60. "Historical geography of Ingham County". Ph.D. Dissertation. Michigan State University. Marks, P. L. and P. A. Harcombe. 1981. Forest vegetation of the Big Thicket, southeast Texas. Ecological Monographs 51:267-305. Martin, H. M. 1955. Map of the peninsula of the southern peninsula of Michigan. Michigan Geological Survey Division Publication 49. Mead, R. 1971. A note on the use and misuse of regression models in ecology. Journal of Ecology 41:215-220. 156 Michigan, Department of conservation. 19 41. Second annual report; Rose Lake Wildlife Experiment Station. Lansing: State of Michigan. Michigan State University, Computer Laboratory. 1978. SPSS 60C0 supplement. East Lansing: Michigan State University Board of Trustees. Mokma, D. L. 1978. "Appendix III: hydrometer method of particle-size analysis". East Lansing: Department of Crop and Soil Sciences, Michigan State University. Mueller-Dombois, D. and H. Ellenburg. 1974. Aims and methods of vegetation ecology. New York: John Wiley and Sons. Nichols, G. E. 1923. A working basis for the ecological classification of plant communities. Ecology 4:11-23, 154-172. Cited in Forest Ecology, 3rd edition, S. H. Spurr and B. V. Barnes, p.413-414. New York: John Wiley and Sons. Nie, N. H., C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. H. Bent. 1975. Statistical package for the social sciences. 2nd ed. New York: McGraw Hill Book Co. Noy-Meir, I. and R. H. Whittaker. 1977. Recent develop­ ments in continuous multivariate techniques. In Hand­ book of vegetation Science, vol. 5:technigues, 2nd ed., ed. R. H. Whittaker, pp.337-378. The Hague: Junk. Niedringhaus, T. E. 1966. "A climatology of Michigan". Ph.D. Dissertation. Michigan State University. Nutter, M. D. 1973. "Soil texture and forest change in northern lower Michigan". M.A. Thesis. Michigan State University. Orloci, L. 19 67. An agglomerative method for classifica­ tion of plant communities. Journal of Ecology 55:193-206. Cited in CLUSTAN, Computer Centre, University College of London, p.38. London: University College of London. ______ . 19 73. Ordination by resemblance matrices. In Ordination and classification of plant communities, ed. R. H. Whittaker, pp.249-286. The Hague: Junk. 157 ________ . 1978. Multivariate analysis in vegetation research, 2nd ed. The Hague: Junk. Parmelee, G. W. 1947. "Postglacial forest succession the Lansing area of Michigan: a study of pollen spectra". M.S. Thesis. Michigan State College. in ________ . 1953. "Oak upland continuum in southern Michigan". Ph.D. Dissertation. Michigan State College. Peet, R. K. and Q. L. Loucks. 1977. A gradient analysis of southern Wisconsin forests. Ecology 58:485-499. Phillips, E. A. 1959. Methods of vegetation study. New York: Holt, Rinehart, and Winston. Pregitzer, K. E. 1978. Soil survey of Clinton County, Michigan. United States Department of Agriculture. Soil Conservation Service. Washington: United States Department of Agriculture. Quick, B. E. 1924. A comparative study of the distribution of the climax association in southen Michigan. Papers of the Michigan Academy of Science, Arts, and Letters 3:211-245. Reimer, C. W. 1953. "Seasonal radial growth of sugar maple as related to certain environmental factors". Ph.D. Dissertation. Michigan State University. Rummel, R. J. 1967. Understanding factor analysis. Journal of Conflict Resolution 2:444-480. Rogers, R. S. 1978. Mature mesophytic hardwood forest: community transitions, by layer, from east-central Minnesota to southeastern Michigan. Ecology 62:1634-1647. ________ . 1970. Applied factor analysis. Northwestern University Press. Evanston: Schneider, G. 19 63. "A twenty-year ecological investiga­ tion in a relatively undisturbed sugar maple-beech stand in southern Michigan". Ph.D. Dissertation. Michigan State University. 158 ________ . 1966. A 20-year investigation in a sugar maplebeech stand in southern Michigan. Michigan Agricul­ tural Station Research Bulletin 15. Schneider, G . , D. P. White, and R. L. Harlan. 1965. Soil moisture regime under old growth hardwoods. Papers of the Michigan Academy of Science, Arts, and Letters 51:13-21. Shickluna, J. C. 1975. Sampling soils. Michigan State University, Cooperative Extension Service Bulletin E-498. Smith, L. R. 1954. "Some effects of vegetal cover upon the hydrology of watersheds at East Lansing, Michigan". Ph.D. dissertation. Michigan State College. Society of American Foresters. 1940. Forest cover types of the eastern United States. Washington: Society of American Foresters. Society of American Foresters. 1954. Forest cover types of North America. Washington: Society of American Foresters. Soil Survey Staff, United States Department of Agriculture, Soil Conservation Service. 1975. Soil Taxonomy. Agriculture Handbook 436. Washington: United States Department of Agriculture. Spurr, S. H. and B. V. Barnes. 1980. Forest Ecology. 3rd edition. New York: John Wiley and Sons. Stone, J. R. 19 80. "untitled BASIC program for calculation of soil fractions". Department of Crop and Soil Sciences. Michigan State University. Tourney, J. W. and R. Kienbolz. 1931. Trenched plots under forest canopies. Yale University School of Forestry Bulletin 30. Toumey, J. W. and C. F. Korstian. 1947. Foundations of silviculture upon an ecological basis. 2nd ed. revised. New York: John Wiley and Sons. Threlkeld, J. E. and J. E. Feenstra. 1974. Soil survey of Shiawassee County, Michigan. United States Department of Agriculture. Soil Conservation Service. Washington: United States Department of Agriculture. 159 Trewartha, G. T. 19 81. The earth's problem climates. ed. Madison: University of Wisconsin Press. 2nd Trewartha, G. T. and L. H. Horn. 1980. An introduction to climate. New York: McGraw-Hill Book Co. United States Army Engineer District, Detroit. 1970a. Comprehensive planning of the Grand River Basin, Michigan, vol. Ill, appendix C, climate. Detroit: United States Army Corps of Engineers. ________ . 1970b. Comprehensive planning study of the Grand River Basin, Michigan, vol. IV, appendix E, geology and ground water. Detroit: United States Army Corps of Engineers. Vankat, J. L.1979. The natural vegetation of America. New York: John Wiley and Sons. North Vanlier, K. E., W. W. Wood, and J. 0. Burnett. 1969. Water-supply development and management alternatives for Clinton, Eaton, and Ingham Counties, Michigan. United States Geological Survey Water-Supply Paper 1969. Veatch, J. 0. 1928. Reconstruction of forest cover based on soil maps. Michigan Agricultural Experiment Station Quarterly Bulletin 10:116-126. Veatch, J. 0. 1959. Presettlement forerst of Michigan map. Department of Resource Development. Michigan State University. Veatch, J. 0., H. G. Adams, E. H. Hubbard, C. Dorman, L. R. Jones, J. W. Moon, and L. H. Wonser. 1941. Soil Survey, Ingham County, Michigan. Washington: Michigan Agricultural Experiment Station and United States Department of Agriculture, Bureau of Plant Industry. Ward, J. H . , Jr. 1963. Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association 58:236-244. Cited in CLUSTAN, Computer Centre, University College of London, p.38. London: University College of London. 160 Wayne, W. J. and J. H. Zumbredge. 19 65. Pleistocene geology of Indiana and Michigan. In The quaternary of the United States, eds. H. E. Wright, Jr. and D. G. Frey, pp. 63-83. Princeton: Princeton University Press. Webb, T. Ill and R. A. Bryson. 1972. Late- and postglacial change in the northern midwest, USA: quantitative estimates derived from fossil pollen spectra by multivariate statistical analysis. Quaternary Research 2:70-115. Westveld, R. H. 1949. Applied silviculture in the United States. New York: John Wiley and Sons. Whiteside, E. P., I. F. Schneider, and R. L. Cook. Soils of Michigan. Michigan State University Cooperative Extension Service Bulletin E-630. 1968. Whitford, H. N. 19 07. The genetic development of the forests of northern Michigan: as study in physio­ graphic ecology. Botanical Gazette 31:289-321. Whittaker, R. H. 1974. York: MacMillan. Communities and ecosystems. New ______ . 1978. Dominance-Types. In Classification of plant communities, ed. R. H. Whittaker, pp.67-110. The Hague: Junk. Whittaker, R. H. and H. G. Gausch, Jr. 1973. Evaluation of ordination techniques. In Ordination and classification of plant communities, ed. R. H. Whittaker, pp.287-321. The Hague: Junk. Williams, W. T. 1971. Principles of clustering. Annual Review of Ecology and Systematics 2:303-326. Cited in Multivariate research in vegetation research, 2nd ed., L. Orloci, p.278. The Hague: Junk. Willis, K. 1975. Regression models of migration Geografiska Annaler 57B:42-54. Wood, L. H. 1914. Geography of Michigan. Horton-Beimer Press. Kalamazoo: Wright, H. E., Jr. 1968. History of the prairie peninsula. In The quaternary of Illinois, Special Bulletin 14, ed. R. E. Bergstrom, pp.78-88. Urbana: University of Illinois. 161 Yeates, M. 1974. An introduction to quantitative analysis in human geography. New York: McGraw-Hill Book Co. Zobeck, T. M. 1976. "The characterization and interpreta­ tion of a complex soil landscape in south-central Michigan". M.S. Thesis. Michigan State University. APPENDIX 162 T a b l e Al. Imp or t an c e value program. 110 ftii 120 PRINT "INPUT FILE NUMBER OF FIELD DATA FOR WOODLOT X" 130 INPUT F 140 FIND F 150 PRINT "INPUT THE NUMBER OF OBSERVATIONS FOR WOODLOT X" 160 INPUT N 170 PAGE 180 DIM A(N)»B 190 REM ArBrC ARE THE THREE ORIGINAL WOODLOT VARIABLES 200 REM A-"NUMERICAL SPECIES ID 2.1.0 REM B=BASAL AREA 220 REM O P O I N T TO PLANT DISTANCE 230 FOR 1= 1 TO N 240 INPUT 033 tA I THEN 440 430 C1=C1+1 440 NEXT J 450 PI. (I)=C1/N 460 REM R1(I) IS THE RELATIVE DENSITY FOR SPECIES I 470 NEXT I 4S0 PRINT "THE RELATIVE DENSITIES HAVE BEEN CALCULATED" 490 RETURN 500 REM RELATIVE FREQUENCY SUBROUTINE 51.0 DIM FI. (38) 520 REM FI. (I >"FREQUENCY OF INDIVIDUAL SPECIES I 530 FOR 1=1 TO 38 540 Cl••~0 550 REM Cl IS A COUNTER 560 Y-=l 163 T a b l e Al. ( C o n t 1d . ). 370 Z=4 580 FDR J=i TO N/4 390 REM N:=OBSERVATIONS PER WOODLOT 600 REM N/4-P0INTS SAMPLED PER WOODLDT 6.1.0 FOR K®Y TO Z 620 IF A (K )<>I THEN 650 630 C1=C1+1 640 IF A C K X I THEN 660 630 NEXT K 660 Y~Y+4 670 Z=Z+4 680 NEXT J 690 F I.CI)“C :l./(N / 4 ) 700 NEXT I 710 S=0 720 REM S=TOTAL FREQUENCIES OF ALL SPECIES 730 FOR 1=1 TO 38 740 S=S+F1(I) 750 NEXT I 760 DIM R 2 (38) 770 REM R 2 ( J >^RELATIVE FREQUENCY FOR EACH SPECIES J 780 FOR J=1 TO 38 790 R2(J)=F1(J)/S 800 NEXT J 810 PRINT -THE RELATIVE FREQUENCIES HAVE BEEN CALCULATED" 820 RETURN 830 REM RELATIVE COVERAGE SUBROUTINE 840 REM THIS PORTION COMPUTES SUM OF BASAL AREAS 850 DIM EC 38) 860 REM E-SUM OF BASAL AREAS FOR ALL SPECIES 870 FOR 1=1 TO 38 880 B2=0 890 REM B2 IS A COUNTER FOR SUM OF BASAL AREAS 900 FOR J=1 TO N 910 IF A (J >< > I THEN 930 920 B2=B2+B(J> 930 NEXT J 940 E C D - B 2 950 NEXT I 960 REM THIS PORTION COMPUTES TOTAL DENSITY 970 C.L-0 980 REM Cl-18 A COUNTER FOR SUM OF POINT/PLANT DISTANCES 990 FOR I 1 TO N 1000 C1=C1+C(I) 1010 NEXT I 1020 D-C 1/N 1030 REM D-MEAN POINT TO PLANT DISTANCE FOR ENTIRE WOODLOT 1040 D1“D"2 /10 *76 164 Tab le Al. ( C o n t 1d . ). 1050 REM .01“MEAN AREA IN SQUARE METERS PER TREE 1060 REM 10»76!=C0NVERSI0N FROM FEET TO METERS 1070 T--J.00/D1 1080 REM T=NUMBER OF INDIVIDUALS PER 100 SQUARE METERS 1090 REM COMPUTES ABSOLUTE DENSITY FOR EACH SPECIES 1100 DIM D3(38> 1110 REM D3«ABSOLUTE DENSITY FOR INDIVIDUAL SPECIES 1120 FOR 1=1 TO 38 1130 D3=R1(I)#T 1140 NEXT I 11S0 REM THIS PORTION COMPUTES COVERAGE 1160 DIM C 3 (38 ) 1170 REM C 3 (I )"COVERAGE FOR INDIVIDUAL SPECIES I .1180 FOR 1 = 1 TO 38 1190 C1=0 1200 REM Cl IS COUNTER FOR TOTAL OF INDIVIDUALS PER SPECIES 1210 FOR J=1 TO N 1220 IF A I THEN 1240 .1230 C1--C1 + 1 1240 NEXT J 1250 IF C1>0 THEN 1280 1260 C 3 (I )=0 127.0. GO TO 1290 1280 C 3 (I )= E (I )* D 3 (I )/Cl 1290 NEXT I 1300 REM THIS PORTION COMPUTES RELATIVE COVERAGE 1310 DIM R 3 (38) 1320 REM R3-RELATIVE COVERAGE FOR INDIVIDUAL SPECIES 1330 X1=0 1340 FOR J=1 TO 38 1350 X1=X1+C3(J> 1360 REM XI WILL BECOME THE TOTAL COVERAGE FOR ALL SPECIES .1370 NEXT J 1380 FOR K ~ .1. TO 38 1390 R3(K)=C3 1450 FOR 1=1 TO 38 1460 R4 (I >• “ R 1 (I > fR2 (I )+R3 (I ) 1470 NEXT I 1480 PRINT "THE IMPORTANCE VALUES HAVE BEEN CALCULATED" 1490 RETURN 1500 REM SUBROUTINE TO COMPUTE IMPORTANCE PERCENTAGE 1510 DIM R5<38> 1520 FOR 1=1 TO 38 165 T ab le Al. ( C o n t ' d . ). 1530 R 5 (I >*< R 1 ( I) + R 2 (I )+ R 3 (I ))/3 .1540 NEXT I 1550 PRINT -THE IMPORTANCE PERCENTAGES HAVE BEEN CALCUIATFD* 1560 RETURN 1570 REM SUBROUTINE TO PRINT CALCULATED VALUES 1580 REM ID=NUMERICAL SPECIES I D ♦ REFER TO MASTER LIST 1590 REM RD-RELATIVE DENSITY 1600 REM RF=RELATIVE FREQUENCY 1610 REM RC-RELATIVE COVERAGE 1620 REM IV=IMPORTANCE VALUE 1630 REM IP" IMP0RTANCE PERCENTAGE 1640 DIM R1 (38) yR2 (38)yR 3 (38) yR 4 (38) yR (58) 1650 FOR 1-1 TO 38 1660 PRINT USING 1670 i "I D " y "RD " y "R F " y "R C " y "IV “ y "IP" 1670 IMAGE / r 2A y1IX y2A y12Xy2A y 12Xy 2A y12X y2A y12X y2A 1680 PRINT USING 1690♦ Iy R 1 (I )yR 2 (I )yR 3 (I )yR 4 (I )yR 5 (I ) 1690 IMAGE 2D y5X y5D *3D y5Xy 5 D .3D y5Xy 5D *3D y5X y5D *3D y5X y5D *3D 1700 NEXT I 1710 RETURN 1720 REM SUBROUTINE TO RECORD DATA ON PRINTER 1730 PRINT "WHAT WOODLOT DATA IS NOW BEING PRINTED'?" 1740 PRINT "INPUT THE WOODLOT NUM BER* “ 1750 INPUT 27 1760 PRINT @ 3 7 y26 J1 1770 PRINT @41♦ USING 178 0J"WOODLOT NUMBER"yZ7 1780 IMAGE 10 /y1OX y14A y1X y2D y2/ 1790 PRINT @41» USING 1800J"I D "y"R D "y"R F "y "R C "y"I V "y"I P " 1800 IMAGE SX y2A y11X y2A y12X y2A y12X y2A y12X y2A y12X y2A y2/ 1310 FOR 1=1 TO 38 1820 P R I N T 04i: 1830 1840 1850 1860 1870 1830 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 USING 1830 1 1 »R 1 yR3(3 8 ) yR4(38)yR5(38) FOR 1=1 TO 38 PRINT @3 3i R l (I )yR 2 (I )yR3(I)yR 4 (I ) *R5(I) NEXT I CLOSE PRINT "THE CALCULATIONS FOR WOODLOT X ARE NOW COMP I FTP <•" PRINT " TYPE RUN TO CONTINUE*" RETURN 166 T ab le A2. 100 110 .120 130 140 150 160 170 130 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400 410 420 430 440 450 460 470 480 490 500 510 520 530 540 550 56 0 570 Soil f ractions program. INIT PAGE DIM T (8) DIM R K 8 ) DIM D<56) DIM M l (20) DIM 87(8) DIM P5<8> READ D DATA 16.31 , 16* 11 >15.98* 15,79-- 15.66,15.46,15.33,15* 14 DATA 15.01,14.89 DATA 14.7,14.51,14.39,14,2,14.01,13.83,13.65,13.52 DATA 13.34,13.16 DATA 12.99,12.87,12.69,12,52,12,34,12.17,12,11.83 DATA 11.66,11.49 DATA 11.33,11.16,11,10.83,10.67,10.51,10.35,10.2 DATA 10.04,9.88 DATA 9.73,9.6,9.5,9.33,9*18,9.02,8.88,S.72,8.53,8,4 DATA 8.2,8.1,7.9,7.78,7.64,7.5 READ 01 DATA 0.01139,0,01109,0.01081,0.01053,0.01027,0.01002 DATA 0.00978,0.00955 DATA 0.00933,0,00911,0.00891,0.00871,0.00851,0.00833 DATA 0.00815,0.00798 DATA 0.00781,0.00765,0.00749,0.00734 PRINT "INPUT THE SOIL SAMPLE ID," INPUT A1> PRINT "INPUT AVG, TEMP. OF THE SUSPENSION, DEGREES C," INPUT Cl PRINT "INPUT A U G » CONTROL HYDROMETER READING," INPUT R2 PRINT "INPUT OVEN DRY W G T . OF SAMPLE (MINUS ORGANIC)," INPUT 0 PRINT "INPUT WGT, OVEN-DRY SAND FRACTION (FROM SIEV E) ,” INPUT Z REM PI IS THE DENSITY OF THE "CALGQN" SOLUTION AT 50G/L PI=0.99949 REM FOLLOWING DATA ARE HYDROMETER READING TIMES (MINUTES) READ T DATA 0,5,1,3,10,30,90,270,480 PRINT "INPUT THE HYDROMETER READINGS OF THE SAMPLE," INPUT R1 W:“C1-~14 V2»V1(W> K::::l PAGE P RIN T U SIN G 6 0 0 i "C 0 M P U T E D D A T A F 0 R SAM P I...E t " ,A $ PRINT 037,26:1 167 T ab le A2. 580 590 600 610 620 630 640 650 660 670 680 690 700 710 720 730 740 750 760 770 780 790 800 810 820 830 840 850 860 870 880 890 900 910 920 930 940 950 960 970 980 ( C o n t ' d . ). PRINT C»41J USING 6 1 0 : "COMPUTED DATA FDR SAMPLE ? " »A4> PRINT @37*26 JO IMAGE SX r25A *1X r3A r 2/ IMAGE 5 / s-5Xv 25A *IX *3A *2/ REM COMPUTES PARTICLE SIZES AND CUMULATIVE PERCENTS FOR 1=1 TO 8 T1=T(K) R3=R1(K> K«K + 1 Y=R3+i H=D(Y> F‘2 “2 ♦65—PI G=30*V2*H/(980.7 * P 2 ) G»G"0*5 G=G#1000 C2=R3-R2 F 5 «I >=100*(C 2 / 0 > T2~T 1 0 *5 S=G/T2 S7 (I ) -Sftl ♦OE-3 PRINT USING 820 J “PARTICLE SIZE = * , S7 (I ) * ■CUM * % =%F'5(I) PRINT 8 3 7 r 2 6 SI PRINT 041 I USING 8 3 0 ? "PAR* SIZE = " * S7 ( I ) r " CUM ♦ "/. ="*P5(T) PRINT 0 3 7 1-26 JO IMAGE 5X*15A*1X*4D*3D*5X*20A*1X*4D.2D*/ IMAGE 5X y15A y1X y4D ♦3D y 5X y14Ay 1X y4D ♦2D y / NEXT I Z7=P5<2>-P5C8) Z3=100-P5(2) REM LINE 672 COMPUTES PERCENT SAND IN SAMPLE Zl=100#(Z/0> PRINT USING 960 J "PERCENT SAND ff numbers in parentheses indicate more than one recorded individual. 171 TableA4. Thickness of overlying fine-textured sediment. Sugar Maple-Beech Woods Woodlot No. Auaer Am. ( f t . ) 1 4.0 (13.0) 1.3 ( 4.1) 2.6 ( 8.6) 2 2.1 ( 7.0) 0.4 ( 1.2) [2.6 ( 8 .5 ) ] a 1.7 ( 5.6) 3 4.0 (13.0) 3.6 (11. 9 ) b 3.8 (12.5) 4 4.0 (13.0) 4.0 (13.0) 4.0 (13.0) 5 1.1 ( 3.6) 2.6 ( 8.5) 1.9 ( 6.1) 6 2.9 ( 9 .5 )c 3.6 (11.8)b 3.3 (10.7) 7 1.0 ( 3.4) .9 ( 2.9) 1.0 ( 3.2) 8 1.3 ( 4.1) 1.3 ( 4.3) 1.3 ( 4.2) 9 3.3 (10.7) 3.3 (10.8) 3.3 (10.8) 10 0.0 ( 0.0) 0.7 ( 2.3) 0.4 ( 1.2) 11 2.8 ( 9.3) 3.0 (10.0) 3.0 ( 9.7) 12 3.2 (10 .4)C 2.7 ( 9. 0 )C 3.0 ( 9.7) 13 2.2 ( 7.3) 1.6 ( 5.4) 2.0 ( 6.4) 14 3.1 (10.2) -- 3.1 (10.2) 15 3.3 (1 0 .8 )b 3.8 (12.6)b 3.6 (11.7) 16 3.6 (1 1 .8 )b 4.0 (13.0) 3.8 (12.4) 17 2.7 ( 9.0) 4.0 (13.0) 3.4 (11.0) 18 2.7 ( 9 .0 )C 2.4 ( 8.0) 2.6 ( 8.5) 19 2.4 ( 8.0) 0.3 ( 1.0 )d 1.4 ( 4.5) 20 4.0 (13.0) 1.8 ( 5.9) 2.9 ( 9.5) 21 4.0 (13.0) 4.0 (13.0) 4.0 (13.0) 22 4.0 (13 .0 )e Auger Bm.(f t.) — Mean Thickness m.( 4.0 (13.0) 172 Table A4. (Cont'd.). Sugar Maple-Beech Moods Woodlot No. Auger Am. ( f t . ) Auger B m.(ft.) Mean Thickness m. ( f t . ) 23 3.1 (10.1) —f 3.1 (10.1) 24 1.6 (5 .4 )d 0.9 ( 3 .1 )d 1.3 ( 4.3) mean = 2.7 (8.7) Oak-Hickory Moods Woodlot No. Auger A m .( f t .) 25 3.4 1 1.0)D 26 2.0 27 Auger B m . ( f t . ) Mean Thickness m.(f t . ) 1.1 ( 3.7) 2.3 7.4) 6.7) 2.4 ( 8 .0 )b 2.3 7 4) 2.4 8.0) 1.9 ( 6.3) 2.2 7.2) 28 2.7 9.0) 2.7 9.0) 29 0.5 1.6) 0.4 ( 1.2) 0.4 1.4) 30 2.0 6.5) 4.0 (13.0) 3.0 9.8) 31 1.9 6.3) 2.7 ( 9.0) 2.3 7.7) 32 2.3 7.5) 2.4 ( 7.8) 2.3 7.7) 33 4.0 13.0) 4.0 13.0) 34 1.5 5.0) 2.0 6.6) 35 3.7 12.2) 3.7 12.2) 36 1.5 5.0) 1.6 5.4) 37 2.9 9.6) ---- 2.9 9.6) 38 1.9 6.3) ---- 1.9 6.3) 39 1.1 3.5) 0.9 3.0) — ---- 2.5 ( 8.2) — 1.8 ( 5.8) 0.8 ( 2.5) 173 Table A4. (Cont'd.). Oak-Hickory Woods Woodlot No. Auger A m . ( f t . ) . Auger B m . ( f t . ) Mean Thickness m . ( f t . ) 40 1.0 (3.4) 1.3 (4.3) 1.2 (3.9) 41 0.6 (2.1) 0.9 (3.1) 0.8 (2.6) 42 1.1 (3 .5) 1.0 (3.2) 1.0 (3.4) 43 0.4 (1.2) 0.8 (2.5) 0.6 (1.9) 44 2.9 ( 9 .5 )c 45 2.7 (9.0) 46 1.6 (5.3) 1.6 (5.3) 1.7 (5.5) 47 2.7 (9 .0 )b 1.4 (4.5) 2.1 (6.8) 48 0.9 (3.1) 1.0 (3.3) 1.0 (3.2) 2.9 (9.5) — 2.7 (9.0) mean = 2.0 (6.6 a Baker Woodlot, 3 holes and h it rock at 8.5 onthird hole b h it rock, s t i l l in t i l l at that depth c overcome by water suction d h it "water sand" at that depth e h it rock at 9.0, recorded at 13.0 based on talk with owner and his house foundation depth f owner allowed only one auger hole Table A5. T-test results. Pooled Variance Estimate N Mean F Value 24 24 .58 1.35 24 24 .37 .37 24 24 .05 24 24 SM-B 0-H % sand-A horizon SM-B 0-H % silt-A horizon SM-B 0-H %clay-A horizon SM-B 0-H % sand-B2 horizon SM-B 0-H 2-Tailed probability T Value DF 2-Tailed probability .473 - .53 46 .602 1.50 .338 .23 46 .823 1.22 .633 .58 46 .566 .60 .60 1.69 .218 - .05 46 .958 24 .27 1.62 .256 - .76 46 .453 24 .29 .59 .04 %si 1t-B2 horizon Table A5. (Cont'd.). Pooled Variance Estimate % cla.y-B2 horizon SM-B 0-H N Mean F Value 24 .13 1.11 24 .11 8 .7 (2 .65m) 1.24 6.6(2.Om) 2-Tailed probability T Value DF 2-Tailed probability .798 1.08 46 .616 2.20 46 (1-tailed = .017) .285 overburden thickness SM-B 0-H 24 24 2-•tailed V Hr where U1 = U2 u1 t u2 = sm-b mean U2 = o-h mean 1-tailed V U1 > g2 Hj: Uj < u2 .033 176 Table A6. Explanation o f variance in edaphic data matrix. Component Eigenvalue Percent of Variance 1 4.5 64.5 64.5 2 1.2 17.3 81.8 Cumulative Percent Table A7. Edaphic component loadings. Component 1 Component 2 A Horizon percent sand -.86 -.45 percent s i l t .92 — percent clay — .88 percent sand -.8 0 -.56 percent s i l t •91 — B Horizon percent clay Overburden Thickness - - .72 .84 177 Table A8. Woodlot No. Component scores. 1 2 3 "Mesic/Xerfc'1 "Disturbance" "Dry-Mesic" 1 .6068 -.0548 -.6440 2 .2846 -.4698 .2424 3 1.6495 -.0676 -.3658 4 -.8280 -.1857 -.6927 5 .7548 -.4582 -.1134 6 .8309 -.3337 -.5800 7 .5653 -.5521 -.2312 8 .5502 -.4439 -.7158 9 2.2036 .1285 -.5571 10 .1758 -.4977 -.4075 11 1.2387 -.4041 .1884 12 -.0320 -.2068 -.3415 13 .9044 -.0291 -.2446 14 1.0069 -.5418 .4218 15 1.6437 1.3467 -1.0057 16 2.4323 -.0078 .4900 17 -.2846 -.7039 -.3540 18 .4167 .0601 -.3177 19 .7605 -.0645 -.3199 20 -.1397 -.4952 -.4108 21 .9039 -.5383 .2347 22 .7754 -.4732 -.3723 23 1.2829 -.4419 -.3342 178 Table A8. Woodlot No. (Cont'd.). 1 "Mesic/Xeric" 2 "Disturbance" 3 "Dry-Mesic" 24 .4079 -.0484 -.5033 25 -1.3107 -.1083 -1.4457 26 -.4392 .6721 .6216 27 -.5693 .8087 -.0947 28 -.1281 .3466 2.7028 29 -1.2438 .2900 -.7176 30 -1.0806 -.4188 -1.2593 31 -.6926 - , 099*2 .3533 32 -.7282 .6106 -.3968 33 -.7373 2.6185 1.7603 34 .1931 4.7466 -.7773 35 -.4267 -.4233 -.8959 36 .4017 -.7856 -1.8155 37 -.1407 -.6130 -1.6912 38 -.8703 1.9540 -.1954 39 -.7964 -.1470 -.2742 40 -.8612 .9207 .7089 41 -1.6335 -.1526 -1.1840 42 -1.2661 -.6902 -1.0152 43 -1.3218 -.6044 -1.0455 44 -1.5220 -1.1028 -.4662 45 -.5460 -.1708 2.7001 46 -1.0877 -.0259 1.8930 47 -.1104 -.6589 1.8367 48 -1.1850 -1.4851 .6184 179 Table A9. Woodlot township/range designation. Woodlot 1 WJs, SEh, SWi, Sec 30, T4N, R1W 2 Nh, NE%, SW%, Sec 19, T4N, R1W 3 lih> NWU> SE%, Sec 36, T4N, R1W and N%, NE%, SVfc, Sec 36, T4N, R1W 4 NW%, NE^, SE%, Sec 30, T4N, R1W 5 NJs, NW%, SEJj, Sec 3, T3N, R1E 6 NWfc, SEh, SW%, Sec 3, T4N, R2E 7 HEh, HEh, Nlfis, Sec 4, T3N, R2E 8 HE%, SW%, NWls, Sec 31, T4N, R1W 9 Eh, SEh, NE*s, Sec 3, T3N, R1W 10 NEh, NWs, SWfc, Sec 31, T4N, R1W 11 UEk, NW%, NE%, Sec 14, T4N, R2E 12 NEJ$, SE%, NE%, Sec 29, T4N, R2E 13 SEh, SWfc, SEh, Sec 29, T4N, R2E 14 NEJj, NVfis, SWfc, Sec 21, T4N, R2E 15 NE%, SE%, SV&, Sec 6, T3N, R1W 16 NVfi$, SE%, SE%, Sec 18, T4N, R2E 17 Hh, SWls, NW^, Sec 4, T3N, R1W 18 NMs, NEJ$, NE%, Sec 11, T4N, R2E 19 NW%, SWU, SW^, Sec 9, T3N, R1W 20 N*s, NW*s, NWJj, Sec 15, t3N, R1W 21 NVAs, SEJj, NVfis, Sec 24, T3N, R1E 22 SEJs, SE%, Nlfij, Sec 34, T4N, R2E 23 NW%, NW**, SWls, Sec 15, T3N, R1E and NEfc, NE%, SV&, Sec 16, T3N.R1E 0' 180 Table A9. (Cont1d . ). Woodlot 24 \H, NE%, SW1«, Sec 9, T4N, R1E 25 NW%, NW%, SW%, Sec 4, T4N, R2E 26 NW^, SW%, SW%, Sec 12, T4N, R1E 27 NW*s, SEH, SEH, Sec 3, T4N, R1E 28 HEH, NW*s, HEH, Sec 3, T4N, R1E 29 Sh, HEH, NW^, Sec 1, T4N, R1W 30 NE%, NW%, NWls, Sec 8, T4N, R2E and SE%, SWH, SW^, Sec 5, T4N, R2E 31 NVfc», NE%, NW%, Sec 2, T4N, R1E 32 NW%, HEh, SVfis, Sec 11, T4N, R1E 33 NE*j, SWls, SW%, Sec 19, T4N, R2E 34 NEJj, NW%, NW%, Sec 5, T4N, R1E 35 Slfe, Slfis, SWfc, Sec 23, T4N, R1W 36 WJs, SEH, NE%, Sec 16, T4N, R1E 37 NW%, SW>s, SW»s, Sec 3, T4N, R1E 38 Eh, HEH, SE%, Sec 33, T5N, R1E 39 HEH, NE%, SWH, Sec 23, T5N, R1W 40 HEH, NWfc, SWfc, Sec 24, T5N, R1W 41 NE*s, NW%, SVfii, Sec 30, T5N, R1W 42 NEJs, HEH, Slfe, Sec 23, T5N, R1E 43 NW%, SW»s, NE*s, Sec 29, T5N, R1W 44 SW»s, SVAs, HEH, Sec 20, T5N, R1W 45 Sh, NW%, NE%, Sec 21, T5N, R1W and NJs, SWH, HEH, Sec 21, T5N, R1W 181 Table A9. (Cont'd.). Woodlot 46 NE%, SVfij, NWSa, Sec 22, T5N, R1W 47 NE%, NW»s, Sec 24, T5N, R1W 48 NE%, SE%, Sec 21, T5N, R1W