<<

An Upland Continuum in the Prairie-Forest Border Region of Author(s): J. T. Curtis and R. P. McIntosh Source: Ecology, Vol. 32, No. 3, (Jul., 1951), pp. 476-496 Published by: Ecological Society of America Stable URL: http://www.jstor.org/stable/1931725 Accessed: 01/07/2008 17:00

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=esa.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

http://www.jstor.org AN UPLAND FOREST CONTINUUM IN THE PRAIRIE- FOREST BORDER REGION OF WISCONSIN1

J. T. CURTIS University of Wisconsin, Madison and

R. P. MCINTOSH Middleburv College, Middlebury, Vermont

INTRODUCTION generously placed his files on forest corn- It has been said that the desirable position at our disposal; to Professors order of ecological research in a given Grant Cottam, H. C. Greene, N. C. Fas- region is first, the study of communities, sett, R. S. Muckenhirn, M. L. Partch, second, the study of the individual spe- and S. A. Wilde for valued advice and cies, and last, the study of the habitat consultation; to Messrs. Orlin Anderson, (Yapp 1922). Sociological investiga- and R. S. Brown for indispensable aid tions of constituent species and their in the collection of data; to Miss numerical relations should precede aute- Margaret Gilbert for aid in the computa- cological studies if the latter are to have tion; to Messrs. A. M. Fuller, Neil maximum meaning. This sequence Harrington and Fred Wilson for help in many should lead to a greater efficiency in re- the location of stands; and to the Serv- search, in that an initial knowledge of any members of the Conservation natural groupings of species may enable ice, the Regional Foresters of the Wis- the autecology of the groups to be studied consin Conservation Department and simultaneously, with a resultant saving in the advisers of the Junior Wisconsin time and travel expenses. The grossly Academy of Sciences, who so willingly condi- inadequate state of autecological knowl- imparted their knowledge of local edge of even our most common species tions in their regions. The authors are indicates that the current haphazard also indebted to Professor K. J. Arnold, method of attack is faulty; if aggrega- of the University Computing Service, for tions of species occur together under aid in the statistical phases of the work. in similar environmental conditions in na- The taxonomic nomenclature used Manual ture, they could be studied in a systematic this paper follows that of Gray's program. A long-term investigation of of Botany, 8th edition, 1950. the communities of Wisconsin has been LITERATURE REVIEW underway since 1946 with a view to de- to the vegeta- termining the existence and the floristic The earliest references were nature of such aggregations (Curtis and tion of southwestern Wisconsin the explorers and pioneer trav- Greene 1949). The present paper deals made by 1698. For with the upland hardwood of the elers in the years following border of the state the most part, they contained only gen- p)rairie-forest region remarks about un- and their interrelations with each other eral descriptions, with first account a and with certain physical factors of the usual species. The by was that of Increase environment. Deep appreciation is ex- competent botanist in l)ut he dealt pri- presse(l to Professor P. B. Whitford, hlio A. Lapham 1852, marily with the distribution of individual 1 This work wvas supported in part by the species. The first description of plant Research Committee of the radiatete School of communities as such was given by T. C. the University of Wisconsin from funds sup- plied by the Wisconsin Alumrni Research Chamiberlainin 1877. This eminent geol- Foundation. ogist, whlo was later to exert such great 476 July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 477 influence on H. C. Cowles at the Univer- and the changes which have taken place sity of , was at that time the State in them since settlement. He attributed Geologist of Wisconsin. He distin- major importance to fire in the mainte- guished three forest communities in nance of openings and to cessation southern Wisconsin, but made clear that of fire as the cause of the subsequent de- no abrupt line of demarcation existed be- velopment of the present closed canopied tween them, with repeated statements stands. illustrating "a gradual transition from The most comprehensive paper on the one to the other" and "an almost imper- forests of southern Wisconsin was by ceptible merging of one into the other." Whitford (1951). Although this study His three communities were the "oak was primarily concerned with the relation group" including Quercus alba (white between the clone size of herbs and the oak), Q. mctcrocarpa (bur oak), Q. rubra age of the forest stands, considerable at- (red oak), and Q. velutina (black oak), tention was devoted to the establishment with Populus tremuloides (aspen), Carya of a method for classifying the stands. ovata (shagbark hickory), Prunus sero- In the course of the investigation, Whit- tina (black cherry) and Pyrus ioensis ford studied a considerable number of (wild crab apple) as lesser members; stands and presented detailed data on the "oak and maple group" in which Acer , shrub, and herb composition. He saccharum (sugar maple) was most con- arranged his series of 28 stands into spicuous but Quercus alba and Q. rubra three groups on the basis of shade toler- were still common, with Acer rubrum ance of the tree species. He assigned (red maple), Ulmus rubra (slippery each species to one of three tolerance elm), and Tilia americana (basswood) in classes, and then determined the relative small numbers; and the "maple group" fraction of all in a stand which be- with A. saccharum dominant, but with longed to each of these classes, on a sum- Tilia amnericanca,Ulmus rubra, and Ostrya mation basis. This method resulted in virginiana (ironwood) as characteristic a sequence of stands from those com- and abundant associates. He comments posed entirely of intolerant species to about the "oak and maple group" as fol- those made up mostly of tolerant species. lows: "This group is not characterized by Whitford divided his stands into three the exclusive presence of any prominent groups-"oak-hickory, intermediate, and plant but by a distinctive association of maple-basswood," but the division was plants common to several groups." made for pragmatic reasons and was not Bruncken (1900) also described three based on any internally apparent separa- forest associations, but only two were tion of the data into three groups. found on uplands the "hemi-xerophytic Studies in , and Minme- oak-hickory" and the mesophyticc bass- sota indicate that a series of forests simi- wood-maple." Additional observational lar to those in Wisconsin are found in at studies on the forests of the region were least portions of these states. No quan- reported by Pammel (1904), Marks titative investigations in northern Illinois (1942) and Stout (1944), but the first have come to our attention, but the gen- account in which quantitative data on eral descriptions of Fuller and Stras- actual composition were presented was baugh (1919), and DeForest (1921) that of Wagner (1947), who studied a indicate a series of forest types from stand in Jefferson County which fits into those dominated by Quercus and Carya Chamberlain's oak and maple group. to those dominated by Acer saccharum. Cottam (1949) investigated an oak In Iowa, the studies of Clark (1926) and woods in Dane County by quantitative of Aikman and Smelser (1938) show a methods and reported in detail on the succession from and historical development of oak openings Q. velutina in the initial stages through 478 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3

Q ucrcus rlibra-Tilia am1tericanastands to Carya ova/a, and Jitglans nigra reach the terminal forests of Acer saccharun their northern boundaries in a zone which and Tilia am erican-a. Daubenmire (1936) also includes the southern limits of such reported on the structure of the Acer- trees as Betfila lutea, A bies balsarnea, Tilia. community and on its relation to and Populls balsamnifera. This same the surrounding Qutercus forests in the zone marks the northward or southward "Big Woods" of . extent of a great many herbaceous spe- While the results of the various inves- cies, both native and exotic; it also sepa- tigations in these four states are by no rates closely related varieties of single means in complete agreement, there does species, as in the case of Acer saccharunm appear to be a general uniformity of saccharum and its subspecies A. s. nigrum opinion as to the topographic distribu- (Desmarais 1948) and Braclyelytrum tion of the forests, with some type of erectau erectur and B. e. septentrionale Quercus forest on the most xeric forested (Salamun 1950). The floristic province sites and an Acer forest in the most mesic southwest of the zone is here termed the habitats. The disagreement seems to lie prairie-forest province, while that to the in the determination of the number of northeast is the northern hardwoods discrete communities that may be recog- province (roughly equivalent to the Lake nized, in the floristic nature of the bound- Forest of Weaver and Clements 1938). aries between them, and in their exact It is recognized that these names may be composition. Many of the early workers of local descriptive value only and are emphasized the lack of sharp lines of considered to be tentative. Perhaps the demarcation, while later workers, who most characteristic feature of the floristic could study only the more completely provinces from the standpoint of vegeta- dissected and isolated stands resulting tion is the occurrence of oak woods in the from the highly agricultural land use of early forest succession south of the zone the present time, were more inclined to in contrast to pine woods in the same describe discrete communities of several successional position north of the zone. types. The terminal forests in both provinces Although the of Wisconsin has are dominated by sugar maple. Within many diverse floral elements, it was early the narrow zone itself various elements of pointed out that there are two major the two provinces occur in puzzling mix- floristic provinces, separated by a di- tures. Their elucidation must await com- agonal line from northwest to southeast pletion of studies on both sides of the (Cheney 1894). The southwestern por- zone. tion of the state contains the prairies and These floristic provinces are similar to hardwood forests, while the northeastern the natural areas of Cain (1947) and part contains both hardwood and conif- the floristic areas of Raup (1947) and erous forests. The boundary line be- Egler ( 1948). They are characterized tween them, actually a zone or band, is by the presence of a uniform flora such comparable to the tension zone proposed that "all stands have equal chances of by Griggs (1914). Its position can be containing all the same species" (Cain determined within fairly narrow limits 1947). It is only within such limitations by the method proposed by Clements that plant associations can possibly (1905: 187) "The limiting line or eco- achieve objective reality. One of the tone of a . . . province is a composite ob- aims of the present paper is to determine tained from the limits of principal species whether discrete communities with defi- and checked by the limits of species typi- nite structure and definable boundaries cal of the contiguous ." actually exist within such a natural floris- Applying this to Wisconsin, a number tic province, or whether plant associa- of trees including Quercus velutina, tions are "metaphysical approximations July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 479 in a field where there are unlimited vari- drift. The underlying bedrock is com- ables, combinations and permutations" posed predominantly of old sedimentary (Cain 1947) even within the narrow rocks-Cambrian sandstones and Lower limits set by the province. Magnesian, Ordovician and Silurian limestones. DESCRIPTION OF AREA Physiography Climate under consideration lies al- The south and west borders of the area The area in the transition zone be- included in this study (the Wisconsin- most wholly regions I and IV of Illinois boundary and the tween climatic such, its climate is River) were prescribed by various non- Borchert (1950). As intermediate between that of the North- ecological but nevertheless compelling of snowy winters reasons. The remaining boundary was eastern forest region summer and that of the based on the biologically more valid pres- and reliable with relatively dry ence of the tension zone discussed above central Prairie region summer . which extends across Wisconsin from winters and frequent forest and northwest to southeast (Fig. 1). The Partaking of both both prairie and major physiographic differentiation of climates, it supports intermediate expression, the region is its separation into glaciated forest and their . Excellent summaries and non-glaciated areas. The Driftless the oak climatic Area in the western portion of the study of the ecologically significant the general region are area is a region of mature pat- variables for Borchert (1950). terns and rugged, almost mountainous, mapped by topography as contrasted with the glaci- Soil ated eastern portion which possesses im- The forest of the area belong to mature drainage and a more or less roll- podzoic soil group and ing topography. The glaciated areas the grey-brown are largely of the series consisted primarily of Wisconsin drift Fayette-Dubuque in the and the Miami- with lesser areas of older (Illinoian) Bellefontaine series in the eastern glaci- ated area. The in both series are vi 0

50 Miles usually composed of less or a mixture of less and residual soil. Profile charac- teristics are similar throughout the study area. In genera, there is a well-marked dark Al layer varying from one-half inch to as many as ten inches in depth, although one to four inches is more typi- cal. The A2 is not well defined in terms of color but may be detected by textural differences; it is usually one-half to two inches in depth. The brownish B layer generally shows a well-defined nut structure. General The southwestern half of Wisconsin lies in the broad ecotonal belt between grassland and summergreen forest. In FIG. 1. Map of Wisconsinshowing location of stands studied. The cross near the center of presettlement times its vegetation con- the map indicates 90? W. Long., 45? N. Lat. sisted of scattered prairies, oak openings 480 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3 of savanna-like aspect, and broadleaf our conviction that variation in floristic forests of varying density and composition is one of the most important composition. The present-day stands of characteristics that may be determined natural vegetation are mostly derived in the study of any vegetation, and this from these three communities, although may be understood only after a great other vegetation types are also present many examples of the type have been in the form of relics (McIntosh 1950), analyzed. It is important that the initial The deciduous forests occur on a range ideas of the type limits be sufficiently of topographic sites, from very dry broad to preclude the dangers of sub- through mesic to very wet. The driest jective selection of only those stands sites, on cliffs, sandy soils and south- which fit a preconceived notion as to what west slopes are characteristically covered a particular community should be. The with scattered stands of small only criteria used for the selection of (Quercus velutina, Q. inacrocarpa) or stands in the present study were: (1) with red cedar () that they be natural forests (i.e., not glades. As the moisture conditions im- artificially planted) of adequate size, fif- prove, the forests become more dense, teen acres being the minimum, with forty with a greater number of important tree acres or more preferred; (2) that they species. The most mesic conditions com- be free from disturbances in the forms monly support sugar maple, basswood or of fire, grazing or excessive cutting; (3) other species of similar requirements. that they be on upland land forms on Wet soil forests are largely restricted to which run-off waters never accumulate. the flood of streams and rivers but If a stand fulfilled these conditions, it are found to a limited extent also on the was studied, regardless of what species shores of glacial lakes in the eastern por- of trees might be present. It might be tion of the area. Their most character- expected that homogeneity would have istic trees are Ulmus americana, Acer been a required characteristic, but, as will saccharinum, Fraxinus pensylvanica lan- be shown below, it was found that this ceolata and Quercus bicolor. could be determined statistically from Most of the original prairie has been the data at a later time. If a stand was placed under cultivation but small rem- then found not to be homogeneous, its nants still occur along railroads and high- data were eliminated from the compila- ways and occasionally elsewhere. Prairies tions. of several sorts occurred on the same A few of the stands were known to range of topographic types as the de- us before the study was begun, but the ciduous forests (Curtis and Greene majority were found in response to a cir- 1949). Prairies formed a part of the cular letter sent out to state and federal presettlement vegetation of all counties forestry agencies, county agents, Soil included in the present study, although Conservation Service personnel,. and the area covered varied from less than others whose work brings them into con- one per cent in Richland County to over tact with . All of the stands 50 per cent in LaFayette County. It is discovered in this way were checked in estimated that 20 per cent of the total the field, and, if they met the criteria of area was in prairie when the first settle- selection, they were incorporated in the ment by Europeans began in the 1830's. study. In this way, an essentially ran- dom selection of 95 stands in 29 counties METHODS was obtained without prejudice as to The plan of the investigation used in kind. The geographical distribution of this work is an adaptation of the mass the samples (Fig. 1) was believed ade- collection method used by taxonomists quate to give an accurate picture of the (Anderson 1941 ; Fassett 1941). It is forests of the area. July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 481

Field methods We knew of no way by which the Each stand was sampled by use of the homogeneity of stands could be deter- random pairs method (Cottam and Cur- mined objectively in the field, so it be- tis 1949). Forty pairs of trees were used came necessary to devise a method of for the determination of frequency, den- checking this after the data were col- sity and dominance. Presence was re- lected. Since 40 separate pairs of trees corded for all vascular species on spe- are measured and recorded in sequence cially prepared forms which contained the in the field, it is possible to segregate names of the 200 plants most commonly the data from successive groups of 10 found in upland forests of the region. pairs, each representing a different areal Frequency of herbs, shrubs and tree seed- portion of the stand. Using these four lings was obtained by 20 quadrats, each groups as separate samples, the Chi- square test for homogeneity (Snedecor ? milacre in size. Density was not de- termined for these forms. 1946) may be used to determine whether Saplings over 1" DBH were counted the actual occurrences of the major trees in a transect between the two trees of in any group deviate significantly from each pair. Since the distance between the occurrences expected on the basis of the trees was variable, the transects re- uniform distribution. The requirement sulted in different total areas in each of an expected value of at least five indi- stand and thus prevented the use of fre- viduals in each sample makes this test quency as a measure. Density, therefore, applicable only to those tree species is the only type of information available represented by at least 20 trees in 40 for the saplings. It appears desirable to points. These species are the ones which in devise some other methods of sampling determine the character of the stand saplings in future studies, for in many any case and if they are uniformly dis- cases the sample derived by the method tributed, deviations on the part of lesser described above was inadequate. tree species are not likely to be signifi- As will be shown in detail elsewhere, cant. When this technique was applied the random pairs method, using 40 pairs to the current data it was found that in of trees per stand, gives results whose ac- 95 cases only five Chi-square values ex- curacy compares very well with that ob- ceeded the expected value at the five per tained by the quadrat method. For ex- cent level. In other words, five of 95 ample, a composite vegetational index differed significantly from a homoge- (see below) by random pairs for three neous distribution. This approximates stands on which adequate quadrat data very closely the expected number (5 out were also available showed an average of 100) of Chi-square figures which deviation of only 2.9% from the index would exceed the expected by chance. calculated from the quadrat figures. Du- On this basis it seemed reasonable to plicate analyses by the random pairs accept these as results of chance and rea- method made in three stands in two con- son for not rejecting the stands. This secutive years by two sets of inivesti- technique may be applied in the field in gators showed an average trial to trial the future; if a stand is found not to be variation in the index of only 1.34%. homogeneous further study will be in This compares with an average deviation order. of 2.11% in five sets of quadrats made in Soil samples were collected separately as many years in one of the same stands. from the AO, Al, and B horizons in most The random pairs method may therefore stands. This was accomplished by the be assumed to give reproducible results use of three small pits located in repre- of sufficient precision for the purposes at sentative portions of the wood, with the hand. three subsamples pooled for later analysis. 482 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3

Treatment of data Intosh 1950). Random pair frequency The data were tabulated and recorded numbers, however, are more limited, in on standard data sheets. Measures of that they may range from equal to the number, size, and distribution for each density, down to one half the density, tree species were calculated from the but they never can be smaller than one figures provided by the random pairs half the density values. As a result, fre- method for each stand. Relative (per quency numbers usually are close to den- cent) density and relative (per cent) sities in absolute magnitude. The im- dominance were determined in the ordi- portance value, therefore, is weighted to- nary manner, but frequency was calcu- ward density, in that the number of trees lated as percentage sum of frequency or present exerts a greater effect on the relative frequency (Raunkiaer 1934, Cur- index than does their size. This seems tis and McIntosh 1950), rather than the proper in a floristic index. more usual simple frequency. The pur- Soil acidity and water-retaining ca- pose of this change was to place fre- pacity were determined in the laboratory quency on the same mathematical basis as on air-dried samples which were ground density and dominance, so that a summa- with a mortar and pestle and sieved tion index (density plus frequency plus through a 2 mm screen. Acidity was de- dominance), with a constant value of 300 termined by the Helige-Truog method, for all species in a stand could be pre- while water-retaining capacity was meas- pared. The relative importance of a tree ured by the Hilgard method. Analyses species is more clearly expressed by this for available phosphorus, potassium, mag- method than by the DFD (density- nesium, and for organic matter were frequency-dominance) index in its origi- made by the Helige-Truog and Wilde nal form (Cottam 1948; Whitford 1949, methods at the State Soils Laboratory on Stearns 1951) where the sum was in- the University campus. Exchangeable determinate. The index is calculated by calcium was determined by volumetric adding the separate values of relative analysis, also at the State Soils Labora- density, frequency and dominance for tory. Nutrient ion concentrations are each species. Its magnitude is an excel- expressed as pounds per acre for the lent indication of the vegetational im- measured depth and weight of the par- portance of a species within a stand, ticular horizon sampled. In the interests since it is sensitive to such variables as of brevity, only the data for the A, apparent contagion or exceptional basal horizon are presented in this paper. area. This summation index of a par- RESULTS ticular species within a stand will be referred to herein as its importance value One of the difficulties inherent in the of in order to differentiate it from the DFD ''mass collection" method ecological study is the problem of handling the large index used by earlier workers. Not quantities of diverse data that accumulate only does it differ because of its use of during the course of the investigation. relative frequency, but because frequency It soon became apparent in the current as determined by the random pairs work that the first requirement was a method is not strictly comparable to fre- scheme for organizing or arranging the quency measured by quadrat sampling. stands so that suitable comparisons of Quadrat frequency values bear a mathe- data could be made. To this end, an at- matical relationship to density values, tempt was made to classify the stands with a fixed minimum ratio and a vari- into groups which had similar structure able maximum depending chiefly on the on the basis of tree composition. How- degree of randomness or contagion ex- ever, detailed examination of the data re- hibited by the species (Curtis and Mc- vealed no clear cut distinctions which JLly, 1951 UPLAND FOREIST CONTINU UM IN \WISCONSIN 483 would readily separate the stands into 'IAL~l~l I. List of maj'for trcc specic's found in 95 upland hardwood stands of southern IVisconsiin such groups without the exercise of a with data illustrating i portancrepotential considerable amount of subjective judg- ment. No. of Average Maximum Species stands of importance importance It is possible to classify the stands ob- occurrence value value jectively on the basis of the leading domi- Quercus rubra 80 90.4 228 nant trees in each. If the single leading Acer saccharum 45 82.8 201 Quercus alba 86 64.1 202 dominant is used for this purpose it Quercus velutina 40 62.8 206 Tilia americana 63 39.6 179 would be necessary to recognize 9 types Ulmus rubra 53 33.4 140 Quercus macrocarpa 25 24.8 170 in the 95 stands. The use of the first Fraxinus americana 28 20.0 94 Prunus serotina 56 17.3 114 two tree species in order would neces- Ostrya virginiana 48 16.2 42 Carya ovata 39 15.7 61 sitate recognizing 30 types; of the first Acer rubrum 18 13.3 58 three species, 75 Juglans nigra 21 12.5 30 types; of the first four, Carya cordiformis 43 12.2 44 95 types. In other words, no two stands Juglans cinerea 23 12.1 53 of the 95 examined had the same ar- rangement of their first four most im- of the upland forests of the region, it was portant tree species. The number of decided to study their interrelations as a units recognized is thus determined en- tirely by the basis of selection and in no possible initial basis for community de- way represents an approximation of any limitation. natural groups. Accordingly, we at- These four species were the leading tempted to devise a method whereby each dominants in 80 of the 95 stands studied. species present in the stand might be When these stands were placed in groups taken into account in the classification of with the same leading dominant, it was the stand. In this way, any tendency found that seven stands had Quercus of the stands to fall into distinct groups velutina as the major tree, 18 had Q. alba, would have maximum opportunity for 34 had Q. rubra, and 21 had Acer sac- expression. charun. The percentage occurrence A study of the importance values for (constancy, based on a sample of 80 trees all species soon showed that only a few in 40 random pairs) of each of the four species ever achieved high levels of im- species and their average importance portance. The majority of tree species values were calculated for each of these never occurred as important elements in groups. The results are shown in Table any of the stands. The ability of a tree II. The stands in which Q. velutina was species to achieve a given importance the leading dominant had a greater aver- value, or what may be termed its im- age amount of Q. alba than of either of portance potential, is the result of vari- the other two tree species, and Q. alba ous hereditary physiological attributes was present in a much greater fraction whose nature is but poorly defined at of the stands. In stands dominated by present. Some species, like Quercus Acer saccharumn,Q. rubra was charac- alba, reach major dominance in many teristically second in importance. On forest stands covering a wide geographic the other hand, stands dominated by area. Others, like Gymnocladus dioica, either Q. alba or Q. rubra each had two apparently never are important members species of about equal rank in second of the forest community in any part of place. Under these circumstances, the their range. Four tree species-Quercus most logical arrangement of stands is that velutina, Q. alba, Q. rubra and Acer sac- indicated in the table, which shows a de- charum-have an importance potential creasing order of Q. veluttina and an in- markedly greater than that of any other creasing order of A. saccharumn,with Q. trees of southern Wisconsin (Table I). alba and Q. rubra in intermediate posi- Since they are the major tree components tions. 484 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3

TABLE II. A verage importance value (I V) and constancy % of trees in stands with given species as the leading dominant (For species with highest importance potential only-80 stands)

Leading dominant in stand Species Q. velutina Q. alba Q. rubra A. saccharum Q. velutina Average IV 165.1 39.6 13.6 0 Constancy % 100.0 72.3 38.3 0 Q. alba Average IV 69.9 126.8 52.7 13.7 Constancy % 100.0 100.0 97.1 66.7 Q. rubra Average IV 3.6 39.2 152.3 37.2 Constancy % 25.0 94.5 100.0 76.3 A. saccharum Average IV 0 0.8 11.7 127.0 Constancy %0 5.6 29.4 100.0

TABLE III. Average importance value (I V) and constancy % of trees in stands with given species as the leading dominant (Eleven species of intermediate importance potential-80 stands)

Leading dominant in stand Species Q. velutina Q. alba Q. rubra A. saccharum Q. macrocarpa IV 15.6 3.5 4.2 0.1 C % 50.0 38.9 20.6 4.8

Prunus serotina IV 21.4 21.8 5.9 1.4 C % 87.5 89.0 64.8 19.0 Carya ovata IV 0.3 8.8 5.2 5.9 C % 12.5 61.2 38.3 33.3

Juglans nigra IV 1.5 1.2 2.2 1.9 C %O 12.5 11.1 20.6 23.8 Acer rubrum IV 3.9 2.3 2.4 1.0 C %O 12.5 33.3 23.5 4.8 Juglans cinerea IV 0 2.7 1.7 4.8 C %0 0 11.1 20.6 47.6 Fraxinus americana IV 0 1.9 5.1 7.6 C % 0 11.1 20.6 42.8

Ulmus rubra IV 4.6 7.7 8.3 32.5 C %O 25.0 27.8 53.3 85.7 Tilia americana IV 0.3 5.9 19.0 33.0 C %0 12.5 16.7 73.5 100.0 Carya cordiformis IV 2.5 5.8 4.1 8.2 C % 12.5 33.3 41.2 66.7

Ostrya virginiana IV 0 2.4 5.5 16.2 C % 0 22.2 41.2 95.3 July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 485

TABLE IV. Average frequency and constancy % of understory species in stands with given tree species as the leading dominant (80 stands)

Leading dominant in stand Species Q. velutina Q. alba Q. rubra A. saccharum Cornus foemina Average frequency % 20.0 23.9 16.2 0.0 Constancy % 66.6 100.0 67.8 12.5 Sanguinaria canadensis Average frequency % 0 7.5 7.8 14.0 Constancy % 0 35.7 46.5 50.0 Laportea canadensis Average frequency % 0 0 0.2 9.4 Constancy % 0 0 3.6 37.5

This ranking of the stands seemed to values for several herbs and shrubs were be in agreement with the general descrip- determined for the same groups of stands tions of the forest series reported for the (Table IV); they showed the same types region and thus gave promise of further of distribution, with optimum values in fruitful investigation. Accordingly, the a particular group for each species. same method (based on groups with Q. Measurements of certain soil characters zVelutina,Q. alba, Q. rubra and A. sac- were also averaged by these groups (Fig. charum as leading dominants) was used 2) and they too showed similar trends. to calculate average constancy percen- The alignment of lesser species with tages and importance values for a series the four major dominant trees shown in of other tree species of intermediate im- Tables III and IV is not very precise, portance potential. The results as shown since it does not separate those species in Table III have been arranged in a which may be associated with the same natural order in that those first in the major dominant. To make further sepa- list (Q. inacrocarpa, Prunus serotina, ration possible, the stands were classed Carya ovata) seem to be allied to Q. into eight groups, based on the two lead- velntiia and Q. alba, while those placed ing dominants in order of their im- last (Carva cordiformis, Tilia americana, portance values. Thus the stands which Ostrva virginicul a) are clearly associated had Q. alba as the main dominant were with Acer saccharum. split into two subgroups-those with Average frequency and constancy Q. velutina as the second most important tree and those with Q. rubra as the sec-

50 ond tree. Constancy percentages and average importance values were calcu- lated for all species in each subgroup. Some of the results are given in Figure 3. They are arranged with the sequence of the groups based on the information 20, in Table II and with the individual spe- cies arranged in order from those most closely associated with Q. vein tina to saccliarumt as Block Oak White Oak Red Oak Sugar Maple those most allied to Acer shown by the position of the stands in FIG. 2. Available potassium and calcium in the A1 soil horizon of stands in which the indi- which they reach optimum importance. cated species arc the leading dominants. On the lasis of this information, the 486 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3

10C, 100 IS0 our Oak Red Oak

75 Z|7 - 120 B75

50r 550 0 25 25 25-

too two 100 is thBecnleaigd tockOak Basswood 10 75- 75 .120 -75

50- 50- s0 60 25 25 5

Basswood firstadSugarMaplWhite t10 sugar Maple - s Oak 75 75 -10 10

50- 50-

60 .50 25- 25-

88 8W We WR RW RB BS SB 88 8W WS WR RW RB 85 SB

FIG. 3. Average importance value (solid bars) and constancy %7(hatched bars) for major species in stands grouped according to the two leading dominants. The scales on the left ordi- nates are constancy percentages, those on the right ordinates are average importance values. The group of stands are indicated by the paired letters below the graphs, where BB represents the stands in which Bur Oak (Q. macrocarpct) is the leading dominant and Black Oak (Q. v elutina) is the second leading dominant; BW represents stands with Black Oak as first dominant and White Oak (Q. a/ba) as second dominant; WB with White Oak leading and Black Oak second; WR with White Oak leading and Red Oak (Q. rubra) second; RW with Red Oak first and White Oak second; RB with Red Oak first and Basswood (Ti/ia amecricana,) second; B S with Basswood first and Sugar Maple (Acer sacceharuni) second; and finally SB with Sugar Maple the leading dominant and Basswood second.

species can be arranged in the following The method of leading dominants thus order, from pioneer to climax species: served to arrange the stands into groups Quercus macrocarpta -* Q. velutinac which showed a reasonable phytosocio- Prunus serotinct-* Q. alba-- Carya ovata logical pattern but it could not demon- -*Juglans nigra -Q. rubra-* Acer ru- strate the existence of discrete communi- brumn Fraxinus americana Tilia ties since it was based on an artificial, americana Juglans cinerea. Carya even though objective, choice of unifying cordiformis > Ulmus rubra Ostrya species. In addition, it suffered from the virgi-niana-> Acer sacclharum. vagaries resulting from the calculation July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 487

of averages from differing numbers of five stands, smoothed by the formula: stands in each group or subgroup. It (a + 2b + c)/4. seemed desirable, therefore, to treat the The other species were not used in es- stands individually, rather than to aver- tablishing this original rank, but they age them in groups. were found to be distributed in character- To facilitate handling this large amount istic positions when their importance of data, the moveable strip method was values in each stand were arranged in the devised. A strip of opaque white cellu- order established above. Figure 6 shows loid, 1/4 inch wide and 10 inches long, was distributions of Q. macrocarpa, Prunus prepared for each stand. The importance serotina, Tilia americana and Ostrya value of each tree species in the stand was virginiana. Quercus macrocarpa and shown on this strip by means of colored Prunus serotina, on the basis of their dis- marks, located on the basis of the scale of tribution relative to the major species, importance values from 0 to 300. By are seen to be more nearly akin to the this method the kinds and relative im- pioneer species, while Ostrya virginiana portance of the tree species in any stand and Tilia americana are seen to approach could be determined at a glance, and by the climax species, thus substantiating the placing the strips side by side in a suit- results of Figure 3. The remainder of able frame, comparison of many stands the species are distributed similarly when could be accomplished by inspection. plotted in this way. Natural groupings of tree species might The relative position in which each be expected to become apparent, if any species approaches its optimum develop- existed. ment can be fixed by means of diagrams The most enlightening arrangement of similar to Figures 5 and 6 where sufficient stands was made on the basis of the in- data are available. The breadth of any formation gained from the leading domi- of the figures and the absolute position nant method. The strips were placed so of the mode would be changed if more that the importance values of Q. velutina, stands were added to the series. How- Q. alba, Q. rubra and A. saccharum.most ever, the relative position remains essen- nearly approached the order of these tially the same once a sufficient number species which was indicated in Figure 3. of points is available. At the end of the This subjective procedure could not be first year's study, with data available for very precise, but a good approximation only 60 stands, the curves for the major was achieved (Fig. 4) as shown more species were well established in the same clearly by Figure 5, where the same data relative positions as given by the final are averaged by successive groups of data on 95 stands. A study of all the figures shows that each species has a minimum, an optimum and a maximum point for development, with the points located by reference to "ax t ^ o o 0 0?I similar states for other species. The too ~~~A 0 only exceptions are Q. macrocarpa and 0 A0 0 0 0 0 A A Acer saccharum', which lack one of the X X X a .0 AA0. dA A 0 0 0 cardinal points due to their location at 10- ee A0*A A 0 0C, 0 opposite ends of the series within the

.A c S0 O*D OF XXD X X 0 A 0 A0 A 0 XA~ A 00 limits of the study. It is to be empha- sized that the optimum points for each

SLACK OAK AWHITE OAK 0 Ago OAK X WOSARMAPLE species are purely relative. The strip method can show that Juglans nigra FIG. 4. Importancevalues of four major spe- cies in individual stands arranged in an order reaches its best development in forest most closely approachingthat shown in Fig. 3. stands which are intermediate between 488 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3

stands dominated by Q. alba and others so / o 0 0.~ .....-* dominated by Q. rubra but it cannot give a precise measure of the difference. 0 0 0 Nevertheless, the arrangement of stands provided by this method showvs that the *20 A , tree species found in the upland hard- woods of southern Wisconsin form a con- 20.~~~~~~~~a tinuous series, with overlapping ranges

but distinct conditions for optimum de- 0 A A- velopment of each species. No discrete O'dwr of VfONs groupings of species are apparent; rather, FIG. 6. Importance values of four lesser spe- the entire assemblage forms a continuum. cies, arranged in the order given in Fig. 4 and averaged by successive groups of 5 stands. Climax adaptation numbers The above results indicate that each may be evaluated by the spatial relations tree species reaches its optimum devel- of their optimum development curves as opment in stands whose position is fixed illustrated in Figures 5 and 6. Thus in a definite relationship to the other tree Tilia americana, whose highest impor- species. The sequence of the species in tance values are reached in stands near this pattern is such that pioneer species the A. sacclarumn end of the pattern, is are at one end and climax species at the better adapted to terminal forest condi- other. Pioneer and climax are here used tions than Q. rubra, with a peak closer in their adjectival sense and refer to the to the middle of the series, while Q. entire syndrome of specific physiological macrocarpa, which is most important in attributes which enables the one to live stands remote from those dominated by in the high light, variable moisture and A. sacchlarum, is the least well adapted. immature soil conditions of initial stands Using Acer sacchlarumn as a standard, and the other in the low light, medium all other species can be rated on the basis moisture and mature soil conditions of of the distance of their optimum stands the terminal forests. from those dominated by Acer. The The results of this study, as well as rating was accomplished by arbitrarily those of other investigations in this area, assigning a relative value of ten to A. indicate that Acer sacchlarum is the tree saccharum and lesser numbers to the species best equipped to persist in the other species according to their position, terminal forests of the area. All other as shown in Table V. Those species designated by an asterisk are species are less efficient in this respect, tentatively placed with an approximate value inas- and their relative degree of "climaxness" much as they occur only rarely in the

200 samples and thus their position is not clearly established by the few points

160 / \, available. The choice of suitable words to desig- nate the two ecologically distinct ends of the series is a troublesome one. The // A . word "climax" for the terminal end is

40 o fS V S especially doubtful, since it already has a plethora of meanings and probably would 10 o0f Stan2 not be defined the same by any two prac- ticing ecologists. It has been suggested FIG. 5. Same data as in Fig 4 averaged by suc- that the term "mesophily," as employed cessive groups of 5 stands. by Dansereau (1943), be used in the July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 489

TABLE V. Scientific and common names of tree species found in stands studied, with the climax adaptation numbers of each Climax Scientific name Common name adaptation number Quercus macrocarpa Michx. Bur oak 1.0 Populus tremuloides Michx. Trembling aspen 1.0 *Acer negundo L. Boxelder 1.0 Populus grandidentata Michx. Large-tooth aspen 1.5 Quercus velutina Lam. Black oak 2.0 Caryaovata (Mill.) K. Koch Shagbark hickory 3.5 Prunus serotina Ehrh. Black cherry 3.5 Quercus alba L. White oak 4.0 Juglans nigra L. Black walnut 5.0 Quercus rubra L. Red oak 6.0 Juglans cinerea L. Butternut 7.0 * Ulmus thomasi Sarg. Rock elm 7.0 *Acer rubrum L. Red maple 7.0 Fraxinus americana L. White ash 7.5 *Gymnocladusdioica (L.) Koch Kentucky coffee tree 7.5 Tilia americana L. Basswood 8.0 Ulmus rubra Muhl. Slippery elm 8.0 *Carpinus caroliniana Walt. Blue beech 8.0 *Celtis occidentalis L. Hackberry 8.0 Caryacordiformis (Wang) K. Koch Yellowbud hickory 8.5 Ostryavirginiana (Mill.) K. Koch. Ironwood 9.0 Acer saccharum Marsh Sugar maple 10.0 * Climax adaptation number of these species is tentative, because of their low frequency of occurrence in this study. present case instead of climax. It is signed to each species "climax adaptation true that the climax plants here discussed numbers." High numbers indicate good are extreme mesophytes, in the sense that adaptation to all of the environmental they achieve optimum development only factors present in terminal stands while in habitats of medium and relatively con- low numbers indicate adaptation to the stant conditions of both atmospheric and conditions of initial stands. This ad- edaphic moisture. However, it is equally mittedly unsatisfactory compromise at true and probably of greater significance least has the merit that it does not un- that these climax plants are shade-loving duely emphasize one factor in an ob- plants (sciophytes) able to grow and re- viously multiple factor situation even produce in conditions of very low light though it does further burden the al- intensity. For the most part they are ready overworked word climax. "Suc- also plants of high nutrient demands, cession number" has been suggested, but with tendencies toward partial hetero- it is inadequate, since an intermediate trophism. They are therefore not me- stage in the series may be reached by dian with respect to light or nutrition and retrogressive as well as by successional the use of the word "mesophily" might processes. carry unfortunate connotations. It would A comparison of these climax adapta- involve an extension of meaning as great tion numbers with the shade tolerance as in the case of "climax." What is ratings of foresters indicates that they perhaps needed is a word to encompass are roughly parallel. This is not un- all of those adaptive physiological and expected, since the characteristics that morphological characters that differen- contribute to great shade tolerance are tiate plants of initial stands from plants largely the same as those that make of terminal stands. In the absence of for persistence in terminal stands. The such a word, we have deemed it best here method used here in establishing the to use the phrase "climax adaptation" climax adaptation numbers may offer as a tentative description of these charac- a supplement to the usual procedures ters, and to call the relative values as- for the subjective determination of 490 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3 shade tolerance. It offers greater pos- silility of refinement than the coarse, qualitative estimates generally used. It must be recognized, however, that the application of this type of rating is only feasible within the bounds of a single :20/ floristic province, since shade tolerance / as measured by the adaptation number of a species may vary in other areas with different component . 500 1000 1500 2000 2500 300

Vegetational continuum index FIG. 8. Average frequency values of three herbs, in order of continuum To establish a basis for evaluating for- arranged index. est stands in terms of their total tree com- numbers is termed the position it was decided to use the climax weighted vegcta- adaptation number of each tree species as tional continiuuml index. A continuum a means of wNeighting its importance value index value was calculated for each stand as determined from the sampling data. in the study. When the celluloid strips The adaptation number of each species for each stand were arranged in the (as given in Table V) was multiplied by numerical sequence given by these index its importance value in a particular stand. values (Fig. 7) it was found that the dis- The products were added and the re- tributional pattern for each of the tree sultant weighted total used as a basis of species retained the same relative posi- placing the stand in its proper relation tion as before. to other stands. The maximum possible The vegetational continuum index is range of these weighted numbers is 2,700 actually a means of utilizing all of the units from 300 to 3,000. A stand Comll- tree species in a stand to express the posed entirely of Q. m1nacrocarpaor a position of that stand on a gradient. mixture of Q. wiacrocarpa and Populus Major importance in assessing the posi- tremutloides would have a sum of 300; a tion of a stand falls upon those species stand composed solely of A. saccharum which have a high importance value but would have a sum of 3,000. A stand each of the other tree species contributes consisting of mixtures of species whose in some degree. climax adaptation numbers lie between Tree is distributed on one an ten would have a sum somewhere reproduction the continuum index in between these extremes. This range of vegetational a pattern similar to that of the mature

200 trees. The reproduction should possibly be used to further weight the position of 160 O.k st. a stand in the vegetational continuum index, but, as indicated earlier, the sam- pling method used in this study gave / X ,00 ! inadequate data on saplings in most cases

0 50 0 00 I0 ~ ~ 00~ ~ 50 30 and therefore such weighting was not <40 /0 0 attempted. Herb and shrub species which played no -part in developing the vegetational continuum index were found to be dis- FIG. 7. Importancevalues of four major spe- tri'butedon itin patterns similar to those cies arranged in numerical order of continuum index and averaged by successive groups of 5 of the tree species. The distributions of stands. certain shrub and herb species is illus- July, 1951 UPLAND FOREST CONTiNUUM iN WISCONSIN 491 trated in Figure 8. They are found to DISCUSSION form curves with characteristic minimum, Three possibilities as to the nature of maximum for devel- optimum and points the upland hardwood forests of this area These curves locate the lesser opment. may be postulated. The first of these is species relative to the tree species and that the forests represent an amorphous to each other. Some herbaceous species, collection of plants in which no units or unlike most trees, are found to cover the patterns are discernable. In other entire range of stands, but it is never- words, they are chance aggregates ac- theless possible to predict with reason- cording to the viewpoint of Mason (1947) able accuracy the important herb species and are dependent solely on a "coimci- to be found in stands in any given range dence of tolerance" between plants and of the vegetational continuum index. environment. The presence of a pattern Soil factors also may be related to the definable in terms of tree composition forest cover by means of the vegetational and the demonstration that trees and continuum index. The relation of ex- other plants are not found together in changeable calcium and of soil reaction chance mixtures but in a rather definite to the index is shown in Figures 9 and pattern would indicate that this interpre- 10. The values ill these graphs were tation is not tenable in the present case. Secondly, one could maintain that these

0, forests represent several discrete com- munities, distinguishable from each other by boundaries which are reasonably dis- a , ?~~~~~~~~~~~/ 84 / tinct in terms of the measurements avail- able to the plant ecologist. If this were E 3~~~~~~~~ / true, then determination of the relative sociological importance of the trees in each stand should show that certain groups of tree species reach their opti- 500 )000 1500 2000 2500 3000

COMI".u. Index nmum development in the same stands, in FIG. 9. while other groups occur together other stands. If the boundaries between corn- are obtained by averaging successive groups munities actually distinct, then the species of one group should never occur of three stands for which data were as important members of another group. available. Similar correlations were ob- The results of this study indicate that with tained the water retaining capacity such a discrete grouping of species does and the organic matter content. not occur in Wisconsin. Rather, each species reaches a high level of importance in only a few stands, with no groups of 0 / two or more species showing optimum / 0 development in the same stands. 0 / It is thus necessary to accept the third 0// .- possibility, namely that the upland hard- woods of southwestern Wisconsin repre- sent a contilmuuitr in which no clearly de- fined subdivisions are discernable but in which a definite gradient is exhibited. 500 1000 1500 20C0 2500 X FIG. - - - 000 If this is truly the case then the entire FIG. 10. Acidity of the A, layer in stands ar- physiognomic upland hardwood forest is ranged on the continuum index. to be interpreted as an entity (or por- 492 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3 tio)n of an entity) wnithout discrete sub- The geographical limits of this con- divisions. The failure of any one or any tinuum outside of Wisconsin are not group of the environmental factors meas- known at present, but it appears that ure(l to indicate any separation sul)- southeastern Minnesota, northeastern stantiates the conclusion of unity 1)ased Iowa and northern Illinois are definitely on the tree data. The presence in cer- included. There are indications that the tain of these environmental factors of upland forests on the southern border of some indication of a gradient or trend re- the prairie peninsula are similar also, es- lated to the tree gradient is further in- pecially those in , dication that these forests represent a (Aikiman 1926), (Arend and continuous dine from initial stages corn- Julander 1948) and Kentucky. This is posed of pioneer species to terminal perhaps to be expected on the basis of stages comlpose(1 of climax species. climatic similarities (Borchert 1950). This conclusion would appear to sul)- The continuum throughout is intimately stantiate the individualistic association associated with the prairie-forest border. hypothesis of Gleason (1926) and its Its major trees are better adapted to enmeidation lby Cain (1947). It differs, xeric conditions, particularly those as- however, in that "mlaliniitcd variables, sociated with periodic severe summer comlbl)inlationsand permutations" (Cain) droughts, than are most other species of (lo not occur. All things are not possi- the deciduous forest. It has borne the 1)le, only some, and these possible permu- brunt of repeated prairie advances in tations seem to follow a pathway pre- post-Wisconsin time and possibly in scrihed by the physiological potentialities earlier interglacial periods as well (Lane of the available flora within the limits of 1941, McComb and Loomis 1944). As the existent physical environment as a result, it is probably simpler in 1)oth mo(lifiable by that flora. Within a given floristic composition and developmental floristic province, the amplitude of en- pattern than the forest aggregations far- vironimental tolerance of each species is ther east and south. It is recognized more limited than the full range of en- that this simplicity was an important vironment in that province. Further- factor in the recognition of the continuum more, although the tolerance range of pattern and that similar patterns may be each species may overlap that of many more difficult to ascertain in other re- other species, the range for optimum de- gions of greater floristic variety. velopment of each species is different If the vegetation of a particular physi- from that of all other species. There are, ognomy on a particular topography in a therefore, no groups of tree species which particular floristic province (i.e., decid- regularly occur together and only to- uous hardwood forests on the uplands gether, except for accidental duplications of southwestern Wisconsin) forms a con- of narrow environmental ranges. Rather, tinuum without identifiable individual tree species occur in a continuously shift- segments, as is the contention of the ing series of combinations with a definite present paper, then certain far-reaching sequence or pattern, the resultant of a implications are apparent. Without at limited floristic complement acting on, and this time entering the perennial argu- acted upon by, a limited range of physi- ment concerning the relative scope of the cal environmental potentialities. Such a word "association" as used by American gradient of communities is here called a and European workers, it is obvious that vegetational continuum, to differentiate both groups usually apply it to discrete it clearly from the controversial concept communities which are thought to be of a plant association. As applied to recognizable as distinct entities in the forests, it is similar in scope to the field. It is on this characteristic that the "sylvan continuum" of Darling (1949). whole series of "synthetic" characters of July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 493 plant communities was erected by Riibel synthetic characters would be still fur- (1922), Braun-Blanquet (1927) and ther reduced in usefulness, since their other European workers. Such well- arithmetic values would be of very low estahlished community measures as pres- magnitude. Presence and constancy may ence, constancy, and fidelity 1)ecome of continue to maintain their useful roles in greatly restricted value when applied to phytosociology if each investigator ac- stands which form a continuum. For companies them with a detailed account example, in the present case, a number of the limits set by him on the range of of herbs are known which may be found communities included within his study. in almost every initial stand in which When used in this way, the synthetic the leading dominant is Quercuts velutina characters may be very valuable, particu- and thus would have a high degree of larly in comparative studies, as between presence or constancy in those stands. grazed and nlon-grazed woodlots or the All of them, however, also occur in lesser like. amounts and in a smaller percentage of It is hoped that the concept of a vege- the stands which are dominated 1)y Q. tational continuum and the methods for allb(a. This diminution continues in the its analysis as developed in this paper will Q. rblfra. stands and most of the species prove to be widely applicable to the forest become verve rare or absent in the termi- stands of other regions. Whether or not nal riccr s(ccliaruuui forest. A similar all floristic provinces have similar con- situation in reverse exists for a group of tinous gradients rather than recogniz- species most common in the Acer stands. ably discrete communities is not known The absolute value of presence or con- at present, although a few studies based stancy, therefore, is determined almost on the analysis of many stands in a sin- solely by the amplitude allowed l)y the gle region (Hanson and Whitman 1938; investigator in his choice of stands to be Whlittaker 1948; Allen 1950; Potzger studied. If a large portion of the con- 1950) clearly indicate that similar con- tinuutm is used, then all species except tinua are present. Dyksterlhuis (1949) the most ubiquitous and least demand- also shows a continuously varying com- ing have very low values of presence and position of grassland tinder the impact constancy. These ubiquitous species are of overgrazing which is essentially simi- commonly found in nearby non-related lar to the forest continuum here de- asseml)lages and hence are of very low scribed. value as community indicators in any In those areas in which its construction case. As the portion of the continuum is possible, the vegetational continuum used becomes restricted in scope, then index offers a new approach to ecological the presence and constancy values in- studies of many types. It may be used to crease in magnitude. As we have seen, describe local deviations in tree compo- however, scarcely any two stands exist sition such as occur when pioneer species which have precisely the same comple- enter spot openings in a mature forest ment of tree species, to say nothing of resulting from a restricted fire or other the more numerous herbs, and hence, disturbance and such other situations as presence and constancy values rarely are included in the "gap phase" concept achieve truly high levels. of Watt (1947). Autecological studies It appears probable from preliminary of an herbaceous species may be con- o1)servations that the lowland forests ducted in stands which are known to be (which receive at least temporary ac- optimum for the development of that cumulations of run-off water) of this species and comparative life history area will eventually prove to be part of studies of groups of species may be made the same continuum herein discussed. readily in stands known to be similar on If this should indeed be the case, then the the basis of trees and soils. The index 494 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3 provides a more detailed framework for 2. The relative ecological importance the study of the relation of random and of each tree species in each stand was contagious distributions of herbaceous expressed by a summation index of the species in relation to age of occupancy as relative frequency, relative density and discussed by XNNlbitford(1949). Mam- relative dominance herein called the i111- mal, bird, insect or other animal popula- portan-ce value. This value has a con- tions, their fluctuations and interrelation- stant total of 300 for each stand. ships, are all inseparably interwoven with 3. Four tree species (Ouercits velutinia, plant communities. One of the major Q. alba, Q. rlfbra and Acer saccharine) difficulties heretofore encountered in attained an average importance value in studies of natural animal populations has all stands much higher than that of any been the lack of an adequate means of other species. When the stands were delimitation of the nature of the plant arranged into groups each dominated segment of the habitat complex. The by one of these four species, it was found vegetational continuum index affords a that each of the lesser tree species was more accurate means of describing this associated with one of the groups. Fur- habitat factor and in addition immedi- ther breakdown of the stands into sub- ately places a stand in its proper relation groups according to the second most to other stands. As further information important species, resulted in a more ac- is collected on the correlations between curate alignment of all species into a the physical factors of the environment series beginning with the pioneer species, and the index and on the relations of the Quercus wnacrocarpa, Popidius grandi- microl)iota or other special groups of dentata, and Quercus velutina, and end- organisms to the index, it will be possible ing with the climax species Ostrya vir- to summarize a very considerable amount gin ianiia and Acer saccelartumt.Each spe- of information by means of an index cies reached optimum development at a numl)er. There is no thought here that given point along this series, but no the index will serve as the final touch- groups of species were found which stone of phytosociologists, nor that the achieved their peak in the same stands. complex of physical and biological agents Thus no discrete communities could be of importance in a community can pos- detected on the basis of the leading domi- silbly be reduced to a single number. The nants in the stands. index rather should serve as a frame- 4. Use of a strip method for visual work upon which a great amount of inde- comparison of the importance values of pendently derived information may be ar- all species in each stand resulted in a rangecl and interrelated. ranking or order of stands which agreed with the order obtained from the leading in SUMMARY dominants. The initial stands the rank were composed of shade-intolerant, 1. The upland hardwood forests of the non-mesic, pioneer species, while the ter- prairie-forest floristic province of south- minal stands contained the very shade- western Wisconsin were studied by tolerant, mesic climax species. Upon the means of a random sampling of undis- assumption that Acer saccharuow was the turbed stands distributed through 29 most successful species in the terminal counties covering an area 160 miles by forests, it was possible to arrange all 230 miles. Data on tree, shrub, and herb other tree species in a descending order composition were obtained in 95 stands according to the position of the stands by the random pairs method of sampling. in which they reached optimum develop- Soil acidity, nutrient concentration and ment, relative to the stands containing moisture constants were also determined Acer sacharums. Arbitrary numbers for these stands. from one to ten, called climax adaptation July, 1951 UPLAND FOREST CONTINUUM IN WISCONSIN 495 nioubcrs, were assigned to the species to ently derived information on animal in(Iicate the degree to which they ap- l)opullations, mnicrobiota,and physical fac- proach the climax Acer in autecological tors of the environment. characteristics. For the more important trees of the region, the adaptation num- LITERATURE CITED bers are as follows: Quecrcus uwcrocarpa Aikman, J. M. 1926. Distribution and struc- -1.0, Qurcius veln tina-2.0, Carva ture of the deciduous forest in eastern Ne- ova ta-2.5, Pri mus serotina-3.5, Qlicr- braska. Univ. Nebraska Studies 26: 1-75. Aikman, J. M., and A. W. Smelser. 1938. clls alba-4.0,Qiterclis riubra-6.0, U/wnits The structure and environment of forest rlublra-7.0, Tilia (a11)lericaca-7.5, Ostrva communities in central Iowa. Ecology 19: virgi)nlimlan-8.0, and Acer sacc/ciarum- 141-170. 10.0. Allan, P. F. 1950. Ecological basis for land 5. The climax adaptation numbers use planning in Gulf marshlands. Jour. Soil and Water Conserv. 5: 57-62, 85. were used as a means of weighting the Anderson, Edgar. 1941. The technique and importance values of each species in a use of mass collections in plant taxonomy. stand, as determined by field sampling. Ann. Missouri Bot. 28: 287-292. The sumn-mation of these weighted num- Arend, J. L., and 0. Julander. 1948. Oak sites in the Arkansas Ozarks. Arkansas bers resulted in an index which served Agr. Exp. Sta. Bull. 484: 1-42. to locate the stand along a gradient. All Borchert, J. R. 1950. The climate of the species present contributed to the index, central North American grassland. Ann. with the greatest effect exerted by the Assoc. Amer. Geog. 40: 1-39. dominant species. On this basis, the 95 Braun-Blanquet, J. 1928. Pflanzensoziologie. Julius Springer. Berlin. stands studied were distributed uni- Bruncken, E. 1900. On the forest conditions formly between an index value of 632 in the vicinity of Milwaukee. Bull. \Vis- and a value of 2,650 (out of a possible consin Nat. Hist. Soc. 1: 179-184. range from 300 to 3,000). No distinct Cain, S. A. 1947. Characteristics of natural areas and factors in their development. groups of stands were apparent-rather Ecological NMonog. 17: 185-200. the entire series of communities formed Chamberlain, T. C. 1877. Native vegetation. a continuum in which a definite gradient Geology of \Wisconsin 2: 176-187. was exhibited from initial stages com- Cheney, L. S. 1894. Is forest culture in Wis- consin \Visconsin posed of species to terminal desirable? Trans. State pioneer Hist. Soc. 24: 163-170. stages coml)ose(l of climax species. Such Clark, 0. R. 1926. An ecological comparison a gradient of communities is here called of two types of . Proc. Iowa Acad. a vegetational con1tinni lnH. Sci. 33: 131-134. 6. Herb and shrub species were found Clements, F. E. 1905. Research methods in ecology. Univ. Publ. Co., Lincoln, Neb. in when to be distributed typical patterns Cottam, G. 1949. The phytosociology of an their frequencies were plotted against the oak woods in southwestern Wisconsin. continuum index. Each of these under- Ecology 30: 271-287. story plants was restricted to a particular Cottam, G., and J. T. Curtis. 1949. A nmethod for range on the index each reached opti- making rapid surveys of woodlands by means of pairs of randomly selected indi- munl developments only within a particu- viduals. Ecology 30: 101-104. lar group of closely related stands. Vari- Curtis, J. T., and H. C. Greene. 1949. A ous measurements of soil properties study of relic \Visconsin prairies by the showed a good correlation with the index. species-presence method. Ecology 30: 83-92. Curtis, J. T., and R. P. McIntosh. 1950. This was especially true of the concen- The interrelations of certain analytic and tration of available calcium, the water synthetic p)hytosociological characters. Ecol- retaining capacity and the organic mat- ogy 31: 434-455. ter content of the A, soil layer. Dansereau, P. 1943. L'erabliere Laurentienne. Can. 21: 7. It is suggested that the vegetational Jour. Res. 66-93. Darling, F. F. 1949. History of the Scottish continuum index may be a useful tool forests. Scottish Geog. Mag. 65: 132-137. for the coordination of much independ- Daubenmire, R. F. 1936. The "Big Woods" 496 CURTIS AND MCINTOSH Ecology, Vol. 32, No. 3

of Minnesota: Its structure, and relation to western WVisconsin. Trans. Wis. Acad. of climate, fire and soils. Ecological Monog. Sci., Arts, Letters. 40: 243-258. 6: 235-268. Pammel, L. H. 1904. Forestry conditions in DeForest, Howard. 1921. The ilant ecology western \Visconsin. Forestry & Irrig. 10: of the Rock River woodlands of Ogle Co., 421-426. Illinois. Ill. State Acad. Sci. Trans. 14: Potzger, J. E. 1950. Forest types in the 152-193. Versailles State Park area, . Amer. Desmarais, Yves. 1948. Dynamics of leaf Midland Nat. 43: 729-746. variation in the sugar maples. Plh.D. thesis. Raunkiaer, C. 1934. The life forms of plant lUniv. of \Yisconsin. and statistical plant geography. Oxford, Dyksterhuis, E. J. 1949. Condition and man- Clarendon Press. London. agement of range land based on quantitative Raup, H. M. 1947. Some natural floristic ecology. Jour. Range NManage. 2: 104-115. areas in boreal America. Ecological Monog. Eggler, W. A. 1938. The maple-basswood 17: 221-234. forest type in W\aslhburnCounty, Wisconsin. Riubel, E. 1922. Geobotanische Untersuchungs- Ecology 19: 243-263. methoden. Berlin. Egler, F. E. 1947. Arid southeast Oalhu Salamun, P. J. 1950. The interpretation of vegetation, Hawvaii. Ecological NIonog. 17: variation in populations of Spiraca[ t[oiicl- 383-435. tosa, I ra1clvc1lfruim crection, Osiiiorh7iz Fassett, N. C. 1941. MXasscollections: Riibius claytom, and Osniiorhizo lio isti lis. Ph.D. odorahls and Re. parv'iflorits. Ann. Missouri thesis. Univ. of \isconsin. Bot. Gard. 28: 299-374. Snedecor, G. W. 1946. Statistical methods. Fuller, G. D., and P. D. Strausbaugh. 1919. Fourth Ed. Iowa State College Press. On the forests of LaSalle Co., Ill. Ill. State Ames. Acad. Sci. Trans. 12: 246-272. Stearns, F. 1951. The composition of the Gleason, H. A. 1926. The individualistic con- sugar maple-lhemlock-yellow birch associa- cept of the l)lant association. Bull. Torrey tion in northern \Wisconsin. Ecology 32: Pot. Club 53: 7-26. 245-265. Griggs, R. F. 1914. Observations on the be- Stout, A. B. 1944. The bur oak openings of lhavior of species at the edge of their range. southern \Viisconsin. Trans. \Visconsin Bull. Torrey Bot. Club 41: 25-49. Acad. 36: 141-161. Hanson, H. C., and W. Whitman. 1938. Wagner, R. D. 1947. The ecological com- Characteristics of major grassland types in position of an oak-maple subclimax forest western . Ecological Mlonog. in Jefferson County, \Visconsin. M.S. thesis. 8: 57-114. Univ. of Wisconsin. Lane, G. H. 1941. Pollen analysis of inter- Watt, A. S. 1947. Pattern and process in glacial p)eats of Iowa. Iowva Geol. Survey the plant community. Jour. Ecology 35: 37: 233-262. 1-22. Lapham, I. A. 1852. \Voods of Wisconsin. Weaver, J. E., and F. E. Clements. 1938. Trans. W\isconsin State Agr. Soc. 2: 419- Plant Ecology. McGraw-Hill. . 434. Whitford, P. B. 1951. Estimation of the Marks, J. B. 1942. Land use and plant suc- ages of hardwood stands in the prairie-forest cession in Coon Valley, \Visconsin. Eco- border region. Fcology 32: 143-146. logical Monog. 12: 113-134. 1949. Distribution of woodland plants Mason, H. L. 1947. Evolution of certain in relation to succession and clonal growth. floristic associations in western Nortlh Amer- Ecology 30: 199-208. ica. Ecological Mionog. 17: 201-210. Whittaker, R. H. 1948. A vegetation analy- McComb, A. L., and W. E. Loomis. 1944. sis of the Great Smoky Mountains. Ph.D. Subclinmax prairie. Bull. Torrey Bot. Club thlesis. Univ. of Illinois. 71: 46-76. Yapp, R. H. 1922. The concept of habitat. McIntosh, R. P. 1950. Pine stands in south- Jour. Ecol. 10: 1-17.