Bulletin OF THE Illinois State Laboratory OF Natural History Uebana, Illinois, U. S. A. STEPHEN A. FORBES, Ph.D., LL.D., DiRBCTOR Vol. IX. October, 1910 Article III. THE VEGETATION OF THE INLAND SAND DEPOSITS OF ILLINOIS BY HENRY ALLAN GLEASON, Ph.D. ERRATA AND ADDENDA Page 54, lines 3 and 2 from bottom, and elsewhere in Article III. for Cassia chainaechrista read Cassia chamaccrista. Page 62, between lines 4 and 5 from bottom of table insert Erigeron annuus. Page loi, table, after Croloii glandulosus read var. septentrionalis: and for Eijuisettim laeinyatum read Eqnisetum bycmale var. inlermedium. Page 131. line 3, for coerulea read caerulca. Page 138, last line, for Zi::a read Zizia. Page 141, line 21 from bottom, dele Diodia teres. Page l6g. between lines 3 and 4, insert as follows : Erigeron annuus (L.) Pers. An interstitial in the liunch-grass association in the Hanover area. Page 177, line 5, for casti^'ard read iveslzvard. Page 209. line 3 from bottom, for copalina read copallina. Page 210, line 13 from bottom, for Diospyrus read Diospyro'i. Page 211, line 5, for Foresteria read Foresticra. Page 256, line 3 of table, for Dr. H. M. Pepoon read H. 5". Pepoon. Page 278, line 16, the fifth word should be in Roman type. Page 286, line 6 (second column), page 295, list of secondary species (second column), and page 353. line 8 from bottom, for hiematis or hiemale read hye- mnte. Page 313, line 4 from bottom (first column), for pedicularis read pcdicularia. Page 315, line 10, second column, for Apoeynum read Apocynum. Page 323, line 3 from bottom, for Cyperus read Scirpus. Page 330, line 14, for virginianum read inrginicum. Page 336. lines 3 and 2 from bottom, for virginicum read virginianum. Page 337. line 2 from bottom, for philadelphicum read philadelphicns. Page 339» in first list of invading species, for Rhus hiria read Rhus typhiua. Page 351. line 4 from bottom, for .verophtic read .verophytic. Page 355, above line 6 from bottom, insert Scirpus heterochaetus Chase. Page 356, line 14 from bottom, for Symlocarpus read Symplocarpus. Page 360, line 14, for Pirus read Pyrus. Page 362, after line 7, insert Acer saccharinum L. Page 363, line 2 from bottom, for quadiflorum read quadriflorum. Page 365, line 14, for Ihapus read thapsus. Page 369, last line, for Tanecelum read Tanaceium. Page 417, line i. dele the. Page 497, line 9 from bottom, for neglible read negligible, and in foot-note, for Auslall read Anstalt. Page 498, line 4 from bottom, for Lockport read Chillicothe. Page 500, line 13 from bottom, after up insert in. Page 5or, line 2 from bottom, for dissolving read dissolved. Page 504, line 23, for gryina read gyrina : line 17, for dentata read knickerbockeri. Page 506, line 11, for vernata read ternala. Page 507, line 3 from bottom, for Macon read tt'asoi. Page 513, line 19, for Nepa read Zaitha; line 18, and page 517, line 13 from bot- tom, page 520, line 12 from bottom, and page 532, line 4, read naid or naids for naiid or naiids. Page 517, line 6 from bottom, for pondiveed read pickcrel-iveed. Page 519. for first sentence of last paragraph read as follows: We have no exactly comparable chemical data for July; but analyses for August give percentages of saturation for Morris and Marseilles as follows : 20.4 per cent, at Morris on the nth and 11 per cent, at Marseilles on the 12th; 16.35 P<^r cent, at Morris on the 22d and 23d and 7.4 per cent, at Marseilles on the 24th and 25th. Page 521, line 6 from bottom, and page 529, line g, for chrysoleucas read cryso- leticas. Page 525, line 22, and page 536, lines 21 and 24, for Ekmann read Ekmnn. Page 532, line i, for Ancyclus read Ancylus. Page 551, line 7, for 00 read 572. Page 615, second line above foot-note, for 106 read 94. Page 616, line I, for the second Biindeln read Bilndel; line 2, for Biindeln read Biindcls; line 3, for ausscrn read ausscren; line 6, for sweierlie read sweierlei. Page 629, line 12, for kciii read kcincn. Page 634, line 9, for untcrnommcn read untcninntincnen; and in line 14 from bottom, after ;;/$ insert is fig. Plate III, Fig. i, after the word mixed in legend insert consocies of the. Plate IX, Fig. 2, dele the legend and read instead ; Root-system of Tephrosia virgiiiiana, exposed by blowing of the sand. Plate X, Fig. 2, dele the legend and read instead ; A blowout almost stabilized by bunch-grasses, especially Leptoloma cognatum. Plate XXXIX. for Calainogrostis read Calamagrostis. Plate LIV. exchange places of cuts, but not the legends. Plate LXXXV, for 7 read ye. CONTENTS Introduction 23 Physiography and Origin 24 Climate 3 ' Ecological Environment 34 General Discussion 35 The Vegetation 42 Classification of the Plant Associations 46 The Prairie Formation 47 The Bunch-grass Association , 47 Successions from the Bunch-grass Association 77 The Panicum f'seudopiibesceiis Association 79 Reversion to the Bunch-grass Association 83 The Blowout Formation 84 The Basin Association 90 The Windward Slope Association 91 The Blowsand Association 92 The Deposit Association 95 Successions between the Associations of the Blowout Formation 100 Stabilization of the Blowouts and their Reversion to Bunch-grass lOS The Hiidsonia Association 106 Successions from the Blowout Formation 107 The Blowout Thicket Association 107 The Stenophyllus Association 108 The Swamp Formation lOQ The Salix and Solidago Associations lOQ The Polytrichum Association 1 12 The Swamp Association 1 14 Succession of the Prairie Formation by the Forest 116 The Forest Formation 121 The Black Oak Association 121 Successions from the Black Oak Association 129 The Bur Oak Association 129 The Mixed Forest Association 13S The River Dunes and their Plant Associations 139 The Perched Dunes 14S Annotated List of Species 143 Bibliography 171 Article HI. — The Vegetation of flic Inland Sand Deposits of Illinois. By Henry Allan GlEason. Introduction* In the rapid development of ecological and phytogeographical knowledge during the past few decades, the vegetation of sand de- posits has been the subject of especially frequent and detailed study. At least three reasons may be mentioned why this type of vegetation has received particular attention. First, sand deposits are usually well developed and form dune complexes of greater or less extent along the shores of the ocean or the larger inland lakes, and in many cases are convenient places for vacation trips. Secondly, the vegeta- tion on sand is usually open and easily studied, and the dynamic na- ture of the environment is emphasized. For this reason the inter- relations of jilant and environment are more easily observed and offer attractive fields for study. Thirdly, sand areas are usually infertile in comparison with their surroundings. They are accordingly fre- quently left uncultivated and constitute temporary natural preserves, in which the original types of vegetation persist and are available for study. In the case of the inland sand regions of Illinois the first state- ment is hardly eftective, and that may explain why they have received relatively little attention from local botanists. At the present time, however, they comprise the largest, and virtually the only, areas of natural vegetation within the state. With the e.xception of parts of the sand deposits, of some small swamp areas, of rock outcrops, of ponds and lakes, and of some small tracts of forest, all the original vegetation of Illinois has been destroyed or greatly modified by clear- ing, planting, or pasturing. The area covered by the last four of these exceptions is \'ery small, but there are still thousands of acres of sand deposits in nearly orig'inal condition and available for study They still contain some virgin prairie that has never been plowed or pastured. These prairies are probably somewhat different from the more representative types of prairie which formerly grew upon more fertile soil, but they are much more nearly typical than the small strips still occurring along the margins of some streams and ponds. * The field work upon which this article is hascd, was carried on by the aid of a grant from the P.otnnical Society of America. Further financial assist- ance was given by the Illinois State laboratory of N;itural History. 24 The study of the vegetation of tlie sand deposits of Illinois is therefore of especial scientific interest because they constitute the only considerable area of natural vegetation in the state, and because their vegetation is closely related to that of the original prairie. It is also of some general value, since it concerns an area which has re- ceived little attention from botanists, and because it affords intel- ligible illustrations of certain ecological principles. The field work upon which the present paper is based was done during the summer of 1908. Reference is also frequently made to the field work at Havana in August, 1903, and August, 1904, the re- sults of which have already been published (Hart and Gleason, /po/). The itinerary during 1908 was as follows: May 28, 29, St. Anne, Kankakee county. May 3 1 -June 3, June 12-24, August 15-18, Hanover, Jo Da- viess county. June 25-29, Shirland, Winnebag'O county. June 30-July 3, August 19, 20, Dixon, Lee county. July 4-15, August 10-14, Oquawka, Henderson county. July 16, Forest Cit)'^, Mason county. August 7, Topeka, Mason county. August 8, Havana, Mason county. August 21, Amboy, Lee county. From June 12 to 19 Mr. Frank C. Gates assisted in the field work. He also identified many of the plants mentioned in the paper. The grasses and sedges were identified through the courtesy of Mrs. Agnes Chase. Mr. H. N. Patterson rendered important assistance in the field work in the Oquawka area. Dr. H. S. Pepoon has supplied valuable information concerning the Hanover area. To each of these the writer extends his thanks for their interest and appreciation. The photographs have been taken by the writer, using a folding film camera, Ansco films, and tank development. Physiography and Origin The chief sand deposits of Illinois lie in the northern half of the state, between latitude 40° and 42° 30'. In the southern half sand occurs only in small local deposits or in bars near the larger rivers, and is never of such extent that a peculiar vegetation is developed upon it. Banks and bars of sand also border the streams of northern Illinois, but their vegetation bears little relation to that of the larger deposits here described. For convenience the sand areas have been given names taken from 25 some geographical feature of the vicinity. Some of these areas are contigiious, and some owe their existence to the same causes. The names, therefore, do not indicate areas which are geologically dis- tinct, but merely general locations in which the field work was prose- cuted. The geography of each of these regions will be described separately. The Havana Area.—The Tazewell sheet of the Field Operations of the Bureau of Soils (Bonsteel, ipo^a) shows the northern extrem- ity of this deposit, and illustrates its relation to the glacial valley of the Illinois river. North of Pekin, in Tazewell county, the Illinois river cuts through the Shelbyville and Bloomington moraines, flow- ing close to high bluffs on its left (eastern) side. From this point southward the river crosses the broad glacial valley diagonally toward the right, exposing a triangular area of lowland between the channel and the east bluffs. The sand is deposited in this glacial flood-plain. At Pekin the plain is about two miles (3 km.) wide; below that city it widens more abruptly, and near Green Valley is 14 miles (22 km.) wide. At some places near the river the plain is covered with modern alluvial deposits, and it is crossed by the Macki- naw river with its broad flood-plain. The remaining area is occupied by sand and by a sandy loam, shown in the Soil Survey as Miami sandy loam. The latter lies at a lower level and represents the original alluvial deposits upon which the sand has been superposed. In this county 22,976 acres (90 sq. km.) are covered with sand. South of Tazewell county the plain retains its maximum width across Mason county, and then becomes gradually narrower toward the south, terminating near Meredosia, Morgan county, with a total length of approximately 75 miles (120 km.). While sand deposits occupy only a portion of this area, their aggregate extent is large and has been estimated (Hart and Gleason, ipo/: 145, 146) at 179,200 acres (700 sq. km.). The Chicago, Peoria and St. Louis railway traverses the areas from Peoria through Havana to Virginia, and a good idea of the general topography may be gained from its trains. The exposed areas of Miami sandy loam, which forms the foundation of the whole, are irregular in shape and extremely variable in size, ranging from a few acres up to several square miles. They are almost entirely under cultivation. Above them rise the low sand hills (PI. I, Fig. i), usu- ally gently undulating at their margins but, if large in extent, fre- quently quite level toward the center. These vaiy in size from mere hills of a few acres up to continuous deposits several miles in extent. Their average height is probably 20-30 feet (6-10 m.), l)Ut isolated 26 dunes rise much higher. One of the highest Hes about four miles (6 km.) north of Topeka, and is probably about 60 feet (18 m.) above the general level. Part of the sand was originally covered with prairie, but most of this has been destroyed by cultivation and pastur- ing, so that only a few small areas remain in their natural condition. A larger portion has been forested, and much of it remains in its virgin state. Particularly large tracts of forest are situated near For- est City and between Kilboume and Bath. The Hanover Area.—This region of sand deposition takes its name from the station of the Chicago, Burlington and Quincy rail- way in Jo Daviess county, which lies near the location of the best development of sand vegetation. As in the Havana area the sand occupies the so-called second bottom, between the bluffs on the east and the Mississippi river on the west. In some places the sand ex- tends to the river's edge, in others a strip of alluvial forested flood- plain intervenes. In the northern portion of the county the bluffs lie close to the river and the sand is limited to small isolated areas. In the southern half the bluffs and river become one to three miles (2-5 km.) apart, affording space for an extensive sand deposit. North of Savanna, in Carroll county, the river again flows directly at the base of the Ijluffs. The area in Jo Daviess county covered by sand is estimated at 5700 acres (22 sq. km.). Unlike the Havana area, the sand deposits here are nearly contin- uous and unbroken by intervening areas of a different soil. The surface of the area is gently rolling, with virtually no extensive level tracts. Its general elevation is about 25 feet (8 m.) above the river, but isolated dunes reach a much greater height. Near the eastern margin of the valley the depth of sand abruptly decreases, leaving a trough-like valley extending for a long distance at the base of the bluffs. The Chicago, Burlington and Quincy railway lies mainly in this depression. The drainage from the hills enters the valley throug'h a number of small spring-fed streams. None of these has sufficient energy to erode a valley through the sand, and their discharge merely accumulates in a series of swamps, which are drained by percolation through the sand into the river beyond. The swamps are not con- tinuous, but are separated by tracts of moist ground, originally prairie (the lower prairie of Pepoon, /pop.' 526) but now almost entirely under cultivation. The sand deposit is chiefly prairie, but a belt of forest lies along the river, and tongues and irregular areas of forest project out into the prairie, in some places e.xtending nearly across. Some of the forest and most of the prairie ha\-e l>een placed under culti\ation, but 27 extensive areas of each are still in their original condition, or but slightly modified by pasturing. Below Sa\anna, sand deposits of the same age again appear and continue intermittently down the Mississippi into Rock Island county, where thev connect with those of the Oquawka area described later. The geological origin of these two sand areas is known with con- siderable accuracy. Both are approximately contemporaneous and are derived from outwash from the Wisconsin glaciers. The method of deposition has been well described by Chamberlin and Salisbury (i88§: 261, 262), with special reference to the Hanover area. "The fringing deposits of glacial zuatcrs.—Outside the moraine lie two classes of deposits which gathered apace with it. The pre- cipitation which fell upon the western slope of the glacial lobe, to- gether with the water which arose from the same part of the glacier by melting, was shed from the edge, except the portion which may have found exit beneath in other directions and the portion lost by evaporation. Copious streams were doubtless the result. It is not difficult to understand that these, as they issued from the glacier, should have been exceptionally charged with silt, sand, and rolling stone, and that, as turbid waters, they poured down the channel-ways that were open to them. Long trains of glacial wash stretching away from the edge of the ice and leading down the several valleys tes- tify to the reality of such streams. "The most notable flood-train originating on the actual border of the driftless region is that which stretches down the valley of the Wisconsin River. The edge of the ice lobe crossed the Wisconsin in the western part of Dane and Sauk counties. In the immediate val- ley of the river the moraine is largely composed of gravelly constitu- ents, disposed in kame-like hills and ridges, or undulatory and pitted plains, showing the combined action of wash and push on the part of the glacier and its waters. Originating from this gravelly mo- raine, there stretches away a flood-train of gravel and sand, reaching down the valley to the Mississippi, and, there joining similar gravel streams originating higher up, it continues down through the drift- less area and l^eyond, though only remnants now remain. This val- ley drift originates at a height of about 90 feet above the present level of the Wisconsin River, and as it stretches down the valley gradually declines, so that, as it leaves the driftless region, it is barely 50 feet above the Mississippi. Near its origin coarse cobbles, bowl- derets, and even occasional bowlders are not infrequent. Farther down, the material becomes finer, and, in the lower stretches, only pebbles and sand are found. The lessening coarseness of the dejwsit 28 seems to show that as the glacial waters issued from the edge of the ice they were overloaded and struggling with a burden too great for their complete mastery; and, while they successfully carried the silt, sand, and even some of the finer gravel far down their courses, the heavier material in large part lodged near its origin and progressively filled the bottom of the channel. "This phenomenon, of which the Wisconsin Valley presents the only complete example lying entirely within the driftless region, finds other examples in several streams which cross the region. The Black River, the Chippewa, the Mississippi, and the Zumbro are all attended by such glacial flood deposits, which may be traced back to their origin on the face of the outer moraine. All these glacial flood plains slope more rapidly than the present streams. The train in the Chip- pewa Valley falls a little more than six feet per mile in the first 40 miles of its course, and over five feet per mile from its source on the face of the moraine to the Mississippi. In crossing the driftless area the glacial flood plain of the Mississippi declines about 50 feet more than the present stream." Their description applies as well to the sands of the Havana area, except that the source of the latter is the outwash through the Bloom- ington moraine in the vicinity, as already described by Hart ( Hart and Gleason, igoj: 139-144). The Aniboy Area.—This name is given in this report to the ir- regular complex of sand ridges and marshes along the Green river in Lee county, well illustrated in the vicinity of Amboy. Near that place the sand occupies a strip about four miles (6 km.) wide on the south (left) bank of the river. It lies usually in comparatively nar- row ridges from 20-50 feet (6-15 m.) above the intervening marshes. Back from the river the ridges are broader and the marshes propor- tionately more limited in size. Numerous small undrained ponds and swamps lie among the ridges. Near Amboy the ridges are either for- ested or under cultivation, but the number of prairie species occupy- ing the roadsides indicates that at least a portion of the sand was originally covered with prairie. Alternating areas of swamp and sand border Green river along its whole course through Lee, Bureau, and Henry counties to its junc- tion with Rock river, a distance of about 70 miles (no km.). They are to be regarded as outwash from the Bloomington morainal sys- tem, which crosses the south part of Lee county from northeast to .southwest (Leverett, tSqq: 277, 492, 493). The drainage of the whole valley is poor, and two large marsh areas, known as the Inlet Swamp and the Winnebago Swamp, are as yet not entirely reclaimed. 29 Probably the present local swamps are the vestiges of large continu- ous marshes which formerly extended the whole length of the river, and the hydrophytic plant associations now between the dunes are doubtless the survivors of an earlier swamp vegetation. Slow drain- age has pennitted the formation of extensive muck deposits, while in the Illinois river valley more rapid and complete dramage has merely left areas of a sandy loam between the dunes. The Dixon Area.—A small outlier of this general area, situated four miles (6 km.) southwest of the city of Dixon, is referred to un- der this name in the subsequent pages. This area is not forested, but the small marshes among the dunes indicate by their vegetation a close similarity to the rest of the area. The Oquaicka Area.—Below the mouth of Rock river the Mis- sissippi turns sharply to the south and follows a generally southerly direction for about 60 miles (100 km.). Through this portion of its course, from Muscatine, Iowa, to Ft. Madison, Iowa, its valley is well filled with sand deposits. These are probably chiefly a continu- ation of those along Green river, derived from outwash from the Bloomington moraine. It is possible that some of the sand is derived from the Wisconsin river outwash, as described above under the Hanover area. At the northern end of this area the principal deposits lie on the Iowa side of the river, where their vegetation has been briefly de- scribed by Pammel (iSgp). In Illinois the sand extends in a strip through the western part of Mercer and Henderson counties, lying usually close to the river, and gradually becoming thinner and less nearly continuous toward the south. A branch of the Chicago, Bur- lington and Quincy railway crosses the deposits between Aledo and New Boston and follows them south from Arpee to the junction with the main line at Gladstone. The town of Oc[uawka is situated on the deposits, and is a convenient location for the study of the sand vegetation. At the north end of Henderson county the sand lies in large, con- tinuous, nearly level areas, with here and there at wide intei^vals a low ridge. Its general height is 30-50 feet (10-15 m.) above the river. The ridges rise a few feet higher and near the Mississippi the river dune reaches a maximum height of about 100 feet (30 m.). Toward the south the sand lies in irregular, gently rolling ridges, not more than 30 feet (9 m.) high, and separated by areas of a sandy loam. South of Oquawka the deposits are broken by the Henderson river, but beyond it low ridges reappear and continue to the southern edge of the county. 30 A large proportion of the area has been forested, and most of that part is not under cultivation. Some fields have been cleared and abandoned, and are now densely covered with a thick growth of small trees. The portion originally covered with prairie is almost entirely under cultivation. Some large areas of blowsand occur, caused in many cases by pasturing or plowing. A conspicuous in- stance may be seen just south of Keithsburg, where the raihoad passes through a blowout complex, with one large traveling dune. The Kankakee Area.—This is undoubtedly the largest sand area rejiresented in the state, but at least three fourths of its total extent lies in Indiana. Leverett (iSpg: 328-338) has given a detailed ac- count of its extent and thickness, and from him the following state- ments are taken. The sand occupies a roughly semiciicular area, with the curved edge to the south. Beginning in western Marshall county, Indiana, the sand margin curves to the south and southwest near the Tippecanoe river, passes westward near the towns of Monti- cello and Kentland into Iroquois county, Illinois, and thence follows the Iroquois river north to the Kankakee ri\'er, which forms the northern boundary of the area. This area includes about 3000 square miles (7500 sq. km.). The deepest deposits lie near the Kan- kakee river, where the sand extends "several feet below the level of the base of the iidg"es." As in other areas the sand is not necessarily continuous. Es- pecially near the border of the area it is heaped into irregular ridges and dunes, probably caused by wind, and between them lie areas of sandy loam or muck. As in the Amboy area, the appearance indi- cates a slow recession of water, with the last of the hydrophytic vegetation still persisting. While the presence of the sand is cer- tainly due to glacial outwash, Leverett does not g'ive more definite conclusions. At the present time all the upland sand ridges are either forested or under cultivation, while the lowlands of peat, muck, or loam are occupied by swamp or meadow associations. Brief notes on the vege- tation, with maps showing the distribution of the sand in Newton and Marshall counties, Indiana, have been pulilislied by the Bureau of Soils (Neill and Tharp, ipo/; Bennett and Ely, 190 'i). But little attention has been given to this area during the present investigation. The Wmnehago Area.—This series of sand deposits lies chiefly in the northern part of Winnebago county, Illinois, and the southern part of Rock county, Wisconsin. It has been mapped and described bv the Soil Survey ( Bonsteel, igo^b; Coffey, Ely, and Mann, 1904). The sand lies between the \alleys of Sugar river and Rock river, in 31 level areas or low ridges with a generally east and west direction, and has a total extent, as estimated by the Bureau of Soils, of 25,088 acres ( 100 sq. km.). This sand differs essentially from the other areas described in its upland position. In some places it forms the bluffs of Sugar river, but in the center of its area it occupies the highest ground between the two river valleys and over 100 feet (30 m.) above them. Its position indicates that it is not of fiuviatile or lacustrine origin, as stated in the Soil Survey report. Leverett's account of its origin (iSpp: 131-138) is the most satisfactory, connecting the sand with the invasion of the lowan glaciers. The western border of the lowan glaciation enters Illinois at the valley of Sugar river, extends south along that river and southwestward along Pecatonica river to the western edge of the county, and thence east to the Rock river. The particular area of sand deposition is thus within the limits of the lowan glaciers, and the sand itself is regarded by Leverett as the drift of the lowan invasion. There is no present evidence of the recent existence of extensive prairies in the Winnebago area. Aside from a few deep depressions with a hydrophytic vegetation, the whole area is either forested or under cultivation. Climate The general climatic conditions of the northern and central parts of Illinois are shown in the following diagrams, taken from Henry (/pod). Dubuque, Iowa, is located on the Mississippi river just north of the Hanover sand area. Beloit, Wisconsin, is situated on the Wisconsin-Illinois state line at the eastern edge of the Winnebago area. Keokuk, Iowa, is on the Mississippi river, near the southern end of the Oquawka area. Springfield, Illinois, is in the central part of the state, east and southeast of the Havana area. Peoria, Illinois, is on the Illinois river, at the northern end of the Havana area. The first four stations, being located at the extreme edges of the general sand areas of the state, will indicate the extremes of climate at the north and south, and the conditions of the intervening region may be approximated by interpolation. Figure i shows that the seasonal dis- tribution of heat is of the continental type, with moderately cold winters and hot summers, and with occasional great extremes of heat and cold. It may again be mentioned that the official temperatures, taken under a shelter of regular pattern, do not reiirescnt the actual temperature to which plants are exposed. This is jjarticularly true of plants growing in exposed sand, where the surface temperature in 32 33 summer may exceed 130° F. (55° C). According to Mosier (igoj) the average date of the last frost in spring is April 29 in northern Illinois and April 21 in the central portion, while the first frosts in autumn occur on October 6 and 10, respectively. The average length of the growing season is accordingly from 160 to 172 days, depending on the latitude. D M M O N 34 DJFMAMJ JASON 70 60 35 The other four are probably responsible chiefly for the selection of a sand flora from those species which are located within invading dis- tance of the area in question. They certainly can not account for the sharp differentiation of the vegetation into definite associations. In this the plants themselves are most concerned, through their modi- fication and control of the physical features of the environment. If their control is lost, a successional series begins. In the field study, it was usually possible to recognize in the dynamic trend of the vege- tation the underlying cause. The descriptions which follow take up the subject from this aspect, discussing in more detail the effect of the dynamic environment, especially the wind, and its partial or com- plete control by the vegetation. General Discussion In the. field study upon which this work is based the observational method has been used almost exclusively. No apology or justification for this method is necessary, for direct obsei-vation has led in the past and will lead in the future to some of the most important results of plant ecology, and must always be the method by which the first ecological work in any region is done. The value of an exact knowl- edge of some of the physical features of the environment is evident, but their evaluation in an area of considerable size is a task not to be undertaken by one man or completed in a single season. Undue emphasis on the environment may lead to the partial neglect of the most important feature of a region, the vegetation. The plant itself is in many cases the controlling agent in the environment; the dif- ferentiation of definite associations is mainly due to the interrelation of the component plants; and the physical environment is as often the result as the cause of the vegetation. The relative importance of the plant covering and the physical environment is happily expressed by Spalding (ipog: 477, 479) : "But little reflection is needed to arrive at the conclusion that the classical question regarding the rel- ative importance of physical constitution and chemical composition of the substratum to plant growth * * * does not, and can not reach the heart of the problem. * * * Th'ts being the case, it would seem that in the future, investigations of the hal)itat relations, of the desert species especially, must be directed mainly to the plant itself. * * * The establishment of a plant in the place which it occupies is conditioned quite as much by the influence of other plants as by that of the physical environment." In the prairies of the sand deposits the two chief dynamic features 36 of the environment are wind, which tends to move the sand, and vegetation, which tends to stabiHze it. These two opposing forces are primarily responsible for the present location of eveiy association. In the forested portion of the sand the water factor is apparently the most important, but it depends chiefly upon the influence of the vegetation rather than upon any truly physical condition. In a broader way, the presence of a particular flora in the sand is due partly to the selection from the surrounding associations of various species with certain physiological requirements, and partly to climatic changes in the past. The latter can only be conjectured; the former are not known for any plants in this state and for veiy few in any place. According to these views, physical factors are relatively little concerned in the development of vegetational structures in this re- gion, while the demands of the plant and the efl^ects of its growth are of chief importance. The delimitation of the various associations concerned is also a matter which must depend, for the present at least, upon direct ob- servation. Jaccard {1Q02) has given a method for comparing dif- ferent associations and stating numerically the degree of difference between them. This has been used frequently and has given some comparisons which are interesting rather than important. It can not, however, be used successfully in the field. The chief difliculty in sep- arating associations lies in the idea of the association itself, which has never been expressed with sufficient clearness.* By some this idea has never been received \\ith favor. It is true that the distinctness of the associations is lost and their character greatly modified by the effects of civilization, but experience in natural conditions jus- tifies the statement that associations are definite organized units and that all vegetation is composed of them, either mature and fully differentiated or in process of organization (cf. Harper, igo6: 33, 34). It is as difficult to formulate a satisfactory definition of an association as of a species, and as unnecessary. For the present it may be considered that it is a homogeneous area of vegetation in which the interrelations of the component individual plants permit them to endure the physical environment. In this work the recognition of associations has been based upon the idea of uniformity, and those areas, whether large or small, * The concepts of the association as expressed hy authors are very variable and frequently conflicting. Some demand that each association shall occupy a definite habitat (Clements, 7905.- 292"), others allow a wide range in environment (Cowles, igoi: 79) ; some consider that the change of a single species affects the nature of the association (Harshberger, 7900; 652). Avhile others permit a large variation in the flora (Warming, 7909; 145, 146). i which are homogeneous (Clements, 1(^04: 11) throughout their ex- tent have been considered members of the same association. This uniformity is shown by the environment, by the behavior of the veg- etation, and, above all, by the plants themselves. In a region of limited size, over which the climate is essentially the same, the physical environment of an association is usually nearly constant, although instances are not lacking of an association living in the same area under widely different conditions (Cowles, ipoi: 79). On the other hand, it is not necessaiy that eveiy area with the same environment should be occupied by one association. It is reg- ularly the case in the sand region, and usually also elsewhere, that if the areas of the associations are conditioned by the environment, a considerable and observable change is necessary to influence the vege- tation (Clements, ipoj: 292). But in no case should the recognition and delimitation of associations be based upon the environment alone, which leads to a classification of habitat rather than \egetation ( Grad- mann, ipop) and may lead to the uniting of radically different types of vegetation.* The behavior of the vegetation with respect to ad- jacent areas is shown by successions which take place between them. If two areas with essentially the same environment show no succes- sional relations it is probable that they represent different consocies of the same association. But the first test of a plant association must be the vegetation itself. No two areas of vegetation are exactly similar, either in species, the relative number of individuals of each, or their spatial ar- rangement, and the smaller the areas to be compared the greater pro- portionately are the differences between them. Also, with continued and more detailed observation the importance of these minor varia- tions is magmified, and tends to lead to the recognition of an un- wieldy number of minor groups unworthy of the rank of association. This introduces the question of how great a variation may occur in the structure of the vegetation without the identity of the associa- tion being changed. Field work shows that the dominant and the sec- ondaiy species may both vary independently. In the same association the dominant species, if more than one, will have the same vegeta- tive fonn, as bunch-grasses, or trees, and will be of nearly the same size. Excessive development of one of them to the partial or com- plete exclusion of the others makes no change in the general ap- *Thc classification of associations by Clements (1905: 302, 303) is largely of this nature, and in some cases leads to tlie wide separation of closely related associations or even to the placing of a particular area in two different groups. Thus a hydrophytic sand-bar (cheradium) may be converted into a new xerophytic "formation" (syriidmm) merely by the fall of the water in the river. 38 pearance of the veg^etation and does not affect the jri-owth of the sec- ondary foniis. There are frequently no successional relations be- tween these local areas, or consocies. Areas characterized by dom- inant species of widely different appearance can not be reg'arded as belonging to the same association unless it can be shown that the areas represent transitory stages of development, as described below for the stabilization of blowouts. If the dominant species have the same general form, but do not tend to mix, except in the tension zone between them, and are accompanied by different groups of secondary species, the occurrence of different associations is suggested, as in the black oak and bur oak forests. The secondary species occupy a comparatively small area in the associations and their number usually depends in some way upon the habits of the dominant species. This is well illustrated by the bunch- grass association, in which the secondary species are absolutely de- pendent upon the dominant bunch-grasses. More species are con- cerned and their distribution is frequently irregular. These irreg- ularities, however, are seldom coincident with any variation in the dominant plants, but are caused chiefly by comjietition for space reg- ulated by seed dispersal and seasonal climatic fluctuations. A con- siderable variation in their quantitative distribution may be expected, unless the}' belong to the derived element of the association (see be- low), in which case they may indicate the lieginning or the end of a succession or some local change in the environment which is never- theless not sufficient to induce a change in the dominant species. Two areas of vegetation dominated by diffeient species are ac- cordingly probably consocies of the same association if ( i ) there is no obvious difference in their environments; (2) if there is no evi- dence of succession between them; (3) if the secondary species are the same for each; (4) if the dominant species are of the same vegetative form or (5) tend to mingle in other areas with the same environment and secondary species. A slight deviation from these criteria may be neglected if there is a preponderance of agreement with them, while a radical deviation would indicate that the areas represent distinct associations. In most cases fin the sand areas, at least) their application in the study of the vegetation leads to definite and unquestionable results. Whether small or large, associations usually contain some species which ai-e more characteristic of other areas. This dcrk'cd clcmeiit can be recognized only by comparison with neighboring associations, where the species in question are more numerous, irroi'e general in distribution, or more luxuriant in growth. They are least abundant J 39 near the center of the association, and tend to increase progressively toward its boundaries. The best idea of the structure of an asso- ciation is accordingly gained at its center. The presence of a derived element is well illustrated in the black oak association, in which every spot of unshaded exposed sand is occupied by interstitial annuals of the bunch-grass association, while near the margin of the forest numerous prairie perennials and grasses also occur. Many species, naturally, are almost equally typical of two or more associations. The boundary of an association is frequently sharp and well de- fined, especially if the dominant species of the adjoining areas are of different vegetation fonns, as between prairie and forest, or if the associations are correlated with considerable and relatively con- stant differences in the environment, as between the windward slope and basin in the blowout formation. In other cases the boundary is broad and more or less indefinite. This is particularly true if the dominant species are of the same vegetation form or if the environ- mental difference is fluctuating, as, for example, between the bunch- grass and the Panicum pseudopubescens associations. The \'egeta- tion of these transition zones is a mixture of usually indefinite and frequently highly variable character. The species in them should be referred as far as possible to their respective associations, and not al- lowed to modify the ideas of structure gained from an examination of more typical localities. Besides these transitions in space, there are also transitions in time. An early stage in the development of an association may re- semble but little its mature condition. Certain members of the asso- ciation with excessive seed production, with more mobile seeds, or better adapted to the somewhat aberrant environment, appear first and for a time dominate the area. Thus, in the stabilization of a blowout, the redevelopment of the bunch-grass association begins with a growth of Lespedesa capitata and Oenothera rhombipetala in large quantities. This condition lasts but a short time before they are replaced by the usual bunch-grasses. Such an area is at first sug- gestive of a distinct association, but examination shows that it has no species, aside from relics of the preceding vegetation, not found also in the bunch-grass, and that the environmental conditions are very similar to those of the spaces between the bunches of grass, where these interstitials (p. 54) grow. A knowledge of the habitat prefer- ences and habits of the component species and of the general dynam- ics of the area is necessary to decide ui)on the proper classification of these transitional stages. In estimating the unifonnity of the vegetation, direct observation 40 is in many cases satis factoiy, especially if the associations are small in area or the component plants low in stature so that a comprehen- sive view of them may be taken. In other cases, simple lists of species, taken in each area of the association, may be compared, and their similarity is a good index. For more accurate work, the quad- rat method proposed by Clements (iQO^: 161-170) may be employed. It gives excellent results but demands much time and labor. In areas of closed vegetation it seems to have its chief value in expressing, rather than detemiining, the structure of the association. A modifi- cation of the quadrat method has been tried with success in this work. It consists in listing, in the approximate order of the space occupied by each (not the number of individuals), the species on an imaginary quadrat of about four square meters situated directly in front of the obsei'ver. Stepping forward two paces brings another quadrat to view, and a series of ten or twenty, extending in a continuous strip or scattered throughout the association, may be listed in a short time. The size of the quadrat used is chosen to suit the character of the vegetation ; two meters square seems adequate in the study of prairie associations. In a forest a quadrat of that size could be used only for the herbaceous vegetation, and one ten meters square would be neces- saiy to show the nature of the forest cover. Quadrats of such size are unwieldy, and in practice it has been found that results are more easily obtained by counting every tree within five meters of the ob- server as he walks through the forest. A new list may be made for each hundred meters or for any area with distinct environment. In investigating the tension zone between associations the tran- sect method (Clements, ipo^: 176-179) may be used, but is subject to the same limitation as the quadrat method. Good results may b; conveniently obtained liy walking back and forth repeatedly from one association to the other, listing the species in the order of their ap- pearance. Carefully conducted studies, as indicated abo\'e, show that the dominant species are unifomily distributed over the whole area of the association or consocies, and that the floral discrepancies are caused by a number of comparatively rare species represented usually by a small number of individuals. The weakness of the whole method lies in the fact that, in a mere list, a rare species, possibly a single in- dividual, is given as much weight as a common one. Actual counts of the individuals of each are diftrcult to make and may give mis- leading results. If each species could be correlated with the pro- portion of the area which it occupies, it would be demonstrated that most parts of an association are highly similar in structure, and the 41 resulting community coef^cient of different areas would probalily be above 0.900 (in the black oak association virtually i.ooo). Unfor- tunately no practical method for this has been devised. The more widely the different areas of an association are sepa- rated, the greater are the floral discrepancies. The dominant species, however, remain constant, and the change lies almost wholly in the secondary species. Many of these are the results of selective migration from neighboring associations, so that a variation in the general nature of the vegetation of an area aft'ects the specific struc- ture of each association. This phenomenon has been discussed briefly by Warming (igo^: 145, 146) under the name of geographical va- riation. It is well illustrated in the sand areas of Illinois by the sec- ondary species in the black oak association. In the Havana area are found some typically southern species, as Quercus marilandica and Galium pilosiiiii, while in the Winnebago area some species of northern or eastern distribution occur, as Pyrus aiuericaua and Lupinus perennis. In comparing areas of such wide geographic separation emphasis must be placed upon the dominant species, which are the fundamental cause of the g"eneral physiognomy of the asso- ciation. The areal distribution of an association may be compared to the distribution of a species. Both are irregular in outline, although coextensive with certain combinations of environmental factors. Both consist of scattered members, independent of each otiier, l)ut related by a common genesis and common demands upon the environ- ment. Both show minor local and broad geographical varieties. The fonner are illustrated in the association by the consocies; the latter, in the species by the subspecies, which in their typical form occupy outlying arms or peninsulas but toward the center of distribution intergrade with the main body of the species. Taxonomic work has shown that the interpretation and classification of these forms is a matter of great difficulty. Proper treatment of the geographical varieties of an association will be a matter of much greater difficulty, since the necessary comparisons must be based entirely upon written description or photographic record. Because of this geographical variation and consequent difficulty of comimrison, few correlations of associations in different parts of America have been made or attempted. Ecological literature con- tains numerous descriptions taken from the few representatives of the associations in a limited locality, I)ut as yet no one has given a gen- eral description of an association, compiled from observations talten 42 throughout its range.* In this respect the present status of systematic ecology resembles that of pre-Linnaean taxonomy, a maze of de- tached facts waiting for a Linnaeus to collate and correlate them into a foundation for future investigation. The associations recognized in the field have been grouped into formations, characterized partly by uniformity in the physiognomy of the vegetation, and partly by uniformity of environment, to which the physiognomy is in some extent due. Formations correspond somewhat to genera in taxonomy, and like them may be limited or comprehensive in their scope, this depending solely upon individual opinion. As far as possible they have been made to coincide with the popular idea of the different types of vegetation. The four forma- tions are all generally known through the sand regions and given the names used here, with the exception of the forest, which is col- loquially known as "timber" or, in some places, as "black-jack." The latter term applies to the particular association rather than to the forest formation in general. The areal extent of a fomiation is ap- proximately coincident with one of the phytogeographical provinces of North America, and fonnations with the same distribution are placed in the same province. It is thus seen that the differentiation of both minor and major ecological groups depends principally upon the plants themselves, the associations being distinguished by the specific composition, the formations by the general appearance, and the province by the distribution of the vegetation. This is an ex- tension of the idea already expressed, that the most important feature of the association is not the habitat but the plant. It is belie\'ed that the regional classification of associations is really genetic and dynamic, bringing together those which are most closely related by origin and succession. The Vegetation The area covered by the state of Illinois occupies a unique posi- tion in respect to the vegetation of the continent, marking the reg-ion of closest approximation of four great floral and vegetational prov- inces. (See maps in Schimper, jpo?; Engler, /po_'; Transeau, /poj, ipofi; Merriam, 1898; Sargent, 1884.) The Austroriparian Province (Merriam, 1898: 45) enters the state at the extreme southern end, and well-developed examples of its dominant hydrophytic vegetation, the cypress swainp (Ta.rodhim distichum), extend northward into the lower valley of the Wabash *The nearest approach to this has been made by Tra'nseau (1903: 190S-6) in his studies of bog floras. J 43 river. Scattered species of Austroriparian affinity extend north in ever decreasing numbers, for some 300 miles (500 km.). One of the most conspicuous plants of this nature is Carya illUwensis, the pecan, which follows the alluvial l30ttom-land of the Mississippi river as far as southeastern Minnesota. Few species of this group occur on the sand deixjsits of northern and central Illinois, although a num- ber occupy the sand-bars of the lower Mississippi. The southern boundary' of the great Northeastern Conifer Prov- ince of the north and northeast passes southeastward across Wis- consin and Michigan, and numerous species persist south of this line. Definite but isolated associations of Piniis Strobus and of Lari.v laricina, each with its usual attendant species, are found in various places in northern Illinois, and many scattered species of northern range, such as Popuhis tremiiloides and Bctula alba, van papyrifera, live in associations of other provinces. Some of them are concerned in the vegetation of the sand deposits. Between the Austroriparian Province on the south and the Northeastern Conifer Province on the north there are extensive plains, reaching from the base of the Appalachian mountain system on the east to Nebraska on the west. This area is known as the Deciduous Forest Province, and is occupied, as its name indicates, by deciduous forests, with Qiicrcus, Acer, Fraximis, Tilia, Fagus, Nyssa, Liriodendron, Acscuhis, and Carya as some of the leading genera. In the eastern part of the province the forest is almost con- tinuous, broken only by minor associations of an edaphic nature. At the west, from Indiana and Illinois to Nebraska, it becomes discon- tinuous, and a portion of the area, becoming proportionately larger westward, is occupied by the prairies. The Prairie Province, last on the list, extends in a long .'^trip north and south through the Great Plains at the eastern base of the Rocky Mountains from Texas to Saskatchewan, and an eastward extension passes across Iowa and Illinois into Indiana, sharing the area with the deciduous forests (Pound and Clements, iSgS). Throughout its whole area the dominant vegetation is prairie. Each of these four provinces is composed of many plant asso- ciations, which occupy usually definite habitats, and which are related to each other Ijy certain successional trends. In each there are asso- ciations occupying limited areas of extreme environment, and these tend to converge, through the effect of various physiographic and biotic agencies, toward the dominant or clima.x vegetation of the region. In each of the provinces the successional extents in the es- tablishment of the dominant vegetation are relatively simple. There 44 is in every case at least a hydrophytic and a xerophytic extreme, forming two general converging lines of succession. In our present knowledge of the subject, it is impossible to state whether there is one definite climax association in each province ; it seems probable that there are several such associations, each characteristic of a lim- ited portion. It is certain that in each province there is a dominant formation, or type of vegetation, deciduous forest, coniferous forest, or prairie, as the case may be. Present evidence seems to indicate that the nature of the dominant type is determined by a long chain of his- torical factors (Adams, Ip03, ipoj) and its present areal distribu- tion by the broader existing climatic factors, notably heat and rain- fall (Transeau, ipo^). The boundaries of the four provinces have been subject to great chang'es in the past, both during and following the glacial period, as the ice swept to the south, overthrowing the previous conditions of climate, and then retreated to the north, uncovering unoccupied ground and throwing it open to plant invasion. The ensuing move- ments of vegetation were among the greatest in the history of the continent, and have been of the greatest moment in determining the present distribution of the biota. These movements have by no means ceased. They are merely less obvious when measured in terms of years and centuries rather than in geological periods. Even now a biotic migration is in prog- ress, which is probably the direct continuation of early postglacial movements, and is doubtless as rapid and as far-reaching in its effects as any of the past. In the present migration the vegetation of the Deciduous Forest Province is the chief factor. It is now pushing out its boundaries to the north and west and enlarging its area at the expense of the Northeastern Conifer Province on one side and the Prairie Province on the other. Some detailed features of the northern extension have been given by Whitford (ipoi), Transeau (igof;-o6), and others, and summarized by Adams (ipoj). The westward migration has been mentioned by many, but scarcely described in detail. The actual steps in the migration of the vegetation are due to a series of successions, by which associations of the pr-airie or of the coniferous forest ai-e replaced by others, with similar environmental demands, from the deciduous forest. Some of the noi-fhem and western associations are succeeded with comparative ease; others resist succession for long periods of time. Because of this the forest extends north and west, not in continuous masses but in long tongues and detached bodies, while relics of the fonner vegetation lag" be- 45 hind as isolated areas in the midst of the forest. Rehcs of the northern coniferous forest persist in Illinois as tamarack swamps and groves of white pine, and both are frequently termed "boreal islands." In a similar way the detached areas of prairie in Illinois may be re- garded as western relics, although they are often miles in extent. The oldest relics, that is, those toward the east or south, are regularly smaller in extent and more mixed with forest species (Bonser, /poj). In the migration of the deciduous forest associations, the greatest advance has always been made in those habitats which most nearly resemble those occupied by the climax formation, and which are therefore most nearly suited to the invading vegetation. On the other hand the relic associations have been left behind in those habi- tats, not necessarily best adapted to the relic vegetation, which are least suited to the invaders. For this reason the boreal associations in Illinois are limited to sandstone hills and to undrained swamps, while the prairies persist chiefly in the upland soils between the stream courses. The successions by which the general migration is consummated are of a type different from that found within the formation and leading merely to the dominance of the climax xeg'etation, since they involve associations of two and sometimes of three provinces. While several descriptions of this type have been published, general con- clusions have not usually been drawn. At the present time it can only be stated that the succession seems to take place between equiv- alent members of the different provincial successional series. Thus, as shown in the following pages, the xerophytic extreme of the prairie, the bunch-grass association, tends to give way to the corre- sponding extreme of the forest, the black oak association. In a sim- ilar way the black oak association may succeed the xerophytic ex- treme of the Northeastern Conifer Province, the jack-pine associa- tion. We find similar relations between the hydrophytic extremes, and in Illinois the succession of the northeastern tamarack associa- tion by the deciduous Ixjttom-land forest may I)e observed. Northern Illinois, therefore, has been and is the scene of im- portant events in the biogeographical history of the continent. The following description of vegetation is designed to be not merely a dis- cussion of static conditions, but rather a portrayal of one phase of this great vegetational movement and of the consequent struggle for supremacy which is still being waged. In the vegetation four distinct formations, or types of vegetation, have been recognized. Each of these consists of several associations, characterized by a distinct group of plants, l)y a distinct habitat, or by 46 both. The subjoined tabular view will express the classification used, while the arrangement of the associations in the descriptive matter follows as nearly as possible their successional relations. Classification of the Plant Associations A. The vegetation is dominated by grasses, occupying a relatively stable habitat with low? water-content. The secondary species occupy the interstices between the stools of grass. j^^ Prairie Formation of the Prairie Province. a. The movement of sand is slow; several species of bunch-grass are pres- ent and many secondary perennial species. The Bunch-grassI Association. b. Sand movement is more rapid ; the area is dominated by Panicuin pseudopubcscens and the secondary species are chiefly annuals. The Panicum psciidopiibesccns Association. B. The vegetation is very sparse and open, occupying usually a very unstable habitat due to rapidly shifting sand ; there is little distinction between dom- inant and secondary species. ^^^ g,^^.^^^^ Formation of the Prairie Province. a. The sand movement is chiefly due to removal by gravity; the vegeta- tion consists of relic grasses and perennials. The Windward Slope Association. b. The sand movement is due chiefly to removal by wind ; the e-xtremeiy sparse vegetation consists of deep-rooted perennials. The Basin Association. c. The sand movement consists chiefly of a mere redistribution or of gradual deposition ; the vegetation is composed chiefly of annuals. The Blowsand Association. d. The sand movement consists chiefly of a mere redistribution; the vege- tation is dominated by the sand-binding perennial, Hudsonia tomentosa. The Hudsonia Association. e. The sand movement consists chiefly of deposition; the vegetation is composed of sand-binding perennials with accessory annuals. The Deposit Association. f. The vegetation occupies a fossil soil, uncovered by continued sand move- ment ; the dominant species is Stenophyllus capiHaris. The Stenophyllus Association. C. The vegetation is dense, closed or nearly so, with grasses only as secondary species ; the sand is stable, with usually relatively high water-content ; the formation is developed in deep depressions. The Swamp Formation of the Deciduous Forest Province. a. The vegetation is semixerophytic, characterized by slender perennials. The Solidago Association. b. The vegetation is truly mesophytic. 1. The vegetation is dominated by willows, -ri o i- a • ^•The i>ahx Association. 2. The vegetation is dominated by mat-forming mosses. The Polytrichum Association. c. The vegetation is hydrophytic The Swamp Association. J 47 D. The vegetation is dense and closed, dominated by trees or avevectent shrubs, or by herbs in the immediate vicinity of shrubs ; the sand is stable, with usually low water-content. The Forest Formation of the Deciduous Forest Province. a. The vegetation is dominated by herbs. ^,^g Smilacina Association. The Physalis Association b. The vegetation is dominated by shrubs, with numerous lianes, the sec- ondary species are of a mesophytic type. 1. On the crests of dunes or other areas of deposition. The Dune Thicket Association. 2. In blowouts The Blowout Thicket Association. c. The vegetation is dominated by trees. 1. The secondary species are generally xerophytic ; avevectent shrubs or lianes are few or absent; the leaf-mold is thin or absent. The Black Oak Association. 2. The secondary species are generally mesophytic; avevectent shrubs and lianes are abundant ; a superficial layer of humus is de- veloped. a. The dominant species are bur oak and white oak. The Bur Oak Association. b. The dominant vegetation is composed largely of black oak, but with numerous other arborescent species. The Mixed Forest Association. The; Prairie Formation THE bunch-grass association The bunch-g"rass association formerly occupied probably more than nine tenths of the unforested portion of the sand areas. It ex- tended over hill and dale, interrupted only by the blowouts and their related associations, and was by far the most important association of the unforested area. Monotonously uniform floristicaUy, its eco- logical structure showed an obvious differentiation into several con- socies, each characterized by the preponderance of one or a few species of grass, and often sharply distinct from its surroundings. These are considered to be consocies instead of associations because they can not 1ie referred to any apparent difference in the environment. and because they exhibit no successional relations to each other. The best development of the bunch-grass association was, and is, in the Hanover sand area. By far the larger portion of the area was originally unforested. Large fields are still in a virgin condition, and hundreds of acres have been but little pastured. The area in- cludes most of the consocies described and offers without doubt the best conditions for ecological study. The Winnebago area includes, so far as observed, but one small area of bunch-grass, not more than an acre in extent, entirely surrounded by forest. It is evidently a 48 relic of a foimer wider extension ot the association. In the Ambo>' area most of the country is either forested or pastured, and the only observed examples of bunch-grass were scattered fragments along the roadsides. The Dixon area formerly contained much bunch-grass, but it is also now largely under cultivation. The Oquawka area is more extensively forested, but some of the bunch-grass still remains in the original condition. The Lcptoloiiia cognatum consocies is es- pecially well represented there. The bunch-grass association formerly occupied thousands of acres in the Havana area, but most of it is now under culti\ation. In the three chief sand areas, at Hanover, Oquawka, and Havana, the sand deposits lie, as has already been noted, on the east side of a river, extending from the water's edg"e to the bluff. The bunch-grass association is always separated from the river by a nar- row or wide marginal forest, but may extend inland to the very base of the bluffs, as at Hanover. It may then be divided into smaller areas by trans\'erse belts of forest, as at Havana. To these smaller tracts local names are sometimes g'iven, as Benton Prairie at Oquawka. The tracts thus delimited are not unifomi, but each may be occupied by two or more consocies. The different prairies of a sand area are, however, occupied in general by the same consocies and have the same flora. But two noteworthy species seem to form an exception to this rule, Brcivcria Pickcringii in Benton Prairie at Oquawka, and LcsqncrcUa argcntca in the Devil's Neck region of the Havana area. The bunch-grasses which give the association its name produce at the base or along- the lower portion of the culm a number of lea\es, which are aggregated into loose or crowded bunches, depending upon their size and number. Rising from their center are the flowering culms, and beneath the living leaves are also the dried dead leaves and culms of the previous season. The height of the bunches, ex- clusive of the culms, is therefore, in most cases, approximatelv equal to the length of the basal and lower leaves. In simple Inmches all the leaves and culms radiate from one center, and a bunch con- sists of one plant, or rather of one stool. The diameter of the bunch is then not more than twice the length of the basal leaves. Such simple bunches are exhibited by Paniciim pcrlonginn and Stipa spartca. With some other species, as Paniciim pscudopiihescens, the culms are also spreading or horizontal, and the diameter of the bunch is about equal to twice the length of the culms. In other cases the individual plants are closely associated, so that the dense bunches may reach any diameter, and are usually very irregular in shape. This & i 49 habit is well illustrated by Lcptoloiiia cognatitin. The bunches of each species are distinct in size, structure, and general appearance, and when in a sterile condition can frequently be recognized by their habit alone. Notes on the individual character of the bunches will follow. The living and dead leaves of the bunches cover the ground in most cases so closely that other plants can not grow among them. The two bunch-forming sedges, Carex Muhlcnbergii and Cypenis Sclra'ciiiitzii, alone produce bunches so loose that various annuals usually grow within them. Stipa spartea also produces loose bunches through which Ambrosia psilosiachya or Teiicnniii occidentalc may grow. A number of small annuals may be found between the radiat- ing culms of Paiiicitin pseitdopiibcscens at some distance from the cen- ter, while the dense compact bunches of Koclcria crisfata and Lcpto- lonia cognatuiii are entirely free from other plants. Besides restricting the growth of other species, and thus retaining the dominance in the association, they act efficiently in preventing the blowing of the sand. The greater proportion of the surface is usually entirely covered, and the small intei"vening spaces are so narrow that the sand is not easily lifted by the wind above the bunches. The tend- ency to blow, if present, is usually shown by the slight elevation of the grasses above the concave or trough-shaped interspaces. Neverthe- less, blowing may sometimes take place to such an extent that the whole association is destroyed, and succeeded l>y another in which Pauicuiii pscudopnbcscens is the dominant grass, as will be described later. It seems probable that in most of these cases the density of the plant covering has been reduced by pasturing or other recent causes, or, conversely, that under strictly natural conditions the bunch- grasses permanently prevent blowing. In some places the surface is entirely covered, either with bunch- grasses alone or with mat-plants in addition, and there is every gra- dation down to cases where but little more than half the actual surface is occupied. It may be arbitrarily assumed that the bunch-grass as- sociation can not exist with more than half the sand exposed, and it is certain that it may disappear with even more of the surface oc- cupied. The proportion of the ground covered by the grasses varies with the species, the habitat, and the stal)ility of the sand. Of the grasses which tend to cover a relatively small part of the surface Koclcria cristata and Andropogon scoparius are good examples, while the bunches of Leptoloma cognatmn show especially a tendency to become confluent and to cover large unbroken areas. The consocies which contain the largest numlser of species of bvinch-formers are also apt to occupy the space most completely. 50 The general appearance of the association, inckiding especially the color-tone and number of secondary plants, depends almost en- tirely upon the specific peculiarities of the bunch-grasses represented and upon the density of the covering. Most bunches are so distinct in size, density, or other features that they are easily recognized, even when sterile. In doubtful cases minor morphological characters may be used, such as pubescence, the structure of the ligule, and other similar vegetative features. Some of the most important bunch- forming species are the following. 1. Koeleria cristata.—Bunches regular, compact, about one foot (3 dm.) in maximum diameter and eight inches (2 dm.) high, with a consideralile accumulation of dead leaves beneath them, forming an elevated central tuft and radiating on the sand; leaves six to ten inches (15-25 cm.) long, mostly straight and erect, glaucous-green or canescent with fine pubescence. The regular close bunches of Koeleria have an appearance of trimness and neatness in which they excel any other species. The gray-green color and the shining spikelike panicles make the grass very conspicuous, especially during the aestival aspect when it is in bloom, or at any season when the dew is still on it in the early morning. The bunches are rarely confluent and tend to leave a con- siderable uncovered area between, especially when not associated with other species. 2. Leptoloiua cognatinn.—Bunches 8-12 inches (20-30 cm.) wide and about eight inches (20 cm.) high, very compact, close and dense, flat-topped, frequently confluent in large irregtilar patches ; leaves short, all erect or radiating-, and freely mixed with the dead leaves of the preceding season, g"iving the whole bunch a yellow-gray ap- pearance. The short leaves are more irregularly arranged than those of Koeleria cristata, and the dead leaves and culms remain for a long time mixed with the living. In the serotinal season the large, but A-ery lax, red-flowered panicles appear and impart a distinct red- dish hue to the consocies in which the plant grows. 3. Stipa spartea.—Bunches loose, few-leaved, but regular in size, I- 1. 5 feet in diameter and about the same height, with a slight ac- cumulation of dead leaves and culms on the sand beneath. The flow- ering culms rise to a height of three feet (i m.). Of all the bunch- forming species of g-rass in the association this species fonns the loosest and most indefinite bunches. 4. Paniciim pseiidopiibesccits (PI. Ill, Fig. 2).—Bunches irregu- larly circular in outline, depressed, 1-1.5 feet (3-5 dm.) in diameter, four to six inches (1-1.5 dm.) high; culms and leaves radiating from 51 the center, straight, barely exceeding the dead cuhns with their spHt and curled leaves. The culms and especially the spikelets are red in color and give a reddish tone to the whole bunch. This species forms one of the most distinctive bunches of the association, due to the depressed or prostrate radiating culms with their erect or almost appressed leaves, and to the persistence on the dead culms of the recurved leaves of the previous season. The culms extend beyond the leaves, and bear small, but conspicuous, panicles of red spikelets. This species is more characteristic of the association to which it gives its name, but is also frequent in the typical bunch-grass, where its peculiar habit makes it conspicuous. 5. Boiitcloua hirsnta.—Bunches low, irregular, two to four inches (5-10 cm.) high, usually confluent in matlike masses 5-12 inches (1-3 dm.) in diameter; leaves short, irregular in position, forming a loose tuft, conspicuously gray-pubescent, and giving a gray color to the whole bunch. The slender culms, 4-12 inches (1-3 dm.) high, appear during the late aestival season. The small bunches are entirely too low to compete with the other grasses for space or to constitute a conspicuous element in the association. They are usually restricted to the intervening spaces, where they have the general habit of mats rather than of bunches. They associate frequently with Sclagun-lla riipestris. 6. Boutcloua ciivtipcndula.—Bunches loose, 6-12 inches (1-3 dm.) in diameter, eight to ten inches (2-3 dm.) high; leaves mostly all erect, six to eight inches ( 15-20 cm.) long. 7. Cypcnis Schivcinitdi.—Bunches very open and loose, basal leaves few in number, ascending; culms several, erect or ascending. The plant frequently has the habit of an interstitial rather than of a bunch-grass. 8. Andropogoii scoparius.—Bunches one to three feet (3-8 dm.) wide, circular, 1-1.5 feet (3-4 dm.) high, compact, and regular; leaves very long and narrow, erect or ascending, the dead leaves persisting as a dense mass at the base, or recurved around the margin of the bunch; culms about two feet (6 dm.) high, the dead culms persisting through the following summer. The bunches are notable for their large size and the dense mass of dead leaves mingled with the li\-ing ones. As the bunches grow older the center dies, and rings are formed which reach a maximum diameter of over a yard (i m.). In such rings the zone of living grass is three to eight inches (1-2 dm.) wide, and the central portion is elevated four to six inches (10-15 cm.) above the general level. It is composed of a dense mass 'of old roots anfl culms mingled with 52 debris of all kinds, and is almost always devoid of any plant growth. 9. Andropogon furcatus.—The bunches of this grass, commonly known as bluejoint, resemble those of the smaller A. scopariiis in general habit, but are taller, 1.5-2.5 feet (4-8 dm.), and frequently larger in diameter, three to four feet (8-12 dm.). The leaves are larger, less densely aggregated, and without the tangle of dead leaves among them. The flowering culms are three to five feet (10-15 dm.) tall or even more, and seldom persist until the following summer. Like A. scoparins, the bluejoint may also fonn rings by the death of the center of the old bunches. These are five to seven feet (15-20 dm.) in diameter and without a conspicuous elevated center. SorgJiastnim nutans and Paiiicum inrgatiini {PI. IV, Fig. 2) form large bunches much resembling those of bluejoint. Calamovilfa longifoUa grows in patches with the individual culms one to six inches (3-15 cm.) apart, forming dense clusters which resemble true bunches. Bragrostis trichodes produces bunches closely resembling those of Andropogon scoparins in general character, but without the mass of dead leaves. The bunches of Panicnm perlongum are very regular, hemispherical in shape, and composed of a dense mass of straight radiating leaves. In general appearance they resemble the bunches of Koclcria crista ta. Paspalimi sctacciiin and Bragrostis pectinacea send up several culms from a common center, on which the leaves are most closely approximated near the base, thus forming a loose irregular bunch. The loose open bunches of Carcx Muhlcn- bcrgii are especially characterized by their leafless, obliquely ascend- ing stems. Since the bunch-grasses virtually exclude other growth beneath them, the secondaiy species are found upon the small aixas of bare sand between the bunches. They may be conveniently divided into four ecological groups based upon their habits and structure. As in most ecological classifications, these groups are not entirely distinct, and some species are of doubtful position. To them may be given the names perennials, mats, interstitials, and parasites. The members of the first group, the perennials, are generally very deep-rooted, and frequently grow in tufts or bunches resembling those of the bunch-grasses. The deep roots are a response to the conditions of water supply, and the bushy habit is possibly correlated with the generally xerophytic environment and exposure to the wind. Re- sembling the bunch-grasses in halnt, thev are able to compete with them for space, and may be found in the center of a' patch of grass, where they lia\'e persisted since the grass surrounded them. Their 53 competition with the bunch-grasses is defensive rather than offensive ; they can resist the encroachment of a grass, but are not able to dis- place it. Some typical plants of this habit are Aster linariifalius, Lithospermiiin Gviclini, Aster scriceiis, Tephrosia virginiana (PI. IX, Fig. i), and Chrysopsis znllosa. Others have more slender stems, several of which arise from a common base and spread divergently, somewhat resembling in habit the looser bunches of Carex Muhlen- bergii or Cypertts Schweinitzii. Good examples of this type are furnished bv Callirhoc triangulata, Petalostenmin purpureuni, and PetalostciiiHiii candiduiii. Still others have erect stems which tend to grow in clusters, as Coreopsis paliuata, Solidago iiiissonrieiisis, Solidago iiciiioralis, and Helianthemum iiiajns. A fourth type is fur- nished by Physalis virginiana, Baptisia bracteata, and Tradescantia reflexa, with solitaiy stems which branch freely or bear widely spreading leaves toward the top. A fifth type is illustrated by Bu- phorbia coroUata or the species of Liatris, whose slender erect stems grow singly and occupy very little g'round space. This type ap- proaches most nearly the third group of interstitials. One member of the group, Breiveria Pickcringii, has veiy numerous long decumbent stems, fomiing an elevated mass at the center, and spreading out in all directions on the sand. The shrubs of the association are for convenience classified in this group. They include Rhus ca)iadensis, var. illinoensis, forming dense rounded masses 3-15 feet (1-5 m. ) across and three feet (i m. ) high, and excluding all other vegetation; Amorpha canescens and Ccanothus americanns, undershrubs with several erect or ascending stems one to three feet i},-'?' dm.) high; and Ccanothus ovatus (PI. II, Fig. 2), with se\-eral ascending stems forming an irregular bushy shrub two or three feet (i m.) tall. The two species of Ceanothus are notable for their immense woody roots, frequently si.x inches (1.5 dm.) in diameter and extending downward to great depths. They are crowned by a few live stems, which are of comparatively short life, and with the dead and decaying bases of many others of previous years. It is needless to say that the vast majority of these plants present obvious xerophytic adaptations, the most general of which are a re- duction of surface to narrow or small leaves, and a protective cover- ing of silvery or canescent hairs or scales. Their general tone is grayish green, amid which the vivid green of Euphorbia corollata and Tradescantia reflexa appears strangely out of place. The various types described do not include all the species of the group, but omit some of the less frequent. Neither do all perennials belong to this 54 group, but some, as LcsqiicrcUa argcutca, are placed among the in- terstitials. The second group inckides the mat-plants, a small group with but three flowering plants, Opitntia Rafiiiesqnii, Opiintia fragiiis, and An- tennaria sp. Selaginella riipestris is also common in the Hanover area. These grow close to the sand and tend to spread annually over a larger area. They are unable to encroach upon either the bunch- grasses or the perennials, and do not survive when covered by mem- bers of these groups. Their number is accordingly largest in the more open consocies. These plants are of the greatest importance in binding sand, and under certain conditions have a prominent part in stabilizing- blowing sand. Selaginella riipestris is especially note- worthy for its habit of circular growth. Extending vegetatively from the center, its growth is so regular that a circular patch is formed. This is soon converted into a ring by the death of the center, leaving a marginal zone of living plants one to two inches (2-5 cm.) wide. This ring gradually increases in size until it may reach a maximum diameter of four feet (1.2 m. ). Additional rings may begin within an old one, or parts of two rings may overlap. Megaspores are produced in enonnous cpiantities, but their successful growth must be rare. The prickly pear, Opiintia Rafinesquii, is much more common in the Havana area than in any of the others. The mats of Selaginella are favorite places for small mats of a dark col- ored crusty species of Cladonia. Small mosses, of unidentified species, are also frequently found, and in many ]Dlaces a dark-colored crust on the sand is fonned by a species of Oscillatorla, which from its habit may also be classified with the mats. The third group, the interstitials, is composed in general of an- nuals, with slender, frecjuently imbranched stems, generally narrow leaves, and fibrous roots. They come up late, principally during the season of heavy rainfall in June or July, and cover the bare areas of sand with prodigious numbers of individuals. Notwithstanding their number, they are of the least ecological importance. Their slender stems occupy little space and take no part in sand-binding, while the veiy existence of the entire group is due to the presence of the bunch- grasses, which act as windbreaks and hold the sand. If the luimber of grasses decreases somewhat, there is a correspondingly larger number of interstitials, but if the bare spots become too large, so that blowing of the sand begins, their number begins to decrease. The most abundant species of interstitials are Oenothera rhonibi- petala, Ambrosia psilostachya, Linaria canadensis, Cassia Chamae- christa, Monarda punctata, and Crofon glandiilosiis, var. septen- trionalis. 55 The fourth group, or parasites, is represented by a single species of seed plant, Orohanche fasckiilata, found on the roots of Artemisia caudata in the Hanover area. While these four groups are sufficiently distinct to serve as eco- logical units, they are not absolutely separate. The perennial Les- pcdcza capitata, with its slender stems and narrow leaves, associates frequently with the true interstitials, and might then well be regarded as one of them. Cypenis Schweinitzii appears now as a bunch-grass, now as an interstitial. Bouteloua hirsiita behaves sometimes as a bunch-grass, producing small tufts two to four inches (5-10 cm.) high, but frequently it functions more as a mat and associates with Sclaginella rupestris and Antennaria sp., or, when growing between larger bunches of Koeleria cristata, it might be regarded also as an interstitial. The number of secondary species and indixiduals is naturally greatest in the more open parts of the association and least in the denser portions. The mats may entirely disappear and the inter- stitials be greatly reduced in number when the bunch-grasses are closely aggregated. The perennials, with their greater resistance to crowding by the grasses, remain throughout and always occupy a prominent place in the association. The close relation between the secondary and dominant plants of an association is seldom better illustrated than in this one, where the presence and disappearance of the interstitials are both correlated with the density and luxuriance of the bunch-grasses. The association as a whole is, as already noted, divided into a number of consocies. Some of these are characterized by a single species of grass and may be called pure consocies. Such are those characterized by Koeleria cristata, Lcptoloma cognatiun, Stipa spartca, and Carex Muhlcnbcrgii. A larger part of the association is occupied by several characteristic species and is here termed the mixed consocies. Although the specific composition of the latter varies somewhat from place to place, its general appearance is so uniform that it does not admit of further subdivision. Besides describing these five, representing natural conditions, it is necessary also to men- tion some of the effects of cultivation, pasturing, and burning. The consocies are described in the reverse order of their importance, and the notes on cultural changes follow. It will be obsen-ed that the fioristic differences between the various consocies are slight. The Carex Muhlenbergii Consocies The only obser\-ed examples of this consocies were in the Ilan- over area, the first in an interdunal depression, the second on the 56 side of a gentle slope. In both cases they were surrounded by othei consocies of the same association, but were sharply separated from them. The dominant species is Carex Muhlenbcrgii. The bunches are separate or rarely confluent and cover about three fourths of the surface^ Since there are few dead leaves beneath the bunches, and the living leaves are mainly erect, there is abundant space for other plants. Although four species of grasses are included, of which three are typical bunch-grasses, they are so sparsely represented that none can at any place be considered dominant. The following sec- ondary species were noted . Bunch-grasses : Lcptolouia cognatum Paiiiciiiii virgatum Perennials : Lithospcriinnii Guiclmi Pentstciuon Iiirsntus Solidago iiciii oralis Mat: Opinitia Rafiiicsqiiii Interstitials : Monarda punctata hinavia canadensis Panicum psciidopithcscens Poa pratensis Hclianthus scaberrimus Hcliaii til us occidentalis Ambrosia psilostachya Lactuca canadensis The vernal aspect is characterized by Lithospcnnum Gincliiii, the serotinal by Monarda punctata, and the autumnal by Helianthus oc- cidentalis. The whole consocies stands out in sharp relief from its surroundings because of the rich dark-green color of the dominant species. Carcx Muhlenbcrgii is also widely distributed throughout the bunch-grass association, and occasionally appears in large numbers on the lee deposits of blowouts, and may take part in their stabiliza- tion. That the consocies does not have this origin is shown by the absence of Diodia teres and Tephrosia x'irginiana, the poor develop- ment of Panicum virgatum, and the presence of Opttntia Rafinesquii. The Stipa spartea Consocies This consocies is developed in but one place in the Hanover area, and is there of limited extent. 57 Slipa spartca is the dominant grass, with Poa pratoisis second in abundance. The bunches of Stipa are here more or less confluent, and the intervening spaces are so occupied by bhie-grass that the sur- face of the sand is completely covered. This leaves no opportunity for the growth of the usual interstitial plants and also tends to limit the number of perennials. But four species occur and they are rep- resented by few individuals. They are Panicn in pscndopubescens, Caliirlwe triangiilata, Coreopsis palinata, and Aster linariifoliiis. All of these are common in other consocies of the same association. At either side the consocies changes rather abruptly into an- other characterized by Koclcria cristata, which has larger open spaces between the bunches and permits the growth of more secondary species. The Koeleria cristata Consocies In the Hanover area this is by far the most important consocies of the bunch-grass association which is characterized by a single species, and in area is second only to the mixed consocies. If its present extent may be taken as an index, it must originally have covered hundreds of acres of the sand prairie, although in scattered patches of rather small size. It is found alike on the sides and tops of the hills, but seldom in the depressions between them. Elsewhere the consocies was not observed. The dominant species is Koeleria cristata. The bunches of this grass are mostly separate and compact, occupying from one half to two thirds of the surface. The dead basal leaves cover the sand be- tween the bunches, and make an efficient protection against blowing. Panicum pseiidopubescens, which flourishes where the sand is largely bare, is also frequenth' well developed. The number of secondary plants is large because of the unusual amount of ground space available, and comparatively many species are represented. The number of individuals of the interstitial species is especially large. The mats of ScIagineUa rupestris reach here their maximum size; regiflar circles up to three feet (i m.) in di- ameter are common, and they may become confluent to fomi solid masses eight to ten feet {2-^ m.) wide. Koeleria cristata lives in- discriminately upon these mats or between them, and so do most of the perennials. Paniciiin pscndopubescens and the annuals are sel- dom found except on the bare sand between them. The centers of the Selaginella mats are usually covered with a black crust, upon which a species of Cladonia is frequently growing. Antennaria s]). may grow on the mats also, or in the al)sence of Selaginella fonn 58 circular patches one to three feet ( conspicuous because of their gray species are the following'. Bunch-grasses : Sorghastrum nutans Panicum virgatnni Pauicitm pseiidopitbescens Perennials : Tradcscantia reflexa Amorpha canescens Petalostonum candidum Petalostciiniiii pitrpurcum Tephrosia lirginiana I'iola pedata CaUirhoe triangulata Lithospcnuitiu Giiiclini Mats : SclagincUa riipcstris Opnntia Rafinesquii Interstitials : Fcstiica octoflora Riiincx Acetosclla Lepidium virginicum Arabis lyrata 3-10 dm.) across, which are very color. The principal secondary Panicum pcdongtim Stipa spartca Pcntsfcnwn liirsutus Solidago nemoralis Aster scriceiis Aster linariifoliiis HcUanthits scabcrriinus Coreopsis palmata Artemisia caudata Opnntia fragilis Antennaria sp. O.valis corniculata Scutellaria parvula Monarda punctata Ambrosia psilostachya The vernal aspect is characterized by the blue flowers of Viola pedata, which were still in bloom when the consocies was first vis- ited in June ; later, Pentstemon hirsutus and Lithospermum Gmelini are conspicuous with their white and yellow flowers. The aestival season is well markd by Tradcscantia reflexa, and the serotinal by Monarda punctata, which is frequently present in immense numbers. The flowers of the prairie clovers (Petalostenmin), the lead plant {Amorpha canescens), and the sand poppy (CaUirhoe) appear at the same season, but the plants are usually too scattered to break the eft'ect of the masses of Monarda. Still later, in the autumnal aspect, the prevailing tone is yellow from the flowers of the goldenrod, Solidago nemoralis. The Koclcria cristata consocies illustrates well the general prin- ciple that an association may be derived from different sources. In 59 some cases it is evidently the result of the stabilization of the Pani- ciiin pscudopiibcsceiis association, in which event it is characterized by the greater abundance of that species, the better development of mats, which are composed of Sclagiiiclla rather than Antcimaria, and the greater abundance of Scutellaria parmtla and Arabis lyrata. In other cases it is entirely independent of any former blow conditions, and then contains less Panicinn pscudopubcscens, mats of Antennaria rather than Selaginella, and a larger proportion of perennials, in- cluding Aster scriceus and Amorpha canescens, which are absent on blo\\ing sand. The two types have the same structure and represent the same consocies, notwithstanding their difference in species. The only ecological difference between them, aside from their origin, is their age, and it may veiy properly be considered that the floral differ- ence will gradually disappear as the various perennials succeed in es- tablishing themselves in the younger type. The order of appearance of the species in this process of stabilization will be considered later. There is usually a gradual change in the appearance of the con- socies at its margin as other grasses appear or as Koeleria disappears. The secondary species vary but little specifically, but the number of individuals naturally increases or decreases according to the density of the grasses. The Leptoloma cogiiatitni Consocies This consocies is extensively developed in the Hanover, the Dixon, and the Oquawka areas, and in the last two is by far the most important consocies characterized by a single species (PI. I, Fig. 2). In the Hanover area it is exceeded in extent by the Koeleria cnstata and the mixed consocies. It has a wider topographic range than the Koeleria cristata consocies, and is found in the interdunal depressions as well as on the hilltops. Leptoloma cogiiatiim is the principal bunch-fonning grass, and its flat-topped bunches are usually so confluent that nine tenths of the surface or more is occupied. The bunches are of such unifomi height and density that the consocies appears as if artificially trimmed, and has a generally gray-green color because of the numerous dead leaves mixed with the living. The other grasses, which are usually of larger size and bright green in color, stand out very prominently against the background. In the serotinal aspect the plants are in bloom, and the large panicles with the red spikelets are so numerous that the whole consocies has a reddish hue. A few other grasses may at some places occupy enough of the surface to affect the gen- eral appearance of the consocies. They are Koeleria cristata at Han- 60 over and Dixon, Sorghastnnn nutans at Hanover, Panicnin pscudo- pubescciis at Hanover, A)idropogon scopariiis at Dixon and Oquawka, and Faspalum sctaccum at Oquawka. At Dixon, Chrysopsis villosa becomes conspicuous and occupies a large amount of space, but it seems prol^able that the conditions there are not quite normal. It has already been noted that the confluent habit of the bunches of Lcptoloma cognatnm restricts the space for secondary species. The great extent of the consocies, on the other hand, tends to in- crease the number of species, even though the number of individuals is relatively small. A list of the secondary species follows. Bunch-grasses : Andropogon scoparius Sorghastrum nutans Faspalum setaccum Panicuni perlongum Panicum Scrihncrianum Panicuni pseudopubcsccns Koeleria cristata Perennials : Tradcscantia rcflcxa Sisyrinchiiun sp. Oxybaphus nyctagineiis Delphinium Pcnardi Baptisia bractcata Amorpha canescens Petalostemum purpureuin Tephrosia zirginiana Lcspedeza capitata Polygala polygama Bnphorbia corollata Rhus canadensis, var. illinoensis Ccan thus aniericanns Callirhoe triangulata Hclianthemnvi majus Viola pedata Mats: Selaginclla rupestris Opuntia Rafinesqnii Bouteloua hirsuta Bontcloua eurtipendula Poa pratensis Cypenis Schweinitzii Carex nnibeUata Carex Miihlcnbergii Asclcpias aniplexicaulis Acerates viridiflora Acerates viridiflora, \'ar lanccolafa Lithospenmim Gnielini Verbena stricta Verba^cnm Thapsus Pentstemon hirsutus Riicllia ciliosa Liatris scariosa Chrysopsis villosa Solidago speciosa, \ar Solidago nem oralis Aster linariifolius Hclianthus scaberrinnis Achillea Millefolium angustaia Antennaria sp. Senecio Balsamitae 61 Interstitials : Festuca octoflora Croton glandulosiis, var. sep- Cypenis iilicuhnis tentrionalis Rniiicx AcctoscUa Oenothera rhombipctala Polygonum tennc Monarda punctata Silene antirrhina Hedeonia hispida Arabis lyrata Linaria canadensis Lepidiuui virginicwn Specularia perfoliata Cassia Chaniacchrisfa Erigeron animus Liniim sulcatum Erigeron ranwsus Polygala incarnata Ambrosia psilostachya Of the grasses in the above Hst, Bouteloua hirsuta is most abun- dant in the Oquawka area, and may usually be found in any of the narrow strips of sand between the bunches of Leptoloma, although its small size makes it very inconspicuous. Near Hanover, Bouteloua does not occur in this consocies, and Panicuni Scribnerianuni and Panicnni perlongum are important secondary species. The others are usually infrequent but are sometimes very conspicuous if the bunches are of larg^e size and overtop the Leptoloma. Those of An- dropogon scoparius and Care.v Muhlenbergii contrast especially with Leptoloma both in size and color. In the Hanover area the most abundant perennials are Euphorbia corollata and Helianthus scaberrimus; in the Oquawka area, Ruellia ciliosa and Baptisia bracteata. Many of the perennials are conspicu- ously taller than the Leptoloma and stand out in prominent relief above it. This is especially true of the bvtshy shrubs Ceanothus americanus, Rhus canadensis, var. illinoensis, and Amorpha can- escens. Antennaria is the most abundant mat, and is frequent throughout the consocies. The mats are small because of the limited space avail- able. Scnccio Balsamitae forms dense patches two to three feet (S-io dm.) across and shows some tendency to resist the encroach- ment of the bunch-grasses. Of the interstitial plants, Ambrosia psilostachya is omnipresent, and is represented by an immense number of individuals. Riimcx Acetosella and Monarda punctata are also very abundant. The lat- ter is one of the most conspicuous features of the serotinal aspect. Oenothera rhombipctala is not so abundant as Monarda, liut is ecpially conspicuous at its blooming season during tlie aestival aspect, be- cause of its taller stems and vivid yellow flowers. Tiic other in- terstitial plants vary greatly from place to place, and almost anv 62 species may in some places or at some seasons appear very conspicu- ously. A striking example of this was given by Linuni sulcatum, which was observed only on one sand-hill near Oquawka, and was so local that it did not appear in any of the quantitative studies made there. The plant has very slender erect unbranched stems, and dur- ing the first days of July was hardly noticeable. A few days later the flowers appeared and brought the plant at once so much into evi- dence that it might have been wrongly considered a characteristic member of the consocies. Table I. — Floristic composition of ten quadrats in the Leptoloma cog- nation Consocies, oquawka area. Leptoloma cognatum Paspalum setaceum Panicum Scribnerianum Bouteloua hirsuta Cyperus filiculmis Carex Muhlenbergii Rumex Acetosella Polyiionum tenue Silene antirrhina Cassia Chamaechrista Baptisia bracteata Amorpha canescens Polygala iiicarnata Oenothera rhombipetala Verbena stricta Monarda punctata Hedeoma hispida Lfinaria canadensis Pentstemon hirsutus Ruellia ciliosa Specularia perfoliata Erig-eron ramosus Antennaria sp. Ambrosia psilostachya Senecio Balsamitae xxxxxxxxxx X X - X X X - X X - - X - - - X X - - - XX-XX X-XXX - - - - - - - X X - X - - XXXX - XXXX - - - - - X - - - X - - X - - - X - - . - X X X - - X X - - X X X . X - - X - X - X X - X - - - - X X - - - - - - X - - X - - X X X - X - - X X X - X - X X - X X - - X------ X - X - . - X ------ - ----- X - - X - XXXXXXXXXX . X - ------ - - - - X X X X - - X . X ------- - XX - XXXXXXX A series of counts (Table I) were made at Oquawka, in the best example of the consocies observed (near the site of Plate I, Fig. 2), to determine the relative frequency of the secondary species. These counts record the presence or absence of the species in each of ten quadrats two meters square, extending in a continuous strip through 63 the consocies. No record was made concerning the abundance of each species. Boiiteloiia liirsiita, Riiellia ciliosa, Ambrosia psilosfacliya, and Rumcx AcctoscUa appear as the most frequent of the secondary species, while Lcptoloma cognatuni naturally appears in every cjuad- rat. The average number of species in each quadrat is 10.6, while the whole number observed in the Oquawka area is 47. The consocies was not under observation during the vernal sea- son, but Baptisia bracteata and Delphinium Pcnardi are probably quite conspicuous at that time. During the aestival aspect of late June and July Oenothera rhonibipctala and Amorpha canescens are much in evidence. These are followed in August by Monarda punc- tata, and the red spikelets of Leptoloma cognatuvi are also veiy conspicuous at that season. In the Hanover area the vernal aspect is characterized by Pentstenion Iiirsutiis, while the aestival and sero- tinal conditions are essentially the same as at Oquawka. In the Oquawka area contact between this consocies and others was not observed. At Hanover it grades into the mixed consocies next to be described. There is no sharp line between the two, but other species of grasses appear, the spaces between the bunches become wider, and a greater number of secondary species occupy the bare sand thus available. The Mixed Consocies In the Hanover area the greatest portion of the sand prairie was originally occupied by a mixed consocies, in which several species of bunch-grass were well represented (PI. II; PI. Ill, Fig. i). The same consocies was also of considerable importance in the Havana area, and was described in a former i)aper ( Hart and Glea- son, ipo/: 158-160). It was also well represented in the Oquawka area, especially in the prairies between Keithsburg and Oquawka. In the Dixon area no estimate can be made at present concerning its former extent. It seems probable that over the sand prairies as a whole at least two thirds of the surface was occu|)ied by this mixed growth. Although now greatly reduced in area because of cultiva- tion, the remnants left show that it grew alike on the higher eleva- tions and on the depressions between the hills ; that there was little difference in the vegetation as the habitat changed; and that the specific composition of the grasses varied considerably from place to place, but that the general appearance of the consocies was remark- ablv uniform. 64 The reason for its wide extent is ob\ious. The bunch-grasses all belong to the same ecological type, and, with the unimportant exception of Boutcloua Iiirsuta, have approximately the same size. Competition between them therefore is largely limited to a struggle for ground space, and of that there is usually an abundance. There is very little possibility of one species shutting off the light from another, either by its size or by making an earlier start in the season. The dead leaves and culms with which each bunch is sur- rounded make a good ground cover which holds the sand and ex- cludes the growth of seedlings of competing species. None of the species is distinguished by a particularly large seed production or by special adaptations for seed dispersal. Few of them spread by underground stems. Taking all these points into consideration, it is clear that there are no particular adaptations which might lead to a monopoly by one species in the consocies. The presence of so many species indiscriminately mixed is caused by their uniform dis- semination and continued by the evenness of their competition for space. For a few species these statements do not hold. The small bunches of Boutcloua Iiirsuta and the flat ones of Panicuui pseudo- pubscens are easily overshadowed and killed by the growth of other species. The loose, few-leaved bunches of Stipa spartca similarly tend to be crowded out by species of denser habit. These three species, accordingly, are not to be found throughout the consocies, but tend to disappear as the surface becomes more completely covered. The number of grasses which occur is large, and includes vir- tually eveiy species of bunch-grass found in the region. Not all of them occur together, or even in the same area, but in most places three or four may be recogiiized as of chief importance, while the others have more of the nature of secondaiy species. The following bunch-grasses were observed : *Andropogon scoparius Calaiuuvilfa loiigifoHa *Andropogon furcatus • *KocIcria cristata Sorghastruni ntitaus * Boutcloua liirsufa *Leptoloiiia cogimtum '^Boutcloua curtipcudnia Paspaluui sctaccuin J^ragrostis trichodcs *Pankuin z'irgatum *Eragrostis pcctinacca Panicuui pcrlonguui Poa pratensis Panicum Scribiicriauinn *Cypenis ScJiwciiiitcii *Panicum pseudopuhescens Carcx nmbcllata *Stip'a spartca Carc.v Muhlcuhcrgii Of these twenty species, eleven, marked with an asterisk, have been noted in some locality as dominant species, that is, so abundant 65 and occupying so much space that their removal would seriously change the nature and appearance of the consocies. A further study shows that three species are so regularly present and so frequently associated with each other that they may be regarded as the most typical grasses of the consocies. They are Lcptoloma cognatiim, Koeleria cristata, and Andropogon scoparius. The remaining nine grasses are always secondary species and never occupy a considerable portion of the ground space. Their huge bunches, as of Sorghastnim nutans, or tall culms, as of CalauwvUfa longifolia, may nevertheless make them ver\' conspicuous in some places. A few other grasses, not bunch-formers, are also mentioned under the proper head. There is a great variety of perennials, interstitials, and a few mats, of which the following were listed. Perennials : Bquisetuiu hycuiale, van inter- nicdinm Tradcscantia refleva Sisyriiicliiinn sp. Anemone cylindrica Baptisia bracfeata Aniorplia cancscens Pctalostcniinn purpnreum Petalostcmitm candidum Tephrosla zirginiana Lespedeca capitata Polygala polygama Euphorbia corollata Rhus canadensis, var. illi)ioensis Ccanothus ainen'canus Ceanothus ovatiis Callirhoe triangulata Hclianthcmu'ii ma jus J'iola pcdata Asclepias a)nplc.vicaulis A cerates viridiflora Acerates viridiflora, var. lanccolata Brewcria Pickcringii Phlox bifida Lithospermu}n Gmclini J'crbena stricta Physostegia denticulata Physalis zirginiana Physalis heterophylla Pentstemon grandiflorus Pcntstemon hirsutus Syntliyris Bullii Kuhuia cupatorioides, var. corymbulosa Liatris cylindracea Liatris scariosa Chrysopsis villosa Solidago speciosa, var. angustata Solidago ncmoralis Solidago missonricnsis Solidago rigida Aster sericens Aster viulfiflorus Aster linariifolius Aster sp. Brauneria pallida Hclianthits scabcrrimus 11elianthus occidentalis Coreopsis palmata A ch illca Millefoliu ni Artemisia caudata 66 Interstitials : Arisfida tuberculosa Fcstuca octoflora Elymns xirginicus Cyperus fHicuhnis StenophyUus capillaris Carc.v pcnusylvamca Carc.v fc.stncacca, var. brevior Commclina virginica Polygonum tcnuc Chcnopodium album Froelichia floridana Mollugo vcrticillata Silcnc antirrhina Talimim rugospcruinm Lcsqnerella argcntea Lcpidium virginicmn Bryxiuium pari'iflornvi Arabis Ixrafa Cassia Chauiacchrista Strophosfylcs sp. Lmum sulcatum Polygala vcrticillata Croton glandulosus, var. scptoitrionalis Crotonopsis linearis Euphorbia Gcycri Oenothera rhombipetala Verbena hracteosa Monarda punctata Hedeoma hispida Linaria canadensis Speciilaria perfoliata Brigeron ramosus Brigeron canadensis Gnaphalium polycephahmi A uibrosia psilostachya Bacfuca canadensis Autennaria sp. Mats : Selaginella ru/^estris Opuntia Rafinesquii Parasite : Orobanchc fasciculata Not eveiy station of the consocies contains all of these secondary species, or even a majority of them. The actual specific compo- sition of the consocies and the frequence and abundance of the species vary so g-reatly from place to place that individual descriptions must be given. Seven distinct areas occupied l)y the mixed consocies, were examined with more or less detail. I. Hanover area, one mile southwest of the railway station. The consocies occupies a flat interdunal depression (PI. Ill, Fig", i). Nine species of bunch-grasses are present, which are named in the approxi- mate order of their abundance : Beptoloma cognatum, Koeleria cris- tata, Stipa spartea, Sorgliastrum nutans, Panicum pcrlongum, Pan- icum Scribneriamim, Carc.v Muhlenbergii, Panicum pseudopubescens, and Bouteloua hirsuta. Of these the first two occupy more space than the other seven together. The bunches are very compact and close and at a little distance resemble a close sod. Manv dead leaves 67 collect tmder the bunches, aiding- in the soil formation, and even the narrow strips between the bunches are frec[uently covered with dead leaves. On the small open spots are mats of Antcnnaria sp. and SclagincUa riipestris, often growing together, and as the Selaginella dies out in the middle of the mats a brownish moss comes in. The result is that there is absolutely no chance for the sand to blow and humus can fonn rapidly. The surface sand is dark brown in color, somewhat loamy in texture, and partially coherent because of the mass of rootlets in it. This condition continues to a depth exceed- ing ten inches (25 cm.). Of the grasses mentioned, Panicum psetidopiibescens, often so abundant in the consocies, is cjuite scarce, because of the absence of flat bare sand areas on which its decumbent bunches may spread. Boittcloua Jiirsuta is also scarce for the same reason. It grows in small tufts two to four inches (5-10 cm.) high on the mats of Selaginella. The presence of Panicum Scribncriannni is of interest, since it occurs only in the densest growth of bunch- grass. Thirty-one secondary species occur in various degrees of freciuency. Ten quadrats of four square meters each were examined, and the numeral following each plant name indicates the number of quadrats in which the species occurred. Eleven species, without numbers, did not appear in the quadrats, but were found elsewhere in the consocies. The secondary species are as follows. Perennials : Bquisctiini hycmalc, var. intcnncdiiiin (2) Poa pratensis Peialostcmum purptireum (i) Tephrosia z'irginiana (4) Polygala polygama Euphorbia corollata (4) Viola pcdata (i) Callirhoe triangulata (7) Lithospcrnium Ginelini (2) Mats: Selaginella ritpcstris (8) Opnntia Rafinesquii (i) Interstitials : Fcstiica octoflora Clicnopodiiiui album Arabis lyrata (i) Accratcs z'iridiflora, var. la)iccolata ( i ) Physalis liyginiana Aster sericeus ( i ) Aster linariifalius (8) Aster sp. Hclianthus scabcrriuius Coreopsis paliuata (2) Artemisia candata (i) Antcnnaria sp. (6) Hcdeoma hispida Liiiaria canadensis Specularia perfolia la (6) 68 Oenothera rhoiiibipctala (i) Brigeron rauiosits Verbena bracteosa Ambrosia psilostachya (9) Monarda punctata (4) It may be noticed that of the twenty species appearing in the quadrats only four were interstitials, while of the eleven more infre- quent species not appearing in the quadrats seven were interstitials. The relative frequency of the perennials is as 172 to 100. This illus- trates and substantiates the general principle that the number of in- dividuals of interstitial plants decreases as the density of the bunch- grass increases. Although Ambrosia psilostachya has a greater frequency than any of the perennials, it actually plays a very unim- portant part in the consocies. Growing up straight and slender, it is quite inconspicuous and really much less important than Aster iinariifoliiis. The high frequency of the latter species and of Cal- lirhoe triangnlata is also of interest 2. Hanover area; up the hill (PI. Ill, Fig. i ) toward the plateau at the southwest of the station just described. The consocies con- tinues without interruption, but is somewhat different in appearance (PL II, Figs. I, 2). On the hillside the dominant species are the same, but Panicnm Scribnerianuin and Selaginella riipestris dis- appear; the mats of Autenuaria are sparse; the ground is not well covered and does not have the loamy texture of the sand in the de- pression. Boiiteloua hirsnta becomes more abundant, corresponding to the larger surface of open sand, and Linaria canadensis, an inter- stitial, is common with it. The remaining secondary species are al- most the same as in the valley. A transect up this hillside is shown in Table II. The table shows that the change from the lower (left) end of the transect to the upper is caused chiefly by the addition of species as the space between the bunches becomes larger. In the first half the average number of species per quadrat of 0.25 sq. m. is 2.5, while in the last half it is increased to 4.35. < Ed > O < a CO w o o CO Z o o Q X w o o ce M K < & en « Q P o o 0} H w .J n <: Eh M ^ f^ rS r^ y. y, y y y y y y y y y y y y y y y y y y y y y y y y y y y y y y y y X y y y y y y y y y y y y y y y y y y y X y X y M 3 'u 'u ?! »- a. ° a w u ^ - be e e\'en hydrophytic in nature. The few cases observed have not made it possible to determine the order in which the vegetation develops, and the dis- cussion must be limited mainly to the simple description of condi- tions as they are. Cowles has mentioned a similar succession at the head of Lake Michigan (iSpp: 308). The deepest of the excavations, measured by the vegetation rather than by actual dimensions, is in the "Devil's Neck"' north of Topeka, in the Havana area. The center of the depression is a sandy loam and probably represents the subsoil, the Miami loam of 110 the Suil Survey, un which the sand is superposed. It is occupied by a sparse vegetation of Lndvigia pahistris and Bleocharis obtusa. Surrounding it is a zone of Salix loiigifolia, now about three feet (I ni. ) high, and Junciis aciiitiinatus. Outside of this are the usual plants of stabilized blowouts, particularly Oenothera vhomhipetala and Lespedeca capitate. This vicinity was studied also in the sum- mer of 1904, and according to the best recollection of the writer, no such assemblage of plants was observed. It is entirely prob- able that the association has developed since that time. Willow seeds may easily ha\-e been blown in by the wind, and the seeds of the herbaceous plants may ha\-e been brought in in mud on the feet of birds, which were attracted by a temporary p*ool of water col- lected after rains. Not far from this depression there was a similar one, but with sandy bottom. In 1904 it was occupied by Polygonum acre, Hyperi- cum inntihim, Cypeius rvvularis, and Jiinctts tenuis. In 1908, after a lapse of four years, the deepest part of the depression had been filled, so that the bottom was generally le^'el. The prevailing vegeta- tion is blue-grass, Poa prafensis, but a few relics of Hypericum, /un- cus, and Cyperus still persist. Just east of Havana, at the " Devil's Hole,'' there is another deep depression with a more luxuriant growth of \egetation. In the deepest part there is a small but dense thicket of Salix longifolia. Under the willows the sand is covered with a carpet of moss and decaying leaves, making a humus layer which must aid greatly in the absorption and retention of water. At the edg"e of the thicket is a narrow zone of Bochmcria cylindrica, Ludz'igia altcrnifolia, and Lycopus americantis. These plants may Idc considered as a part of the Saliv association but do not live within the thicket because of the weak light. Around the willows, and extending partl_\- up the hill, is a dense growth of Solidago graminifolia and Bquisetum hyemale, var. intermedium, constituting the Solidago association. The two alternate, Bquisetum occupying about one third of the zone. Ming'led with these are a few plants of Cacalia atriplicifolia, Vernonia fasciculata, and Asclepias syriaca, as well as a number of the usual blowsand plants, as Paspalum sctaceum, Cas- sia Chamaechrista, Cristatella Janiesii, Croton glandulosus, var. sep- tenlrioimlis, Ambrosia psilostachya, and Monarda punctata. Bra- grostis trichodes, Chrysopsis villosa, and Lithospermuui angustifoliuui also occur, but are rare. This zone has a vertical width of about six feet (2 m.). Near its upper (outer) margin are a number of old bunches of Sporoholiis cryptaudrus and a few of Andropogon fur- Ill catiis and J'aniciiin lirgatiiiii. Tliese, as well as the appearance uf the goldenrod, indicate that the zone is migrating up hill over the partially stabilized sand. The rounded contour of the willow thicket with the youngest plants at the edge, show that it is also enlarging and occupying successively higher levels. This movement is corre- lated with the development of the retentive layer of humus under the thicket, but the movement of the Solidago, already six feet (2 m.) above the willows, depends more upon the g-eneral vegetative activ- ity of the plant itself. These two associations, Solidago and Salix, may then be expected to appear in any deep blowout. Naturally the deeper ones only can support the willow, which recjuires a larger supply of moisture. This zonal relation of willow and goldenrod is by no means local, but may be observed in many localities in the eastern states. The absence of the goldenrod zone around the willows in the first depres- sion described is merely one of those chance instances of distribu- tion for which no explanation can be given. Possibly the presence of Juncus acurninatus indicates the first stage in its formation. It is evident that with the establishment of the dense growth of Solidago the movement of the sand must cease, and it may be that it does not appear until the sand has first become static. In either case, if its depth is not sufficient to reach moist layers of sand the willow can not develop and the blowout will be occupied by Solidago alone. On the other hand, if the conditions are suitable for the growth of willows, the Solidago association can develop simulta- neously around it. The willows therefore, requiring the deeper excavation, can not follow the goldenrod, but must appear before or with it. When both are established, the former becomes dominant because of its greater control of the j^hysical conditions and tends to succeed the latter. This condition of affairs is pectdiar in two respects. First, the development of the dominant Salix association can not follow in time that of the minor Solidago association. Second, the general movement of the zones is centrifugal, extending progressively fur- ther up the sides of the blowout, and the direction of succession is apparentl}' toward a hydrophytic climax. It must not be presumed that in this case a hydrophytic clima.x will appear. It is probable that the water-retaining humus does not become thick enough to hold standing water, and it is still more probable that a new movement of sand from the west may overwhelm the whole association. As in another case already mentioned, there is at present no association in the region that could possibly be 112 referred back to this for its origin, indicating' that aU foiTner asso- ciations of this type have sooner or later been destroyed. THE POLYTRICHUM ASSOCIATION In the Dixon area the hydrophytic series is carried further, and a new association, characterized by Polytrichum juniperinum, also appears. The Solidago and Salir associations are also represented. In all, six depressions show one or more of these associations and illustrate not only the successions between them but their develop- ment as well. For convenience they will be referred to by letters. These depressions are near the Northwestern tracks about four miles west of Dixon. Blowout A is on the north side of the track ; B is near the track on the south side; C is east of the deposits of a large blowout south of the track; D, B, and F are in this blowout or its southern extension. Blowout A is a. shallow depression, but with rather moist sand. It is occupied mainly by a dense growth of Solidago graniinifolia, with Equisctum arvcnsc, Carcx sp., and Spiraea salicifoHa as acces- soiy species. Blowout B is smaller in width and length, but deeper and with steeper sides. On the outside there is a ring of Solidago graniinifolia with an abundant growth of Aristida tuberculosa. The most abundant accessory species is Lcspcdr::a capifafa, and others of less frequency are Aiidropogon fiircafiis, Cassia Clianiacchrista, and a few sterile grasses which could not be identified. Aristida and 5*0//- dago are almost equally abundant except at the inner margin, where the fonner is slightly in excess. This zone extends up the hillside to the typical bunch-grass and is rather sharply delimited from it. In the center of this ring is the Polytrichum association. The moss grows in dense mats, occupying all the surface in the deepest part of the depression. These mats are very thick and spongy and sink beneath the feet several inches. The dead stems grade off beneath into a thick, brown, moist, spongy laver of a somewhat peaty texture. The mats are sparsely occupied bv solitary plants of the several acces- sory species. Solidago graniinifolia is the most abundant of these and extends entirely across, but the plants are much smaller than in the association outside. The others are Lycopus anicricanns, Hyper- icum uiajus, Sali.v pcdiccllaris, and Aster sn. There are also a few depauperate relic bunches of Panicum virgatun\. There is a narrow tension zone between the two associations, in \\bich the mats of moss are less close and the stand of Solidago less pure. The moss is 113 encroaching upon the Solidago. There is a difference of about 1.5 feet (5 dm.) in the upper and lower levels of the Polytrichum zone,' and the Solidago association is somewhat broader vertically. The deepest part is about 16 feet (5 m. ) above the drainage level 200 yards (200 m. ) away and all the intervening territory is sand. The mesophytic nature of the association must be due to the action of the moss in developing a retentive layer of humus, rather than to any feature of drainage. Blowout C is small and flat and most of it is occupied by a dense carpet of Polytrichum, with many low shrubs of Salix pcdiceUaris and some seedlings of Popnliis dcltoides. The surrounding zone consists chiefly of Solidago grainiiiifolia and Lespedcsa capilata. This blowout is about 12 feet (4 m.) above the drainage level and about 6 feet (2 m. ) above the cultivated field just east of it. It is 14.5 ft. (4.8 m.) below the crest of the deposits of blowout D at the west. D is a large blowout still active on the north, east, and west. A low oblong area enters the blowout from the southwest and is now almost entirely stabilized. Most of this represents a recent deposit of sand from the rear, but the deepest part, nearest the cen- ter of the blowout, is the extinct basin. It is now 11 ft. (3.3 m.) below the crest of the deposits. In this basin both the Salix and Solidago associations are now developing. The latter is represented by a plentiful growth of Solidago grainiiiifolia and Aristida tuber- culosa, with some Juncus aciciniiiatus; the former by abundant young plants of Salix longifolia, with Ludvigia palustris, some small plants of Panicuvd virgatuni, four or five plants of Populus dcltoides, and four bunches of Scirpus cyperinus. The sand is wet and well cov- ered with a layer of dead vegetable matter. On pure sand back of this basin and 6-12 inches (1-3 dm.) above it is a mixture of the Solidago and Polytrichum associations. The ground is partially covered with dense or open mats of Polytrichum, with Hypericum gcntianoidcs, Rlicxia virginica, Juncus acuminatus, Polygala sanguinea, and seedlings of Salix pediccllaris. Solidago graminifolia and Aristida tuberculosa are abundant, but as usual are cons])icuously smaller when growing on the moss mats. Rhexia may live in the middle of the mats, but Hypericum gentian- oidcs grows only in the bare sand in the immediate vicinity of the moss. Hach of these parts of the blowout illustrates early stages in the development of the associations, before their zonal relations have I)een established. The Salix association apparently demands moist 114 sand aiid conies in only in the deepest part of the blowout. Polytri- chiim, on the other hand, may colonize in relatively dry sand, where it at once produces moist conditions by its dense growth. Still farther in the rear and also somewhat higher is a deposit of sand representing a later stage in the refilling of the basin. It has been stabilized by Panicum virgatuin and Lespedeza capitata. Besides these, Solidago neiiioraHs, Hiidsonia tomentosa, and Pajiicum pseudopuhcscens indicate a reversion to bunch-grass. Over this whole area mats of Polytrichum are appearing. Some mats are large, confluent, and dense ; others small, regularly circular, and with very small plants near the margin. These are coming in even'where, even under the bunches of Panicum virgatuui and Panicum pscudopn- besccns, or surrounding Hiidsonia tomentosa or Solidago )iciiioralis. Coming up with it are many plants of Solidago graminifoUa, Aris- tida tuberculosa, Rhcxia virginica, Polygala sanguinca, and in the deepest parts a few plants of Scirpus cyperinus. On the larger and older mats these species are small or absent, and other species more characteristic of the association occur. These are Sali.v pedicellaris, Viola lancculata, and Spirauthcs ccrnua. A few depauperate plants of Panicum pseudopuhcscens persist even in the dense mats of the moss. Along the south and west margins of this area almost pure mats of Polytrichum extend to the very edge of the blowsand (PI. XII, Fig. I; XII, Fig. 2), reaching a height of three feet (9 dm.) above the Sali.v association already described. They are associated only with the three typical species just mentioned. Throughout this area Solidago graminifoUa and Aristida tuberculosa occur, but they are most abundant on the bare sand between the mats. As the Polytri- chiim increases and finally occupies all the surface, these will be forced into a marginal zone, as in blowouts B and C. THE SWAMP ASS0CI.\TI0N. In the Dixon area, just south of blowout D, is a long north and south excavation (PI. XIII, Fig. i), with rather steep walls of bare sand on either side. These walls represent a partially stabilized wind- ward slope, and are occupied by Carex umbellata, Aristida tuberculosa, Panicum pseudopubesccus, and Solidago ncmoralis. In the deepest parts are two ponds, B and F, surrounded by definite zones of vege- tation. The bottom of the ponds is a black muck well mixed with sand. The water-le\'cl fluctuates with the weather. When visited in August it was one foot (3 dm.) above the basin of blowout D, 6.4 feet (2.1 m.) above the country to the east, and 16 feet (5.4 m.) 115 above the drainage level at the northeast. The inner zone of vege- tation is characterized by Scirpus cypcrinns, Elcochavis obfiisa, Liul- vigia palustris, Junciis nodosus, and a few relics of Paiiicum znrga- tum. Outside this is a regular but narrow zone of Polytrichum, with Polygala saiigniuca, Juncus acuininatiis, Hypericum gentianoides, and Rhc.via virginica. Next is the zone of Solidago graiiiinifolia with its usual associate Aristida tuberculosa, and, as accessor)^ plants, Rhexia virginica, Juncus acuniinatus, Gerardia purpurea, and Poly- gala saiigiiiiica. Either of these outer zones may be absent for short inter\^als, but are usually vei-y distinct. The second pond, F, was almost dry when visited in August, 1908, and its mucky bottom was about 15 inches (4 dm.) below the water- level in pond E. Its vegetation had been badly destroyed bv cattle. On the windward slope near B a mat of Polytrichum is develop- ing in a very shallow, fiat depression three feet (9 dm.) above the water-level and one foot (3 dm.) above its nearest neighbors. It is surrounded by a large patch of Hypericum gentiauoidcs extending one to ten feet beyond it. From this detailed description it is seen that the vegetation of the foreg'oing series of depressions of the swamp formation comprises four associations of the following species. 1. The Solidago association. Solidago graminifolia, Aristida tuberculosa, Equisetuin arircuse, Spiraea salicifolia, Carcx sp.. Poly- gala saiiguiiiea, Gerardia purpurea, /uncus acuminatus, Rhexia vir- ginica. Accessory or relic species: Steuophyllus capillaris, Lcspedeza capitata, Andropogon furcatus. Cassia Chamaechrista. 2. The Salix association. Salix longifolia, Salix nigra, Populus deltoides. Accessory or relic species: Patiicum virgatuin, Scirpus cyperinus, Juncus acuminatus, Ludvigia palustris. 3. The Polytrichum association. Polytrichum juniperinuni, Hy- pericum gentianoides, Salix pediceUaris, Aster sp., Lycopus anieri- cauu.s, Hypericum uiajiis, J'iola lauceolata, Rhexia virginica, Spi- ranthcs cernua. Accessory or relic species : Polygala sanguinea, Pan- icuin z'irgatum, Juncus acuniinatus, Ludvigia palustris. 4. The swamp association. Scirpus cyperinus, Juncus nodosus, Eleocharis obtusa, Ludvigia palustris. Relic species: Panicum vir- gatum. The Salix and Polytrichum associations occupy parallel positions, but de\'elop under different conditions. The fonner demands a con- siderable supply of moisture and is restricted to the deeper depres- sions, whereas the latter may develop at almost any level in the blow- 116 out. The appearance of the Salix association can not follow in time that of the Solidago association, as has already been explained. Poly- irichuiii may colonize not only under the Solidago but also under a more xerophytic type of vegetation as well. Both indicate moist sand : the mosses by retention of moisture, the willows by retention and depth of position. Consequently each develops contemporaneously with the Solidago association, and in the early stages the associations are not differentiated. Later the Solidago association is forced to the outside. In both cases the succession is in a xerophytic-hydrophytic di- rection. Nothing has been observed to succeed the Salix association, even in the oldest and deepest blowouts. The mats of Polytrichum, on the other hand, produce a peaty layer over the sand, which be- comes so thick that it retains standing water and admits of the devel- opment of a pond society. These ponds must be held by a water- tight bottom, otherwise their water would soon drain out through the sandy subsoil. As it is, they are conspicuously higher than the general level of the country. The zones surrounding the ponds move outward and upward and permit the continued growth of the pond. This is evidenced not only by the position of young mats of Poly- trichum, but also by relic bunches of such typical sand plants as Paiiicinn virgatitw, now actually in the standing water. If continued far enough the increase of the pond might ultimately lead to the establishment of other associations, such as pondweeds or water-lilies. Its growth is retarded, however, by the gradual deposition of wind- blown sand, by the accumulation of soil by the acjuatic plants, and by loss of water because of increased pressure on the mucky bottom. Most important of these is the deposition of sand, which will mix with the peat and eventually raise the level of the soil somewhat above the water-table. The later stages in the succession are prob- ably similar to the meadows in the Kankakee area, except that the latter represent primary successions on a large scale, instead of sec- ondary successions in a small area. Succession of the Prairie Formation by the Forest It was a matter of great interest to the first explorers and settlers in Illinois that so much of the surface was occupied by prairie, and that the forests were confined to certain physiographic divisions, es- pecially the stream valleys. In seeking to account for this natural feature, the earlier generation of scientists, and to some extent even the modem ones as well, were influenced, or even prejudiced, by two 117 wrong- ideas. In the first place, as they and their ancestors had hved for generations in a forested country, the forest came to be re- garded as the only possible natural covering, and any other type of vegetation was considered extraordinary. In the second place, they did not at first recognize that the forests were everywhere encroach- ing slowly upon the prairies, or that the encroachment became meas- urable as soon as the prairie fires were checked. The prairie is not an extraordinary thing, to be explained only by some strange or fanciful causes; it owes its origin to ages of arid climate in the west and southwest ( Harve_\', ipoS: 84). The forest also owes its origin to ages of humid climate in the east and southeast (Adams, ig02). These great climatic types acting upon the plant world through evo- lution and elimination, gradually developed the two extreme types of vegetation, each of which was especially adapted to its own en- vironment. After the close of the glacial period migration of each of these types brought them in contact in Illinois and the neighboring states, and a struggle for supremacy began between them. The out- come is decided mainly by two sets of factors; first, the control of the environment by the vegetation, and second, the climatic condi- tions of temperature and rainfall. In the first case, the prairie vege- tation, by virtue of its close sod, tends to prevent the proper germina- tion and growth of the forest-tree seedlings (Harvey, ipo8: 86; Rob- bins and Dodds, igoS: 35). Prairie fires, following- the advent of man, also tend to restrict the growth of the forest. On the other hand, the forest has control of the light supply for the herbaceous layers and the well-established trees are resistant to fire. Above all, the climatic conditions are favorable to forest ( Schiniper, jpoj." 162-173; Tran- seau, 1905). The balance has been in general in favor of the forest and it has advanced slowly upon the prairie.* The greatest speed of advance has been along the lines of least resistance, the wacercourses, and has resulted in long strips of forest, paralleling the streams, and usually widest on the east side of streams or marshes where they were better protected from fire. In the sand regions t!ie forest distribu- tion is not regulated in that way, because of the absence of small streams, but it does show a possible relation to fires. Where the sand lies in disconnected ridges, separated by strips of moist or swampy ground acting as fire-breaks, as in the Havana, Ambov, and Kankakee areas, there is a good growth of forest on the higher ground. Where the sand lies in large continuous masses, as in the * It is probable that at certain places and during certain periods the influence of fires has turned the balance in favor of the prairie, but this has not interfered with the general advance of the forest. 118 Oquawka and Hanover areas, there are large tracts of prairie. The Winnebago area lies protected on three sides by streams of con- siderable size, and is almost entirely forested, except the cultivated fields. In the Havana area, there is a belt of forest along the Illinois river, and large forest masses at the south and north ends, particularly near Forest City and Kilbourne. In other parts of the territory the broader deposits of sand are usually prairie, and the forest is re- stricted to the narrow ridges. These extend north and south and mark the location of old sand-bars. In the Amboy area the dis- tribution is similar, but the ridges run generally east and west. They have probably all been forested except those nearest the margin of the deposits. In the Oquawka area there is a belt of forest along the Mississippi river and another inland near the bluff line. These are connected by broad bands of forest which separate several areas of prairie. The Hanover area has a similar belt along the Mississippi, and a number of transverse strips extend inland. These have been partially cleared, but probably none of them crossed to the bluffs except at the extreme northern end. The Winnebago area was en- tirely forested except a few small areas of marsh and islands of prairie. It is difficult to estimate the proportion of the area covered with forest. It was probably considerably more than half in the Oquawka area, about a third in the Hanover area, and about a half in the Havana area. The regular belt of forest along the rivers in the last three areas may be correlated with the effect of fire. The transverse bands across the area in the Hanover and Oquawka areas follow the most irregular portion of the surface, where the effect of fire was possibly limited. The large grove northwest of Hanover station, in partic- ular, follows a line of steep-sided irregTilar dunes totally unlike the gently rolling prairie. The encroachment of the forest is caused by the slow migration of the forest trees in every direction. The open structure of the bunch-grass does not prevent the proper germination of seeds or growth of seedlings as does the close sod of a nonnal prairie. Few species of trees, however, are able to withstand in their seedling stages the extreme conditions of the physical environment. These are especially the shifting nature of the sand, the hot surface layer, which may be almost totally dry to a depth of more than a decimeter, and the lack of protection against wind during the winter. Still another restricting influence is the absence of ready means of dis- pers?l. The trees composing the early stages of the forest are oaks. Thei' heavy acorns have no means of dispersal except gravity and 119 the agency of animals. There are few animals to carry the acorns out on the prairie. The majority of such acorns are eaten, and many of the remainder decay. Some trees produce exceedingly heavy crops of acorns, which lie in layers an inch or two (3-4 cm.) deep be- neath the tree, but of a large number examined, not one was sound. The life of a tree seedling is at best precarious, and in an unusual environment, with full exposure to wind and sun, few of them may be expected to survive. It is possible that some seasons are more favorable than others, and that after intervals of several years a succession of two or three favorable seasons may lead to a con- siderable extension of the forest. This condition has been described by Ramaley (ipoS: 30) and is probably of wide application. Establibhment of the forest makes at first very little difference in the environment. The trees are relatively far apart, and sufficient light comes through the foliage to permit the growth of many species of the original bunch-grass. The edge of the forest, there- fore, shows, not a change in the flora Ijut merely the addition of a few other species. There are at present few places where the con- tact between forest and prairie can be observed. Of these, the best is in the Hanover area (PI. XIII, Fig. 2). The ground co\"er is the usual climax growth of the mixed consocies of bunch-grass, consist- ing particularly of Koderia cristata and Andropogon scoparius. With these are Boutcloua liirsiila, Aster linariifolins, Aster seri- ceiis, CaUirhoc triangitlata, and other common species. The sand is in apparently the same condition as upon the prairie. The fallen oak leaves have either blown away completely or have been collected in piles around fallen branches and in thickets of Rhus canadensis, var. illinociisis. There is none of the additional herbaceous species typ- ical of the older established forest. In the Winnebago area there are a few small open spots within the forest, which represent the last stages of a prairie. In the first of these there are Carex Mithlcn- bergii, Kocleria cristata, Liatris cylindracca, Lespcdeza capitata, Viola pedata, Polygala polygama, and Artemisia caiidata. Oak seed- lings one or two years old were also present. In a larger opening (PI. XVII, Fig. i) the prairie character is more oln-ious. The domi- nant species consist of a mixed growth of Paiiiciiiii Scrib)icriaiiuiii, P. perlongum, P. pseiidopubcsccns, P. virgatvui, and Carex Miihlcii- bergii. Between their branches the ground is well matted with Cla- donia. Some of the accessor}r species are Tephrosia z'irgiiiiana, Amorpha cancscens, Lcspedcca capitata, Solidago ncinoralis, Asclepias aiuplexicaiiUs, Poteutilla argiila, Aceratrs riridiflora, var. linearis, Viola pedata, and Ambrosia psilostachya. There are no forest relics. 120 This fact, together with the pure yellow sand of which the substratum is composed, indicates that it never has been forested. The boundary between the forest and prairie differs from the usual forest margin in the absence of a tension zone and a definite vegetation. Thickets of hazel, of sassafras, or of sumach, w'hich surround the typical Illinois forests, are absent. There is no sharp distinction of flora within and without the forest edge, and no mass- ing of a large number of species near the margin. The whole suc- cession is of a type rarely mentioned or described, in which there is at first no essential change in the environment. There is no first-hand evidence concerning the rate at which the extension of the forest is proceeding. The first settlements were usually made near the edge of the forest, where clearing and culti- vation at once stopped any advance. Historical evidence is not al- ways of value, because complete dependence can not be placed on statements of a scientific nature made by travelers or casual observ- ers. A note by Patrick Ivennedy (Imlay, lygy: 508), however, is suggestive, and probably at least partially correct. "About sun-set we passed the river Demi-Quian.* It comes in on the western side of the Illinois river (165 miles from the Missis- sippi) ; is 50 yards wide, and navigable 120 miles. We encamped on the south-eastern side of the Illinois river, opposite to a large savanna, belonging to, and called, the Demi-Quian swamp. The lands on the southeastern side are high and thinly timbered ; but at the place of our encampment are fine meadows, extending farther than the eye can reach, and aft'ording a delightful prospect. The low lands on the western side of the Illinois river extend so far back from it, that no high grounds can be seen. Here is plenty of buffalo, deer, elk, turkies, etc." Kennedy's whole narrative seems reliable, and we may believe that at least in some directions his camp commanded an uninter- rupted view^ of the prairie. At the present time, however, the mar- ginal belt of timber along the river in the vicinity of Havana is from 100 yards to a quarter of a mile (100 to 500 m.) wide, while the fringing woods along Quiver creek and large tracts of black oak completely cut off a view of the prairies. If Kennedy's state- ment is correct, then large areas of timber have developed withi^ the last century. It is difficult to explain the migration of the oalres- ence in large quantities may in some degree be considered as one in- dication of the approaching succession. On some ridges there is a well-defined zone of Pteris along the slopes, extending neither into the xerophytic lilack oak timber above, nor into (he mesopiiytic bur 130 oak forest below. This type of distribution has been obsei-ved only on the narrower ridges, along the crests of which the xerophytic habitat is more emphasized and where the accumulation of humus takes place more slowly. On the broader uplands Pteris is usually common. Some of the latter show a transition to the bur oak type in the presence of Pyrola elliptka, Amphicarpa PitcJicri, J'itis vulpiua, and Agriinonia mollis, typical members of the latter association. On the slopes from the black oak into the bur oak association there is usually a well-defined tension zone (PI. XV, Fig. 2) where the plants of both groups mingle. Among these the most abundant are Pteris aqiiiliiia and Sviilacina stcllata of the black oak association and Geranium maciilatiun and Primus virginiana of the bur oak. The shaqMiess of the tension line, coupled with the slow devel- opment of the dominant species of the two associations, indicates a condition approaching an equilibrium between the two associations. Their common boundary on the steeper slopes seems to depend upon the water content of the sand as influenced by the height above the water-table, while on the broader uplands the incipient succession may depend not only upon the depth of the water-table, but to a greater extent upon the increase in water capacity through the de- velopment of humus. The present location of small ponds in the sand deposits shows that the actual depth of the water-table is sev- eral yards, and in all probability too great to explain the sharp tension line already noted. Its origin must accordingly be referred back to a past condition in which the general water-level was higher. Trees of Quercus vehitina in the lower portions of the bur oal< association are very few in number and usually small in size, showing that they are not relics, but recent sporadic invaders, and there is an unusually small number of herbaceous relics. In the upland portions of the bur oak association relic trees of black oak are numerous and frequently of large size, while many relic herbaceous species also occur. Throughout the black oak association pioneers of the bur oak group are well represented, as is shown by the list given in the discussion of the former association. All these peculiarities lead to a choice of two conclusions : the depressions have never been occupied by black oak, or the succession by the bur oak has been extraordinarily com- plete. Further e\-idence leading to the acceptance of the first alter- native is afforded by conditions in the Amboy area and to some extent also in the Kankakee area. In the Amboy area the sand lies in similar ridges mostly parallel to Green river and not over 60 feet (20 m.) above it. The inter- 131 veiling valle3'S are for the most part filled with extensive deposits of muck overlying sand and occupied by swamp vegetation, with Iris versicolor, Typlia latifolia, Rhexia virginica, Houstonia coernlea, Populus trevudoidcs, and other species of similar habitat prefer- ences. Outside these depressions, and accordingly above them, lies the bur oak association, above which in turn is the black oak asso- ciation, occupying the crests of the ridge. In eveiy case the bur oak type is characteristic of the more mesophytic sand near the water- level. According to all established principles of succession the drain- age of the intervening swamps would cause a downward migration of the bur oak association, provided other features of the environ- ment were favorable to it. It is, of course, hardly probable that the bur oak would extend very far out upon the deposits of muck. If the swamps were composed of sand instead of muck, it is very prob- able that the whole area would be occupied by the bur oak association as rapidly as the lowering of the water-level permitted. In the Kankakee area the interdunal depressions are occupied by meadows, which are doubtless very similar to the Amboy swamps, having a number of species in common, and probably representing a further stage in the succession on muck or peat. The ridges are all covered with forest, but in the short trip made through the area the distinc- tions made between the black oak and bur oak associations were not recognized. The whole leads to the first alternative mentioned, that the de- pressions in the \A'innebago area have never been occupied by the black oak association, and that the bur oak association, which now occu]jies them, represents the present culmination of a past hydro- 13h\-tic to mesophytic succession, which has been so far completed that scarcely a trace of it is now in existence. This conclusion is supported by the presence of a single small pond occupying a depression in a partially cleared field. The few plants remaining indicate that the surrounding vegetation was of the bur oak type. The swamp vegetation at its margin is scanty, consisting of Scirpus validus and Stcirojiciiia lanceolatniii, outside of which are successive zones of Populus trcniiiloides and Solidago graiiiiiiifolia. Further details were not noted. The development of the muck soil in the depressions of the Am- boy and Kankakee areas and its absence in those of the Winnebago area must also be explained. The latter areas are essentially fluviatile; their depressions are not far above the beds of the Green and the Kankakee rivers, respectively, and the lowering of the water-level is entirely dependent upon changes in the river level and u]X)n deposi- 132 tion of soil. A slow change in the water-level, for such it must ha\e been along- these comparatively sluggish streams, would per- mit the long-continued existence of swamps and the consequent ac- cumulation of large deposits of muck. In the Winnebago area, on the other hand, the deposits lie many feet above the Rock, Sugar, and Pecatonica rivers ; swamps would be of short duration and the accumulation of muck would not take place. Consequently, the de- pressions have been occupied almost entirely with the bur oak asso- ciation, with the exception of the single pond already mentioned, of the hydrophytic extreme is, as usual, chiefly due to changes in the water factor, while that of the xerophytic extreme is in this case The development of the bur oak association represents, therefore, another case of the interpolation of a mesophytic mean association between a hydrophytic and a xerophytic extreme. The succession caused primarily lay the development of humus. The whole probable successional history is indicated upon the diagram (Fig. 6.) showing the relation of the various associations. The preceding statements concerning the development of the bur oak association following the swamp vegetation does not imply that it was the first type of forest to appear. On the other hand, there is some fragmentary e\-idence that an entirely difl^erent forest association preceded it. Within the bur oak association (PI. XVI, Fig. i), Oiicrcus iiiacro- carpa is everywhere the prevailing tree. In the Amboy area, it was the onl)' arborescent species in the small areas examined. In the Winnebago area it is mixed with white oak, Qucrciis alba, and shell- bark hickory, Carya ovata. The former composes 25-50 per cent, of the whole, while there is seldom over 2 per cent, of hickory. The trees are larger and straighter than those in the black oak association, but still much inferior to those of their own species g-rowing on a more fertile soil. A large number of shrubs form a second layer beneath them, and are often aggregated into dense tliickets. Fniiuis scrotina, Pninits virginiana, and Corylus americana are by far the most abundant, with several other occasional species. Lianes are not com- mon, and consist of scattered individuals of Vitis vulpina, Siuilax liispida, and Rluis Toxicodendron, with a few other species of less importance. The forest cover is dense and the light diffuse. This prevents the growth of most grasses, except where the forest has been partially cleared. The ground cover is composed chiefly of a dense luxuriant growth of herbaceous plants. They are of a larger average size and much more mesophytic appearance than those of the black oak ridges. 133 THE FOREST FOFaiATlON THE SWAMP FORJilATlON Deposit -i Blowsand < Baaln Windward Slope Assoc. * Asaoc. * Assoc. Assoc. Panicum pseudoputescens Assoc. Bunch Grass Assoc. THE PRAIRIE THE PRAIRIE PORUATIOH Assoc. PROVmCE THE BLOWOUT PORUATIOJJ Pi(j. 6. Dia(,'ram showing- the plant associations of llie inland sand deposits of llliiioiSf and the principal successions bft\veen them. U'A Although coniparativel}' few in number, their distribution is remark- ably uniform, so that the various areas occupied by the association are strikingly similar in their vegetational appearance. On the up- lands, it is more or less mixed with relics of the preceding black oak association. An association greatly resembling this, and possibly identical with it, occupies in the Winnebago area the upland areas of clay overlying limestone, and indicates that, in the future development of the vegetation, the associations on sand and clay will become grad- ually similar. This is in accordance with the views of Cowles {igoi: 7), that all the veg'etation of a region "is tending toward an ultimate common destiny." The specific composition of the association is shown by the fol- lowing list. A. Species characte 1. Trees: Quercns uiacrocarpa Querciis alba 2. vSmaller trees and shrubs Popiihis treninhides Populus grandidcntata Coryliis aiiiericana Rnbiis idaeiis, var. aculcatissiiinis 3. Lianes : Smilax hcrbacea Sinilax ecirrhata Smilax hispida Rhus Toxicodendron 4. Herbs : Botrychiuin ternatiiin, var. intermedium Botrychium virginianiim Polygonatnm commntatum Smilacina racemosa Cypripediitm parviflorum, var. pubescens Silcne stellafa Arcnaria lateriflora Heiichera hispida Agrimonia mollis ristic of the association Car\'a ovata Rubus occidentalis Prunus serotina Prumis virgijiiaiia Connis Baileyi Vitis zndpina Psedera quinqiiefolia Lonicera Sidlivantii Desinodiuin grandiflorum Amphicarpa Pitcheri Geranium maculatum - Circaea liifetiaiia Sanicula canadensis Pyrola elliptica Dodecathcon Mcadia Veronica inrginica Galium concinnnm Prcnanthcs alba 135 B. Species more typical of preceding associations Ptcris aquilina Ceanothus antericanus Poa prateiisis Apocynum androsaemifolium Tradescaiitia reflexa Monarda mollis Sali.v tristis Synthyris Biillii Quercus velutina Gerardia grandiflora Frasaria virsiniana, Pedicularis canadensis var. illinoensis Antennana sp. Rosa huinili-s Helianthus stntmosus Amorpha canesccns Coreopsis palmata Euphorbia corollata Cacalia atriplicifolia THE MIXED FOREST ASSOCIATION In the Hanover, Oquawka, and Havana areas the succession from the black oak association is dependent primarily upon a general and gradual increase in the water content of the sand and a correspond- ing decrease in the light, without the concurrence of historical fac- tors as in the two areas previously described. The succeeding vegeta- tion is derived wholly from the surrounding associations. Since other types of vegetation are developed best near the rivers which border these sand areas, invasion begins near the river and gradu- ally extends back toward the center of the sand deposits, so that the most advanced stages in the succession are always found near the river. The development of this succession is greatest in the Ha- vana and least in the Hanover area; a feature which is perhaps cor- related with the general southeastern origin of the forest formation. In the Havana area, as in the Winnebago deposits, the narrower ridges of sand are the last to be affected by this succession, while the broader ridges or those near the river generally show some indica- tion of it. Certain species are soon recognized as the normal pio- neers in the succession, and while their order of appearance is not constant, their presence is always connected with the development of a thin, fibrous laver of leaf-mold over the surface of the sand. It is frequently possible to observe nearly all stages in the succession in a distance of a mile, passing from the edge of the forest toward its center. In some places, adjacent ridges of sand represent differ- ent stages of the succession, and jiermit an easy comparison of the vegetation. This is especially well shown on the first two forested ridges east of Havana, about two miles (3 km.) from that city. The first of these represents an advanced stage of the succession, while the second is occupied by a nearly typical black oak association. 136 Excluding the two species of oaks, the ridges have 29 and 31 species, respectively, of which only ten are common to both. This gives a community coefficient (Jaccard 1902: 351) of 0.200, indicating at once the great floral dissimilarity. There are, on the other hand, thousands of acres still occupied by the black oak association, with as yet no indication of the approaching succession. Rhus canadensis, var. illinocnsis, Tephrosia znrginiana, and Opitn- tia Rafinesqiiii are the chief species concerned in acctimulating fallen leaves for conversion into leaf-mold. With the simultaneous de- crease in light, the succession begins, and occasional plants of Poly- gonatttm couivnitatiim and Silcne stcllata appear as pioneers. Agri- monia mollis comes in somewhat later, and young plants of three lianes appear. These are Vitis vulpina, Rhus Toxicodendron, and Psedcra qiiinqucfolia. The last is especially common and valuable as a succession index. Its long", slender stems trail for several feet along the ground, unless by chance they encounter a tree trunk to climb. Rhus Toxicodendron seldom trails, but usually grows directly at the base of some tree. Following these six species, which are easily rec- ognized as pioneers, a number of others appear in irregular order. The arborescent flora remains essentially the same, except for occa- sional trees of Celtis occidenfalis or Prunus serotina, or, near the river, Quercus rubra, Quercus macrocarpa, Juglans nigra, . Ulmus americana, Morns rubra, and Gymnocladus dioica. The undergrowth is frequently dense, Avith numerous thickets of shrubs, and the her- baceous growth is tall and luxuriant. The following additional species are especially characteristic : Asparagus officinalis Sanicula canadensis Smilacina raccmosa Cornus Baileyi Sniilax herbacea Asclcpias phytolaccoides Dioscorea villosa Lappula virginiana Anemone virginiana Scrophularia leporella Ribes gracile Galium concinnum Celastrus scandens Bupatorium pnrpureum Oenothera biennis Bupatorium urticaefolium It will be noted that a majority of these species have unusually efficient means of seed dispersal. When an association develops de novo some distance from the nearest existing area of it, the most mobile species may naturally be expected to appear first, while the less mobile species follow after greater intervals of time. The species in the preceding list accordingly represent the mobile pioneers of an asso- ciation, the usual dominant species of which have as yet not ap- 137 peared. Near the Illinois river there are some rather extensive sand fields occupied by a forest characterized particularly by bur oak and white oak, with several other arborescent species, such as red oak, Qitercus rubra, elm, Ulmiis americana, hackberry, Celtis occidentalis, and white ash, Fraximis americana. This probably represents the complete succession, the beginning of which has been indicated above. Intermediate stages, however, have not been observed. In the Oquawka area the succession is found only on the long dunes nearest the river. The first indication of it is given by SUene stellata, Polygonatwn commutatum, and Psedera quinqucfolia, which are followed by a number of additional species, including several trees. The whole leads to the highest type of mesophytic forest (PI. XX, Fig. 2, background) that occurs on the sand deposits. The succession is best seen along the dune nearest the river, north of the town of Oquawka, and may be traced through various stages from north to south for a distance of about 1.5 miles (2 km.). This dune has a maximum height of about 100 feet (30 m.), indicating a very strong and continued wind action* at some time in the past. It is now completely covered with trees, and the surface layers of sand are well mixed with organic matter. At the north end a bayou of the Mississippi lies at its base, and the margin of the water is marked by a line of elms and willows, with Rumex verticillatus, Physostegia virginiana, and other species of hydrophytic tendencies. Equisetum hyemale is the only one of these which extends much above the water- level, where it mingles with the usual sand-dune species. Along the north end of this ridge the prevailing trees are Quercus velutina and Quercus marilandica, with occasional trees of Quercus rubra, espe- cially on the lower part of the slope. Besides the pioneer herbaceous species mentioned above, there are also Strophostyles heh'ola, Mo- anrda Ustulosa, Aquilegia canadensis, and Vitis viilpina. Somewhat farther toward the south Juglans nigra appears near the base of the hill not far above water-level; farther along it extends higher and even appears at the top of the dune. Cercis canadensis is usually found with it. In the same way the river birch, Betitla nigra, the elm, Uhnus americana, the green ash, Fraximis pennsvhanica, var. lance- olata, and finally the soft maple, Acer saccharinum, appear first at the bottom of the dune and as mesophytic conditions increase toward the southward extend higher and higher above water-level, until they finally appear at the top. Each one of these species is more moisture- loving than its predecessors, until tlie climax is readied in the soft maple, a characteristic tree of river-bottom swamps, here growing inany feet above the water. 138 The change in the herbaceous and shrubby flora and ground cover is no less manifest. Corniis Baileyi and Scropliularia Icporclla soon appear ; Celastnis sca)ideiis becomes a common Hane ; and Zanthoxy- I'um americaniim grows high above the river. Following these, dense mats of moss and Pcltigera cover the sand and aid in the increase and conservation of soil moisture. With them come such pronounced mesophytes as Pariefaria pennsylvanica, Aster oblongifoUns, Any- chia canadensis, and finally IVoodsia obtiisa and Anemone canaden- sis. The last species, together with the soft maple, is sufficient proof of the extraordinary change that has taken place in the water factor. Bare sand is seldom exposed, but is covered with the dense mats of moss and Pcltigera, and shaded by the luxuriant tangle of herba- ceous plants and shrubs. The few bare spots are still occupied with the typical black oak vegetation of Artemisia caudata, Rudbeckia hirta, and other similar species. The further fate of this association will be described later in con- nection with the vegetational history of the river dune as a physio- graphic form. In the Hanover area, the conditions which lead to succession are virtually the same as in the Havana area, but a much smaller area has been affected. Pscdera qninqucfolia is one of the pioneers, as usual, and is followed by a considerable number of species of meso- phytic character. Among tliese are the following. Trees : Betula nigra Fraxinus pomsylvanica, Ulmtis amcricana var. lanceolata Celtis occidrnfaHs Shrubs : Ribes gracile Pnimts virginiana Riibus idaciis, Rhus glabra var. aciileatissimus Ceanothus aincricanus Riibus occidcntalis Cornus Baileyi Lianes : Smilax ccirrhafa Rhus Toxicodendron Smilax herbacea Vitis vidpina Menispermum canadense Herbs : Polygonatnm commutatum Amphicarpa Pitcheri Silene stellata Ziza- aurea 139 Aqnilegia canadensis Asclepias phytolaccoidcs Rannnciiliis ahortivus Apocynum androsaemifolium Heuchcra hispida Monarda mollis Fragaria virginiana, Bupatorium serotinum van illinoensis Bupatorium urticaefolinm Geimi canadense Antennaria plantaginifolia The River Dunes and their Plant Associations In the preceding pages those various associations have been de- scribed which comprise most of the vegetation of the sand areas. The chief physical factors concerned in molding their topography or differentiating their associations have been wind and soil moisture. There remains to be discussed the narrow strip of dunes which lies close along the Alississippi river and which is affected also bv water action. The river dunes are well developed in the Hanover and Oquawka areas along the Mississippi river, whose swift current and shifting channel have been chiefly responsible for their origin. They are much less prominent along the Illinois river, whose sluggish cur- rent possesses but little power of erosion. The first stages in the vegetational history of the dunes were observed only in the Hanover area; the last, from and including the development of the oak forest, only in the Oquawka area. In the first two areas, the sand deposits lie at an average height of 15-30 feet (5-10 m.) above the swampy, alluvial flood-plain. The river meanders across its flood-plain from side to side, and in some places flows directly at the foot of a sand hill. Under these condi- tions a river dune may be formed. Erosion by the river carries away the sand from below, and that portion of the sand above the high-water mark of the river, and consequently removed from the direct erosive action, stands at a steep slope, the angle of which de- pends upon the wind, the rate of erosion, and the vegetative cover-- ing. The surface sand on this slojje is exposed to the full sun and keeps loose and dry. Below ordinary high-water mark the sand is worked over by the water and lies at a gentle slope, forming a broad or narrow beach. The wind, which is generally from the west, re- moves sand from the lower slope, and to a less extent also from the steeper portion, carries it up the slope, and jiilcs it in a long dune parallel with the river and several feet higher than the general level of the sand. As long as the river continues erosion on that part of its banks, the whole slope moves gradually back; if the wind constructs the dune as rapidly, or more rapidly, than the river erodes 140 it, the whole complex migrates slowly landward. If the river erodes more rapidly than the wind piles up the sand, the dune will soon be destroyed and only a bare slope remain. On the other hand, if the river shifts its channel, or for some other reason ceases erosion, the whole exposed surface will ultimately be fixed with vegetation and be- come static. From a physiographical standpoint, therefore, the whole dune consists typically of two divisions (PI. XVII, Fig. 2) : the lower, termed the middle slope, consists of sand now being uncovered and re- moved by the wind and erosion ; and the upper, called the upper slope, of sand dei:)Osited by the wind, and removed by undennining through erosion. From the standpoint of vegetation, several associations may be distinguished which are in part correlated with the physiography. The lowest portion of the dune, near the river and within reach of high water, is marked by vegetation of a semi-hydrophytic nature. When visited in June, 1908, the river was very high, and only the tops of the half-submerged plants could be seen. These were Populns dcltoides, Salix longifoUa, Fraxinus pennsylvanica, var. lanceolata, Gleditsia triacanthos, and Uliiiits aiiicricaiia, typical sand-bar or river-bottom plants. The herbaceous vegetation of a later season is doubtless of the same ecological nature, probably including Bupato- rium serotiiiiim, Xanthium commune, and other species of similar habitat. This vegetation has no relation to the typical dune vegeta- tion above it, except in the presence of a few individuals of Panicimi virgaUmi, which had probably slid down from the slope above. The vegetation of the middle slope clearly belongs to the blow- sand association, as described under the blowout fonnation. The plant covering is sparse, becoming somewhat dense toward the bot- tom. It consists chiefly of Cassia Chamaechrista and Diodia teres, with smaller numbers of Croton glandiilosus, var. scptentrionalis, Ambrosia psilostacliya, and Cristatclla Jamesii. At wide intervals are tufts of perennials, including Accrates viridiflora, var. lanceolata, Paniciim virgatnm, Lithosperuuim Gmclini, Euphorbia coroUata, Cy- periis Sch-weinitsii, and Tephrosia virginiana. The surface of the sand is dotted with numerous pebbles, sometimes as much as 2 inches in diameter. They apparently do not affect the vegetation, and there are not enough of them to be called gravel. They evidently repre- sent the accumulation of pebbles left by the sand blowing up into the deposits above. The top of the middle slope is marked by the outcrop of a layer of loamy sand (PI. XVII, Fig. 2), very dark brown in color, rather fine-grained, and conspicuously earthy in texture. The top of this layer is well marked, but it gradually passes below into the typical orange- 141 brown sand of the middle slope. It is caused by past generations of plants which occupied this surface before the dune was formed, or at least before it had migrated so far inland. Since this soil blows, weathers, or dries out less rapidly or easily than the pure sand, the outcrop is marked by a slightly steeper slope and by dark-colored patches. Digging behind shows that the stratum extends indefinitely beneath the sand. For long distances the outcrop line is very dis- tinct, but not perfectly level. Its elevation varies gradually, but irregularly, and in some places the whole outcrop disappears, cor- responding to irregularities in the original level of the sand, or to locations of fomier blowouts. This soil stratum is on the same level as the country behind the dune and illustrates plainly the continued inward migration of the dune. (Cf. also PI. XVIII, Fig. 2.) It is characterized now by a line of Blymtts canadensis. The upper slope consists of fine sand piled at an average angle of about 20 degrees. The vegetation is much like that of the middle slope, but denser and with many additional species (PI. XVII, Fig. 2). It is likewise referred to the blowsand association. Cassia Chainae- christa and Diodia teres are again the most abundant species, and Aristida tuberculosa is also conspicuous. Other less characteristic species are Monarda punctata, Lithospennu-m Gmelini, Buphorhia coroUata, Diodia teres, Cristatella Jaincsii, Ambrosia psilostachya, Oenothera rhonibipetala, Linaria canadensis, Kuhnia eupatoriodes, var. corymbidosa, Acerates vvridiHora, van lanceolata and var. line- aris, Oxybaphus nyctagineus, Teucriuin occidentale, Tradescantia re- flcxa, Lespede::a capitata, and Rmnex Acetosella. In somewhat shel- tered places Scrophularia leporella, Draba caroliniana, and Coryd- alis micrantha occur. In some places, near the top of the slope, adjacent to the thickets described later, are Associations of Physalis heterophylla (PL XVIII, Fig. i). The individual plants grow in large patches and are very loosely aggregated, with a large amount of open sand be- tween them, yet the patches are remarkably free from other plants. Even Cassia and Diodia, so abundant on the upper slope, are almost entirely absent from these patches. This peculiarity of distribution leads to the inference that they are more closely related ecologically to the thickets which crown the dunes than to the slope below. They may bear the same relation to the blowsand association that Sniilacina stellata on the lee slope bears to the deposit association. The crest of the river dune is primarily an area of deposit, and is consequently occupied at first by the regular deposit association, already described in connection with the blowout formation. Rhus 142 canadensis, van illiiiocnsis is again the most important member of the association. The perpetuation and vertical growth of the river ckme arc chietly due to its efficiency as a sand-binder. The general height of the dune thus held is from 15-30 feet (5-10 m.) above the general level of the sand, but a maximum height of about 80 feet (25 m.) is attained in the Oquawka area, or fully 100 feet (30 m.) above the high-water level of the river. It is noteworthy that this highest point is occupied by a loose patch of Rhus, evidently of great age. Asso- ciated with Rhus on these dunes are similar dense patches of Ceano- thus ovatus and, occasionally, of Rhus Toxicodendron. This shiaibby habit of the last species was not observed elsewhere in the region, but is very common along the dunes of Lake Michigan, particularly toward the north, and is reported from Lake Erie by Jennings (/pop). There are also the usual bunches of Tephrosia virginiana. Panicuni virgatinn (PI. XX, Fig. i), and, more rarely, Bragrostis trichodes and Sporoholus cryptandrus. Litervening spaces of open sand are occupied by the usual members of the blowsand association. If the erosion by the river proceeds at such a rate that the crest of the dunes remains relatively stable for some years, opportunity is given for the development of a higher type of vegetation. The first step in this succession depends ujion the introduction of seeds by wind or animals from the alluvial bottom-lands. The species most frequently introduced in this way are Ulinus auicricana and Fraxi- nus pennsylvanica, var. laiiceolata, both of which have light winged seeds. Pods of Glcditsia triacanthos are blown up the slope from the trees on the river bank below, and more rarely Jnglans nigra devel- ops from seeds probably carried by animals. Seeds of Acer saccha- rinuni w-ere also found on the dunes, but they probably do not germi- nate, since no young plants were seen. These trees are not numerous, and never reach a large size, partly because of the unfavorable habi- tat, but chiefly because of the general movement of the dune. All the older trees have portions of their root systems exposed. One ash tree, 8 inches (2 dm.) in diameter, had the base of its stem 3 feet (i m.) above the surface and 15 feet (q m.) behind the present crest of the dunes (PI. XVIII, Fig. 2). The ash, which is by far the most abundant of the trees, usually branches freely from the base, forming a complex of stems. The trees offer a roosting place for birds, which in turn serve as agents in the dispersal of several shrubs and lianes. These at once spring up beneath the trees, and develop the dune thicket associa- tion. The mature thickets (PI. XIX, Fig. i) are exceedino-lv dense, impenetrable tangles of shrubs and lianes, with an occasional tree, half 143 smothered with, vines, rising above them. Eight species of shrubs or small trees and seven species of lianes are concerned and, with a single exception, all have seeds adapted to dispersal l>y birds. They are as follows : Sinilax herbacca Primus seroiiiia Smilax hispida F riinus sp. (plum) Salix longifolia Rhus Toxicodendron Celtis occidentalis Celastrus scandens Menispermum canadense Pscdcra quinqiiefolja Ribes gracile Vitis vulpina Pynis ioensis Coniits Baileyi Priinus zirginiana The choke-cherry {P. virgiiiiaiia), plum, and crab (Pyrits ioensis) are the most abundant shrubs. The plum has running roots which send out shoots at short intervals, so that it tends to spread out upon the blowsand. The lianes are usually luxuriant and cover the shrubs with such masses tliat the supports are almost hidden. Within the thicket the light is very low; many of the branches are leafless or dead, and the herbaceous vegetation is scanty. It consists of Teii- criiiui occidentale, Scrophtdaria leporella, Polygonatum commutatum, and Sniilacina raccnwsa, with seedlings of Psedera quinquefolia. These thickets occupy the crest of the dunes and usually extend also some distance down the lee side. In some places the advance of the dune is sufficiently rapid to bring a portion of the thickets over to the windward side, where they are soon undermined (PI. XIX, Fig. 2). The further fate 'of these thickets is not known. It is worthy of note that they are somewhat similar in floristic composition to the thickets developing in certain blowouts, as described elsewhere in this paper (p. 107) and also in an earlier article (Hart and Gleason, iQoy: 168). Many of the species concerned are also characteristic of the mixed forest association and indicate a possible succession in that direction. Just at the margin of the thickets on the lee side, and partially shaded by them, patches of Smilacina stellata frequently occur. Tlie plant spreads by running rootstocks, but is not efficient as a sand- binder. The few patches on the windward side of the thickets are very soon undermined and destroyed. This small association en- croaches upon the deposit association in advance of the thickets, and is dependent upon the thickets for a partial protection from sunlight. It illustrates a peculiar case of succession in which an early stage is dependent upon a later stage for its existence and appears only after 144 the later stage (in this case the dune thicket association) is well de- veloped. Blowouts may be formed on the crest of the river dune in the usual way, and extend transversely through it. They seldom reach below the old soil bed which marks the limits of the middle slope. Their vegetation is of the usual type, except that the lateral slopes are frequently held by the plums and crabs of the thicket association. In some places in the Hanover area the river dune is occupied by the black oak association. The erosion there is generally feeble and the dune relatively stable. It seems probable that the oaks would also develop on the dunes stabilized by the ordinary deposit association if the thickets did not encroach upon them so rapidly. Smilacbia stcUata, as already mentioned, is a characteristic member of the black oak association, and its position on the dunes between the thicket and the deposit associations possibly indicates a potential development of the black oak forest at this place. In the Oquawka area the greater portion of the river dune is forested, and in parts of it the development of humus and the in- creased density of the ground cover has led to the establishment of a mesophytic type of forest, described already (p. 137) under the mixed forest association. This portion of the dune is no longer washed by the river itself, but by some sluggish bayous representing a former channel of the river and separated from the present channel by a number of densely wooded alluvial islands. At the foot of these islands the channel bends eastward against the foot of the dune and erosion is now proceeding rapidly. The plant covering is an efficient protection against wind erosion, and the dune would be completely stable if it was farther inland, but it can not resist the undermining efifect of the water. On a strip several hundred yards long the forest has been completely destroyed (PI. XX, Fig. i), and the vegetation now consists entirely of the blowsand and deposit associations. At the north end of the deforested portion the destruction of the forest is still proceeding. The effect of the erosion is first manifested at the foot of the dune, and its influence gradually extends higher until eventually the trees at the top are imdermined. There is thus pro- duced a triangular extension of the blowsand, extending like a wedge along the river between the water below and the forest above. It is now seen that the principal root development of the herbaceous vege- tation extends but one or two feet (3-5 dm.) below the surface, and binds the sand into a coherent stratum resting on the loo^e sand be- neath (PI. XX, Fig. 2). The loose sand rests at as steep an anele as possible, and irregular blocks of the surface layer become detached 145 and slide slowly down the incline toward the river. Their sides are nearly vertical, and by their detachment the margin of the remaining forest association is left as a prominent vertical wall of coherent sand. The motion of these detached blocks is of course very slow; but that they are loose is at once demonstrated by stepping on one, which then immediately starts down the slope and in a short time comes to rest on the flat beach at the base of the dune. Their plant population is a relic of the former mesophytic vegetation, and consists largely of perennials with a root system extensive enough to bind the mass to- gether. Some of the commoner species are Lespcdem capitata, Tradescantia reHexa, Monarda mollis, Solidago nemoralis, and Ar- tciiiisia caudata. The more pronounced mesophytes of course disap- pear with the removal of the protecting trees. The general trend of vegetation on the river dunes is therefore always toward stabilization, but their permanence is never certain because of the constant changes in the channel of the river. With the destruction of the higher types of vegetation by erosion, the pioneer blowsand association reappears and the successional cycle begins anew. The Perched Dunes In the Hanover area wind-blown sand has collected on top of the high bluffs which border the sand areas, and fonns miniature dunes and blowouts. A number of typical sand plants have colonized upon them, and are usually accompanied by the more resistant species of the uplands or of the rocky hillsides. In the blowouts, which are al- ways small, the vegetation represents the blowsand association and consists of Scutellaria parvula, Linaria canadensis, Monarda punc- tata, Verbena bracteosa. Ambrosia psilostachya, Fcstuca octodora, and Hedeonia hispida. On the stabilized dunes there are also Opnn- tia Rafinesqiiii, Artemisia caudata, Amorpha cancscens, Lithospcr- muin Guiclini, Rhus canadensis, var. ilUnoensis, Pauicuin pscudopu- besccns, Viola pedata, and Lespedeza capitata. In the sandy soil un- der the oaks are Cacalia atripUcifoUa, Hypo.vis hirsuta, Lithosper- muiii Ginclini, Phlo.r pilosa, Antennaria sp.. Anemone patens, var. lVolf(^angiana, Erigeron pulchellus, Poa pratcnsis, Corylus amcr- icana, and Junipcrns znrginiana. Annot.\ted List or Species No attempt was made to secure a complete collection or a com- plete list of the plants living in the sand regions, and the list given 146 here could be greatly extended by further observation. Only the seed-plants, ferns, and fern-allies are included, and the usual habitat of each species is given by associations. Many imusual locations of species are omitted. The nomenclature follows the Vienna Code, as exemplified in the seventh edition of Gray's Manual. Folypodiaccac Ptcris aqiiilina L. Winnebago, Amboy, and Havana areas, in the black oak association; Kankakee area, very abundant in the black oak forest and the inten'ening marshy meadows ; sometimes persist- ing as a relic in the bur oak association in the Winnebago area. Woodsia obfiisa (Spreng. ) Torr. Ocpawka area, in the mixed forest association, growing in dense shade on mats of moss on the mesophytic portions of the river dune. Not observed elsewhere in the sand region. Ophioglossaceae Botrychinm tcrnaimn (Thunb.) Sw., var. intermedium D. C. Eaton. Winnebago area, in the upland portions of the bur oak as^ sociation. Botrychiuiii I'irginiamim (L.) Sw. With the last species. Bquisetaceae Bqidsetiiin arvense L. Dixon area, in the Solidago association in extinct blowouts. Bquisctum hyemale L. Oquawka area, an invader from the al- luvial flood-plain vegetation into the mixed forest association on the river dune. Bquisctum hyemale L., var. inteniiediitm A. A. Eaton. Han- over, Dixon, and Havana areas, usually in the bunch-grass associa- tion ; sometimes growing in dense masses and aiding in the stabiliza- tion of blowout deposits ; abundant in the Solidago association in the Havana area. Selaginellaeeae Selagiiiella rupexlris (L.) Spring. Hanover area, in the bunch- grass association. It is frequently concerned in the fixation of sand and the re-establishment of the bunch-grass, and sometimes appears in the windward slo]5e association of the blowouts. The growth rings formed by this plant have been described in the text. 147 Pinaceae Juniperus virginiana L. Hanover area, frequent on the rocky exposed blufifs and from them invading the perched dunes. Graiiiincac Andropogon scoparius Michx. One of the most typical sand grasses in the Hanover, Amboy, Dixon, Ocjuawka, and Havana areas ; very frequent in the bunch-grass association and persisting from it as a rehc in the Paiiiciiiii pscudopiibescciis and the black oak associations. Andropogon fitrcatits I\Iuhl. An abundant and important grass, but by no means as common as the preceding species. Hanover, Amboy, Dixon, Oquawka, and Havana areas, normally in the bunch- grass association, but persisting as a relic in the black oak and Solidago associations, and sometimes appearing on blowout deposits. Sorghastruui nutans (L.) Nash. Hanover, Amboy, Oquawka, and Havana areas, in the bunch-grass association, and as a relic in the edge of the black oak association. Digitaria filifonnis (L. ) Koeler. Hanover, Oquawka, and Ha- vana areas, apparently not native, but coming in as a weed along roadsides or in too closely cropped pastures. Lcptoloina cognatitvi (Schultes) Chase. Abundant in each area except Winnebago and Kankakee, chiefly in the bunch-grass, where it may be dominant, also as a relic in the edge of the black oak association, in the Panicum pscudopubescens association, and on the windward slope of blowouts; it also appears early on blowout de- posits. Paspalinn sctaccuut Michx. Hanover, Dixon, Oquawka, and Havana areas, typically in the blowsand association, and continuing on the deposits, also as an interstititial in the bunch-grass and in bare spots at the edge of the black oak association. Panicum pscudopubescens Nash. Abundant in each of the five areas in a variety of situations ; common in the bunch-grass asso- ciation but usually as a secondary species ; characteristic of the association to which it gives its name; persisting as a relic in the blowout succession in the windward slope and deposit associations; frequent in open sunny places in the black oak forest ; rare in the Ijlowsand association; and, in the Dixon area, depauperate plants per- sist in the mats of Polytrichum. Panicum virgaium L. Common throughout but not abundant, usually in the bunch-grass association, but in the Hanover area one of the commonest dune-fonners on the blowout deposits or the crest 148 of the river dunes; rare in open places in the black oak association, or as a relic in other situations. Panicum pcrlongum Nash. Hanover, Havana, and Winnebago areas, common in the bunch-grass and Panicum pseudopuhescens associations, or in open places in the black oak association. Panicum Scribncrianum Nash. Only in the most mesophytic sta- tions of the bunch-grass association in the Havana, Hanover, Dixon, and Oquavvka areas; along roadsides and at the edge of the black oak forest in the Amboy and Winnebago areas. Sctaria glaiica (L. ) Beauv. Naturalized from Europe. Hanover area, a weed in pastured bunch-grass. Ccnchnis carolinianns Walt. Hanover, Dixon, Havana, and Oquawka areas ; regularly in the blowsand association or as an in- terstitial on blowout deposits. Stipa spartea Trin. In the bunch-grass association in the Han- over, Dixon, Havana, and Oquawka areas, more rarely on deposits or at the edge of black oak woods; also in a pastured field in the Winnebago area. Aristida basiramea Engelm. Oquawka area, according to Patter- son. Aristida tuberculosa Nutt. Hanover, Dixon, Havana, and Oquawka areas, common as an interstitial in the bunch-grass and Panicum pscudopubescens associations, very abundant and charac- teristic in the blowsand association, common at the edge of the black oak forest, and in the Dixon area abundant in the Solidago asso- ciation. Sporobolus cryptandrus (Torr. ) Gray. Chiefly in the blowsand and deposit associations of the Hanover, Havana, and Oquawka areas, sometimes in bare sunny spots in the black oak forest. Sporobolus heterolepis Gray. Oquawka area, according to Pat- terson. CalaniovUfa longifolia (Hook.) Hack. Dixon, Havana, and Oquawka areas, in the bunch-grass association or persisting as a relic in open places in the black oak association. Koclcria cristata (L.) Pers. Hanover, Winnebago, Dixon, and Oquawka areas, abundant and conspicuous in the bunch-grass ; per- sisting as a relic in the black oak and Panicum pscudopubescens associations; and in rare cases appearing on blowout deposits. Spartina Michauxiana Hitchc. Amboy area, along roadsides, doubtless adventive from the swampy meadows below. Boutcloua liirsuta Lag. Hanover, Havana, and Oquawka areas, common but inconspicuous in the bunch-grass, where it grows as an 149 interstitial between the larger grasses, rarely persisting in the Pani- cmn pseitdopiibescens and black oak associations. Bontcloiia oligostachya (Nutt.) Torr. Hanover area, according to Pepoon. Bonteloua curtipendula (Michx.) Torr. Havana and Oquawka areas, in the bunch-grass association. Tridciis i^az'us (L.) Hitchc. Havana area, chiefly in the bunch- grass, but also in the black oak association and rarely as a relic in the Panicum pseudopiibcscens association. Triplasis purpurea (Walt.) Chapm. Oquawka area, in the blow- sand association, according to Patterson. Eragrostis trichodes (Nutt.) Nash. Havana and Oquawka areas, typically in the bunch-grass but also in the black oak and deposit associations. Eragrostis pcctinacca (]\Iichx.) Steud. Hanover, Havana, and Oquawka areas, always in the bunch-grass association. Poa compressa L- Sunny places in the black oak association, Havana area. Poa pratcnsis L. In a large variety of situations in each area, but chiefly where the land has been pastured or along roadsides. Fcstnca octoftora Walt. Hanover, Winnebago, and Oquawka areas, chiefly as an interstitial in tlie bunch-grass, but also common in the blowsand association. Hordcnin pusillum Nutt. A weed along the roadsides in the Oquawka area. Elymils virginiciis L. Oquawka area, occasional in the bunch- grass association. Ely 111 us canadensis L. Hanover and Oquawka areas, in the bunch-grass association and sometimes as a relic in the Panicum pscudopuhescens association. Elymus striatus Willd. Oquawka area, a typically mesophytic species of the mixed forest association on the river dune. Cyperaceae Cyperus rividaris Kunth. Havana area, in the swamp association of an extinct blowout. Cyperus Sch'ii'cinit::ii Torr. Hanover, Dixon, Havana, and Oquawka areas, chiefly in the bunch-grass ; common also in the Panicum pscudopuhescens association and rarely in the blowsand. Cyperus Uliculmis Vahl. Hanover, Winnebago, Havana, and Oquawka areas, a common interstitial of the bunch-grass, frequent 150 in the blowsand and Panicuiii pseitdopubcscens associations, and more rarely in open places in the black oak woods. Eleocharis obtitsa (Willd.) Schultes. In the swamp association in the Dixon and Havana areas. Stciiophylliis capillaris (L.) Britton. The characteristic species of the Stcnophyllus association in the bottoms of partially stabilized blowouts in the Hanover, Havana, and Oquawka areas, rarely in the SoUdago and bunch-grass associations. Scirpus validiis Vahl. Around the margin of a pond in a depres- sion between the dunes, Winnebago area. Scirpus cypcriniis (L.) Kunth. Common in the swamp associa- tion and occasional in the Salix association in the Dixon area. Carex festucacea Schkuhr, var. brcvior (Dewey) Fernald. An interstitial in the bunch-grass association in the Hanover and Oquawka areas; not common. Cai-ex Mnhlenbcrgii Schkuhr. Abundant in the bunch-grass as- sociation in the Hanover, Havana, Dixon, and Oquawka areas, and sometimes becoming the dominant species ; one of the commoner bunch-grass relics in the Panicum pseiidopubesccns association ; in- frequent on the deposits and windward slopes of blowouts; in the black oak association in the Hanover, Winnebago, Havana, and Oquawka areas. Carcx umbcllata Schkuhr. Hanover, Havana, and Dixon areas, most abundant in the Panicum pseudopubescens association, persist- ing as a relic on the windward slopes, occasional in the bunch-grass association, and rare on the deposits of blowouts. Carex pennsylvanica Lam. Oquawka area, in the bunch-grass association ; Winnebago area, in open places in the islack oak asso- ciation. Carcx sp. Dixon area, in the Solidago association. CouniicUnaccac Cominclina virgiiiica L. Ha\ana and Oquawka areas ; one of the most abundant interstitial species, growing- in a wide variety of associations, but probably most abundant on blowout deposits. Tradescanlia reflcxa Raf. Hanover, Kankakee, Winnebago, Am- boy, Havana, and Oquawka areas ; common in the bunch-grass and black oak associations and persisting as a relic in the bur oak and Panicum pseudopubescens associations. 151 Jiincaceae JuncHs tenuis Willd. Havana area, in the Solidago asr-ociation ; Winnebago area, in open places in the black oak association. Juncxis nodosus L. Dixon area, in the swamp association. Jiincus aciiiiiinatiis Alichx. Dixon and Havana areas, in the Solidago association, and also less frequently in the Polytricliiini and Salix associations. Liliaceae Lilinm phUadclphicmn L., van, aiidininii (Nutt.) Ker. Winne- bago area, in the bur oak association. Asparagus officinalis L. Havana area, in the mixed forest asso- ciation. Smilacina racciiiosa- ( L. ) Desf. Hanover area, in the dune tliicket association ; Winnebago area, in the bur oak association ; Havana area, in the mixed forest association, and occasional!}' in the black oak association. Siuilacina stellata (L.) Desf. Hanover, Winnebago, Amboy, and Havana areas, characteristic of the black oak association, and in the last area rarely also in the mixed forest association. Polygonatum coinmutatitm (R. & S.) Dietr. Hanover, Winne- bago, Amboy, Oquawka, and Havana areas, characteristic of the bur oak and mixed forest associations, one of the earliest pioneers in the black oak forest, indicating the succession, and occasional in the blowout and dune thickets. Siuila.v hcrhacca L. Hanover, Winnebago, Amboy, and Havana areas, chiefly in the liur oak and mixed forest associations, occasional as a pioneer in the black oak association, and in the Hanover area in the dune thickets. Smilax ccirrhata (Engelm.) Wats. TIano\-er, Winnebago, and Amboy areas, in the bur oak association, or as a pioneer in the black oak forest. Smilax hispida Muhl. In the dune tliickets in the Hanover area, mixed forest in the Oquawka area, and in the lilack oak and bur oak forest in the Amboy area. Dioscovcaccac Dioscorca zillosa L. Havana area, in the nn'xed forest asso- ciation. Auiaryllidaceac Hypoxis hirsnta (L.) Coville. Hanover area, under oaks on the perched dunes. 152 Iridaceae Sisyrinchiiiin sp. Hanover and Oquawka areas, in the bunch- grass association and persisting as a reHc in the Pankum pscudopu- bescens association. Orchidaceae Cypripedinm parviHorum SaHsb., var. pubescens (Willd. ) Knight. Winnebago area, in the bur oak association. Spiranthes ccrniia (L. ) Richard. Dixon area, in the Polytrichmii association. Salica£eae Salix nigra Marsh. Dixon area, in the Salix association. Salix longifolia Muhl. Dixon and Havana areas, in the Salix association; Hanover area, in the dune thickets. Salix pedicellaris Pursh. Dixon area, in the Polytrichwn asso- ciation. Salix tristis Ait. Winnebago and Amboy areas, in the black oak association and as a rehc in the bur oak forests; Havana area, in the mixed forest association. Populus alba L. Oquawka area, frequently planted and escaped along roadsides and fence-rows. Populus tremuloidcs Michx. Winnebago, Amboy, and Kankakee areas, in the swamps and meadows between the sand hills, and oc- casional in the bur oak association. Populus grandidcntata Michx. Amboy area, in the bur oak as- sociation, or occasionally as a pioneer in the black oak woods. Popidus deltoides Marsh. Hanover area, in the dune thickets; Dixon area, in the Salix association; Oquawka and Havana areas, in blowout thickets. Juglandaceae Juglans nigra L. Hanover area, in the dune thickets; Oquawka area, in the mixed forest association on the river dune. Carya ovata (Mill.) K. Koch. Winnebago area, in the bur oak association. Carya cordifonnis (Wang.) K. Koch. Hanover area, in the mixed forest ; Oquawka and Havana areas, in the black oak association and persisting in the mixed forest. 153 Betulaceae Corylits amcricana Walt. Hanover area, on the perched dunes; Winnebago and Amboy areas, common in the bur oalv association and occasional as a pioneer in the black oak forest. Bctula nigra L- Hanover and Oquawka areas, in the mixed for- est association near the river. Bctula alba L., var. papyrifcra (Marsh.) Spach. Hanover area, on the perched dunes. Fagaceae Quercus alba L. Winnebago area, in the bur oak association. Quercus macrocarpa Michx. Winnebago and Amboy areas, the characteristic species of the bur oak association. Quercus rubra L. Oquawka area, in the mixed forest association. Quercus velutina Lam. Hanover, Winnebago, Amboy, Kanka- kee, Havana, and Oquawka areas, the characteristic species of the black oak association, persisting commonly as a relic in the mixed forest association and less commonly in the bur oak forest; common on the perched dunes in the Hanover area. Quercus marilandica Muench. Havana and Oquawka areas, abundant in the black oak association, and persisting in the mixed forest. Urticaceae Ulmus americana L. Hanover and Oquawka areas, in the mixed forest association and in the dune thickets. Celtis occidentalis L. Hanover and Havana areas, in the mixed forest association, and in the blowout and dune thickets. Morus rubra L. Havana area, in the mixed forest association. Boehmcria cylindrica (L.) Sw. Havana area, in the Sali.v as- sociation. Parictaria pennsylvanica Muhl. Oquawka area, in the mixed for- est association. Santalaceae Comandra iinihellata (L.) Nutt. Hanover, Winnebago, Amboy, Oquawka, and Hanover areas, in the black oak association, or oc- casionally along roadsides. Polygonaceae Rumex altissimus Wood. Hanover area, in the black oak asso- ciation. 154 Rumex Acctosclla L. Hanover, Winnebago, Dixon, and Oquawka areas, a common interstitial in the bunch-grass association, and in cultivated ground, less frequently in the Fankiim psciidopiibcsccns association. Rutnc.v sp. Hanover area, in the blowsand association. Polygonum avicularc L. Oquawka area, a weed along roadsides and in yards. Polygoniiin erectimi L. Oquawka area, a common roadside weed. Polygoiniin tcmtc Michx. Hanover, Oquawka, and Havana areas, a common interstitial in the bunch-grass association and less commonly also in the Paniciim pseudopuhcscens association. Polygonella articiilata (L. ) Meisn. Hanover and Dixon areas, in the blowsand association ; Hanover and Oquawka areas, in open places in the black oak association. Chenopodiaceac Cydoloma atriplicifolium (Spreng.) Coult. Havana and Oquaw- ka areas, in the blowsand and deposit associations. Chcnopodiiim album L. Hanover, Havana, and Oquawka areas, common as a weed, occasional in the bunch-grass association and in open places in the black oak forest. Ainamnthaceae Froclichia floridana (Nutt.) Moq. Dixon, Oquawka, and Havana areas, usually in the blowsand association, occasionally an interstitial in the bunch-grass association or open blowout deposits ; Hanover area, along- the railroad track, appearing as if introduced. Phytolaccaceae Phytolacca decandra L. Oquawka area, in waste places under the shade of trees. Nyctaginaceae O.vybaphus nyctagincus (Michx.) Sweet. Hanover, Dixon, and Oquawka areas, in the blowsand and blowout thicket associations, an interstitial in the bunch-grass, and frequent as a weed in waste places and along roads. Illcccbvaceae Anycliia polygonoidcs Raf. Hanover, Havana, and Oquawka areas, in the black oak association. 155 Anycliia canadensis (L-) BSP. Oquawka area, in the mixed for- est association. Aisoaceae Mollugo verticillata L. Hanover, Dixon, Oquawka, and Havana areas, especially common in the blowsand association and occasional as an interstitial in the bunch-grass and Panicum pseudopubcsccns associations. Caryophyllaceae Arenaria lateriflora L. Winnebago area, in the bur oak associa- tion. Silene antirrhina L- Hanover, Winnebago, Dixon, Havana, and Oquawka areas, an abundant weed in fields and a common inter- stitial in the bunch-grass and Panicum pseudopithesccns associations. Silene stellata (L.) Ait. f. Hanover, Winnebago, Havana, and Oquawka areas, characteristic of the bur oak and mixed forest asso- ciations and a pioneer in the black oak association. Saponaria officinalis L. Hanover area, in the black oak associa- tion near dwellings. Portulacaceae Talinum rugosperniiiin Holzinger. Winnebago, Oquawka, and Havana areas, in the black oak association; Hanover area, in the bunch-grass association. Raniniciilaccac Ranunculus abortivus L. Hanover area, in the mixed forest as- sociation. Ancuione patens L., var. IVoIfgangiana (Bess.) Koch. Hanover area, on the perched dunes. Anemone caroliniana Walt. Hanover area, in the Iwnch-grass association. Anemone cylindrica Gray. Hanover and Oquawka areas, in the bunch-grass and black oak associations; Winnebago and Amboy areas, in the black oak association. Anemone I'irginiana L. Havana area, in the mixed forest as- sociation and appearing as a pioneer in the black oak association. Anemone canadensis L. Oquawka area, in the mixed forest as- sociation. AquiJcgia canadensis L. Hanover and Oquawka areas, in the mixed forest association. 156 Delphinium Penardi Huth. Oquawka area, in the bunch-grass association. Menispennaceae Memspennum canadensc L. Hanover area, in the dune thicket and mixed forest associations; Havana area, in the blowout thicket association. Pumariaccac Corydalis micrantha (Engelm.) Gray. Hanover area, in the blowout association on the river dune. Cruciferae Draba caroliniaiia Walt. Hanover area, in the bunch-grass as- sociation, and in sheltered places in the blowsand association on the river dune. Lesqnerella argentea (Pursh) MacM. Havana area, in the bunch- grass association. Lepidinm virginicum L. Hanover, Winnebago, Dixon, Oquawka, and Havana areas, especially abundant in the Panicum pseudopn- bescens association, a common interstitial in the bunch-grass associa- tion, occasional in the blowsand association, and common in open places in the black oak association. Brysimiwi parviftorum Nutt. Havana area, in the biuich-grass association. Arabis lyrafa L- Hanover, Winnebago, and Oquawka areas, most abundant as an interstitial in the bunch-grass and Panicum psciidopiibcscciis associations! in the Hanover area, also in open places in the black oak association. Capparidaceae Polanisia graveolens Raf. Hanover, Oquawka, and Havana areas, frequent in the blowsand association and in the black oak for- est. Crisfatclla Jamesii T. & G. Hanover and Havana areas, in the blowsand association. Saxifragaceae Hcnchera hispida Pursh. Winnebago area, in the bur oak asso- ciation ; Hanover and Oquawka areas, in the mixed forest associa- tion. 157 Ribes gracile Michx. Hanover and Havana areas, in the mixed forest association, occasional in the dune thickets and the black oak forest, or along fence-rows on the prairie. Rosaceae Spiraea salicifolia L. Dixon area, in the Solidago association. Pynts ioensis (Wood) Bailey. Hanover area, in the dune thickets. Pyrus Mains L. Hanover area, in the blowout thicket association. Pyriis auicricana (Marsh.) DC. Winnebago area, in the black oak association. Fragaria virginiana Duchesne, var. illinoensis (Prince) Gray. Winnebago and Amboy areas, in the black oak association and as a relic in the bur oak association ; Hanover area, in the mixed forest. Fragaria vesca L., var. auiericana Porter. Winnebago area, in the bur oak association ; Hanover area, on the perched dunes. Potciitilla argufa Pursh. Winnebago and Amboy areas, in the black oak association and along roadsides. Potcntilla argcntea L. Winnebago area, in a pastured field. Potentilla canadensis L. Winnebago area, in the black oak asso- ciation. Geiim canadcnse Jaccj. Hanover area, in the mixed forest asso- ciation. Riibns idaciis L-, var. aculeatissimus (C. A. Mey.) Regel & Til- ing. Hanover area, in the mixed forest association ; Winnebago area, in the bur oak association. Riibus occidcntalis h. Hanover area, in the black oak and mixed forest associations and on the perched dunes; Winnebago area, in the black oak and bur oak associations. Rubiis sp. (Blackberry). Amboy area, in the bur oak association. Agrinionia mollis' (T. & G.) Britton. Winnebago and Amboy areas, in the bur oak association ; Havana area, in the mixed for- est association. Rosa humilis Marsh. Hanover area, in the black oak and blow- out thicket associations ; Winnebago area, in the black oak forest and as a relic in the bur oak association ; Dixon area, in the bunch- grass association. Primus serotina Ehrh. Hanover area, in the dune thickets; Winnebago area, in the bur oak association and as a pioneer in the black oak forest; Havana area, in the mixed forest association. Primus virginiana L. Hanover area, in the mixed forest and 158 dune thicket associations; Winnebago area, in the bur oak asso- ciation ; in both areas as a pioneer in the black oak association. Pniiiiis sp. (Plum). Hanover area, in the dune thickets. Leguminosae Gyiiinoclailiis dioica (L.) Koch. Havana area, in the mixed for- est association. Gleditsia triacanthos L. Hanover area, in the dune thicket asso- ciation ; Dixon area, planted on the deposits of a blowout ; Oquawka area, along roadsides. Cassia Chaiiiacchrisfa L. Most abundant in the Havana and Ocjuavvka areas, in the bunch-grass, blowsand, and black oak asso- ciations ; Hanover area, in the blowsand association ; Amboy area, in the black oak forest ; Dixon area, in the Solidago association. Ccrcis canadensis L. Ocjuawka area, in the mixed forest asso- ciation. Baptisia bractcafa (Muhl.) Ell. Winnebago area, in the black oak association ; Oquawka area, in the bunch-grass and Paiiicmn pscudopitbcscens associations. Liipimis pcrennis L. Winnebago and Amboy areas, in the black oak association ; Kankakee area, in mucky meadows at the base of sand hills. Trifoliiun pratcnsc L. Oquawka area, roadsides. Trifolimn repens L. Oquawka area, along roadsides. Anwrpha caiicsccns Pursh. Hanover, Winnebago, Amboy, Dixon. Kankakee, Oquawka, and Havana areas, most abundant in and typi- cal of the bunch-grass association, persistent as a relic and common in the black oak association, and occasional in the bur oak and mixed forest associations ; in the Planover area, also on the perched dunes. Pctalostcuium purpuvcuui (Vent.) Rydb. Hanover and Oquawka areas, abundant in the bunch-grass ; also in the black oak association in the Amboy and Oquawka areas ; along roadsides in the Winne- bago area. Petalostcinum candid inn Michx. Hanover and Oquawka areas, in the bunch-grass association ; Winnebago and Amboy areas, in the black oak association. TcpJirosia tirginiana (L.) Pers. In all seven areas; abundant in the bunch-grass and black oak associations, frequent on blowout deposits, and occasional in the blowsand and Panicnin pscndo- pubesccns associations. Robinia Pseudo-Acacia L. Oquawka area, commonly planted as 159 a sand-binder and escaping into the mixed forest and blowout thicket associations. Dcsnwdiiim grandiflonim (Walt.) DC. Winnebago area, in the bur oak association. Desmodium UlUwensc Gray. Oquawka area, in tlie Inmch-grass association ; Amboy area, in the black oak forest. Lcspcdeza capitata Michx. Common in all seven areas, chiefly in the bunch-grass and black oak associations ; a relic in the Pani- cum pseudopubescens association, active in the stabilization of all parts of the blowouts ; in the Dixon area it appears in the Solidago association ; and in the Hanover area, on the perched dunes. Strophostyles helvola (L.) Britton. Havana area, in the black oak forest; Oquawka area, in the bunch-grass, blowsand, black oak, and mixed forest associations. Strophostyles sp. Havana area, in the bunch-grass association. I4mphicarpa Pitcheri T. & G. Hanover area, in the mixed for- est association ; Winnebago area, in the bur oak association. Linaceac Linuni sulcatum Riddell. Hanover and Oquawka areas, in the bunch-grass association. Oxalidaceae Oxalis cornicidata L. Hanover area, in the bunch-grass and black oak associations. Geraniaceae Geranium luaculatum L. Winnebago area, in the bur oak as- sociation. Rutaccac Zanthoxylum aniericanum Mill. Oquawka area, in the mixed forest association. Polygalaccac Polygala polygama Walt. Hanover, Winnebago, Dixon, and Oquawka areas, in the bunch-grass, Panicum pscudopuhcscens, and black oak associations. Polygala htcarnata L. Oquawka area, in the bunch-grass asso- ciation. Polygala sangttinca L. Dixon area, in the Solidago association, 160 and as a relic also in the Polytrichmn association; Amboy area, in muck meadows at the base of the dunes. Polygala verticiUata L. Hanover, Havana, and Oquawka areas, in the bunch-grass and Panicum pseitdopubesccns associations. Bii'phorhiaccae Croton glandulosiis L., van septcntrionalis Muell. Arg. Han- over, Havana, and Oquawka areas, an interstitial in the bunch-grass and Panicum pscudopubcscens associations, common in the blowsand association, and occasional on blowout deposits. Crotonopsis linearis Michx. Havana area, a common interstitial in the bunch-grass, Panicum pscudopubcscens, blowsand, deposit, and black oak associations. Euphorbia Gcycri Engelm. Hanover, Dixon, Havana, and Oquawka areas, most abundant in the blowsand association, occa- sional as an interstitial in tlie bunch-grass and the black oak forest. Euphorbia corollata L. Very abundant in all seven areas, chiefly in the bunch-grass and black oak association, frequent on blowout deposits, occasional in the blowsand association, and rare in the bur oak association. Anacardiaceae Rhus glabra L. Winnebago and Amboy areas, in the bur oak as- sociation ; Hanover area, in the mixed forest association. Rhus Toxicodendron L. Hanover, Oquawka, and Havana areas, in the mixed forest association, in the dune thickets, and occasionally a pioneer in the black oak forest; Winnebago area, in the bur oak and black oak associations. Rhus canadensis Marsh., yar. illinoensis (Greene) Fernald. Han- over, Havana, and Oquawka areas, in the bunch-grass, deposit, and black oak associations. Celastraceae Cclastrus scandcns L. Havana and Oquawka areas, in the mixed forest; Hanover area, in the dune thickets. Aceraccae Acer saccharinum L. Oquawka area, in the mixed forest associa- tion on the river dune. Acer Negundo L. Havana area, in the blowout thicket asso- ciation. 161 Rhamnaceae Ceanothtis aincricantis L. Hanover, Winnebago, Amboy, Oquawka, and Hanover areas, in the black oak, bur oak, and mixed forest associations; occasional in the bunch-grass in the Oquawka area. Ceanothns ovatiis Desf. Hanover area, in the bunch-grass, de- posit, and black oak associations. f 'itaceae Psedera quinquefolia (L.) Greene. Winnebago area, in the bur oak association ; Hanover, Havana, and Ocjuawka areas, in the mixed forest, blowout thicket, and dune thicket associations, and one of the most frequent pioneers in the black oak association. Vitis z'ldpbia L. Winnebago and Amboy areas, in the bur oak association; Havana, Oquawka, and Hanover areas, in the mixed forest, the blowout thickets, and the dune thickets, and occasional as a pioneer in the black oak association. Malvaceae CaUirhoc triaiigulata (Leavenw.) Gray. Hanover and Havana areas, chiefly in the bunch-grass and the black oak forests, and occa- sionally in the Panicmn psendopnbcscens association. Hypcricaccac Hypericum cistifolium Lam. Oquawka area, in the mixed for- est association along the river dune. Hypericum mufilum L. Havana area, in the SoHdago asso- ciation. Hypericum mcijits (Gray) Britton. In the Polytrichum asso- ciation in the Dixon area. Hypericum gentianoides (L.) BSP. Dixon area, in the Poly- trichum association, or perhaps more common in a zone just outside of it. Cistaceae Hclianthciuuiii majus BSP. Hanover, Winnebago, Amboy, Dixon, Havana, and Oquawka areas, probably most widely distributed in the black oak association, but also abundant in the bunch-grass. Hudsonia iomcntosa Nutt. Hanover and Dixon areas, cliarac- 162 teristic of the Hudsonia association, and occasional npon blowout deposits and in black oak woods. Lcchea sp. In the black oak association in the Winnebago and Oquawka areas. Violaccae Viola pcdaia L. Hanover, Oquawka, and Winnebago areas, most abundant in the bunch-grass prairie and in the black oak woods, oc- casional in the Panicwn pseudopithcsccns association and on blowout deposits. VioJa lanceolata L. Dixon area, in the Polytrichuin association ; Kankakee area, in the wet meadows. Cactaceae Opiintia RaiiJicsqiiii Engelm. Hanover, Oquawka, and Havana areas, usually very abundant in the bunch-grass and the open parts of the black oak forest. Opiintia fragilis (Nutt.) Haw. Hanover area, in the bunch-grass and Paiiiciini psciidapiibcscciis associations. Melasfoiiiaccac Rhe.via virginica L. Dixon area, in the Polytrichum and Sol- idago associations ; Amboy area, in the wet meadows. Onagraccac Litdz'-igia alternifolia L. Havana area, in the Salix association. Liidvigia paliisfris (L.) Ell. Dixon and Havana areas, in the swamp association, and persistent as a relic in the Solidago and Polytrichum associations. Oenothera biennis L. In the l:)ur oak association in the Winne- bago area, and in the mixed forest in the Havana area. Oenothera rhonibipctahi Nutt. One of the most common inter- stitials, occurring in all seven areas in a wide variety of associations, but most abundant in the bunch-grass and in the Panicinn pseiidopu- bescens association. Circaca lufetiana L. In the bur oak forest in the Winnebago area. Uiiibciliferac Saniciila canadensis L. Winnebago and Havana areas, in the bur oak and mixed forest associations. 168 Zisia aurea (L.) Koch. Winnebago area, in the black oak for- est ; Hanover area, in tlie mixed forest. Cornaceae Corniis Baileyi Couh. & Evans. Hanover, Winnebago, Amboy, Havana, and Oquawka areas, in the bur oak and mixed forest asso- ciations ; also in the dune thickets and the blowout thickets. Ericaceae Pyrola eltipika Nutt. In the bur oak association in the Winne- bago area. Monotropa uiiiflora L. Winnebago area, in the black oak forest, probably a pioneer from the bur oak association. Priiiiiilaceae Steironema lanceolatinn (Walt.) Gray. Winnebago area, in the swamp association. Dodecatheon Mcadia L. Winnebago area, in the bur oak asso- ciation. Oleaeeae Fraxiniis pcnnsylvanica Marsh., var. lanceolata (Borkh.) Sarg. Hanover and Oc|uawka areas, in the mixed forest and dune thicket associations and as a pioneer in the black oak forest. Apocynaceae Apocynmn aiidrosaemifolium L. Winnebago and Amboy areas, in the black oak forest and persisting as a relic in the bur oak asso- ciation ; Hanover area, in the mixed forest. Apocyiiuiii caiinabiniim L-, var. liypericifoliuni (Ait.) Gray. In blowsand in the Dixon area. Asclepiadaceae Asclepias tiiberosa L. Hanover, Havana, and Amlx)y areas, in the black oak forest ; Oquawka area, in the mixed forest ; Winne- bago area, along a sandy roadside. Asclepias syriaca L. Hanover and Oquawka areas, near culti- vated grounds around the river dune; Havana area, in the black oak and Solidago associations. 164 Asclepias amplcxicaulis Sm. In the black oak forest in the Han- over, Winnebago, Amboy, Havana, and Oquawka areas; also in the bunch-grass association in the Dixon and Oquawka areas. Asclcpias phytolaccoidcs Pursh. Hanover and Havana areas, in the mixed forest association. Asdcpias verticillata L. In the black oak and bur oak forests in the Havana area, and along roadsides in the Hanover area. Aceratcs Horidana (Lam.) Hitchc. In the windward slope as- sociation in the Dixon area, probably a relic from the bunch-grass association. Aceratcs znridiflora Ell. Hanover and Oquawka areas, in the bunch-grass and basin associations, and occasional in the Panicum psciidopuhesccns association; Winnebago area, in an open place in the black oak forest. Aceratcs viridiHora Ell., var. lan^colata (Ives) Gray. Hanover and Havana areas, characteristic of the basin association, and occa- sional in the bunch-grass and Panicwn pseitdopitbcscciis associations. Acerates viridHiora Ell, var. linearis Gray. Hanover and Oquawka areas, in the basin and blowsand associations. Convolvidaccae Bretveria Pickcringii (M. A. Curtis) Gray. In the bunch-grass prairies of the Oquawka area. Iponioea hederacea Jacq. A weed in cultivated fields in the Oquawka area. Polcinoniaccac Phlox pilosa L. Perched dunes in the Hanover area. Phlox bifida Beck. In the bunch-grass and black oak associations in the Havana and Winnebago areas. Boraginaceae Lappida virginiana (L.) Greene. Havana area, in the mixed forest association. Lithospermnm Gmclini (Michx.) Hitchc. Abundant in all seven areas, chiefly in the bunch-grass and black oak associations, more rarely in the basin, blowsand, and Panicum psciidopuhesccns asso- ciations. Lithospcrnniui angnstifolium Michx. In the bunch-grass asso- ciation in the Havana area. _ 165 Verbenaceae Verbena stricta Vent. Havana, Oquawka, Dixon, and Hanover areas, occasional in the bunch-grass and black oak forest, and a weed along roadsides. Verbena bracteosa Michx. Perched dunes and bunch-grass in the Hanover area. Labiatae Teiicrium canadense L. Havana and Hanover areas, in the black oak and Solidago associations. Teiicriuni occidentale Gray. Hanover area, an interstitial in the bunch-grass, and common in tlie blowsand association along the river dune. Scutellaria parvula Michx. In tlie black oak association in the Hanover, Amboy, and Havana areas ; also an interstitial in the bunch-grass in the Hanover area, and in the blowsand association in the Dixon area. Nepeta Cataria L. Hanover area, in the black oak forest near dwellings. Physostegia denticidata (Ait.) Britton. Oquawka, Havana, and Hanover areas, in the bunch-grass and black oak associations. Leonurus Cardiaca L. Hanover area, near dwellings in the black oak forest. Monarda iisfulosa L. Oquawka area, in the black oak and mixed forest associations. Monarda mollis L. Hanover, Winnebago, Amboy, Havana, and Oquawka areas, usually in the bur oak forest, but occasionally as a pioneer in the black oak association. Monarda punctata L. Hanover, Dixon, Havana, Oquawka, and Kankakee areas, one of the most abundant interstitials in the bunch- grass, and common also in the Panicum pseudopubcsccns and blow- sand associations and in open places in the black oak forest; very common as a weed in pastured ground. Hedeonia hispida Pursh. Hanover, Dixon, and Oquawka areas, an interstitial in the bunch-grass, and common also in the Panicum pseudopubcsccns and blowsand associations. Lycopus americanus Muhl. Havana area, in the Salix associa- tion ; Dixon area, in the Polytrichum association ; Amboy area, in the wet meadows. 166 Soliuiaccae Solatium nigrum L. Hanover, Oquawka, and Havana areas, un- der the shade of trees. Solamiui caroliiicusc L. Hanover and Oquawka areas, in the black oak forest and a weed in cuUivated fields. Physalis heterophylla Nees. Hanover, Winnebago, Dixon, and Havana areas, in the bunch-grass and black oak associations. Physalis virginiana Mill. Hanover and Ocjuawka areas, in the black oak and bunch-grass associations. Scrophulariaceae Verhascuyii Thapsiis L. Hanover, Winnebago, Oquawka, and Havana areas, usually in the black oak forest, but occasionally in the bunch-grass association. Liuaria canadensis (L.) Dumont. Hanover, Dixon, Oquawka, and Havana areas, a common interstitial in the bunch-grass, fre- quent in the Panicum pscudopubesccns and blowsand associations, and occasional on blowout deposits. Scrophidaria Icporclla Bicknell. Hanover, Winnebago, Oquawka, and Havana areas, most abundant in the mixed forest and in the; dune thickets, less frequent in the black oak forest. Pcntstciiion Iiirsutits (L.) Willd. Hanover, Amboy, Havana, and Oquawka areas, common in the bunch-grass and the black oak forest, and occasional in the Panicum pscudopubesccns and bur oak associations. Pcntstcmon graudiflorus Xutt. In the bunch-grass and black oak associations of the Oquawka area. Veronica virginica L. Winnebago and Amboy areas, in the bur oak association. Syntliyris Bidlii (Eaton) Heller. Hanover, Winnebago, and Oquawka areas, characteristic of the black oak association and oc- casional as a relic in the bur oak forest, rare in the bunch-grass. Gerardia grandiflora Benth. In the black oak association in the Winnebago and Amboy areas, and as a relic in the bur oak forest. Gerardia- purpurea L. In the Solidago association in the Dixon area. Castillcja coccinea (L.) Spreng. Winnebago area, in the black oak association. Pcdicularis canadensis L. Winnebago area, in the black oak and bur oak associations. 1G7 Orobanchaceac Orobanche fasciculata Nutt. Parasitic on Artemisia caudata in the bunch-grass association of the Hanover area. Acanthaceae Ruellia ciliosa Pursh. Havana and Oquawka areas, in the bunch- grass and black oak associations. Plantaginaceae Plantago RugcUi Dene. Along roadsides in the Oquawka area. Riibiaceae Galium pilosum Ait. In the black oak forest in the Havana area. Galium concinnum T. & G. Winnebago area, in the bur oak association ; Havana area, in the mixed forest. Diodia teres Walt. Hanover and Havana areas, common in the blovvsand association, and occasional in the black oak forest and as a weed in cultivated ground. Caprifoliaceae Lonicera Sullivantii Gray. Winnebago area, in the bur oak asso- ciation. Campamdaceae Specularia perfoliata (L.) A. DC. Hanover, Havana, and Win- nebago areas, a common interstitial in the bunch-grass, occasional in open places in the black oak forest and one of the most abundant weeds in sandy fields. Coinposilae Vcnw)iia fasciculata Michx. Havana area, in the Solidago asso.- ciation ; Amboy area, in the wet meadows between the dunes ; Han- over area, in pastured bunch-grass. Eupatorium purpureiim L. Havana area, in the mixed forest association ; Amboy area, in the wet meadows at the base of the dunes. Eupatorium scrotimim Michx. Hanover area, in the black oak and mixed forest associations. Eupatorium urticacfolium Reichard. In the mixed forest asso- ciation in the Hanover and Havana areas. Kuhnia eupatorioides^ L., var. corymbulosa T. & G. Hanover, Oquawka, and Havana areas, chiefly in the bunch-grass, but occa- sional in the blowsand and dune thicket associations. 168 Liatris cyUndracca Michx. Hanover area, in the bunch-grass association ; Winnebago and Amboy areas, in the black oak forest. Liatris scariosa Willd. In the bunch-grass association in the Hanover, Dixon, Oquawka, and Havana areas ; in the black oak for- est in the Winnebago and Amboy areas. Chrysopsis viUosa Nutt. Hanover, Dixon, and Havana areas, common in the bunch-grass association ; along roadsides in the Amboy area. Solidago speciosa Nutt., var. angusiata T. & G. Winnebago, Amboy, Hanover, and Havana areas, in the black oak forest; Han- over and Oquawka areas, in the bunch-grass association. Solidago missouriensis Nutt. In the bunch-grass association in the Havana and Hanover areas. Solidago nemoralis Ait. Abundant in all seven areas, in the black oak and bunch-grass associations ; in the Hanover area, also in the Paniciiiii psciidopnbcsccns association. Solidago serotina Ait. Hanover and Havana areas, in the black oak association. Solidago rigida L. In the bunch-grass association in the Han- over area; along roadsides in the Amboy area. Solidago graminifolia (L.) Salisb. Characteristic of the Soli- dago association in the Dixon and Havana areas; along sandy road- sides in the Amboy area; in a swamp between the dunes in the Win- nebago area. Aster oblongifolius Nutt. Oquawka area, in the mixed forest on the river dune. Aster sericeus Vent. In the Ininch-grass association in the Han- over, Havana, and Oquawka areas; in the black oak forest in the Hanover and Amboy areas; in the mixed forest on the river dune in the Oquawka area; in a cleared field in the Winnebago area. Aster a::iireits Lindl. Winnebago, Amboy, Havana, and Oquawka areas, in the black oak association. Aster vinlfifJonis Ait. Havana, Oquawka, and Hanover areas, in the bunch-grass association. lAstcr Vuiariifoliits L. Common in all seven areas in the bunch- grass and black oak associations, and occasionally persisting as a relic in the Panicum pscudopitbcsccns association. Aster sp. In the bunch-grass association in the Hanover area. Aster sp. Dixon area, in the Polytrichum association. Brigeron pulchellns Michx. On the perched dunes in the Han- over area. 169 Brigeron ramosus (Walt.) BSP. Hanover and Oquawka areas, in the bunch-grass and Panicnm pscndopubcsccns associations; Win- nebago area, in the black oak forest. Brigeron canadensis L. An interstitial in the bunch-grass in the Hanover area, and on blowout deposits in the Oquawka area. Antennaria plantaginifolia (L.) Richards. Hanover area, in the mixed forest association. Antennaria sp. One or more unidentified species of Antennaria are common in the bunch-grass and black oak forests of the Hanover, Winnebago, Amboy, Dixon, Oquawka, and Havana areas. Gnaphaliiini polyccphalum Michx. Hanover area, in the bunch- grass ; Amboy and Oquawka areas, in the black oak forest ; in the latter area also in the Stcnophyllns and blowout thicket associations. Silphinm laciiiiatnm L. Along roadsides in the Amboy area. Silphimn integrifolium Michx. Amboy area, in the black oak forest. Parthcninm integrifolium L- Hanover area, in the mixed forest association; Amboy area, along sandy roadsides. Ambrosia artemisiifolia L. A weed in the waste grounds in the Oquawka and Winnebago areas. Ambrosia psilostachya DC. Observed in the Hanover, Winne- bago, Dixon, Oquawka, and Havana areas, and probably in the oth- ers as well ; a common interstitial in the bunch-grass, black oak, and Panicnm pseudopitbcsccns associations, and abundant in the blow- sand association. Xantliium coinmunc Britton. A weed in sandy fields in the Oquawka area. Rudbcckia hirta L. Hanover, Winnebago, Havana, and Oquawka areas, abundant in the black oak forest, and occasional in the mixed forest and bunch-grass. Branneria pallida (Nutt.) Britton. Hanover, Dixon, and Oquawka areas, in the bunch-grass association ; Amboy area, along roadsides. Lcpachys pinnata (Vent.) T. & G. Winnebago area, in a clear- ing in the black oak forest. Hciianthiis lenticnlaris Dougl. Ocjuawka area, in the blowsand association. Helianthus scaberrimus Ell. Hanover and Oquawka areas, com- mon in the bunch-grass and Panicnm psendopuhcscens associations; Amboy area, along sand roadsides. Helianthus occidentalis Riddell. Common in all seven areas in the Ininch-grass and black oak associations. 170 Helianthiis occidentalis Riddell, van ilUnocnsis (Gleason) Gates. With the species, especially in more shaded places ; occasional in the mixed forest association. Hclianthus strumosiis L. Winnebago, Amboy, and Havana areas, in the black oak forest, and persisting as a relic in the bur oak and mixed forest associations. Coreopsis palmata Nutt. Hanover, Winnebago, Amboy, and Oquawka areas, in the bunch-grass and black oak associations, rare as a relic in the bur oak forest ; Havana area, in the mixed forest. Hyiiiciwpappus carolincnsis (Lam.) Porter. In the black oak association in the Kankakee area. Achillea Millefoliimi L. Winnebago, Amboy, Dixon, and Oquawka areas, in the bunch-grass and black oak associations. Anthemis Cotnla L. A weed in the Oquawka area. Artemisia caudata Michx. Very common in all seven areas in the bunch-grass and black oak associations, occasional as a relic in the Paiiiciiiii pscndopiibcscens association. Artemisia ludoviciana Nutt. Hanover area, occasional in the bunch-grass and black oak forest; more abundant in shaded places along fence-rows and thickets. Cacalia atriplicifolia L. Hanover area, on the perched dunes; Winnebago area, in the bur oak forest ; Havana area, in the black oak and mixed forest associations, and occasionally in the bunch- grass. Scnccio Balsainitac Muhl. Hanover and Oquawka areas, in the bunch-grass and black oak associations. Krigia virginica (L.) Willd. Havana area, in the black oak as- sociation. Krigia amplexicauUs Nutt. In the bur oak forest in the Winne- bago area. Lacfiica scariola L., var. integrata Gren. & Godr. Oquawka area, in the Stcnophyllus association. Lactuca canadensis L. Hanover, Havana, and Oquawka areas, in the bunch-grass, occasional in the blowout thickets and the black oak forest. Prenanfhcs alba L. Winnebago area, in the bur oak association, or as a pioneer in the black oak forest. Hieracitim longipilum Torr. In the black oak forest in the Win- nebago area. Hicraciuni canadcnse Michx. Hanover area, in the black oak as- sociation. BIBLIOGRAPHY Adams, C. C. 1902. Southeastern United States as a center of geographical distribution of flora and fauna. Biol. Bull, j: 115-131. 1905. The postglacial dispersal of the North American biota. Biol. Bull, p: 53-71. Bennett, F., and Ely, C W. 1905. Soil survey of Marshall county, Indiana. Field Opera- tions of the Bureau of Soils, 6: 689-706, map jo. Bonser, T. A. 1903. Ecological study of Big Spring prairie, Wyandot county, Ohio. Ohio State Acad. Sci., Special Paper 7. Bonsteel, J. A. 1903a. Soil survey of Tazewell count)-, Illinois. Field Opera- tions of the Bureau of Soils, ./ : 465-489, map 2^. 1903&. Soil survey of the Janesville area, Wisconsin. Field Op- erations of the Bureau of Soils, 4. : 549-570, map 2j. Britten, W. E. 1903. Vegetation of the North Haven sand plains. Bull. Torr. Bot. Club jo: 571-620, pi. 23-28. Chamberlin, T. C, and Salisbury, R. D. 1885. The driftless area of the upper Mississippi. U. S. Geol. Surv., Ann. Rep. 6: 205-322, pi. 2j-2(), f. 26-48. Clements, F. E. 1904. The development and structure of vegetation. Bot. Surv. Nebraska, 7: 3-175. 1905. Research methods in e.colog\'. Lincoln, Neb. Coffey, G. N., Ely, C. W., and Mann, C. J. 1904. Soil survey of Winnebago county, Illinois. Field Opera- tions of the Bureau of Soils, 5: 753-775, map 59. Cowles, H. C. 1899. The ecological relations of the vegetation on the sand dunes of Lake Michigan. Bot. Gaz. 2^: 95-117, 167-202, 281- 308, 361-391, /. 1-26. 1901. The physiographic ecology of Chicago and vicinity; a study of the origin, development, and classification of plant societies. Bot. Gaz. j/: 73-108, 145-182,/. i-25- 171 172 Engler, A. 1902. Die pflanzengeographische Gliederung Nordamerikas. Notizbl. d. Konigl. Bot. Gart., Appendix IX. Gradmann, R. 1909. Ueber Begriffsbildung in der Lehre von den Pflanzenfor- mationen. Engler's Jb. 4^, Beibl. 99: 91-103. Harper, R. M. 1906. A phytogeographical sketch of the Altamaha grit region of the coastal plain of Georgia. Ann. N. Y. Acad. Sci. //: 1-415, pl- 1-28, f. 1-17, map. Harshberger, J. W. 19CX). An ecological study of the New Jersey strand flora. Proc. Acad. Nat. Sci. Philadelphia 1900: 623-671. Hart, C. A., and Gleason, H. A. 1907. On the biology of the sand areas of Illinois. Bull. 111. State Lab. Nat. Hist. 7: 135-273, pi. 8-2j, map. Harvey, L. H. 1908. Floral succession in the prairie-grass formation of south- eastern South Dakota. Bot. Gaz. 46: 81-108, 277-298, /. ^-3, 1-4- Henry, A. J. 1906. Climatology of the United States. U. S. Dept. Agr., Weather Bureau, Bull. Q. Imlay, G. 1797. A topographical description of the western territory of America. Third edition. London. Jaccard, P. 1902. Gesetze der Pflanzenvertheilung in der alpinen Region Flora 90 : 349-377- Jennings, O. E. 1908. An ecological classification of the vegetation of Cedar Point. Ohio Nat. 8: 291-340, /. 1-22. 1909. A botanical survey of Presque Isle, Erie county, Pennsyl- vania. Ann. Carnegie Mus. 5: 289-421, pi. 22-^1. Leverett, F. 1899. The Illinois glacial lol^e. U. S. Geol. Sui-v., Monograph XXXVIII. Livingston, B. E. 1903. The distribution of the upland plant societies of Kent county, Michigan. Bot. Gaz. 55: 36-55, map. 173 Merriam, C. H. 1898. Life zones and crop zones of the United States. U. S. Dept. Agr., Div. Biol. Surv., Bull. 10. Hosier, J. G. 1903. Climate of Illinois. Bull. 111. Agr. Exper. Station 6: 45-76. Neill, N. P., and Tharp, W. E. 1907. Soil survey of Newton county, Indiana. Field Operations of the Bureau of Soils, 7: 747-779, map ^2. Olsson-Seffer, P. 1909. Relation of soil and vegetation on sandy sea shores. Bot. Gaz. _//: 85-126, /. 7-1-'. Pammel, L. H. 1899. Some ecological notes on the Muscatine flora. Plant World 2: 1 8 1- 1 86. Pepoon, H. S. 1909. An ecological survey of the driftless area of Illinois and Wisconsin. School Sci. and Math, p: 441-446, 522-527. Pound, R., and Clements, F. E. 1898. The vegetation regions of the prairie province. Bot. Gaz. -'5: 381-394, pi- 21. 1900. The phytogeography of Nebraska. Second ed. Lincoln, Neb. Ramaley, F. 1908. Climatology of the mesas near Boulder, Colorado. Univ. of Col. Studies 6: 19-31. Reid, C. 1899. The origin of the British flora. London. Robbins, W. W., and Dodds, G. S. 1908. Distribution of conifers on the mesas. Univ. of Col. Studies 6: 31-36. Rydberg, P. A. 1895. Flora of the sand hills of Nebraska. Contr. U. S. Nat. Herb. ? : 133-203. Sargent, C. S. 1884. Report on the forests of North America. U. S. Tenth Census Report, p : Schimper, A. W. 1903. Plant geography upon a physiological basis. Oxford. Spalding, V. M. iQOf). Problems of local distribution in arid regions. Amer. Nat. 43 : 472-486. 174 Transeau, E. N. 1903. On the geographic distribution and ecological relations of the bog plant societies of northern North America. Bot. Gaz. 36: 401-420, /. 1-3. 1905. Forest centers of eastern America. Amer. Nat. jp : 875- 889, /. 1-6. 1905-06. The bogs and bog floras of the Huron river valley. Bot. Gaz. 40: 351-375, 418-448; 41- 17-42; /. 1-16. Warming, E. 1909. Oecology of plants. Oxford. Whitford, H. N. 1901. The genetic development of the forests of northern Michi- gan; a study in physiographic ecology. Bot. Gaz. 31: 289- 325, /. 1-18. October, 1910, ERRATA Plate III, Fiy. 1, after llie wofd >/i/xt\f in lecrend insert consochs of the. Plate IX, Fig. 2. dele the leg-end and read instead: Root-s:ystem of Tt'phrosia vir^iniana^ exposed by bJowing- of the sand. Plate X, Fig*. 2, dele the leg-end and read instead: A blowout almost stabilized by bunch-yrasscs, especially Lcptoloma cognatum. Plate I. I V Fig-. 1. (General view of the sand-dunes near Havana. The isolated trees are ^iwrrus I't-Zu- tiiia \ at the right is a grove of 'Jnglans nigra. Fig. 2. Lt-ptoluma cogiiaiiim consocics ofjlhe bunch-^-'rass associalioii, OquuwUa area. Large bunches of Andropogon scoparius in the rear, and a flowering- Achillea Millefolium in the center. Plate II. Fiff. 1. Mixed consocies of the bunch-grass association scopariiis most abundant. in tlie Hanover area, Audropogou Fij^. 2. Mixed consocies of the bunch-grass association in the Hanuver area. Various species of grasses and perennials in the foreground, and a society of Ccatwthiis ovattts behind. Plate III. Fiy. 1. Luxuriant development of the iween dunes in the Hanover area. ixed bunch-g-rass association in a depression be- V\^. 2. Typical dt:velopiii»_-iii ..i tin; Punicn in /•scudopiibesceiis association, Hanover area. Platk IV. .•^ Fiff. 1. Typical blowout in the Hanover area, looking- west. In the foreground, the deposit association, with /*. Fi^'. 2. Seedlings of Diodia teres coming up in wagron tracks. Thickets of Rhus and buiiclie of PaniiHui virgatum at the rear. Plate IX. Vi^. 1. Characteristic y-rowlh ut '/'../•/u osm .•i/oi/iiiiua Fig'. 2. A blowout almost completely stabilized by biiiiih-f,>Tasscs;, especially Leptolomu rog- natitm. Plate X. ••b*Ssi»(J*-.'^-' :^^l^^m Fifj-. 1. Extensive tract of blowsatid in the 0-/iotlicra rhomhipetala. the Hanover Fi(?. 2. Blowout in the Oi|uavvUa area. Slabilization is beiriniiin^', as shown b> ihe persist- ence of the (grasses at the base of the windward slope, the conspicuous plants of Z,f.s/C(/(va cap- ttatay patches of moss, and by the thicket of Populus dcltoidcs in the backt'''ou"ti, at the left. Plate XII. Fig. 1. Contact of the Polylricliiim association witli bare blowsand. Fi(f. 2. Contact of tlie Polvtriclium (see foreground) and bunch-grass associations. Plate XIII. Fig-. 1. Pond ".E," in a depression between dunes in the Dixon area, showinpr the zones of veg-etation. Fiif. 2. Invasion of the bunch-grass by the black oak association. HanoTer area. Plate XIV. Fig-. 1. Black oak association, Oquawka area. Typical habitat of .Sy/ithvns Builii. Fig-. 2. Typical opeuintr in the black cak association, Ocjuawka area, with Tepluosia vtrgiu- lana^ Monarda punctata^ Opniitia Rajinestjuii^ etc. Plath XV. 'Fig. 1. Characteristic g^rowth of rie. lion, Wtnnebag^o area. tii//f/7inii near the mary^in of the black i.iak associa- Fitr. 2. Hillside in ilit- \ViDnebat,'o area, sliowing- the trausiliuu iroin the black oak associa- tion on the upland (leftj to the bur oak association in the lowland (right). The shrubbery is chiefly Prunus yirginiana. Plate XVI. Fig. 1. Bur oak association, Winnebapo area. liubiis sp. in the foresrround, at the left'; Ptcris aqitilinti conspicuous under the trees. Fig-. 2. Blaclc oalc association, Oquawka area. A few young viiies of Pscdcra and Cdastrus have appeared, indicating the beginning of the succession to the mi.xed forest. Plate XVII. Fifi-. 1. Natural opeuiag in the black oak association, Winnebago area, occupied by the bunch-fi-rass association. Fiji". 2. Face of the river dune, Hanover area, showinK" the upper and middle slopes, separa- ted by the outcrop of an old soil layer. The thicket association caps the dune in the background. Plate XVIII. Fig-. 1. Windward margin of the thickets on the river dune, Hanover area. The outermost tree at the left is a green ash. A small Physalis heterophylla association in the foreground, at the right. Fig. 2. Dune thickets on the slopt- of the river dime, Haiiovt system of the ash at the left indtcaies the migration of the dune area. The exposed root Platk XTX. Fifi". 1. Associati"us on the river dune, Hanover area. In llie rig-ht foreg-round, tbedeposit association, with a large bunch of Panicum virgatum and abundant Aristida tuberculosa', behind it, the .S'w//Vflr/>// association ; ia the backgronnd the dune thickets, with a dense tangle of lianes. Marg'in of the dune thickets on the windward sid the mit'rati m of the dune is shown in the expos.;d Fi?. 2. -__ The effect of the ....^.„ „ _ _ river floud-plain in the background ;ide of tlie river dune, Havana area. ^ roots. Forests of the Mississippi Plate XX. Fig. 1. Destruction of tlie stabilized river-dune and its mesophjtic veg-etation by river ero- sion, and the reversion of the vegretatioii to the pioneer blowsand associatioa. Oquawka area. Fig. 2. Destruction of tlie mesophyiic vegeta'ion of the river duae by river erosion, showing llie coherent surface-layer of sand, and the sliding masses.